* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Summer climate and heatwaves in Europe
Climate change adaptation wikipedia , lookup
Climate governance wikipedia , lookup
Citizens' Climate Lobby wikipedia , lookup
Fred Singer wikipedia , lookup
Economics of global warming wikipedia , lookup
Climatic Research Unit documents wikipedia , lookup
Urban heat island wikipedia , lookup
Politics of global warming wikipedia , lookup
Climate change in Tuvalu wikipedia , lookup
Media coverage of global warming wikipedia , lookup
Numerical weather prediction wikipedia , lookup
Climate change and agriculture wikipedia , lookup
Scientific opinion on climate change wikipedia , lookup
Climate sensitivity wikipedia , lookup
Global warming wikipedia , lookup
Atmospheric model wikipedia , lookup
Global warming hiatus wikipedia , lookup
Physical impacts of climate change wikipedia , lookup
Effects of global warming wikipedia , lookup
Solar radiation management wikipedia , lookup
Climate change and poverty wikipedia , lookup
Public opinion on global warming wikipedia , lookup
Attribution of recent climate change wikipedia , lookup
Surveys of scientists' views on climate change wikipedia , lookup
Effects of global warming on human health wikipedia , lookup
Climate change in the United States wikipedia , lookup
Climate change in Saskatchewan wikipedia , lookup
Effects of global warming on humans wikipedia , lookup
Climate change feedback wikipedia , lookup
IPCC Fourth Assessment Report wikipedia , lookup
Instrumental temperature record wikipedia , lookup
Summer climate and heatwaves in Europe Annemiek Stegehuis To cite this version: Annemiek Stegehuis. Summer climate and heatwaves in Europe. Biodiversity and Ecology. Université Paris-Saclay, 2016. English. <NNT : 2016SACLV052>. <tel-01480310> HAL Id: tel-01480310 https://tel.archives-ouvertes.fr/tel-01480310 Submitted on 1 Mar 2017 HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. NTT:2016SACLV052 T HÈSE DE D OCTORAT DE L ’U NIVERSITÉ PARIS -S ACLAY PRÉPARÉE À L ’U NIVERSITÉ V ERSAILLES S AINT-Q UENTIN EN Y VELINES E COLE D OCTORALE N O. 129 S CIENCES DE L ’ ENVIRONNEMENT D ’I LE - DE -F RANCE M ÉTÉOROLOGIE , OCÉANOGRAPHIE ET PHYSIQUE DE L ’ ENVIRONNEMENT PAR A NNEMIEK I RENE S TEGEHUIS S UMMER CLIMATE AND HEATWAVES IN E UROPE T HÈSE PRÉSENTÉE ET SOUTENUE AU LSCE, LE 7 JUILLET 2016 PhD director : PhD co-director : PhD co-director : President of the jury : Reporters : Examiner : Robert Vautard Philippe Ciais Adriaan J. Teuling Directeur de recherche (LSCE/CNRS) Philippe Bousquet Bart J. J. M. van den Hurk Jesus Fernandez Wolfgang Cramer Professeur (UVSQ/IPSL) Directeur de recherche (LSCE/CEA) Assistant Professor (Wageningen University) Directeur de recherche (KNMI) Assistant Professor (Universidad de Cantabria) Professeur / Directeur de recherche (CNRS) Acknowledgments Merci pour toutes ces belles années ! i Contents Acknowledgments i Contents iii Abstract vii Introduction 1 I 5 6 6 8 8 8 II General introduction 1 European summer climate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Oceanic influence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Large scale circulation . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.1 North Atlantic Oscillation . . . . . . . . . . . . . . . . . . . . 1.2.2 Principal summer weather regimes . . . . . . . . . . . . . . . 1.2.3 Weather regimes associated with warm summer temperature anomalies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Land-atmosphere interactions influencing the summer climate . . . . . 1.3.1 Soil moisture-temperature feedback . . . . . . . . . . . . . . 1.3.2 Soil moisture-precipitation feedback . . . . . . . . . . . . . . 1.3.3 Regions for strong land-atmosphere feedbacks . . . . . . . . 1.3.4 Influence of vegetation . . . . . . . . . . . . . . . . . . . . . 2 Heatwaves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Characteristics of the 2003 heatwave . . . . . . . . . . . . . . . . . . . 2.2 Mechanisms behind the 2003 heatwave . . . . . . . . . . . . . . . . . 2.3 Consequences of the 2003 heatwave . . . . . . . . . . . . . . . . . . . 3 Impact of elevated temperatures and drought on vegetation . . . . . . . . . . 10 10 11 12 13 14 16 16 16 18 18 Uncertainties in the future climate 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 22 iii iv CONTENTS 2 3 4 5 The European climate under a 2◦ C global warming . . . . . . . . . . . . . . . 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.1 Temperature and precipitation change . . . . . . . . . . . . . 2.3.2 Land heat fluxes and radiation . . . . . . . . . . . . . . . . . 2.4 Summary and conclusion . . . . . . . . . . . . . . . . . . . . . . . . . Future European temperature change uncertainties reduced by using land heat flux observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion and additional results . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 Model ensemble . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Uncertainty reduction method . . . . . . . . . . . . . . . . . . . . . . . 4.3 The choice for the 2◦ C limit . . . . . . . . . . . . . . . . . . . . . . . . Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III Atmospheric processes 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 2 WRF model description . . . . . . . . . . . . . . . . . . 2.1 Microphysics . . . . . . . . . . . . . . . . . . . . 2.2 Planetary Boundary Layer . . . . . . . . . . . . . 2.3 Surface Layer . . . . . . . . . . . . . . . . . . . . 2.4 Radiation . . . . . . . . . . . . . . . . . . . . . . 2.5 Convection . . . . . . . . . . . . . . . . . . . . . 2.6 Land surface . . . . . . . . . . . . . . . . . . . . 3 An observation-constrained multi-physics WRF ensemble pean mega heat waves . . . . . . . . . . . . . . . . . . . 4 Discussion and additional results . . . . . . . . . . . . . 4.1 The importance of the convection scheme . . . . 4.2 What triggers convection? . . . . . . . . . . . . . 5 Summary and conclusions . . . . . . . . . . . . . . . . . IV V . . . . . . . . . . . . . . . . for . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . simulating Euro. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 24 24 24 24 26 27 28 33 33 33 35 35 39 40 41 41 42 43 43 43 44 44 68 68 70 71 The influence of soil moisture on summer temperatures 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Model description . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Control simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Attribution of temperature anomalies . . . . . . . . . . . . . . . . . . 3 Summer warming induced by early summer soil moisture changes in Europe 3.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Main text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Discussion and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 74 75 75 75 75 77 77 77 82 Drought impacts on European forest species 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Drought and heat stress on 5 main European forest tree species 2.1 Picea sp. . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Pinus sylvestris . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 87 88 88 91 . . . . . . . . . . . . . . . . . . . . . . . . . . . . CONTENTS . . . . . . . . . . . . . . . . . . . . . . . 93 95 96 99 99 100 100 100 100 100 101 101 102 105 106 107 109 110 111 112 112 112 114 Discussion & concluding remarks 1 Climate models in general . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Regarding the future climate . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Impact studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 118 119 121 3 4 5 6 VI 2.3 Fagus sylvatica . . . . . . . . . . . . . . . . . . . 2.4 Quercus robur & petraea . . . . . . . . . . . . . . 2.5 Quercus ilex . . . . . . . . . . . . . . . . . . . . . Model description . . . . . . . . . . . . . . . . . . . . . . 3.1 ORCHIDEE . . . . . . . . . . . . . . . . . . . . . 3.2 ORCHIDEE-CAN . . . . . . . . . . . . . . . . . . 3.2.1 From PFT to species . . . . . . . . . . . 3.2.2 Carbon allocation and canopy structure 3.2.3 Hydraulic architecture . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 Puéchabon . . . . . . . . . . . . . . . . . . . . . 4.1.1 GPP . . . . . . . . . . . . . . . . . . . . 4.1.2 Soil hydraulics . . . . . . . . . . . . . . 4.1.3 Transpiration . . . . . . . . . . . . . . . 4.2 Other sites . . . . . . . . . . . . . . . . . . . . . 4.2.1 Brasschaat . . . . . . . . . . . . . . . . 4.2.2 Tharandt . . . . . . . . . . . . . . . . . 4.2.3 Hesse & Collelongo . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1 Site-specific characteristics . . . . . . . . . . . . . 5.2 Plant hydraulics . . . . . . . . . . . . . . . . . . 5.3 Other processes . . . . . . . . . . . . . . . . . . . Summary and concluding remarks . . . . . . . . . . . . Bibliography Annexes A v The European climate under a 2dC global warming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 153 155 Abstract vii Titre: Le climat estival et des vagues de chaleur en Europe Mots clés: Vagues de chaleur en Europe, Modélisation la méthode, Forets en Europe, Sécheresse Résumé: L'objectif de ce travail de thèse est de contribuer à la compréhension des rôles joués par les interactions sol-atmosphère et par la circulation à grande-échelle dans la formation d’anomalies de températures estivales en Europe. Cela constitue un double défi du fait de la rareté des observations d’une part et des incertitudes liées aux paramétrisations des modèles atmosphériques et de leurs schémas de surface d’autre part. Ce travail est centré sur 4 sujets principaux: 1) Comment les interactions sur les processus sol-atmosphère liés à l’eau du sol disponible pour les plantes et aux flux d’énergie échangés entre la surface et l’atmosphère influencent-elles les projections climatiques de température d’été et leurs incertitudes dans le climat européen? 2) Comment les paramétrisations atmosphériques influencent les simulations des vagues de chaleur extrêmes en Europe? 3) Quelle importance joue l’humidité du sol comparée aux circulations atmosphériques sur les anomalies de température estivales? Et 4) Quels sont les impacts des sécheresses et de la chaleur sur la végétation? 1) Comment les interactions sol-atmosphère influencent-elles les projections climatiques et leurs incertitudes? Il existe des incertitudes considérables dans la simulation du climat actuel en Europe, qui peuvent se propager dans la simulation du climat futur et de sa sensibilité à l’augmentation des forçages anthropiques. Cela soulève plusieurs questions, dont deux sont traitées dans ce chapitre, présentés dans deux parties différentes. Premièrement je discute des incertitudes dans les projections climatiques Européennes pour un échauffement global moyen de 2˚C. Je présente les incertitudes dans le climat futur liée pour différents horizons temporels. Un accent particulier est mis sur les températures estivales et les flux d’énergie échangés entre la surface et l’atmosphère. Cette étude fait partie du projet Européen IMPACT2C et les résultats sont publiés par Vautard et al. (2014). Je ne montre que les résultats du travail personnel accompli dans cette étude. L’article complet se trouve dans l’Annexe A. Deuxièmement j’examine le lien heuristique entre les incertitudes des simulations des flux d’énergie échangés entre la surface et l’atmosphère pour l’actuel et les incertitudes de température dans un climat futur. Les rétroactions entre surface et atmosphère jouent un rôle important dans la genèse de vagues de chaleur et de sécheresses en l’été dans le climat tempéré en Europe. Cela peut également être analysé dans un ensemble de sorties de modèles où les rétroactions plus fortes donnent lieu à des températures plus élevées. Stegehuis et al. (2013b) ont montré que, en utilisant un ensemble de modèles, les flux de chaleur sensible les plus importants sont associés à des températures estivales plus élevées (Fig. II.I). Ensuite, nous avons mis en évidence de grandes différences dans la répartition des flux d’énergie échangés entre la surface et l’atmosphère entre différents modèles, ce qui conduit à des incertitudes dans les simulations pour le climat actuel. Cependant, comment ces différences se traduisent en différences pour le climat futur reste mal connu. Les incertitudes dans des simulations du climat sont mieux représentées dans des ensembles de modèles que dans chaque modèle pris individuellement. Dirmeyer et al. (2006) ont montré que rétroactions entre terre et atmosphère sont relativement mal représentées dans chaque modèle, alors que la moyenne de l’ensemble montre de meilleures performances. Ceci est en partie dû à la rareté des observations spatialement explicites à long terme pour certaines des variables importantes dans les rétroactions entre surface et atmosphère, tels que les flux de chaleur sensible et latente. Récemment, certains produits satellitaires sont devenus disponibles comme GLEAM (Miralles et al. 2011). Un autre produit, basé sur une combinaison de mesures de télédétection et des données de tours à flux du réseau FLUXNET est le ‘Model Tree Ensemble’ de Jung et al. (2010), utilisé dans l’étude présentée au Chapitre 2. Cependant, en utilisant ce type de produit dérivé des observations, on doit toujours garder à l’esprit que les données ont également des incertitudes systématiques. Mueller et al. (2011) ont montré que la dispersion des produits de flux de chaleur latente dérivés d’observations est comparable à celle des flux simulés avec des ensembles de modèles de surface. En raison de la meilleure représentation des ensembles de modèles par rapport à celle de modèles individuels, plusieurs ensembles de modèles, régionaux et globaux, ont été réalisés par différents groupes de recherche. Un exemple d’un ensemble avec des modèles climatiques régionaux (RCM) européen est celui du projet FP6-ENSEMBLES (Hewitt et al., 2004; Van der Linden et al., 2009). Dans cette étude j’ai utilisé 15 RCMs, conduits par 6 modèles climatiques globaux, du 4ème rapport du GIEC, pour le scénario A1B. En ce qui concerne la première question posée, nous avons d’abord constaté que les expériences de simulation européennes ont un réchauffement moyen qui est supérieur à la moyenne mondiale, dépassant 2°C au niveau régional. Le réchauffement est le plus grand en Scandinavie, la région méditerranéenne, la péninsule ibérique et les Alpes. Les iles britanniques subissent un réchauffement plus faible. Les flux de chaleur sensible plus élevés confirment ce plus fort réchauffement dans le sud, associé à une diminution des flux de chaleur latente et une augmentation des flux de chaleur sensible dans les régions du sud pendant l’été. Dans la région où le régime d’évapotranspiration est limité par l’énergie (Seneviratne et al., 2010), Scandinavie, les températures plus élevées conduisent à l’augmentation du flux de chaleur latente pendant l’été, alors que le flux de chaleur sensible est diminué. Cela peut être dû à la formation des nuages qui diminue le rayonnement solaire et améliore le rayonnement des ondes longues. En ce qui concerne la deuxième question posée, utilisant des observations de flux de chaleur sensible de Jung et al. (2009), nous avons pu réduire l’écart inter-modèles des RCMs sur la prédiction du changement de température d’été en Europe jusqu’à 40%. Les réductions d’incertitude les plus importantes ont été trouvé sur la France et les Balkans, des régions déjà mentionnées pour leur importance des rétroactions entre terre et atmosphère (e.g. Seneviratne et al., 2006a). Dans le sud et le nord de l’Europe, d’autres variables observées pourraient être utilisées pour réduire l’incertitude dans les projections de température, par exemple les rétroactions entre les précipitations et la température (Quesada et al., 2012) ou la rétroaction neige-albédo-température (Hall & Qu, 2006). grand écart entre les modèles. La source d’une telle surestimation n’est pas évidente. Parmi les différents modèles, le choix d’un jeu de paramétrisations des processus physiques conduit à des résultats différents (Bellprat et al., 2012). Une autre source d’incertitude dans les RCMs peut aussi découler des forçages. En outre, la dynamique créée par les modèles, pourrait être une source de différenciation par la variabilité naturelle chaotique, provoqué par l’instabilité de l’atmosphère (Lorenz, 1963). 2) Dans le chapitre 2 l’accent était mis sur les rétroactions entre surface et atmosphère dans des simulations de RCMs et leur influence sur les températures estivales futures. Ce chapitre se concentre sur les processus atmosphériques (physique et microphysique des nuages, convection, transfert radiatif) dans un modèle RCM très largement utilisé (WRF) et leurs impacts sur les simulations des températures estivales pendant les vagues de chaleur en Europe. L’objectif principal est d’essayer de répondre à la question de savoir si et comment différentes formulations physiques différentes disponibles pour le modèle WRF influencent la simulation des vagues de chaleurs et l’incertitude des températures d’été en Europe. Dans une étude sur les phénomènes climatiques extrêmes, Good et al. (2006) ont constaté que les projections des modèles sont nécessairement limitées par des incertitudes, en particulier en raison de la paramétrisation des processus physiques dans les modèles atmosphériques. Des ensembles ‘multi-physiques’ ont été construits pour estimer cette incertitude (Awan et al., 2011). Avec cette approche, un seul modèle est utilisé, mais des physiques différentes sont testées pour trouver les combinaisons les plus réalistes. Bien que cette méthode ne soit pas vraiment nouvelle (Fernandez et al., 2007; Awan et al., 2011 ; Evans et al., 2012, Garcia-Diez et al., 2013 ; Mooney et al., 2013), elle n’a pas encore été utilisée pour analyser les vagues de chaleur. Pourtant, cela peut être d’une grande importance en ce qui concerne leurs conséquences et leur fréquence et gravité plus élevées attendu dans le climat futur de l’Europe. Il n’est pas évident à priori que les RCMs peuvent représenter les vagues de chaleur extrêmes dans les régions où les observations sont rares (Perkins et al., 2013). Même si la physique était exactement connue, des incertitudes sont liées au manque de données, par exemple de réseaux d’observations de stations météorologiques. peuvent conduire à une sousestimation des températures maximales (Halenka et al., 2006), soit trop peu de vagues de chaleurs simulées. Cependant, Vautard et al. (2013) ont constaté que la fréquence et l’amplitude des vagues de chaleur sont souvent surestimées dans les RCMs en Europe, en particulier dans la région méditerranéenne. Toutefois ils ont montré un Dans le chapitre 3 j’ai utilisé le modèle ‘Weather and Research Forecast’ (WRF) (Skamarock et al., 2008) et j’ai testé 216 combinaisons différentes de la paramétrisation des processus physiques atmosphériques. Ils se composent de 3 physiques de rayonnement, 4 physiques de convection, 3 microphysiques et 6 physiques de couche limite planétaire. J’ai simulé les vagues de chaleur de 2003 et 2010, et l’été plus humide de 2007. Pour mettre l’accent sur la physique seulement, j’ai guidé les simulations au-dessus de la couche limite atmosphérique avec des observations. Bien que les interactions surfaceatmosphère jouent également un rôle important dans les canicules d’été (voir Chapitre 2), seules les physiques de l’atmosphère ont été testées en raison du choix limité entre différents schémas de surface dans le modèle WRF au moment de l’étude. Les 216 simulations ont été évaluées par rapport à différents jeux d’observations. J’ai trouvé un fort écart de la température, la précipitation et le rayonnement simulé avec les observations, même si les simulations ont été guidées. Les différences de température entre modèle et observations étaient localement jusqu’à 10˚C, avec les plus grandes incertitudes en Europe centrale, probablement en raison de représentations différentes des interactions surface-atmosphère (Stegehuis et al., 2013a). Les températures moyennes et maximales ont été systématiquement sous-estimées quelle que soit la paramétrisation des processus physiques atmosphériques utilisée. Les précipitations et le rayonnement solaire incident ont été, la plupart de temps, surestimés. J’ai montré que la convection domine la dispersion de l’ensemble de simulations de WRF avec différentes physiques, probablement en induisant des incertitudes dans les nuages que affectent à la fois l’énergie de surface et le budget de l’eau avant et pendant les vagues de chaleur. Un petit ensemble a été sélectionné, qui produit le meilleur accord avec les données d’observation. Nous suggérons que cet ensemble de configurations pourrait être utilisé pour d’autres recherches sur les vagues de chaleur et les températures estivales en Europe avec le modèle WRF. 3) Les résultats des chapitres 2 et 3 ont permis de montrer que les températures estivales en Europe sont contrôlées par les interactions surfaceatmosphère et par la circulation de l’atmosphère (physique). Pour les vagues de chaleur, la circulation atmosphérique doit favoriser le transport de l’air chaud vers le continent. Le sol doit favoriser une rétroaction entre la baisse de l’humidité du sol et l’augmentation des températures, à travers la dissipation d’énergie par les flux de chaleur sensible. Bien que des études antérieures ont établi l’importance de ces deux processus (Black et al., 2004 ; Miralles et al., 2014), elles n’ont pas cherché à séparer quantitativement leur contribution individuelle a des anomalies de température d’été pour les vagues de chaleur. Dans cette étude, j’ai choisi d’étudier l’effet de humidité du sol au début de l’été seulement sur la production d’anomalies de température en été, car l’humidité du sol plus tard dans la saison estivale est aussi dépendent de la circulation atmosphérique, donc moins formellement séparable. L’augmentation de la température prédite en Europe à la fin de ce siècle est estimée entre 2˚C et 5˚C, selon le scénario et la région (van der Linden & Mitchell, 2009 ; Boberg & Christensen, 2012 ; Stegehuis et al., 2013a). Cette augmentation devrait être plus forte en Europe du Sud (up to 5˚C), que dans la partie nord (Fig. 3, Stegehuis et al., 2013a). En outre, une plus grande variabilité des températures est attendue, notamment en Europe centrale. Cette augmentation globale de la température sur le continent Européen pourrait conduire à plus de dessèchement du sol et plus de variabilité, créant des situations favorables pour des rétroactions positives entre le sol et la température. Une compréhension plus précise, et la quantification du rôle de l’humidité du sol et de ses tendances au cours des dernières années, pourraient donc être critiques dans la compréhension de la variabilité des températures estivales et événements canicule dans le climat futur. Les régimes de circulation atmosphérique associés au développement des canicules sont les deux régimes de Blocage Atmosphérique et Atlantic Low (Cassou et al., 2005). L’apparition des différents régimes de circulation ont changé au cours du siècle dernier (Horton et al., 2015 ; Coumou et al., 2015 ; Alvarez-Castro et al., 2016). Bien que l’impact de ces régimes aient été étudiés (van Haren et al., 2015), il y a encore beaucoup d’incertitude quant à savoir si et comment ils peuvent changer dans le climat futur (Cattiaux et al., 2012). Néanmoins, une meilleure compréhension de leur contribution aux vagues de chaleur en été est utile pour prédire le changement climatique. Dans le chapitre 3, j’avais sélectionné un nombre limité de configurations de paramétrisations des processus physiques atmosphériques dans le RCM WRF qui sont capables de simuler les vagues de chaleur observées de 2003 et 2010, ainsi que l’année d’été relativement humide de 2007. Bien que le chapitre 4 ne traite pas uniquement des vagues de chaleur, ces configurations seront choisies pour créer un ensemble d’étés basés sur différentes initialisations de l’humidité du sol au début de l’été et différents régimes atmosphériques. Toutes ces conditions initiales sont réalistes, comme les valeurs choisies sont celles des années d’été modélisées de 1980-2011. Nous avons constaté que la contribution de l’humidité du sol au début de l’été dans les vagues de chaleur de 2003 et 2010 a été importante et a contribué à environ 1˚C au maximum. Cependant, le conditionnement par le régime de circulation atmosphérique (facteur de grande échelle) explique statistiquement la plus grande partie des anomalies de température, avec une contribution maximum de 3˚C en 2003 et jusqu’à 6˚C en 2010. Un résultat intéressant est que la contribution de l’humidité du sol initiale a augmenté en importance au cours des trois dernières décennies en Europe centrale, en France et dans certaines parties de la Russie, alors que celle des régimes de temps est restée à peu près stable. Ces régions coïncident avec les régions avec une tendance négative significative de l’humidité du sol dans les simulations du modèle WRF. L’influence des régimes de temps sur les anomalies de température d’été n’a augmenté au cours des trois dernières décennies qu’en l’Europe de l’Est. 4) Dans les chapitres précédents l’accent a été mis sur le climat de l’été et les vagues de chaleur en Europe, en analysant les processus atmosphériques et les rétroactions surfaceatmosphère. Un aspect important dans les rétroactions liées aux processus de surface est la végétation, qui peut soit avoir un effet modérateur ou un effet d’amplification sur les anomalies de température régionales. Ceci dépend en grande partie du type de végétation et sur la disponibilité de l’humidité du sol qui influence l’état de la végétation couplé à l’état de l’atmosphère par les flux de chaleur sensible et latente. L’impact des étés chauds et des vagues de chaleur sur l’état hydrique et le fonctionnement de la végétation est donc un aspect important dans la compréhension du climat, jusqu’ici peu ou pas abordé dans les RCMs qui ont des schémas de surface assez simples (pas de phénologie, pas de physiologie, pas de mortalité ou dépérissement de la végétation). Les forêts peuvent émettre plus de chaleur latente dans l’atmosphère, et plus longtemps durant les vagues de chaleur, en raison de la profondeur de l’enracinement, et donc un meilleur accès à l’eau souterraine, conduisant à plus d’évapotranspiration que les prairies pendant le développement des vagues de chaleur. L’albédo plus faible des forêts peut en revanche provoquer un effet de réchauffement radiatif (Naudts et al., 2016). L’effet d’albédo semble dominer l’effet de l’évaporation pendant le début des vagues de chaleur (Teuling et al., 2010). Cet équilibre peut cependant être inversé pendant le développement ultérieur des vagues de chaleur ou lorsque la plupart des arbres ferment leurs stomates à cause du stress hydrique. Pour séparer ces effets, les connaissances sur la réaction de la végétation au cours du stress de l’eau est cruciale. Pour les effets du climat sur la végétation sur des plus grandes échelles, les modèles de végétation peuvent être utilisés couplés à des modèles atmosphériques. La plupart de ces modèles ne distinguent pas entre des espèces différentes, mais utilisent des ‘groupes fonctionnels’ (PFT). Cependant, une nouvelle version d’ORCHIDEE (Krinner et al., 2005) inclut les caractéristiques des espèces de la plupart des essences d’arbre européennes. Cette version, ORCHIDEE-CAN (Naudts et al., 2015), a en outre une description explicite des potentiels d’eau dans le continuum sol-plante-atmosphère, que peut mieux représenter en principe le stress hydrique des plantes pendant les périodes chaudes et sèches. Dans le chapitre 5, je présente d’abord une revue de littérature sur les effets de la sécheresse et de la chaleur sur le stress hydrique de la végétation. Cette revue est concentrée sur cinq espèces d’arbres en Europe. En outre, je présente des résultats préliminaires de l’évaluation du modèle de la végétation ORCHIDEE-CAN pour simuler la baisse de transpiration et de photosynthèse du couvert végétal pendant des épisodes de sécheresse. Les données d’évaluation sont celles de tours à flux (Baldocchi et al., 1996) pour 5 sites en Europe. Le site sur lequel j’ai essayé de calibrer le modèle ORCHIDEE-CAN pour Quercus ilex (chêne vert) est celui de Puéchabon, soumis à un fort stress hydrique du sol au printemps et en été. J’ai montré qu’ORCHIDEECAN ne permet pas de simuler la baisse de photosynthèse (GPP) et de transpiration de ce site. La GPP de la végétation est restée trop élevée pendant l’été, probablement causée par des niveaux d’humidité du sol trop élevé. Par contre, pour les espèces tempérées Fagus sylvatica, Picea sp., Pinus sylvestris et Quercus robur et petraea la simulation de GPP a été beaucoup mieux représenté, bien que ces espèces n’aient pas subi beaucoup de stress hydriques pendant les années observées. Certains processus à l’échelle de l’arbre ou du couvert ne sont pas représentés dans le modèle de la végétation, ce qui peut entrainer des différences entre la production primaire modélisé et observé. Pour simuler l’impact du stress hydrique sur la végétation de l’augmentation prévue de la fréquence et de la gravité des vagues de chaleur, certaines adaptations du modèle que ORCHIDEE-CAN doivent être apportées. Tout d’abord, les processus spécifiques qui agissent le plus fortement en période de sécheresse doivent être ajoutés. Un exemple est la résistance entre le sol et la racine. Outre ces processus spécifiques à la sécheresse, il serait également nécessaire d’introduire des effets décalés. La perte de conductivité (hydraulic failure) n’est pas reproduite par ORCHIDEE-CAN pour le moment. En réalité, la mortalité des arbres peut se produire même plusieurs années après un événement de sécheresse extrême. Une solution relativement simple serait d’ajouter une forme de mortalité lorsque l’embolie devient trop importante, si ce processus peut être reproduit dans le modèle utilisé. Le temps n’a pas permis de déboucher sur un résultat finalisé, c’est à dire une version satisfaisante d’ORCHIDEE-CAN capable de reproduire la baisse de potentiel hydrique du sol, et celle des différents organes de la plante. Title: Summer climate and heatwaves in Europe Keywords: Heatwaves, Europe, Modeling, Forest, Summer climate Abstract: Through this work I aimed to improve the understanding of the role of land-atmosphere feedbacks and large-scale circulation that lead to warm summer temperatures in Europe. This is challenging due to the scarcity of observations and the uncertainties of parameterized atmospheric processes. I focused on four main issues: 1) How do land-atmosphere feedbacks affect climate projections and their uncertainties? 2) How do different physical parameterizations affect the simulation of extreme heatwaves? 3) How large are the roles of soil moisture and atmospheric circulation in the development of European summer temperature anomalies? And 4) What are the impacts of heat and drought stress on vegetation? Regarding the first question I found that the different partitioning of land heat fluxes between models leads to spatially different warming over Europe in the future. The uncertainty of future climate change was especially high in central Europe, largely due to the uncertainty in heat flux partitioning, while in Southern Europe the models mostly agreed. The use of observationbased sensible heat fluxes allowed to reduce this climate change uncertainty regionally up to 40%. While studying different atmospheric parameterizations for the extreme heatwaves of 2003 and 2010, I found a large temperature spread between the simulations. Compared to observations, temperature was mostly underestimated. Shortwave radiation and precipitation were generally overestimated. I selected a reduced model ensemble of well performing configurations compared to observations, to perform future studies on warm summer temperatures over Europe. The best physics configuration was consequently used to quantify the role of early summer soil moisture and large-scale drivers on summer temperature anomalies. The contribution of soil moisture was up to maximum 1ºC during the heatwaves of 2003 and 2010. The contribution of large-scale drivers was larger, reaching up to 3ºC in 2003 and up to 6ºC in 2010. However, the contribution of early summer soil moisture to the temperature anomalies has been increasing over the last decades over parts of central Europe and Russia, corresponding to the regions with a significant negative trend of soil moisture. Largescale drivers showed an increasing importance in the Eastern European region. Lastly, I studied the impacts of drought and heat stress on several European forest tree species. I found an overestimation of modeled GPP at a local scale in the Mediterranean region during summer with ORCHIDEE. This indicates that the vegetation model does not well reproduce the complicated consequences of drought stress. To model future, possibly more severe impacts of drought, the model may need to be adapted with drought-specific processes and lagged effects. Introduction The European summer climate is characterized by a large variability, both in space and time. Although our understanding of this climate variability has been increasing, some important questions are still unanswered. One example is the question of whether European summer conditions can be predicted in advance, over time periods of weeks or even decades. Such a prediction could bring enormous benefits in many different sectors, but requires detailed and precise knowledge on underlying mechanisms. Local observations are adequate for understanding processes at local scales. However, some of the processes and feedbacks occur at larger regional or continental scales for which models may be a more suitable tool. The first numerical climate models only calculated the large-scale circulation patterns on a global scale. Nowadays climate models exist on both regional and global scale and contain numerous routines and parameterizations to describe the current state of the land and the atmosphere. Their increasing complexity and resolution are advantageous in that more processes can be described, but the uncertainty of these processes might lead to error compensations, which may lead to ’the right temperature for the wrong reason’. Such model deficiencies often become apparent when looking at interannual variability. This is problematic in current-climate simulations, but might be even more of an issue for the simulation of the future climate, as interannual variability is expected to increase (Schär et al., 2004). The last decades have shown extreme examples out of the historical range of variability, in the form of extreme warm and dry summers such as in 2003 in Western Europe and in 2010 in Russia (Luterbacher et al., 2004; Barriopedro et al., 2011). While warm weather does bring advantages, these events were characterized by high financial costs, increased mortality rates and decreased ecosystem productivity (Ciais et al., 2005; García-Herrera et al., 2010). The importance of the ability of models to correctly simulate interannual variability and extreme events such as heatwaves has not been recognized for some time, and there was more focus on obtaining the right climatology. Now both our knowledge on climatic processes and computing capacity have increased, more effort has been invested in the representation of the above mentioned issues, especially after the two heatwave episodes. 1 Two processes central for European heatwaves are large-scale atmospheric circulation and land-atmosphere coupling on a local to regional scale. Warm dry air imported from Eastern or Southern Europe has the potential to reduce precipitation and increase clear sky conditions. This increases evaporative demand resulting in enhanced soil drying. Reduced soil moisture alters the partitioning of land heat fluxes and favors sensible over latent heat flux which can result in a positive feedback loop promoting high temperatures. However, large-scale observations of the most important variables of these processes, such as soil moisture and land heat fluxes, are scarce. This limits their precise understanding and thus the simulation of heatwave events. Through this work, I aim to improve the understanding of the role of land-atmosphere feedbacks and large-scale circulation that lead to warm summer temperatures in Europe. This is challenging due to the scarcity of observations and the uncertainties of parameterized atmospheric processes. I will first focus on some important variables of the energy budget and land-atmosphere feedbacks, and study how these, and their uncertainties, affect climate projections. A second step is to evaluate different models to simulate the extreme heatwaves we have witnessed in 2003 in Europe and 2010 in Russia. I will pay special attention to different atmospheric physics to find the uncertainty that physical formulations can induce. After the evaluation of the models, I will study the specific mechanisms leading to higher summer temperatures, with a direct focus on soil moisture legacy effects and atmospheric circulation patterns and try to separate their importance. An important part of land-atmosphere feedbacks during summer is the vegetation covering the land surface. It can either have an amplifying or a moderating effect on the climate, depending largely on the availability of soil moisture which influences the status of the vegetation, and the vegetation type. Forests were found to have an amplifying effect on high temperatures at the beginning of a heatwave, while grasslands have a moderating effect (Teuling et al., 2010). The lower albedo and evapotranspiration rates of forest were responsible for these differences. On longer time scales however, this effect may be reversed due to the higher capacity of trees to store water and extract it from the soil. To better understand the influence of vegetation on climate, we also need to study how climate influences vegetation. Here too, models are a very suitable tool for investigating the impacts of climate on vegetation as the processes involved can act over large scales. A large uncertainty from the modeling of forests is upscaling of processes from single trees and leaves to whole forest ecosystems. In addition, not all processes on leaf level might be known, inducing more uncertainty. A start to answer the full question about the impact of future heatwaves is the evaluation of vegetation models on the simulation of current summer stress. I will use a recently developed vegetation model with a specific description of water potentials in the soil-plant-atmosphere continuum, and asses its capability to simulate summer stress at different sites in Europe. This model could then be used to upscale to the whole European region. Besides this first/limited model assessment, I will discuss different drought-affected processes on tree-level described in the literature. To summarize, the research questions addressed in this work are: 1) How do the important land-atmosphere processes in regional climate models lead to different summer temperatures? How do their uncertainties affect future summer temperatures predictions? And how can present-day observations be used to reduce a part of the 2 CONTENTS 3 uncertainty in these temperature projections? 2) Can current regional climate models correctly simulate heatwaves in Europe? How do different atmospheric parameterizations influence the simulation of heatwaves? And can we design a model ensemble that can correctly simulate European heatwaves? 3) How large is the role of atmospheric circulation and soil moisture feedbacks in the development of European summer temperatures? Can we quantify and separate their importance? And is the importance of land-atmosphere feedbacks becoming more important? 4) Can a state of the art vegetation model correctly simulate summer stress in forests? What are some of the missing processes? And can the model simulate extreme stresses as have been experienced in 2003, with a focus on green oak ecosystems in the Mediterranean region? C HAPTER I General introduction 5 6 1 CHAPTER I. GENERAL INTRODUCTION European summer climate Europe can be separated in broadly four climatic zones: a boreal climate in the north, a Mediterranean climate in the south, a continental climate in the central east and a maritime or oceanic climate central west. In the Alps and the Pyrenees some regions have an alpine climate. These different climates are characterized by different annual and seasonal temperature and precipitation values. The Mediterranean climate has hot and dry summers but milder and wet winters (Fig. I.1). The former is partially due to the sinking air from the Hadley cells, whilst the latter is partially caused by the Mediterranean Sea that moderates the temperatures during the winter and storms that form over the Atlantic Ocean. The maritime or oceanic temperate climate in Europe is characterized by warm summers and mild winters (Fig. I.1). The annual temperature range is relatively small as it is moderated by the relative warm water from the ocean North Atlantic drift. Precipitation occurs the whole year round. The continental climate of Eastern Europe has warm to hot summers and cold winters (Fig. I.1). Most winter precipitation falls in the form of snow. The attenuating effect of the Atlantic Ocean does not apply in this part of Europe. The boreal climate is characterized by long and cold winters while the summers are short and cool. Winter temperatures can drop to -40◦ C. The average summer temperature is approximately 10◦ C (Fig. I.1). Precipitation is usually low. To better understand model processes and the different drivers for interannual summer variability, it is important to identify these different climatic zones and to understand what drives the differences between the varying temperatures. In the next sections I will discuss the influences of the Atlantic Ocean and large-scale circulation on the European summer climate. This is followed by a section on land-atmosphere feedbacks and their importance on the summer climate. Heatwaves will be discussed in section 1.2, and the last part is a concise description on the influence of summer climate on vegetation. Besides the influences of the ocean and land-atmosphere feedbacks, the land itself can also play a role in the local or regional climate. For example, at high altitudes in mountainous areas the climate is colder even in low latitudes. Snow albedo feedbacks can play a role here. Furthermore, a difference between a north- or south-facing slope can cause a significant difference in local temperatures. Orography also has an influence on precipitation. As air is pushed up against a mountain slope, it loses heat with altitude and might precipitate mostly on one side of the mountain, while the other side remains much drier throughout the year. This however, will not be discussed in further detail. 1.1 Oceanic influence The ocean generally has a moderating effect on the European climate through the northward transport of warm tropical water by the thermohaline circulation and the Gulf Stream. Besides these currents, the Atlantic Multidecadal Oscillation (AMO) also exerts an influence. This is a mode of variability of sea surface temperatures (SST) in the North Atlantic, which is detrended for climate change. The AMO is characterized by cold and warm phases, with a cool phase from 1905 to 1925 and from 1965 to 1990 and a warm phase between 19311960 (Sutton and Hodson, 2005). Besides its effects on the winter weather, Sutton and Hodson (2005) have shown its influence on Europe’s summer climate. They found that during the warm phases, a relatively low pressure area is formed west of the British Isles. This 1. EUROPEAN SUMMER CLIMATE 7 Winter (a) Summer (b) C 25 20 15 10 5 0 −5 Winter (a) Summer (b) mm 5 4 3 2 1 0 Figure I.1: European winter and summer climatology. a) DJF temperature; b) JJA temperature; c) DJF precipitation; and d) JJA precipitation. Data is from E-OBS (Haylock et al., 2008). 8 CHAPTER I. GENERAL INTRODUCTION is linked with reduced precipitation over Western Europe and increased temperatures over central Europe (Sutton and Hodson, 2005). Confirming these result, Mariotti and Dell’Aquila (2012) find a positive correlation between the AMO index and air temperature patterns that is, besides over Western Europe, extending over the Mediterranean during summer. However, these temperature differences might also be related to dimming and brightening as described by Wild et al. (2005). 1.2 1.2.1 Large scale circulation North Atlantic Oscillation The variability of the European climate is strongly influenced by large scale atmospheric patterns. The main mode of variability over Europe and the North Atlantic Ocean is the North Atlantic Oscillation (NAO). This index is based on the average difference in surface sea-level pressure between the Icelandic low and the Subtropical (Azores) High, describing the largest fluctuations of the North Atlantic flow. It has a large influence on Europe especially during the winter. The positive phase of the NAO (NAO+ ) is characterized by a strengthening of both the Icelandic low and the Azores High. This pattern increases the strength of westerlies, transporting humid and relatively warm air towards Northern Europe and generating warmer and more humid winter weather than normal in temperate Northern Europe (Bladé et al., 2012; Gimeno et al., 2003). In contrary, Southern Europe experiences drier winters. During summer, this region experiences drought conditions due to reduced precipitation in the preceding months (López-Moreno and Vicente-Serrano, 2008). As a result, Roig et al. (2009) was able to link tree-ring width with NAO indices in the Iberian Peninsula. Also Gouveia et al. (2008) found reduced vegetation activity after a winter characterized by NAO+ in Southern Europe. Another consequence of the reduced winter and spring precipitation in Southern Europe is a possible enhancement of land-atmosphere coupling, that may favor even drier conditions due to different positive feedback loops (further discussed in section 3.3) (Wang et al., 2011). During the negative phase of the NAO (NAO− ), the westerlies are reduced and an opposite effect on the weather is observed over Europe. Mediterranean winters are wetter and warmer than average, while in Northern Europe it is drier and colder. The extreme cold weather in March 2013 was characterized by NAO− . Opposite from the NAO+ , the NAO− reduces precipitation over northern Europe, causing summers to be drier on average (López-Moreno and Vicente-Serrano, 2008). The NAO index however, is based on the data of a few stations only (Lisbon, the Azores or Gibraltar and Reykjavik), and might therefore not observe all movement of the action centers entirely (Roig et al., 2009). 1.2.2 Principal summer weather regimes During the summer, the main atmospheric weather regimes or circulation patterns are studied in detail by Cassou et al. (2005). They revealed four main states of the atmosphere by analyzing the geopotential height anomalies at 500 hPa (Fig. I.2). Both the NAO+ and NAO− patterns were observed together with the Atlantic Low and the Atlantic Ridge. The Atlantic Low exhibits a negative pressure anomaly above the North Atlantic Ocean and a weak positive anomaly above Europe. The Atlantic ridge is characterized by a negative anomaly above Scandinavia and Greenland, while an anticyclonic pattern is located over the North 1. EUROPEAN SUMMER CLIMATE 9 Figure I.2: (a-d) Four main atmospheric circulation patterns over the North Atlantic-European sector from 1950 to 2003 at Z500 (m). The figure is taken from Cassou et al. (2005). 10 CHAPTER I. GENERAL INTRODUCTION Atlantic Ocean east from Western Europe. Only Blocking and the Atlantic Low will be further discussed, as they are associated with warm European summer temperatures. 1.2.3 Weather regimes associated with warm summer temperature anomalies 1.2.3.1 Blocking One of the two patterns associated with exceptional warm weather in Europe summers is a Blocking pattern. The anticyclonic conditions over northern Europe, associated with Atmospheric Blocking, transports air from the east into Europe. This air has had sufficient time over land to warm up, and therefore transports warmer air especially into central and northern Europe. Because of the strength of the positive pressure anomaly, local convective instabilities are suppressed, diminishing wind strengths and increasing drought and cloudless skies even more (Cassou et al., 2005). Warm and dry conditions are especially found over the British Isles, central Europe and a part of the Baltic (Linderholm et al., 2009; Bladé et al., 2012). In the Mediterranean region this circulation pattern has less effect because the high pressures zone is too far north (Bladé et al., 2012; Mariotti and Dell’Aquila, 2012). However, on general wetter conditions can be found (Linderholm et al., 2009). The extreme heatwave in August 2003 and exceptionally warm July 2013 were characterized with an atmospheric Blocking pattern. 1.2.3.2 Atlantic Low The Atlantic Low is the second pattern that has been related with warm summer temperatures and heatwaves in Europe. It has most influence over the southern regions. Warm air from northern Africa and the Mediterranean basin is transported northwards (Cassou et al., 2005). The recent extreme hot summer in southern Europe of 2015 is characterized by an Atlantic Low pressure regime. 1.3 Land-atmosphere interactions influencing the summer climate The influence the land surface has on climate is mostly through the effect of soil moisture controlling the partitioning of the land heat fluxes. Wet soils have the potential for evaporation and high evaporative fractions (ratio of latent heat flux to net radiation), while sensible heat flux dominates over dry soils. Budyko defined two evapotranspiration regimes; a soil moisture limited and an energy limited regime (Fig. I.3). In the former regime, enough energy is available for evapotranspiration but due to a dry soil, there is very limited water accessible for latent heat flux. This regime prevails in the Mediterranean region during summer. In the energy limited regime however, soil moisture is plenty while the energy to evaporate the moisture is limited. This situation occurs in Northern Europe. Both regimes are quite stable and the land surface has very limited ability to influence the atmosphere (Schär et al., 1999; Seneviratne et al., 2010). In between these regions both an energy- or a soil moisture limited situation can be found, where wet soils may lead to increased evapotranspiration and relative low temperatures, whilst dry soils might favor higher temperatures with less precipitation. This can lead to the large interannual summer variability found in Europe. The coming two sections (I.3.1 and I.3.2) describe soil moisture-temperature feedbacks and soil moisture-precipitation feedbacks in more detail. This is followed by a section where regions of strong land-atmosphere coupling are discussed. I.3.4 describes the role of vegetation in land-atmosphere feedbacks. 1. EUROPEAN SUMMER CLIMATE EF=λE/Rn EFmax 0 DRY 11 TRANSITIONAL SOIL MOISTURE LIMITED θWILT WET ENERGY LIMITED θCRIT Soil moisture content Figure I.3: Definition evapotranspiration regimes. This definition follow the framework described in section I.3. EF denotes the evaporative fraction, and EFmax its maximal value. The dotted lines indicate possible pathways, as changes in EF are not necessarily linear. Adapted from Seneviratne et al. (2010). 1.3.1 Soil moisture-temperature feedback Soil moisture has the potential to influence temperature through the partitioning of land heat fluxes in latent, sensible and ground heat flux. The latter accounts only for a very small percentage and does not exert a strong influence on air temperature. The two remaining heat fluxes can affect temperature through multiple positive and negative feedback loops. A very direct consequence of enhanced sensible heat flux is its immediate effect on temperature. Besides that, a lower Bowen ratio due to drier soils leads to a drier atmosphere and reduced cloudiness increasing the net radiation. Intensified radiation and a drier atmosphere raise the atmospheric demand and consequently induce more evapotranspiration and even drier soils. A reduced convection may furthermore lead to a development of anticyclonic circulation conditions in the upper air (Zampieri et al., 2009). Also, in the beginning of a warm period, when soil moisture is still available, an increased atmospheric demand and an increased latent heat flux may lead to more relative humidity and thus a stabilizing situation. The positive feedback loops described above act mostly non-linearly, and a higher impacts of soil moisture on temperatures are found during dry conditions on temperature maxima (or a decrease of temperatures with initial wetter soil conditions) (Jaeger and Seneviratne, 2011). Ford and Quiring (2014) and Mueller and Seneviratne (2012) find similar results in that soil moisture has most impact on extreme warm temperatures. The latter find a higher probability on hot days after dry soil conditions, e.g. > 70% in the Iberian Peninsula, South America and Eastern Australia; > 60% in most of North US and Eastern Europe, while it decreased till < 30-40% after wet soil conditions. Quesada et al. (2012) and Hirschi et al. (2011) find similar results from observations over Europe, and Lorenz et al. (2015) find a stronger impact for soil moisture variability on the warmest temperatures. During the 2003 heatwave a positive coupling between soil moisture and temperature is found through reduced evaporation, a drier atmosphere and reduced cloud cover by Stéfanon et al. (2014). They find that this feedback contributed up to 20% of the extreme temperatures in Eastern France and Western Germany, while its contribution reached up to 40% over Western France and Northern Spain. However, although these processes can account for an important part to the temperature anomalies during the hot summer of 2003, the authors note that the initial cause of the extreme event was a Blocking situation over Europe. 12 CHAPTER I. GENERAL INTRODUCTION While these feedbacks are mostly local and on shorter time scales, non-local impacts with longer timescales are also possible. A well described mechanism is the impact of reduced winter and spring precipitation in southern European on elevated summer temperatures in Europe (Vautard et al., 2007; Zampieri et al., 2009; Quesada et al., 2012). A probable underlying process is the local inhibition of convective cloud formation and increased sensible heat flux due to reduced precipitation. These warm and dry air masses can be advected over central Europe due to southerly flows. The reduced cloudiness associated with these air masses increases atmospheric demand of soil moisture, reducing the water reservoir, enhancing sensible heat flux and higher temperatures (Zampieri et al., 2009). A negative feedback loop can occur when latent heat flux is favored over sensible heat flux by wetter initial soils. This might occur when temperatures start rising in the beginning of the summer. The higher temperatures increase latent heat flux, which leads to an increased atmospheric humidity. This in turn enhances cloud formation by lowering the dew point, leading to reduced net radiation and decreased temperatures, reduced atmospheric demand and thus relatively more soil moisture. The soil albedo may also play a role as it usually bears higher values with a lower soil humidity. With an increase in longwave radiation due to dry and warm soil, they can lead to a reduction of net radiation and consequently lower temperatures (Eltahir, 1998). A last negative coupling may be through a destabilization of the atmosphere due to very strong thermal activity. This can result in cloud formation when the PBL reaches the lifting condensation level (Boé, 2013). 1.3.2 Soil moisture-precipitation feedback Soil moisture can influence precipitation in different ways. The most direct mechanism might be moisture recycling, in which regional evaporated soil moisture contributes to precipitation. Comparing different studies Eltahir and Bras (1996) find that 10% of the precipitation may be derived from regional recycled water over Eurasia, while in the tropics this can be 25%-60% (Spracklen et al., 2012). Two important processes that determine the magnitude that soil moisture has on precipitation are: 1) the amount of soil moisture that is evaporated; and 2) the amount of evaporation that is precipitated. While the former is relatively easy to measure, the latter is harder to quantify. It is mostly considered that the effect of soil moisture on evaporation is larger than the effect of evaporation on precipitation (Wei and Dirmeyer, 2012; Boé, 2013). Another direct mechanism is the influence of soil moisture on the intensity of precipitation. Lorenz et al. (2015) showed that in mid-latitudes, and especially in the Mediterranean, soil moisture variability increases heavy precipitation events both in amount of rainfall and in duration. Besides this direct soil moisture-precipitation feedback, three indirect feedbacks have been described through the modification of soil moisture on boundary layer stability and on the potential for convective precipitation by Schär et al. (1999). 1) Starting with dry soils, an increase of the Bowen ratio and a deepening of the daytime PBL is expected. This causes an increased entrainment of dry air. The changes in PBL and entrainment together induce a decrease in the density of water vapor and in the moist static energy (MSE) that can lead to reduced potential for convective rainfall. 2) An increase of longwave radiation due to dry soils and an increase in albedo induces a decrease in net radiation. This decreases the total turbulent energy flux transferred to the PBL that decreases the MSE. The decrease in MSE in a deeper PBL with increased entrainment causes again a reduced potential for convective 1. EUROPEAN SUMMER CLIMATE 13 precipitation (Schär et al., 1999). 3) The level of free convection decreases due to dry soils, also stabilizing the atmosphere and reducing the potential for convective rainfall. These processes remain positive feedback mechanisms when starting with wet soils instead of the dry case examples, thus enhancing the potential for convective precipitation (Schär et al., 1999). The indirect mechanisms have been subject of many modeling and observational studies. Findell et al. (2011); Guillod et al. (2014) found for example that previous wetter conditions enhance the probability of afternoon rainfall in Mexico and eastern US. In central and southwestern US however, where a soil moisture limited regime prevails, soil moisture is the driver for positive feedbacks on precipitation. Rios-Entenza and Miguez-Macho (2014) find a large impact of evapotranspiration on precipitation in the Iberian Peninsula, especially when there is the potential for convection. The feedback mechanisms are thus also dependent on largescale circulation patterns (Ford et al., 2015a; Guillod et al., 2015; Boé, 2013). Soil moisture may exert an influence on the persistence of certain large-scale circulation patterns, however, this has so far only been demonstrated for SST anomalies (Guemas et al., 2010). Some studies also mention negative feedbacks mechanisms between soil moisture and precipitation. Boé (2013) find a weak but negative feedback during warm conditions with a Blocking pattern. They assign this to a destabilization of the atmosphere over drier soils. Likewise Taylor et al. (2011); Taylor (2015); Ford et al. (2015b); Guillod et al. (2015) find that convective precipitation initiation has a preference over dry soils. During the 2003 heatwave many studies have investigated the soil moisture-temperature feedbacks, but Zaitchik et al. (2006) studied the soil moisture-precipitation feedback in detail during this event. They find evidence for a negative feedback due to a reduced local production of MSE (the second indirect mechanism) only during April. Non-local feedback also occur, where the advection of evaporated soil moisture affects neighboring regions Beljaars et al. (1996); Rowntree and Bolton (1983). The latter found that moister or drier air was advected from central Europe into Scandinavia and from southern Europe into North Africa. This has also been shown to occur for soil moisture-temperature coupling during hot European summers, where reduced southern European precipitation during winter and spring induces warmer conditions over Europe (Zampieri et al., 2009). While there is substantial evidence for soil moisture-precipitation feedbacks, precaution has to be taken with the interpretation of results. Different observation-based data sets and different convection parameterizations in climate models can change the sign and the magnitude of the feedbacks (Hohenegger et al., 2009; Guillod et al., 2014). 1.3.3 Regions for strong land-atmosphere feedbacks Regions of strong soil moisture-temperature and -precipitation feedbacks, as discussed in the previous sections, have been studied extensively in the Global Land-Atmosphere Coupling Experiment (GLACE) (Koster et al., 2004). These so called ’hot spots’ are mostly identified for their strong feedbacks between soil moisture and precipitation. Although when soil moisture has a substantial influence on precipitation and thus on the atmosphere, we can assume that it is similar for temperature. Simulations with varying and climatologically prescribed soil moisture performed with an AGCM ensemble (GLACE), showed that strongest land-atmosphere feedback strength was 14 CHAPTER I. GENERAL INTRODUCTION over the Great Plains in Northern US, the Sahel, equatorial Africa and India. Less coupling strength was found over parts of the Amazon, Eastern China and some parts over Russia and Canada. Using GCMs over Europe Koster et al. (2004) did not find land-atmosphere coupling strength, besides a very small region north of Portugal and around Poland. Using the absolute values of precipitation instead of its natural logarithm, as was done in Koster et al. (2004), Koster et al. (2006) found that hot spots occurred with a very similar pattern but with a decreased magnitude, indicating some form of robustness. Using only the eight best performing models for summer precipitation, they found that the magnitude of coupling with soil moisture increased in most areas, including a small region in South-East Europe. With these same eight models, Seneviratne et al. (2006a) obtained an estimate of seasonal forecasting potential by combining the coupling strength with soil moisture memory. While the coupling strength is not enormously high, some hot spots in Europe appear: in the Mediterranean regions, the southern part of France and in Eastern Europe around the Danube. Including more detailed analysis on the surface heat fluxes, Dirmeyer (2011) found positive coupling over almost the entire Mediterranean area, while there is a negative coupling over Northern Europe and over the Alps. The latter is probably related to late snow melt. Using a regional climate model over Europe, Seneviratne et al. (2006b) found different feedback hot spots than were found by Koster et al. (2006). While Koster et al. (2006) did not find any sign of strong land-atmosphere coupling over Europe, Seneviratne et al. (2006b) find that in the present climate strong coupling is found over the Mediterranean region, which is more similar to Dirmeyer (2011), although the regions found by the latter is extended more west-ward. They assign the changes to the different models, a different spatial resolution and a different representation of SSTs. In the future the coupling hot spots might change. These ’new’ hot spots might occur more north and east-ward, extending over central and Eastern Europe (Seneviratne et al., 2006b). The difference in coupling strength between the separate models is mostly caused by the different sensitivities of evaporation rate to soil moisture. Models with a stronger correlation between these variables show stronger feedback strength (Guo et al., 2006). The ensemble model mean of the coupling strength is shown to be fairly well simulated (Dirmeyer et al., 2006). However, interpretations of single models must be done with care, because their skill in correctly simulating the relation between land and atmosphere is limited. Comparing certain model variables such as latent heat flux and surface soil moisture with observations, Dirmeyer et al. (2006) find large biases. 1.3.4 Influence of vegetation A very clear example of vegetation-atmosphere feedbacks is the work of Aleina et al. (2013). They model a world starting with different vegetation fractions and different initial deep soil moisture conditions. They find three stable equilibria: a warm desert without vegetation, a cold desert with little vegetation and a temperate vegetated state. The vegetated state can only be reached when starting with a sufficient initial vegetation fraction and sufficient initial deep soil moisture. Whenever the initial vegetation fraction is too small, the generated water cycle is not strong enough to maintain the vegetation, and thus the water cycle, alive (Aleina et al., 2013). This extreme case is very unlikely to happen on a global scale, but it illustrates the importance of impact of vegetation on the water cycle on longer time scales. Some ecosystems seem to have multiple alternative stable states, and the loss 1. EUROPEAN SUMMER CLIMATE 15 of only a small fraction of vegetation may shift the system from a vegetated to a desert state (Scheffer et al., 2001). This is also shown by some modeling studies (Wang and Eltahir, 2000; Baudena et al., 2008). Other feedback mechanisms between vegetation and atmosphere occur on much shorter time-, and smaller spatial-scales. A well-studied topic is the difference between forests and croplands in their effects on temperature. Intuitively forests may seem to better moderate temperatures, but in reality this is not always the case (Teuling et al., 2010). Two processes that can explain differences between the two vegetation types are the albedo and evapotranspiration. While evapotranspiration is mostly found to have a mitigating effect on temperature, as explained in section I.1.3.1, the lower albedo of forests might increase local and regional temperatures (Teuling et al., 2010). Stéfanon et al. (2012) presented the mitigating effect of evapotranspiration in June 2003, which was increased due to enhanced spring greening. During August however, the vegetation stress and early leaf fall reduced evapotranspiration, thereby increasing sensible over latent heat flux and amplifying the already above average temperatures Lorenz et al. (2013). Detto et al. (2006) found similar mitigating results for Sardinia, where the natural woody vegetation evaporated more and longer during drought periods than grasses. Vegetation is also found to enhance evapotranspiration and accordingly the water cycle in the semi-arid regions of Northern China (Jiang and Liang, 2013). Although the enhanced evapotranspiration may have a moderating effect on temperature, the faster depletion of the soil water reservoir (Zhang and Schilling, 2006) may lead to enhanced sensible heat flux during longterm droughts. Although Baudena et al. (2008) find that natural vegetation can balance these effects through the adaptation to lower soil moisture. The negative effect of albedo is especially pronounced over the boreal regions, where a decrease in albedo occurs by forests that cover the snow layer. This induces a rise in temperature (Snyder and Liess, 2014; Bonan, 2008). Furthermore, forests usually have a lower albedo than grass or crop land, retaining more sunlight energy that also increases the temperature (Jackson et al., 2008; Naudts et al., 2016). This effect is also found by Teuling et al. (2010) during the heatwave of 2003. The air temperature over grasslands was found to be cooler than over forests due to a lower Bowen ratio. However, the different rooting depths between trees and grasses or crops may inverse these effects during periods of prolonged drought. Other than the non-local effects found for both soil moisture-temperature and soil moistureprecipitation feedbacks, the effects of the biosphere on the atmosphere seem to be only local (Pitman et al., 2009). Also Lorenz et al. (2013) found that vegetation only has a minor effect on mean spring and summer climate in Europe. In the future, reforestation has the potential to reduce global warming till approximately 2050. Increased CO2 concentrations can cause an increase in leaf area index which enhances evapotranspiration. However, at the end of the 21st century the cooling by evapotranspiration becomes less important due to CO2 fertilization (Pitman and Narisma, 2005). Similar results are found over central Europe by Wramneby et al. (2010), although in Southern Europe increased droughts may limit plant growth, reducing evapotranspiration and consequently increase temperatures. Vilà-Guerau De Arellano et al. (2012) emphasizes the CO2 fertilization effect on stomatal closure, and its possible impact on the partitioning of heat fluxes and cloud 16 CHAPTER I. GENERAL INTRODUCTION formation. 2 Heatwaves The processes described in the previous sections lead to interannual and seasonal variability and extreme events such as heatwaves. While definitions vary, heatwaves are usually described as prolonged periods of extreme high temperatures relative to the usual weather. The World Meteorological Organization defines a heatwave as a period of at least five consecutive days with a daily maximum temperature that exceeds the average maximum temperature of the period 1961-1990 by 5◦ C. Although all heatwaves are characterized by above normal temperatures, they differ temporally and spatially. Consequently it is impossible to describe one mechanism that describes all heatwaves. For this reason I will discuss the recent 2003 European heatwave as an illustrative example, describe its characteristics, possible mechanisms and consequences. This specific heatwave is selected as it is very well described in literature and because of the varying processes that resulted to the anomalous hot temperatures. 2.1 Characteristics of the 2003 heatwave The European heatwave in 2003 broke many temperature records, both in June and August. Exceptionally warm periods with daily maxima well over 40◦ C were marked across Spain and Portugal, 36-38◦ C over Southern and central France and 32-36◦ C over Northern France (García-Herrera et al., 2010). The monthly average of temperature anomalies were +2◦ C, +4.2◦ C, +2◦ C and +3.8◦ C from May to August, respectively (Black et al., 2004). The temperature anomaly of early August was up to 8.5◦ C (Zaitchik et al., 2006). The summer mean exceeded the climatological mean by approximately 3◦ C in some regions, corresponding to five standard deviations (Schär et al., 2004). Most exceptional temperatures were recorded in France and Switzerland (Black et al., 2004), with 40-60 hot days, which is 5 to 6 times the long term average (Fischer et al., 2007a). The rareness of the temperatures over Switzerland are shown in their very long return period of 46 000 year (Schär et al., 2004). This summer was likely the warmest in the last 500 years (Luterbacher et al., 2004). 2.2 Mechanisms behind the 2003 heatwave The main mechanisms of the heatwave were the winter and spring precipitation deficit, an atmospheric Blocking pattern which caused anticyclonic conditions, warm SSTs and landatmosphere interactions. Teleconnections are also mentioned in several studies, but will not be discussed in detail. For a comprehensive overview I refer to García-Herrera et al. (2010). Precipitation deficit Over large parts of Europe, the average precipitation was more than 50% lower in the preceding winter and spring months compared with the long-term average (Fischer et al., 2007a). This deficit probably amplified the evolution of the temperature anomalies during spring and summer (Loew et al., 2009). The cause of the anomalous rainfall might have been anomalous 500 hPa geopotential heights (García-Herrera et al., 2010). 2. HEATWAVES 17 Atmospheric circulation Besides, and maybe related to the precipitation deficit, there was a persistent atmospheric Blocking pattern. This pattern was associated with anticyclonic situations during almost the entire summer season (Loew et al., 2009; Cassou et al., 2005; Trigo et al., 2005), being most pronounced in June, and in August over France (Black et al., 2004). The anticyclonic conditions might have been related with the Icelandic low that was displaced more southward and the intensification and northward shift of the Azores anticyclone (Black et al., 2004). The almost stagnant atmospheric flow resulted in the possibility of the build-up of a warm and deep boundary layer (Black et al., 2004), remaining relatively constant over night time (Miralles et al., 2014). Besides the promotion of land-atmosphere feedbacks, the synoptic patterns also induced clear skies and a positive downward net radiative flux, which were preserved by the interactions between land and atmosphere (Black et al., 2004). SSTs Although SSTs may not be able to induce a heatwave, they can prolong or favor certain large-scale atmospheric patterns, which can form the foundation of extreme climatic events (Guemas et al., 2010). During the spring and summer of 2003 the Mediterranean SST was anomalously high (Feudale and Shukla, 2007). The sea started warming in April. In June, both the Mediterranean Sea and the Atlantic Ocean were hotter than normal (Black et al., 2004). Feudale and Shukla (2007) suggests that global SSTs were responsible for the anticyclonic blocking, whilst local SSTs were more important in maintaining the heatwave (Fennessy and Kinter III, 2011). A combination of anomalously global SSTs and those in the Atlantic Ocean may have caused a reduction in the meridional gradient, leading to a northward shift of the polar yet, that allowed an enlargement of the anticyclone. The impact of the SSTs on the 2003 temperatures may have been in the order of 1-2◦ C (Fennessy and Kinter III, 2011). Land-atmosphere coupling Land-atmosphere feedbacks have been mentioned in many studies for their contribution to the extreme 2003 summer temperatures. The soil drying probably contributed to local heating (Schär et al., 2004) and caused a increase of 1-2◦ C to the surface temperature (Fennessy and Kinter III, 2011), with 40-50% to the maximum temperatures (Fischer et al., 2007a). It furthermore enhanced dry convection and entrainment that deepened the boundary layer (Black et al., 2004; Miralles et al., 2014). Due to the stable synoptic conditions, the dry and warm air prevailed inducing further soil desiccation. The early vegetation green-up (Zaitchik et al., 2006) caused positive latent heat flux in the beginning of June (Black et al., 2004), dampening the heatwave (Stéfanon et al., 2012; Lorenz et al., 2013). However, the depletion of the soil moisture caused negative evapotranspiration anomalies later in summer (Fischer et al., 2007b). This in turn increased temperatures, leading to further decreased vegetation production and an amplification of the hot temperatures (Stéfanon et al., 2012; Lorenz et al., 2013). A change in albedo due to soil drying could possibly have a negative effect on temperature, although this was largely masked by the water-stressed vegetation that increased the albedo (Teuling and Seneviratne, 2008). Besides the direct effect on temperature, the drought conditions may also have amplified the anticyclonic circulation patterns (Fischer et al., 2007b). The total soil moisture-temperature coupling might have induced 50-80% of the number of hot days during the 2003 heatwave (Fischer et al., 2007a), and was strongest even before the peak of the temperature in France (Miralles et al., 2012). 18 2.3 CHAPTER I. GENERAL INTRODUCTION Consequences of the 2003 heatwave The impacts of the heatwave were widespread in many different sectors. The total estimated human mortality related to the event is estimated at over 40 000, particularly elderly people (García-Herrera et al., 2010). In France alone the excess mortality was 14 800 during August (Fouillet et al., 2006). Besides the mortality rate, the failure of the health systems to such extreme events was evident (García-Herrera et al., 2010). Air pollution was also problematic. Exceptionally high ozone concentrations have been measured during an extensive period (Vautard et al., 2005). The increased UV radiation due to decreased cloud cover and precipitation, stagnant atmospheric circulation, reduced dry deposition, increased biogenic emissions and forest fires have been mentioned as cause (Tressol et al., 2008). Although forest fires were not extreme (García-Herrera et al., 2010), they resulted in an economic loss of $1.6 billion (Schär and Jendritzky, 2004). The fires in Portugal in the beginning of August might have resulted in the highest aerosol load of this month (Hodzic et al., 2006). The reduced dry deposition was caused by the stomatal closure of vegetation (Emberson et al., 2000). The biogenic emission came mostly from isoprene, emitted by trees as a defense against the high temperature and increased solar irradiation (Solberg et al., 2008). The agricultural sector also experienced losses both in crops and animals (García-Herrera et al., 2010). Crop losses were unprecedented (INSEE, 2004; Zaitchik et al., 2006), accounting for an economical loss of $12.6 billion (Schär and Jendritzky, 2004). The largest losses were mainly of non-irrigated summer crops. Some wine regions however, were positively affected (García-Herrera et al., 2010). The GPP of natural vegetation was severely affected and a reduction was estimated around 30% (Ciais et al., 2005; Reichstein et al., 2007). Only mountainous vegetation experienced increased growth rates. In spring the early green-up (Zaitchik et al., 2006) led to higher GPP in some regions, although the reduced precipitation also caused negative NDVI in other areas (Lorenz et al., 2013). In August most vegetation activity ceased (Stéfanon et al., 2012). 3 Impact of elevated temperatures and drought on vegetation The summer climate usually is the period of most extensive growth and production in many parts of Europe, the Mediterranean being an exception. The higher temperatures are favorable for most tree species if the soil water level is sufficiently high. The latter is generally the bottleneck for vegetation development during summers in Southern Europe. Given the predicted increase in interannual summer climate variability in central Europe (Schär et al., 2004), this might become a relevant issue in much larger regions of Europe. Especially after the 2003 heatwave has shown a decrease in vegetation production of 30% (Ciais et al., 2005). Similar results were found for the Great Plains after the 2012 summer drought, although the enhanced spring temperatures compensated the summer carbon loss (Wolf et al., 2016). In the section below a description is given of the most common and general responses of vegetation to drought or heat stress. The focus will be on European forest tree species. Note however, that in the various climate zones, different species occur with different responses to water shortage. There is also considerable intra-species variability, which depends on genotypic variation and phenotypic plasticity (Corcuera et al., 2011). A more detailed description 3. IMPACT OF ELEVATED TEMPERATURES AND DROUGHT ON VEGETATION 19 of the effects of drought and heat of the most common European tree species will be given in chapter V. The main challenge for plants during warm and dry summers is to continue carbon assimilation while water is limited. Carbon is taken up by plants in the form of CO2 through their stomata, but water is evaporated through the stomata at the same time. Broadly, two types of drought response exist: drought avoidance or isohydric species with quick stomatal closure with beginning drought, and drought tolerance or anisohydric species, which decrease stomatal conductance and water potential and can keep their stomata open for a longer time period by increasing root water uptake. The disadvantage or the latter one is the risk of cavitation (disrupted water columns) and embolism (air-filling), or hydraulic failure (Hartmann et al., 2013), while the problem of the first response mechanism is carbon starvation (Sevanto et al., 2014). Embolism is not necessary exceptional and detrimental (Lo Gullo and Salleo, 1992). It can occur on a daily basis during summer, and refilling of the vessels can occur within minutes and during the night (McCully et al., 1998; Salleo et al., 2004; Zufferey et al., 2011). The problem develops when too many xylem vessels are embolized or when ’runaway embolism’ occurs (increased frequency of cavitation due to a reduced number of functioning vessels with a constant transpiration rate) (Tyree and Sperry, 1988). Cavitation occurs more easily in wider vessels, in where the hydraulic conductivity is larger (Tyree and Zimmermann, 2002). A higher conductivity increases the fluency of the water transport. It is therefore believed that a trade-off exists between the resistance to cavitation and the specific hydraulic conductivity (Martínez-Vilalta and Piñol, 2002). Most tree species seem to operate very close to their ’catastrophic’ cavitation limit and have small safety margins (Tyree and Sperry, 1988; Choat et al., 2012). Roots seem to act more closely to their limits than other parts of the tree (Sperry and Saliendra, 1994; Hacke et al., 2000). When the soil is wet this reduces the resistance of water transport, but when the soil is dry it disconnects the plant from the soil (Sperry and Ikeda, 1997; Martínez-Vilalta and Piñol, 2002). Furthermore, root construction is relatively cheap in carbon compared to root maintenance. Thus whenever xylem embolism becomes too high, roots can be shed without major consequences (Bouma et al., 2001). Isohydric species close their stomata much faster in case of droughts. Stomatal conductance may be reduced even before stem water potential decreases (Gollan et al., 1992). Roots are believed to release ascorbic acid (ABA) that induces this process (Cochard et al., 1996b; Beck et al., 2007). Because of carbon reserves trees can survive for a while without carbon assimilation. Prolonged droughts however, may lead to carbon starvation (McDowell et al., 2008; Galiano et al., 2010). Although there has been some controversy against this theory (Rowland et al., 2015). But even if trees would not die from carbon starvation, carbon reserves have been shown to decline during droughts leading to reduced growth rates in subsequent years (Galiano et al., 2010; Gruber et al., 2012) and reduced susceptibility to pathogens and insect attacks (Rebetez and Dobbertin, 2004; Bigler et al., 2006; Dobbertin et al., 2007). Furthermore, trees seem to be more vulnerable to successive droughts when growth rates have been reduced (Bigler et al., 2007). Another disadvantage of stomatal closure are photoinhibition and photorespiration. The former occurs when an imbalance between the electron transport rate and the available CO2 occurs. Electron transport, which is driven by light, is often increased during prolonged periods of drought and clear skies, whilst carbon assimilation is reduced. This can lead to the formation of reactive oxygen species 20 CHAPTER I. GENERAL INTRODUCTION (ROS) (Hendry et al., 1992), which may damage proteins in membranes or in the photosynthetic apparatus (Perdiguero et al., 2013). Antioxidants can reduce this problem by oxidizing or scavenging ROS (Costa et al., 1998). Photorespiration occurs when Rubisco binds to O2 instead of CO2 . The plant is not able to use the formed molecule and its breakdown costs energy. Besides stomatal closure, trees can maintain water potentials with osmotic adjustment (Osonubi and Davies, 1978; Harfouche, 2003). This is the active accumulation of solutes in the vacuole of cells and is a mechanism to maintain turgor pressure (Thomas and Hartmann, 1996; López et al., 2009; Sergeant et al., 2011). Turgor pressure is important in among others cell expansion and growth, and stomatal aperture (Nguyen-Queyrens et al., 2002). Tree morphology can also be adapted in response to drought stress. Leaf thickness may be increased while their size may be smaller to limit evapotranspiration (Bansal et al., 2015). Increased cuticular wax leads to a similar result (Le Provost et al., 2013). Leaf abscission has also been observed in response to drought stress to reduce evapotranspiration rates, and to redistribute assimilates (Bréda et al., 2006). The number of trichomes can be increased for larger sunlight reflectivity (Sardans et al., 2013). To avoid cavitation xylem vessels could be adapted for higher cavitation resistance (Hacke and Sperry, 2001; Paiva et al., 2008). Biomass allocation can be shifted to the roots to maximize the uptake of limiting resources (López et al., 2009; Barthel et al., 2011; Matías et al., 2014). C HAPTER II Uncertainties in the future climate 21 22 1 CHAPTER II. UNCERTAINTIES IN THE FUTURE CLIMATE Introduction There are considerable uncertainties in simulating the present climate of Europe that can propagate into the future climate. This raises several questions, of which I answer two in this chapter with two different studies presented in two parts. The first part discusses uncertainties of future projections for Europe for a given 2◦ C average global warming. The uncertainties constraining the future climate by temperature instead of by time are presented. A specific focus is placed on summer temperatures and land heat fluxes. This study is part of the European IMPACT2C project and the results are published by Vautard et al. (2014). I will only show the results of the work done by myself in this study. The complete paper can be found in Appendix A. The second part of this chapter investigates how the uncertainty of modeled present-day land heat fluxes links with uncertainties in the future climate. Feedbacks between land and atmosphere play an important role in generating summer heat in temperate climates. This can also be seen in models, where stronger feedbacks result in higher temperatures. Stegehuis et al. (2013b) showed that, using an ensemble of models, larger sensible heat flux is associated with elevated summer temperature (Fig. II.1). Furthermore they found large differences in the heat flux partitioning between the models, leading to uncertainties in present-day simulations. However, how these differences and uncertainties in land heat flux partitioning translate into future climate projections is unknown. The use of model ensembles has been shown to better represent uncertainties than single model simulations. Dirmeyer et al. (2006) showed for example that land atmosphere feedbacks are relative poorly represented in single models, while the ensemble model mean showed much better skill. This is partly due to the scarcity of long-term gridded observationbased data of some of the important variables related to land-atmosphere feedbacks such as the land heat fluxes. Recently some satellite-based products have become available such as GLEAM (Miralles et al., 2011). Another product, based on a combination of remote sensing measurements and FLUXNET data is the Model Tree Ensemble (MTE) of Jung et al. (2010). The latter data set is used in the present study. However, by using these kind of data sets, one always has to keep its uncertainty in mind. Mueller et al. (2011) showed that the spread in observation based latent heat flux products was as large as that from a model ensemble. Because of the better representation of model ensembles compared to that of single models, different regional and global model ensembles have been constructed. The European RCM ensemble from the FP6-ENSEMBLES project (Hewitt and Griggs, 2004; van der Linden and Mitchell, 2009) is one example on a regional scale. In this study 15 of its RCMs, driven by 6 different GCMs from the 4th IPCC assessment report for the A1B scenario, were used. The two studies presented below use this ensemble. Regional European impacts of a 2◦ C average global warming In the first study, we investigate regional European effects on temperature, land heat fluxes and the radiation budget with an average global warming of 2◦ C since the pre-industrial period. This has been done in context of the European IMPACT2C project in the view of the 2◦ C global warming target. The target has been agreed upon during the last United Nations Conference on Climate Change (COP21). While in climate studies usually a temporal range in the future is considered, a global temperature metric for climate change is often used as a measure of impacts and risk of potential tipping points, which has also been looked upon in the present study. 1. INTRODUCTION 23 Figure II.1: Annual cycle of latent heat flux (a), sensible heat flux (b) and 2-m temperature (c) over Europe, from 1961 to 2000 for RT3-models, ERA-40 and E-OBS, and from 1983 to 2008 from MTE and ERA-I. The two colored models are examples of the different feedback strengths between models. The observations are in black. From Stegehuis et al. (2013b). The ENSEMBLES models are used to study the above mentioned climatic consequences and uncertainties in Europe when the global temperature modeled by the driving GCMs reaches 2◦ C. These periods range between 2014 and 2067. The full study can be found in Appendix A (Vautard et al., 2014). We find that Europe experiences on average a higher warming than the global average, thus exceeding the 2◦ C regionally. Warming is largest in Scandinavia, the Mediterranean region, the Iberian Peninsula and the Alps. The British Isles experience a weaker warming. The land heat fluxes confirm this warming in the south, by showing less latent and more sensible heat flux in the southern regions during summer. In the energy limited region of Scandinavia, the higher temperatures lead to higher latent heat flux during summer, while sensible heat flux is diminished. This might be due to cloud formation that diminishes incoming shortwave radiation and enhances longwave radiation. Reduction of temperature change uncertainties by observational constraints In the second study we constrain future temperature projections by using land heat fluxes. Using the ENSEMBLES model ensemble we hypothesize that summer biases in the partitioning of the land surface energy balance in the current climate will be amplified under future (drier) summer conditions. Models with stronger feedbacks will then produce hotter and drier future temperature. Information on the present-day may then help constrain temperature change projections. This concept has been used by Hall and Qu (2006), who used the seasonal cycle of the snow albedo feedback to reduce uncertainty in temperature predictions. Boberg and Christensen (2012) related the future temperature bias of models to their present-day bias to constrain projections. They found the models’ bias especially in the warm end of the temperature distribution. Quesada et al. (2012) used a similar approach, but related winter and spring precipitation to summer temperatures. Although land heat fluxes are a key component of land-atmosphere feedbacks, these have not yet been used for the reduction of temperature change projections. Using observations of sensible heat flux we were able to constrain the ensemble and reduce the summer temperature change predictions regionally by up to 40%. Largest reductions were found over France and the Balkan, regions already mentioned for their impor- 24 CHAPTER II. UNCERTAINTIES IN THE FUTURE CLIMATE tance of land-atmosphere feedbacks (Seneviratne et al., 2006a). In Southern and Northern Europe, other metrics may be used to reduce uncertainty in temperature projections, such precipitation-temperature coupling (Quesada et al., 2012) or the snow-albedo feedback (Hall and Qu, 2006). In the next section I discuss some of the European climatic consequences of an average global warming of 2◦ C. This is followed by the peer reviewed paper on reducing future summer temperature change predictions by using land heat flux observations. In section II.4 the discussion will focus on model ensembles, the used method of uncertainty reduction and the 2◦ C limit. This is followed by a summary and some concluding remarks. 2 The European climate under a 2◦ C global warming 2.1 Introduction In most climate change studies, only the temporal scale in the future is considered and investigated. However, impact studies are often based on temperature change above a certain threshold instead of a period in the future. A limit of two degree global warming (since the pre-industrial period) has been advocated by the IPCC and has now been agreed upon as target during the last United Nations Conference on Climate Change (COP21). This limit is based on the IPCC 2nd assessment Report (IPCC, 1995). The potential risk of disastrous events to happen by exceeding this temperature threshold is substantial. In this section some of the climatic consequences over Europe were studied, with a focus on land atmosphere interactions, when the average global warming reaches 2◦ C. This has been done in the context of the European IMPACT2C project. The results presented here are part of a peer reviewed paper (Vautard et al., 2014 - Appendix A). 2.2 Methods We used 15 RCMs of the FP6 ENSEMBLES project (Hewitt and Griggs, 2004; van der Linden and Mitchell, 2009). The 2◦ C average global warming period for each RCM was based on its driving GCM. These 30-year periods lay in the range between 2014 and 2067 (Table 1 in Appendix A). A reference period from 2031 to 2060 is taken to remove the uncertainty of GCM sensitivity to GHG as compared with fixed time periods. A specific attention is payed to the land heat fluxes, and the difference between the ’2◦ C’ and the control period (1971-2000). A more detailed description of the methodology can be found in Appendix x. 2.3 2.3.1 Results and discussion Temperature and precipitation change The annual mean temperature increase over Europe, with a global warming of 2◦ C compared to the pre-industrial period, is higher than the 2◦ C global average, except from the British Isles. Most warming occurs in autumn and winter, while spring warms least (Fig. II.2). The difference of temperature change between the 2◦ C and 2031–2060 is small, but the ensemble spread of the period between 2031–2060 is larger than the spread of the 2◦ C period. This is expected because the models are centered on a similar temperature instead of a similar time-period in the future. 25 15 ∆TAS [°C] 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75 2. THE EUROPEAN CLIMATE UNDER A 2◦ C GLOBAL WARMING 10 ● ● ● ∆PR [%] 0 5 ● −5 ● ● ● Eastern Europe Mediterranean Alps Scandinavia Mid−Europe France 10.0 Iberian Peninsula ● British Isles Eastern Europe Mediterranean Alps Scandinavia Mid−Europe France Iberian Peninsula 2031−2060 2−degree period 5.0 ∆PR [%] 0.0 2.5 −2.5 0.00 −7.5 −5.0 2.50 1.00 0.50 ∆TAS [°C] 1.50 2.00 7.5 3.00 3.50 British Isles −15 ● −10 ● Spring (MAM) Summer Autumn (JJA) (SON) Winter (DJF) Spring (MAM) Summer Autumn (JJA) (SON) Winter (DJF) Figure II.2: (a) Yearly averaged change —relative to the reference period 1971–2000 — in yearly mean temperature in the different European regions for periods corresponding to +2C of global average change. The global temperature change (1.54◦ C) between 1971–2000 and the 2◦ C period is marked as a dotted line. (b) Same as (a) for precipitation in % of change. The solid red line indicates no change for precipitation and the red dotted line a change of 1.54 C for temperature, corresponding to a global warming of 2◦ C relative to pre-industrial. (c) Spatial average over land of changes in temperature and (d) precipitation, together with the range of changes for the GCM–RCM ensemble (median, 25–75% range and min and max values). The open bars refer to fixed time future period (2031–2060), the gray bars to the temperature controlled (+2◦ C) period. 26 CHAPTER II. UNCERTAINTIES IN THE FUTURE CLIMATE Most warming is found to occur over Scandinavia (Fig. II.2). The large warming might be partly due to a decrease in albedo due to snow melt and the general arctic warming (IPCC, 2013). Furthermore, the albedo might also be influenced by trees that cover snow. The snowalbedo feedback may explain the large winter warming signal in panel c (Fig. II.2). After Scandinavia, most warming is found in the Alps, the Iberian Peninsula, the Mediterranean region and Eastern Europe. Over France and central Europe it is somewhat lower. The large spread found over central Europe might be due to the possible differences between models in soil moisture or energy limited regimes, caused by land-atmosphere feedbacks and the uncertainty in the partitioning of modeled land heat fluxes during summer. Least warming is found over the British Isles. The large influence of the ocean on this region and the increases in clouds and precipitation may make less sensible for temperature change. Precipitation is more variable in the 2◦ C warmer world (Fig. II.2). There is a tendency of drier summers in Southern central Europe but wetter ones in Scandinavia, although the spread between the models is large. The other three seasons might experience more rainfall, especially autumn and winter, with a mean increase of approximately 2.5% and 5% respectively. Also here, the difference between the 2◦ C and 2031–2060 is small, but ensemble spread is large in both cases. Precipitation is expected to increase most in Scandinavia, followed by mid-Europe, the British Isles and Eastern Europe. This might be caused by an increase in evaporation due to elevated temperatures, causing a more humid atmosphere and more cloud formation. A precipitation decrease if found especially in the south of Europe. Drier soils result in the opposite pattern as found in the north: less humid atmospheres and less cloud formation. 2.3.2 Land heat fluxes and radiation Land heat fluxes show a consistent pattern of change in the future with previous studies, indicating a clear change in North and South Europe and a disagreement in sign between models in central Europe (Schär et al., 2004; Stegehuis et al., 2013b). In spring, latent heat fluxes increase over almost entire Europe, although the sign of change in most of Southern Europe can be negative, but the magnitude is small (Fig. II.3). In spring, soil moisture usually is still non-limiting and elevated temperatures may in that case an increase in evaporation. However, if the regime is already limited by soil moisture today, as seems to be the case in the most southern parts of Europe, an increase in temperature would cause a decrease in latent, and an increase in sensible heat flux (Fig. II.3). This is also evident from the radiation fluxes. Almost all models show an increase in shortwave radiation in the south, while longwave radiation decreases. In Northern Europe however, the average longwave radiation increases, although due to little model agreement, the area is not colored (Fig. II.3). In summer, the spring pattern of especially sensible heat flux is intensified, indicating more drying in Southern Europe and a possibly wetter climate in the north (also shown by its increase in precipitation - Fig. II.2). The increase in latent heat flux in the north causes more cloud formation leading to an increase in longwave radiation, while the amount of shortwave radiation decreases. Summer precipitation is also intensified. Central Europe is a region where climate change seems to be most uncertain and the variability is highest (Seneviratne et al., 2006b). The region can either stay in an energy limited regime as usually is the case in spring, or change into a soil moisture limited regime when enough energy is available to 2. THE EUROPEAN CLIMATE UNDER A 2◦ C GLOBAL WARMING 27 Figure II.3: Changes, between the 1971–2000 and the +2◦ C periods, in surface energy fluxes. Left panels: latent heat fluxes; second left panels: sensible heat fluxes; second right panels: net short-wave fluxes; right panels: net long-wave fluxes. Upper panels: MAM, Lower panels: JJA. Only areas with at least 12 models agreeing on change sign are colored. Areas where at least 14 models agree on change sign are highlighted with dots (except for temperature where almost all areas satisfy this). cause soil drying (Seneviratne et al., 2010). As shown by Stegehuis et al. (2013b), some models in the ENSEMBLES project overestimate land-atmosphere feedbacks based on their performance for present day latent heat flux observations. These models will probably make a switch between the regimes, while others might not do so. 2.4 Summary and conclusion Using an ensemble of RCMs, I showed that with a global temperature rise to a 2◦ C, Europe will experience an overall more intense warming than the global mean. The magnitude depends on the region and the season, with most warming in Northern and Southern Europe and during the winter season. Precipitation will increase except for the southern areas that tend to become drier. Both latent and sensible heat flux from model simulations confirm these findings. Latent heat flux shows an increase over almost entire Europe in spring and over Northern Europe in the summer. Whilst sensible heat flux increases especially in Southern Europe. Although there seems to be some robustness in the results (smaller model spread) over the most Northern and Southern parts of Europe, large uncertainty remains over central Europe. This can both be illustrated in the large error bars of Fig. II.2 and in the absence of color in Fig. II.3, in both land heat fluxes and radiation. The temperature increase in the Mediterranean area and the Iberian Peninsula, together with a drying of these regions may lead to a shift in vegetation. Trees species in this are adapted to hot and dry summers, but their resistance differs between species. Less droughtadapted species might disappear or shift north and to higher altitudes, while better adapted species may expand their distribution range. A temperature increase over Scandinavia may favor a northern expansion of broadleaf species, that could possible outcompete the native 28 CHAPTER II. UNCERTAINTIES IN THE FUTURE CLIMATE needleleaf trees in that region. 3 Future European temperature change uncertainties reduced by using land heat flux observations GEOPHYSICAL RESEARCH LETTERS, VOL. 40, 2242–2245, doi:10.1002/grl.50404, 2013 Future European temperature change uncertainties reduced by using land heat flux observations Annemiek I. Stegehuis,1 Adriaan J. Teuling,2 Philippe Ciais,1 Robert Vautard,1 and Martin Jung3 Received 4 January 2013; revised 16 March 2013; accepted 21 March 2013; published 28 May 2013. [1] The variability of European summer climate is expected to increase in the next century due to increasing levels of atmospheric greenhouse gases, likely resulting in more frequent and more extreme droughts and heatwaves. However, climate models diverge on the magnitude of these processes, due to land-surface coupling processes which are difficult to simulate, and poorly constrained by observations. Here we use gridded observation-based sensible heat fluxes to constrain climate change predictions from an ensemble of 15 regional climate models. Land heat flux observations suggest that temperature projections may be underestimated by up to 1 K regionally in Central to Northern Europe, while slightly overestimated over the Mediterranean and Balkan regions. The use of observation-based heat flux data allows significant reductions in uncertainty as expressed by the model ensemble spread of temperature for the 2071–2100 period. Maximal reduction is obtained over France and the Balkan with values locally reaching 40%. Citation: Stegehuis, A. I., A. J. Teuling, P. Ciais, R. Vautard, and M. Jung (2013), Future European temperature change uncertainties reduced by using land heat flux observations, Geophys. Res. Lett., 40, 2242–2245, doi:10.1002/grl.50404. 1. Introduction [2] The greenhouse gases forced increase in radiative forcing causes global warming. However, not all regions of the globe are expected to warm at the same rate. Europe is a sensitive region where natural climate variability is large, and land-atmosphere coupling is important in large parts of the continent. In Central-Western Europe, a projected increase in summer temperature and temperature variability [Schär et al., 2004; Seneviratne et al. 2006; Fischer et al. 2007] has been attributed to an increase in this coupling. In this study, we hypothesize that any summer anomaly in the partitioning of the land surface energy balance in the current climate will be amplified on average under (drier) future summer climate conditions. It follows that information to reduce uncertainty on the land surface energy balance partitioning in the current climate may help to constrain future climate projections [Fischer et al. 2012]. This type of approach (see Hall and Qu [2006]) is topical because the magnitude of land surface feedback processes is uncertain and differs considerably between climate models because the parameterization of these models are diverse, and rather poorly constrained. Most commonly, model ensemble means and spread are used to quantify model ensemble predictions of future climate variables. [3] A method that leads to local reduction in uncertainty in the future by using present-day observations and models was proposed by Hall and Qu [2006]. They linked empirically the seasonal cycle of present-day snow albedo feedback to the magnitude of this feedback in the future. Following a similar rationale, Boberg and Christensen [2012] showed that inter-climate models’ future summer temperature predictions over Europe were related to the present-day temperature bias of models, and this property was used to constrain temperature predictions in regions with dry and warm climate such as the Mediterranean. Quesada et al. [2012] reduced uncertainty in future European summer climate projections by the CMIP3 global climate model (GCM) ensemble by selecting a subset of models that best represent the current relationships between spring precipitation and summer temperature. They found that the CMIP3 models underpredict summer temperatures in Central Europe. However, constraining future projections using surface heat fluxes observations (latent or sensible) has not been done, and yet these fluxes controlling land-atmosphere feedbacks were identified to be a main source of model spread and uncertainty [De Noblet-Ducoudré et al., 2012]. [4] A method to improve future projections based upon heat flux measurements requires long-term observations, which are generally not available; the FLUXNET network of sites where latent and sensible heat is measured online only contains rather few multi-annual time series. Here we used a recent gridded observation-based dataset of sensible heat fluxes (hereafter H) based on the time and space extrapolation of point-scale FLUXNET observations by machine-learning algorithm [Jung et al., 2009]. The purpose of this study is to investigate the use of the Jung et al observation-based land sensible heat flux dataset for selecting the most realistic climate models among an ensemble of future regional climate model (RCM) projections over Europe. 1 LSCE/IPSL, Laboratoire CEA/CNRS/UVSQ, Gif-sur-Yvette, France. Hydrology and Quantitative Water Management Group, Wageningen University, The Netherlands. 3 Max-Planck Institute for Biogeochemistry, Jena, Germany. 2 Corresponding author: A.I. Stegehuis, LSCE/IPSL, Laboratoire CEA/CNRS/UVSQ, 91191 Gif-sur-Yvette CEDEX, France. (annemiek. [email protected]) ©2013. American Geophysical Union. All Rights Reserved. 0094-8276/13/10.1002/grl.50404 2. Data and Methods [5] 15 RCMs of the FP6 ENSEMBLES project are used [Hewitt and Griggs, 2004; van der Linden and Mitchell, 2009]. The RCMs are driven by GCMs used for the 4th IPCC assessment report for the A1B scenario. We use only those RCMs that were run for the period 1961–2099 or longer. All model outputs are provided on a 25 km spatial 2242 STEGEHUIS ET AL.: EUROPEAN TEMPERATURE CHANGE UNCERTAINTY ρ(H,ΔT) 1.0 0.8 0.6 0.4 0.2 0.0 −0.2 −0.4 Figure 1. Spatial distribution of the Pearson correlation coefficient between simulated summer sensible heat flux (W m2) and simulated temperature change (K) of 15 RCMs, defined by the difference in average air temperature between future and present-day periods. The black dot represents the grid point center location for Figure 2. resolution, but in this study, they have been projected onto a common 50 km resolution grid using bilinear interpolation to be compared to the observational-based data. Monthly mean model output are extracted for sensible heat flux at surface (Wm2) and 2-m mean temperature (K). The following GCMs are used from the institutes (in parentheses) HadCM3Q16 (C4I), ARPEGE (CNRM, DMI), ECHAM5-r3 (DMI, ICTP, KNMI, MPI, SMHI), BCM (DMI, SMHI), HadCM3Q0 (ETHZ, HC), HadCM3Q3 (HC, SMHI), and HadCM3Q16 (HC). [6] We used observationally derived gridded sensible heat fluxes [Jung et al., 2011] (hereafter MTE-H). This data product was obtained by applying a machine learning algorithm Model Tree Ensemble to estimate global monthly heat fluxes from 1982 to 2008 on a resolution of 0.5 (for a detailed description of the data product and underlying methodology, see Jung et al., [2009]). The spread (standard deviation, sH), between ensemble members being taken as a measure of uncertainty on observation-derived H. The advantage of using H instead of latent heat flux is that (1) H is more directly linked with air temperature, and (2) the measurement method by the eddy covariance technique is more direct than for latent heat fluxes [Wilson et al., 2002]. [7] The spread of simulated summer temperature change (ΔT, difference between 2071–2100 and 1971–2000) is rather large in the FP6 ENSEMBLES ensemble of RCMs. We will refer to this spread as the a priori spread (i.e., before adding information from flux observations). If a tight correlation exists between RCM-modeled ΔT and H (1971–2000) (Figure 1), then a better (a posteriori) estimate of ΔT could be obtained through the linear regression ΔT = aH + b at every grid point. The a posteriori uncertainty of ΔT then depends on the uncertainty of the observation (sH), the goodness-of-fit of the regression (through the residual spread sres), and the slope of the regression (a) (Figure 2a). The a posteriori uncertainty of ΔT, given sH, is given by (see Appendix A): sΔT jH ¼ (1) [8] Note that adding information on H (predictor) will only result in a reduction of the a priori uncertainty if the observational uncertainty sH is (much) smaller than the spread of RCM-modeled H. The a posteriori uncertainty of ΔT in equation (1) can be reduced by taking the standard error of the prediction of this correlation (see Appendix A). Where correlation is low, uncertainty reduction will only be minor. This also applies for a larger uncertainty in the observation-derived H data product (Figure 2b). (b) 0 0 2 2 4 4 6 6 8 8 (a) Summer ΔT (K) qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi s2res þ a2 s2H 0 20 40 60 80 100 Present day summer H (Wm−2) Figure 2. (a) Simulated summer sensible heat flux (W m2) versus simulated air temperature change (ΔΤ in K) in summer for the grid box indicated in Figure 1. Vertical lines indicate the MTE-H (solid) observation derived estimate, plus and minus one standard deviation (dashed). The gray area indicates the prediction interval for ΔΤ (sΔT|H) that is compatible with the MTE-H observations. The red lines show the linear regression between RCM-modeled ΔΤ and H (thick solid) and the 1-s prediction interval (dashed). (b) Estimated prior and posterior normal distributions with modeled mean and SD; SD of ΔΤ a priori (solid), SD of ΔΤ given a perfect H observation (dotted), and SD of ΔΤ a posteriori given uncertain H observations (sΔT|H) (dashed). 2243 STEGEHUIS ET AL.: EUROPEAN TEMPERATURE CHANGE UNCERTAINTY ΔT a priori (a) ΔT a posteriori a posteriori − a priori K (b) 6 K (c) 1.5 5 1.0 4 0.5 3 0.0 2 −0.5 −1.0 1 −1.5 σΔT a priori (d) 0 σΔT a posteriori a posteriori − a priori K (e) 2.0 (f) % 20 10 1.5 0 1.0 −10 −20 0.5 0.0 −30 −40 Figure 3. (a) Mean predicted summer temperature change (ΔΤ in K) of 15 RCMs, i.e., a priori simulations, (b) predicted temperature change that is compatible with observed sensible heat flux, i.e., a posteriori, and (c) the estimated RCM model ΔΤ bias, diagnosed from the difference between b and a (K). Hashed areas indicate a significant difference between observation-based H and modeled H. (d) Standard deviation of mean temperature change of 15 RCMs a priori, (e) standard error of predicted temperature change based on the subset of RCM models that are compatible with present-day sensible heat flux, i.e., a posteriori, and (f) the relative change of ΔΤ (s difference between d and e in percentage). 3. Results [9] We found that there is a significant positive correlation between simulated average summer H in the current climate and mean summer ΔT (Figure 1) which reflects the existence of a physical link between summer H and near-surface temperature change. Positive values are found over most of Europe, with exception of the Iberian Peninsula (Figure 1). The strongest positive correlations are found over central France and Hungary, regions which were also highlighted as strong coupling regions in other studies on soil moisturetemperature coupling [e.g., Seneviratne et al., 2006]. The low correlation over the Iberian Peninsula might indicate different controls than land-atmosphere coupling behind model differences in this region. [10] In general, the RCM-modeled ΔT (Figure 3a), is close to ΔT constrained by MTE-H (Figure 3b), leading us to diagnose only a small temperature bias when using MTE-H to restrain temperature change. Over central Europe and the southern part of Scandinavia, we found a positive ΔT bias (Figure 3c), locally up to 1 K, suggesting an underestimation of modeled H with regard to MTE-H. Over the southeastern part of Europe, however, we found a negative bias, suggesting that models have somewhat unrealistic drying, resulting into too low latent heat emissions (soil moisture limitations) and too high H [Stegehuis et al., 2012; Boberg and Christensen, 2012]. This is confirmed by subtracting mean simulated H from the observation based MTE-H (figure not shown). Areas with significant difference (a = 0.1) between the two are indicated in Figure 3c. [11] The ensemble spread of simulated temperature change is largest over France and the Balkan (Figure 3d), which are two regions with predicted high interannual variability of summer temperature due to a strong soil moisture-temperature coupling [Seneviratne et al., 2006; Hirschi et al., 2011]. Also, in these regions, the correlation between present-day H and temperature change was highest (suggesting a strong coupling, indirectly explained by soil moisture limitations on evapotranspiration), thus resulting in a smaller standard error of the prediction (Figure 3e). The difference between the a priori and a posteriori, the uncertainty reduction of temperature change predictions, is displayed in Figure 3 f. As can be seen, uncertainty (RCM model spread) is reduced by up to 40% in regions with highest correlations, by using H observations to constrain ensemble projections. This approach is only possible because we had access to an ensemble of RCM results with marked differences in model skills to reproduce presentday H, allowing to establish a strong positive relationship between H and temperature change, like in Figure 2. 4. Concluding Remarks [12] In this study, we proposed an approach to constrain summer temperature change predictions by using the existing correlation between simulated present-day H and temperature change in combination with present-day H observation-based gridded datasets over Europe. The fact that the uncertainty in observed H is smaller than the range spanned by the regional models output made it possible to reduce regional uncertainty of future climate change predictions. We chose to use sensible rather than latent heat flux as a predictor of future temperature change in the present study, because of the uncertainty between different observationbased data products [Mueller et al., 2011]. [13] Our results indicate that in Central and Northern Europe, the ENSEMBLES RCM projections underestimate future temperature change, but that they overestimate it in Mediterranean regions. In contrast, Fischer et al. [2012] found that an ensemble constrained by observed present-day interannual summer variability predicts lower temperature change over Central Europe than the ensemble of all models. While this could be caused by an overestimation of MTE-H, a 40% lower MTE-H would be required to obtain a similar reduction in ΔT. Using reanalysis data from ERA-interim 2244 STEGEHUIS ET AL.: EUROPEAN TEMPERATURE CHANGE UNCERTAINTY [Dee et al., 2011] instead of MTE-H as a reference dataset results in an overestimation of ΔT in Central to Northern Europe (like in Fischer et al. [2012]) of the ensemble of all models, and an underestimation over France (similar to MTE-H), the Balkan, and the Mediterranean regions, with on average twice the magnitude of MTE-H (not shown). Using our method with interannual summer variability instead of H results in a similar difference between a priori and a posteriori ΔT as found in Fischer et al. [2012], indicating that the difference is caused by the selection of datasets and not methodology. The apparent inconsistency can be explained by the fact that the relations used are statistical rather than physical, since correlation does not guarantee causality. [14] In the case of MTE-H constraint, the reduction of uncertainty in regional temperature change predictions has a heterogeneous amplitude. In regions where surface fluxes form a loose constraint on projections, other variables might be used. In addition to snow cover [Hall and Qu, 2006], temperature-precipitation relationships [Quesada et al., 2012], or summer temperature variability [Fischer et al., 2012], we hypothesize that other variables such as convective precipitation, soil moisture, moisture convergence, wind speed and direction, and indices of weather regimes that influence the transport of heat from North Africa can be used. In future studies, a multi-linear empirical regression approach could be developed where the target variable (temperature change) is empirically related to these different predictors. Since our study highlights the impact of land surface conditions on European summer temperatures, one key implication is that land cover and land cover change must be better accounted for by RCMs for robust predictions of the future summer climate in Europe. Appendix A [15] The results of the regression between H and ΔT can be written as: E ½ΔT jH ¼ aH þ b (A1) [16] With the residual standard deviation independent of H: E½ðΔT aH bÞ2 H ¼ s2res (A2) E ½H ¼ mH (A3) h i E ðH mH Þ2 ¼ s2H (A4) [17] Let: and and denote the a posteriori expectations of ΔT by: E½ΔT jmH ¼ mΔT jH 2 s2ΔT jH ¼ E ΔT mΔT jH (A5) mΔT jH ¼ amH þ b (A7) (A6) [18] With: [19] In rewriting (A6), the cross-term cancels out because E[ΔT aH b|H] = 0 [20] The conditional expectation E[(ΔT E[ΔT])2|H] = sΔT|H then becomes: sΔT jH ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi s2res þ a2 s2H (A8) [21] Acknowledgments. The ENSEMBLES data used in this work was funded by the EU FP6 Integrated Project ENSEMBLES (Contract number 505539) whose support is gratefully acknowledged. We acknowledge financial support from The Netherlands Organization for Scientific Research through Veni grant 016.111.002. This work was partly supported by the FP7 CARBOEXTREME project. We thank Paul Torfs for his help on the uncertainty propagation. References Boberg, F., and J. H. Christensen (2012), Overestimation of Mediterranean summer temperature projections due to model deficiencies, Nat. Clim. Change, 2, 433–436, doi:10.1038/NCLIMATE1454. Dee, D. P., et al. (2011), The ERA-Interim reanalysis: Configuration and performance of the data assimilation system, Q. J. R. Meteorol. Soc., 137, 553–597, doi:10.1002/qj.828. De Noblet-Ducoudré, N., et al. (2012), Determining Robust Impacts of Land-Use-Induced Land Cover Changes on Surface Climate over North America and Eurasia: Results from the First Set of LUCID Experiments, J. Climate, 25, 3261–3281. Fischer, E. M., S. I. Seneviratne, D. Luthi & C. Schär (2007) Contribution of land-atmosphere coupling to recent European summer heat waves. Geophy. Res. Lett., 34, L06707, doi:10.1029/2006GL029068. Fischer, E. M., J. Rajczak & C. Schär (2012) Changes in European summer temperature variability revisited. Geoph. Res. Lett., 39, L19702, doi:10.1029/2012GL052730. Hall, A. & X. Qu (2006) Using the current seasonal cycle to constrain snow albedo feedback in future climate change. Geoph. Res. Lett., 33, L030502, doi:10.1029/2005GL025127. Hewitt, C. D., and D. J. Griggs (2004), Ensembles-based predictions of climate changes and their impacts, Eos, 85, 566. Hirschi, M., S. I. Seneviratne, V. Alexandrov, F. Boberg, C. Boroneant, O. B. Christensen, H. Formayer, B. Orlowsky, and P. Stepanek (2011), Observational evidence for soil-moisture impact on hot extremes in southeastern Europe, Nature Geosci., 4, 17–21, doi:10.1038/NGEO1032. Jung, M., M. Reichstein, and A. Bondeau (2009), Towards global empirical upscaling of FLUXNET eddy covariance observations: Validation of a model tree ensemble approach using a biosphere model, Biogeosciences, 6, 2001–2013. Jung, M., et al. (2011) Global patterns of land-atmosphere fluxes of carbon dioxide, latent heat, and sensible heat derived from eddy covariance, satellite, and meteorological observations. J. Geophys. Res., 116, G00J07, doi:10.1029/2010JG001566. Mueller, B., et al. (2011) Evaluation of global observations-based evapotranspiration datasets and IPCC AR4 simulations. Geoph. Res. Lett., 38, L06402, doi:10.1029/2010GL046230. Quesada, B., R. Vautard, P. Yiou, M. Hirschi, and S. I. Seneviratne (2012), Asymmetric European summer heat predictability from wet and dry winters and springs, Nat. Clim. Change, 2, 736–741, doi:10.1038/NCLIMATE1536. Schär, C., P. L. Vidale, D. Luthi, C. Frei, C. Haberli, M. A. Liniger, and C. Appenzeller (2004), The role of increasing temperature variability in European summer heatwaves, Nature, 427, 332–336, doi:10.1038/ nature02300. Seneviratne, S. I., D. Luthi, M. Litschi, and C. Schär (2006), Landatmosphere coupling and climate change in Europe, Nature, 443, 205–209, doi:10.1038/nature05095. Stegehuis, A. I., R. Vautard, P. Ciais, A. J. Teuling, M. Jung, and P. Yiou (2012) Summer temperatures in Europe and land heat fluxes in observation-based data and regional climate model simulations. Clim. Dyn., in press, doi:10.1007/s00382-012-1559-x. Van der Linden, P., and J. F. B. Mitchell (Eds) (2009), ENSEMBLES: climate change and its impacts: summary of research and results from ENSEMBLES project, Met Office, Hadley Centre, Exeter. Wilson, K., et al. (2002), Energy balance closure at FLUXNET sites, Agric. For. Meteorol., 113, 223–243, doi:10.1016/S0168-1923(02)00109-0. 2245 4. DISCUSSION AND ADDITIONAL RESULTS 4 33 Discussion and additional results In this chapter we used the ENSEMBLES set of RCMs to analyze a few key aspects of the summer climate in Europe and its change with climate warming. We first investigated the changes in the land energy budget, precipitation and temperature over Europe under a global warming scenario of 2◦ C. Furthermore, with the modeled relationship between sensible heat flux and temperature change, the uncertainty of summer temperature change predictions were reduced by using an observation-based data set of sensible heat flux. While model ensembles are found to better represent observations than a single model (Dirmeyer et al., 2006), they might add different uncertainty. This will be discussed in section 4.1. In section 4.2 I will discuss the method that was used to reduce temperature change uncertainties. This is followed by a discussion on the 2◦ C limit. 4.1 Model ensemble An objective of this study was to use the differences in land-atmosphere feedbacks between models to reduce uncertainty in future temperature changes. While I did not have as objective to study each model in detail, it is important to keep in mind that within the ensemble, the models might not be completely independent. Although models of different institutes were used, they might share for example similar convective calculations or parameterizations which could cause similar biases (Knutti et al., 2013). Furthermore, the uncertainty in climate change projections comes from both the driving and the forced models. In this study, the 15 RCMs are driven by only 6 GCMs, which might also result in similar biases. To overcome and understand such biases, very extensive ensembles should be created and analyzed, in which the RCMs are driven by all different GCMs. Furthermore all models should be thoroughly studied and compared with one another to verify their independence. Whenever different models share similar parameterizations or parts of codes, weighting could be applied to avoid that similar biases are multiplied in the ensemble. 4.2 Uncertainty reduction method We used the differences between the models to reduce uncertainty in temperature prediction. The models that simulate a sensible heat flux closer to the observation-based heat flux, are considered more important than models that simulate a heat flux that is far from the observations. The latter model values are less accounted for in the calculation of the a posteriori temperature and uncertainty prediction. However, this method can be discussed. All models are used for the calculation of the relationship between present-day sensible heat flux and summer temperature change, and the models being far from the observations are not necessarily wrong when considering changes. They might be less credible. Conversely, models that well simulate the present climate, do not necessarily correctly simulate a future different climate (Oreskes et al., 1994). Therefore it can be argued that a method with model weighting might be better justified. To ascertain a robustness in our results, we applied Bayesian Model Averaging (BMA) (van Oijen et al., 2013b) on the same data sets. The probability of each model is calculated based on the observation-based values of sensible heat flux. This probability is then used to obtain a model ensemble weighted average. 34 CHAPTER II. UNCERTAINTIES IN THE FUTURE CLIMATE Figure II.4: (a) Mean predicted summer temperature change of 15 RCMs, i.e., a priori simulations, (b) predicted temperature change that is compatible with observed sensible heat flux, i.e., a posteriori based on Bayesian Model Averaging, and (c) the estimated RCM model temperature ∆T bias, diagnosed from the difference between b and a (K). (d) Standard deviation of mean temperature change of 15 RCMs a priori, (e) standard deviation of predicted temperature change based on BMA, and (f) the relative change of ∆T (σ difference between d and e in percentage). Although differences between both methods are found, the overall patterns of temperature change and uncertainty reduction are similar (compare Fig. II.3 with Fig. 3 from section 2.3). The estimated model ∆T bias differs most in Southern Europe. While BMA only shows a negative ∆T bias in some of the south eastern parts, the method used in section 2.3 had more negative and more widespread negative values. Most parts of Italy seemed to experience too much drying, whilst using the BMA weighting approach it seems to be more realistic. Over the Iberian Peninsula the pattern is opposite: models show not enough drying, or an underestimation of modeled H compared to the observations using BMA, while almost no bias was found previously. Regarding the uncertainty reduction of temperature predictions (Fig. II.3 f) most difference between the two methods is found in the magnitude of the reduction. BMA shows much more reduction, up till 80% (not shown in figure), whilst before the reduction showed maxima of 40%. The pattern is also slightly different. While previously a larger uncertainty (positive values) was found over Scandinavia and parts of Great Britain, BMA reduces the model prediction uncertainty also over parts of these regions. While some differences between the two methods exist, the overall similarity in the patterns improves our confidence in the previously obtained results. 5. SUMMARY AND CONCLUSIONS 4.3 35 The choice for the 2◦ C limit The 2◦ C global warming limit has now been agreed upon as target during the last United Nations Conference on Climate Change (COP21). However, this target has been debated by Hare et al. (2011); Knutti et al. (2016). They question its safety and argue that scientific evidence is lacking that legitimize this specific threshold. Figure II.4 shows the average European temperature for a control period (1971-2000), and that for both a 1.5◦ C and a 2◦ C global warming. Both future periods are GCM dependent and their details can be found in Appendix A. In spring and summer the temperature of the 2◦ C period is higher than that of the 1.5◦ C period. However, during autumn and winter the mean temperature is similar, while there is more spread in the 1.5◦ C period. This suggests non-linear warming and confirms the necessity of studying different ’warming limits’. For precipitation most differences between the two periods is found in winter, where there is significantly less precipitation change in the 1.5◦ C period. In summer there seems to be less drying compared to the 2◦ C period, but there is a large inter-model uncertainty. 5 Summary and conclusions In this chapter we showed some of the patterns on the European climate under a 2◦ C global warming. Furthermore we proposed an approach to constrain summer temperature projections by using the strong relationship between present-day sensible heat flux and summer temperature change together with present-day observation-based data. We showed that on average, a 2◦ C global temperature rise results in a higher European warming than 2◦ C. Precipitation changes strongly depend on the region, with highest increases over Scandinavia and drying over the Iberian Peninsula and the Mediterranean region. The reason for these changes may be due to a different partitioning of the land heat fluxes. During summer, sensible heat flux increases strongly in Southern Europe, while latent heat flux decreases over this region. This pattern is opposite in Northern Europe where the increase in latent heat flux may favor precipitation. The largest model differences for both heat fluxes lies in central Europe, where models do not agree on the sign of change. This translates into a large uncertainty in the simulated temperature projections. However, using observations of the land heat fluxes and their uncertainty, we were able to reduce the spread of models’ summer temperature change predictions regionally up to 40%. Yet, this approach relies on strong assumptions about the interpretation of model bias. In particular, whether a model with a smaller bias for the present climate will be more realistic in the future remains unproven. While temperature change will increase over entire Europe in the future, its magnitude and regional patterns remain uncertain. While land heat fluxes helped to constrain model spread in central Europe, in other regions different methods may need to be used, as the sensitivity of evapotranspiration to changes is soil moisture might be too small. Better projections of future temperature changes may be important for impact studies, for example for the estimation of vegetation distribution and subsequent management practices. When better predictions for climate change are available, management can be adapted for the conservation of species, especially those that are sensitive to elevated temperatures and grow on the southern limit of their distribution. For example for extensive thinning may help survive Pinus sylvestris and Fagus sylvatica on xeric sites (Giuggiola et al., 2013; van der CHAPTER II. UNCERTAINTIES IN THE FUTURE CLIMATE −5 0 5 TAS [°C] 10 15 20 36 1971−2000 1.5−degree period 2−degree period Spring (MAM) Summer (JJA) Autumn (SON) Winter (DJF) Figure II.5: Temperature over Europe per season from 15 RCM from the ENSEMBLES project (spatial average over land pixels) for the control period (1971-2000) (open bars), the 1.5◦ C period (light grey bars) and the 2◦ C period (dark grey bars). 37 0.0 −7.5 −5.0 −2.5 ∆PR [%] 2.5 5.0 7.5 10.0 5. SUMMARY AND CONCLUSIONS 1.5−degree period 2−degree period Spring (MAM) Summer (JJA) Autumn (SON) Winter (DJF) Figure II.6: Spatial average over land of changes in precipitation over Europe relative to the control period (1971-2000), together with the range of changes for the GCM-RCM ensemble (median, 25-75% range and min and max values). The open bars refer the 1.5◦ C period, the grey bars to the 2◦ C period. 38 Maaten, 2013). CHAPTER II. UNCERTAINTIES IN THE FUTURE CLIMATE C HAPTER III Atmospheric processes 39 40 1 CHAPTER III. ATMOSPHERIC PROCESSES Introduction In the previous chapter the focus was on land-atmosphere feedbacks in models and their influence on future summer temperatures. This chapter is concentrated on atmospheric processes and physics in models and their impact on summer temperatures and heatwaves. The main aim is to try to answer the question how different physical formulations influence simulated heatwaves and the spread and uncertainty of modeled summer temperatures. It is not self-evident that RCMs can well represent extreme heatwaves in regions where they have only been scarcely observed (Butler, 2003; Perkins and Fischer, 2013). Even if the physics of such events would be known, uncertainties could arise by the poor restriction by observations, leading to underestimation of maximum temperatures (Halenka et al., 2006). This would directly lead to too few simulated heatwaves. Vautard et al. (2013), however, found that the heatwave frequency and magnitude were overestimated in RCMs, especially in the Mediterranean region. But simultaneously they found a large spread between the models. The source of such a spread however, is not always clear. Among various models, different physics can be used. But changing parameters can cause different model outputs even if the physics are similar (Bellprat et al., 2012). Another source of uncertainty in RCMs can arise from boundary conditions, either from models or reanalyses data. Additionally, the dynamics created by the model might be a source of differentiation through natural chaotic variability which arises from the ’butterfly’ effect (Lorenz, 1963). In a study on climate extremes, Good et al. (2006) found that model projections are necessarily limited by uncertainties, particularly due to the model physics. Multi-physic ensembles have been used to reduce such this specific uncertainty (Awan et al., 2011). With this approach, a single model is used, but different physics are tested to find the most realistic combinations of physical formulations. Although this method is not unique (Fernández et al., 2007; Awan et al., 2011; Evans et al., 2012; García-Díez et al., 2013; Mooney et al., 2013), multi-physics ensembles have not yet been used for analyzing heatwaves. Yet this may be of high importance regarding their consequences and their expected elevated frequency and severity in the future European climate. In this chapter we use the Weather and Research Forecast model (WRF) (Skamarock et al., 2008) and test 216 different atmospheric physics combinations. They consist out of 3 radiation, 4 convection, 3 microphysics and 6 planet boundary-surface layer schemes. All possible combinations were used. Both the heatwaves of 2003 and 2010 were simulated together with the wetter summer of 2007. To focus on physics only, we nudge the model simulations above the atmospheric boundary layer. Although land-atmosphere interactions are known to play an important role in heatwaves, only the atmospheric physics were tested because of the limited choice between different land surface schemes at the time of the study. All 216 simulations were evaluated against different sets of observations and the best configurations, as compared to observations, were chosen to perform further studies. We found a large spread in simulated temperature, precipitation and incoming shortwave radiation, even though the simulations were nudged. Temperature differences were locally up to 10◦ C, with the largest uncertainties in central Europe, probably due to differences in landatmosphere feedback representations (Stegehuis et al., 2013a). Mean and maximum temperatures were systematically underestimated. Precipitation and shortwave radiation were mostly overestimated. Convection was found to dominate the ensemble spread, probably by 2. WRF MODEL DESCRIPTION 41 inducing uncertainties in clouds that affect both the surface energy and water budget prior to, and during the heatwaves. A small ensemble was selected, constrained by different observational data sets. We suggest that this model ensemble could be used for further research on heatwaves and summer temperatures in Europe. In the following section I describe the model and the used physics, which could not be included in the peer reviewed paper on the WRF multi-physics ensemble, presented in section III.3. In the discussion I focus especially on the convection physics, as it was shown to be most important in the model spread. The chapter is completed with a summary and concluding remarks in section 5. 2 WRF model description WRF (Skamarock et al., 2008) was developed by a collaboration between the National Center for Atmospheric Research (NCAR), the National Oceanic and Atmospheric Administration (represented by the National Centers for Environmental Prediction (NCEP) and the (then) Forecast Systems Laboratory (FSL)), the Air Force Weather Agency (AFWA), the Naval Research Laboratory, the University of Oklahoma, and the Federal Aviation Administration (FAA) (website WRF). It can both be used for weather forecasting as for climate research. In this study we used it for analyzing European summer heatwaves. WRF is a model platform that allows different physic packages to be interchanged. In version 3.3.1, the used version for this study, more than 50 different physics options were available. These cover microphysics, planet boundary layer physics, cumulus or convection physics, radiation physics, land surface and surface layer physics. The aim of this study is to analyze the spread due to model physics, and to find some that well represent European summer climate and heatwaves specifically. Due to the large number of available physics schemes however, it was impossible to test all combinations. Possible non-linear effects between different schemes advocates not only testing the different physics, but also the inter-physic interactions. We therefore chose to investigate 3 radiation, 4 convection, 3 microphysics and 6 planet boundary-surface layer schemes, with all possible combinations. The selection of schemes was made to avoid variants of the same physics and to maximize the difference of temperatures and precipitation outputs in preliminary experiments. However, as some physics gave unrealistic results, a subset was selected within a reasonable range of the observed variables (Fig. III.1). In this study, a total of 216 configurations was tested. A short description with the main characteristics and differences between the schemes is described below. 2.1 Microphysics The microphyics scheme plays an important role in cloud representation and non-convective rain (the convective rain is determined by the convection scheme). The cloud effects impact the radiation physics directly and the land surface is affected by the amount and distribution of non-convective rain. Differences between microphysics schemes can lay among others in their number of mass variables, number variables and formulation of different processes such as saturation and nucleation. The three schemes used in this study, WRF-SM6 (Hong and Lim, 2006), New Thompson (Thompson et al., 2008) and Morrison DM (Morrison et al., 42 CHAPTER III. ATMOSPHERIC PROCESSES Figure III.1: Maximum temperatures over Europe using different microphysics, PBL-surface layer, cumulus and land surface physics. The numbers correspond to the different options of the physics available in WRF. The black curve corresponds to the observations. 2009), have 6 particle classes: water vapor, clouds, rain, ice, snow and graupel. New Thompson and Morrison DM are both double-moment schemes, predicting both mass and number concentrations of the particles, while the single-moment WRF-SM6 only predicts the mass of the variables. Double-moment schemes may be more realistic but might be more difficult to validate because of their more specific cloud properties (Lee and Donner, 2011; Molthan and Colle, 2012). 2.2 Planetary Boundary Layer The Planetary Boundary Layer (PBL) physics describe entrainment, vertical diffusion and local and/or non-local mixing processes. They interact mostly with the surface layer physics, as it impacts surface temperature, wind and exchanges of momentum and moisture fluxes. The land heat fluxes from the surface physics affect the PBL. PBL schemes tested in this study are Yonsei Uni (YSU) (Hong et al., 2006), Mellor–Yamada–Janjic (MYJ) (Janjić, 1994, 2002), Quasi–Normal Scale Elimination (QNSE) (Sukoriansky et al., 2005), Mellor–Yamada 2. WRF MODEL DESCRIPTION 43 Nakanishi Niino (MYNN) (Nakanishi and Niino, 2006, 2009) and the Asymmetric Convection Model 2 Scheme (ACM2) (Pleim, 2007). Differences between schemes can come from among others the prognostic and diagnostic variables, the formulation for vertical mixing and closure schemes, and the (non)-locality. MYJ, QNSE and MYNN have local TKE-based vertical mixing, while YSU has non-local mixing. ACM2 describes both local and non-local mixing. Furthermore, YSU and ACM2 have first order closure, whilst MYJ and QNSE have 1.5-order closure and MYNN both 1.5- and second order closure schemes. This can lead for example to differences in PBL depth, which have been found to be too deep for both YSU and ACM2 and to shallow for MYJ, QNSE and MYNN (Cohen et al., 2015). 2.3 Surface Layer The surface layer physics are linked very tightly to the PBL as they describe the momentum, heat and moisture fluxes from the surface into the PBL. They are often coupled to specific PBL physics, and are therefore considered as pairs. Besides its relation with the PBL, they interact with the radiation scheme through short- and longwave radiation, and with the microphysics and the convection scheme through non-convective and convective rain, respectively. All four surface layer schemes are based on the Monin-Obukhov similarity theory. They differ among others in the calculation of roughness length. The four schemes used are: 1) MM5 (Beljaars, 1994) coupled with YSU and ACM2 PBL schemes; 2) ETA (Janjić, 1996), coupled with MYJ and MYNN; 3) QNSE (Sukoriansky et al., 2005), coupled with QNSE; and 4) MYNN (Nakanishi and Niino, 2006, 2009), coupled with MYNN. 2.4 Radiation Radiation physics are important for the calculation of the energy budget. Short- and longwave radiation are calculated as well as reflection and scattering. The radiation scheme interacts with the microphysics for the effects of clouds on radiation. It is connected with the surface layer for the downward radiation and upward surface emission and the albedo. It receives the cloud fraction from the convection scheme. In this study the radiation schemes for longwave and shortwave radiation are used simultaneously. The three used schemes are the Community Atmosphere Model (CAM) radiation (Collins et al., 2004), the RRTGM scheme (Iacono et al., 2008) and the New Goddard radiation (Chou and Suarez, 1999). Differences between the radiation physics can arise from e.g., a different interaction with cloud fractions, different spectral methods and a different number of spectral bands (8, 16 and 10 bands for longwave and 19, 14 and 11 bands for shortwave radiation in CAM, RRTGM and New Goddard, respectively) and its interaction with aerosols. Because of its effect on the radiation, these schemes can strongly affect temperature (Evans et al., 2012). 2.5 Convection Convection is very important in the coupling of land and atmosphere as it describes convective rainfall over land and water recycling and determines cloud cover. Furthermore it is involved in the heat, moisture and momentum transport and cloud cover. The four convection schemes studied are Kain-Fritsch (Kain, 2004), Grell-Devenyi (Grell and Freitas, 2014), Tiedtke (Tiedtke, 1989; Zhang et al., 2011) and the New Simplified Arakawa–Schubert (New SAS) scheme (Han and Pan, 2011). Differences between convective parameterizations can 44 CHAPTER III. ATMOSPHERIC PROCESSES arise among others from different closure formulations, different triggering functions and differences in entrainment and detrainment. Some of the main biases of convection physics are a wrong diurnal cycle, missing cold pools, missing mesoscale circulations and a convection that is triggered too easily (Lynn et al., 1995; Hohenegger et al., 2009; Hohenegger and Stevens, 2013; Hourdin et al., 2013; Marsham et al., 2013). 2.6 Land surface The land surface scheme is the most important scheme for land-atmosphere interactions, as it describes the soil moisture, and thereby impacts the partitioning of the land heat fluxes and temperature (Jin et al., 2010; Zeng et al., 2011; Ramarohetra et al., 2015). In this study only the NOAH land surface scheme was used (Tewari et al., 2004). It has four soil layers where both soil moisture and soil temperature are predicted. It furthermore includes vegetation effects, snow cover and canopy moisture. In the used WRF version more options were available, but they were not appropriate for our study. Tests with the Rapid Update Cycle (RUC) (Benjamin et al., 2004) scheme revealed unrealistic results, and technical problems made us decide not to further use this scheme. A more detailed description of the choice of a single land surface scheme is presented in the next section. 3 An observation-constrained multi-physics WRF ensemble for simulating European mega heat waves Geosci. Model Dev., 8, 1–14, 2015 www.geosci-model-dev.net/8/1/2015/ doi:10.5194/gmd-8-1-2015 © Author(s) 2015. CC Attribution 3.0 License. An observation-constrained multi-physics WRF ensemble for simulating European mega heat waves A. I. Stegehuis1 , R. Vautard1 , P. Ciais1 , A. J. Teuling2 , D. G. Miralles3,4 , and M. Wild5 1 LSCE/IPSL, Laboratoire CEA/CNRS/UVSQ, Gif-sur-Yvette, France and Quantitative Water Management Group, Wageningen University, the Netherlands 3 Department of Earth Sciences, VU University Amsterdam, Amsterdam, the Netherlands 4 Laboratory of Hydrology and Water Management, Ghent University, Ghent, Belgium 5 ETH Zurich, Zurich, Switzerland 2 Hydrology Correspondence to: A. I. Stegehuis ([email protected]) Received: 8 September 2014 – Published in Geosci. Model Dev. Discuss.: 19 November 2014 Revised: 5 May 2015 – Accepted: 23 June 2015 – Published: Abstract. Many climate models have difficulties in properly reproducing climate extremes, such as heat wave conditions. Here we use the Weather Research and Forecasting (WRF) regional climate model with a large combination of different atmospheric physics schemes, in combination with the NOAH land-surface scheme, with the goal of detecting the most sensitive physics and identifying those that appear most suitable for simulating the heat wave events of 2003 in western Europe and 2010 in Russia. In total, 55 out of 216 simulations combining different atmospheric physical schemes have a temperature bias smaller than 1 ◦ C during the heat wave episodes, the majority of simulations showing a cold bias of on average 2–3 ◦ C. Conversely, precipitation is mostly overestimated prior to heat waves, and shortwave radiation is slightly overestimated. Convection is found to be the most sensitive atmospheric physical process impacting simulated heat wave temperature across four different convection schemes in the simulation ensemble. Based on these comparisons, we design a reduced ensemble of five well performing and diverse scheme configurations, which may be used in the future to perform heat wave analysis and to investigate the impact of climate change during summer in Europe. 1 Introduction An increasing number of simulations and studies project a higher frequency of several types of extreme weather events in the future (e.g., Schär et al., 2004; Meehl et al., 2004; Della-Marta et al., 2007; Beniston et al., 2007; Kuglitsch et al., 2010; Fischer and Schär, 2010; Seneviratne et al., 2012; Orlowsky and Seneviratne, 2012). Since summer heat waves are among the most impacting of such phenomena – threatening society and ecosystems – climate models used for future projections must provide accurate simulations of these phenomena, or at least their uncertainties should be documented. Even if climate models have been evaluated using observed weather in past decades, it is unclear whether they will be able to simulate extreme heat waves in future climates that may not have analogues in the historical record. At a minimum, models should be able to reproduce the conditions measured during recent extreme heat wave cases, some of them having been shown to be unprecedented when considering the climate over the past 5 or 6 centuries (Chuine et al., 2004; Luterbacher et al., 2010; García-Herrera et al., 2010; Barriopedro et al., 2011; Tingley and Huybers, 2013). Given the importance of forecasting summer heat waves well in advance, many studies have analyzed their predictability, which remains poor in seasonal forecasts. For instance, the 2003 European heat wave was not simulated realistically (neither timing nor intensity) by the operational European Centre for Medium-Range Weather Forecasts (ECMWF) system, but improvements were clear with Published by Copernicus Publications on behalf of the European Geosciences Union. 2 A. I. Stegehuis et al.: Multi-physics WRF ensemble for simulating European mega heat waves the use of a new land-surface hydrology, convection and radiation schemes (e.g., Weisheimer et al., 2011; Dole et al., 2011; Koster et al., 2010; van den Hurk et al., 2012). However, seasonal forecasting experiments do not straightforwardly allow for the assessment of a model’s physical processes underlying extreme temperatures during heat waves because it is difficult to separate model biases due to deficiencies in the model representation from sensitivity to initial conditions. These may inhibit the effect of the representation of physical processes in reproducing the exact atmospheric circulation when starting simulations at the beginning of the season. From a statistical perspective, extreme temperatures have been found to be reasonably well represented in global simulations of the current climate (IPCC, 2013), as well as in regional simulations (Nikulin et al., 2010). In recent regional modeling evaluation experiments, using an ensemble of state-of-the-art regional models guided by re-analysis at the boundaries of a European domain, summer extreme seasonal temperatures were shown to be simulated with biases in the range of a few degrees (Vautard et al., 2013). Individual mega heat waves (2003 in western Europe, 2010 in Russia) were reproduced by most models. However, it was difficult to infer whether these models could also simulate associated processes leading to the extreme heat waves. The exact same events with similar atmospheric flow and persistence could not be reproduced due to internal variability (internal degrees of freedom) of the models. A comprehensive assessment of simulations of recent mega heat waves has only been the object of a limited number of such studies. Process-oriented studies of high extreme temperatures over Europe have focused on land–atmosphere feedbacks (e.g., Seneviratne et al., 2006, 2010; Fischer et al., 2007; Teuling et al., 2009; Stegehuis et al., 2013; Miralles et al., 2014) because, beyond atmospheric synoptic circulation, these feedbacks are known to play an important role in summer heat waves. However, the sensitivity of simulated heat wave conditions to physical processes in models has not yet been explored in a systematic way. This could be important because error compensation among processes that involve land–atmosphere interactions, radiation and clouds may cause high temperatures for the wrong reasons (Lenderink et al., 2007). The goal of the present study is threefold. First we examine the ability of a regional climate model, Weather Research and Forecast (WRF, Skamarock et al., 2008), to simulate recent European mega heat waves with a number of different model configurations. Analysis of these experiments then allows understanding which physical parameterizations are prone to reproduce the build up of extreme temperatures and thus the need for carefully constraining them in order to simulate these events properly. Finally, using observational constraints of temperature, precipitation and radiation, we select a reduced ensemble of WRF configurations that best simulates European heat waves with different sets of physical Geosci. Model Dev., 8, 1–14, 2015 schemes combinations. This constrained multi-physics ensemble aims therefore at spanning a range of possible physical parameterizations in extreme heat wave cases while keeping simulations close to observations. Our multi-physics regional ensemble approach contrasts with the classical multi-model ensembles that are constructed by the availability of model simulations in coordinated experiments (see e.g., Déqué et al., 2007, and references therein) or by arbitrarily configured combinations of parameterizations selected by different groups using the same model system (García-Díez et al., 2015). In the latter ensemble, the lack of overall design strategy may lead the uncertainty estimation to be biased and the models to be farther from observations. In addition, the real cause of model spread is difficult to understand because of shortcomings in the representation of physical processes and their interactions. Regional perturbed-physics or multi-physics ensembles could help understand and constrain uncertainties more effectively, but so far they have been seldom explored. García-Díez et al. (2015) showed that even a small multi-physics ensemble confronted to several climate variable observations can help diagnose mean biases of a regional climate model. Bellprat et al. (2012) showed that a well-constrained perturbed-physics ensemble may encompass the observations. Their perturbedphysics ensemble was designed by varying the values of a number of free parameters and selecting only the configurations that were closest to the observations; however, the number of combinations of different physical parameterization schemes was limited to a total of eight different configurations. The WRF model offers several parameterization schemes for most physical processes, and is thus suitable for a multiphysics approach. In fact, a WRF multi-physics approach has been used in several studies (e.g., García-Díez et al., 2011; Evans et al., 2012; Awan et al., 2011; Mooney et al., 2013), also with its predecessor MM5, but not specifically to simulate extreme heat waves. Here we run an ensemble of 216 configurations of WRF physical parameterizations and compare each simulation with a set of observations of relevant variables in order to select a reduced set of five configurations that best represent European summer mega heat waves. The evaluation is made over the extreme 2003 and 2010 events. The ensemble is also evaluated for a more regular summer (2007) in order to test the model configurations under non-heat wave conditions. 2 2.1 Methods Simulations and general model setup We use the WRF version 3.3.1 and simulate the 3 summers (2003, 2007, 2010) using an ensemble of physics scheme combinations. We first test the time necessary to initialize the soil moisture on a limited number of cases. Soil conditions www.geosci-model-dev.net/8/1/2015/ A. I. Stegehuis et al.: Multi-physics WRF ensemble for simulating European mega heat waves 3 Table 1. Physics schemes used in this study (with references). All possible permutations are made, yielding a total of 216 simulations. The numbers in the table refer to the number the schemes have in the Weather Research and Forecasting (WRF) model. Microphysics (MP) PBL + surface (PBLSF) Radiation (RA) Convection (CU) Soil (6) WRF-SM6 (Hong and Lim, 2006) (1-1) Yonsei UniMM5 (Hong et al., 2006; Beljaars, 1994) (3) CAM (Collins et al., 2004) (1) Kain-Fritsch (Kain, 2004) (2) NOAH (Tewari et al., 2004) (8) New Thompson (Thompson et al., 2008) (2-2) MYJ-ETA (Janjic, 1994, 2002) (4) RRTMG (Iacono et al., 2008) (3) Grell–Devenyi (Grell and Devenyi, 2002) (10) Morrison DM (Morrison et al., 2009) (4-4) QNSE-QNSE (Sukoriansky et al., 2005) (5) Goddard (Chou and Suarez, 1999) (6) Tiedtke (Tiedtke, 1989; Zhang et al., 2011) (14) New SAS (Han and Pan, 2011) (5-2) MYNN-ETA (Nakanishi and Niino, 2006, 2009; Janjic, 2002) (5-5) MYNN-MYNN (Nakanishi and Niino, 2006, 2009) (7-1) ACM2-MM5 (Pleim, 2007; Beljaars, 1994) are initialized using the ERA-Interim (Dee et al., 2011) soil moisture and temperatures; thereafter, soil moisture and air temperature are calculated as prognostic variables by WRF. For the August 2003 case, we find that temperatures differ by less than 0.5 ◦ C among one another when starting experiments before 1 May. Thus, in the current study, each simulation is run from the beginning of May to the end of August for the years 2003, 2007 and 2010. The regional domain considered is the EURO-CORDEX domain (European Coordinated Downscaling Experiment; Jacob et al., 2014; Vautard et al., 2013) and the low-resolution setup of 50 km × 50 km (∼ 0.44◦ on a rotated lat–long grid) is used – note that Vautard et al. (2013) recently concluded that a higher spatial resolution did not provide a substantial improvement in heat wave simulations. We use a vertical resolution with 32 levels for WRF. Boundary conditions come from ERA-Interim, including sea surface temperatures, initial snow cover, and soil moisture and temperature. In order to focus on physical processes in the boundary layer and the soil–atmosphere interface, and to avoid chaotic evolution of large-scale atmospheric circulation, we constrain the model wind fields with ERA-Interim re-analyses above model level #15 (about 3000 m), similar to previous studies (Vautard et al., 2014), using grid nudging, with a relaxation coefficient of 5.10−5 s−1 corresponding to a relaxation time approximately equivalent to the input frequency (every 6 h) (Omrani et al., 2013). Tem- www.geosci-model-dev.net/8/1/2015/ perature and water vapor were not constrained, to allow feedbacks to fully develop. 2.2 Physics schemes We test 216 combinations of physics schemes. We consider different physics of the planetary boundary layer and surface layer (PBL; six schemes), microphysics (MP; three schemes), radiation (RA; three schemes) and of convection (CU; four schemes). For each type of scheme, a few options were selected among the ensemble of possibilities offered in WRF. The selection was made to avoid variants of the same scheme and to maximize the difference of temperature and precipitation outputs in preliminary experiments. At the time of study and model development stage, different land-surface schemes were available in WRF: five-layer thermal diffusion scheme (Dudhia, 1996), NOAH (Tewari et al., 2004), Rapid Update Cycle (RUC) (Benjamin et al., 2004) and Pleim–Xiu (Gilliam and Pleim, 2010). We decided however to only use one, the NOAH land-surface scheme in order to focus our study on atmospheric processes, while limiting the number of simulations, and because the NOAH scheme is the most widely used in WRF applications. This was also motivated by the poor performance and extreme sensitivity of the RUC land-surface scheme for the land latent and sensible heat fluxes as compared with local observations in 2003. It simulates strong latent heat fluxes in the beginning of the season Geosci. Model Dev., 8, 1–14, 2015 4 A. I. Stegehuis et al.: Multi-physics WRF ensemble for simulating European mega heat waves ATneu (47N,11E) − GRA LH (Wm−2) 250 200 150 100 50 0 DEtha (51N,4E) − ENF ● ● ● 300 ● ● ● FLUXNET ● ●● ● ● NOAH ● ● ●● ● ● RUC ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 120 140 160 180 200 220 300 250 200 150 100 50 0 240 120 140 ATneu (47N,11E) − GRA 200 SH (Wm−2) 150 100 50 0 −50 −100 140 160 180 200 220 200 EF (Wm−2) 1.5 1.0 0.5 0.0 150 100 50 0 −50 −100 240 140 160 180 200 200 220 250 200 150 100 50 0 240 120 140 120 140 160 180 200 220 240 2.0 1.5 1.0 0.5 0.0 220 120 140 160 180 200 180 200 220 240 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 150 100 50 0 −50 −100 240 120 140 160 180 200 220 240 DKsor (55N,11E) − DBF ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● Time (DOY) 160 DKsor (55N,11E) − DBF 200 DEtha (51N,4E) − ENF ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 120 180 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ATneu (47N,11E) − GRA 2.0 160 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 300 DEtha (51N,4E) − ENF ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 120 DKsor (55N,11E) − DBF ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 220 240 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 2.0 1.5 1.0 0.5 0.0 120 Time (DOY) 140 160 180 200 220 240 Time (DOY) Figure 1. Time series of daily land heat fluxes in 2003 from May to the end of August on three different FLUXNET sites, with latent heat flux (LH) on the first row, sensible heat flux (SH) on the second row, and evaporative fraction (EF – latent heat flux divided by the sum of latent and sensible heat flux) on the last row (DOY is day of year). The three columns represent three sites, with Neustift/Stubai (Austria – ATneu 47◦ N, 11◦ E) in the first column, Tharandt (Germany – DETha, 51◦ N, 4◦ E) in the second, and Soroe-LilleBogeskov (Denmark – DKsor, 66◦ N, 11◦ E) in the third column. Vegetation types on the three sites are respectively grassland (GRA), evergreen needleleaf forest (ENF), and deciduous broadleaf forest (DBF). In gray all 216 simulations with the NOAH scheme. Observational data is shown in black (FLUXNET). The green line is one configuration with NOAH, while the blue line represents the same configuration but with RUC instead of NOAH. and an extreme drying at the end, while sensible heat flux is overestimated. The NOAH scheme appeared more realistic and robust in the tests that were done for capturing both latent and sensible heat fluxes during the 2003 heat wave at selected flux tower sites in western Europe (Fig. 1). Furthermore, the Pleim–Xiu scheme is especially recommended for retrospective air quality simulations and is developed with a specific surface layer scheme as coupled configuration (Gilliam and Pleim, 2010). The last possible option is the five-layer thermal diffusion scheme (Dudhia, 1996) which predicts ground and soil temperatures but no soil moisture and is therefore also not suitable for our study. Table 1 describes the physical Geosci. Model Dev., 8, 1–14, 2015 schemes that were combined to simulate the weather over the 3 summer seasons. 2.3 Observational data In order to evaluate the ensemble and to rank and select its best-performing simulations, we use gridded observed daily temperature and precipitation from E-OBS with a 0.25◦ resolution (version 7.0) (Haylock et al., 2008). Bilinear interpolation is used to regrid E-OBS data and the model output to the same grid. Furthermore, we use station data of monthly global radiation from the Global Energy Balance Archive (GEBA) network (Wild et al., 2009). For France 2003 the www.geosci-model-dev.net/8/1/2015/ 20 30 40 50 60 70 A. I. Stegehuis et al.: Multi-physics WRF ensemble for simulating European mega heat waves −20 −10 0 10 20 30 Figure 2. Domains used in this study: France, Iberian Peninsula, Russia and Scandinavia. data of 21 stations were available, for 2007 this number was 20. Observations over Russia were too scarce and have therefore not been considered. Model data are interpolated to these stations using the nearest neighbor method. In addition, in order to check land–atmosphere fluxes and the partitioning of net radiation into sensible and latent heat fluxes, we use the satellite observation-driven estimates of daily latent heat fluxes from GLEAM (Miralles et al., 2011). Since the latter is not a direct measurement we do not use them to rank the model configurations. Furthermore, latent and sensible heat flux measurements are used from three FLUXNET sites (Neustift/Stubai – Austria (Wohlfahrt et al., 2010); TharandtAnchor station – Germany (Grünwald and Bernhofer, 2007); and Soroe-LilleBogeskov – Denmark (Pilegaard et al., 2011), from the Carbon-Extreme database), for the evaluation of the land-surface schemes. 2.4 Evaluation and ranking of model simulations For ranking, we set up several measures of model skill based on the differences between observed and simulated spatial averages over two domains: France for 2003 and 2007 (5◦ W–5◦ E, 44–50◦ N), and Russia for 2007 and 2010 (25– 60◦ E and 50–60◦ N) (Fig. 2). A first scheme selection is made based on the skill to reproduce air temperature dynamics, since this is the primary impacted variable, while corresponding observations are reliable. Because we are interested in heat waves, we select only those simulations that are within a 1 K regional average difference between simulated and observed temperature, for heat wave periods; these pewww.geosci-model-dev.net/8/1/2015/ 5 riods are defined as 1–15 August for France (in 2003), and 1 July till 15 August for Russia (in 2010). The 1 K threshold was arbitrarily chosen and is used to avoid processing a large number of simulations that have unrealistic temperatures. Only 55 of the 216 simulations meet this criterion and are further considered. Then, the ranking of the retained simulations is done based on (i) the daily temperature difference between simulations and observations during the heat wave periods (as above for 2003 and 2010), and during the period 1–31 August for the normal year 2007; and (ii) the root mean square error of monthly precipitation and radiation for the months July, June and August. The GEBA data set only contains scarce radiation observations over Russia and therefore we could not consider this region for ranking models against incoming shortwave radiation. As a final step, an overall ranking is proposed by averaging the ranks obtained from the three variables (temperature, precipitation and radiation). From this final ranking, and in order to select an elite of multi-physics combinations, we selected the top-five highest-ranked configurations. Note that observational uncertainty is not considered in this study, which was shown to potentially impact model ranking over Spain (Gomez-Navarro et al., 2012). 3 Results 3.1 Large systematic errors found during heat wave periods Figure 3 shows the large temperature range spanned by the 216 ensemble members for the spatial average over the heat wave areas. The min–max range between ensemble members is up to 5 ◦ C during heat wave periods (Fig. 3). Locally, at 50 km resolution, the difference between the warmest and the coldest simulation during a heat wave is larger, reaching more than 10 ◦ C in 2003 (Fig. 3d). In 2007, when summer temperatures were not extreme, the range is about twice as small. Only a few simulations match the observed high temperatures (Fig. 3a–c). In Fig. 3a, we select two extreme configurations (blue and red lines) based on daily mean temperature over France during the 2003 heat wave. Interestingly, they are extreme in all regions and years, indicating that each configuration tends to induce a rather large systematic bias. This bias, however, is different for the “warm” and the “cold” configurations. It seems not to be due to a misrepresentation of the diurnal cycle, since they remain when analyzing time series of maximum and minimum daily temperatures independently (see Fig. S1a–f in the Supplement). However, minimum temperatures show a less consistent bias than maximum daily temperatures. A systematic temperature underestimation by WRF simulations over Europe has also been found in other multi-physics ensemble studies over Europe (e.g., Awan et al., 2011; García-Díez et al., 2011, 2015). Geosci. Model Dev., 8, 1–14, 2015 6 A. I. Stegehuis et al.: Multi-physics WRF ensemble for simulating European mega heat waves FRANCE 2003 (a) Temperature (°C) 25 20 15 10 5 120 140 160 180 FRANCE 2007 (b) 30 ●● ●● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● 200 220 25 Temperature (°C) 30 20 15 10 5 240 ●● ●● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● 120 140 160 Time (Day Of Year) RUSSIA 2010 (c) 30 Temperature (°C) 25 20 15 10 5 140 160 180 200 220 240 (d) K ●● ●● ●● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● 120 180 Time (Day Of Year) 200 220 70 10 65 60 8 55 6 50 4 45 2 40 35 240 −20 0 20 40 Time (Day Of Year) Figure 3. Time series of daily mean temperature over France in 2003 (a) and 2007 (b) and Russia in 2010 (c). Every simulation is shown in gray and observations of E-OBS in black. The blue and red lines are the coldest and the warmest simulations over France during the heat wave. These lines have the same set of physics in all the figures (3, 4, 5). Panel (d) shows the simulated temperature min–max range during the heat wave of 2003 (1–15 August). The range is calculated as the difference between the warmest simulation and the coldest simulation during the heat wave period between the 216 members of the ensemble. 5 4 4 Precipitation (mm) Precipitation (mm) 5 3 ● 2 ● ● ● ● ● ● ● 1 ● FRANCE 2007 (b) ● ● 5 4 ● ● 3 ● ● ● ● ● 2 ● ● ● ● July 2 ● 1 ● ● ● ● ● August ● ● 1 ● ● ● 0 June 3 ● 0 May RUSSIA 2010 (c) Precipitation (mm) FRANCE 2003 (a) 0 May June July August May June July August Figure 4. Monthly precipitation over France in 2003 (a) and 2007 (b) and Russia 2010 (c). The box plots show the extremes, 25th, 50th, and 75th percentiles. The blue and red dots are the coldest and the warmest simulations over France during the heat wave (as in Fig. 3). For monthly precipitation we obtain a large range of simulated values, with most configurations overestimating monthly summer rainfall (JJA) during heat wave years, and to a lesser extent during the wetter 2007 season (Fig. 4a–c). This is in line with the findings reported by Warrach-Sagi Geosci. Model Dev., 8, 1–14, 2015 et al. (2013) and Awan et al. (2011) and with the overestimation of precipitation by many EURO-CORDEX models shown by Kotlarski et al. (2014). The two selected extreme configurations (based on temperature, as explained above) are reproducing precipitation overall without a major bias. www.geosci-model-dev.net/8/1/2015/ A. I. Stegehuis et al.: Multi-physics WRF ensemble for simulating European mega heat waves FRANCE 2003 (a) 24 Temperature (°C) 23 22 FRANCE 2003 (b) 25 ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● cu=1 cu=3 cu=6 cu=14 ● ● ● ● ● ● ● ● 20 ● ● 23 ●● ● ● ● ● ● ● ● ● ● ● ● ● mp=6 mp=8 mp=10 ● ● ●● ● ● 21 ● 24 Temperature (°C) 25 7 ● ● ● ● ● ● ● 22 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 21 ● ● ● ●● ● ● ● ● 20 ● ● ● ● ● ● ● ● ● ● 19 19 18 18 0.18 0.20 FRANCE 2003 ● ● ● ● ● ● ● ● ● ● ● Temperature (°C) ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● 22 ● ● ● ● ●● ● ●● ● ● ● ● ●● ● ● 20 0.24 FRANCE 2003 ● ● 24 0.22 ● ● ● ● ● ●● ● ● ● ● ● 23 ● 21 ra=3 ra=4 ra=5 ● ● 0.20 Soil moisture (m3m−3) (d) 25 ● ● 23 0.18 Temperature (°C) 24 0.24 Soil moisture (m3m−3) (c) 25 0.22 ● ● ● ● ● ● ● ● ● 22 bpl−sf=1−1 bpl−sf=2−2 bpl−sf=4−4 bpl−sf=5−2 bpl−sf=5−5 bpl−sf=7−1 ● ● ● ● ● ● ●● 21 ● ● ● ● ● 20 ● ● ● 19 19 18 18 0.18 0.20 0.22 Soil moisture (m3m−3) 0.24 0.18 0.20 0.22 Soil moisture (m3m−3) 0.24 Figure 5. Scatter plot of soil moisture content on 31 July, and temperature in August. Every point is one simulation. Different colors and symbols represent different physics for convection (CU) (a), microphysics (MP) (b), radiation (RA) (c) and planetary boundary layer–surface (PBL-SF) (d). This suggests that the temperature bias in these two extreme simulations is not explicitly caused by a misrepresentation of the atmospheric water supply from precipitation. However, soil moisture (the soil moisture over the whole column) does show a strong relation to temperature biases in model simulations. Figure 5a–d shows soil moisture at the end of July versus temperature in August 2003 for each model configuration. Configurations with low soil moisture level are associated with higher temperatures and vice versa, confirming the role of land–atmosphere feedbacks during heat waves, already pointed out by previous studies. This indicates that the evapotranspiration from spring to summer depleting soil moisture can be a critical process during summer for the development of heat waves, and that this process is not simply related to summer precipitation. For solar radiation, the mean differences between our simulations over France 2003 and 2007 reach approximately 100 W m−2 (Fig. 6a, b). Observations for France (black dots) are found below the median value of the simulations, so a slight overestimation of the ensemble is obtained. The first www.geosci-model-dev.net/8/1/2015/ (warmest) extreme configuration (red dot) is associated with an overestimated radiation of 10–50 W m−2 while the other (coldest, blue dot) extreme configuration exhibits an underestimated radiation by about the same amount. Since the warmest simulation agrees better with temperature observations than the coldest simulation, one may therefore suspect that it contains a cooling mechanism that partly compensates for the overestimated solar radiation. 3.2 Sensitivity of temperatures to physical parameterizations and sources of spread In order to identify the physics schemes to which the development of heat waves is most sensitive, we examine how resulting temperatures are clustered as a function of the scheme used. We find that the spread between all simulations – both in terms of temperature and soil moisture – is mostly due to the differences in convection scheme (clustering of dots with the same color in Fig. 5a). For instance the Tiedtke scheme (blue dots) systematically leads to higher temperatures and lower soil moisture, while the Kain–Fritsch scheme Geosci. Model Dev., 8, 1–14, 2015 8 A. I. Stegehuis et al.: Multi-physics WRF ensemble for simulating European mega heat waves FRANCE 2003 (a) FRANCE 2007 (b) 350 350 ● ● ● 300 Global radiation (Wm−2) Global radiation (Wm−2) 300 ● ● ● 250 ● ● ● ● ● 200 ● ● ● ● 250 ● ● ● ● ● 200 ● ● ● ● 150 150 May June July August May June July August Figure 6. Monthly radiation over France in 2003 (a) and 2007 (b); no radiation data being available in Russia for 2010. The box plots show the extremes, 25th, 50th, and 75th percentiles. The blue and red dots are the coldest and the warmest simulations over France during the heat wave (as in Fig. 3). Table 2. The five best-performing configurations of physics in ranked from the best to the fifth best. Microphysics PBL-surface Radiation Convection Soil Rank Morrison DM WRF-SM6 WRF-SM6 New Thompson New Thompson Yonsei Uni-MM5 MYNN-MYNN ACM2-MM5 MYNN-MYNN ACM2-MM5 RRTMG RRTMG Goddard RRTMG RRTMG Tiedtke Grell–Devenyi Tiedtke New SAS Tiedtke NOAH NOAH NOAH NOAH NOAH 1 2 3 4 5 (green dots) leads to wetter soils and lower temperatures, inhibiting heat waves. Microphysics and radiation schemes are also contributing to the spread of simulated temperature and soil moisture values (Fig. 5b, c), although their effect is less marked than for convection. Heat wave temperatures and soil moisture seem to be least sensitive to the planetary boundary layer and surface layer physics schemes. The sensitivity of the convection scheme in WRF has already been mentioned in previous studies (Jankov et al., 2005; Awan et al., 2011; Vautard et al., 2013; García-Díez et al., 2015). Note that the soil moisture simulated in early August 2003 is better correlated with preceding radiation than with precipitation (compare Figs. S2 and S3), indicating that the way clouds, and particularly convective clouds, affect radiation prior to the onset of heat waves is a major driver of the spread for the development of heat waves – higher radiation leading to drier soils and higher temperatures during heat waves. 3.3 A constrained reduced ensemble of best simulations Focusing only on the 55 selected simulations that differ less than 1 ◦ C from the observations during the heat waves, we apply the ranking method introduced in Sect. 2 based on temperature, precipitation and radiation model–observation comparison metrics. The five highest-ranked simulations are given in Table 2 and are actually the numbers 1–5 in Table S1 Geosci. Model Dev., 8, 1–14, 2015 in the Supplement. Figure 7a confirms the ranking by showing that these simulations also perform well in terms of temperature, during the months prior to the heat wave. The same is furthermore found for the years 2007 in France (Fig. S5) and 2010 in Russia (Fig. S4), and also for other regions such as the Iberian Peninsula and Scandinavia (Fig. S6a, d). The selected simulations however performed less well for precipitation over France in 2003 (Fig. 7b), but do not show a large overestimation of precipitation either. Precipitation over Russia for the five highest-ranked simulations does show good performance (Fig. S4b), as well as for other European regions (Fig. S6). The mean radiation of the ensemble of the five best simulations is closer to the GEBA observations than in the case of the original ensemble (Fig. 7c). Nonetheless, the better match of the reduced ensemble of the five highest-ranked simulations to the observations of temperature, precipitation and radiation is to a very large degree unsurprising: the selection was based on the fit to observations. However, it is still satisfactory to see that some simulations are capable of matching all three variables. Conversely, we also compare simulations against another key variable that was not used for evaluating and ranking simulations, namely the latent heat flux (Fig. 7d). Albeit somehow reduced compared to the full-ensemble spread, the spread of the five best simulations for the latent heat flux remains large over the whole period, on average between 50 and www.geosci-model-dev.net/8/1/2015/ A. I. Stegehuis et al.: Multi-physics WRF ensemble for simulating European mega heat waves 9 Table 3. Cross-comparison between France 2003 and Russia 2010. The (5, 10, 15, 20 and 25) best simulations, when only using one heat wave to select the best configurations and vice versa, are taken and compared with their ranking for the other heat wave. If there would be no correlation between the 2 years, the average ranking would lay approximately at half of the total number of simulations for both years that lay within a first selection of 1 K (column eight). In bold the rankings that are lower than this number. Because observations of radiation are lacking over Russia, we tested France with and without including radiation in the ranking. With radiation With radiation Without radiation Without radiation average rank Fr–Ru average rank Ru–Fr average rank Fr–Ru average rank Ru–Fr (a) Number of simulations 5 10 15 20 25 within 1 ◦ C 22.6 15.75 53 20.25 21.8 15.2 37 16.8 25.3 14.7 28.4 18.1 23.1 13 27.6 17 27.5 39.3 25.5 19.9 104 58 104 58 FRANCE 2003 25 20 15 10 5 120 140 160 180 FRANCE 2003 (b) ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ● ● 1 2 3 4 5 EOBS 200 220 ● 3.5 3.0 ● Precipitation (mm) 30 Temperature (°C) Average ranking of 5, 10, 15, 20 and 25 best simulations 1 2 3 4 5 EOBS 2.5 ● 2.0 ● ● ● 1.5 ● 1.0 ● ● ● July August 0.5 May 240 June Time (DOY) ● ● 1 2 3 4 5 GEBA 300 ● ● ● ● 250 ● ● ● ● FRANCE 2003 200 Latent heat flux (Wm−2) 350 Global radiation (Wm−2) (d) FRANCE 2003 (c) 150 100 50 200 0 May June July August ●● ●● ● ●● 1 ●● ●● 2 ●● ●● 3 ●●●●● ● 4●●● ● ●● ●● 5 ●● ●● ● GLEAM ● ●● ●● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ●● 120 140 160 180 200 220 240 Time (DOY) Figure 7. Daily time series of temperature (a) and latent heat flux (c), monthly time series of precipitation (b) and incoming shortwave radiation (d). Observations are shown in black and the five best-performing runs in colors. Gray lines indicate other simulations. All figures are a spatial average over France during summer 2003. 120 W m−2 (observed values are around 75 W m−2 ). However, during the 2003 heat wave over France, three of the five best simulations exhibit a close resemblance to the latent heat observations (approximately 5–10 W m−2 ) (Fig. 7d). The two simulations that are found to considerably overestimate latent heat flux by approximately 30–40 W m−2 (as www.geosci-model-dev.net/8/1/2015/ compared to GLEAM) are those that use a different convection scheme than the Tiedtke scheme. The overestimation of latent heat fluxes in these schemes is however not generalized for other regions and years (Figs. S4c, S5d, S6c, f–h), for which the latent heat flux was fairly well simulated within the range of uncertainty of GLEAM. Geosci. Model Dev., 8, 1–14, 2015 10 A. I. Stegehuis et al.: Multi-physics WRF ensemble for simulating European mega heat waves 1 −2 −1 0 Temperature (°C) 2 3 FRANCE 2003 ● ● ● ● ● ● ● ● ● ● ● ● ● 120 ● ● ● ● ● ● ● ● ● 140 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 160 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 180 ● ● ● ● ● ● ● ● ● ● ● ● 200 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 220 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 240 Time (DOY) Figure 8. Sensitivity test of the initialization of soil moisture. Difference between the perturbed simulations (red indicates 20 % reduction of initial soil moisture, blue 20 % enhancement) performed with the five highest-ranked configurations compared to their corresponding “control” simulations. The darkest lines refer to the simulation conducted with the best-ranked configuration (1), while descending color shade agrees with descending ranking (1–5). A cross-comparison for the years 2003 and 2010, that is, using only the 2010 heat wave to select schemes and verify the performance of the selected schemes over 2003 and vice versa, yields some promising results. Table 3 shows the average ranking of the best (5, 10, 15, 20 and 25) simulations. When only using one heat wave to select the best configurations, they all lie in the top-ranked half, and even higher in the ranking in the case of the 2010 heat wave over Russia being used to select the best configurations. This suggests that the selection based upon one heat wave in one region should also provide better simulations for other heat waves or heat waves in other areas, i.e., that the bias of a member of the WRF ensemble is not local but at least regional at the scale of western Europe. Geosci. Model Dev., 8, 1–14, 2015 4 Concluding remarks In this study we designed and analyzed a large multi-physics ensemble with the WRF model. It is made of all possible combinations of a set of different atmospheric physics parameterization schemes. They were evaluated for their ability to simulate the European heat waves of 2003 and 2010 using the regional climate model WRF based on temperature, precipitation and shortwave radiation. Even though the simulations were constrained by grid nudging, the multi-physics ensemble contained a large spread in temperature, precipitation and incoming shortwave radiation, the three variables we used to create an overall configuration ranking. Most simulations systematically underestimate temperature and overestimate precipitation during heat waves, a model pattern that was already found in previous studies dealing with much smaller ensembles (e.g., Awan et al., 2011; García-Díez et al., 2011; Warrach-Sagi et al., 2013). The spread among ensemble members is amplified during the two extreme heat www.geosci-model-dev.net/8/1/2015/ A. I. Stegehuis et al.: Multi-physics WRF ensemble for simulating European mega heat waves waves of study. Since we only considered a single landsurface scheme, it is possible that the ensemble spread would considerably increase when incorporating the uncertainty associated with modeling land-surface processes. Nevertheless, considering only atmospheric processes, the magnitude of the spread still reaches 5 ◦ C during the peak of the heat waves. We also showed that among atmospheric process parameterizations, the choice of a convection scheme appears to dominate the ensemble spread. We found indications that the large differences between convection schemes seem to occur mostly through radiation and therefore the way convective clouds affect the surface energy and water budget prior to and during heat waves. Changes in incoming radiation cause changes in evapotranspiration and therefore soil moisture, which may subsequently feed back on air temperature. From this ensemble, we selected a small sub-ensemble with the five best configurations of atmospheric physics schemes based on the fit to observations. These configurations capture well the temperature dynamics during the mega heat waves of France and Russia and they perform better than other configurations in other regions of Europe. In addition, they are consistent with independent latent heat flux data used for cross-validation. This indicates that the constraints set for the selection reduce the uncertainty across the whole European continent and point towards the creation of an optimized ensemble of WRF configurations specific for heat waves, with reduced error compensations. A sub-ensemble that outperforms a larger ensemble was also found by Herrera et al. (2010). The sub-ensemble based on mean precipitation showed better results for extreme precipitation as well. However, a limitation of this study is the use of only one land-surface scheme; the five selected WRF configurations may actually all be affected by systematic errors of the NOAH land-surface scheme. The importance of the selected land-surface scheme is further confirmed by the larger spread of the “best” ensemble for latent heat (in W m−2 ) than for shortwave radiation. In order to mimic radically different land-surface processes, sensitivity tests in which the initial absolute amount of soil moisture was artificially increased and decreased by 20 % all along the soil column have been conducted. Results confirm the sensitivity of the temperature simulations to soil moisture, a variable partly controlled by the land-surface scheme (Fig. 8). The full answer to this question is left for a future study in which different atmospheric schemes and surface schemes will be jointly permuted. Although our ensemble is trained on only summer conditions, our results have several implications for climate modeling. First, the constrained WRF ensemble may be used in future studies of climate change; each of the five members may exhibit a different sensitivity to future climate change conditions, leading to a constrained exploration of the uncertainty. Then it is important to notice that our study pinpoints the need to carefully design or adjust the convection scheme www.geosci-model-dev.net/8/1/2015/ 11 for a proper representation of the summer climate during heat waves. This is particularly important in order to evaluate the impacts of climate change on ecosystems, health, carbon cycle, water and cooling capacity of thermal energy plants, since heat waves in the mid latitudes are expected to be one of the most impacting phenomena in a human-altered climate. Therefore, impact studies can be designed based on the selected configurations. The Supplement related to this article is available online at doi:10.5194/gmd-8-1-2015-supplement. Acknowledgements. A. I. Stegehuis acknowledges CEA for funding as well as the GHG-Europe FP7 project. A. J. Teuling acknowledges financial support from the Netherlands Organisation for Scientific Research through Veni grant 016.111.002. P. Ciais acknowledges support of the ERC-SYG project P-IMBALANCE. The authors acknowledge K. Pilegaard, A. Ibrom, C. Bernhofer, G. Wohlfahrt and CarboEurope for sharing FLUXNET data. We would like to thank the reviewers for their useful comments and suggestions for improving the manuscript. Edited by: A. Colette References Awan, N. K., Truhetz, H., and Gobiet, A.: Parameterization-induced error characteristics of MM5 and WRF operated in climate mode over the Alpine region: an ensemble-based analysis, J. Climate, 24, 3107–3123, doi:10.1175/2011JCLI3674.1, 2011. Barriopedro, D., Fischer, E. M., Luterbacher, J., Trigo, R., and Garcia-Herrera, R.: The hot summer of 2010: redrawing the temperature record map of Europe, Science, 332, 220–224, doi:10.1126/science.1201224, 2011. Beljaars, A. C. M.: The parameterization of surface fluxes in largescale models under free convection, Q. J. Roy. Meteorol. Soc., 121, 255–270, 1994. Bellprat, O., Kotlarski, S., Luthi, D., and Schär, C.: Exploring perturbed physics ensembles in a regional climate model, J. Climate, 25, 4582–4599, doi:10.1175/JCLI-D-11-00275.1, 2012. Beniston, M., Stephenson, D. B., Christensen, O. B., Ferro, C. A. T., Frei, C., Goyette, S., Halsnaes, K., Holt, T., Jylha, K., Koffi, B., Palutikof, J., Scholl, R., Semmler, T., and Woth, K.: Future extreme events in European climate: an exploration of regional climate model projections, Clim. Change, 81, 71–95, doi:10.1007/s10584-006-9226-z, 2007. Benjamin, S. G., Grell, G. A., Brown, J. M., and Smirnova, T. G.: Mesoscale weather prediction with RUC hybrid isentropicterrain-following coordinate model, Mon. Weather Rev., 132, 473–494, 2004. Chou, M.-D. and Suarez, M. J.: A solar radiation parameterization for atmospheric studies, NASA Tech. Memo 104606, 40, Greenbelt, Maryland, 1999. Geosci. Model Dev., 8, 1–14, 2015 12 A. I. Stegehuis et al.: Multi-physics WRF ensemble for simulating European mega heat waves Chuine, I., Yiou, P., Viovy, N., Seguin, B., Daux, V., and Ladurie, E. L.: Historical phenology: grape ripening as a past climate indicator, Nature, 432, 289–290, doi:10.1038/432289a, 2004. Collins, W. D., Rasch, P. J., Boville, B. A., Hack, J. J., McCaa, J. R., Williamson, D. L., Kiehl, J. T., Brieglib, B., Bitz, C., Lin, S.-J., Zhang, M., and Dai, Y.: Description of the NCAR Community Atmosphere Model (CAM 3.0), NCAR Tech. Note NCAR/TN464+STR. 214 pp., 2004. Dee, D. P., Uppala, S. M., Simmons, A. J., Berrisford, P., Poli, P., Kobayashi, S., Andrae, U., Balmaseda, M. A., Balsamo, G., Bauer, P., Bechtold, P., Beljaars, A. C. M., van de Berg, L., Bidlot, J., Bormann, N., Delsol, C., Dragani, R., Fuentes, M., Geer, A. J., Haimberger, L., Healy, S. B., Hersbach, H., Hólm, E. V., Isaksen, L., Kållberg, P., Köhler, M., Matricardi, M., McNally, A. P., Monge-Sanz, B. M., Morcrette, J. J., Park, B. K., Peubey, C., de Rosnay, P., Tavolato, C., Thépaut, J. N., and Vitart, F.: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system, Q. J. Roy. Meteorol. Soc., 137, 553– 597, 2011. Della-Marta, P. M., Haylock, M. R., Luterbacher, J., and Wanner, H.: Doubled length of western European summer heat waves since 1880, J. Geophys. Res., 112, D15103, doi:10.1029/2007JD008510, 2007. Déqué, M., Rowell, D. P., Luthi, D., Giorgi, F., Christensen, J. H., Rockel, B., Jacob, D., Kjellstrom, E., de Castro, M., and van den Hurk, B.: An intercomparison of regional climate simulations for Europe: assessing uncertainties in model projections, Clim. Change, 81, 53–70, doi:10.1007/s10584-006-9228-x, 2007. Dole, R., Hoerling, M., Perlwitz, J., Eischeid, J., Pegion, P., Zhang, T., Quan, X.-W., Xu, T. Y., and Murray, D.: Was there a basis for anticipating the 2010 Russian heat wave?, Geophys. Res. Lett., 38, L06702, doi:10.1029/2010GL046582, 2011. Dudhia, J.: A multi-layer soil temperature model for MM5. Sixth Annual PSU/NCAR Mesoscale Model Users’ Workshop, Boulder CO, July 1996, 49–50, 1996. Evans, J. P., Ekstrom, M., and Ji, F.: Evaluating the performance of a WRF physics ensemble over South-East Australia, Clim. Dynam., 39, 1241–1258, doi:10.1007/s00382-011-1244-5, 2012. Fischer, E. M. and Schär, C.: Consistent geographical patterns of changes in high-impact European heat waves, Nat. Geosci., 3, 398–403, 2010. Fischer, E. M., Seneviratne, S. I., Luthi, D., and Schär, C.: Contribution of land-atmosphere coupling to recent European summer heat waves, Geophy. Res. Lett., 34, L06707, doi:10.1029/2006GL029068, 2007. García-Díez, M., Fernández, J., Fita, L., and Yague, C.: Seasonal dependence of WRF model biases and sensitivity to PBL schemes over Europe, Q. J. Roy. Meteorol. Soc., 139, 501–514, doi:10.1002/qj.1976, 2011. García-Díez, M., Fernandez, J., and Vautard, R.: An RCM multi-physics ensemble over Europe: multi-variable evaluation to avoid error compensation, Clim. Dynam., in press, doi:10.1007/s00382-0152529-x, 2015. García-Herrera, R., Diaz, J., Trigo, R. M., Luterbacher, J., and Fischer, E. M.: A review of the European summer heat wave of 2003, Crit. Rev. Env. Sci. Tec., 40, 267–306, doi:10.1080/10643380802238137, 2010. Geosci. Model Dev., 8, 1–14, 2015 Gilliam, R. C. and Pleim, J. E.: Performance assessment of new land surface and planetary boundary layer physics in the WRF-ARW, J. Appl. Meteorol. Climatol., 49, 760–774, 2010. Gomez-Navarro, J. J., Montávez, J. P., Jerez, S., Jimenez-Guerrero, P., and Zorita, E.: What is the role of the observational dataset in the evaluation and scoring of climate models?, Geophys. Res. Lett., 39, L24701, doi:10.1029/2012GL054206, 2012. Grell, G. A. and Devenyi, D.: A generalized approach to parameterizing convection combining ensemble and data assimilation techniques, Geophys. Res. Lett., 29, 1693, doi:10.1029/2002GL015311, 2002. Grünwald, T. and Bernhofer, C.: A decade of carbon, water and energy flux measurements of an old spruce forest at the Anchor Station Tharandt, Tellus, 59B, 387–396, 2007. Han, J. and Pan, H.: Revision of convection and vertical diffusion schemes in the NCEP Global Forecast System, Wea. Forecasting, 26, 520–533, 2011. Haylock, M. R., Hofstra, N., Klein Tank, A. M. G., Klok, E. J., Jones, P. D., and New, M.: A European daily highresolution gridded data set of surface temperature and precipitation for 1950–2006, J. Geophys. Res., 113, D20119, doi:10.1029/2008JD010201, 2008. Herrera, S., Fita, L., Fernández, J., and Gutierrez, J. M.: Evaluation of the mean and extreme precipitation regimes from the ENSEMBLES regional climate multimodel, J. Geophys. Res., 115, D21117, doi:10.1029/2010JD013936, 2010. Hong, S.-Y. and Lim, J.-O. J.: The WRF single-moment 6-class microphysics scheme (WSM6), J. Korean Meteor. Soc., 42, 129– 151, 2006. Hong, S.-Y., Noh, Y., and Dudhia, J.: A new vertical diffusion package with an explicit treatment of entrainment processes, Mon. Weather Rev., 134, 2318–2341, 2006. Iacono, M. J., Delamere, J. S., Mlawer, E. J., Shephard, M. W., Clough, S. A., and Collins, W. D.: Radiative forcing by long-lived greenhouse gases: Calculations with the AER radiative transfer models, J. Geophys. Res., 113, D13103, doi:10.1029/2008JD009944, 2008. IPCC: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, edited by: Stocker, T. F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S. K., Boschung, J., Nauels, A., Xia, Y., Bex, V., and Midgley, P. M., Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 1535 p., doi:10.1017/CBO9781107415324, 2013. Jacob, D., Petersen, J., Eggert, B., Alias, A., Christensen, O. B., Bouwer, L. M., Braun, A., Colette, A., Deque, M., Georgievski, G., Georgopoulou, E., Gobiet, A., Menut, L., Nikulin, G., Haensler, A., Hempelmann, N., Jones, C., Keuler, K., Kovats, S., Kroner, N., Kotlarski, S., Kriegsmann, A., Martin, E., Van Meijgaard, E., Moseley, C., Pfeifer, S., Preuschmann, S., Radermacher, C., Radtke, K., Rechid, D., Rounsevell, M., Samuelsson, P., Somot, S., Soussana, J. F., Teichmann, C., Valentini, R., Vautard, R., Weber, B., and Yiou, P.: EURO-CORDEX: new-highresolution climate change projections for European impact research, Reg. Environ. Change, 14, 563–578, 2014. Janjic, Z. I.: The Step–Mountain Eta Coordinate Model: Further developments of the convection, viscous sublayer, and turbulence closure schemes, Mon. Wea. Rev., 122, 927–945, 1994. www.geosci-model-dev.net/8/1/2015/ A. I. Stegehuis et al.: Multi-physics WRF ensemble for simulating European mega heat waves Janjic, Z. I.: Nonsingular implementation of the Mellor-Yamada Level 2.5 Scheme in the NCEP Meso model, NCEP Office Note No. 437, 61 p., National Centers for Environmental Prediction, College Park, MD, 2002. Jankov, I., Gallus, W. A., Segal, M., Shaw, B., and Koch, S. E.: The impact of different WRF model physical parameterizations and their interactions on warm season WCS rainfall, Wea. Forecasting, 20, 1048–1060, doi:10.1175/WAF888.1, 2005. Kain, J. S.: The Kain–Fritsch convective parameterization: An update, J. Appl. Meteor., 43, 170–181, 2004. Koster, R. D., Mahanama, S. P. P., Yamada, T. J., Balsamo, G., Berg, A. A., Boisserie, M., Dirmeyer, P. A., Doblas-Reyes, F. J., Drewitt, G., Gordon, C. T., Guo, Z., Jeong, J. H., Lawrence, D. M., Lee, W. S., Li, Z., Luo, L., Malyshev, S., Merryfield, W. J., Seneviratne, S. I., Stanelle, T., Van den Hurk, B. J. J. M., Vitart, F., and Wood, E. F.: Contribution of land surface initialization to subseasonal forecast skill: First results from a multi-model experiment, Geophys. Res. Lett., 37, L02402, doi:10.1029/2009GL041677, 2010. Kotlarski, S., Keuler, K., Christensen, O. B., Colette, A., Déqué, M., Gobiet, A., Goergen, K., Jacob, D., Lüthi, D., van Meijgaard, E., Nikulin, G., Schär, C., Teichmann, C., Vautard, R., Warrach-Sagi, K., and Wulfmeyer, V.: Regional climate modeling on European scales: a joint standard evaluation of the EUROCORDEX RCM ensemble, Geosci. Model Dev., 7, 1297–1333, doi:10.5194/gmd-7-1297-2014, 2014. Kuglitsch, F. G., Toreti, T., Xoplaki, E., Della-Marta, P. M., Zerefos, C. S., Turkes, M., and Luterbacher, J.: Heat wave changes in the eastern Mediterranean since 1960, Geophys. Res. Lett., 37, L04802, doi:10.1029/2009GL041841, 2010. Lenderink, G., van Ulden, A., van den Hurk, B., and van Meijgaard, E.: Summertime inter-annual temperature variability in an ensemble of regional model simulations: analysis of the surface energy budget, Clim. Change, 81, 233–247, 2007. Luterbacher, J., Koenig, S. J., Franke, J., Van der Schrier, G., Zorita, E., Moberg, A., Jacobeit, J., Della-Marta, P. M., Kuttel, M., Xoplaki, E., Wheeler, D., Rutishauser, T., Stossel, M., Wanner, H., Brazdil, R., Dobrovolny, P., Camuffo, D., Bertolin, C., Van Engelen, A., Gonzalez-Rouco, F. J., Wilson, R., Pfister, C., Limanowka, D., Nordli, O., Leijonhufvud, L., Soderberg, J., Allan, R., Barriendos, M., Glaser, R., Riemann, D., Hao, Z., and Zerefos, C. S.: Circulation dynamics and its influence on European and Mediterranean January-April climate over the past half millennium: results and insights from instrumental data, documentary evidence and coupled climate models, Clim. Change, 101, 201–234, doi:10.1007/s10584-009-9782-0, 2010. Meehl, G. A. and Tebaldi, C.: More intense, more frequent, and longer lasting heat waves in the 21st century, Science, 305, 994– 997, doi:10.1126/science.1098704, 2004. Miralles, D. G., Holmes, T. R. H., De Jeu, R. A. M., Gash, J. H., Meesters, A. G. C. A., and Dolman, A. J.: Global land-surface evaporation estimated from satellite-based observations, Hydrol. Earth Syst. Sci., 15, 453–469, doi:10.5194/hess-15-453-2011, 2011. Miralles, D. G., Teuling, A. J., van Heerwaarden, C. C., and VilàGuerau de Arellano, J.: Mega-heatwave temperatures due to combined soil desiccation and atmospheric heat accumulation, Nature Geosci., 7, 345–349, doi:10.1038/ngeo2141, 2014. www.geosci-model-dev.net/8/1/2015/ 13 Mooney, P. A., Mulligan, F. J., and Fealy, R.: Evaluation of the Sensitivity of the Weather Research and Forecasting Model to Parameterization Schemes for Regional Climates of Europe over the Period 1990–95, J. Climate, 26, 1002–1017, 2013. Morrison, H., Thompson, G., and Tatarskii, V.: Impact of Cloud Microphysics on the Development of Trailing Stratiform Precipitation in a Simulated Squall Line: Comparison of One- and TwoMoment Schemes, Mon. Weather Rev., 137, 991–1007, 2009. Nakanishi, M. and Niino, H.: An improved Mellor–Yamada level 3 model: its numerical stability and application to a regional prediction of advecting fog, Bound. Layer Meteor., 119, 397–407, 2006. Nakanishi, M. and Niino, H.: Development of an improved turbulence closure model for the atmospheric boundary layer, J. Meteor. Soc. Japan, 87, 895–912, 2009. Nikulin, G., Kjellstrom, E., Hansson, U., Strandberg, G., and Ullerstig, A.: Evaluation and future projections of temperature, precipitation and wind extremes over Europe in an ensemble of regional climate simulations, Tellus A, 63, 41–55, 2010. Omrani, H., Dobrinski, P., and Dubos, P. T.: Optimal nudging strategies in regional climate modelling: investigation in a Big-Brother experiment over the European and Mediterranean regions, Climate Dynam., 41, 2451–2470, 2013. Orlowsky, B. and Seneviratne, S. I.: Global changes in extreme events: regional and seasonal dimension, Clim. Change, 110, 669–696, doi:10.1007/s10584-011-0122-9, 2012. Pilegaard, K., Ibrom, A., Courtney, M. S., Hummerlshøj, P., and Jensen, N. O.: Increasing net CO2 uptake by a Danish beech forest during the period from 1996 to 2009, Agric. For. Meteorol., 151, 934–946, 2011. Pleim, J. E.: A Combined Local and Nonlocal Closure Model for the Atmospheric Boundary Layer. Part I: Model Description and Testing, J. Appl. Meteor. Climatol., 46, 1383–1395, 2007. Schär, C., Vidale, P. L., Luthi, D., Frei, C., Haberli, C., Liniger, M. A., and Appenzeller, C.: The role of increasing temperature variability in European summer heatwaves, Nature, 427, 332– 336, doi:10.1038/nature02300, 2004. Seneviratne, S. I., Luthi, D., Litschi, M., and Schär, C.: Landatmosphere coupling and climate change in Europe, Nature, 443, 205–209, doi:10.1038/nature05095, 2006. Seneviratne, S. I., Corti, T., Davin, E. L., Hirschi, M., Jaeger, E. B., Lehner, I., Orlowsky, B., and Teuling, A. J.: Investigating soil moisture-climate interactions in a changing climate: a review, Earth Sci. Rev., 99, 125–161, 2010. Seneviratne, S. I., Nicholls, N., Easterling, D., Goodess, C. M., Kanae, S., Kossin, J., Luo, Y., Marengo, J., McInnes, K., Rahimi, M., Reichstein, M., Sorteberg, A., Vera, C., and Zhang, X.: Changes in climate extremes and their impacts on the natural physical environment, in: Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation, edited by: Field, C. B., Barros, V., Stocker, T. F., Qin, D., Dokken, D. J., Ebi, K. L., Mastrandrea, M. D., Mach, K. J., Plattner, G.-K., Allen, S. K., Tignor, M., and Midgley, P. M., a Special Report of Working Groups I and II of the Intergovernmental Panel on Climate Change (IPCC), Cambridge University Press, Cambridge, UK, and New York, NY, USA, pp. 109–230, 2012. Skamarock, W. C., Klemp, J. B., Dudhia, J., Gill, D. O., Barker, D. M., Duda, M. G., Huang, X.-Y., Wang, W., and Powers, J. G.: A description of the Advanced Research WRF version 3, Geosci. Model Dev., 8, 1–14, 2015 14 A. I. Stegehuis et al.: Multi-physics WRF ensemble for simulating European mega heat waves NCAR Tech. Note 1–125, available at: http://nldr.library.ucar. edu/repository/collections/TECH-NOTE-000000-000-855 (last access: 20 July 2015), 2008. Stegehuis, A., Vautard, R., Ciais, P., Teuling, A. J., Jung, M., and Yiou, P.: Summer temperatures in Europe and land heat fluxes in observation-based data and regional climate model simulations, Clim. Dynam., 41, 455–477, doi:10.1007/s00382-012-1559-x, 2013. Sukoriansky, S., Galperin, B., and Perov, V.: Application of a new spectral model of stratified turbulence to the atmospheric boundary layer over sea ice, Bound.-Lay. Meteor., 117, 231–257, 2005. Teuling, A. J., Hirschi, M., Ohmura, A., Wild, M., Reichstein, M., Ciais, P., Buchmann, N., Ammann, C., Montagnani, L., Richardson, A. D., Wohlfahrt, G., and Seneviratne, S. I.: A regional perspective on trends in continental evaporation, Geophys. Res. Lett., 36, L02404, doi:10.1029/2008GL036584, 2009. Tewari, M., Chen, F., Wang, W., Dudhia, J., LeMone, M. A., Mitchell, K., Ek, M., Gayno, G., Wegiel, J., and Cuenca, R. H.: Implementation and verification of the unified NOAH land surface model in the WRF model, 20th conference on weather analysis and forecasting/16th conference on numerical weather prediction, pp. 11–15, Seattle, WA, American Meteorological Society, 2004. Thompson, G., Field, P. R., Rasmussen, R. M., and Hall, W. D.: Explicit Forecasts of Winter Precipitation Using an Improved Bulk Microphysics Scheme. Part II: Implementation of a New Snow Parameterization, Mon. Weather Rev., 136, 5095–5115, 2008. Tiedtke, M.: A comprehensive mass flux scheme for cumulus parameterization in large-scale models, Mon. Weather Rev., 117, 1779–1800, 1989. Tingley, M. P. and Huybers, P.: Recent temperature extremes at high northern latitudes unprecedented in the past 600 years, Nature, 496, 201–205, doi:10.1038/nature11969, 2013. Van den Hurk, B., Doblas-Reyes, F., Balsamo, G., Koster, R. D., Seneviratne, S. I., and Camargo, H.: Soil moisture effects on seasonal temperature and precipitation forecast scores in Europe, Clim. Dynam., 38, 349–362, doi:10.1007/s00382-010-0956-2, 2012. Geosci. Model Dev., 8, 1–14, 2015 Vautard, R., Gobiet, A., Jacob, D., Belda, M., Colette, A., Deque, M., Fernandez, J., García-Díez, M., Goergen, K., Guttler, I., Halenka, T., Karacostas, T., Katragkou, E., Keuler, K., Kotlarski, S., Mayer, S., Van Meijgaard, E., Nikulin, G., Patarcic, M., Scinocca, J., Sobolowski, S., Suklitsch, M., Teichmann, C., Warrach-Sagi, K., Wulfmeyer, V., and Yiou, P.: The simulation of European heat waves from an ensemble of regional climate models within the EURO-CORDEX project, Clim. Dynam., 41, 2555–2575, doi:10.1007/s00382-013-1714-z, 2013. Vautard, R., Thias, F., Tobin, I., Breon, F.-M., Devezeaux de Lavergne, J.-G., Colette, A., Yiou, P., and Ruti, P. M.: Regional climate model simulations indicate limited climatic impacts by operational and planned European wind farms, Nature Communications, 5, 3196, doi:10.1038/ncomms4196, 2014. Warrach-Sagi, K., Schwitalla, T., Wulfmeyer, V., and Bauer, H. S.: Evaluation of a climate simulation in Europe based on the WRFNOAH model system: precipitation in Germany, Clim. Dynam., 41, 755–774, doi:10.1007/s00382-013-1727-7, 2013. Weisheimer, A., Doblas-Reyes, F. J., Jung, T., and Palmer, T. N.: On the predictability of the extreme summer 2003 over Europe, Geophys. Res. Lett., 38, L05704, doi:10.1029/2010GL046455, 2011. Wild, M., Trussel, B., Ohmura, A., Long, C. N., Konig-Langlo, G., Dutton, E. G., and Tsvetkov, A.: Global dimming and brightening: An update beyond 2000, J. Geophys. Res., 114, D00D13, doi:10.1029/2008JD011382, 2009. Wohlfahrt, G., Pilloni, S., Hörtnagl, L., and Hammerle, A.: Estimating carbon dioxide fluxes from temperate mountain grasslands using broad-band vegetation indices, Biogeosciences, 7, 683–694, doi:10.5194/bg-7-683-2010, 2010. Zhang, C., Wang, Y., and Hamilton, K.: Improved representation of boundary layer clouds over the southeast pacific in ARW–WRF using a modified Tiedtke cumulus parameterization scheme, Mon. Weather Rev., 139, 3489–3513, 2011. www.geosci-model-dev.net/8/1/2015/ Supplement of Geosci. Model Dev., 8, 2285–2298, 2015 http://www.geosci-model-dev.net/8/2285/2015/ doi:10.5194/gmd-8-2285-2015-supplement © Author(s) 2015. CC Attribution 3.0 License. Supplement of An observation-constrained multi-physics WRF ensemble for simulating European mega heat waves A. I. Stegehuis et al. Correspondence to: A. I. Stegehuis ([email protected]) The copyright of individual parts of the supplement might differ from the CC-BY 3.0 licence. 1 Supplementary material 2 Table S1: Final ranks. Final rank 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 3 4 Physic combinations MP PBL 10 1 6 5 6 7 8 5 8 7 8 7 10 5 6 1 10 2 10 5 8 4 8 5 6 7 8 5 6 7 6 4 6 7 6 5 8 5 8 1 8 2 6 2 8 5 6 5 8 2 8 5 6 5 8 1 8 5 6 5 8 7 8 1 6 2 6 7 6 5 6 2 8 2 10 5 6 1 10 5 6 1 10 1 6 7 6 5 8 7 8 2 6 5 8 5 8 5 8 6 8 5 6 5 6 1 8 1 8 2 SF 1 5 1 5 1 1 2 1 2 5 4 5 1 2 1 4 1 5 5 1 2 2 2 2 2 2 5 1 5 2 1 1 2 1 5 2 2 5 1 2 1 1 1 5 1 2 5 2 5 6 2 2 1 1 2 RA 4 4 5 4 4 5 4 3 4 4 4 4 4 4 3 4 4 4 5 5 4 5 5 4 5 3 5 5 3 3 4 4 3 4 3 4 3 4 5 4 4 4 4 5 4 4 4 4 4 4 4 5 4 4 5 CU 6 3 6 14 6 6 6 6 6 6 6 3 6 14 6 6 3 14 6 6 14 6 6 14 6 6 6 3 6 6 14 14 6 14 6 14 6 1 14 1 14 1 1 1 1 1 1 1 1 1 1 1 1 1 1 SU 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 5 Figure S1a-d: Timeseries over France 2003 (a,b) and 2007 (c,d) and Russia (e,f) with maximum (a,c,e) 6 and minimum (b,d,f) daily temperatures. Every simulation is shown in gray and observations of E-OBS 7 in black. The blue and red lines are the coldest and the warmest simulations over France during the 8 heatwave. These lines have the same set of physics in all the figures. 9 10 11 12 Figure S2a-d: Scatter plot with soil moisture content at July 31st and precipitation in the preceding 13 months of June-July. Every point is one simulation. Different colors and symbols represent different 14 physics for convection (CU) (a), microphysics (MP) (b), radiation (RA) (c) and planet boundary layer- 15 surface (PBL-SF) (d). 16 17 18 19 Figure S3a-d: Scatter plot with soil moisture content at the end of July and shortwave radiation during 20 the preceding months of June-July. Every point is one simulation. Different colors and symbols 21 represent different physics for convection (a), microphysics (b), radiation (c) and planet boundary 22 layer-surface (d). 23 24 25 26 27 28 Figure S4a-c: Timeseries of temperature (a), precipitation (b) and latent heatflux (c) over Russia 2010. 29 Figure S5a-d: Timeseries of temperature (a), precipitation (b), radiation (c) and latent heatflux (d) over 30 France 2007. 31 32 33 34 35 Figure S6a-h: Timeseries of temperature (a,d), precipitation (b,e) and latent heatflux (c,f,g,h) over the 36 Iberian Peninsula 2003 (a,b,c), Scandinavia 2003 (d,e,f), the Iberian Peninsula 2010 (g) and 37 Scandinavia 2010 (h). 38 39 40 41 42 68 4 CHAPTER III. ATMOSPHERIC PROCESSES Discussion and additional results Convection was found to play a very important role for summer temperatures, precipitation and soil moisture, as was shown in the previous section. The specific role of the convection however, needs further analyses. In this discussion I try to investigate how the temperature and other variables are influenced by the different convection physics. 4.1 The importance of the convection scheme The large spread due to the convection schemes (Fig. 5, S2 and S3 of the paper) raises the question whether the choice of this scheme in the beginning of the summer was still of influence for August temperatures. Low soil moisture level in the beginning of the summer favors warm summer conditions and heatwaves (Vautard et al., 2007; Quesada et al., 2012). If a ’dry’ convection scheme would initialize dry soils, it could possibly determine the values of late summer temperatures through positive soil moisture-temperature feedbacks. In a ’wet’ convection scheme however, the wetter soils may not lead to positive soil moisturetemperature feedbacks, leading to cooler August temperatures. Here we investigate whether initial dry conditions can be turned into wetter conditions by changing the convection physics. This may elucidate the relative importance of land-atmosphere feedbacks and convection for August temperatures. In a sensitivity experiment on the 2003 heatwave, four control simulations were run from May to August with different convection physics. All other physics schemes were kept constant. The four convection schemes were those used in the previous section (Kain-Fritsch, Grell-Devenyi, Tiedtke and New-SAS). The control simulation was restarted at the end of May, 15 and 30 June, and 15 and 30 July. At these dates the simulation was restarted with the other three convection schemes, resulting in three new simulations at every restart. The August temperatures were then evaluated. Changing the convection physics during a simulation may lead to inconsistencies in the simulation due to the interaction with the other physic schemes. However, we did not find any abnormal behavior in analyzing the model outputs. The Grell-Devenyi scheme was found to be the wettest scheme overall, with high soil moisture values and high precipitation. Tiedtke was the driest scheme, with lower values for both soil moisture and precipitation (Fig. III.2 and figures 5, S2 and S3 of the paper). In Fig. III.2 all colored dots represent a simulation and the different colors represent different convection physics. All grey dots represent a simulation from the sensitivity test. Their different sizes represent different restart dates; the later the restart date, the bigger the dots. Their different shades represents the different restart convection physics. The orange circles represent the 4 control simulations. Interestingly, the range of the four different convection physics (vertical range of the grey dots) is almost similar. This suggests that the convection mostly determines the temperatures at the end of the summer, and that land atmosphere feedbacks seem to be less important in some convection schemes. Because, even when starting with very dry soils in May, temperatures can decrease substantially, probably due to enhanced precipitation in certain convection physics. 4. DISCUSSION AND ADDITIONAL RESULTS 69 Figure III.2: Scatter plot of soil moisture content in May, and temperature in August over France 2003. Every point is one simulation. Different colors and symbols represent different physics for convection. The orange circle is the original simulation. Different shades of grey represent the convection scheme for the restart and different size shows the time of the restart date, the later the restart, the larger the circle. ● ●● 0 20 40 60 80 Relative humidity 500 hPa (%) 0.6 0.4 0.2 Precipitation (mm) ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● 0.8 ● Kain−Fritsch ●● ●● Grell−Devenyi ●● ●● ● Tiedtke ●● ●● SAS New ●● 0.0 0.015 0.010 Density 0.005 0.000 1.0 CHAPTER III. ATMOSPHERIC PROCESSES 0.020 70 100 ● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● 0 20 40 60 80 100 Relative humidity 500 hPa (%) Figure III.3: PDF of the relative humidity at 500 hPa (left) and precipitation as a function of relative humidity at 500 hPa (right). The different colors correspond to different convection schemes. 4.2 What triggers convection? In both section 3.3 and in the sensitivity experiment the importance of the convection scheme on temperature and humidity was emphasized. However, the question of how the convection physics modify the temperature is still unanswered. It is not possible to solve this issue completely without the precise knowledge of the physics in the convection schemes and the interactions with all other physics. Here I will give some brief ideas and show some preliminary results. While the source of the exact differences between the schemes is uncertain and are probably multiple, some simple tests may show some possible mechanisms between the different physics. In Fig. III.2 it was shown that even starting with a dry soil, cold August temperatures could be obtained by changing the convection scheme. This implies that the physics have a different sensibility regarding humidity. One can imagine for example, that if one scheme rains more with a certain relative humidity, it may get wetter overall through feedbacks with enhanced precipitation and soil moisture. Fig. III.3 shows a probability density function (PDF) of the relative humidity at 500 hPa for the 4 different convection schemes. Although the schemes are quite similar, Grell-Devenyi exhibits a higher density between 5% and 30% relative humidity, while it shows lower values between 50% and 70%, compared to the other three convection schemes. This fact alone does not necessarily proves a general higher wetness of this scheme. The right panel shows the amount of precipitation at a certain relative humidity. With values between 10% and 30%, both Kain-Fritsch and Grell-Devenyi have relatively high precipitation compared with the other two schemes. This coincides with the higher occurrence of relative humidity in the PDF of Grell-Devenyi. It could explain why Grell-Devenyi is wetter overall, and why even when starting with dry initial soils, it can get much colder and wetter than with other convection physics. 5. SUMMARY AND CONCLUSIONS 71 In a model three main questions should be answered regarding convection: 1) Will there be convection? 2) How much convection will there be? and 3) What are the consequences of convection; will there be cloud formation? (Hohenhegger, personal communication). The first question depends on the trigger function which identifies the potential updraft source layers for convective clouds and determines when and where convection occurs (Charabi, 2010; Warner et al., 2010). An often used method is the parcel lifting. A parcel with a certain temperature is lifted until it is warmer than its environment. In this case convection is triggered. If the parcels’ temperature remains less than the temperature of the environment, convection will not take place. Kain and Fritsch (1992) found considerable differences between simulations where only the trigger function was changed. The amount of convection depends on the closure scheme. Three closure schemes are moisture convergence, convective available potential energy (CAPE) closure and boundary layer closure. Moisture convergence is a relatively old method and is used in Grell-Devenyi (Grell and Freitas, 2014) and in the first version of the Tiedtke scheme (Tiedtke, 1989). The latter has been adapted with a CAPE closure (Zhang et al., 2011), a scheme that is now widely used. This method is based on a quasi-equilibrium between convection and large scale processes. It is also used in Kain-Fritsch (Kain, 2004), the New SAS scheme (Han and Pan, 2011) and Grell-Devenyi (Grell and Freitas, 2014; Pohl et al., 2011). Boundary layer closure is relatively new and not used in any of the tested convection schemes (Hohenhegger, personal communication). The amount of clouds is determined by the cloud model that simulates the effect of single clouds and cloud ensembles. Arakawa (1974) developed a spectral cloud model that permitted different clouds in each grid cell of a model. However, because of its complexity simpler cloud models exist as well. In these only one single cloud model is applied to each grid cell. This is for example used in the original Tiedtke scheme (Tiedtke, 1989). For the entrainment and detrainment rate different formulation exist that can have a large impact on precipitation rates. A comparison between these formulations however, is out of the scope of this study. 5 Summary and conclusions In this chapter a large multi-physics ensemble in WRF was designed and tested on its ability to simulate the heatwaves of 2003 and 2010 and the relatively wet summer of 2007 to 1) identify the sensitivity to different atmospheric physics and 2) to create a well-performing reduced ensemble for future studies especially focusing on summer climate and heatwaves. The 216 different configurations were evaluated against observations of temperature, precipitation and incoming radiation. Despite wind nudging above the boundary layer, a large temperature spread was found at the end of the summer between different model configurations. Almost all simulations underestimated temperature, while precipitation and shortwave radiation were mostly overestimated. This is important as future heatwaves might be of even larger magnitude than the ones of 2003 and 2010 that were evaluated here. A good model structure with appropriate physics must therefore be chosen to be able to reach such extreme temperatures or droughts. Convection was found to be the most sensitive process leading to most model spread. There seem to be relatively ’dry’ schemes, such as the Tiedtke scheme, that result in warmer 72 CHAPTER III. ATMOSPHERIC PROCESSES temperature and little precipitation. On the other hand, the relatively ’wet’ schemes, such as Grell-Devenyi, induce higher soil moisture values, more precipitation and, probably more cloudy skies, less shortwave radiation and therefore relatively low temperatures. The positive feedback mechanism between dry soils, less clouds and higher temperatures might not always be of importance. Dry soils in the summer can easily be reversed to wetter soils by changing the convection scheme. This suggests that the triggering of precipitation might be decisive for climate conditions at the end of the summer in some model simulations. Although the land surface physics were not changed within the 216 simulations, its sensitivity was tested. It was shown that temperature could be in- or decreased till up to 3◦ C by perturbing the initialization of soil moisture by 20%. At the time of the study however, only one suitable land surface scheme was available. In the current version of WRF, more schemes are available which might be more appropriate for future studies. Though soil moisture plays an essential role in the partitioning of fluxes which are described in the land surface parameterization, the convection scheme may be able to kill positive soil moisture-temperature feedbacks. Thus the scheme determining eventual positive land-atmosphere feedbacks is not only the land surface scheme, but also the convection, radiation, microphyics and boundary layer schemes. Evaluated and constrained by several observation-based data products, we created a 5member multi-physics ensemble to perform future heatwave studies. This reduced ensemble contains 3 different microphysic schemes, 3 PBL-surface layer combinations, 2 radiation schemes and 3 convection schemes. The high variety of these schemes is important, as different physics might take slightly different pathways to reach similar climatic conditions. For assessing uncertainty in future climate simulations, very similar models would not be helpful, because the full range of uncertainty would not be displayed. Also uncertainty assessment for impact studies may benefit from more diverse simulations. Vegetation for example, can react very differently to the different precipitation rates and soil humidity levels that are present in our 5-member multi-physics ensemble. C HAPTER IV The influence of soil moisture on summer temperatures 73 74 1 CHAPTER IV. THE INFLUENCE OF SOIL MOISTURE ON SUMMER TEMPERATURES Introduction Summer temperatures in Europe are controlled by both atmospheric circulation and landatmosphere interactions. For elevated temperatures and heatwaves, the atmospheric circulation needs to favor the transport of warm air towards the continent and stabilize atmospheric conditions. The soil must relate to a positive soil moisture-temperature feedback mechanism as discussed in chapter I. Although previous studies have established the importance of these two processes (Black et al., 2004; Miralles et al., 2014), they did not try to quantify their individual contribution to warm summer temperature anomalies. In this study we chose to investigate early summer soil moisture only, as soil moisture later in the summer season may be much harder to separate from the circulation. The work in the chapter was initiated by the internship of Martha Vogel, and I contributed to the supervision. She performed most of the simulations and made a preliminary analysis of the results. The predicted temperature increase over Europe at the end of this century is estimated between 2◦ C and 5◦ C, depending on the scenario and the region (van der Linden and Mitchell, 2009; Boberg and Christensen, 2012; Stegehuis et al., 2013a). This increase is projected to be stronger in Southern Europe (up to 5◦ C), than in the northern part (up to 2◦ C) (Fig. 3 Stegehuis et al., 2013a). Furthermore, a larger temperature variability is expected, especially in central Europe. This overall increase in temperature might lead to more soil drying, creating favorable situations for positive soil moisture-temperature feedbacks. A more precise understanding, and the quantification of the role of soil moisture and its trends over the recent years, might therefore be critical in understanding the variability in summer temperatures and heatwave events in the future climate. The atmospheric circulation patterns associated with heatwaves are atmospheric Blocking and Atlantic low, as discussed in section I.1.2.3. The occurrence of the different circulation patterns have changed over the last century (Horton et al., 2015; Coumou et al., 2015; Alvarez-Castro et al., 2016). Although the impact of these patterns have been studied (van Haren et al., 2015), there is still considerable uncertainty about whether and how they may change in the future climate (Cattiaux et al., 2012). Nonetheless, a better understanding of their contribution to extreme heatwaves can be useful in predicting climate change. In the previous chapter we have selected a limited number of model configurations that were able to simulate the heatwaves of 2003 and 2010 well, as well as the relatively wet summer year of 2007. Although we do not focus solely on heatwaves in this chapter, these configurations will be used here to create an ensemble of summers based on different initializations of soil moisture and atmospheric variables. All these initial conditions are realistic, as the chosen values are those from modeled summer years from 1980-2011. We find that the contribution of early summer soil moisture in the heatwaves of 2003 and 2010 was important and contributed to approximately 1◦ C at maximum. The large scale drivers however, explains the largest part of the temperature anomaly, with up to 3◦ C in 2003 and up to 6◦ C in 2010. An interesting result is that the contribution of initial soil moisture has been increasing in importance over the last three decennia in central Europe and France and in parts of Russia. These regions coincide with the regions with a significant negative trend in soil moisture. The influence of large-scale drivers on the other hand, has increased only over Eastern Europe. 2. METHODS 75 In the following section the methods are presented, followed by the results of ongoing work. Some additional results of the evaluation of the used control simulation are presented afterwards. The chapter is finished with a discussion and concluding remarks. 2 2.1 Methods Model description We used the regional climate model Weather Research and Forecast (WRF) (Skamarock et al., 2008) over the EURO-CORDEX domain (Jacob et al., 2014) with a spatial resolution of 50 km. For more specific details of the model setup we refer to (Stegehuis et al., 2015). The physics that were used in this study correspond to some of those tested for heatwaves in the same study. The main physics used was ranked number 7, with microphysics Morission (Morrison et al., 2009), PBL-Surface layer MY-ETA (Nakanishi and Niino, 2006), RRTGM radiation citepIacono2008, the convection scheme Tiedtke (Tiedtke, 1989; Zhang et al., 2011) and the NOAH land surface (Tewari et al., 2004). A more detailed description of the physic schemes can be found in chapter III. 2.2 Control simulation The physics configuration that was used in this study, after Stegehuis et al. (2015), was only evaluated by them for summer conditions. For a multi-year simulation however, it is preferable that the model also performs well during other seasons. A bias in winter or spring might still be visible in summer, as for example soil moisture conditions have a long-term memory. The correct performance over all seasons is not imperative, as different processes are important in different times of the year. Snow melt for example, is a typical winter/spring process, and might influence temperature and soil moisture even over spring and summer. In this section I present an evaluation of the control simulation. For this study configuration 7 was used from the supplementary table 1 from Stegehuis et al. (2015). Although it was not the best performing configuration over the summer, preliminary tests showed better performance during the winter. Boundary conditions are from ERA-Interim (Dee et al., 2011). The control simulation performs best over the summer, the season on which the configuration was selected (Fig. IV.7). While the minimum temperature is slightly too warm in large parts of Europe (up to ∼0.5◦ C-1◦ C), the mean and maximum temperature are slightly too cold and thus underestimating observed temperatures. Maximum temperatures have a cold bias in all seasons except from the autumn. The largest cold bias was found in spring, which was similar for the minimum and mean temperatures (Fig. IV.7). Whilst a temperature bias in spring can sometimes be attributed to a misrepresentation in snowmelt, we do not assume this is the case here, as the cold bias is almost similar over entire Europe, and not just over the coldest regions. The warm biases in minimum and mean temperature in Northern Scandinavia however, are most likely due to the incorrect simulation of snow. 2.3 Attribution of temperature anomalies Two experiments were performed in order to analyze and quantify the effect of initial soil moisture (ESSM) and large-scale drivers (LS) on summer temperature anomalies. In the first experiment, initial soil moisture conditions of one year were replaced in the end of June, 76 CHAPTER IV. THE INFLUENCE OF SOIL MOISTURE ON SUMMER TEMPERATURES Figure IV.1: Minimum, mean and maximum temperature difference between the control simulation and EOBS over the winter, spring, summer and autumn averaged over 1980-2011. with the soil moisture conditions of the 31 other years, to attribute temperature anomalies to anomalies in soil moisture. In the second experiment, we replaced all other variables of each of the 31 years for the attribution to large-scale drivers, which include SST, large-scale circulation and boundary conditions. WRF was then run over the two following months, till the end of August. The temperature response to the soil moisture anomaly can therefore be estimated as: Ry’y = Ty – Ty’yESSM (1) Where Ty is the temperature of the control run for year y, and Ty’yESSM the average temperature over the 32 ensemble simulations (the 31 perturbed years plus the control year). A similar equation could be written to express the response of the temperature anomaly to LS contribution: Ryy’ = Ty – Tyy’LS (2) Where Ty is the temperature of the control year and Tyy’LS the average temperature of the 32 simulations. We have verified that the this decomposition is almost additive, i.e. that T’y (Ty – the long-term climatology) and T”y (Ryy’ + Ry’y ) are close to one another, meaning that there is no strong nonlinear interaction among ESSM and initial SM effects. 3. SUMMER WARMING INDUCED BY EARLY SUMMER SOIL MOISTURE CHANGES IN EUROPE 3 3.1 77 Summer warming induced by early summer soil moisture changes in Europe Abstract Heatwaves and warm summer temperatures are often associated with a deficit in soil moisture. Heat can develop more easily during and after drought conditions, fostered by landatmosphere feedbacks (Seneviratne et al., 2010; Stéfanon et al., 2014). Summer drought changes have also been shown to be a major driver of the fraction of heat attributable to climate change (Orth and Seneviratne, 2014) as for instance in the case of the Russian heatwave. However, the role of soil moisture changes in recent warming trends has not been established so far. Here we show that early summer soil moisture (ESSM) deficit explains most of the rapid recent warming trend in Western/central Europe [∼0.5◦ C.decade-1], where extreme heatwaves took place in 2003, 2006 and 2015. This result comes in contrast to the relatively weak relative contribution of ESSM to interannual variability of summer temperature, reaching only at most 20% on average, while more than 80% is explained by atmospheric circulation and other large-scale drivers. Such a process attribution was made possible by using a regional climate model and a systematic array of sensitivity experiments to estimate and separate the temperature response of ESSM anomaly from other contributions on a seasonal basis. We show as an example that the spring drought preceding the 2003 heatwave contributed to about 1◦ C, while large scale drivers were dominant. We also found that the different contributions of circulation and soil moisture effects approximately add up to the anomaly, suggesting little nonlinear interactions among the two processes. Our results illustrate the complexity of processes involved in regional warming signal, not directly attributable to changes in large-scale drivers. 3.2 Main text Mega-heatwaves and exceptional warm summers are very likely to become more frequent over Europe in the coming century (Schär et al., 2004; Barriopedro et al., 2011). Such extreme climatic events can cause great damage and loss to society and ecosystems, as witnessed in for example 2003 in Europe and 2010 in Russia (e.g. Ciais et al., 2005; Fouillet et al., 2006; Barriopedro et al., 2011). Although adaptation can mitigate such impacts, early warnings and seasonal predictions might be far more efficient. However, such predictions in Europe had limited success in practice so far. Heatwaves, and warm summers in general, are associated with specific atmospheric circulation patterns, especially atmospheric blocking and south-westerly flows (Cassou et al., 2005), but their predictability on the seasonal time scale is very low. Hopes of seasonal predictability come from soil moisture in early summer conditions which may precondition temperature anomalies in subsequent months (Wang and Kumar, 1998; Kanamitsu et al., 2003; Vautard and Yiou, 2009). In Europe, indications of such a causal link have been provided in several observational studies (Vautard et al., 2007; Hirschi et al., 2011). In particular dry early summer soil moisture (ESSM) has been shown to make temperature more sensitive to the occurrence anticyclonic weather regimes (Quesada et al., 2012). The establishment of dry summer conditions induce land-atmosphere feedbacks helping amplification of heat in such cases (Seneviratne et al., 2010; Hirschi et al., 2011; Mueller and Seneviratne, 2012). However, the contribution of ESSM to late summer temperature interannual variability has 78 CHAPTER IV. THE INFLUENCE OF SOIL MOISTURE ON SUMMER TEMPERATURES not been quantified so far in a systematic way. Long-term changes in ESSMs, due to greenhouse gases or natural interdecadal variability may also contribute to the summer warming trend, which has not been estimated so far. We use regional climate sensitivity simulations to quantify the contribution of ESSM. A control simulation from 1979 to 2011 with ERA-Interim (Dee et al., 2011) boundary conditions and grid nudging (wind only) above the boundary layer was performed with the Weather Research and Forecasting (WRF) (Skamarock et al., 2008) regional climate model. The model physics were optimized for the simulation of the European summer climate (Stegehuis et al., 2015). The sensitivity simulations consist of more than 900 additional simulations starting at June 30 of each year from 1980 to 2011. At June 30, in year x, the soil moisture is initialized with the values of all other 31 years, where after the simulation is run until the end of August, keeping the other conditions as in the control experiment. A similar set of experiments is done with the large-scale drivers (LS, here defined as winds above the lowest model layers, SST and boundary conditions). The contribution of ESSM and LS to late summer temperature is the difference between the temperature anomaly of year x and the ensemble mean of the 31 sensitivity simulations + year x, for all years (see methods). This method works best in absence of feedback of initial soil moisture on later large-scale circulation and SST. This seems realistic as no significant correlation was found (P < 0.1) between ESSM and observed sea level pressure or global z500 in July and August (not shown). The model reproduces well the 2003 July and August mean temperature anomaly, as the bias over Europe remains lower than about one degree, giving confidence in the process analysis (Fig. IV.1a). Such a model skill was also found for other summers. The average summer temperature bias over 1980 to 2011 was of similar magnitude with the exception of some small regions as in the Alps, the Pyrenees and parts of the Northern Iberian Peninsula. Summer precipitation is mostly underestimated over Southern Europe, where rain is essentially absent during the summer, with largest deviations (up to 60%) in the southern part of the Iberian Peninsula, and overestimations in Scandinavia (∼20%). In the build-up of the temperatures of the summer 2003, the contribution of ESSM was significant but not dominant in the temperature anomaly of July and August 2003. It explained approximately 0.5 - 1◦ C of the temperature anomaly in central Europe (∼20%) (Fig. IV.1b). This area corresponds well with the area where ESSM was most anomalously low during summer (-10% to -50%) (Fig. IV.1d). The contribution of ESSM was of the same magnitude in 2010 in Russia, but contributed 10% to 15% to the total temperature anomaly (Fig. IV.1f). The importance of LS such as atmospheric circulation and SST was larger in the heatwave area, reaching 2-3◦ C in France and central Europe in 2003 and 5-6◦ C in Russia in 2010 (Fig. IV.1c,g). An essential component of the heatwaves therefore lies in the large-scale drivers. The sum of the two contributions is nearly equal to the modeled anomaly itself, indicating weak or no nonlinear interactions between ESSM-driven mechanisms and flow-driven mechanisms (Fig. IV.1a,e). Compared to the contribution of LS, the contribution of ESSM in 2003 and 2010 might have been relatively small. However, the influence of ESSM on summer temperatures has increased over central Europe over the last 30 years with approximately 0.2◦ C per decade (Fig. IV.2c). Thus, the temperature warming is induced by drier initial soils, a result that is 3. SUMMER WARMING INDUCED BY EARLY SUMMER SOIL MOISTURE CHANGES IN EUROPE 79 Figure IV.2: Difference of July + August (JA) temperature anomaly between the model and the observations for 2003 (a) and 2010 (e). The contribution of ESSM (b,f) and LS (OTH) (c,g) to the JA temperature anomaly of 2003 (b,c) and 2010 (f,g). Simulated JA soil moisture anomaly of 2003 (d) and 2010 (h). The shaded area in panel a and e indicates the regions where the joint contribution of ESSM and LS is more than 0.5◦ C different from the total temperature anomaly. in agreement with Seneviratne et al. (2006b), who showed an increasing importance of soil moisture towards the end of the century. A large part of France, Germany and other parts of central Europe have experienced a more important trend in the ESSM contribution (p-value < 0.05) (Fig. IV.2c). In this area, the trend in ESSM-induced warming is almost equal to the warming trend itself, despite that the model slightly underestimates this trend as compared to observations (Fig. IV2a,b). An underestimation of the warming trend was also found in the ENSEMBLES models (Lorenz and Jacob, 2010). A central result here is therefore that the early summer drying trend (Fig. IV.3) explains a major part of the July and August warming along the past three decades in many Western European areas in our WRF simulations. In other areas where the ESSM contribution is significant (e.g. Ukraine), it explains a smaller fraction of the overall warming. The contribution of LS also shows an overall warming trend over large parts over Europe (Fig. IV.2d). It is not surprising that this contribution is large as it contains the response to the more direct effects of climate change in SST warming, temperature and humidity in the boundary conditions of the domain. The trend in LS in complementary areas of where the soil moisture is dominating (Fig. IV.2d). This is partly due to the trend in the SLP, that is positive over Eastern Europe, while it is negative over central and Western Europe (Fig. IV.4). The latter suggests that the overall global warming trend could be contra-balanced by the negative influence of the SLP. The contribution of ESSM to recent warming is larger than the one of circulation in Western Europe. This comes in contrast to its relative contribution in explained interannual variance of the summer temperature anomalies, which, as for the summer in 2003 and of 2010, only reaches a small fraction (Fig. IV.1b,d). At maximum, only 20% of the interannual variance of temperature is explained by ESSM variability. As for the case of 2003, we do not find an overall strong nonlinear interaction between ESSM-driven warming, and warming induced 80 CHAPTER IV. THE INFLUENCE OF SOIL MOISTURE ON SUMMER TEMPERATURES Figure IV.3: Temperature trend of observations (a) and WRF (b) in July and August from 1980-2011. Trend of the contribution of soil moisture (c) and circulation (d). Shaded areas indicate a significance level of 95%. 3. SUMMER WARMING INDUCED BY EARLY SUMMER SOIL MOISTURE CHANGES IN EUROPE 81 65 0.001 55 0.002 60 70 mm/yr 50 0.000 45 −0.001 40 −0.002 −10 0 10 20 30 40 Figure IV.4: Modeled trend of the soil moisture level at 30 June (ESSM) over 1980-2011 (mm/yr). by large-scale drivers as their sum nicely add up to the total simulated warming trend. The correlation between the temperature anomaly and the ESSM contribution is relatively low (Fig. IV.5a) compared to that between the temperature anomaly and the LS contribution (Fig. IV.5c). The maximum of the correlation with ESSM is located in central-southern and central-eastern areas, which are areas where strong soil-atmosphere interactions take place in several studies (e.g. Seneviratne et al., 2006b). The correlation between the two drivers is very weak (Fig. IV.5b), while their sum adds up to almost 1 (Fig. IV.5d). The role of soil moisture to the temperature anomaly in France and central Russia is most important in the warmest years, in this case 2003 and 2010 (compare Fig. IV.5a with IV.6). Which is understandable as in these areas the regime can shift from energy- to soil moisture limited. In colder years (Fig. IV.6), soil moisture is important in Italy and areas around the Black Sea, which might be caused by enhanced precipitation. The high correlation between the LS and the temperature anomaly in Russia, there were the heatwave was strongest, is especially due to the specific 2010 heatwave, as the elimination of this year causes a decrease in correlation from > 0.9 to 0.8 or smaller (Figs. IV.5,6). Note however, that a correlation over 10 year cannot easily be compared with that over 32 years. 82 CHAPTER IV. THE INFLUENCE OF SOIL MOISTURE ON SUMMER TEMPERATURES 70 15 60 10 5 50 0 40 −5 30 −10 −15 −40 −20 0 20 40 60 Figure IV.5: Trend of NCEP sea level pressure over 1980-2011 (hPa/yr). 4 Discussion and conclusions In this chapter we attempted to quantify the relative role of early summer soil moisture (ESSM) and of large-scale drivers (LS) to summer temperatures. We find that the contribution of ESSM in the heatwaves of 2003 and 2010 was important and contributed to approximately 1◦ C at maximum. LS however, explains the largest part of the temperature anomaly, with up to 3◦ C in 2003 and up to 6◦ C in 2010. This finding uses the assumption of no feedback from ESSM onto subsequent circulation. Interestingly, we found that ESSM deficit explains most of the rapid recent warming trend in Western/central Europe [∼0.5◦ C.decade-1]. Regions that coincide with the areas with a significant negative trend in ESSM. The influence of LS on the other hand, has increased over Eastern Europe. The relatively warm summer years of 2012 to 2015 may stabilize or strengthen the trend, but time was missing to extend the experiment. An assumption that we made in this study was the absence of feedback of initial soil moisture to large-scale circulation and SST. To test this, we correlated the initial soil moisture averaged over France, there were the soil moisture anomaly was largest in 2003, to observation-based data of global sea level pressure and z500. We found no correlation and thus concluded the validity of the assumption. Another assumption has to do with the model settings. We nudged the wind above the boundary layer that might have blocked the circulation to develop totally freely, especially in response to ESSM anomalies. In the future a test with a non-nudged run could be done for the validation of this second assumption. It 4. DISCUSSION AND CONCLUSIONS 83 Figure IV.6: Correlation between ESSM contribution and July-August (JA) temperature anomaly (a), ESSM and circulation contribution (b), LS contribution and JA temperature anomaly (c) and the added contribution of ESSM and LS with JA temperature anomaly. Note that the JA temperature anomaly is not detrended. 84 CHAPTER IV. THE INFLUENCE OF SOIL MOISTURE ON SUMMER TEMPERATURES 10 warmest years 10 warmest years without 2003 and 2010 10 coldest years 10 coldest years without 1980 and 1984 Figure IV.7: Same as Fig. IV.5 but for the 10 warmest years (averaged over the domain), the 10 warmest years without 2003 and 2010, the 10 coldest years and the 10 coldest years without 1980 and 1984. 4. DISCUSSION AND CONCLUSIONS 85 seems that there is only little or no feedback between soil moisture and large-scale drivers. We can however, not be fully sure with the method we have chosen. This may partly be due to our decision to describe these drivers as a combination of SST, boundary conditions and large-scale circulation. The increasing contribution of early summer soil moisture deficit to the warming trend in Western/central Europe and parts of Russia may provide more accurate seasonal predictions and early warning signals in the future. The large-scale circulation does not have a longterm memory, thus knowledge about its state at the beginning of the summer season is little helpful. The increasing importance of soil moisture, and its longer-term memory, may support the possibility of early seasonal summer predictions. Furthermore, a better insight to the different drivers of summer temperatures could improve our understanding of impacts, such as vegetation responses, to elevated temperatures. Vegetation can respond differently to soil drought and atmospheric drought (Chaves, 1991; Sellin, 1997; Darlington et al., 1997; MaierMaercker, 1998). Thus knowledge on both the soil status and the atmosphere may help to better predict future vegetation responses and distribution. C HAPTER V Drought impacts on European forest species 1 Introduction In the previous chapters the main focus was on summer climate and heatwaves in Europe, discussing both atmospheric processes and land-atmosphere feedbacks. An important aspect in land-atmosphere feedbacks is the vegetation, which can either have a moderating or an amplifying effect on the climate. This depends largely on the vegetation type and on the availability of soil moisture which influences the status of the vegetation (Section I.1.3.4) and can regulate the repartitioning of land heat fluxes, also through the vegetation (Section I.1.3.1). The impact of warm summers and heatwaves on vegetation is therefore an important aspect in understanding the climate and will be investigated in this chapter. Forests can transport more moisture into the atmosphere due to deeper rooting depths, and thus a better access to groundwater, leading to more evapotranspiration than grasslands during droughts when the topsoil is dry. The lower albedo of forests may however cause a micro-climate above forests that is warmer compared to short vegetation (Naudts et al., 2016). The albedo effect seems to outrun the evaporation effect for starting heatwaves (Teuling et al., 2010). This may however be reversed after longer or more severe heatwaves when most trees close their stomata due to water stress. To untangle these effects, knowledge on the reaction of vegetation during water stress is crucial. For the effect of climate on vegetation on larger scales, vegetation models can be used. Most of these models do not distinguish between separate species, but use plant functional types (PFTs). A new version of ORCHIDEE (Krinner et al., 2005) however, does include the species characteristics of most European tree species. This version, ORCHIDEE-CAN (Naudts et al., 2015), additionally has a more detailed description of the water potentials in the soilplant-atmosphere continuum, which may better represent the water stress that is experienced by the plants during dry and warm summers. Because effects on drought and heatwave on vegetation is often local and species dependent, it is important to review ecological literature about these effects before starting with modeling. I will focus on 5 of the main forest tree species in Europe. Furthermore, I will 87 88 CHAPTER V. DROUGHT IMPACTS ON EUROPEAN FOREST SPECIES present some preliminary results of the evaluation of the vegetation model for simulating drought and heat stress on vegetation. These are studies on a local scale, using eddy covariance data (Baldocchi et al., 1996) for the evaluation of ORCHIDEE-CAN. This can be considered as a first step for simulating vegetation stress on a large, regional or global, scale. For the latter, the vegetation model needs to be coupled to, or driven by, reanalyses data or a regional or global climate model. However, these last steps have not yet been performed. In the next section a detailed description of heat and drought stress on some of the main tree species in Europe is presented based on a literature review. This is followed by a description of the vegetation model ORCHIDEE-CAN, which is used to evaluate the simulation of drought and heat stress on vegetation presented in section 5.4. In the next section, I will discuss on some of the missing processes in vegetation models for better simulating the impact of heatwaves on vegetation. The chapter is concluded by a summary and concluding remarks. 2 Drought and heat stress on 5 main European forest tree species This section describes some of the drought stress-induced characteristics of five main forest tree species in Europe (Picea sp., Pinus sylvestris, Fagus sylvatica, Quercus robur & petraea, and Quercus ilex). The literature review shows the complexity of drought stress mechanisms in different species, highlighting the difficulty of modeling impacts of drought stress on vegetation. 2.1 Picea sp. Spruce is an important economically dominant forest tree species in many parts of Europe. While the genus Picea consists out of multiple species, in this study they are all grouped together. Spruce is relatively sensitive to drought and is a water saving, or isohydric species. The drought sensitivity is often attributed to its relatively shallow root system (Schlyter et al., 2006). Although some studies contradict this and point out that deep vertical roots develop at an older age (Puhe, 2003). Decline and dieback in different regions are often associated with drought and high summer temperatures (Nillson, 1993; Modrzyński and Eriksson, 2002; Allen et al., 2010). Photosynthesis Photosynthesis of Picea species greatly decreases or even stops during drought stress (Silim et al., 2001; Bigras, 2005; Zweifel et al., 2009). Stomatal closure, a mechanism to limit water stress during drought, may occur at shoot water potentials below -2 MPa (Bigras, 2005). Although this is also dependent on the atmospheric demand (MaierMaercker, 1998; Ditmarová et al., 2010). During less severe water stress, a significant decline in stomatal conductance was found up to 50% (Townend, 1993; Stewart et al., 1995; Lu et al., 1996; Tan and Blake, 1997; Silim et al., 2001; Pepin et al., 2002; Lu et al., 2007; Ditmarová et al., 2010). It might however, not be the principal limitation of net photosynthesis (Toivonen and Vidaver, 1988; Stewart et al., 1995). The inter-cellular CO2 concentration, which is coupled to the stomatal conductance, may be more important for the reduction in net photosynthetic rate. Stomatal conductance was found to contribute only to approximately 20% to 2. DROUGHT AND HEAT STRESS ON 5 MAIN EUROPEAN FOREST TREE SPECIES 89 the reduction in net photosynthesis (Stewart et al., 1995). The recovery of stomatal conductance is debated, and both quick (Stewart et al., 1995) and incomplete recovery (Blake and Li, 2003) are found. The incomplete recovery may be due to membrane damage preventing the reopening of stomata (Blake and Li, 2003). The differences found between the studies might be dependent on species and the experimental design. Mesophyll limitation seemed to increase with seedling age, but was not affected by soil drought (Stewart et al., 1995). Water use efficiency (WUE) increased during drought stress (Silim et al., 2001; Duan et al., 2007). Another method to measure photosynthetic performance is chlorophyll fluorescence. Chlorophyll fluorescence is the excess energy of absorbed light that is re-emitted by chlorophyll. Absorbed light can furthermore be used to drive photosynthesis or it can be dissipated as heat. Drought or heat stress can influence the partitioning of absorbed light energy into these components. Some studies on photosynthetic performance nowadays use this technique (Maxwell and Johnson, 2000). The paragraph below describes some of the results concerning Picea sp. Chlorophyll fluorescence mostly provides information on photosystem II (PSII) because it is generally assumed that the variability of the fluorescence of photosystem I is not significant at room temperature. Maximum efficiency of PSII decreased quadratically with decreasing shoot water potential (Bigras, 2005). This decrease seemed to be due to a decrease in maximal fluorescence (Fm) rather than an increase in dark limited fluorescence, the latter being related to damage to the reaction centers (Bigras, 2005). The decrease in Fm may be the start of photoprotective mechanisms (Demmig and Björkman, 1987). Pukacki and Kamińska-Rozek (2005); Ditmarová et al. (2010) however, found that the decline of the maximum efficiency of PSII only decreased significantly after ∼1 month of water stress, while mild stress did slightly, or not at all, decrease the maximal photochemical efficiency of PSII (Eastman and Camm, 1995; Duan et al., 2005; Blödner et al., 2007; Fossdal et al., 2007; Ditmarová et al., 2010). Effective quantum yield also decreased significantly, to about 50%, with increasing drought (Bigras, 2005; Blödner et al., 2005; Lu et al., 2007; Fossdal et al., 2007; Blödner et al., 2007), even before photoinhibition occurred (Eastman and Camm, 1995). The reduction however, was largely reversible in darkness. Furthermore, the efficiency of open reaction centers decreased significantly with increasing drought (Eastman and Camm, 1995; Bigras, 2005). This can be due to damaged reoxidation of the primary electron acceptor (QA ) of PSII by light induced protein damage in the adjacent acceptor (QB ), the second electron accepting plastoquinone of PSII (Kyle et al., 1984). Photochemical quenching also decreased with low soil moisture levels (Bigras, 2005), but only under high irradiances (Eastman and Camm, 1995). Non-photochemical quenching (qN) increased (Eastman and Camm, 1995; Bigras, 2005; Lu et al., 2007). This indicates an enhanced capacity to dissipate excess energy at PSII. The increase in qN was higher in sun than in the shade (Duan et al., 2005), which corresponds to the location where the dissipation of energy is most necessary. After ∼3 weeks, photosynthetic processes recovered to their pre-stress values (Pukacki and Kamińska-Rozek, 2005). Drought stress resulted in alterations in protein composition. Several proteins involved in photosynthesis were increased, for example those of the oxygen-evolving complex and of Rubsico (Blödner et al., 2007). Also chitinase transcripts and dehydrins of the stressed plants increased (Blödner et al., 2005; Fossdal et al., 2007; Eldhuset et al., 2013). 90 CHAPTER V. DROUGHT IMPACTS ON EUROPEAN FOREST SPECIES Antioxidants Plants have different antioxidants that can have a protective role during drought stress. Both chlorophyll a and b decreased in severely drought stressed Picea seedlings, but their ratio remained constant (Wallin et al., 2002; Pukacki and KamińskaRozek, 2005; Barsi et al., 2009; Ditmarová et al., 2010). Mild stress did not affect the chlorophyll content (Lu et al., 2007; Barsi et al., 2009; Ditmarová et al., 2010). Carotenoid concentration also decreased (Pukacki and Kamińska-Rozek, 2005; Barsi et al., 2009). This lead to a lower chlorophyll:carotenoid ratio that indicates a higher capacity to dissipate excess energy (Munné-Bosch and Alegre, 2000). After ∼3 weeks of recovery, chlorophyll a and carotenoid concentrations reached their control values, while chlorophyll b concentration remained ∼20% lower (Pukacki and Kamińska-Rozek, 2005). Proline accumulated in stressed spruce plants after 2 weeks (Yang et al., 2010; Ditmarová et al., 2010). This antioxidant has osmoprotective function that can help to preserve the protein structures and enzyme activities (Arndt et al., 2001). It can furthermore scavenge hydroxyl and other free radicals that may increase during drought (Smirnoff and Cumbes, 1989; Kishor et al., 1995; Kamińska-Rozek and Pukacki, 2004), even as ascorbate peroxidase (APX), catalase (CAT), peroxidase (POD) and superoxide dismutase (SOD), that also increased under drought stress (Duan et al., 2005; Kamińska-Rozek and Pukacki, 2004; Lu et al., 2007; Duan et al., 2007; Yang et al., 2010). The concentration of malondialdehyde (MDA) may reflect the degree of lipid peroxidation and membrane damage due to drought stress by active free radicals (Menconi et al., 1995). Its concentration increased significantly with increasing stress (Kamińska-Rozek and Pukacki, 2004; Lu et al., 2007). It may cause an enhanced membrane permeability leading to electrolyte leakage (Campos et al., 2003). Other changing components due to droughts stress were: ABA concentration that increased (Lu et al., 2007; Duan et al., 2007), whilst glutathione, ascorbic acid, alpha-tocopherol, guaiacol peroxidase and flavolnoids declined (Kamińska-Rozek and Pukacki, 2004). Xylem embolism Xylem embolism is a phenomenon where the water column in conduits is filled with air or water vapor and becomes disrupted. The xylem potential at which photosynthesis declines, lays between -1 and -1.6 MPa (Townend, 1993; Eastman and Camm, 1995; Tan and Blake, 1997), and stomata begin to close at approximately -1.6 MPa (Grossnickle and Blake, 1987; Blake and Li, 2003), although this can fluctuate depending on the level of soil drought and air humidity (Sellin, 1998). Significant xylem embolism occurred at values lower than -2.5 MPa (Sperry and Tyree, 1990; Cochard, 1992; Eastman and Camm, 1995), and P50 was found to be between -2.3 and -3.5 MPa, depending on the age and the properties of the wood (Lu et al., 1996; Mayr and Cochard, 2003; Rosner et al., 2006). Most embolism seemed to occur in the soil and trunk compartments, as no embolism was detected in the branches (Lu et al., 1996). Recovery of embolism may be difficult or even impossible in spruce, as conifers are not able to refill embolized xylem due to the sealing of the pit membranes to the pit pores (Sperry and Tyree, 1990). Also Blake and Li (2003) found that hydraulic and stomatal conductance did not recover after rehydration. Morphological changes The maintenance of turgor is crucial for the viability of plants. Due to water loss, the turgor can be lost. The elasticity of the cell wall can overcome this problem and it has been found to increase with drought stress (Major and Johnsen, 1999). Although another study reports no such changes (Blake and Li, 2003). The maintenance of turgor can also be provided by active osmotic adjustment (Major and Johnsen, 1999). 2. DROUGHT AND HEAT STRESS ON 5 MAIN EUROPEAN FOREST TREE SPECIES 91 Although drought can also cause osmotic stress, which can damage cell membranes (Major and Johnsen, 1999, 2001). Drought can also cause morphological changes in wood which was also observed for Picea sp. Water stress caused less early and late wood tracheids, decreasing the radial increase (Jyske et al., 2010; Eldhuset et al., 2013). Furthermore the cell wall thickness was higher due to drought, mostly in the late wood (Jyske et al., 2010). This increased wall thickness resulted in an increased wood density. This may, in the future, lead to changed wood properties (Jyske et al., 2010). Root growth, mass and density may decrease during and after drought (Beier et al., 1995; Blanck et al., 1995; Darlington et al., 1997; Eldhuset et al., 2013). Although increased fine root biomass and production were found by Gaul et al. (2008); Major et al. (2012), compared to a control plot during and after drought treatment. There was no increased dieback of fine roots during drought and post-drought periods (Blanck et al., 1995). Leaves chlorosis increased after a drought treatment (Dambrine et al., 1993; Fan and Blake, 1997), and needle biomass decreased during severe drought (Eldhuset et al., 2013). Needle loss was not enhanced (Dambrine et al., 1993). Further changes were observed in the nutrient status of the needles, with decreased calcium and magnesium concentration (Blanck et al., 1995) and increased needle leakage (Blake and Li, 2003). Other morphological changes were found on sub-cellular level with decreased chloroplast and mitochondria volume (Zellnig et al., 2010), and a decreased specific leaf area (Duan et al., 2011). Height and radial increment decreased with decreasing soil water levels (Spiecker, 1990; Blanck et al., 1995; Beier et al., 1995; Wallin et al., 2002; Dohrenbusch et al., 2002; Büntgen et al., 2006; Fossdal et al., 2007; Chhin and Wang, 2008; Jyske et al., 2010; Duan et al., 2011; Sohn et al., 2012; Eldhuset et al., 2013). Maybe partly due to limited carbon allocation (Sohn et al., 2012). The limitation in height and radial increments depends on the length of the stress period (Blake and Li, 2003). Furthermore the position and structure of the tree may also influence the drought impacts as trees with larger crowns were found to be less affected that those with shorter crowns (Spiecker, 1990). Lagged effects and acclimation Although many direct drought and heat effects are found, lagged effects from up till 5 years after the drought are observed in elevated mortality rates and vulnerability, either due to desiccation, insect attacks (Spiecker, 1990; Berg et al., 2006; Bigler et al., 2007), or decreased growth rates (Dohrenbusch et al., 2002). Management strategies could be applied by choosing provenance from drier and warmer sites. As individuals from drought-tolerant populations were better able to maintain turgor and photosynthesis under drought stress (Tan and Blake, 1997; Major et al., 2012). Furthermore, the decrease in chlorophyll due to drought was lower in plants from a dry population, and also their antioxidant level was higher (Duan et al., 2005). Tree resilience may also be increased by extensive thinning (Misson et al., 2003; Sohn et al., 2012; Kohler et al., 2010), as was the case for Quercus ilex. 2.2 Pinus sylvestris The next species that will be discussed is Pinus sylvestris, or Scots pine. Pinus sylvestris is one of the most abundant tree species in Europe and has a geographical range from the Mediterranean region to the boreal forests (Martín et al., 2010; Matías and Jump, 2012). It is a drought avoiding, or isohydric species (Cochard, 1992; Irvine et al., 1998; Gea-Izquierdo 92 CHAPTER V. DROUGHT IMPACTS ON EUROPEAN FOREST SPECIES et al., 2014). Its southern limits are mostly determined by water availability and elevated temperatures (Rebetez and Dobbertin, 2004; Brunner et al., 2009; Matías and Jump, 2012). In the southern regions it is often outcompeted by more drought-resistant species (Sykes and Prentice, 1995). Photosynthesis Drought induced a quick decrease in stomatal conductance and stomatal closure in Scots pine (Irvine et al., 1998; Croisé et al., 2001; Cregg and Zhang, 2001; Zweifel et al., 2007, 2009), thereby limiting both the occurrence on xylem embolism (Irvine et al., 1998; Perks et al., 2004), and net photosynthesis (Holst et al., 2008). Besides the stomatal limitations, also non-stomatal limitations to net photosynthesis were found, that were most important during severe water stress (Kellomäki and Wang, 1996). The photochemical efficiency was only affected when water stress was very high (Pearson et al., 2013). This may explain the rapid recovery after the period of stress (Pearson et al., 2013; Bansal et al., 2013). Due to stomatal closure, WUE increased with increasing drought stress (Cregg and Zhang, 2001; Bansal et al., 2013). Xylem embolism Because of the fast stomatal closure, xylem embolism does not occur frequently. The water potential that caused significant xylem embolism was around -2.5 MPa and P50 was approximately 3 MPa (Cochard, 1992). Water potentials recovered and xylem vessels were refilled in a few weeks after rewatering (Perks et al., 2004; Croisé et al., 2001). Morphological changes Lumen diameters were larger in drought stressed trees. They are cheaper in carbon and increased conductivity, but also more vulnerable to xylem embolism (Eilmann et al., 2011). Radial increment was lower, and wood formation ended earlier than in non-stressed trees (Gruber et al., 2010; Eilmann et al., 2011). Decreased soil water level also decreased the fine root length and biomass (Brunner et al., 2009; Pearson et al., 2013; Bansal et al., 2013), together with needle biomass. The latter mostly because of senescence of older needles during and after the stress period (Croisé et al., 2001; Rebetez and Dobbertin, 2004; Zweifel et al., 2007; Martínez-Vilalta et al., 2008; Galiano et al., 2010). Stem diameter increase was reduced during moisture deficit (Croisé et al., 2001; Bigler et al., 2006; Eilmann et al., 2011; Pearson et al., 2013), and persistent drought stress resulted in smaller ring widths on the long-term (Bigler et al., 2006). Although other studies find little or positive effects of warm and dry summers on radial growth (Polacek et al., 2006; Bansal et al., 2013). Moderate water stress may not cause a decrease in stored carbohydrates (Croisé and Lieutier, 1993; Gruber et al., 2012). However, Galiano et al. (2010) found a decrease in carbohydrate reserves. Lagged effects and acclimation Several studies have reported Scots pine die-back due to drought stress (Bigler et al., 2006; Galiano et al., 2010; Hereş et al., 2012; Vilà-Cabrera et al., 2014). However other factors may also determine the survival chances of stressed trees. Manion (1991) defined three categories of stress: predisposing, inciting and contributing factors. Predisposing factors act on the long term and could be related to site characteristics such as poor nutrient availability. An inciting factor acts on a short time-scale, such as drought stress. Insect attack could be a contributing factors. The latter does not kill the tree if it has not been exposed before to the predisposing and inciting factors (Manion, 1991). In the case of Pinus sylvestris, different studies have reported the importance of both the drought stress 2. DROUGHT AND HEAT STRESS ON 5 MAIN EUROPEAN FOREST TREE SPECIES 93 and secondary pathogens or insect outbreaks (Croisé et al., 2001; Rebetez and Dobbertin, 2004; Dobbertin et al., 2007; Polomski and Rigling, 2010; Krams et al., 2012). Thinning may be a proper management tool to increase drought resistance of trees, especially through diminishing competition (Bigler et al., 2006; Giuggiola et al., 2013). Acclimation may also occur on a longer time scale, as individuals from warmer and drier sites were found to be more drought resistant than individuals from cooler and wetter sites (Cregg and Zhang, 2001; Martín et al., 2010; Matías et al., 2014). 2.3 Fagus sylvatica Beech, Fagus sylvatica, is a very important broad-leaved species in Western and central Europe (García-Plazaola and Becerril, 2000; Peuke et al., 2002). Beech is relatively drought sensitive, but can both exhibit an isohydric or an anisohydric strategy. The species is sometimes replaced in the Mediterranean region by more drought resistant species (Aranda et al., 2012). Photosynthesis One of the first effects to soil water depletion of drought sensitive species such as Fagus sylvatica (Zapater et al., 2013; Bréda et al., 2006) is a decrease of stomatal conductance Tognetti et al. (1995); Sánchez-Gómez et al. (2013), which down-regulates water use (Chaves, 1991) and decreases net photosynthesis. Net photosynthesis was found to decrease from about 50% to almost complete inhibition after two weeks of drought (Tognetti et al., 1995; Gallé and Feller, 2007; Robson et al., 2009; Simpraga et al., 2011). The inhibition of the photosynthesis was probably caused by the almost complete closure of the stomata, and an irreversible change in their structure (Gallé and Feller, 2007). However, Aranda et al. (2012) found, although with lower water stress, that stomatal closure only contributed with 33% to the reduction in net photosynthesis, while the largest part (45%) was due to biochemical limitation. Mesophyll limitation contributed with 22%. After 4 weeks after rewatering, after severe drought stress, (Gallé and Feller, 2007; Simpraga et al., 2011) found restored net photosynthesis values, whilst the restoration of stomatal conductance was incomplete. Similar results were found by Tognetti et al. (1995), indicating that the intrinsic WUE was, and stayed, higher in stressed trees (Peuke et al., 2002; Gallé and Feller, 2007; Aranda et al., 2012; Sánchez-Gómez et al., 2013). The maximum photochemical efficiency of PSII was found to decrease due to water stress (García-Plazaola and Becerril, 2000), but increased again within two weeks after rewatering (Gallé and Feller, 2007). The decrease was attributed to the quenching of Fm and initial fluorescence (F0), where a reduction in Fm is most important in the early stages of drying, but during more severe desiccation the increase in F0 is most critical. Gallé and Feller (2007); Sánchez-Gómez et al. (2013) found a non-significant effect of drought on the effective quantum yield, while Robson et al. (2009) found a small but significant decrease. The efficiency of open reaction centers also decreased slightly (Robson et al., 2009). Non-photochemical quenching did not change significantly (Gallé and Feller, 2007), but to fight the decline of photochemical efficiency and the risk of photooxidation, antioxidant levels were increased, such as violoxanthin, antheraxanthin, zeoxanthin (VAZ) and a-tocopherol (García-Plazaola and Becerril, 2000). Gallé and Feller (2007) however, found the pool of xanthophyll cycle pigments (VAZ) to be decreased during severe drought. Chlorophyll and carotenoids decreased due to drought (Gallé and Feller, 2007; García-Plazaola and Becerril, 2000). Both 94 CHAPTER V. DROUGHT IMPACTS ON EUROPEAN FOREST SPECIES the rate of electron transport and the maximum carboxylation rate decreased in August, although this might be due to leaf aging instead of water stress (Aranda et al., 2012). Xylem embolism P50 is between -2.5 and -3.2 MPa (Lemoine et al., 2002; Wortemann et al., 2011). The water potential with significant loss of hydraulic conductivity and a chance on the occurrence of xylem embolism is in between -1.5 and -2.25 MPa (Fotelli et al., 2001; Lemoine et al., 2002). Differences between individuals and branches can be due to different irradiance levels, where sun-exposed branches are less vulnerable than shaded branches (Cochard et al., 1999; Lemoine et al., 2002). Furthermore they can be due to different soils (Čermák et al., 1993). The reduction of water potential due to drought is different in the various tree parts. Water potential was reduced more strongly in the roots than in the shoots (Peuke et al., 2002). Irrigation caused a fast response in water stressed trees and indicates resistance of short-term water stress to cavitation (Čermák et al., 1993; Gallé and Feller, 2007). Lemoine et al. (2002) also concluded that embolism remained low at the end of the summer. They assigned this to the fast stomatal closure that inhibits evaporation and thus embolism. Morphological changes Decreasing precipitation reduced the fine root biomass in different soil layers (Meier and Leuschner, 2008). The fine roots were also thinner in drier stands. So it seems that beech trees maintain root water absorption by lowering their root potential and thus increasing absorption per unit root surface during mild drought periods (Meier and Leuschner, 2008). Biomass allocation to the roots was increased (Barthel et al., 2011). Also the composition of the roots was found to be different after drought; more fibrous roots were found in declining trees (Čermák et al., 1993). The specific leaf area decreased under water deficit (Sánchez-Gómez et al., 2013), but nitrogen content increased. Furthermore, sclerophyllous adaption to mild drought was found (Bussotti et al., 1995a), and leaf shedding during severe drought (Bréda et al., 2006). The growth of beech seedlings was lowered during moderate water stress (Madsen and Larsen, 1997; Fotelli et al., 2001; Piovesan et al., 2008). Furthermore, the stem diameter can decline during summer (Ježík et al., 2011). While regular variations in shrinkage and swelling of the stem diameter occur the whole year around, drought causes a sharper midday decline and a depletion of internal water reserves on the longer term (Simpraga et al., 2011). This suggests that in areas exposed to regular drought, reserves, both in water and nutrients, are important for plant survival and resilience (Scartazza et al., 2013). Lagged effects and acclimation Several studies show that beech can survive moderate and short-term droughts (Čermák et al., 1993), but that longer and more severe droughts may decrease their survival chances, which has already been showed in the Mediterranean region (Piovesan et al., 2008; Robson et al., 2009). This may partly be due to drought induced nitrogen shortage (Geßler et al., 2004). The nitrogen shortage may be a result of an altered water balance of the tree that negatively affects the capacity of mycorrhizal fungi to take up nitrogen (Geßler et al., 2005). Although the ectomycorrhiza fungi can allow for more water uptake during mild stress (Pena et al., 2013). Acclimation may occur as plants from different provenances and different climates respond differently to drought stress (Peuke et al., 2002; García-Plazaola and Becerril, 2000; Peuke and Rennenberg, 2011). Beech trees from drier regions were on average more sensitive to drought, and so might be better adapted to future drier conditions (Tognetti et al., 1995; Fotelli et al., 2009; Rose et al., 2009; Stojnić et al., 2. DROUGHT AND HEAT STRESS ON 5 MAIN EUROPEAN FOREST TREE SPECIES 95 2012; Sánchez-Gómez et al., 2013). A management strategy for increased survival chances of drought stress is thinning (van der Maaten, 2013). Also a reduced competition with other species may favor survival (Fotelli et al., 2002; Geßler et al., 2007). 2.4 Quercus robur & petraea Pedunculate (Quercus robur) and sessile oak (Quercus petraea) are, after beech, the most common summer green broadleaved species in central Europe (Thomas and Gausling, 2000). On average, Quercus species are relatively deeply rooted and relatively well resistant to drought stress, and are isohydric (Abrams, 1990; Cochard, 1992). However, their susceptibility to defoliation by insects may decrease their overall resistance (Gieger and Thomas, 2005). Overall, Quercus petraea seems to be somewhat more drought resistant than Quercus robur (Cochard, 1992; Dickson and Tomlinson, 1996; Demeter et al., 2014). Photosynthesis Stomatal conductance and photosynthesis decreased with drought stress (Epron and Dreyer, 1993b; Bréda et al., 1993; Thomas and Gausling, 2000; Grassi et al., 2005). The reduction in photosynthesis was happening slowly and was still positive at water potentials of -3 MPa (Epron and Dreyer, 1993a,b), whilst the reduction in stomatal conductance occurred more suddenly (Epron and Dreyer, 1993a; Cochard et al., 1996a; Čater and Batič, 2006). The latter could have resulted in enhanced survival by avoiding xylem embolism (Cochard et al., 1996a). Epron and Dreyer (1993b); Simonin et al. (1994) did not find a significantly decrease due to drought stress of maximum photochemical efficiency of PSII, effective quantum yield, and the efficiency of open reaction centers. However, some studies found these variables to decrease (Epron and Dreyer, 1993a). The inhibition of photosynthesis of severely stressed seedlings may have been a result from both excess light energy at PSII and drought stress. This excess light energy also decreased photochemical quenching and the primary electron acceptors of PSII (Epron and Dreyer, 1993a). Antioxidants Active oxygen compounds were found to increase due to drought stress. This probably led to an increased level of antioxidants, such as gamma-tocopherol (Hendry et al., 1992). Also osmoprotectants, such as proline, methylproline and hydroxyproline, were found to increase (Spieß et al., 2012; Hu et al., 2013). Osmotic adjustment may be greater in natural occurring droughts that develop slowly (Collet and Guehl, 1997). Ascorbic acid however, was found to decrease (Hendry et al., 1992), even as glutathione (Hu et al., 2013). The latter was not detected by Hendry et al. (1992) with increased water stress. ABA concentrations did not change significantly in leaf tissues (Triboulot et al., 1996; Fort et al., 1997), however it did increase in the root xylem (Fort et al., 1997). Xylem embolism The vulnerability to cavitation was greater in Quercus robur than in Q. petraea (Cochard, 1992; Bréda et al., 1993). This is also apparent in the P50 , which was between -2.7 and -3.5 MPa in Q. robur and between -3.3 and -4.3 MPa in Q. petraea (Cochard, 1992; Higgs and Wood, 1995). Whereas cavitation becomes significantly higher around -2.5 MPa in the former species Simonin et al. (1994) and around -2 to -2.5 MPa in the latter (Thomas and Hartmann, 1996; Cochard et al., 1996a). Most stomata were closed at about -1.8 MPa, while net photosynthesis was almost zero at -2.8 MPa for both oak species (Vivin et al., 1993; Čater and Batič, 2006). The somewhat higher vulnerability to cavitation of Q. robur makes the species less drought resistant (Cochard, 1992; Higgs and Wood, 1995). The 96 CHAPTER V. DROUGHT IMPACTS ON EUROPEAN FOREST SPECIES petioles were found to be more vulnerable than the twigs (Cochard, 1992), although another study found similar vulnerabilities (Simonin et al., 1994). The water potentials were restored till pre-drought conditions after rewatering (Cochard, 1992). Morphological changes Early wood seems to be most important for the transport of water in trees and wider vessels transport water more efficiently (Hagen-Poiseulle). This may be important for the fitness of the trees, as dead oak trees had much smaller vessel widths than healthy oak trees (Tulik, 2014). The leaf area and biomass of drought stressed trees was decreased (Gieger and Thomas, 2002; Arend et al., 2011), while the fine root length and biomass were unaffected by drought (Arend et al., 2011). This resulted in a decreased leaf-fine root ratio during drought (Thomas and Gausling, 2000). The absence of response of the roots might be due to the relatively deep root system that enables the access to water even during periods of drought (Arend et al., 2011). Leaf abscission increased during the drought stress, starting some weeks after the stress period began (Rust and Roloff, 2004). This may be a method to reduce transpiration. Growth was slowed down or even ceased in stems and shoot during drought (Saveyn et al., 2007; Arend et al., 2011; Spieß et al., 2012). Osmotic pressure was lowered in drought stressed trees (Thomas and Gausling, 2000). Soluble carbohydrates can contributed to osmoregulation during drought (Epron and Dreyer, 1996; Sergeant et al., 2011). This may explain the increased level of leaf carbohydrates, glucose, fructose, galactose, mannitol and sucrose found by (Spieß et al., 2012). Lagged effects and acclimation Mortality of oak trees can occur from one up to several years after the drought event (Becker and Becker, 1982; Vivin et al., 1993). Acclimation may take place as trees from warmer and drier sites have a higher frequency of drought-tolerant alleles. An increased frequency and severity of extreme droughts or heatwaves may thereby lead to reduced diversity loss, making populations less adaptive to other stresses (Borovics and Mátyás, 2013). 2.5 Quercus ilex Holm oak, or Quercus ilex, is a widespread species occupying large parts of the Mediterranean basin (Terradas and Savé, 1992; Ogaya and Peñuelas, 2007). It is anisohydric and highly tolerant to drought stress (Tognetti et al., 1998; Martínez-Ferri et al., 2000). This is partly due to its sclerophylly and its deep root system (Terradas and Savé, 1992; Filella et al., 1998; Gratani and Varone, 2004). Because of the usually very dry and hot summers, holm oak usually has two peaks of vegetation growth, one in the spring and one in the autumn (Ogaya and Peñuelas, 2004; Gratani et al., 2013). Due to its resprouting ability, it was planted in large parts of Spain and Italy as coppice systems (Serrada et al., 1992). Photosynthesis Stomatal control is efficient in Quercus ilex, and the conductance is significantly lowered during periods of water stress, together with net photosynthesis (Terradas and Savé, 1992; Peñuelas et al., 1997; Filella et al., 1998; Llusià and Peñuelas, 1998; Peñuelas and Llusià, 1999a; Fotelli et al., 2000; Loreto et al., 2001; Gulías et al., 2002; Martínez-Vilalta and Piñol, 2002; Pesoli et al., 2003; Serrano and Peñuelas, 2005; Baquedano and Castillo, 2006; Gimeno et al., 2009; Limousin et al., 2010b; Misson et al., 2010; Rodríguez-Calcerrada et al., 2011; Gratani et al., 2013; Martin-Stpaul et al., 2013; Tsonev et al., 2014). During autumn and winter, net photosynthesis and stomatal conductance increase again (Peñuelas 2. DROUGHT AND HEAT STRESS ON 5 MAIN EUROPEAN FOREST TREE SPECIES 97 and Llusià, 1999a; Serrano and Peñuelas, 2005; Vaz et al., 2010; Misson et al., 2010; Gratani et al., 2013), although the rate of recovery may be dependent on leaf age (Limousin et al., 2010b). Mesophyll conductance also decreases significantly during drought (Limousin et al., 2010b), although net photosynthesis may be more limited by stomatal than by mesophyll conductance (Misson et al., 2010; Limousin et al., 2010b). Fleck et al. (2010); Gallé et al. (2011) however, find that mesophyll conductance is the main factor limiting net assimilation, but this might be dependent on the severity of the drought stress (Misson et al., 2010). Biochemical limitation was found to have the lowest impact (Limousin et al., 2010b; Gallé et al., 2011), or similar to that of mesophyll limitation (Misson et al., 2010). WUE is usually higher during drought due to less water loss because of stomatal closure. This was also found for holm oak (Gratani and Varone, 2004; Gimeno et al., 2009; Limousin et al., 2010b). Although (Pesoli et al., 2003; Serrano and Peñuelas, 2005) found a reduced WUE under water stress. This may be due to increased respiration due to elevated temperature (Reichstein et al., 2002). Maximum photochemical efficiency of PSII decreased during water stress (Filella et al., 1998; Martínez-Ferri et al., 2000; Peña-Rojas et al., 2004; Baquedano and Castillo, 2006; Ogaya et al., 2011; Valero-Galván et al., 2013). This may indicate down-regulation of the non-cycling electron transport (Chaumont et al., 1995). Rambal et al. (1994); Méthy et al. (1996); Valladares et al. (2005) however, find stable values for water potentials up to -4 to -5 MPa, where after a significant decrease is found. The values were measured to be close to 0 at water potentials of -7 MPa. These results suggest that stomatal regulation is responsible for the reduced net assimilation (Rambal et al., 1994). Effective quantum yield, the non-cycling electron transport, also showed a significant decrease during soil water depletion (MartínezFerri et al., 2000; Peña-Rojas et al., 2004; Baquedano and Castillo, 2006; Valero-Galván et al., 2013). Although other studies find a small, non-significant decrease (Echevarría-Zomeño et al., 2009). The efficiency of open reaction centers decreased as well during drought stress (Peña-Rojas et al., 2004; Baquedano and Castillo, 2006). A decrease in this efficiency may imply increased excitation energy of quenching processes in the light-harvesting antennae of PSII (Horton et al., 1994). Such a mechanism could protect the plant from photodamage. Furthermore, photochemical quenching decreased with increasing drought stress and decreasing effective quantum yield (Peña-Rojas et al., 2004; Baquedano and Castillo, 2006). Non-photochemical quenching increased with decreasing maximum photochemical efficiency and drought stress (Martínez-Ferri et al., 2000; Gulías et al., 2002). Chlorophyll and antioxidants The change of chlorophyll concentration under drought stress is debated, and some studies find a decrease (Baquedano and Castillo, 2006), an increase (Gulías et al., 2002), or no change (Vaz et al., 2010). Furthermore, the concentration of carotenoids were decreased (Baquedano and Castillo, 2006). Other antioxidants, such as alpha-tocopheral, ascorbic acid, malondialdehyde (MDA) and VAZ (violaxanthin, antheraxanthin and zeaxanthin), the latter related to non-photochemical quenching, increased during low soil moisture levels (Gulías et al., 2002; Munné-Bosch et al., 2004; Baquedano and Castillo, 2006; Camarero et al., 2012). Xylem embolism P50 , the pressure causing a 50% loss of hydraulic conductivity, was found to be around -3 and -5.6 MPa (Corcuera et al., 2004; Limousin et al., 2010a; Martin- 98 CHAPTER V. DROUGHT IMPACTS ON EUROPEAN FOREST SPECIES Stpaul et al., 2013), and stomatal closure occurred at water potentials between -4 and -4.5 MPa (Pesoli et al., 2003). Water potentials of -6 MPa were found during summer (Tognetti et al., 1998). And while no injuries occurred at potentials lower than -3 MPa (Pigott and Pigott, 1993), longer-term exposures may lead to critical damage (Terradas and Savé, 1992). Although there are differences in vulnerability by different tree parts (Martínez-Vilalta and Piñol, 2002). Morphological adaptations Biomass allocation to the roots is found to increase during drought (Baquedano and Castillo, 2006). However Manes et al. (2006); Corcobado et al. (2014) find no change in root:shoot ratio, root biomass and root length. Drought can also cause morphological changes in the wood. Corcuera et al. (2004) found narrower xylem vessels in a dry summer year. Changing leaf properties have also been observed. Although Rodríguez-Calcerrada et al. (2011); Martin-Stpaul et al. (2013) did not find a change in leaf mass per area (LMA) during drought stress, Ogaya and Peñuelas (2006) found a small decrease and Limousin et al. (2010b); Camarero et al. (2012) found an increase in leaf mass per area. These contradicting results may be due to different experimental set-ups or the use of different subspecies. Furthermore, LMA may be a more important protection against cold than against drought stress (Ogaya and Peñuelas, 2007). Also the position of the leaves on the tree plays a role, as sun leaves have higher LMA, thickness and lower area than shade leaves (Ogaya and Peñuelas, 2006). Leaf area also decreased with water stress (Baquedano and Castillo, 2006; Limousin et al., 2010b; Martin-Stpaul et al., 2013; Rodríguez-Calcerrada et al., 2011), as well as leaf production, resulting in a lower leaf area index LAI (Ogaya and Peñuelas, 2006, 2007; Rodríguez-Calcerrada et al., 2011; Martin-Stpaul et al., 2013). Furthermore, the leaf to sapwood ratio decreased (Martin-Stpaul et al., 2013). Defoliation and leaf damage occurred after severe drought stress (Peñuelas et al., 2000; Pesoli et al., 2003; Corcuera et al., 2004). The non-structural carbon reserves were reduced due to low soil moisture levels (Galiano et al., 2012), although such a reduction was not found by (SanzPérez et al., 2009). Furthermore, the buds of water stressed seedlings were found to be bigger and the bud-burst was altered (Sanz-Pérez and Castro-Díez, 2010; Misson et al., 2010). Ogaya and Peñuelas (2004) however, did not find a change in flowering and fruit growth in water stressed individuals. BVOC Another process that may act on the carbon balance, although with a very limited effect, is the emission of biogenic volatile organic compounds. Quercus ilex emitted more terpenes under moderate, but not under severe water stress. This might be because the substances cannot be stored (Llusià and Peñuelas, 1998; Blanch et al., 2009), which makes the emission rate more dependent on the availability of recently fixed carbon than on temperature or soil moisture availability (Peñuelas and Llusià, 1999a; Staudt et al., 2002). The percentage of emitted newly synthesized carbon was between 6% and 13% (Peñuelas and Llusià, 1999b). Highest concentrations of emitted substances were found to be alpha- and beta-pinene, limonene, sabinene and myrcene (Llusià and Peñuelas, 1998; Loreto et al., 2001; Staudt et al., 2002; Plaza et al., 2005; Blanch et al., 2009). Less emitted substances were camphene, alpha-tujene, delta3-carene, 2-carene, p-cymene, alpha- and betaphellandrene, cis and trans-beta-ocimene, gamma-terpinene, linalool, ocimene, caryophyllene and alpha-caryophyllene (Staudt et al., 2002; Blanch et al., 2007, 2009). Both alphaand beta-pinene emissions were stronger during the warm summer temperatures (Peñuelas 3. MODEL DESCRIPTION 99 and Llusià, 1999a). Terpene emissions may have a protecting role for the photosynthetic functioning or tissue damage of environmental stresses, as has been suggested for isoprene (Litvak et al., 1996). Plasticity Drought stress also caused differentiation in the composition of proteins (Jorge et al., 2006; Echevarría-Zomeño et al., 2009). The ones increasing are proteins involved in photosynthesis, several proteins involved in storage, possibly because of increased need for stored carbohydrates, a protein involved in carbohydrate metabolism, and a protein involved in stress regulation (Jorge et al., 2006; Echevarría-Zomeño et al., 2009). The results may be dependent of the duration of the water stress, as different stress protection mechanisms play a role at different time scales (Jorge et al., 2006). Lagged effects and acclimation Severe drought stress was found to lead to permanent damage (Peñuelas et al., 2000; Galiano et al., 2012), long-term loss of reserves (Galiano et al., 2012), or mortality (Filella et al., 1998; Peñuelas et al., 2000; Valladares et al., 2005; Galiano et al., 2012) up to several years after the drought event took place. But acclimation can also take place as plants from warm and dry regions were found to be more drought resistant than individuals from colder and wetter climates (Bussotti et al., 1995b; Pesoli et al., 2003; Villar-Salvador et al., 2004; Sánchez-Vilas and Retuerto, 2007; Bonito et al., 2011; Corcobado et al., 2014). For example, osmotic adjustment (Villar-Salvador et al., 2004), acorn size (Bonito et al., 2011), and photosynthetic capacity (Sánchez-Vilas and Retuerto, 2007) were enhanced. Although Limousin et al. (2010b); Rodríguez-Calcerrada et al. (2011) did not find long-term acclimation during several years of study. Intensive thinning was found to reduce competition and improve the resistance of holm oak forests to severe water stress (Galiano et al., 2012). All the drought stress effects in this section are mainly based on observations at local scales and small scale experiments. To study impacts on larger scales a different approach is needed, such as the use of vegetation models. However, before simulating drought impacts on the European scale, simulations on site scales are necessary to evaluate the model. This will be done in the coming sections. First the vegetation model will be described, followed by the simulation results at 5 different forested sites in Europe (Table V.1). The sites are chosen for their species composition and localization. A discussion and conclusion will end the chapter. 3 3.1 Model description ORCHIDEE The vegetation model used in this study is ORCHIDEE-CAN (Naudts et al., 2015), an adapted version of the Organising Carbon and Hydrology In Dynamic Ecosystems (ORCHIDEE) model (Krinner et al., 2005). ORCHIDEE consists out of two modules: a hydrological module SECHIBA and a carbon module STOMATE. The former describes the energy and water exchanges in the soil-vegetation-atmosphere continuum. The latter simulates the vegetation processes such as carbon dynamics and the phenology. The vegetation in the model is represented by 13 plant functional types (PFT): 1 bare ground, 2 crop types, 2 grass types and 8 forest types. 100 CHAPTER V. DROUGHT IMPACTS ON EUROPEAN FOREST SPECIES 3.2 ORCHIDEE-CAN In ORCHIDEE-CAN some structural changes have been implemented. I will briefly discuss the changes that are most relevant for my study. For a detailed description of all changes I refer to Naudts et al. (2015). 3.2.1 From PFT to species A first important adaptation is the change from PFTs to species. The main European forest trees species have been parameterized. These are: Fagus sylvatica, Pinus pinaster, Quercus ilex, a group with Quercus robur and Q. petraea, Betula sp., Pinus sylvestris and Picea sp. The last three species are parameterized for both the boreal and temperate region. Furthermore, Populus sp. was described for short rotation coppice and observations from Larix sp. were included for a better description of boreal needleleaf deciduous forest. 3.2.2 Carbon allocation and canopy structure In ORCHIDEE-CAN the canopy structure is much more detailed than in the standard version of ORCHIDEE (trunk). Different circumference classes are added, which supports a social position of the trees in the canopy. This allows for intra-tree competition as dominant trees capture more light, resulting in more photosynthesis and biomass accumulation for the most dominant trees. Furthermore, the different height classes also allow for a more consistent calculation of the absorbed light. In the trunk version this was based on LAI only, and not the physical height or canopy depth of the trees. The photosynthesis module in the ORCHIDEE-CAN includes a modified Arrhenius function. This allows for a temperature dependence accounting for a decrease in electron transport and carboxylation capacity at high temperatures. 3.2.3 Hydraulic architecture In the trunk version of ORCHIDEE, the soil hydrology had a direct influence on the water stress of the plant. In the new version, a specific hydraulic architecture with plant water supply is implemented in order to more realistically simulate water stress on plants. A drought stress factor is calculated as the ratio between the transpiration based on water supply and the transpiration based on atmospheric demand. When the former is exceeded by the transpiration based on energy supply, stomatal conductance is restricted. The water supply of the plants is a pressure difference between the soil and the leaves, and between the total hydraulic resistance of the roots, sapwood and leaves. Cavitation is calculated based on an s-shaped vulnerability curve and increases the sapwood resistance. ORCHIDEE-CAN uses an 11-layer hydrology scheme. 4 Results This section presents the preliminary results of model experiments with ORCHIDEE-CAN, trying to answer the main question of this chapter whether this state-of-the-art vegetation model can simulate summer drought stress on different tree species. Five sites with different species were chosen to represent the main tree forest species of Europe (Table V.1). There is a specific focus on the simulation of Quercus ilex, a drought-resistant tree species. 4. RESULTS 101 Table V.1: Site characteristics of the 5 sites used in this study. EB = Evergreen broadleaf, EN = Evergreen needleleaf, DB = Deciduous broadleaf. Puéchabon Brasschaat Tharandt Hesse Collelongo Country Latitude Longitude Altitude Species France 43◦ 44’N 3◦ 35’E 270 m Quercus ilex Germany 50◦ 95’N 13◦ 57’E 380 m Picea abies France 48◦ 40’N 7◦ 05’E 300 m Fagus sylvatica Italy 41◦ 51’N 13◦ 35’E 1550 m Fagus sylvatica Phenology EB Belgium 51◦ 18’N 4◦ 31’E 16 m Pinus sylvestris & Quercus robur EN & DB EN DB DB 4.1 Puéchabon A first site that will be evaluated is Puéchabon, a forest site located in the south of France with a Mediterranean climate. The choice of this specific site was made because of our interest in heat and drought stress on vegetation, which occurs every summer in this area. Furthermore, the Mediterranean region is expected to experience even warmer and drier summers in the future climate (Giorgi, 2006; Giorgi and Lionello, 2008; Somot et al., 2008; Nikulin et al., 2011), with probable negative effects for vegetation productivity (Hickler et al., 2012; van Oijen et al., 2014). The main tree species at Puéchabon is Quercus ilex or holm oak, a broadleaf evergreen species. The average canopy height is between 4 and 6 meters (Reichstein et al., 2002; Allard et al., 2008; Limousin et al., 2010a; Misson et al., 2010; Rambal et al., 2014) with a basal area of approximately 28 m2 ha−1 (Misson et al., 2010; Rodríguez-Calcerrada et al., 2014). Holm oak is a widespread species occupying large parts of the Mediterranean basin (Terradas and Savé, 1992; Ogaya and Peñuelas, 2007). It is highly tolerant to drought stress as described in section V.2.5. The site of Puéchabon has a very rocky soil. More than 75% of the upper 50 cm of soil is occupied by rocks, where 90% of the roots are found (Allard et al., 2008). Below 50 cm, rocks occupy more than 90% of the soil (J.-M. Ourcival, personal communication). The deepest roots can reach till 4.5 m depth (Allard et al., 2008). 4.1.1 GPP Modeled GPP is on average too low during spring, while during summer it is too high (Fig. V.1). This suggests that the drought stress is simulated incorrectly, which can have consequences when looking at a higher frequency and severity of heatwaves in a future climate. In reality, multiple climate extreme such as heatwaves may lead to forest die-off, but with insufficient water stress in the model this cannot be simulated. Incorrect GPP values can also be the result of inaccurate biomass allocation. Although, besides fruit mass, the allocation seemed to be reasonably well simulated. A study by Rambal et al. (2014) estimated fruit mass around 26.4 g cm−2 , while it was simulated much lower with ORCHIDEE-CAN. Increasing the fruit mass allocation in the model led to better fruit mass results (from approximately 10 to 23 g cm−2 ). The tree height slightly decreased (from ∼8 to 7.5 m), as less biomass was available for growth, and became closer to the observed values as well (5.5 m). This indicates that the hydraulics might be simulated incorrectly. A comparison of soil moisture with daily measurements with a neutron probe up to 4 m soil depth confirms this (Fig. V.2). While the Rainfall (kg m−2 s−1) 3 1 −3 −1 4e−08 20 −2e−08 Time Time ORCHIDEE−CAN FLUXNET 5 4 3 1 2 GPP (gC m−2 day−1) 6 7 −30 SW net (W m−2) CHAPTER V. DROUGHT IMPACTS ON EUROPEAN FOREST SPECIES 0 Temperature (°C) 102 ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ● ● ●● ●● ●● ●● ●● ●● ● ●● 2000 2001 2002 ● ●● ●● ●● ●● ●● ● Time ●● ●● ●● ●● ●● ●● ● ●● ●● ●● ●● ●● ●● ● ●● ●● ●● 2003 2004 2005 2006 Time (years) Figure V.1: Temperature (blue), rainfall (light blue) and shortwave net radiation (purple) anomalies and gross primary productivity (GPP) in Puéchabon from 2000 to 2006. Observed GPP in red and simulated in black. The grey vertical bars represent the summer months (June, July and August), where the model is in general not able to reproduce the observed summer stress in GPP. observed soil moisture level is around 90 mm in summer, the modeled soil moisture does not reach values below 110 mm, presumably leading to a lack of modeled water stress, and too high GPP. Two years with extreme summer stress are 2003 and 2006. Figure V.1 shows that this is mostly due to the high temperatures. Both precipitation and shortwave net radiation seem to have a lesser effect, as their values are not anomalous. 4.1.2 Soil hydraulics For calculating the root water potential in ORCHIDEE-CAN, it is assumed that the root potential is similar to the soil potential. Furthermore, there is no capacitance in the model. This resulted in too much water stress in some broadleaf tree species, leading to the addition of a tuning parameter (PST), to be able to reduce stress. The root water potential is therefore a function of the root density, the soil water potential and the tuning parameter. The default PST value is 0 MPa for Quercus ilex. We tried different values, from releasing stress (positive PST) to adding stress (negative PST). We found no changes for PST values as low as -2 103 200 100 150 Soil moisture (mm) 250 300 4. RESULTS ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● Observations ORCHIDEE−CAN 2004 2005 2006 Time (years) Figure V.2: Soil moisture in Puéchabon from 2004 to 2006, observed (red) and simulated (black). MPa. Below this value GPP started to react by showing more summer stress (lower GPP). An additional stress of -3.3 MPa resulted in the most realistic GPP values (Fig. V.3; note that only several PST values are shown for readability). Adding more stress resulted in zero production (Fig. V.4 darkbrown dotted line). Testing the same parameter values at Castelporziano in Italy, another site with Quercus ilex, led to better results with a tuning value of -3.2 MPa, thus very similar to the one found for Puéchabon. Although the oak trees on this site suffered much less from drought stress, probably due to its location close to the sea. Another reason why the soil moisture and the summer stress could be estimated wrongly is the vertical distribution of the soil moisture into the different soil layers and its evolution after a precipitation event. If the soil moisture would penetrate too fast into the deepest layers, the trees would not be able to subtract water from the soil. While the reverse could also be true. Two interesting seasons for studying the soil moisture are the very wet spring of 2004 and the very dry summer of 2006. Figure V.4 shows the water distribution per layer after several rain events in the spring of 2004. The distribution shows to be reasonable, with the top layer becoming wettest and drying out fastest. The lowest three layers react differently due to the saturated hydraulic conductivity that decreases exponentially below 30 cm, that corresponds to the depth below layer 8. The total modeled soil moisture is higher than the observations, from approximately 50 CHAPTER V. DROUGHT IMPACTS ON EUROPEAN FOREST SPECIES 8 104 PST 0 (default) PST −3.2 PST −3.3 PST −3.4 FLUXNET 6 ● ● ● ● ● ● ● ● ● 4 ● ● ● ● ● ● ● ● 0 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 2 GPP (gC m−2 day−1) ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 2002 ● ● ● ● 2001 ● ● ● ● 2000 ● 2003 2004 ● ● 2005 ● 2006 Time (years) Figure V.3: GPP in Puéchabon, from 2000 to 2006, observed (red) and simulated. The different colors and symbols correspond to different values of a tuning parameter for root water potential (see main text for more details). The grey vertical bars represent the summer months. ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ● ●●● ● ●●● ● ●● ● ● ● ●● ● ● ●● ● ● ● ● ●● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ●● ● ●● ● ●● ● ●● ● ● ●● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ●● ●● ●● ●● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ●● ●● ● ● ● ●● ● ●● ●● ● ● ● ● ● ●● ●● ● ●● ● ● ● ●●● ●● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ●●● ● ●● ● ● ● ●● ● ● ● ●●● ● ●● ● ●● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ●● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●●● ● ● ● ●● ● ●● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● Layer 1 Layer 2 Layer 3 Layer 4 Layer 5 Layer 6 Layer 7 Layer 8 Layer 9 ● Layer 10 Layer 11 Rainfall March April 1.0e−06 5.0e−07 Rainfall (kg m−2 s−1) 1.5e−06 ● ● 0.0e+00 0.3 0.2 0.1 Soil moisture (m3 m−3) 0.4 ● May Time (months) Figure V.4: Soil moisture per soil layer in Puéchabon, from March to May 2004. The different colors correspond to the 11 different soil layers of the model, layer 1 being the top layer. Precipitation is shown in blue. March April 1.0e−06 Rainfall (kg m−2 s−1) 5.0e−07 0.0e+00 250 200 150 Soil moisture (mm) ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ●● ●● ● ●● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 1.5e−06 105 300 4. RESULTS May Time (months) Figure V.5: Total soil moisture in Puéchabon, from March to May 2004, in observations (red) and the model (black). Precipitation is shown in blue. mm in the beginning, reaching to about 100 mm at the end of the spring of 2004 (Fig. V.5). Furthermore, the modeled soil moisture reacts less strongly to water input by precipitation and decays slower after rainfall. This may partly be responsible for the lack of summer stress in the trees. During the dry summer of 2006 we observe similar differences, with higher soil water levels in the model and a reduced reaction to precipitation (Fig. V.6). Furthermore it seems that there is a threshold value under which the soil moisture cannot go. A possibility in future work would be to force the summer soil moisture to different levels (e.g. 70-90 mm), or increase the infiltration. This could lead to higher vegetation stress. The soil water level with an adapted tuning parameter for water potentials is higher than that of the original model. This is induced by the increased difficulty for trees to take up soil moisture. 4.1.3 Transpiration Another variable to assess the model for its ability to simulate the right water stress is the transpiration, or the sapflow. Figure V.7 shows the evolution of sapflow during 2006, from both observations and the model. While during winter and spring the model simulates the transpiration quite well compared to the observations, during summer the model transpires too much. This is probably due to the lack of summer stress and the constant availability of soil moisture. With the water potential tuning parameter, the transpiration decreases, but is still higher than the observations. This shows that even by tuning the water potential 106 CHAPTER V. DROUGHT IMPACTS ON EUROPEAN FOREST SPECIES Figure V.6: Total soil moisture in Puéchabon, from June to August 2006, in observations (red) and the model (black). The brown symbols correspond to the adapted model parameter for soil water potential (see text for more details). Precipitation is shown in blue. parameter and thereby better simulating the GPP, the model has still a limited ability to simulate summer drought stress. A possible improvement may be to include a more physically based model to simulate the soil and root water potential. 4.2 Other sites The other sites (Brasschaat, Tharandt, Collelongo and Hesse) that are studied are grouped together, as so far there was no time to do a thorough investigation to their ecological water balance and drought periods. The seasonal cycles are on average well simulated. The largest differences between the model and observations occurred in Hesse and Collelongo, both sites with Fagus sylvatica. ORCHIDEE-CAN underestimated the GPP during some summers with over 4 gC m− 2 day− 1. By adjusting a parameter related to the photosynthesis, the GPP was much better simulated for most summers. One exception occurred for a summer with water stress that with the new parameterization did not show stress anymore. This shows the need for a more in depth study to the simulation of GPP. 4. RESULTS 107 Figure V.7: GPP (upper panel) and transpiration (lower panel) in Puéchabon in 2006, in observations (red) and the model (black). The green and brown symbols correspond to the adapted model parameters for soil water potential (see main text for more details). 4.2.1 Brasschaat The first site that will be discussed is Brasschaat, a FLUXNET site in Northern Belgium, close to Antwerp. The forest is composed of a mixture with mostly Pinus sylvestris (∼40%) and Quercus robur (∼25%). Furthermore there are some other needleleaf species (∼10%), deciduous species (∼5%), C3 grasses (∼15%) and some bare soil (Carrara et al., 2003, 2004). The tree height of Scots pine is ∼21 m, while pedunculate oak reaches about 17 m (Xiao et al., 2003; Carrara et al., 2003). LAI is in between 2 and 3, depending on the species mixture and the season (Xiao et al., 2003; Carrara et al., 2004). Scots pine Figure V.8 shows the GPP of Pinus sylvestris from 2000-2002 in Brasschaat. It is simulated too high in all seasons except for the spring, where the observations and the model correspond relatively well. The overestimation in winter may be due to the LAI that remains too high, although this should not be the main problem in the summer. The soil texture is sandy till 1.5 m depth. Below this layer, till 2 m, is a clay layer with poor drainage (Carrara et al., 2003). While the soil depth in ORCHIDEE-CAN is 2 m, the soil texture cannot 108 CHAPTER V. DROUGHT IMPACTS ON EUROPEAN FOREST SPECIES Pinus sylvestris ● ORCHIDEE−CAN FLUXNET ● ● 8 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 4 GPP (gC m−2 day−1) 6 ● ● ● ● ● ● ● ● 2 ● ● ● ● ● ● ● 0 ● ● 2000 2001 2002 Time (years) Figure V.8: Gross primary productivity of Pinus sylvestris in Brasschaat, from 2000 to 2002, observed (red) and simulated (black). be changed with increasing depth, and sand has been used as texture class throughout the whole vertical soil column. Although this may favor the model relative to the observations because the water may be more easily available. Lastly, the species combination may cause a difference between the model and the observations. Brasschaat is a mixed forest, consisting out of needleleaf and broadleaf species. Although in the simulation also different vegetation fraction are applied, in this example I only show the results of one species which represents 40% of the vegetation cover, that may not well reflect the reality. Oak Summer GPP of oak at Brasschaat is, even as spruce, too high (Fig. V.9). This may be caused by a modeled overestimation of LAI. The winter GPP is slightly too low. Pedunculate oak is a summergreen species, and loses all leaves during winter resulting in a LAI of zero. The observations however, do not differentiate between the different tree species, leading to a slightly higher winter GPP. Extreme drought stress was not observed. A better comparison between the model and observations may be made by comparing a mono-culture of oak, instead of this mixed forest, that only has a coverage of 25& for oak. In Brasschaat the seasonal cycle of the GPP was in general quite well simulated both for Scotch pine and oak. We found some inconsistencies with the observations during summer where the modeled GPP was somewhat too high (∼1-2 gC m− 2 day− 1) for several years. During winter, simulated GPP was slightly too high for Pinus sylvestris while it seemed to 4. RESULTS 109 Quercus robur ORCHIDEE−CAN FLUXNET ● ● ● 8 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 4 GPP (gC m−2 day−1) 6 ● ● ● ● ● ● ● ● 2 ● ● ● ● ● ● ● 0 ● ● 2000 2001 2002 Time (years) Figure V.9: Gross primary productivity of Quercus robur in Brasschaat, from 2000 to 2002, observed (red) and simulated (black). be too low for Quercus robur. The latter is due to the species mixture at the site, with an evergreen needleleaf covering ∼40% and a summergreen broadleaf species covering ∼25%. A solution could be to test different percentages of both species at the site to study the sensitivity of the model for species mixtures. 4.2.2 Tharandt Tharandt is a forest site in central-East Germany. The main tree species are 120-year old evergreen needleleaf Picea abies, which grows together with Pinus sylvestris (∼10%), deciduous summer green trees (∼10%), C3 grasses (∼10%), leaving ∼5% bare soil. The forest floor is mainly covered by young Fagus sylvatica and Deschampsia flexuosa (Grünwald and Bernhofer, 2007). The average tree height is about 27 m with an LAI of ∼8.5 (Grünwald and Bernhofer, 2007; Granier et al., 2007). Tharandt has a shallow soil of approximately 1 m, and the main rooting depth is ∼35 cm (Grünwald and Bernhofer, 2007). Simulated GPP is too low in summer for all modeled years (1996-2006), while in winter it is too high (Fig. V.10). Although the species is sensitive to drought, extreme summer stress as in Quercus ilex was not observed. This may be due to the different climates between both sites. The difference in winter GPP may be due to the species mixture which contains both evergreen and summer green trees. The LAI during winter may stay too high in the model, CHAPTER V. DROUGHT IMPACTS ON EUROPEAN FOREST SPECIES Picea abies 12 110 6 0 2 4 GPP (gC m−2 day−1) 8 10 ORCHIDEE−CAN FLUXNET 2002 2003 2004 Time (years) Figure V.10: Gross primary productivity of Picea abies in Tharandt, from 2002 to 2004, observed (red) and simulated (black). possibly resulting in more carbon assimilation. The average modeled LAI however, is slightly too low, possibly explaining the summer GPP deficit. 4.2.3 Hesse & Collelongo Beech, Fagus sylvatica, is a very important broad-leaved species in Western and central Europe (García-Plazaola and Becerril, 2000; Peuke et al., 2002) and is present at two of the studied FLUXNET sites, Collelongo and Hesse. Hesse, located in the north-east of France, is a deciduous broadleaf forest. Approximately 90% of the vegetation cover exists out of beech. The remaining 10% is a mixture of Carpinus betulus, Betula pendula, Quercus petraea, Larix decidua, Prunus avium and Fraxinus excelsior (Granier et al., 2000b). The stand age is about 45 years and the average tree height was 13 m in 2000. The understory vegetation is very sparse (Granier et al., 2000a). After thinning in 1999, the LAI was ∼4.8 m2 m−2 (Granier et al., 2002). Collelongo is beech forest site in central Italy, approximately 80 km east from Rome. About 90% of the ground is covered with beech, while the remaining part is bare soil. Gallium makes the understory vegetation. The average tree height was 25 m 2003, with an LAI of 4 (Valentini, 2003). The stand age is ∼110 years (Valentini et al., 1996). The soil at both sites is not very deep, with 0.8 m in Collelongo and 1 m in Hesse (Valentini, 2003). The modeled GPP of Fagus sylvatica is too low during summer at both test sites (Fig. V.11). 5. DISCUSSION 111 Collelongo 14 14 Hesse ● ORCHIDEE−CAN ORCHIDEE−CAN tuned FLUXNET ● ● ● ● ● ● ● 12 12 ● ● ORCHIDEE−CAN ORCHIDEE−CAN tuned FLUXNET ● ● ● ● ● ● ● 10 ● ● ● ● ● ● ● ● ● ● ● ● ● 8 GPP (gC m−2 day−1) ● 8 ● ● ● ● ● ● ● 6 10 ● ● ● ● 6 GPP (gC m−2 day−1) ● ● ● ● ● ● ● ● ● 4 4 ● ● ● ● ● ● ● ● ● ● 2 2 ● ● ● ● ● ● ● ● 0 0 ● ● 2003 2004 Time (years) 2005 ● 1999 2000 2001 Time (years) Figure V.11: Gross primary productivity of Fagus sylvatica in Hesse from 2003 to 2005 and in Collelongo from 1999 to 2001. Comparison between observations (red) and ORCHIDEE-CAN (black). The solid black curve represents the standard model, the dotted line the model with an adapted parameter for enhanced photosynthesis. In Hesse, the observed tree height was 13 m in 2000. The modeled tree height is 21 m. The modeled LAI may have been slightly too high as well. This would indicate that the GPP would be higher than observed, however, this is not the case. In Collelongo, both the tree height and the LAI in the model are very close to the measurements. This suggests that the incorrect GPP has another reason. By tuning a parameter for the efficiency of photosynthesis (LLALPHA from 0.3 to 0.5) the GPP is much better simulated, implying that this may be the reason for the too low simulated GPP. However, by increasing photosynthetic efficiency, both LAI and height increase as well, inducing errors in these simulated variables. Furthermore, drought stress seems not to be simulated by increasing the photosynthetic efficiency (Fig. V.11 - Hesse year 2004). This shows that underlying physical mechanisms of drought stress need to be better understood, before we are able to correctly model them. 5 Discussion In this chapter we analyzed the impact of drought stress on five European tree species in literature, observational data and in a vegetation model. Overall the model simulated the seasonal cycle of the growth, represented by the GPP, quite well. Drought stress however, mostly experienced by Quercus ilex in Puéchabon, was not so well simulated. The production of the vegetation remained too high during summer, and the soil water level did not decrease enough as compared to observations. In a future, possibly warmer climate, it is especially important to simulate drought stress, as more severe and frequent heatwaves and droughts are expected to come that may cause permanent damage to the vegetation. In this section we review the model results of Quercus ilex in more detail to better understand the lack of stress response in the simulations. Furthermore we will discuss possible model adjustments with 112 CHAPTER V. DROUGHT IMPACTS ON EUROPEAN FOREST SPECIES some of the missing processes as described in the literature (section V.2). 5.1 Site-specific characteristics Some characteristics of Puéchabon make the site difficult to model and should be discussed first. The rock fraction in the soil is very high. Ingelmo et al. (1994) already remarked that rocky soil are hard to model due to their heterogeneity. In ORCHIDEE-CAN, the coarsest stone or sand fraction is smaller than 2 mm. This has consequences for the availability of water for the tree, the run-off and the root distribution. The modeled root distribution is represented by an exponential decrease that is grown in a homogeneous soil. In rocky soils however, deep roots can only grow in cracks that are not represented by the model. The run-off and infiltration rate are also strongly affected by the presence of rocks (Poesen and Lavee, 1994), decreasing the total soil water availability (Saxton and Rawls, 2006). This can lead to overestimation of the available soil moisture (Cousin et al., 2003). To overcome these issues, it would be necessary to decrease the soil depth or to include rocky soil fraction in models, and anisotrope soils, with preferential flow. Several of such models exist, although the more complex models may still have limited use due to the scarcity of observations for correct parameterizations (Simunek et al., 2003). A more simplistic option may be to describe the soil in only two layers, in where the deepest layer is only available for a small fraction of the roots. This method seems to work relatively well for Puéchabon (S. Rambal, personal communication). It is important to keep in mind that ORCHIDEE-CAN is designed as a global vegetation model. Therefore it cannot be expected that all site-specific characteristics are well represented and simulated. A better infiltration parameterization however, could be useful in many vegetation models. 5.2 Plant hydraulics Besides the rocky soil, a process that is missing in the current version of ORCHIDEE-CAN is a soil-root resistance. This is an additional resistance that can act very strongly in the case of severe droughts (H. Cochard, personal communication). When the soil dries out, it will be more difficult for the root to subtract water. Furthermore, the root may shrink, causing a gap between the root and the soil (Carminati et al., 2013). Bonan et al. (2014) developed a vegetation model with a soil-to-root resistance, which adds an additional stress factor during drought. Preliminary results indicate that the addition of such a resistance reduces the GPP during the vegetative period, but does not explain all of the missing summer stress. This may be caused by the soil water level that is not correctly simulated. In the model, the soil moisture level during summer is too high and an additional resistance would only increase the water level further, as the trees would have more difficulty to uptake water. This may be overcome by adapting the whole tree hydraulic architecture, by adding for example capacitance, as in Bonan et al. (2014), that is currently missing in the model. This is ongoing work by Emilie Joetzjer (personal communication). 5.3 Other processes Aside from the soil, different specific mechanisms exist that can help trees to survive periods of drought. These will be discussed in the same order as presented in the literature 5. DISCUSSION 113 review in section V.2, starting with photosynthesis, BVOCs, and plasticity, lagged effects and acclimation. Photosynthesis When drought stress increases, stomatal conductance and photosynthesis decrease. However these processes do not only depend on the climatic factors but can also be determined by factors related to the tree. Limousin et al. (2010b) found for example that leaf age influences the recovery rate of photosynthesis. Such mechanisms are often not implemented in large scale vegetation models that are based on PFTs instead of species. In ORCHIDEE-CAN the implementation may be easier, as this model defines different tree species and the Vcmax (maximum rate of carboxylation) is related to leaf age. Another process influencing photosynthesis rate is the mesophyll conductance that is, opposite from stomatal conductance, not always present in models (Naudts et al., 2016). This may be a limiting factor in explaining gas exchange during drought stress (Keenan et al., 2010). The WUE can also change during drought stress, and usually increases during and sometimes after periods of stress. A changing WUE however, is not necessarily present in vegetation models (Traore et al., 2014). Besides these relatively well known processes, some mechanisms are still unknown or only partly known. For example those of non-photochemical quenching (Farquhar et al., 2001). Furthermore, different observational studies do not always agree on their results. Peña-Rojas et al. (2004); Limousin et al. (2010b); Vaz et al. (2010); Misson et al. (2010), for example, find a decrease in both the rate of electron transport and the maximum carboxylation rate due to drought stress, but Gallé et al. (2011) did not find such a change. These differences in results may be related to the severity of the water stress or the age of the studied plants (Peña-Rojas et al., 2004), but complicate the modeling effort on larger scales. Lastly, during drought stress, leaf damage can be heterogeneous (Sardans et al., 2010) and the closure of stomata may be un-even (Terashima et al., 1988). This could also lead to an incorrect calculation of the total gas exchange. Another process related to photosynthesis that is missing in large-scale vegetation models is the change in chlorophyll concentrations and antioxidant substances. Although chlorophyll concentration and antioxidants under drought stress is debated (section V.2), this is not at all accounted for in models. These possible missing protection mechanisms for photosynthesis may lead to an underestimation of GPP in models during drought stress. BVOC During drought or heat stress many species emit BVOC, such as terpenes or isoprene. Quercus ilex is a terpene emitting species. The precise mechanisms of the emissions are not known, so they are often not represented in models or only in a very simplistic way. As BVOC are carbon compounds, their emission may alter the carbon balance of the tree. During drought stress, when carbon uptake is limited, this may be an important mechanisms through which the tree loses carbon. The impact of BVOC on the carbon balance of Quercus ilex is probably minimal. The substances cannot be stored in the tree thus newly synthesized carbon molecules are used for BVOC. As photosynthesis is very low during drought stress, BVOC emissions become close to zero. For other species this process may be more important, and could possibly play a role on the carbon balance of the tree on longer time scales. Acclimation & lagged effects Process that are mostly not accounted for in global vegetation models are plasticity, acclimation and lagged effects. During drought stress the composition of proteins may be changed. These proteins can for example be related to the pho- 114 CHAPTER V. DROUGHT IMPACTS ON EUROPEAN FOREST SPECIES tosynthesis apparatus or stress regulation. Although such detailed genetic knowledge is very useful for understanding and predicting impacts, it is too detailed for the use in large-scale vegetation models. Furthermore, trees can adapt to drought stress by allocating carbon to different plant parts such as the roots, or by defoliation. This is also currently lacking in most models. Regarding the future climate, maybe most important processes would be permanent damage, mortality, acclimation and management. As more frequent and more severe heatwaves are expected in the future climate (Schär et al., 2004), such effects may have detrimental impacts that could play a significant role in the global carbon cycle. For example, severe drought stress was found to lead to permanent damage (Peñuelas et al., 2000; Galiano et al., 2012), long-term loss of reserves (Galiano et al., 2012), or mortality (Filella et al., 1998; Peñuelas et al., 2000; Valladares et al., 2005; Galiano et al., 2012) up to several years after the drought event took place. Although embolism is sometimes considered in vegetation models, it may not always lead to permanent damage and carry-over effects. Acclimation is also not accounted for while trees originating from different climates may respond differently to drought and heat stress. Plants from warm and dry regions were found to be more drought resistant than individuals from colder and wetter climates (Bussotti et al., 1995b; Pesoli et al., 2003; Villar-Salvador et al., 2004; Sánchez-Vilas and Retuerto, 2007; Bonito et al., 2011; Corcobado et al., 2014). For example, osmotic adjustment (Villar-Salvador et al., 2004), acorn size (Bonito et al., 2011), and photosynthetic capacity (Sánchez-Vilas and Retuerto, 2007) were enhanced. These results suggest acclimation, which may be important in a future warmer and drier climate (Misson et al., 2010). Although Limousin et al. (2010b); Rodríguez-Calcerrada et al. (2011) did not find long-term acclimation during several years of study. Lastly, management options may be included. This has already be done in ORCHIDEECAN (Naudts et al., 2015), but is not present in all vegetation models. Intensive thinning for example, was found to reduce competition and improve the resistance of holm oak forests to severe water stress (Galiano et al., 2012). It is important to keep in mind that some specific events, such as extreme heatwaves for climate models, are not easy to simulate. Furthermore, some of the uncertainties in vegetation models may come from larger scale processes, such as the incorrect simulation of the evapotranspiration, the absence of nutrient cycles, or other inaccuracies in the water or carbon cycle. It may be more relevant to focus on these processes than on the small scale processes that were mentioned previously. A next step could be an extension with mortality. This will probably be one of the most important processes for simulating the response of vegetation to more frequent and more severe heatwaves in a future climate. 6 Summary and concluding remarks In this chapter I reviewed the consequences of drought and heat stress on five main forest tree species in Europe. Furthermore, some preliminary results of simulating summer stress at a local scale are presented. The main goal was to analyze the ability of the global vegetation model ORCHIDEE-CAN to correctly reproduce drought stress in order to use the model to predict future, possibly more severe impacts. At Puéchabon, the site with most recurrent drought stress, ORCHIDEE-CAN was not able to correctly simulate the affected GPP of Quercus ilex. The production of the vegetation remained too high during summer, probably caused by 6. SUMMARY AND CONCLUDING REMARKS 115 too high soil moisture levels. For Fagus sylvatica, Picea sp., Pinus sylvestris and Quercus robur and Q. petraea the simulation of GPP was much better represented, although these species did not experience much drought stress. Some processes on a tree- and leaf-scale are not represented in the vegetation model, which may cause the differences between the modeled and observed primary production. To simulate the impact of drought stress on vegetation of the projected increase in frequency and severity of heatwaves, some adaptations should be made to the model. Firstly, specific processes that act most strongly during drought conditions could be added. An example is soil-root resistance. Besides such drought-specific processes, it would also be necessary to introduce lagged effects. Hydraulic failure is hardly possible to simulate at the moment, as embolism is restored completely during autumn and winter. In reality, tree mortality can occur even several years after an extreme drought event. A relatively simple solution would be to add a form of mortality when embolism becomes too extensive. Lastly, it may be interesting to add insects or pathogens to the model. Drought stress can increase the susceptibility of trees to insect attacks. Furthermore, increasing temperatures may favor the population density of harmful insects. However, this may be much more complicated and may remain an issue for a future study. C HAPTER VI Discussion & concluding remarks Through this work I aimed to improve the understanding of the role of land-atmosphere feedbacks and large-scale circulation that lead to warm summer temperatures in Europe. This is challenging due to the scarcity of observations and the uncertainties of parameterized atmospheric processes. I focused on four main issues: 1) How do land-atmosphere feedbacks affect climate projections and their uncertainties? 2) How do different physical parameterizations affect the simulation of extreme heatwaves? 3) How large are the roles of soil moisture and atmospheric circulation in the development of European summer temperature anomalies? And 4) What are the impacts of heat and drought stress on vegetation? Regarding the first question I found that the different partitioning of land heat fluxes between models leads to spatially different warming over Europe in the future. The uncertainty of future climate change was especially high in central Europe, largely due to the uncertainty in heat flux partitioning, while in Southern Europe the models mostly agreed. The use of observation-based sensible heat fluxes allowed to reduce this climate change uncertainty regionally up to 40%. While studying different atmospheric parameterizations for the extreme heatwaves of 2003 and 2010, I found a large temperature spread between the simulations. Compared to observations, temperature was mostly underestimated. Shortwave radiation and precipitation were generally overestimated. I selected a reduced model ensemble of well performing configurations compared to observations, to perform future studies on warm summer temperatures over Europe. The best physics configuration was consequently used to quantify the role of early summer soil moisture and large-scale drivers on summer temperature anomalies. The contribution of soil moisture was up to maximum 1◦ C during the heatwaves of 2003 and 2010. The contribution of large-scale drivers was larger, reaching up to 3◦ C in 2003 and up to 6◦ C in 2010. However, the contribution of early summer soil moisture to the temperature anomalies has been increasing over the last decades over parts of central Europe and Russia, corresponding to the regions with a significant negative trend of soil moisture. Large-scale drivers showed an increasing importance in the Eastern European region. Lastly, I studied the impacts of drought and heat stress on several European forest tree species. I found an overestimation of modeled GPP at a local scale in the Mediterranean region during summer with ORCHIDEE. This indicates that the vegetation model does not 117 118 CHAPTER VI. DISCUSSION & CONCLUDING REMARKS well reproduce the complicated consequences of drought stress. To model future, possibly more severe impacts of drought, the model may need to be adapted with drought-specific processes and lagged effects. In the following section I discuss three topics concerning the presented work in this manuscript. The first regards climate models in general. The second part discusses the future climate. To conclude I will discuss the modeling of vegetation impacts. 1 Climate models in general The use of climate models is becoming increasingly important. With the growing understanding of the functioning of the climate, new processes are added into the current models which are becoming more sophisticated. This has both advantages and disadvantages. It often allows for better physical mechanisms and can lead to higher spatial and temporal accuracy, but at the same time it can more easily lead to error compensation and to too complex models for a full comprehension. The expanding number of parameters must be tuned while observations are not always abundant. What is then the meaning of using a climate model? How is it then possible to obtain a correct understanding of the model outputs? Chapter 3 serves as a suitable example. Starting to model the heatwave of 2003, I required the model to give me the correct temperatures. At that point a first question emerged: which physics to choose? Using WRF I had many options for the physics, but they did not necessarily mean a lot to me. There I discovered that modeling is not so simple, at least, modeling and trying to understand what I was doing. I have attempted during my work to better understand why different configurations of the one model I used provided me with such different climates. I think I partly succeeded, although some questions remain unanswered. I did not have the intention to completely understand the exact differences between all physics options, and even if I had, I am not sure I would have succeeded during my PhD. I also do not think it is necessary to understand all processes for using a model. I want to argue here that in using models, it is important to at least consider the different physics and parameterizations of models. To try to understand why a certain model produces higher temperatures and less rain than another one. How otherwise, would it be possible to deduce significant information on our climate? In this work we obtained a better understanding of some of the physics of the WRF model and created a small well-performing multi-physics ensemble that was used to study the summer climate in Europe. Furthermore, this ensemble could be used for studying impacts of heat waves on vegetation for example. Another interesting study that could be done concerns the atmospheric circulation patterns in the different physics configurations from chapter 3. As shown in chapter 4, the largescale drivers increased their influence on summer temperatures across almost all of Eastern Europe from 1980-2011. It may be interesting if this was due to a change in atmospheric circulation, for example, an increase in Blocking situations. This has recently been studied by Horton et al. (2015); Alvarez-Castro et al. (2016), who found a change in atmospheric patterns looking at reanalyses data. Although the simulations are nudged above the boundary layer, differences in strength of certain atmospheric features, due to the diversity of the physics, may be discovered. Subsequently, the method could be applied to future simulations to better understand possible climates that we may encounter. 2. REGARDING THE FUTURE CLIMATE 2 119 Regarding the future climate Modeling the future climate may require different kind of questions than using a climate model for representing and understanding the current climate. Although the issue of including a good representation of the mechanisms is still important, another question we should ask ourselves is to what extend we put our trust in the models. Models that well simulate the current climate, do not necessarily accurately model another climate. For example our ’best’ physics configuration, is not able to correctly simulate the monsoon rain in Africa (C. Klein, personal communication). Also the past climate may not be well represented because other processes may have been more important than in the present climate. So why put trust in simulations about the future climate? A climate that is very uncertain if we consider the range of temperature possibilities at the end of the century (IPCC, 2013). A climate for which we have no measurements and observations. A climate with a possible sea level rise of several meters (Hansen et al., 2016). Is it then legitimate to say we have reduced the uncertainty of predictions (Chapter 2) by using a small set of regional climate models? Does this mean we cannot say anything about the future climate? That all the predictions are just random guesses? No, probably not. Our current knowledge on the climate is extensive, and with the information on different periods of past climate, the predictions are much more than a guess, and built on physically-based mechanisms. However, it is important to keep in mind that current processes may change and therefore generate unpredictability in the future climate that may not be foreseen. Another issue that could be important in the simulation of the future climate is the resolution of climate models. Although many large-scale processes can be reasonably well simulated, we do not yet perfectly understand how small-scale processes, that are not represented in these models, influence the processes at large-scale. These small-scale processes may arise from future land-use changes for example. At the moment, all climate models operate on too large scales for the investigation of such issues. The reduced ensemble of well performing physics configurations that was created in Chapter 3 could be used for studying the future climate. Although a model that is well performing in the present climate does not necessarily correctly simulate the future, our configurations can at least handle high summer temperatures. This may be essential for simulating the end of the current century. Preliminary results of 150-year climate simulations, from 1951-2100, show differences between the three ensemble members (Fig. VI.1). The numbers of the physics correspond to the model ranking from Stegehuis et al. (2015). In total I performed 4 simulations. Two with the best performing physics, one with the second best, and one with the third (chapter 3.3). The difference between the simulations with the best performing physics is that one is nudged above the boundary layer, while the second one is not nudged. All four future simulations are forced by the climate of the LMDZ model with RCP8.5. There is an almost constant temperature difference between the three physics schemes, with a range of almost 1◦ C over Europe during the entire simulation. The simulation that is not nudged shows somewhat lower temperatures than the one that is nudged, but the differences seem small. Interestingly, the different physics do not diverge much even at the end of the century. This may suggest that the boundary conditions play a more important role than the different physics that were used. However, the simulations need to be analyzed in more detail. A focus could be on the summer seasons, to investigate whether there will be more frequent and more severe heatwaves as expected. The bias of the LMDZ model should also be investigated and taken into account. A comparison between simulations driven by the 120 CHAPTER VI. DISCUSSION & CONCLUDING REMARKS Figure VI.1: Mean annual temperature over Europe from 1950 to 2099. Different colors represent different model configurations. The numbers of the physics correspond to the model ranking in Stegehuis et al. (2015). All simulations were nudged above the boundary layer except from one (black line). 3. IMPACT STUDIES 121 LMDZ model and driven by reanalyses data, from approximately 1980 to 2012, could show some of the main biases of LMDZ. Although, of course, some bias may also arise from the interaction between the forcing and the forced climate model. 3 Impact studies Concerning vegetation impact studies both the uncertainty of the climate model and that of the vegetation model should be taken into account. As discussed in chapter 5, vegetation models are far from perfect, and many processes at both a tree- and leaf-level are not included. However, the models are repeatedly calibrated and evaluated against observations and they are essential for impact studies on a large spatial and temporal resolution. To increase confidence in simulations a similar approach as has been done for WRF (chapter 3) could be meaningful: using different models of biochemical and physical processes to simulate various events and analyze the output. This has been done for some processes such as photosynthesis (Misson et al., 2004), but not yet in a systematic way. This may be challenging, as a model platform such as WRF does not exist for vegetation models. For assessing the impact of climate events on vegetation different methods have been proposed. van Oijen et al. (2013a) propose a probabilistic method to estimate the expected loss of ecosystem variables such as net primary productivity. They define a risk as expected loss, a hazard as a potentially damaging phenomenon and the probability of the hazard actually occurring, and the vulnerability as the degree of loss that results if the phenomenon occurs (van Oijen et al., 2013a). Through this approach it is possible to assess how and when extreme climate events such as heatwaves jeopardize ecosystems on a large scale. Furthermore the method allows for the detection of the underlying causes that can either be a change in the frequency of the climate event, or a change in the vulnerability of the ecosystem (van Oijen et al., 2013a). Another interesting reversed approach is proposed by (Zscheischler et al., 2013), in where first the impact is detected, followed by the attribution to a climate event. This approach allows for a deeper understanding of different causes of loss of ecosystem productivity. Because of the increasing temperature and the higher chance of heatwaves, it may be interesting to use such methods for future climate projections. A multi-physics ensemble as proposed in this work may serve as an uncertainty measure on the vegetation stress. The different simulations (Fig. VI.1) all lead to warmer temperatures, but there may be differences in for example the partitioning of the land heat fluxes, leading to different vegetation responses. However, in this case it may be necessary to include lagged effects in the vegetation model, as carbon starvation or hydraulic failure caused by extreme climate events may lead to mortality on a longer time scale. Bibliography Abrams, M. D.: Adaptation and responses to drought in Quercus species of North America., Tree Physiology, 7, 227–238, 1990. Aleina, F. C., Baudena, M., D’Andrea, F., and Provenzale, A.: Multiple equilibria on planet Dune: climate vegetation dynamics on a sandy planet., Tellus, Series B: Chemical and Physical Meteorology and Oceanography, 65, 1–12, doi:10.3402/tellusb.v65i0.17662, 2013. Allard, V., Ourcival, J. M., Rambal, S., Joffre, R., and Rocheteau, A.: Seasonal and annual variation of carbon exchange in an evergreen Mediterranean forest in southern France., Global Change Biology, 14, 714–725, doi:10.1111/j.1365-2486.2008.01539.x, 2008. Allen, C. D., Macalady, A. K., Chenchouni, H., Bachelet, D., McDowell, N., Vennetier, M., Kitzberger, T., Rigling, A., Breshears, D. D., Hogg, E. H. T., Gonzalez, P., Fensham, R., Zhang, Z., Castro, J., Demidova, N., Lim, J. H., Allard, G., Running, S. W., Semerci, A., and Cobb, N.: A global overview of drought and heat-induced tree mortality reveals emerging climate change risks for forests., Forest Ecology and Management, 259, 660–684, doi:10.1016/j.foreco.2009.09.001, 2010. Alvarez-Castro, C., Faranda, D., and Yiou, P.: Changes in the atmospheric dynamics leading to West-European heatwaves since 1900., Submitted, 2016. Aranda, I., Rodriguez-Calcerrada, J., Robson, T. M., Cano, F. J., Alte, L., and Sanchez-Gomez, D.: Stomatal and non-stomatal limitations on leaf carbon assimilation in beech (Fagus sylvatica L .) seedlings under natural conditions., Forest Systems, 21, 405–417, 2012. Arend, M., Kuster, T., Günthardt-Goerg, M. S., and Dobbertin, M.: Provenance-specific growth responses to drought and air warming in three European oak species (Quercus robur, Q. petraea and Q. pubescens)., Tree Physiology, 31, 287–297, doi:10.1093/treephys/tpr004, 2011. Arndt, S. K., Clifford, S. C., Wanek, W., Jones, H. G., and Popp, M.: Physiological and morphological adaptations of the fruit tree Ziziphus rotundifolia in response to progressive drought stress., Tree physiology, 21, 705–715, doi:10.1093/treephys/21.11.705, 2001. Awan, N. K., Truhetz, H., and Gobiet, A.: Parameterization-induced error characteristics of MM5 and WRF operated in climate mode over the alpine region: An ensemble-based analysis., Journal of Climate, 24, 3107–3123, doi:10.1175/2011JCLI3674.1, 2011. Baldocchi, D., Valentini, R., Running, S., Oechel, W., and Dahlman, R.: Strategies for measuring and modelling carbon dioxide and water vapour fluxes over terrestrial ecosystems., Global Change Biology, 2, 159–168, doi: 10.1111/j.1365-2486.1996.tb00069.x, 1996. 123 124 BIBLIOGRAPHY Bansal, S., Hallsby, G., Löfvenius, M. O., and Nilsson, M. C.: Synergistic, additive and antagonistic impacts of drought and herbivory on Pinus sylvestris: leaf, tissue and whole-plant responses and recovery., Tree Physiology, 33, 451–463, doi:10.1093/treephys/tpt019, 2013. Bansal, S., Harrington, C. a., Gould, P. J., and St.Clair, J. B.: Climate-related genetic variation in droughtresistance of Douglas-fir (Pseudotsuga menziesii)., Global Change Biology, 21, 947–958, doi:10.1111/gcb. 12719, 2015. Baquedano, F. J. and Castillo, F. J.: Comparative ecophysiological effects of drought on seedlings of the Mediterranean water-saver Pinus halepensis and water-spenders Quercus coccifera and Quercus ilex., Trees, 20, 689– 700, doi:10.1007/s00468-006-0084-0, 2006. Barriopedro, D., Fischer, E. M., Luterbacher, J., Trigo, R. M., and García-Herrera, R.: The hot summer of 2010: redrawing the temperature record map of Europe., Science, 332, 220–224, doi:10.1126/science.1201224, 2011. Barsi, D. C., Major, J. E., Mosseler, A., and Campbell, M.: Genetic variation and control of chloroplast pigment concentrations and related needle-level traits in Picea rubens, Picea mariana, and their hybrids: Moisture and light environmental effects., Trees, 23, 555–571, doi:10.1007/s00468-008-0301-0, 2009. Barthel, M., Hammerle, A., Sturm, P., Baur, T., Gentsch, L., and Knohl, A.: The diel imprint of leaf metabolism on the d13C signal of soil respiration under control and drought conditions., New Phytologist, 192, 925–938, doi:10.1111/j.1469-8137.2011.03848.x, 2011. Baudena, M., D’Andrea, F., and Provenzale, A.: A model for soil-vegetation-atmosphere interactions in waterlimited ecosystems., Water Resources Research, 44, 1–9, doi:10.1029/2008WR007172, 2008. Beck, E. H., Fettig, S., Knake, C., Hartig, K., and Bhattarai, T.: Specific and unspecific responses of plants to cold and drought stress., Journal of Biosciences, 32, 501–510, doi:10.1007/s12038-007-0049-5, 2007. Becker, M. and Becker, M.: Le dépérissement du chêne en forêt de Tronçais les causes écologiques., Annales des Sciences Forestieres, 39, 439–444, 1982. Beier, C., Gundersen, P., Hansen, K., and Rasmussen, L.: Experimental manipulation of water and nutrient input to a Norway spruce plantation at Klosterhede, Denmark. II. Effects on tree growth and nutrition., Plant and Soil, 168-169, 613–622, 1995. Beljaars, A. C. M.: The parametrization of surface fluxes in large-scale models under free convection., Quarterly Journal of the Royal Meteorological Society, 121, 255–270, doi:10.1002/qj.49712152203, 1994. Beljaars, A. C. M., Viterbo, P., Miller, M. J., and Betts, A. K.: The anomalous rainfall over the United States during July 1993: sensitivity to land surface parameterization and soil moisture anomalies., Monthly Weather Review, 124, 362–383, 1996. Bellprat, O., Kotlarski, S., Lüthi, D., and Schär, C.: Objective calibration of regional climate models., Journal of Geophysical Research: Atmospheres, 117, 1–13, doi:10.1029/2012JD018262, 2012. Benjamin, S., Grell, G., Brown, J., Smirnova, T., and Bleck, R.: Mesoscale weather prediction with the RUC hybrid isentropic-terrain-following coordinate model., Monthly Weather Review, 132, 473–494, 2004. Berg, E. E., David Henry, J., Fastie, C. L., De Volder, A. D., and Matsuoka, S. M.: Spruce beetle outbreaks on the Kenai Peninsula, Alaska, and Kluane National Park and Reserve, Yukon Territory: relationship to summer temperatures and regional differences in disturbance regimes., Forest Ecology and Management, 227, 219–232, doi:10.1016/j.foreco.2006.02.038, 2006. Bigler, C., Bräker, O. U., Bugmann, H., Dobbertin, M., and Rigling, A.: Drought as an inciting mortality factor in scots pine stands of the Valais, Switzerland., Ecosystems, 9, 330–343, doi:10.1007/s10021-005-0126-2, 2006. Bigler, C., Gavin, D. G., Gunning, C., and Veblen, T. T.: Drought induces lagged tree mortality in a subalpine forest in the Rocky Mountains., Oikos, 116, 1983–1994, doi:10.1111/j.2007.0030-1299.16034.x, 2007. BIBLIOGRAPHY 125 Bigras, F. J.: Photosynthetic response of white spruce families to drought stress., New Forests, 29, 135–148, doi:10.1007/s11056-005-0245-9, 2005. Black, E., Blackburn, M., Harrison, R. G., Hoskins, B. J., and Methven, J.: Factors contributing to the summer 2003 European heatwave., Weather, 59, 217–223, doi:10.1256/wea.74.04, 2004. Bladé, I., Liebmann, B., Fortuny, D., and van Oldenborgh, G. J.: Observed and simulated impacts of the summer NAO in Europe: implications for projected drying in the Mediterranean region., Climate Dynamics, 39, 709– 727, doi:10.1007/s00382-011-1195-x, 2012. Blake, T. J. and Li, J.: Hydraulic adjustment in jack pine and black spruce seedlings under controlled cycles of dehydration and rehydration., Physiologia plantarum, 117, 532–539, doi:10.1034/j.1399-3054.2003.00059.x, 2003. Blanch, J. S., Peñuelas, J., and Llusià, J.: Sensitivity of terpene emissions to drought and fertilization in terpenestoring Pinus halepensis and non-storing Quercus ilex., Physiologia Plantarum, 131, 211–225, doi:10.1111/j. 1399-3054.2007.00944.x, 2007. Blanch, J. S., Peñuelas, J., Sardans, J., and Llusià, J.: Drought, warming and soil fertilization effects on leaf volatile terpene concentrations in Pinus halepensis and Quercus ilex., Acta Physiologiae Plantarum, 31, 207– 218, doi:10.1007/s11738-008-0221-z, 2009. Blanck, K., Lamersdorf, N., Dohrenbusch, A., and Murach, D.: Response of a Norway spruce forest ecosystem to drought/rewetting experiments at Solling, Germany., Water, Air, and Soil Pollution, 85, 1251–1256, doi: 10.1007/BF00477153, 1995. Blödner, C., Skroppa, T., Johnsen, Ø., and Polle, A.: Freezing tolerance in two Norway spruce (Picea abies [L.] Karst.) progenies is physiologically correlated with drought tolerance., Journal of Plant Physiology, 162, 549– 558, doi:10.1016/j.jplph.2004.09.005, 2005. Blödner, C., Majcherczyk, A., Kües, U., and Polle, A.: Early drought-induced changes to the needle proteome of Norway spruce., Tree physiology, 27, 1423–1431, 2007. Boberg, F. and Christensen, J. H.: Overestimation of Mediterranean summer temperature projections due to model deficiencies., Nature Climate Change, 2, 433–436, doi:10.1038/nclimate1454, 2012. Boé, J.: Modulation of soil moisture-precipitation interactions over France by large scale circulation., Climate Dynamics, 40, 875–892, doi:10.1007/s00382-012-1380-6, 2013. Bonan, G. B.: Forests and climate change: forcings, feedbacks, and the climate benefits of forests., Science, 320, 1444–1449, doi:10.1126/science.1155121, 2008. Bonan, G. B., Williams, M., Fisher, R. A., and Oleson, K. W.: Modeling stomatal conductance in the Earth system: linking leaf water-use efficiency and water transport along the soil-plant-atmosphere continuum., Geoscientific Model Development, 7, 2193–2222, doi:10.5194/gmdd-7-2193-2014, 2014. Bonito, A., Varone, L., and Gratani, L.: Relationship between acorn size and seedling morphological and physiological traits of Quercus ilex L. from different climates., Photosynthetica, 49, 75–86, doi:10.1007/ s11099-011-0014-2, 2011. Borovics, A. and Mátyás, C.: Decline of genetic diversity of sessile oak at the retracting (xeric) limits., Annals of Forest Science, 70, 835–844, doi:10.1007/s13595-013-0324-6, 2013. Bouma, T. J., Yanai, R. D., Elkin, A. D., Hartmond, U., Flores-Alva, D. E., and Eissenstat, D. M.: Estimating age-dependent costs and benefits of roots with contrasting life span: comparing apples and oranges., New Phytologist, 150, 685–695, doi:10.1046/j.1469-8137.2001.00128.x, 2001. Bréda, N., Cochard, H., Dreyer, E., and Granier, A.: Field comparison of transpiration, stomatal conductance and vulnerability to cavitation of Quercus petraea and Quercus robur under water stress, Annales Des Sciences Forestieres, 50, 571–582, doi:10.1051/forest:19930606, 1993. 126 BIBLIOGRAPHY Bréda, N., Huc, R., Granier, A., and Dreyer, E.: Temperate forest trees and stands under severe drought: a review of ecophysiological responses, adaptation processes and long-term consequences., Annales of Forest Science, 63, 625–644, doi:10.1051/forest:2006042, 2006. Brunner, I., Graf Pannatier, E., Frey, B., Rigling, A., Landolt, W., Zimmermann, S., and Dobbertin, M.: Morphological and physiological responses of Scots pine fine roots to water supply in a dry climatic region in Switzerland, Tree Physiology, 29, 541–550, doi:10.1093/treephys/tpn046, 2009. Büntgen, U., Frank, D. C., Schmidhalter, M., Neuwirth, B., Seifert, M., and Esper, J.: Growth/climate response shift in a long subalpine spruce chronology., Trees, 20, 99–110, doi:10.1007/s00468-005-0017-3, 2006. Bussotti, F., Bottacci, A., Bartolesi, A., Grossoni, P., and Tani, C.: Morpho-anatomical alterations in leaves collected from beech trees (Fagus sylvatica L.) in conditions of natural water stress., Environmental and Experimental Botany, 35, 201–213, doi:10.1016/0098-8472(94)00040-C, 1995a. Bussotti, F., Cenni, E., Ferretti, M., Cozzi, A., Brogi, L., and Mecci, A.: Forest condition in Tuscany (Central Italy). Field surveys 1987-1991., Forestry, 68, 11–24, doi:10.1093/forestry/68.1.11, 1995b. Butler, D.: Heatwave underlines climate-model failures., Nature, 424, 867, doi:10.1038/424867a, 2003. Camarero, J. J., Olano, J. M., Arroyo Alfaro, S. J., Fernández-Marín, B., Becerril, J. M., and García-Plazaola, J. I.: Photoprotection mechanisms in Quercus ilex under contrasting climatic conditions., Flora, 207, 557–564, doi:10.1016/j.flora.2012.06.003, 2012. Campos, P. S., Quartin, V., Ramalho, J. C., and Nunes, M. A.: Electrolyte leakage and lipid degradation account for cold sensitivity in leaves of Coffea sp. plants., Journal of plant physiology, 160, 283–292, doi:10.1078/ 0176-1617-00833, 2003. Carminati, A., Vetterlein, D., Koebernick, N., Blaser, S., Weller, U., and Vogel, H. J.: Do roots mind the gap?, Plant and Soil, 367, 651–661, doi:10.1007/s11104-012-1496-9, 2013. Carrara, A., Kowalski, A. S., Neirynck, J., Janssens, I. A., Yuste, J. C., and Ceulemans, R.: Net ecosystem CO2 exchange of mixed forest in Belgium over 5 years, Agricultural and Forest Meteorology, 119, 209–227, doi: 10.1016/S0168-1923(03)00120-5, 2003. Carrara, A., Janssens, I. A., Yuste, J. C., and Ceulemans, R.: Seasonal changes in photosynthesis, respiration and NEE of a mixed temperate forest., Agricultural and Forest Meteorology, 126, 15–31, doi:10.1016/j.agrformet. 2004.05.002, 2004. Cassou, C., Terray, L., and Phillips, A. S.: Tropical Atlantic influence on European heat waves., Journal of Climate, 18, 2805–2811, doi:10.1175/JCLI3506.1, 2005. Čater, M. and Batič, F.: Groundwater and light conditions as factors in the survival of pedunculate oak (Quercus robur L.) seedlings., European Journal of Forest Research, 125, 419–426, doi:10.1007/s10342-006-0134-6, 2006. Cattiaux, J., Yiou, P., and Vautard, R.: Dynamics of future seasonal temperature trends and extremes in Europe: a multi-model analysis from CMIP3., Climate Dynamics, 38, 1949–1964, doi:10.1007/s00382-011-1211-1, 2012. Čermák, J., Matyssek, R., and Kučera, J.: Rapid response of large, drought-stressed beech trees to irrigation., Tree Physiology, 12, 281–290, 1993. Chaumont, M., Morot-Gaudry, J. F., and Foyer, C. H.: Effects of photoinhibitory treatment on CO2 assimilation, the quantum yield of CO2 assimilation, D-1 protein, ascorbate, glutathione and xanthophyll contents and the electron transport rate in vine leaves., Plant Cell and Environment, 18, 1358–1366, 1995. Chaves, M. M.: Effects of water deficits on carbon assimilation., Journal of Experimental Botany, 42, 1–16, 1991. Chhin, S. and Wang, G. G.: Climatic response of Picea glauca seedlings in a forest-prairie ecotone of western Canada., Annal Forest Science, 65, 1–8, 2008. BIBLIOGRAPHY 127 Choat, B., Jansen, S., Brodribb, T. J., Cochard, H., Delzon, S., Bhaskar, R., Bucci, S. J., Feild, T. S., Gleason, S. M., Hacke, U. G., Jacobsen, A. L., Lens, F., Maherali, H., Martínez-Vilalta, J., Mayr, S., Mencuccini, M., Mitchell, P. J., Nardini, A., Pittermann, J., Pratt, R. B., Sperry, J. S., Westoby, M., Wright, I. J., and Zanne, A. E.: Global convergence in the vulnerability of forests to drought., Nature, 491, 752–755, doi:10.1038/nature11688, 2012. Chou, M.-D. and Suarez, M. J.: A solar radiation parameterization for atmospheric studies., Tech. Rep. June, NASA, 1999. Ciais, P., Reichstein, M., Viovy, N., Granier, A., Ogée, J., Allard, V., Aubinet, M., Buchmann, N., Bernhofer, C., Carrara, A., Chevallier, F., De Noblet, N., Friend, A. D., Friedlingstein, P., Grünwald, T., Heinesch, B., Keronen, P., Knohl, A., Krinner, G., Loustau, D., Manca, G., Matteucci, G., Miglietta, F., Ourcival, J. M., Papale, D., Pilegaard, K., Rambal, S., Seufert, G., Soussana, J. F., Sanz, M. J., Schulze, E. D., Vesala, T., and Valentini, R.: Europe-wide reduction in primary productivity caused by the heat and drought in 2003., Nature, 437, 529–533, doi:10.1038/nature03972, 2005. Cochard, H.: Vulnerability of several conifers to air embolism., Tree physiology, 11, 73–83, doi:10.1093/treephys/ 11.1.73, 1992. Cochard, H., Bréda, N., and Granier, A.: Whole tree hydraulic conductance and water loss regulation in Quercus during drought: evidence for stomatal control of embolism?, Annales des Sciences Forestieres, 53, 197–206, 1996a. Cochard, H., Ridolfi, M., and Dreyer, E.: Responses to water stress in an ABA-unresponsive hybrid poplar (Populus koreana x trichocarpa cv. Peace) II. Hydraulic properties and xylem embolism., New Phytologist, 134, 455–461, doi:10.1111/j.1469-8137.1996.tb04362.x, 1996b. Cochard, H., Lemoine, D., and Dreyer, E.: The effect of acclimation to sunlight on the xylem vulnerability to embolism in Fagus sylvatica L., Plant, Cell and Environment, 22, 101–108, 1999. Cohen, A. E., Cavallo, S. M., Coniglio, M. C., and Brooks, H. E.: A review of planetary boundary layer parameterization schemes and their sensitivity in simulating a Southeast U.S. cold season severe weather environment., Weather and Forecasting, 30, 591–612, doi:10.1175/WAF-D-14-00105.1, 2015. Collet, C. and Guehl, J. M.: Osmotic adjustment in sessile oak seedlings in response to drought., Annales des Sciences Forestieres, 54, 389–394, 1997. Collins, W. D., Rasch, P. J., Boville, B. A., Hack, J. J., McCaa, J. R., Williamson, D. L., Kiehl, J. T., Briegleb, B., Bitz, C., Lin, S.-J., Zhang, M., and Dai, Y.: Description of the NCAR Community Atmosphere Model (CAM 3.0)., Tech. Rep. June, National Center for Atmospheric Research, 2004. Corcobado, T., Cubera, E., Juárez, E., Moreno, G., and Solla, A.: Drought events determine performance of Quercus ilex seedlings and increase their susceptibility to Phytophthora cinnamomi, Agricultural and Forest Meteorology, 192-193, 1–8, doi:10.1016/j.agrformet.2014.02.007, 2014. Corcuera, L., Camarero, J. J., and Gil-Pelegrín, E.: Effects of a severe drought on Quercus ilex radial growth and xylem anatomy., Trees, 18, 83–92, doi:10.1007/s00468-003-0284-9, 2004. Corcuera, L., Cochard, H., Gil-Pelegrín, E., and Notivol, E.: Phenotypic plasticity in mesic populations of Pinus pinaster improves resistance to xylem embolism (P50) under severe drought., Trees, 25, 1033–1042, doi: 10.1007/s00468-011-0578-2, 2011. Costa, P., Bahrman, N., Frigerio, J. M., Kremer, A., and Plomion, C.: Water-deficit-responsive proteins in maritime pine., Plant Molecular Biology, 38, 587–596, doi:10.1023/A:1006006132120, 1998. Coumou, D., Lehmann, J., and Beckmann, J.: The weakening summer circulation in the Northern Hemisphere mid-latitudes., Science, 348, 324–327, doi:10.1126/science.1261768, 2015. Cousin, I., Nicoullaud, B., and Coutadeur, C.: Influence of rock fragments on the water retention and water percolation in a calcareous soil., Catena, 53, 97–114, doi:10.1016/S0341-8162(03)00037-7, 2003. 128 BIBLIOGRAPHY Cregg, B. M. and Zhang, J. W.: Physiology and morphology of Pinus sylvestris seedlings from diverse sources under cyclic drought stress., Forest Ecology and Management, 154, 131–139, doi:10.1016/S0378-1127(00)00626-5, 2001. Croisé, L. and Lieutier, F.: Effects of drought on the induced defense reaction of Scots pine to bark beetleassociated fungi., Annals of Forest Science, 50, 91–97, 1993. Croisé, L., Lieutier, F., Cochard, H., and Dreyer, E.: Effects of drought stress and high density stem inoculations with Leptographium wingfieldii on hydraulic properties of young Scots pine trees., Tree physiology, 21, 427– 436, doi:10.1093/treephys/21.7.427, 2001. Dambrine, E., Carisey, N., Pollier, B., and Granier, A.: Effects of drought on the yellowing status and the dynamics of mineral elements in the xylem sap of declining spruce (Picea abies L.)., Plant and Soil, 150, 303–306, 1993. Darlington, A. B., Halinska, A., Dat, J. F., and Blake, T. J.: Effects of increasing saturation vapour pressure deficit on growth and ABA levels in black spruce and jack pine., Trees, 11, 223–228, doi:10.1007/s004680050079, 1997. Dee, D. P., Uppala, S. M., Simmons, a. J., Berrisford, P., Poli, P., Kobayashi, S., Andrae, U., Balmaseda, M. a., Balsamo, G., Bauer, P., Bechtold, P., Beljaars, A. C. M., van de Berg, L., Bidlot, J., Bormann, N., Delsol, C., Dragani, R., Fuentes, M., Geer, A. J., Haimberger, L., Healy, S. B., Hersbach, H., Hólm, E. V., Isaksen, L., Kållberg, P., Köhler, M., Matricardi, M., Mcnally, A. P., Monge-Sanz, B. M., Morcrette, J. J., Park, B. K., Peubey, C., de Rosnay, P., Tavolato, C., Thépaut, J. N., and Vitart, F.: The ERA-Interim reanalysis: configuration and performance of the data assimilation system., Quarterly Journal of the Royal Meteorological Society, 137, 553–597, doi:10.1002/qj.828, 2011. Demeter, Z., Kanalas, P., Máthé, C., Cseke, K., Szollosi, E., M-Hamvas, M., Jámbrik, K., Kiss, Z., and Mészáros, I.: Osmotic stress responses of individual white oak (Quercus section, Quercus subgenus) genotypes cultured in vitro., Journal of Plant Physiology, 171, 16–24, doi:10.1016/j.jplph.2013.09.013, 2014. Demmig, B. and Björkman, O.: Comparison of the excessive light on chlorophyll fluorescence (77K) and photon yield of O2 evolution in leaves of higher plants, Planta, 171, 171–184, 1987. Detto, M., Montaldo, N., Albertson, J. D., Mancini, M., and Katul, G.: Soil moisture and vegetation controls on evapotranspiration in a heterogeneous Mediterranean ecosystem on Sardinia, Italy., Water Resources Research, 42, 1–16, doi:10.1029/2005WR004693, 2006. Dickson, R. E. and Tomlinson, P. T.: Oak carbon metabolism in response to water stress., Annales des Sciences Forestieres, 53, 181–196, 1996. Dirmeyer, P. A.: The terrestrial segment of soil moisture-climate coupling., Geophysical Research Letters, 38, 1–5, doi:10.1029/2011GL048268, 2011. Dirmeyer, P. A., Koster, R. D., and Guo, Z.: Do global models properly represent the feedback between land and atmosphere?, Journal of Hydrometeorology, 7, 1177–1198, doi:10.1175/JHM532.1, 2006. Ditmarová, L., Kurjak, D., Palmroth, S., Kmet, J., and Střelcová, K.: Physiological responses of Norway spruce (Picea abies) seedlings to drought stress., Tree physiology, 30, 205–213, doi:10.1093/treephys/tpp116, 2010. Dobbertin, M., Wermelinger, B., Bigler, C., Bürgi, M., Carron, M., Forster, B., Gimmi, U., and Rigling, A.: Linking increasing drought stress to Scots pine mortality and bark beetle infestations., The Scientific World JOURNAL, 7, 231–239, doi:10.1100/tsw.2007.58, 2007. Dohrenbusch, A., Jaehne, S., Bredemeier, M., and Lamersdorf, N.: Growth and fructification of a Norway spruce (Picea abies L. Karst) forest ecosystem under changed nutrient and water input., Annal Forest Science, 59, 359–368, doi:10.1051/forest, 2002. Duan, B., Lu, Y., Yin, C., Junttila, O., and Li, C.: Physiological responses to drought and shade in two contrasting Picea asperata populations., Physiologia Plantarum, 124, 476–484, doi:10.1111/j.1399-3054.2005.00535.x, 2005. BIBLIOGRAPHY 129 Duan, B., Yang, Y., Lu, Y., Korpelainen, H., Berninger, F., and Li, C.: Interactions between water deficit, ABA, and provenances in Picea asperata., Journal of Experimental Botany, 58, 3025–3036, doi:10.1093/jxb/erm160, 2007. Duan, B., Ran, F., Zhang, X., Zhang, Y., Korpelainen, H., and Li, C.: Long-term acclimation of mesophyll conductance, carbon isotope discrimination and growth in two contrasting Picea asperata populations exposed to drought and enhanced UV-B radiation for three years., Agricultural and Forest Meteorology, 151, 116–126, doi:10.1016/j.agrformet.2010.09.008, 2011. Eastman, P. A. K. and Camm, E. L.: Regulation of photosynthesis in interior spruce during water stress: changes in gas exchange and chlorophyll fluorescence., Tree Physiology, 15, 229–235, 1995. Echevarría-Zomeño, S., Ariza, D., Jorge, I., Lenz, C., Del Campo, A., Jorrín, J. V., and Navarro, R. M.: Changes in the protein profile of Quercus ilex leaves in response to drought stress and recovery., Journal of Plant Physiology, 166, 233–245, doi:10.1016/j.jplph.2008.05.008, 2009. Eilmann, B., Zweifel, R., Buchmann, N., Graf Pannatier, E., and Rigling, A.: Drought alters timing, quantity, and quality of wood formation in Scots pine., Journal of Experimental Botany, 62, 2763–2771, doi:10.1093/jxb/ erq443, 2011. Eldhuset, T. D., Nagy, N. E., Volařík, D., Børja, I., Gebauer, R., Yakovlev, I. A., and Krokene, P.: Drought affects tracheid structure, dehydrin expression, and above- and belowground growth in 5-year-old Norway spruce., Plant and Soil, 366, 305–320, doi:10.1007/s11104-012-1432-z, 2013. Eltahir, E. A. B.: A soil moisture-rainfall feedback mechanisms 1. Theory and observations., Water Resources Research, 34, 765–776, doi:doi:10.1029/97WR03499, 1998. Eltahir, E. A. B. and Bras, R. L.: Precipitation recycling., Reviews of Geophysics, 34, 367–378, doi:10.1029/ 96RG01927, 1996. Emberson, L. D., Simpson, D., Tuovinen, J. P., Ashmore, M. R., and Cambridge, H. M.: Towards a model of ozone deposition and stomatal uptake over Europe., Tech. rep., Det Norske Meteorologiske Instituit, 2000. Epron, D. and Dreyer, E.: Long-term effects of drought on photosynthesis of adult oak trees [Quercus petraea ( Matt .) Liebl . and Quercus robur L .] in a natural stand., New Phytologist, 125, 381–389, 1993a. Epron, D. and Dreyer, E.: Photosynthesis of oak leaves under water stress: maintenance of high photochemical efficiency of photosystem II and occurrence of non-uniform CO2 assimilation., Tree Physiology, 13, 107–117, 1993b. Epron, D. and Dreyer, E.: Starch and soluble carbohydrates in leaves of water-stressed oak saplings., Annales des Sciences Forestieres, 53, 263–268, doi:10.1051/forest:19960209, 1996. Evans, J. P., Ekström, M., and Ji, F.: Evaluating the performance of a WRF physics ensemble over South-East Australia., Climate Dynamics, 39, 1241–1258, doi:10.1007/s00382-011-1244-5, 2012. Fan, S. and Blake, T. J.: Comparison of polyethylene glycol 3350 induced osmotic stress and soil drying for drought simulation in three woody species., Trees, 11, 342–348, 1997. Farquhar, G. D., von Caemmerer, S., and Berry, J. A.: Models of photosynthesis., Plant physiology, 125, 42–45, doi:10.1104/pp.125.1.42, 2001. Fennessy, M. J. and Kinter III, J. L.: Climatic feedbacks during the 2003 European heat wave., Journal of Climate, 24, 5953–5967, doi:10.1175/2011JCLI3523.1, 2011. Fernández, J., Montávez, J. P., Sáenz, J., González-Rouco, J. F., and Zorita, E.: Sensitivity of the MM5 mesoscale model to physical parameterizations for regional climate studies: annual cycle, Journal of Geophysical Research: Atmospheres, 112, 1–18, doi:10.1029/2005JD006649, 2007. Feudale, L. and Shukla, J.: Role of Mediterranean SST in enhancing the European heat wave of summer 2003., Geophysical Research Letters, 34, 2–5, doi:10.1029/2006GL027991, 2007. 130 BIBLIOGRAPHY Filella, I., Llusià, J., Piñol, J., and Peñuelas, J.: Pistacia lentiscus and Quercus ilex saplings in severe drought and high temperature conditions., Environmental and Experimental Botany, 39, 213 – 220, 1998. Findell, K. L., Gentine, P., Lintner, B. R., and Kerr, C.: Probability of afternoon precipitation in eastern United States and Mexico enhanced by high evaporation., Nature Geoscience, 4, 434–439, doi:10.1038/ngeo1174, 2011. Fischer, E. M., Seneviratne, S. I., Lüthi, D., and Schär, C.: Contribution of land-atmosphere coupling to recent European summer heat waves., Geophysical Research Letters, 34, 1–6, doi:10.1029/2006GL029068, 2007a. Fischer, E. M., Seneviratne, S. I., Vidale, P. L., Lüthi, D., and Schär, C.: Soil moisture-atmosphere interactions during the 2003 European summer heat wave., Journal of Climate, 20, 5081–5099, doi:10.1175/JCLI4288.1, 2007b. Fleck, I., Peña-Rojas, K., and Aranda, X.: Mesophyll conductance to CO2 and leaf morphological characteristics under drought stress during Quercus ilex L. resprouting., Annals of Forest Science, 67, 1–10, doi:10.1051/ forest/2009114, 2010. Ford, T. W. and Quiring, S. M.: In situ soil moisture coupled with extreme temperatures: A study based on the Oklahoma Mesonet, Geophysical Research Letters, 41, 1–8, doi:10.1002/2014GL060949, 2014. Ford, T. W., Quiring, S. M., Frauenfeld, O. W., and Rapp, A. D.: Synoptic conditions related to soil moistureatmosphere interactions and unorganized convection in Oklahoma., Journal of Geophysical Research : Atmospheres, 120, 11 519–11 535, doi:10.1002/2015JD023975, 2015a. Ford, T. W., Rapp, A. D., Quiring, S. M., and Blake, J.: Soil moisture-precipitation coupling: Observations from the Oklahoma Mesonet and underlying physical mechanisms., Hydrology and Earth System Sciences, 19, 3617– 3631, doi:10.5194/hess-19-3617-2015, 2015b. Fort, C., Fauveau, M. L., Muller, F., Label, P., Granier, A., and Dreyer, E.: Stomatal conductance, growth and root signaling in young oak seedlings subjected to partial soil drying., Tree Physiology, 17, 281–289, 1997. Fossdal, C. G., Nagy, N. E., Johnsen, Ø., and Dalen, L. S.: Local and systemic stress responses in Norway spruce: similarities in gene expression between a compatible pathogen interaction and drought stress., Physiological and Molecular Plant Pathology, 70, 161–173, doi:10.1016/j.pmpp.2007.09.002, 2007. Fotelli, M. N., Radoglou, K. M., and Constantinidou, H.-I. A.: Water stress responses of seedlings of four Mediterranean oak species., Tree physiology, 20, 1065–1075, doi:10.1093/treephys/20.16.1065, 2000. Fotelli, M. N., Geßler, A., Peuke, A. D., and Rennenberg, H.: Drought affects the competitive interactions between Fagus sylvatica seedlings and an early successional species, Rubus fruticosus: responses of growth, water status and delta13C composition., New Phytologist, 151, 427–435, doi:10.1046/j.1469-8137.2001.00186.x, 2001. Fotelli, M. N., Rennenberg, H., and Geßler, A.: Effects of drought on the competitive interference of an early successional species (Rubus fruticosus) on Fagus sylvatica L. seedlings: 15N uptake and partitioning, responses of amino acids and other N compounds., Plant Biology, 4, 311–320, doi:10.1055/s-2002-32334, 2002. Fotelli, M. N., Nahm, M., Radoglou, K., Rennenberg, H., Halyvopoulos, G., and Matzarakis, A.: Seasonal and interannual ecophysiological responses of beech (Fagus sylvatica) at its south-eastern distribution limit in Europe., Forest Ecology and Management, 257, 1157–1164, doi:10.1016/j.foreco.2008.11.026, 2009. Fouillet, A., Rey, G., Laurent, F., Pavillon, G., Bellec, S., Guihenneuc-Jouyaux, C., Clavel, J., Jougla, E., and Hémon, D.: Excess mortality related to the August 2003 heat wave in France, International Archives of Occupational and Environmental Health, 80, 16–24, doi:10.1007/s00420-006-0089-4, 2006. Galiano, L., Martínez-Vilalta, J., and Lloret, F.: Drought-induced multifactor decline of Scots pine in the Pyrenees and potential vegetation change by the expansion of co-occurring oak species, Ecosystems, 13, 978–991, doi: 10.1007/s10021-010-9368-8, 2010. BIBLIOGRAPHY 131 Galiano, L., Martínez-Vilalta, J., Sabaté, S., and Lloret, F.: Determinants of drought effects on crown condition and their relationship with depletion of carbon reserves in a Mediterranean holm oak forest., Tree Physiology, 32, 478–489, doi:10.1093/treephys/tps025, 2012. Gallé, A. and Feller, U.: Changes of photosynthetic traits in beech saplings (Fagus sylvatica) under severe drought stress and during recovery., Physiologia Plantarum, 131, 412–421, doi:10.1111/j.1399-3054.2007.00972.x, 2007. Gallé, A., Florez-Sarasa, I., El Aououad, H., and Flexas, J.: The Mediterranean evergreen Quercus ilex and the semi-deciduous Cistus albidus differ in their leaf gas exchange regulation and acclimation to repeated drought and re-watering cycles., Journal of Experimental Botany, 62, 5207–5216, doi:10.1093/jxb/err233, 2011. García-Díez, M., Fernández, J., Fita, L., and Yagüe, C.: Seasonal dependence of WRF model biases and sensitivity to PBL schemes over Europe., Quarterly Journal of the Royal Meteorological Society, 139, 501–514, doi:10. 1002/qj.1976, 2013. García-Herrera, R., Díaz, J., Trigo, R. M., Luterbacher, J., and Fischer, E. M.: A review of the European summer heat save of 2003., Critical Reviews in Environmental Science and Technology, 40, 267–306, doi:10.1080/ 10643380802238137, 2010. García-Plazaola, J. I. and Becerril, J. M.: Effects of drought on photoprotective mechanisms in European beech (Fagus sylvatica L.) seedlings from different provenances., Trees, 14, 485–490, doi:10.1007/s004680000068, 2000. Gaul, D., Hertel, D., Borken, W., Matzner, E., and Leuschner, C.: Effects of experimental drought on the fine root system of mature Norway spruce., Forest Ecology and Management, 256, 1151–1159, doi:10.1016/j.foreco. 2008.06.016, 2008. Gea-Izquierdo, G., Viguera, B., Cabrera, M., and Cañellas, I.: Drought induced decline could portend widespread pine mortality at the xeric ecotone in managed mediterranean pine-oak woodlands., Forest Ecology and Management, 320, 70–82, doi:10.1016/j.foreco.2014.02.025, 2014. Geßler, A., Keitel, C., Nahm, M., and Rennenberg, H.: Water shortage affects the water and nitrogen balance in central European beech forests., Plant Biology, 6, 289–298, doi:10.1055/s-2004-820878, 2004. Geßler, A., Jung, K., Gasche, R., Papen, H., Heidenfelder, A., Börner, E., Metzler, B., Augustin, S., Hildebrand, E., and Rennenberg, H.: Climate and forest management influence nitrogen balance of European beech forests: microbial N transformations and inorganic N net uptake capacity of mycorrhizal roots., European Journal of Forest Research, 124, 95–111, doi:10.1007/s10342-005-0055-9, 2005. Geßler, A., Keitel, C., Kreuzwieser, J., Matyssek, R., Seiler, W., and Rennenberg, H.: Potential risks for European beech (Fagus sylvatica L.) in a changing climate., Trees, 21, 1–11, doi:10.1007/s00468-006-0107-x, 2007. Gieger, T. and Thomas, F. M.: Effects of defoliation ans drought stress on biomass partitioning and water relations of quercus robur and quercus petraea., Basic and Applied Ecology, 3, 171–181, doi:10.1078/1439-1791-00091, 2002. Gieger, T. and Thomas, F. M.: Differential response of two Central-European oak species to single and combined stress factors., Trees, 19, 607–618, doi:10.1007/s00468-005-0424-5, 2005. Gimeno, L., de la Torre, L., Nieto, R., García, R., Hernández, E., and Ribera, P.: Changes in the relationship NAO-Northern hemisphere temperature due to solar activity., Earth and Planetary Science Letters, 206, 15–20, doi:10.1016/S0012-821X(02)01090-7, 2003. Gimeno, T. E., Pas, B., Lemos-Filho, J. P., and Valladares, F.: Plasticity and stress tolerance override local adaptation in the responses of Mediterranean holm oak seedlings to drought and cold., Tree Physiology, 29, 87–98, doi:10.1093/treephys/tpn007, 2009. Giorgi, F.: Climate change hot-spots., Geophysical Research Letters, 33, 1–4, doi:10.1029/2006GL025734, 2006. 132 BIBLIOGRAPHY Giorgi, F. and Lionello, P.: Climate change projections for the Mediterranean region., Global and Planetary Change, 63, 90–104, doi:10.1016/j.gloplacha.2007.09.005, 2008. Giuggiola, A., Bugmann, H., Zingg, A., Dobbertin, M., and Rigling, A.: Reduction of stand density increases drought resistance in xeric Scots pine forests., Forest Ecology and Management, 310, 827–835, doi:10.1016/j. foreco.2013.09.030, 2013. Gollan, T., Schurr, U., and Schulze, E. D.: Stomatal response to drying soil in relation to changes in the xylem sap composition of Helianthus annuus. I. The concentration of cations, anions, amino acids in, and pH of, the xylem sap., Plant, Cell and Environment, 15, 551–559, doi:10.1111/j.1365-3040.1992.tb01489.x, 1992. Good, P., Bärring, L., Giannakopoulos, C., Holt, T., and Palutikof, J.: Non-linear regional relationships between climate extremes and annual mean temperatures in model projections for 1961-2099 over Europe., Climate Research, 31, 19–34, doi:10.3354/cr031019, 2006. Gouveia, C., Trigo, R. M., DaCamara, C. C., Libonati, R., and Pereira, J. M.: The North Atlantic Oscillation and European vegetation dynamics., International Journal of Climatology, 28, 1835–1847, doi:10.1002/joc.1682, 2008. Granier, A., Biron, P., and Lemoine, D.: Water balance, transpiration and canopy conductance in two beech stands., Agricultural and Forest Meteorology, 100, 291–308, doi:10.1016/S0168-1923(99)00151-3, 2000a. Granier, A., Ceschia, E., Damesin, C., Dufrêne, E., Epron, D., Gross, P., Lebaube, S., Le Dantec, V., Le Goff, N., Lemoine, D., Lucot, E., Ottorini, J. M., Pontailler, J. Y., and Saugier, B.: The carbon balance of a young beech forest., Functional Ecology, 14, 312–325, 2000b. Granier, A., Pilegaard, K., and Jensen, N. O.: Similar net ecosystem exchange of beech stands located in France and Denmark., Agricultural and Forest Meteorology, 114, 75–82, doi:10.1016/S0168-1923(02)00137-5, 2002. Granier, A., Reichstein, M., Bréda, N., Janssens, I. A., Falge, E., Ciais, P., Grünwald, T., Aubinet, M., Berbigier, P., Bernhofer, C., Buchmann, N., Facini, O., Grassi, G., Heinesch, B., Ilvesniemi, H., Keronen, P., Knohl, A., Köstner, B., Lagergren, F., Lindroth, A., Longdoz, B., Loustau, D., Mateus, J., Montagnani, L., Nys, C., Moors, E., Papale, D., Peiffer, M., Pilegaard, K., Pita, G., Pumpanen, J., Rambal, S., Rebmann, C., Rodrigues, A., Seufert, G., Tenhunen, J., Vesala, T., and Wang, Q.: Evidence for soil water control on carbon and water dynamics in European forests during the extremely dry year: 2003., Agricultural and Forest Meteorology, 143, 123–145, doi:10.1016/j.agrformet.2006.12.004, 2007. Grassi, G., Vicinelli, E., Ponti, F., Cantoni, L., and Magnani, F.: Seasonal and interannual variability of photosynthetic capacity in relation to leaf nitrogen in a deciduous forest plantation in northern Italy., Tree Physiology, 25, 349–360, doi:10.1093/treephys/25.3.349, 2005. Gratani, L. and Varone, L.: Adaptive photosynthetic strategies of the Mediterranean maquis species according to their origin., Photosynthetica, 42, 551–558, doi:10.1007/S11099-005-0012-3, 2004. Gratani, L., Catoni, R., and Varone, L.: Morphological, anatomical and physiological leaf traits of Q. ilex, P. latifolia, P. lentiscus, and M. communis and their response to Mediterranean climate stress factors., Botanical Studies, 54, 1–12, doi:10.1186/1999-3110-54-35, 2013. Grell, G. A. and Freitas, S. R.: A scale and aerosol aware stochastic convective parameterization for weather and air quality modeling., Atmospheric Chemistry and Physics, 14, 5233–5250, doi:10.5194/acp-14-5233-2014, 2014. Grossnickle, S. C. and Blake, T. J.: Water relations and morphological development of bare-root jack pine and white spruce seedlings: seedling establishment on a boreal cut-over site., Forest Ecology and Management, 18, 299–318, 1987. Gruber, A., Strobl, S., Veit, B., and Oberhuber, W.: Impact of drought on the temporal dynamics of wood formation in Pinus sylvestris., Tree Physiology, 30, 490–501, doi:10.1093/treephys/tpq003, 2010. BIBLIOGRAPHY 133 Gruber, A., Pirkebner, D., Florian, C., and Oberhuber, W.: No evidence for depletion of carbohydrate pools in Scots pine (Pinus sylvestris L .) under drought stress., Plant Biology, 14, 142–148, doi:10.1111/j.1438-8677. 2011.00467.x, 2012. Grünwald, T. and Bernhofer, C.: A decade of carbon, water and energy flux measurements of an old spruce forest at the Anchor Station Tharandt., Tellus, 59B, 387–396, doi:10.1111/j.1600-0889.2007.00259.x, 2007. Guemas, V., Salas-Mélia, D., Kageyama, M., Giordani, H., Voldoire, A., and Sanchez-Gomez, E.: Summer interactions between weather regimes and surface ocean in the North-Atlantic region., Climate Dynamics, 34, 527–546, doi:10.1007/s00382-008-0491-6, 2010. Guillod, B. P., Orlowsky, B., Miralles, D., Teuling, A. J., Blanken, P. D., Buchmann, N., Ciais, P., Ek, M., Findell, K. L., Gentine, P., Lintner, B. R., Scott, R. L., van den Hurk, B., and Seneviratne, S. I.: Land-surface controls on afternoon precipitation diagnosed from observational data: uncertainties and confounding factors., Atmospheric Chemistry and Physics, 14, 8343–8367, doi:10.5194/acp-14-8343-2014, 2014. Guillod, B. P., Orlowsky, B., Miralles, D. G., Teuling, A. J., and Seneviratne, S. I.: Reconciling spatial and temporal soil moisture effects on afternoon rainfall., Nature Communications, 6, 6443, doi:10.1038/ncomms7443, 2015. Gulías, J., Flexas, J., Abadía, A., and Medrano, H.: Photosynthetic responses to water deficit in six Mediterranean sclerophyll species: possible factors explaining the declining distribution of Rhamnus ludovici-salvatoris, an endemic Balearic species., Tree physiology, 22, 687–697, doi:10.1093/treephys/22.10.687, 2002. Guo, Z., Dirmeyer, P. A., Koster, R. D., Bonan, G., Chan, E., Cox, P., Gordon, C. T., Kanae, S., Kowalczyk, E., Lawrence, D., Liu, P., Lu, C.-H., Malyshev, S., McAvaney, B., McGregor, J. L., Mitchell, K., Mocko, D., Oki, T., Oleson, K. W., Pitman, A., Sud, Y. C., Taylor, C. M., Verseghy, D., Vasic, R., Xue, Y., and Yamada, T.: GLACE: The Global Land–Atmosphere Coupling Experiment. Part II: analysis., Journal of Hydrometeorology, 7, 611–625, 2006. Hacke, U. G. and Sperry, J. S.: Functional and ecological xylem anatomy., Perspectives in Plant Ecology, Evolution and Systematics, 4, 97–115, doi:10.1078/1433-8319-00017, 2001. Hacke, U. G., Sperry, J. S., Ewers, B. E., Ellsworth, D. S., Schäfer, K. V. R., and Oren, R.: Influence of soil porosity on water use in Pinus taeda., Oecologia, 124, 495–505, doi:10.1007/PL00008875, 2000. Halenka, T., Kalvová, J., Chládová, Z., Demeterová, A., Zemánková, K., and Belda, M.: On the capability of RegCM to capture extremes in long term regional climate simulation – comparison with the observations for Czech Republic., Theoretical and Applied Climatology, 86, 125–145, 2006. Hall, A. and Qu, X.: Using the current seasonal cycle to constrain snow albedo feedback in future climate change., Geophysical Research Letters, 33, 1–4, doi:10.1029/2005GL025127, 2006. Han, J. and Pan, H.-L.: Revision of convection and vertical diffusion schemes in the NCEP global forecast system., Weather and Forecasting, 26, 520–533, doi:10.1175/WAF-D-10-05038.1, 2011. Hansen, J., Sato, M., Hearty, P., Ruedy, R., Kelley, M., Masson-Delmotte, V., Russell, G., Tselioudis, G., Cao, J., Rignot, E., Velicogna, I., Kandiano, E., von Schuckmann, K., Kharecha, P., Legrande, a. N., Bauer, M., and Lo, K.-W.: Ice melt, sea level rise and superstorms: evidence from paleoclimate data, climate modeling, and modern observations that 2C global warming could be dangerous., Atmospheric Chemistry and Physics, 16, 3761–3812, doi:10.5194/acpd-15-20059-2015, 2016. Hare, W. L., Cramer, W., Schaeffer, M., Battaglini, A., and Jaeger, C. C.: Climate hotspots: key vulnerable regions, climate change and limits to warming., Regional Environmental Change, 11, 1–13, doi: 10.1007/s10113-010-0195-4, 2011. Harfouche, A.: Retrospective early test for adult vigor of Pinus pinaster families grown under two water regimes. Implications for early selection., Annales of Forest Science, 60, 539–547, doi:10.1051/forest, 2003. Hartmann, H., Ziegler, W., Kolle, O., and Trumbore, S.: Thirst beats hunger - declining hydration during drought prevents carbon starvation in Norway spruce saplings., New Phytologist, 200, 340–349, doi:10.1111/nph. 12331, 2013. 134 BIBLIOGRAPHY Haylock, M. R., Hofstra, N., Klein Tank, a. M. G., Klok, E. J., Jones, P. D., and New, M.: A European daily highresolution gridded data set of surface temperature and precipitation for 1950-2006., Journal of Geophysical Research Atmospheres, 113, 1–12, doi:10.1029/2008JD010201, 2008. Hendry, G. a. F., Finch-Savage, W. E., Thorpe, P. C., Atherton, N. M., Buckland, S. M., Nilsson, K. a., and Seel, W. E.: Free radical processes and loss of seed viability during desiccation in the recalcitrant species Quercus robur L., New Phytologist, 122, 273–279, doi:10.1111/j.1469-8137.1992.tb04231.x, 1992. Hereş, A. M., Martínez-Vilalta, J., and López, B. C.: Growth patterns in relation to drought-induced mortality at two Scots pine (Pinus sylvestris L.) sites in NE Iberian Peninsula., Trees - Structure and Function, 26, 621–630, doi:10.1007/s00468-011-0628-9, 2012. Hewitt, C. D. and Griggs, D. J.: Ensembles-based predictions of climate changes and their impacts., Eos, 85, 566, 2004. Hickler, T., Vohland, K., Feehan, J., Miller, P. A., Smith, B., Costa, L., Giesecke, T., Fronzek, S., Carter, T. R., Cramer, W., Kühn, I., and Sykes, M. T.: Projecting the future distribution of European potential natural vegetation zones with a generalized, tree species-based dynamic vegetation model., Global Ecology and Biogeography, 21, 50–63, doi:10.1111/j.1466-8238.2010.00613.x, 2012. Higgs, K. H. and Wood, V.: Drought susceptibility and xylem dysfunction in seedlings of 4 European oak species., Annales des Sciences Forestieres, 52, 507–513, doi:10.1051/forest:19950509, 1995. Hirschi, M., Seneviratne, S. I., Alexandrov, V., Boberg, F., Boroneant, C., Christensen, O. B., Formayer, H., Orlowsky, B., and Stepanek, P.: Observational evidence for soil-moisture impact on hot extremes in southeastern Europe., Nature Geoscience, 4, 17–21, doi:10.1038/ngeo1032, 2011. Hodzic, A., Vautard, R., Chepfer, H., Goloub, P., Menut, L., Chazette, P., Deuzé, J. L., Apituley, A., and Couvert, P.: Evolution of aerosol optical thickness over Europe during the August 2003 heat wave as seen from CHIMERE model simulations and POLDER data., Atmospheric Chemistry and Physics, 6, 1853–1864, doi: 10.5194/acpd-6-1853-2006, 2006. Hohenegger, C. and Stevens, B.: Controls on and impacts of the diurnal cycle of deep convective precipitation., Journal of Advances in Modeling Earth Systems, 5, 801–815, doi:10.1002/2012MS000216, 2013. Hohenegger, C., Brockhaus, P., Bretherton, C. S., and Schär, C.: The soil moisture-precipitation feedback in simulations with explicit and parameterized convection., Journal of Climate, 22, 5003–5020, doi:10.1175/ 2009JCLI2604.1, 2009. Holst, J., Barnard, R., Brandes, E., Buchmann, N., Gessler, A., and Jaeger, L.: Impacts of summer water limitation on the carbon balance of a Scots pine forest in the southern upper Rhine plain., Agricultural and Forest Meteorology, 148, 1815–1826, doi:10.1016/j.agrformet.2008.06.008, 2008. Hong, S.-Y. and Lim, J.-O. J.: The WRF single-moment 6-class microphysics scheme (WSM6)., Journal of the Korean Meteorological Society, 42, 129–151, 2006. Hong, S.-Y., Noh, Y., and Dudhia, J.: A new vertical diffusion package with an explicit treatment of entrainment processes., Monthly Weather Review, 134, 2318–2341, doi:10.1175/MWR3199.1, 2006. Horton, D. E., Johnson, N. C., Singh, D., Swain, D. L., Rajaratnam, B., and Diffenbaugh, N. S.: Contribution of changes in atmospheric circulation patterns to extreme temperature trends., Nature, 522, 465–469, doi: 10.1038/nature14550, 2015. Horton, P., Ruban, A. V., and Walters, R. G.: Regulation of light harvesting in green plants. Indication by nonphotochemical quenching of chlorophyll fluorescence., Plant physiology, 106, 415–420, doi:10.1146/annurev. arplant.47.1.655, 1994. Hourdin, F., Grandpeix, J. Y., Rio, C., Bony, S., Jam, A., Cheruy, F., Rochetin, N., Fairhead, L., Idelkadi, A., Musat, I., Dufresne, J. L., Lahellec, A., Lefebvre, M. P., and Roehrig, R.: LMDZ5B: The atmospheric component of the IPSL climate model with revisited parameterizations for clouds and convection., Climate Dynamics, 40, 2193–2222, doi:10.1007/s00382-012-1343-y, 2013. BIBLIOGRAPHY 135 Hu, B., Simon, J., and Rennenberg, H.: Drought and air warming affect the species-specific levels of stressrelated foliar metabolites of three oak species on acidic and calcareous soil., Tree Physiology, 33, 489–504, doi:10.1093/treephys/tpt025, 2013. Iacono, M. J., Delamere, J. S., Mlawer, E. J., Shephard, M. W., Clough, S. A., and Collins, W. D.: Radiative forcing by long-lived greenhouse gases: calculations with the AER radiative transfer models., Journal of Geophysical Research Atmospheres, 113, 1–8, doi:10.1029/2008JD009944, 2008. Ingelmo, F., Cuadrado, S., Ibañez, A., and Hernandez, J.: Hydric properties of some Spanish soils in relation to their rock fragment content: implications for runoff and vegetation., Catena, 23, 73–85, doi:10.1016/ 0341-8162(94)90054-X, 1994. INSEE: L’agriculture en 2003 en Europe et en France. Les grands pays agricoles affectés par la sécheresse., INSEE, p. 4, 2004. IPCC: IPCC Second Assessment: Climate Change 1995: A report of the Intergovernmental Panel on Climate Change. Full report., Tech. rep., Intergovernmental Panel on Climate Change, URL http://www.ipcc.ch/pdf/ climate-changes-1995/ipcc-2nd-assessment/2nd-assessment-en.pdf, 1995. IPCC: IPCC Fifth Assessment: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change., Tech. rep., Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, doi:10.1017/CBO9781107415324, 2013. Irvine, J., Perks, M. P., Magnani, F., and Grace, J.: The response of Pinus sylvestris to drought: stomatal control of transpiration and hydraulic conductance., Tree physiology, 18, 393–402, doi:10.1093/treephys/18.6.393, 1998. Jackson, R. B., Randerson, J. T., Canadell, J. G., Anderson, R. G., Avissar, R., Baldocchi, D. D., Bonan, G. B., Caldeira, K., Diffenbaugh, N. S., Field, C. B., Hungate, B. A., Jobbágy, E. G., Kueppers, L. M., Nosetto, M. D., and Pataki, D. E.: Protecting climate with forests., Environmental Research Letters, 3, 044 006, doi:10.1088/ 1748-9326/3/4/044006, 2008. Jacob, D., Petersen, J., Eggert, B., Alias, A., Christensen, O. B., Bouwer, L. M., Braun, A., Colette, A., Déqué, M., Georgievski, G., Georgopoulou, E., Gobiet, A., Menut, L., Nikulin, G., Haensler, A., Hempelmann, N., Jones, C., Keuler, K., Kovats, S., Kröner, N., Kotlarski, S., Kriegsmann, A., Martin, E., van Meijgaard, E., Moseley, C., Pfeifer, S., Preuschmann, S., Radermacher, C., Radtke, K., Rechid, D., Rounsevell, M., Samuelsson, P., Somot, S., Soussana, J. F., Teichmann, C., Valentini, R., Vautard, R., Weber, B., and Yiou, P.: EURO-CORDEX: new high-resolution climate change projections for European impact research., Regional Environmental Change, 14, 563–578, doi:10.1007/s10113-013-0499-2, 2014. Jaeger, E. B. and Seneviratne, S. I.: Impact of soil moisture-atmosphere coupling on European climate extremes and trends in a regional climate model., Climate Dynamics, 36, 1919–1939, doi:10.1007/s00382-010-0780-8, 2011. Janjić, Z. I.: The step-mountain eta coordinate model: further developments of the convection, viscous sublayer, and turbulence closure schemes., Monthly Weather Review, 122, 927–945, doi:10.1175/1520-0493(1994)122, 1994. Janjić, Z. I.: The surface layer in the NCEP ETA model., in: Eleventh conference on numerical weather prediction, Norfolk, VA, 19-23 August. American Meteorological Society, pp. 354–355, Boston, MA, 1996. Janjić, Z. I.: Nonsingular implementation of the Mellor-Yamada level 2.5 Scheme in the NCEP meso model, Tech. rep., National Centers for Environmental Prediction, URL http://www.emc.ncep.noaa.gov/officenotes/ newernotes/on437.pdf, 2002. Ježík, M., Blaženec, M., Střelcová, K., and Ditmarová, L.: The impact of the 2003-2008 weather variability on intra-annual stem diameter changes of beech trees at a submontane site in central Slovakia., Dendrochronologia, 29, 227–235, doi:10.1016/j.dendro.2011.01.009, 2011. 136 BIBLIOGRAPHY Jiang, B. and Liang, S.: Improved vegetation greenness increases summer atmospheric water vapor over Northern China., Journal of Geophysical Research: Atmospheres, 118, 8129–8139, doi:10.1002/jgrd.50602, 2013. Jin, J., Miller, N. L., and Schlegel, N.: Sensitivity study of four land surface schemes in the WRF model., Advances in Meteorology, 2010, 1–11, doi:10.1155/2010/167436, 2010. Jorge, I., Navarro, R. M., Lenz, C., Ariza, D., and Jorrín, J.: Variation in the holm oak leaf proteome at different plant developmental stages, between provenances and in response to drought stress., Proteomics, 6, S207– S214, doi:10.1002/pmic.200500364, 2006. Jung, M., Reichstein, M., Ciais, P., Seneviratne, S. I., Sheffield, J., Goulden, M. L., Bonan, G., Cescatti, A., Chen, J., de Jeu, R., Dolman, A. J., Eugster, W., Gerten, D., Gianelle, D., Gobron, N., Heinke, J., Kimball, J., Law, B. E., Montagnani, L., Mu, Q., Mueller, B., Oleson, K., Papale, D., Richardson, A. D., Roupsard, O., Running, S., Tomelleri, E., Viovy, N., Weber, U., Williams, C., Wood, E., Zaehle, S., and Zhang, K.: Recent decline in the global land evapotranspiration trend due to limited moisture supply., Nature, 467, 951–954, doi:10.1038/nature09396, 2010. Jyske, T., Hölttä, T., Mäkinen, H., Nöjd, P., Lumme, I., and Spiecker, H.: The effect of artificially induced drought on radial increment and wood properties of Norway spruce., Tree Physiology, 30, 103–115, doi:10.1093/ treephys/tpp099, 2010. Kain, J. S.: The Kain–Fritsch convective parameterization: an update., Journal of Applied Meteorology, 43, 170– 181, 2004. Kain, J. S. and Fritsch, J. M.: The role of the convective "trigger function" in numerical forecasts of mesoscale convective systems., Meteorology and Atmospheric Physics, 49, 93–106, doi:10.1007/BF01025402, 1992. Kamińska-Rozek, E. and Pukacki, P. M.: Effect of water deficit on oxidative stress and degradation of cell membranes in needles of Norway spruce (Picea abies)., Acta Physiologiae Plantarum, 26, 431–442, doi: 10.1007/s11738-004-0034-7, 2004. Kanamitsu, M., Lu, C. H., Schemm, J., and Ebisuzaki, W.: The predictability of soil moisture and near-surface temperature in hindcasts of the NCEP Seasonal Forecast model., Journal of Climate, 16, 510–521, doi:10. 1175/1520-0442(2003)016<0510:TPOSMA>2.0.CO;2, 2003. Keenan, T., Sabate, S., and Gracia, C.: Soil water stress and coupled photosynthesis-conductance models: bridging the gap between conflicting reports on the relative roles of stomatal, mesophyll conductance and biochemical limitations to photosynthesis., Agricultural and Forest Meteorology, 150, 443–453, doi:10.1016/j.agrformet. 2010.01.008, 2010. Kellomäki, S. and Wang, K.-y.: Photosynthetic responses to needle water potentials in Scots pine after a four-year exposure to elevated CO2 and temperature., Tree Physiology, 16, 765–772, 1996. Kishor, P. B. K., Hong, Z., Miao, G. H., Hu, C.-A. A., and Verma, D. P. S.: Overexpression of [delta]-pyrroline5-carboxylate synthetase increases proline production and confers osmotolerance in transgenic plants., Plant physiology, 108, 1387–1394, doi:10.1104/pp.108.4.1387, 1995. Knutti, R., Masson, D., and Gettelman, A.: Climate model genealogy: Generation CMIP5 and how we got there., Geophysical Research Letters, 40, 1194–1199, doi:10.1002/grl.50256, 2013. Knutti, R., Rogelj, J., Sedláček, J., and Fischer, E. M.: A scientific critique of the two-degree climate change target., Nature Geoscience, 9, 13–18, doi:10.1038/ngeo2595, 2016. Kohler, M., Sohn, J., Nägele, G., and Bauhus, J.: Can drought tolerance of Norway spruce (Picea abies (L.) Karst.) be increased through thinning?, European Journal of Forest Research, 129, 1109–1118, doi:10.1007/ s10342-010-0397-9, 2010. Koster, R. D., Dirmeyer, P. a., Guo, Z., Bonan, G., Chan, E., Cox, P., Gordon, C. T., Kanae, S., Kowalczyk, E., Lawrence, D., Liu, P., Lu, C.-H., Malyshev, S., McAvaney, B., Mitchell, K., Mocko, D., Oki, T., Oleson, K., Pitman, A., Sud, Y. C., Taylor, C. M., Verseghy, D., Vasic, R., Xue, Y., and Yamada, T.: Regions of strong coupling between soil moisture and precipitation, Science, 305, 1138–1140, doi:10.1126/science.1100217, 2004. BIBLIOGRAPHY 137 Koster, R. D., Guo, Z., Dirmeyer, P. A., Bonan, G. B., Chan, E., Cox, P. M., Gordon, C. T., Kanae, S., Kowalczyk, E., Lawrence, D. M., Liu, P., Lu, C. H., Malyshev, S., MacAvaney, B., McGregor, J. L., Mitchell, K., Mocko, D., Oki, T., Oleson, K. W., Pitman, A., Sud, Y. C., Taylor, C. M., Verseghy, D., Vasic, R., Xue, Y., and Yamada, T.: GLACE: The Global Land – Atmosphere Coupling Experiment. Part I: overview., Journal of Hydrometeorology, 7, 590–610, doi:10.1175/JHM511.1, 2006. Krams, I., Daukšte, J., Kivleniece, I., Brumelis, G., Cibulskis, R., Abolinš-abols, M., Rantala, M. J., Mierauskas, P., and Krama, T.: Drought-induced positive feedback in xylophagous insects: Easier invasion of Scots pine leading to greater investment in immunity of emerging individuals., Forest Ecology and Management, 270, 147–152, doi:10.1016/j.foreco.2012.01.012, 2012. Krinner, G., Viovy, N., de Noblet-Ducoudré, N., Ogée, J., Polcher, J., Friedlingstein, P., Ciais, P., Sitch, S., and Prentice, I. C.: A dynamic global vegetation model for studies of the coupled atmosphere-biosphere system., Global Biogeochemical Cycles, 19, 1–33, doi:10.1029/2003GB002199, 2005. Kyle, D. J., Ohad, I., and Arntzen, C. J.: Membrane protein damage and repair: selective loss of a quinone-protein function in chloroplast membranes., Proceedings of the National Academy of Sciences of the United States of America, 81, 4070–4074, doi:10.1073/pnas.81.13.4070, 1984. Le Provost, G., Domergue, F., Lalanne, C., Campos, P. R., Grosbois, A., Bert, D., Meredieu, C., Danjon, F., Plomion, C., and Gion, J.-M.: Soil water stress affects both cuticular wax content and cuticle-related gene expression in young saplings of maritime pine (Pinus pinaster Ait)., BMC plant biology, 13, 1–12, doi: 10.1186/1471-2229-13-95, 2013. Lee, S. S. and Donner, L. J.: Effects of cloud parameterization on radiation and precipitation: a comparison between single-moment microphysics and double-moment microphysics., Terrestrial, Atmospheric and Oceanic Sciences, 22, 403–420, doi:10.3319/TAO.2011.03.03.01(A), 2011. Lemoine, D., Cochard, H., and Granier, A.: Within crown variation in hydraulic architecture in beech (Fagus sylvatica L): evidence for a stomatal control of xylem embolism., Annal Forest Science, 59, 19–27, doi:10. 1051/forest, 2002. Limousin, J.-M., Longepierre, D., Huc, R., and Rambal, S.: Change in hydraulic traits of Mediterranean Quercus ilex subjected to long-term throughfall exclusion., Tree Physiology, 30, 1026–1036, doi:10.1093/treephys/ tpq062, 2010a. Limousin, J.-M., Misson, L., Lavoir, A.-V., Martin, N. K., and Rambal, S.: Do photosynthetic limitations of evergreen Quercus ilex leaves change with long-term increased drought severity?, Plant, Cell and Environment, 33, 863– 875, doi:10.1111/j.1365-3040.2009.02112.x, 2010b. Linderholm, H. W., Folland, C. K., and Walther, A.: A multicentury perspective on the summer North Atlantic Oscillation (SNAO) and drought in the eastern Atlantic Region., Journal of Quaternary Science, 24, 415–425, doi:10.1002/jqs.1261, 2009. Litvak, M. E., Loreto, F., Harley, P. C., Sharkey, T. D., and Monson, R. K.: The response of isoprene emission rate and photosynthetic rate to photon flux and nitrogen supply in aspen and white oak trees., Plant Cell and Environment, 19, 549–559, doi:10.1111/j.1365-3040.1996.tb00388.x, 1996. Llusià, J. and Peñuelas, J.: Changes in terpene content and emission in potted Mediterranean woody plants under severe drought., Canadian Journal of Botany, 76, 1366–1373, 1998. Lo Gullo, M. A. and Salleo, S.: Water storage in the wood and xylem cavitation in 1-year-old twigs of Populus deltoides Bartr., Plant, Cell and Environment, 15, 431–438, 1992. Loew, A., Holmes, T., and de Jeu, R.: The European heat wave 2003: early indicators from multisensoral microwave remote sensing?, Journal of Geophysical Research: Atmospheres, 114, 1–14, doi:10.1029/ 2008JD010533, 2009. López, R., Aranda, I., and Gil, L.: Osmotic adjustment is a significant mechanism of drought resistance in Pinus pinaster and Pinus canariensis., Investigacion Agraria-Sistemas Y Recursos Forestales, 18, 159–166, 2009. 138 BIBLIOGRAPHY López-Moreno, J. I. and Vicente-Serrano, S. M.: Positive and negative phases of the wintertime North Atlantic Oscillation and drought occurrence over Europe: a multitemporal-scale approach., Journal of Climate, 21, 1220–1243, doi:10.1175/2007JCLI1739.1, 2008. Lorenz, E. N.: Deterministic nonperiodic flow., Journal of the Atmospheric Sciences, 20, 130–141, doi:10.1175/ 1520-0469(1963)020, 1963. Lorenz, P. and Jacob, D.: Validation of temperature trends in the ENSEMBLES regional climate model runs driven by ERA40., Climate Research, 44, 167–177, doi:10.3354/cr00973, 2010. Lorenz, R., Davin, E. L., Lawrence, D. M., Stöckli, R., and Seneviratne, S. I.: How important is vegetation phenology for european climate and heat waves?, Journal of Climate, 26, 10 077–10 100, doi:10.1175/ JCLI-D-13-00040.1, 2013. Lorenz, R., Argüeso, D., Donat, M. G., Pitman, A. J., van den Hurk, B., Berg, A., Lawrence, D. M., Chéruy, F., Ducharne, A., Hagemann, S., Meier, A., Milly, P. C. D., and Seneviratne, S. I.: Influence of land-atmosphere feedbacks on temperature and precipitation extremes in the GLACE-CMIP5 ensemble., Journal of Geophysical Research: Atmospheres, 121, 607–623, doi:10.1002/2015JD024053, 2015. Loreto, F., Fischbach, R. J., Schnitzler, J.-P., Ciccioli, P., Brancaleoni, E., Calfapietra, C., and Seufert, G.: Monoterpene emission and monoterpene synthase activities in the Mediterranean evergreen oak Quercus ilex L. grown at elevated CO2 concentrations., Global Change Biology, 7, 709–717, doi:10.1046/j.1354-1013.2001.00442.x, 2001. Lu, P., Biron, P., Granier, A., and Cochard, H.: Water relations of Norway Spruce (Picea abies (L.) Karst) under soil drought in the Vosges mountains: whole-tree hydraulic conductance, xylem embolism and water loss regulation., Annals of Forest Science, 53, 117–129, 1996. Lu, Y., Duan, B., and Li, C.: Physiological responses to drought and shade in two contrasting Picea asperata populations., Canadian Journal of Forest Research, 37, 1253–1262, doi:10.1111/j.1399-3054.2005.00535.x, 2007. Luterbacher, J., Dietrich, D., Xoplaki, E., Grosjean, M., Wanner, H., Dartois, E., Wiesemeyer, H., and Jones, a. P.: European seasonal and annual temperature variability, trends, and extremes since 1500., Science, 303, 1499– 1503, doi:10.1126/science.1093877, 2004. Lynn, B. H., Rind, D., and Avissar, R.: The importance of mesoscale circulations generated by subgrid-scale landscape heterogeneities in general circulation models., Journal of Climate, 8, 191–205, 1995. Madsen, P. and Larsen, J. B.: Natural regeneration of beech (Fagus syhatica L .) with respect to canopy density, soil moisture and soil carbon content., Forest Ecology and Management, 97, 95–105, doi:10.1016/S0378-1127(97) 00091-1, 1997. Maier-Maercker, U.: Dynamics of change in stomatal response and water status of Picea abies during a persistent drought period: a contribution to the traditional view of plant water relations., Tree Physiology, 18, 211–222, 1998. Major, J. E. and Johnsen, K. H.: Shoot water relations of mature black spruce families displaying a genotype x environment interaction in growth rate. II. Temporal trends and response to varying soil water conditions., Tree Physiology, 19, 375–382, 1999. Major, J. E. and Johnsen, K. H.: Shoot water relations of mature black spruce families displaying a genotype x environment interaction in growth rate. III. Diurnal patterns as influenced by vapor pressure deficit and internal water status., Tree Physiology, 21, 579–587, 2001. Major, J. E., Johnsen, K. H., Barsi, D. C., and Campbell, M.: Needle parameter variation of mature black spruce displaying genetic x soil moisture interaction in growth., Canadian Journal of Forest Research, 42, 1926–1938, doi:10.1007/s00468-013-0865-1, 2012. Manes, F., Vitale, M., Donato, E., Giannini, M., and Puppi, G.: Different ability of three Mediterranean oak species to tolerate progressive water stress., Photosynthetica, 44, 387–393, doi:10.1007/s11099-006-0040-7, 2006. BIBLIOGRAPHY 139 Manion, P. D.: Tree disease concepts., Prentice Hall, Englewood Cliffs, New Jersey, 2 edn., 1991. Mariotti, A. and Dell’Aquila, A.: Decadal climate variability in the Mediterranean region: roles of large-scale forcings and regional processes., Climate Dynamics, 38, 1129–1145, doi:10.1007/s00382-011-1056-7, 2012. Marsham, J. H., Hobby, M., Allen, C. J. T., Banks, J. R., Bart, M., Brooks, B. J., Cavazos-Guerra, C., Engelstaedter, S., Gascoyne, M., Lima, A. R., Martins, J. V., McQuaid, J. B., O’Leary, A., Ouchene, B., Ouladichir, A., Parker, D. J., Saci, A., Salah-Ferroudj, M., Todd, M. C., and Washington, R.: Meteorology and dust in the central Sahara: observations from Fennec supersite-1 during the June 2011 intensive observation period., Journal of Geophysical Research Atmospheres, 118, 4069–4089, doi:10.1002/jgrd.50211, 2013. Martín, J. A., Esteban, L. G., de Palacios, P., and Fernández, F. G.: Variation in wood anatomical traits of Pinus sylvestris L. between Spanish regions of provenance., Trees, 24, 1017–1028, doi:10.1007/s00468-010-0471-4, 2010. Martin-Stpaul, N. K., Limousin, J. M., Vogt-Schilb, H., Rodríguez-Calcerrada, J., Rambal, S., Longepierre, D., and Misson, L.: The temporal response to drought in a Mediterranean evergreen tree: comparing a regional precipitation gradient and a throughfall exclusion experiment., Global Change Biology, 19, 2413–2426, doi: 10.1111/gcb.12215, 2013. Martínez-Ferri, E., Balaguer, L., Valladares, F., Chico, J. M., and Manrique, E.: Energy dissipation in droughtavoiding and drought-tolerant tree species at midday during the Mediterranean summer., Tree Physiology, 20, 131–138, doi:10.1093/treephys/20.2.131, 2000. Martínez-Vilalta, J. and Piñol, J.: Drought-induced mortality and hydraulic architecture in pine populations of the NE Iberian Peninsula., Forest Ecology and Management, 161, 247–256, doi:10.1016/S0378-1127(01)00495-9, 2002. Martínez-Vilalta, J., López, B. C., Adell, N., Badiella, L., and Ninyerola, M.: Twentieth century increase of Scots pine radial growth in NE Spain shows strong climate interactions., Global Change Biology, 14, 2868–2881, doi:10.1111/j.1365-2486.2008.01685.x, 2008. Matías, L. and Jump, A. S.: Interactions between growth, demography and biotic interactions in determining species range limits in a warming world: the case of Pinus sylvestris., Forest Ecology and Management, 282, 10–22, doi:10.1016/j.foreco.2012.06.053, 2012. Matías, L., González-Díaz, P., and Jump, A. S.: Larger investment in roots in southern range-edge populations of Scots pine is associated with increased growth and seedling resistance to extreme drought in response to simulated climate change., Environmental and Experimental Botany, 105, 32–38, doi:10.1016/j.envexpbot. 2014.04.003, 2014. Maxwell, K. and Johnson, G. N.: Chlorophyll fluorescence - a practical guide., Journal of Experimental Botany, 51, 659–668, doi:10.1093/jexbot/51.345.659, 2000. Mayr, S. and Cochard, H.: A new method for vulnerability analysis of small xylem areas reveals that compression wood of Norway spruce has lower hydraulic safety than opposite wood., Plant, Cell and Environment, 26, 1365–1371, doi:10.1046/j.0016-8025.2003.01060.x, 2003. McCully, M. E., Huang, C. X., and Ling, L. E. C.: Daily embolism and refilling of xylem vessels in the roots of field-grown maize., New Phytologist, 138, 327–342, doi:10.1046/j.1469-8137.1998.00101.x, 1998. McDowell, N., Pockman, W. T., Allen, C. D., Breshears, D. D., Cobb, N., Kolb, T., Plaut, J., Sperry, J., West, A., Williams, D. G., and Yepez, E. A.: Mechanisms of plant survival and mortality during drought: why do some plants survive while others succumb to drought?, New Phytologist, 178, 719–739, doi:10.1111/j.1469-8137. 2008.02436.x, 2008. Meier, I. C. and Leuschner, C.: Belowground drought response of European beech: fine root biomass and carbon partitioning in 14 mature stands across a precipitation gradient., Global Change Biology, 14, 2081–2095, doi: 10.1111/j.1365-2486.2008.01634.x, 2008. 140 BIBLIOGRAPHY Menconi, M., Sgherri, C. L. M., Pinzino, C., and Navari-lzzo, F.: Activated oxygen production and detoxification in wheat plants subjected to a water deficit programme., Journal of Experimental Botany, 46, 1123–1130, doi:10.1093/jxb/46.9.1123, 1995. Méthy, M., Damesin, C., and Rambal, S.: Drought and photosystem II activity in two Mediterranean oaks., Annal Forest Science, 53, 255–262, doi:10.1051/forest:19960208, 1996. Miralles, D. G., de Jeu, R. A. M., Gash, J. H., Holmes, T. R. H., and Dolman, A. J.: Magnitude and variability of land evaporation and its components at the global scale., Hydrology and Earth System Sciences, 15, 967–981, doi:10.5194/hess-15-967-2011, 2011. Miralles, D. G., van den Berg, M. J., Teuling, A. J., and De Jeu, R. A. M.: Soil moisture-temperature coupling: a multiscale observational analysis., Geophysical Research Letters, 39, 1–6, doi:10.1029/2012GL053703, 2012. Miralles, D. G., Teuling, A. J., van Heerwaarden, C. C., and Vilà-Guerau de Arellano, J.: Mega-heatwave temperatures due to combined soil desiccation and atmospheric heat accumulation., Nature Geoscience, 7, 345–349, doi:10.1038/ngeo2141, 2014. Misson, L., Nicault, A., and Guiot, J.: Effects of different thinning intensities on drought response in Norway spruce (Picea abies (L.) Karst.)., Forest Ecology and Management, 183, 47–60, doi:10.1016/S0378-1127(03) 00098-7, 2003. Misson, L., Panek, J. a., and Goldstein, A. H.: A comparison of three approaches to modeling leaf gas exchange in annually drought-stressed ponderosa pine forests., Tree physiology, 24, 529–541, doi:10.1093/treephys/24.5. 529, 2004. Misson, L., Limousin, J.-M., Rodriguez, R., and Letts, M. G.: Leaf physiological responses to extreme droughts in Mediterranean Quercus ilex forest., Plant, Cell and Environment, 33, 1898–1910, doi:10.1111/j.1365-3040. 2010.02193.x, 2010. Modrzyński, J. and Eriksson, G.: Response of Picea abies populations from elevational transects in the Polish Sudety and Carpathian mountains to simulated drought stress., Forest Ecology and Management, 165, 105– 116, doi:10.1016/S0378-1127(01)00651-X, 2002. Molthan, A. L. and Colle, B. a.: Comparisons of single- and double-moment microphysics schemes in the simulation of a synoptic-scale snowfall event., Monthly Weather Review, 140, 2982–3002, doi:10.1175/ MWR-D-11-00292.1, 2012. Mooney, P. a., Mulligan, F. J., and Fealy, R.: Evaluation of the sensitivity of the weather research and forecasting model to parameterization schemes for regional climates of europe over the period 1990-95., Journal of Climate, 26, 1002–1017, doi:10.1175/JCLI-D-11-00676.1, 2013. Morrison, H., Thompson, G., and Tatarskii, V.: Impact of cloud microphysics on the development of trailing stratiform precipitation in a simulated squall line: comparison of one- and two-moment schemes., Monthly Weather Review, 137, 991–1007, doi:10.1175/2008MWR2556.1, 2009. Mueller, B. and Seneviratne, S. I.: Hot days induced by precipitation deficits at the global scale., Proceedings of the National Academy of Sciences, 109, 12 398–12 403, doi:10.1073/pnas.1204330109, 2012. Mueller, B., Seneviratne, S. I., Jimenez, C., Corti, T., Hirschi, M., Balsamo, G., Ciais, P., Dirmeyer, P., Fisher, J. B., Guo, Z., Jung, M., Maignan, F., McCabe, M. F., Reichle, R., Reichstein, M., Rodell, M., Sheffield, J., Teuling, A. J., Wang, K., Wood, E. F., and Zhang, Y.: Evaluation of global observations-based evapotranspiration datasets and IPCC AR4 simulations., Geophysical Research Letters, 38, 1–7, doi:10.1029/2010GL046230, 2011. Munné-Bosch, S. and Alegre, L.: Changes in carotenoids, tocopherols and diterpenes during drought and recovery, and the biological significance of chlorophyll loss in Rosmarinus officinalis plants., Planta, 210, 925–931, doi: 10.1007/s004250050699, 2000. Munné-Bosch, S., Peñuelas, J., Asensio, D., and Llusià, J.: Airborne ethylene may alter antioxidant protection and reduce tolerance of holm oak to heat and drought stress., Plant physiology, 136, 2937–2947, doi:10.1104/pp. 104.050005, 2004. BIBLIOGRAPHY 141 Nakanishi, M. and Niino, H.: An improved Mellor-Yamada Level-3 model: its numerical stability and application to a regional prediction of advection fog., Boundary-Layer Meteorology, 119, 397–407, doi:10.1007/ s10546-005-9030-8, 2006. Nakanishi, M. and Niino, H.: Development of an improved turbulence closure model for the atmospheric boundary layer, Journal of the Meteorological Society of Japan, 87, 895–912, doi:10.2151/jmsj.87.895, 2009. Naudts, K., Ryder, J., McGrath, M. J., Otto, J., Chen, Y., Valade, A., Bellasen, V., Berhongaray, G., Bönisch, G., Campioli, M., Ghattas, J., de Groote, T., Haverd, V., Kattge, J., MacBean, N., Maignan, F., Merilä, P., Peñuelas, J., Peylin, P., Pinty, B., Pretzsch, H., Schulze, E. D., Solyga, D., Vuichard, N., Yan, Y., and Luyssaert, S.: A vertically discretised canopy description for ORCHIDEE (SVN r2290) and the modifications to the energy, water and carbon fluxes., Geoscientific Model Development, 8, 2035–2065, doi:10.5194/gmd-8-2035-2015, 2015. Naudts, K., Chen, Y., McGrath, M. J., Ryder, J., Valade, A., Otto, J., and Luyssaert, S.: Europe’s forest management did not mitigate climate warming., Science, 351, 597–601, 2016. Nguyen-Queyrens, A., Costa, P., Loustau, D., and Plomion, C.: Osmotic adjustment in Pinus pinaster cuttings in response to a soil drying cycle., Annales of Forest Science, 59, 759–799, doi:10.1051/forest, 2002. Nikulin, G., Kjellström, E., Hansson, U., Strandberg, G., and Ullerstig, A.: Evaluation and future projections of temperature, precipitation and wind extremes over Europe in an ensemble of regional climate simulations., Tellus, Series A: Dynamic Meteorology and Oceanography, 63, 41–55, doi:10.1111/j.1600-0870.2010.00466.x, 2011. Nillson, L. O.: Carbon sequestration in Norway spruce in south Sweden as influenced by air pollution, water availability, and fertilization., Water, Air, and Soil Pollution, 70, 177–186, doi:10.1007/s13398-014-0173-7.2, 1993. Ogaya, R. and Peñuelas, J.: Phenological patterns of Quercus ilex, Phillyrea latifolia, and Arbutus unedo growing under a field experimental drought., Ecoscience, 11, 263–270, 2004. Ogaya, R. and Peñuelas, J.: Contrasting foliar responses to drought in Quercus ilex and Phillyrea latifolia., Biologia Plantarum, 50, 373–382, doi:10.1007/s10535-006-0052-y, 2006. Ogaya, R. and Peñuelas, J.: Leaf mass per area ratio in Quercus ilex leaves under a wide range of climatic conditions. The importance of low temperatures., Acta Oecologica, 31, 168–173, doi:10.1016/j.actao.2006. 07.004, 2007. Ogaya, R., Peñuelas, J., Asensio, D., and Llusià, J.: Chlorophyll fluorescence responses to temperature and water availability in two co-dominant Mediterranean shrub and tree species in a long-term field experiment simulating climate change., Environmental and Experimental Botany, 71, 123–127, doi:10.1016/j.envexpbot.2011. 08.004, 2011. Oreskes, N., Shrader-Frechette, K., and Belitz, K.: Verification, validation, and confirmation of numerical models in the Earth sciences., Science, 263, 641–646, doi:10.1126/science.263.5147.641, 1994. Orth, R. and Seneviratne, S. I.: Using soil moisture forecasts for sub-seasonal summer temperature predictions in Europe., Climate Dynamics, 43, 3403–3418, doi:10.1007/s00382-014-2112-x, 2014. Osonubi, O. and Davies, W. J.: Solute accumulation in leaves and roots of woody plants subjected to water stress., Oecologia, 32, 323–332, 1978. Paiva, J. A. P., Garnier-Géré, P. H., Rodrigues, J. C., Alves, A., Santos, S., Graça, J., Le Provost, G., Chaumeil, P., Da Silva-Perez, D., Bosc, A., Fevereiro, P., and Plomion, C.: Plasticity of maritime pine (Pinus pinaster) wood-forming tissues during a growing season., New Phytologist, 179, 1080–1094, doi:10.1111/j.1469-8137. 2008.02536.x, 2008. Pearson, M., Saarinen, M., Nummelin, L., Heiskanen, J., Roitto, M., Sarjala, T., and Laine, J.: Tolerance of peat-grown Scots pine seedlings to waterlogging and drought: morphological, physiological, and metabolic responses to stress., Forest Ecology and Management, 307, 43–53, doi:10.1016/j.foreco.2013.07.007, 2013. 142 BIBLIOGRAPHY Pena, R., Simon, J., Rennenberg, H., and Polle, A.: Ectomycorrhiza affect architecture and nitrogen partitioning of beech (Fagus sylvatica L.) seedlings under shade and drought., Environmental and Experimental Botany, 87, 207–217, doi:10.1016/j.envexpbot.2012.11.005, 2013. Peña-Rojas, K., Aranda, X., and Fleck, I.: Stomatal limitation to CO2 assimilation and down-regulation of photosynthesis in Quercus ilex resprouts in response to slowly imposed drought., Tree physiology, 24, 813–822, doi:10.1093/treephys/24.7.813, 2004. Peñuelas, J. and Llusià, J.: Seasonal emission of monoterpenes by the Mediterranean tree Quercus ilex in field conditions: relations with photosynthetic rates, temperature and volatility., Physiologia Plantarum, 105, 641– 647, doi:10.1034/j.1399-3054.1999.105407.x, 1999a. Peñuelas, J. and Llusià, J.: Short-term responses of terpene emission rates to experimental changes of PFD in Pinus halepensis and Quercus ilex in summer field conditions., Environmental and Experimental Botany, 42, 61–68, doi:10.1016/S0098-8472(99)00018-0, 1999b. Peñuelas, J., Llusià, J., Piñol, J., and Filella, I.: Photochemical reflectance index and leaf photosynthetic radiationuse-efficiency assessment in Mediterranean trees., International Journal of Remote Sensing, 18, 2863–2868, doi:10.1080/014311697217387, 1997. Peñuelas, J., Filella, I., Lloret, F., Piñol, J., and Siscart, D.: Effects of a severe drought on water and nitrogen use by Quercus ilex and Phillyrea latifolia., Biologia Plantarum, 43, 47–53, 2000. Pepin, S., Plamondon, A. P., and Britel, A.: Water relations of black spruce trees on a peatland during wet and dry years., Wetlands, 22, 225–233, doi:10.1672/0277-5212(2002)022, 2002. Perdiguero, P., Barbero, M. d. C., Cervera, M. T., Collada, C., and Soto, A.: Molecular response to water stress in two contrasting Mediterranean pines (Pinus pinaster and Pinus pinea)., Plant physiology and biochemistry, 67, 199–208, doi:10.1016/j.plaphy.2013.03.008, 2013. Perkins, S. E. and Fischer, E. M.: The usefulness of different realizations for the model evaluation of regional trends in heat waves., Geophysical Research Letters, 40, 5793–5797, doi:10.1002/2013GL057833, 2013. Perks, M. P., Irvine, J., and Grace, J.: Xylem acoustic signals from mature Pinus sylvestris during an extended drought., Annal Forest Science, 61, 1–8, doi:10.1051/forest, 2004. Pesoli, P., Gratani, L., and Larcher, W.: Responses of Quercus ilex from different provenances to experimentally imposed water stress., Biologia Plantarum, 46, 577–581, 2003. Peuke, A. D. and Rennenberg, H.: Impacts of drought on mineral macro- and microelements in provenances of beech (Fagus sylvatica L.) seedlings., Tree Physiology, 31, 196–207, doi:10.1093/treephys/tpr007, 2011. Peuke, A. D., Schraml, C., Hartung, W., and Rennenberg, H.: Identification of drought-sensitive beech ecotypes by physiological parameters., New Phytologist, 154, 373–387, doi:10.1046/j.1469-8137.2002.00400.x, 2002. Pigott, C. D. and Pigott, S.: Water as a determinant of the distribution of trees at the boundary of the Mediterranean zone., Journal of Ecology, 81, 557–566, 1993. Piovesan, G., Biondi, F., Di Filippo, A., Alessandrini, A., and Maugeri, M.: Drought-driven growth reduction in old beech (Fagus sylvatica L.) forests of the central Apennines, Italy., Global Change Biology, 14, 1–17, doi: 10.1111/j.1365-2486.2008.01570.x, 2008. Pitman, A. J. and Narisma, G. T.: The role of land surface processes in regional climate change: a case study of future land cover change over south western Australia., Meteorology and Atmospheric Physics, 89, 235–249, doi:10.1007/s00703-005-0131-1, 2005. Pitman, A. J., De Noblet-Ducoudré, N., Cruz, F. T., Davin, E. L., Bonan, G. B., Brovkin, V., Claussen, M., Delire, C., Ganzeveld, L., Gayler, V., van den Hurk, B. J. J. M., Lawrence, P. J., Van Der Molen, M. K., Müller, C., Reick, C. H., Seneviratne, S. I., Strengen, B. J., and Voldoire, A.: Uncertainties in climate responses to past land cover change: first results from the LUCID intercomparison study., Geophysical Research Letters, 36, 1–6, doi:10.1029/2009GL039076, 2009. BIBLIOGRAPHY 143 Plaza, J., Núñez, L., Pujadas, M., Pérez-Pastor, R., Bermejo, V., García-Alonso, S., and Elvira, S.: Field monoterpene emission of Mediterranean oak (Quercus ilex) in the central Iberian Peninsula measured by enclosure and micrometeorological techniques: observation of drought stress effect., Journal of Geophysical Research D: Atmospheres, 110, 1–12, doi:10.1029/2004JD005168, 2005. Pleim, J. E.: A combined local and nonlocal closure model for the atmospheric boundary layer. Part I: model description and testing., Journal of Applied Meteorology and Climatology, 46, 1383–1395, doi: 10.1175/JAM2539.1, 2007. Poesen, J. and Lavee, H.: Rock fragments in top soils: significance and processes., Catena, 23, 1–28, doi:10.1016/ 0341-8162(94)90050-7, 1994. Pohl, B., Crétat, J., and Camberlin, P.: Testing WRF capability in simulating the atmospheric water cycle over Equatorial East Africa., Climate Dynamics, 37, 1357–1379, doi:10.1007/s00382-011-1024-2, 2011. Polacek, D., Kofler, W., and Oberhuber, W.: Radial growth of Pinus sylvestris growing on alluvial terraces is sensitive to water-level fluctuations., New Phytologist, 169, 299–308, doi:10.1111/j.1469-8137.2005.01589.x, 2006. Polomski, J. and Rigling, D.: Effect of watering regime on disease development in Pinus sylvestris seedlings inoculated with Bursaphelenchus vallesianus and B. Mucronatus., Plant Disease, 94, 1055–1061, 2010. Puhe, J.: Growth and development of the root system of Norway spruce (Picea abies) in forest stands—a review., Forest Ecology and Management, 175, 253–273, doi:http://dx.doi.org/10.1016/S0378-1127(02)00134-2, 2003. Pukacki, P. M. and Kamińska-Rozek, E.: Effect of drought stress on chlorophyll a fluorescence and electrical admittance of shoots in Norway spruce seedlings., Trees, 19, 539–544, doi:10.1007/s00468-005-0412-9, 2005. Quesada, B., Vautard, R., Yiou, P., Hirschi, M., and Seneviratne, S. I.: Asymmetric European summer heat predictability from wet and dry southern winters and springs., Nature Climate Change, 2, 736–741, doi: 10.1038/nclimate1536, 2012. Ramarohetra, J., Pohl, B., and Sultan, B.: Errors and uncertainties introduced by a regional climate model in climate impact assessments: example of crop yield simulations in West Africa., Environmental Research Letters, 10, 1–16, doi:10.1088/1748-9326/10/12/124014, 2015. Rambal, S., Methy, M., and Damesin, C.: Fluorescence signatures applied to climate change effects: case studies with Mediterranean oak ecosystems., in: Geoscience and Remote Sensing Symposium, pp. 631–633, IEEE, Pasasena, CA, USA, 1994. Rambal, S., Lempereur, M., Limousin, J. M., Martin-StPaul, N. K., Ourcival, J.-M., and Rodriguez-Calcerrada, J.: How drought severity constrains GPP and its partitioning among carbon pools in a Quercus ilex coppice?, Biogeosciences, 11, 6855–6869, doi:10.5194/bgd-11-1-2014, 2014. Rebetez, M. and Dobbertin, M.: Climate change may already threaten Scots pine stands in the Swiss Alps., Theoretical and Applied Climatology, 79, 1–9, doi:10.1007/s00704-004-0058-3, 2004. Reichstein, M., Tenhunen, J. D., Roupsard, O., Ourcival, J.-M., Rambal, S., Miglietta, F., Peressotti, A., Pecchiari, M., Tirone, G., and Valentini, R.: Severe drought effects on ecosystem CO2 and H2O fluxes at three Mediterranean evergreen sites: revision of current hypotheses?, Global Change Biology, 8, 999–1017, doi: 10.1046/j.1365-2486.2002.00530.x, 2002. Reichstein, M., Ciais, P., Papale, D., Valentini, R., Running, S., Viovy, N., Cramer, W., Granier, A., Ogée, J., Allard, V., Aubinet, M., Bernhofer, C., Buchmann, N., Carrara, A., Grünwald, T., Heimann, M., Heinesch, B., Knohl, A., Kutsch, W., Loustau, D., Manca, G., Matteucci, G., Miglietta, F., Ourcival, J.-M., Pilegaard, K., Pumpanen, J., Rambal, S., Schaphoff, S., Seufert, G., Soussana, J. F., Sanz, M. J., Vesala, T., and Zhao, M.: Reduction of ecosystem productivity and respiration during the European summer 2003 climate anomaly: a joint flux tower, remote sensing and modelling analysis., Global Change Biology, 13, 634–651, doi:10.1111/j.1365-2486.2006. 01224.x, 2007. 144 BIBLIOGRAPHY Rios-Entenza, A. and Miguez-Macho, G.: Moisture recycling and the maximum of precipitation in spring in the Iberian Peninsula., Climate Dynamics, 42, 3207–3231, doi:10.1007/s00382-013-1971-x, 2014. Robson, T. M., Rodríguez-Calcerrada, J., Sánchez-Gómez, D., and Aranda, I.: Summer drought impedes beech seedling performance more in a sub-Mediterranean forest understory than in small gaps., Tree Physiology, 29, 249–259, doi:10.1093/treephys/tpn023, 2009. Rodríguez-Calcerrada, J., Jaeger, C., Limousin, J.-M., Ourcival, J.-M., Joffre, R., and Rambal, S.: Leaf CO2 efflux is attenuated by acclimation of respiration to heat and drought in a Mediterranean tree., Functional Ecology, 25, 983–995, doi:10.1111/j.1365-2435.2011.01862.x, 2011. Rodríguez-Calcerrada, J., Martin-StPaul, N. K., Lempereur, M., Ourcival, J.-M., del Rey, M. d. C., Joffre, R., and Rambal, S.: Stem CO2 efflux and its contribution to ecosystem CO2 efflux decrease with drought in a Mediterranean forest stand., Agricultural and Forest Meteorology, 195-196, 61–72, doi:10.1016/j.agrformet. 2014.04.012, 2014. Roig, F. A., Barriopedro, D., García Herrera, R., Patón Dominguez, D., and Monge, S.: North atlantic oscillation signatures in western iberian tree-rings., Geografiska Annaler, Series A: Physical Geography, 91, 141–157, doi:10.1111/j.1468-0459.2009.00360.x, 2009. Rose, L., Leuschner, C., Köckemann, B., and Buschmann, H.: Are marginal beech (Fagus sylvatica L.) provenances a source for drought tolerant ecotypes?, European Journal of Forest Research, 128, 335–343, doi:10.1007/ s10342-009-0268-4, 2009. Rosner, S., Klein, A., Wimmer, R., and Karlsson, B.: Extraction of features from ultrasound acoustic emissions: A tool to assess the hydraulic vulnerability of Norway spruce trunkwood?, New Phytologist, 171, 105–116, doi:10.1111/j.1469-8137.2006.01736.x, 2006. Rowland, L., da Costa, A. C. L., Galbraith, D. R., Oliveira, R. S., Binks, O. J., Oliveira, A. A. R., Pullen, A. M., Doughty, C. E., Metcalfe, D. B., Vasconcelos, S. S., Ferreira, L. V., Malhi, Y., Grace, J., Mencuccini, M., and Meir, P.: Death from drought in tropical forests is triggered by hydraulics not carbon starvation., Nature, 528, 119–122, doi:10.1038/nature15539, 2015. Rowntree, P. R. and Bolton, J. A.: Simulation of the atmospheric response to soil moisture anomalies over Europe., Quarterly Journal of the Royal Meteorological Society, 109, 501–526, doi:10.1002/qj.49710946105, 1983. Rust, S. and Roloff, A.: Acclimation of crown structure to drought in Quercus robur L. - Intra- and inter-annual variation of abscission and traits of shed twigs., Basic and Applied Ecology, 5, 283–291, doi:10.1016/j.baae. 2004.03.003, 2004. Ryder, J., Polcher, J., Peylin, P., Ottlé, C., Chen, Y., van Gorsel, E., Haverd, V., McGrath, M. J., Naudts, K., Otto, J., Valade, A., and Luyssaert, S.: A multi-layer land surface energy budget model for implicit coupling with global atmospheric simulations., Geoscientific Model Development, 9, 223–245, doi:10.5194/gmdd-7-8649-2014, 2016. Salleo, S., Lo Gullo, M. a., Trifilò, P., and Nardini, A.: New evidence for a role of vessel-associated cells and phloem in the rapid xylem refilling of cavitated stems of Laurus nobilis L., Plant, Cell and Environment, 27, 1065–1076, doi:10.1111/j.1365-3040.2004.01211.x, 2004. Sánchez-Gómez, D., Robson, T. M., Gascó, A., Gil-Pelegrín, E., and Aranda, I.: Differences in the leaf functional traits of six beech (Fagus sylvatica L.) populations are reflected in their response to water limitation., Environmental and Experimental Botany, 87, 110–119, doi:10.1016/j.envexpbot.2012.09.011, 2013. Sánchez-Vilas, J. and Retuerto, R.: Quercus ilex shows significant among population variability in functional and growth traits but maintains invariant scaling relations in biomass allocation., International Journal of Plant Sciences, 168, 973–983, doi:10.1086/518943, 2007. Sanz-Pérez, V. and Castro-Díez, P.: Summer water stress and shade alter bud size and budburst date in three mediterranean Quercus species., Trees, 24, 89–97, doi:10.1007/s00468-009-0381-5, 2010. BIBLIOGRAPHY 145 Sanz-Pérez, V., Castro-Díez, P., and Joffre, R.: Seasonal carbon storage and growth in Mediterranean tree seedlings under different water conditions., Tree Physiology, 29, 1105–1116, doi:10.1093/treephys/tpp045, 2009. Sardans, J., Peñuelas, J., and Lope-Piedrafita, S.: Changes in water content and distribution in Quercus ilex leaves during progressive drought assessed by in vivo 1H magnetic resonance imaging., BMC plant biology, 10, 1–12, doi:10.1186/1471-2229-10-188, 2010. Sardans, J., Rivas-Ubach, A., Estiarte, M., Ogaya, R., and Peñuelas, J.: Field-simulated droughts affect elemental leaf stoichiometry in Mediterranean forests and shrublands., Acta Oecologica, 50, 20–31, doi:10.1016/j.actao. 2013.04.002, 2013. Saveyn, A. N., Steppe, K., and Lemeur, R.: Drought and the diurnal patterns of stem CO2 efflux and xylem CO2 concentration in young oak (Quercus robur)., Tree Physiology, 27, 365–374, 2007. Saxton, K. E. and Rawls, W. J.: Soil water characteristic estimates by texture and organic matter for hydrologic solutions., Soil Science Society of America Journal, 70, 1569–1578, doi:10.2136/sssaj2005.0117, 2006. Scartazza, A., Moscatello, S., Matteucci, G., Battistelli, A., and Brugnoli, E.: Seasonal and inter-annual dynamics of growth, non-structural carbohydrates and C stable isotopes in a Mediterranean beech forest., Tree Physiology, 33, 730–742, doi:10.1093/treephys/tpt045, 2013. Schär, C. and Jendritzky, G.: Climate change: hot news from summer 2003., Nature, 432, 559–560, doi:10.1038/ 432559a, 2004. Schär, C., Lüthi, D., Beyerle, U., and Heise, E.: The soil-precipitation feedback: a process study with a regional climate model., Journal of Climate, 12, 722–741, 1999. Schär, C., Vidale, P. L., Lüthi, D., Frei, C., Haberli, C., Liniger, M. A., and Appenzeller, C.: The role of increasing temperature variability in European summer heatwaves., Nature, 427, 332–336, doi:10.1038/nature02300, 2004. Scheffer, M., Carpenter, S., Foley, J. A., Folke, C., and Walker, B.: Catastrophic shifts in ecosystems., Nature, 413, 591–596, 2001. Schlyter, P., Stjernquist, I., Bärring, L., Jönsson, A. M., and Nilsson, C.: Assessment of the impacts of climate change and weather extremes on boreal forests in northern Europe, focusing on Norway spruce., Climate Research, 31, 75–84, doi:10.3354/cr031075, 2006. Sellin, A.: Variation in shoot water status of Picea abies (L.) Karst. trees with different life histories., Forest Ecology and Management, 97, 53–62, doi:10.1016/S0378-1127(97)00060-1, 1997. Sellin, A.: The dependence of water potential in shoots of Picea abies on air and soil water status., Annales Geophysicae, 16, 470–476, doi:10.1007/s005850050617, 1998. Seneviratne, S. I., Koster, R. D., Guo, Z., Dirmeyer, P. A., Kowalczyk, E., Lawrence, D., Liu, P., Mocko, D., Lu, C.-H., Oleson, K. W., and Verseghy, D.: Soil moisture memory in AGCM simulations: analysis of Global Land–Atmosphere Coupling Experiment (GLACE) data, Journal of Hydrometeorology, 7, 1090–1112, doi: 10.1175/JHM533.1, 2006a. Seneviratne, S. I., Lüthi, D., Litschi, M., and Schär, C.: Land-atmosphere coupling and climate change in Europe., Nature, 443, 205–209, doi:10.1038/nature05095, 2006b. Seneviratne, S. I., Corti, T., Davin, E. L., Hirschi, M., Jaeger, E. B., Lehner, I., Orlowsky, B., and Teuling, A. J.: Investigating soil moisture-climate interactions in a changing climate: a review., Earth-Science Reviews, 99, 125–161, doi:10.1016/j.earscirev.2010.02.004, 2010. Sergeant, K., Spiess, N., Renaut, J., Wilhelm, E., and Hausman, J. F.: One dry summer: a leaf proteome study on the response of oak to drought exposure., Journal of Proteomics, 74, 1385–1395, doi:10.1016/j.jprot.2011.03. 011, 2011. 146 BIBLIOGRAPHY Serrada, R., Allué, M., and San Miguel, A.: The coppice system in Spain. Current situation, state of the art and major areas to be investigated., Annali dell’Instituo Sperimentale per la Selvicoltura, 23, 266–275, 1992. Serrano, L. and Peñuelas, J.: Contribution of physiological and morphological adjustments to drought resistance in two Mediterranean tree species., Biologia Plantarum, 49, 551–559, doi:10.1007/s10535-005-0049-y, 2005. Sevanto, S., McDowell, N. G., Dickman, L. T., Pangle, R., and Pockman, W. T.: How do trees die? A test of the hydraulic failure and carbon starvation hypotheses., Plant, Cell and Environment, 37, 153–161, doi:10.1111/ pce.12141, 2014. Silim, S., Guy, R., Patterson, T., and Livingston, N.: Plasticity in water-use efficiency of Picea sitchensis, P. glauca and their natural hybrids., Oecologia, 128, 317–325, doi:10.1007/s004420100659, 2001. Simonin, G., Cochard, H., Delatour, C., Granier, A., and Dreyer, E.: Vulnerability of young oak seedlings (Quercus robur L.) to embolism: Responses to drought and to an inoculation with Ophiostoma querci (Georgevitch) Nannf., Annales des Sciences Forestieres, 51, 493–504, doi:10.1051/forest:19940505, 1994. Simpraga, M., Verbeeck, H., Demarcke, M., Joó, E., Pokorska, O., Amelynck, C., Schoon, N., Dewulf, J., van Langenhove, H., Heinesch, B., Aubinet, M., Laffineur, Q., Muller, J.-F., and Steppe, K.: Clear link between drought stress, photosynthesis and biogenic volatile organic compounds in Fagus sylvatica L., Atmospheric Environment, 45, 5254–5259, doi:10.1016/j.atmosenv.2011.06.075, 2011. Simunek, J., Jarvis, N. J., van Genuchten, M. T., and Gärdenäs, A.: Review and comparison of models for describing non-equilibrium and preferential flow and transport in the vadose zone., Journal of Hydrology, 272, 14–35, doi:10.1016/S0022-1694(02)00252-4, 2003. Skamarock, W. C., Klemp, J. B., Dudhi, J., Gill, D. O., Barker, D. M., Duda, M. G., Huang, X.-Y., Wang, W., and Powers, J. G.: A description of the advanced research WRF version 3, Tech. Rep. June, National Center for Atmospheric Research, Boulder, Colorado, USA, doi:10.5065/D6DZ069T, 2008. Smirnoff, N. and Cumbes, Q. J.: Hydroxyl radical scavening activity of compatible solutes., Phytochemistry, 28, 1057–1060, 1989. Snyder, P. K. and Liess, S.: The simulated atmospheric response to expansion of the Arctic boreal forest biome., Climate Dynamics, 42, 487–503, doi:10.1007/s00382-013-1746-4, 2014. Sohn, J. A., Kohler, M., Gessler, A., and Bauhus, J.: Interactions of thinning and stem height on the drought response of radial stem growth and isotopic composition of Norway spruce (Picea abies)., Tree Physiology, 32, 1199–1213, doi:10.1093/treephys/tps077, 2012. Solberg, S., Hov, Ø., Søvde, A., Isaksen, I. S. A., Coddeville, P., de Backer, H., Forster, C., Orsolini, Y., and Uhse, K.: European surface ozone in the extreme summer 2003., Journal of Geophysical Research Atmospheres, 113, 1–16, doi:10.1029/2007JD009098, 2008. Somot, S., Sevault, F., Déqué, M., and Crépon, M.: 21st century climate change scenario for the Mediterranean using a coupled atmosphere-ocean regional climate model., Global and Planetary Change, 63, 112–126, doi: 10.1016/j.gloplacha.2007.10.003, 2008. Sperry, J. S. and Ikeda, T.: Xylem cavitation in roots and stems of Douglas-fir and white fir., Tree Physiology, 17, 275–280, doi:10.2144/05384CI01, 1997. Sperry, J. S. and Saliendra, N. Z.: Intra- and inter-plant variation in xylem cavitation in Betula occidentalis., Plant, Cell and Environment, 17, 1233–1241, doi:10.1111/j.1365-3040.1994.tb02021.x, 1994. Sperry, J. S. and Tyree, M. T.: Water-stress-induced xylem embolism in 3 species of conifers., Plant Cell and Environment, 13, 427–436, doi:10.1111/j.1365-3040.1990.tb01319.x, 1990. Spiecker, H.: Growth variation and environmental stresses: long-term observations on permanent research plots in Southwestern Germany., Water, Air, and Soil Pollution, 54, 247–256, doi:10.1007/BF00298669, 1990. BIBLIOGRAPHY 147 Spieß, N., Oufir, M., Matušíková, I., Stierschneider, M., Kopecky, D., Homolka, A., Burg, K., Fluch, S., Hausman, J.-F., and Wilhelm, E.: Ecophysiological and transcriptomic responses of oak (Quercus robur) to long-term drought exposure and rewatering., Environmental and Experimental Botany, 77, 117–126, doi: 10.1016/j.envexpbot.2011.11.010, 2012. Spracklen, D. V., Arnold, S. R., and Taylor, C. M.: Observations of increased tropical rainfall preceded by air passage over forests., Nature, 489, 282–285, doi:10.1038/nature11390, 2012. Staudt, M., Rambal, S., and Joffre, R.: Impact of drought on seasonal monoterpene emissions from Quercus ilex in southern France., Journal of Geophysical Research Atmospheres, 107, 1–9, doi:10.1029/2001JD002043, 2002. Stéfanon, M., Drobinski, P., D’Andrea, F., and De Noblet-Ducoudré, N.: Effects of interactive vegetation phenology on the 2003 summer heat waves., Journal of Geophysical Research Atmospheres, 117, 1–12, doi:10.1029/ 2012JD018187, 2012. Stéfanon, M., Drobinski, P., D’Andrea, F., Lebeaupin-Brossier, C., and Bastin, S.: Soil moisture-temperature feedbacks at meso-scale during summer heat waves over Western Europe., Climate Dynamics, 42, 1309–1324, doi:10.1007/s00382-013-1794-9, 2014. Stegehuis, A. I., Teuling, A. J., Ciais, P., Vautard, R., and Jung, M.: Future European temperature change uncertainties reduced by using land heat flux observations., Geophysical Research Letters, 40, 1–4, doi: 10.1002/grl.50404, 2013a. Stegehuis, A. I., Vautard, R., Ciais, P., Teuling, A. J., Jung, M., and Yiou, P.: Summer temperatures in Europe and land heat fluxes in observation-based data and regional climate model simulations., Climate Dynamics, 41, 455–477, doi:10.1007/s00382-012-1559-x, 2013b. Stegehuis, A. I., Vautard, R., Ciais, P., Teuling, A. J., Miralles, D. G., and Wild, M.: An observation-constrained multi-physics WRF ensemble for simulating European mega heat waves., Geoscientific Model Development, 8, 2285–2298, doi:10.5194/gmd-8-2285-2015, 2015. Stewart, J. D., Zine El Abidine, A., and Bernier, P. Y.: Stomatal and mesophyll limitations of photosynthesis in black spruce seedlings during multiple cycles of drought., Tree physiology, 15, 57–64, 1995. Stojnić, S., Orlović, S., Pilipović, A., Vilotić, D., Sijačić-Nikolić, M., and Miljković, D.: Variation in leaf physiology among three provenances of European beech (Fagus sylvatica L.) in provenance trial in Serbia., Genetika, 44, 341–353, doi:10.2298/GENSR1202341S, 2012. Sukoriansky, S., Galperin, B., and Perov, V.: Application of a new spectral theory of stably stratified turbulence to the atmospheric boundary layer over sea ice., Boundary-Layer Meteorology, 117, 231–257, doi:10.1007/ s10546-004-6848-4, 2005. Sutton, R. T. and Hodson, D. L. R.: North Atlantic forcing of North American and European summer climate., Science, 309, 115–118, doi:10.1126/science.1109496, 2005. Sykes, M. T. and Prentice, I. C.: Boreal forest futures: modelling the controls on tree species range limits and transient responses to climate change., Water, Air and Soil Pollution, 82, 415–428, doi:10.1007/BF01182851, 1995. Tan, W. and Blake, T. J.: Gas exchange and water relations responses to drought of fast- and slow growing black spruce families., Canadian Journal Of Botany, 75, 1700–1706, 1997. Taylor, C. M.: Detecting soil moisture impacts on convective initiation in Europe., Geophysical Research Letters, 42, 4631–4638, doi:10.1002/2015GL064030.Received, 2015. Taylor, C. M., Gounou, A., Guichard, F., Harris, P. P., Ellis, R. J., Couvreux, F., and de Kauwe, M.: Frequency of Sahelian storm initiation enhanced over mesoscale soil-moisture patterns., Nature Geoscience, 4, 430–433, doi:10.1038/ngeo1173, 2011. 148 BIBLIOGRAPHY Terashima, I., Wong, S.-C., Osmond, C. B., and Farquhar, G. D.: Characterisation of non-uniform photosynthesis induced by abscisic acid in leaves having different mesophyll anatomies., Plant and Cell Physiology, 29, 385– 394, 1988. Terradas, J. and Savé, R.: The influence of summer and winter stress and water relationships on the distribution of Quercus ilex L., Vegetatio, 99-100, 137–145, 1992. Teuling, A. J. and Seneviratne, S. I.: Contrasting spectral changes limit albedo impact on land-atmosphere coupling during the 2003 European heat wave., Geophysical Research Letters, 35, 1–5, doi:10.1029/ 2007GL032778, 2008. Teuling, A. J., Seneviratne, S. I., Stockli, R., Reichstein, M., Moors, E., Ciais, P., Luyssaert, S., van den Hurk, B., Ammann, C., Bernhofer, C., Dellwik, E., Gianelle, D., Gielen, B., Grünwald, T., Klumpp, K., Montagnani, L., Moureaux, C., Sottocornola, M., and Wohlfahrt, G.: Contrasting response of European forest and grassland energy exchange to heatwaves., Nature Geoscience, 3, 722–727, doi:10.1038/ngeo950, 2010. Tewari, M., Chen, F., Wang, W., Dudhia, J., LeMone, M. A., Mitchell, K., Ek, M., Gayno, G., Wegiel, J., and Cuenca, R. H.: Implementation and verification of the unified noah land surface model in the WRF model., in: 20th conference on weather analysis and forecasting/16th conference on numerical weather prediction, pp. 11–15, 2004. Thomas, F. M. and Gausling, T.: Morphological and physiological responses of oak seedlings (Quercus petraea and Q . robur) to moderate drought., Annal Forest Science, 57, 325–333, 2000. Thomas, F. M. and Hartmann, G.: Soil and tree water relations in mature oak stands of northern Germany differing in the degree of decline., Annales des Sciences Forestieres, 53, 697–720, doi:10.1051/forest:19960247, 1996. Thompson, G., Fielk, P. R., Rasmussen, R. M., and Hall, W. D.: Explicit forecasts of winter precipitation using an iImproved bulk microphysics scheme. Part II: implementation of a new snow parameterization., Monthly Weather Review, 136, 5095–5115, doi:10.1175/1520-0493(2004)132<0519:EFOWPU>2.0.CO;2, 2008. Tiedtke, M.: A comprehensive mass flux scheme for cumulus parameterization in large-scale models., Monthly Weather Review, 117, 1179–1800, doi:10.1175/1520-0493(1989)117, 1989. Tognetti, R., Johnson, J. D., and Michelozzi, M.: The response of European beech (Fagus sylvatica L .) seedlings from two Italian populations to drought and recovery., Trees, 9, 348–354, 1995. Tognetti, R., Longobucco, A., and Raschi, A.: Vulnerability of xylem to embolism in relation to plant hydraulic resistance in Quercus pubescence and Quercus ilex co-occurring in a Mediterranean coppice stand in central Italy., New Phytologist, 139, 437–447, 1998. Toivonen, P. and Vidaver, W.: Variable chlorophyll a fluorescence and CO2 uptake in water-stressed white spruce seedlings., Plant physiology, 86, 744–748, doi:10.1104/pp.86.3.744, 1988. Townend, J.: Effects of elevated carbon dioxide and drought on the growth and physiology of clonal Sitka spruce plants (Picea sitchensis (Bong.) Carr.)., Tree Physiology, 13, 389–399, 1993. Traore, A. K., Ciais, P., Vuicard, N., Poulter, B., Viovy, N., Guimberteau, M., Jung, M., Myneni, R., and Fisher, J. B.: Evaluation of the ORCHIDEE ecosystem model over Africa against 25 years of satellite-based water and carbon measurements., Journal of Geophysical Research: Biogeosciences, 119, 1554–1575, doi:10.1002/ 2014JG002638.Received, 2014. Tressol, M., Ordonez, C., Zbinden, R., Brioude, J., Thouret, V., Mari, C., Nedelec, P., Cammas, J.-P., Smit, H., Patz, H.-W., and Volz-Thomas, A.: Air pollution during the 2003 European heat wave as seen by MOZAIC airliners., Atmospheric Chemistry and Physics, 8, 2133–2150, doi:10.5194/acpd-8-2133-2008, 2008. Triboulot, M. B., Fauveau, M. L., Breda, N., Label, P., and Dreyer, E.: Stomatal conductance and xylem-sap abscisic acid (ABA) in adult oak trees during a gradually imposed drought., Annales Des Sciences Forestieres, 53, 207–220, doi:10.1051/forest:19960204, 1996. BIBLIOGRAPHY 149 Trigo, R. M., García-Herrera, R., Díaz, J., Trigo, I. F., and Valente, M. A.: How exceptional was the early August 2003 heatwave in France?, Geophysical Research Letters, 32, 1–4, doi:10.1029/2005GL022410, 2005. Tsonev, T., Wahbi, S., Sun, P., Sorrentino, G., and Centritto, M.: Gas exchange, water relations and their relationships with photochemical reflectance index in quercus ilex plants during water stress and recovery., International Journal of Agriculture and Biology, 16, 335–341, 2014. Tulik, M.: The anatomical traits of trunk wood and their relevance to oak (Quercus robur L.) vitality., European Journal of Forest Research, 133, 845–855, doi:10.1007/s10342-014-0801-y, 2014. Tyree, M. T. and Sperry, J. S.: Do woody-plants operate near the point of catastrophic xylem dysfunction caused by dynamic water-stress - answers from a model., Plant Physiology, 88, 574–580, doi:10.1104/pp.88.3.574, 1988. Tyree, M. T. and Zimmermann, M. H.: Xylem structure and the ascent of sap., Springer-Verlag Berlin Heidelberg, 2 edn., 2002. Valentini, R.: EUROFLUX: an integrated network for studying the long-term responses of biospheric exchanges of carbon, water, and energy of European forests., in: Fluxes of Carbon, Water and Energy of European Forests, edited by Valentini, R., vol. 163, pp. 1–8, Springer-Verlag Berlin Heidelberg, Heidelberg, 1 edn., 2003. Valentini, R., De Angelis, P., Matteucci, G., Monaco, R., Dore, S., and Scarascia Mugnozza, G. E.: Seasonal net carbon dioxide exchange of a beech forest with the atmosphere., Global Change Biology, 2, 199–207, doi: 10.1111/j.1365-2486.1996.tb00072.x, 1996. Valero-Galván, J., González-Fernández, R., Navarro-Cerrillo, R. M., Gil-Pelegrín, E., and Jorrín-Novo, J. V.: Physiological and proteomic analyses of drought stress response in Holm oak provenances., Journal of Proteome Research, 12, 5110–5123, doi:10.1021/pr400591n, 2013. Valladares, F., Dobarro, I., Sánchez-Gómez, D., and Pearcy, R. W.: Photoinhibition and drought in Mediterranean woody saplings: scaling effects and interactions in sun and shade phenotypes., Journal of Experimental Botany, 56, 483–494, doi:10.1093/jxb/eri037, 2005. van der Linden, P. and Mitchell, J. F. B.: Climate change and its impacts., Tech. rep., Met Office, Hadley Centre, Exeter, doi:PNR61,SIGE, 2009. van der Maaten, E.: Thinning prolongs growth duration of European beech (Fagus sylvatica L.) across a valley in southwestern Germany., Forest Ecology and Management, 306, 135–141, doi:10.1016/j.foreco.2013.06.030, 2013. van Haren, R., Haarsma, R. J., de Vries, H., van Oldenborgh, G. J., and Hazeleger, W.: Resolution dependence of circulation forced future central European summer drying., Environmental Research Letters, 10, 1–9, doi: 10.1088/1748-9326/10/5/055002, 2015. van Oijen, M., Beer, C., Cramer, W., Rammig, A., Reichstein, M., Rolinski, S., and Soussana, J.-F.: A novel probabilistic risk analysis to determine the vulnerability of ecosystems to extreme climatic events., Environmental Research Letters, 8, 1–7, doi:10.1088/1748-9326/8/1/015032, 2013a. van Oijen, M., Reyer, C., Bohn, F. J., Cameron, D. R., Deckmyn, G., Flechsig, M., Härkönen, S., Hartig, F., Huth, A., Kiviste, A., Lasch, P., Mäkelä, A., Mette, T., Minunno, F., and Rammer, W.: Bayesian calibration, comparison and averaging of six forest models, using data from Scots pine stands across Europe., Forest Ecology and Management, 289, 255–268, doi:10.1016/j.foreco.2012.09.043, 2013b. van Oijen, M., Balkovi, J., Beer, C., Cameron, D. R., Ciais, P., Cramer, W., Kato, T., Kuhnert, M., Martin, R., Myneni, R., Rammig, A., Rolinski, S., Soussana, J. F., Thonicke, K., van der Velde, M., and Xu, L.: Impact of droughts on the carbon cycle in European vegetation: a probabilistic risk analysis using six vegetation models., Biogeosciences, 11, 6357–6375, doi:10.5194/bg-11-6357-2014, 2014. Vautard, R. and Yiou, P.: Control of recent European surface climate change by atmospheric flow., Geophysical Research Letters, 36, 1–6, doi:10.1029/2009GL040480, 2009. 150 BIBLIOGRAPHY Vautard, R., Honoré, C., Beekmann, M., and Rouil, L.: Simulation of ozone during the August 2003 heat wave and emission control scenarios., Atmospheric Environment, 39, 2957–2967, doi:10.1016/j.atmosenv.2005.01.039, 2005. Vautard, R., Yiou, P., D’Andrea, F., de Noblet, N., Viovy, N., Cassou, C., Polcher, J., Ciais, P., Kageyama, M., and Fan, Y.: Summertime European heat and drought waves induced by wintertime Mediterranean rainfall deficit., Geophysical Research Letters, 34, 1–5, doi:10.1029/2006GL028001, 2007. Vautard, R., Gobiet, A., Jacob, D., Belda, M., Colette, A., Déqué, M., Fernández, J., García-Díez, M., Goergen, K., Güttler, I., Halenka, T., Karacostas, T., Katragkou, E., Keuler, K., Kotlarski, S., Mayer, S., van Meijgaard, E., Nikulin, G., Patarčić, M., Scinocca, J., Sobolowski, S., Suklitsch, M., Teichmann, C., Warrach-Sagi, K., Wulfmeyer, V., and Yiou, P.: The simulation of European heat waves from an ensemble of regional climate models within the EURO-CORDEX project., Climate Dynamics, 41, 2555–2575, doi:10.1007/s00382-013-1714-z, 2013. Vautard, R., Gobiet, A., Sobolowski, S., Kjellström, E., Stegehuis, A. I., Watkiss, P., Mendlik, T., Landgren, O., Nikulin, G., Teichmann, C., and Jacob, D.: The European climate under a 2C global warming, Environmental Research Letters, 9, 1–11, doi:10.1088/1748-9326/9/3/034006, 2014. Vaz, M., Pereira, J. S., Gazarini, L. C., David, T. S., David, J. S., Rodrigues, A., Maroco, J., and Chaves, M. M.: Drought-induced photosynthetic inhibition and autumn recovery in two Mediterranean oak species (Quercus ilex and Quercus suber)., Tree Physiology, 30, 946–956, doi:10.1093/treephys/tpq044, 2010. Vilà-Cabrera, A., Martínez-Vilalta, J., and Retana, J.: Variation in reproduction and growth in declining Scots pine populations., Perspectives in Plant Ecology, Evolution and Systematics, 16, 111–120, doi:10.1016/j.ppees. 2014.02.005, 2014. Vilà-Guerau De Arellano, J., van Heerwaarden, C. C., and Lelieveld, J.: Modelled suppression of boundary-layer clouds by plants in a CO2-rich atmosphere., Nature Geoscience, 5, 701–704, doi:10.1038/ngeo1554, 2012. Villar-Salvador, P., Planelles, R., Oliet, J., Peñuelas-Rubira, J. L., Jacobs, D. F., and González, M.: Drought tolerance and transplanting performance of holm oak (Quercus ilex) seedlings after drought hardening in the nursery., Tree physiology, 24, 1147–1155, doi:10.1093/treephys/24.10.1147, 2004. Vivin, P., Aussenac, G., and Levy, G.: Differences in drought resistance among 3 deciduous oak species grown in large boxes., Annales des Sciences Forestieres, 50, 221–233, doi:10.1051/forest:19930301, 1993. Wallin, G., Karlsson, P. E., Selldén, G., Ottosson, S., Medin, E.-L., Pleijel, H., and Skärby, L.: Impact of four years exposure to different levels of ozone, phosphorus and drought on chlorophyll, mineral nutrients, and stem volume of Norway spruce, Picea abies., Physiologia plantarum, 114, 192–206, doi:10.1034/j.1399-3054.2002. 1140205.x, 2002. Wang, G. and Eltahir, E. A. B.: Biosphere—atmosphere interactions over West Africa. II: multiple climate equilibria., Quarterly Journal of the Royal Meteorological Society, 126, 1261–1280, doi:10.1002/qj.49712656504, 2000. Wang, G., Dolman, A. J., and Alessandri, A.: A summer climate regime over Europe modulated by the North Atlantic Oscillation., Hydrology and Earth System Sciences, 15, 57–64, doi:10.5194/hess-15-57-2011, 2011. Wang, W. and Kumar, A.: A GCM assessment of atmospheric seasonal predictability associated with soil moisture anomalies over North America., Journal of Geophysical Research, 103, 28 637–28 646, doi:10.1029/ 1998JD200010, 1998. Wei, J. and Dirmeyer, P. A.: Dissecting soil moisture-precipitation coupling., Geophysical Research Letters, 39, 1–6, doi:10.1029/2012GL053038, 2012. Wild, M., Gilgen, H., Roesch, A., Ohmura, A., Long, C. N., Dutton, E. G., Forgan, B., Kallis, A., Russak, V., and Tsvetkov, A.: From dimming to brightening: decadal changes in solar radiation at Earth’s surface., Science, 308, 847–850, doi:10.1126/science.1103215, 2005. BIBLIOGRAPHY 151 Wolf, S., Keenan, T., Fisher, J., Baldocchi, D., Desai, A., Richardson, A., Scott, R., Law, B., Litvak, M., Brunsell, N., Peters, W., and van der Laan-Luijkx, I.: Warm spring reduced carbon cycle impact of the 2012 summer drought., Proceedings of the National Academy of Sciences of the United States of America, Early edit, 1–6, doi:10.1073/pnas.1519620113, 2016. Wortemann, R., Herbette, S., Barigah, T. S., Fumanal, B., Alia, R., Ducousso, A., Gomory, D., Roeckel-Drevet, P., and Cochard, H.: Genotypic variability and phenotypic plasticity of cavitation resistance in Fagus sylvatica L. across Europe., Tree Physiology, 31, 1175–1182, doi:10.1093/treephys/tpr101, 2011. Wramneby, A., Smith, B., and Samuelsson, P.: Hot spots of vegetation-climate feedbacks under future greenhouse forcing in Europe., Journal of Geophysical Research Atmospheres, 115, 1–12, doi:10.1029/2010JD014307, 2010. Xiao, C.-W., Yuste, J. C., Janssens, I. A., Roskams, P., Nachtergale, L., Carrara, A., Sanchez, B. Y., and Ceulemans, R.: Above- and belowground biomass and net primary production in a 73-year-old Scots pine forest., Tree Physiology, 23, 505–516, doi:10.1093/treephys/23.8.505, 2003. Yang, Y., Liu, Q., Wang, G. X., Wang, X. D., and Guo, J. Y.: Germination, osmotic adjustment, and antioxidant enzyme activities of gibberellin-pretreated Picea asperata seeds under water stress., New Forests, 39, 231–243, doi:10.1007/s11056-009-9167-2, 2010. Zaitchik, B. F., Macalady, A. K., Bonneau, L. R., and Smith, R. B.: Europe’s 2003 heat wave: a satellite view of impacts and land - atmosphere feedbacks., International Journal of Climatology, 26, 743–769, doi:10.1002/ joc.1280, 2006. Zampieri, M., D’Andrea, F., Vautard, R., Ciais, P., De Noblet-Ducoudré, N., and Yiou, P.: Hot European summers and the role of soil moisture in the propagation of mediterranean drought., Journal of Climate, 22, 4747–4758, doi:10.1175/2009JCLI2568.1, 2009. Zapater, M., Bréda, N., Bonal, D., Pardonnet, S., and Granier, A.: Differential response to soil drought among co-occurring broad-leaved tree species growing in a 15- to 25-year-old mixed stand., Annals of Forest Science, 70, 31–39, doi:10.1007/s13595-012-0233-0, 2013. Zellnig, G., Perktold, A., and Zechmann, B.: Fine structural quantification of drought-stressed Picea abies (L.) organelles based on 3D reconstructions., Protoplasma, 243, 129–136, doi:10.1007/s00709-009-0058-3, 2010. Zeng, X. M., Wu, Z. H., Xiong, S. Y., Song, S., Zheng, Y. Q., and Liu, H. Q.: Sensitivity of simulated short-range high-temperature weather to land surface schemes by WRF., Science China Earth Sciences, 54, 581–590, doi: 10.1007/s11430-011-4181-6, 2011. Zhang, C., Wang, Y., and Hamilton, K.: Improved Representation of Boundary Layer Clouds over the Southeast Pacific in ARW-WRF Using a Modified Tiedtke Cumulus Parameterization Scheme., Monthly Weather Review, 139, 3489–3513, doi:10.1175/MWR-D-10-05091.1, 2011. Zhang, Y.-K. and Schilling, K. E.: Effects of land cover on water table, soil moisture, evapotranspiration, and groundwater recharge: a field observation and analysis., Journal of Hydrology, 319, 328–338, doi:10.1016/j. jhydrol.2005.06.044, 2006. Zscheischler, J., Mahecha, M. D., Harmeling, S., and Reichstein, M.: Detection and attribution of large spatiotemporal extreme events in Earth observation data., Ecological Informatics, 15, 66–73, doi:10.1016/j.ecoinf.2013. 03.004, 2013. Zufferey, V., Cochard, H., Ameglio, T., Spring, J. L., and Viret, O.: Diurnal cycles of embolism formation and repair in petioles of grapevine (Vitis vinifera cv. Chasselas)., Journal of Experimental Botany, 62, 3885–3894, doi:10.1093/jxb/err081, 2011. Zweifel, R., Steppe, K., and Sterck, F. J.: Stomatal regulation by microclimate and tree water relations: interpreting ecophysiological field data with a hydraulic plant model., Journal of Experimental Botany, 58, 2113–2131, doi:10.1093/jxb/erm050, 2007. 152 BIBLIOGRAPHY Zweifel, R., Rigling, A., and Dobbertin, M.: Species-specific stomatal response of trees to drought - a link to vegetation dynamics?, Journal of Vegetation Science, 20, 442–454, doi:10.1111/j.1654-1103.2009.05701.x, 2009. Annexes A PPENDIX A The European climate under a 2dC global warming 155 Home Search Collections Journals About Contact us My IOPscience The European climate under a 2 °C global warming This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 Environ. Res. Lett. 9 034006 (http://iopscience.iop.org/1748-9326/9/3/034006) View the table of contents for this issue, or go to the journal homepage for more Download details: IP Address: 132.166.73.161 This content was downloaded on 10/12/2015 at 15:28 Please note that terms and conditions apply. Environmental Research Letters Environ. Res. Lett. 9 (2014) 034006 (11pp) doi:10.1088/1748-9326/9/3/034006 The European climate under a 2 ◦C global warming Robert Vautard1 , Andreas Gobiet2 , Stefan Sobolowski3 , Erik Kjellström4 , Annemiek Stegehuis1 , Paul Watkiss5 , Thomas Mendlik2 , Oskar Landgren6 , Grigory Nikulin4 , Claas Teichmann7,8 and Daniela Jacob7 1 Laboratoire des Sciences du Climat et de l’Environnement (CEA/CNRS/UVSQ), Institut Pierre-Simon Laplace, Orme des Merisiers, Gif sur Yvette, France 2 Wegener Center for Climate and Global Change, University of Graz, Austria 3 Uni Research, Bjerknes Center for Climate Research, Bergen, Norway 4 Swedish Meteorological and Hydrological Institute, Norrköping, Sweden 5 Paul Watkiss Associates, Oxford, UK 6 Norwegian Meteorological Institute, Oslo, Norway 7 Climate Service Center (CSC), Helmholtz-Zentrum Geesthacht, Fischertwiete 1, D-20095 Hamburg, Germany 8 Max Planck Institute for Meteorology (MPIM), Bundesstr. 53, D-20146 Hamburg, Germany E-mail: [email protected] Received 22 December 2013, revised 8 February 2014 Accepted for publication 11 February 2014 Published 6 March 2014 Abstract A global warming of 2 ◦ C relative to pre-industrial climate has been considered as a threshold which society should endeavor to remain below, in order to limit the dangerous effects of anthropogenic climate change. The possible changes in regional climate under this target level of global warming have so far not been investigated in detail. Using an ensemble of 15 regional climate simulations downscaling six transient global climate simulations, we identify the respective time periods corresponding to 2 ◦ C global warming, describe the range of projected changes for the European climate for this level of global warming, and investigate the uncertainty across the multi-model ensemble. Robust changes in mean and extreme temperature, precipitation, winds and surface energy budgets are found based on the ensemble of simulations. The results indicate that most of Europe will experience higher warming than the global average. They also reveal strong distributional patterns across Europe, which will be important in subsequent impact assessments and adaptation responses in different countries and regions. For instance, a North–South (West–East) warming gradient is found for summer (winter) along with a general increase in heavy precipitation and summer extreme temperatures. Tying the ensemble analysis to time periods with a prescribed global temperature change rather than fixed time periods allows for the identification of more robust regional patterns of temperature changes due to removal of some of the uncertainty related to the global models’ climate sensitivity. Keywords: regional climate change, extreme events, European climate 1. Introduction Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. 1748-9326/14/034006+11$33.00 Internationally, there is an ambition to limit global average surface temperature to 2 ◦ C relative to pre-industrial levels. 1 c 2014 IOP Publishing Ltd Environ. Res. Lett. 9 (2014) 034006 R Vautard et al a 2◦ This is in broad alignment with Article 2 of the objectives of the United Nations Framework Convention on Climate Change (UNFCCC 1992), i.e. ‘stabilization of greenhouse gas (GHG) concentrations in the atmosphere at a level that would prevent dangerous anthropogenic interference with the climate system’. The 2 ◦ C goal was initially advocated (WBGU 1995) on the basis of the evidence of the IPCC 2nd Assessment Report (IPCC 1995) with the aim of avoiding serious adverse effects to water resources, ecosystems, biodiversity and human health. More recent IPCC assessments broadly reinforce the goal. The Third Assessment Report (TAR: IPCC 2001) outlined greater negative impacts and more widespread and greater risks with rising temperature (as presented in the reasons for concern ‘burning embers’ diagram), while the Fourth Assessment Report (AR4: IPCC 2007) stated it was ‘very likely that all regions will experience either declines in net benefits or increases in net costs for increases in temperature greater than about 2–3 ◦ C’. The European Union agreed to the proposed goal (CEU 1996, 2004, CEC 2005, 2007), and at the UNFCCC Conference of the Parties in Cancun (UNFCCC 2010), there was international agreement to ‘establish clear goals and a timely schedule for reducing human-generated GHG emissions over time to keep the global average temperature rise below two degrees’. One element of these review assessments—and the 2 ◦ C goal—is the potential risk of catastrophic events (global or regional discontinuities), known as tipping points or tipping elements (Lenton et al 2008). While information on the likelihood of such events remain subjective, and the critical threshold temperatures that might trigger them are highly uncertain, previous studies (Smith et al 2008, Kriegler et al 2009) indicate potential concerns of shifting too far away from the present climate, and especially for moving above 2 ◦ C of warming. However, even the achievement of the 2 ◦ C goal will be accompanied by a significantly changed climate from today, and will necessitate adaptation. In order to have a comprehensive picture of the consequences of a 2 ◦ C warmer climate for Europe, climate projections with a higher spatial resolution than global climate projections (such as provided by the World Climate Research Program ‘Climate Model Intercomparison Project’ CMIP3 (Meehl et al 2007) and CMIP5 (Taylor et al 2012)) are needed, with a rigorous assessment of uncertainties. These goals can be achieved by ensembles of climate projections using regional, limited-area models to downscale global climate projections. Such ensembles have been produced in recent studies dedicated to Europe as in the EU FP5 project PRUDENCE (Christensen et al 2007) and EU FP6 project ENSEMBLES (van der Linden and Mitchell 2009). However, none of these downscaled studies specifically investigated the climate associated with 2 ◦ C warming. Instead, they investigated climate change and its uncertainty in fixed future timed periods. Also, most of the socio-economic scenarios used for these projections (SRES A1B, Nakićenović et al 2000) were not designed to reach a stabilized 2 ◦ C warming and therefore, reach a warmer climate over the century. While a small number of simulations have investigated stabilization scenario (e.g. the ENSEMBLES E1 scenario, van der Linden and Mitchell 2009, Jacob and Podzun 2010), their small number does not allow for robust uncertainty estimation. New simulations carried out in the framework of the CMIP5 and EURO-CORDEX (Jacob et al 2013) have used a scenario that drives to a likely warming lower than 2 ◦ C (RCP2.6), but at the time of writing the number of simulations using this scenario also remains too small to study uncertainty. The identification of changes corresponding to the 2 ◦ C global warming thus requires using scenarios overpassing this target value with a snapshot approach. Here we use the ENSEMBLES regional simulations of the A1B scenario (Nakićenović et al 2000), which are now well studied (Kjellström et al 2013, Déqué et al 2012). The GCMs driving these regional simulations have different sensitivities to natural and anthropogenic climate forcing and reach the target warming at different times. Our method is to collect changes in climate parameters associated with these different times and a reference period for each simulation and gather them in a ‘2 ◦ C ensemble’. This ensemble thus includes uncertainties in the simulation of regional processes simulations and their responses to the global warming and reduces some of the uncertainty due to driving GCM sensitivity. There are limitations to this approach as it does not account for the contributions to uncertainty from systems that have response times longer than the 2 ◦ C time period. This letter reports the most likely changes and their uncertainties calculated from this ensemble. It also aims to estimate the part of the uncertainty that is removed due to considering a period defined by a fixed global warming target instead of fixed time target. We focus on main variables such as temperature, precipitation, sea level pressure and winds, changes in their average and extremes, and on less classical but more explanatory variables such as surface fluxes. By doing so, we provide a unique assessment of what the 2 ◦ C goal might mean for Europe’s climate, overall and across regions, and how this compares to the global average. 2. GCM and RCM simulations used In the subsequent analysis, we analyze 15 out of 22 RCMs from ENSEMBLES with a horizontal resolution of about 25 km. The 22 RCMs are driven by 6 different A1B GCMs, however, not each of the regionalized climate simulations has a sufficiently long time series to reach +2 ◦ C warming. This leaves 15 RCM simulations driven by 6 different GCMs (table 1). Two of the GCMs (bccr bcm2 0-r1, mpi echam5-r3) are realizations from the CMIP3 multi-model dataset (Meehl et al 2007), three models (HadCM3Q0, HadCM3Q3, HadCM3Q16) stem from the Hadley Centre perturbed physics GCM ensemble ‘QUMP’ (Quantifying Uncertainty in Model Predictions) (Collins et al 2011). One GCM (ARPEGE; Salas y Mélia et al 2005) is a spectral model with a stretched grid (Fox-Rabinovitz et al 2008). In order to account for RCM and GCM errors, we used a model output statistic (MOS) approach (Maraun et al 2010), namely quantile mapping (QM) as described by Themeßl et al (2011), based on Déqué (2007). The observational reference was the E-OBS version 5 dataset on a regular 25 km × 25 km 2 Environ. Res. Lett. 9 (2014) 034006 R Vautard et al Table 1. Time period for which +2 ◦ C and +1.5 ◦ C compared to pre-industrial times was reached in ENSEMBLES A1B global climate projections. GCM RCM +2 ◦ C central year +2 ◦ C period +1.5 ◦ C central year +1.5 ◦ C period bccr bcm2 0-r1 HadCM3Q0 HadCM3Q16 HadCM3Q3 mpi echam5-r3 ARPEGE RCA HIRHAM RRCMCLM HadRM RCA HadRM RCA HadRM RegCM REMO HIRHAM RACMO RCA ALADIN HIRHAM 2052 2035 2028 2047 2048 2043 2038–2067 2021–2050 2014–2043 2033–2062 2034–2063 2029–2058 grid (Haylock et al 2008) in the period 1965–2010. We used bias corrected data whenever available, i.e. for daily mean, minimum, and maximum temperature and daily precipitation sum. Themeßl et al (2012) demonstrated the successful application of QM to future scenarios of daily precipitation and Wilcke et al (2013) for other meteorological variables. This implementation is very stable and flexible and has been demonstrated to have higher skill in systematically reducing RCM biases than parametric methods (Gudmundsson et al 2012). In order to avoid the suppression of new extremes in the future periods (i.e. values outside the calibration range), our implementation uses the correction terms of minimum and maximum values of the calibration range outside of the calibration range. Although this simple heuristic extrapolation can probably be improved by using methods of the extreme value theory, it proved to be stable and to lead to better results than the uncorrected model output (Themeßl et al 2012). However it has to be kept in mind that while quantile mapping is very successful in removing biases and adjusting distributions, it cannot substantially improve the temporal structure of time series from RCMs (e.g., Maraun 2012 and Wilcke et al 2013) or properly correct biases in atmospheric circulation (Eden et al 2012). Further, several studies show that bias correction can moderately modify the climate change signal of a simulation (Christensen et al 2008, Themeßl et al 2012, Boberg and Christensen 2012, Dosio et al 2012). However, we conducted a parallel analysis based on raw RCM output (not shown), which led to similar qualitative conclusions as the presented study for mean changes. 2039 2022 2016 2028 2035 2028 2025–2053 2008–2037 2002–2031 2014–2043 2021–2050 2014–2043 temperature rise in a 30-year running mean from 1881–1910 to 1971–2000 (figure 1). The three datasets show an average past warming from the pre-industrial period until the base period of 0.46 K (GISS LOTI: 0.437 K, HadCRUT3: 0.475 K, NOAA NCDC: 0.477 K). Thirty-year running means, starting from the base period 1971–2000, are calculated for the 6 GCMs used. The +2 ◦ C period is determined by the year when the 30-year running mean crosses the +2 ◦ C threshold. The projected +2 ◦ C periods show considerable spread, reaching from 2014– 2043 (HadCM3Q16) to 2038–2067 (BCM) (figure 1) with corresponding central years at 2028 and 2052, respectively. The subset of 6 GCMs used in this analysis (table 1) still spans the same range for global temperature, so no considerable information should be lost in this respect compared to the full set of ENSEMBLES GCMs, even though possibly reducing the range for regional variables. Also, compared to the entire CMIP3 A1B ensemble with 53 simulations, the ENSEMBLES GCMs miss some lower sensitivity simulations. The CMIP3 simulations project +2 ◦ C warming from 2029 up to 2075, with a median of 2049, whereas the selected ENSEMBLES GCMs reach +2 ◦ C around 2045. Figure 1 (bottom panels) shows for comparison global warming in the new CMIP5 simulations for the representative emission pathways RCP2.6, RCP4.5 and RCP8.5. While most of the RCP2.6 simulations don’t reach +2 ◦ C at all, the RCP4.5 simulations reach it around 2050 and the RCP8.5 simulations around 2042 (median). This means, except for RCP2.6, all shown emission scenarios most likely to lead to +2 ◦ C warming in a relatively narrow time window between 2042 and 2050, while much stronger differences between the scenarios can be expected in the second half of the 21st century. 3. When is global climate likely to reach a 2 ◦ C warming? 4. Robustness assessment In this study the +2 ◦ C period is defined as the time when the 30-year average global mean temperature reaches +2 ◦ C, compared to a ‘pre-industrial’ period 1881–1910. To define the +2 ◦ C period, we analyzed past observed and future projected temperatures. The following global observational datasets have been analyzed for this purpose: GISS LOTI (1880–2011) (http://data.giss.nasa.gov/gistemp/), HadCRUT3 (1850–2011) (www.cru.uea.ac.uk/cru/data/temperature/), and NOAA NCDC (1880–2011) (www.ncdc.noaa.gov/cmb-faq/a nomalies.php). The time period common to all datasets matching best the pre-industrial period is 1881–1910. Thus, we consider past pre-industrial warming until the base period as the Recently a number of approaches to assess and communicate robustness of projected climate change have been proposed (Tebaldi et al 2011), distinguishing model agreement in sign and some indication of statistical significance of individual models changes. They also generally make some attempt to show areas where the climate change signal is low relative to internal variability, but may still contain useful information for policy makers (e.g. a projected change is small and not statistically significant but the models agree on the sign). Here we simply define robustness based on agreement between models. In order for the ensemble change to be considered robust 12 of the 15 models at least must agree on the sign 3 Environ. Res. Lett. 9 (2014) 034006 R Vautard et al Figure 1. Global mean temperature (30-year running mean; gray lines) for the SRES A1B ensemble (top panel) and for the RCP2.6, RCP4.5 and RCP8.5 CMIP5 simulations (bottom panels) exceeding the +2 ◦ C threshold (bold red horizontal line). The average observed temperature compared to pre-industrial (1881–1910) is depicted in the upper panel as black line. The CMIP3 and CMIP5 ensemble median years of reaching the 2 ◦ C target for each emission scenario are shown as black vertical lines, whereas the red vertical line represents the median year of the six driving GCMs of this study, which are highlighted in red. Since most RCP2.6 simulations stabilize below +2 ◦ C, no median exceedance year is shown. (threshold based on the 95% confidence interval of a binomial test with 50% chance of success). In subsequent figures, areas where such an agreement is not obtained are filled by gray color. A more in-depth investigation of uncertainty will be carried out in a future study using a broader ensemble of EURO-CORDEX simulations and multiple socio-economic scenarios. measured as the standard deviation between individual model changes, ranges between about 0 and 1 ◦ C depending on season (not shown). The removal of the uncertainty related to transient climate response to radiative forcing in the GCMs can be readily seen when comparing the spread of simulated temperature changes at +2 ◦ C to those from a fixed time period of 2031–2060, which has a roughly equivalent mean temperature change (figure 2(c)). Figure 3 (left panels) shows the mean temperature changes simulated by the RCMs between the control period and the +2 ◦ C period, in winter and summer separately. This provides information on the pattern of warming across Europe. Average changes have similar patterns as those described for fixed time periods in several regional ensemble studies (Fischer and Schär 2010, Kjellström et al 2011): a temperature increase is found everywhere for all models, with enhancement in North-Eastern and Eastern parts of Europe in winter (2–3 ◦ C) and in Southern Europe in summer (2–3 ◦ C). The regional warming exceeds the global warming in most areas except the British Isles and Iceland, where the influence of the moderate warming of the North Atlantic is seen in all seasons. In summer, a relatively small warming is also seen close to the North Sea and the 5. Changes in Europe at 2 ◦ C average global warming Under a 2 ◦ C global warming (+0.46 ◦ C from pre-industrial to 1971–2000 and +1.54 ◦ C from 1971–2000 to the 2 ◦ C period), ensemble-averaged projected European regional warming generally ranges between 1.5 and 2.0 ◦ C depending on the region. European temperatures therefore mainly exceed the global warming after 1971–2000 (figure 2(a)). Only North-Western Europe (British Isles and France in particular) witness a lower relative increase in warming, i.e. below 1.54 ◦ C. The warming is higher in winter (mean = 1.99 ◦ C, 25% = 1.65 ◦ C, 75% = 2.34) than in summer (mean = 1.72 ◦ C, 25% = 1.45 ◦ C, 75% = 1.99) (figure 2(c)). The spread of simulated changes, 4 Environ. Res. Lett. 9 (2014) 034006 R Vautard et al Figure 2. (a) Yearly averaged change—relative to the reference period 1971–2000—in yearly mean temperature in the different European regions for periods corresponding to +2 ◦ C of global average change. The global temperature change (1.54 ◦ C) between 1971–2000 and the 2 ◦ C period is marked as a dotted line. (b) Same as (a) for precipitation in % of change. The solid red line indicates no change for precipitation and the red dotted line a change of 1.54 ◦ C for temperature, corresponding to a global warming of 2 ◦ C relative to pre-industrial. (c) Spatial average over land of changes in temperature and (d) precipitation, together with the range of changes for the GCM–RCM ensemble (median, 25–75% range and min and max values). The open bars refer to fixed time future period (2031–2060), the gray bars to the temperature controlled (+2 ◦ C) period. Baltic Sea. All areas undergo robust warming (robustness not shown for temperature because it covers all areas). The ensemble standard deviation of the changes remains much smaller than the amplitude of changes (about 3–10 times smaller) everywhere (not shown). These spatial differences are important with respect to subsequent impacts. Higher summer temperature changes are found in the Iberian Peninsula and the Mediterranean region, and will thus compound existing temperature-related impacts such as energy use for cooling (EEA 2012). However, the higher winter warming in Northern Europe will have a mix of positive as well as negative effects, including reduced winter heating. This reveals important distributional consequences 5 Environ. Res. Lett. 9 (2014) 034006 R Vautard et al Figure 3. Seasonal mean changes of temperature (left panels), precipitation (middle panels) and sea level pressure (right panels), between the 1971–2000 and the +2 ◦ C periods. Top panels show wintertime changes and bottom panels show summertime changes. Only areas where at least 12 models agree on the sign of the change are colored and areas where at least 14 models agree on the sign of the change are dotted areas. For temperature the agreement on the sign of the change was found everywhere and is not shown. in the Northeast, but the signal is not very robust. This is however consistent with the temperature and precipitation changes and suggests expansion of the subtropical dry zone into Southern Europe and an enhanced hydrological cycle in Northern Europe and Scandinavia. The summer signal is more robust with most of the models agreeing on modest decreases in SLP across Southern Europe. This modest response could be indicative of localized thermal low pressure due to heating. This is a common feature in the Iberian Peninsula under present conditions but any future increase of said phenomenon requires further elucidation than the present study allows. The summer pattern also indicates an increase in the North–South pressure gradient over the Northern part of the North Atlantic in the domain as pressure increases over the British Isles and decreases over Iceland. Such an increase may help to explain the increase in precipitation in parts of Scandinavia as partly being a consequence of enhanced moisture transport from the North Atlantic. across Europe, in terms of the patterns of likely impacts, even under the 2 ◦ C goal. A similar analysis has been undertaken for precipitation. When averaged over the PRUDENCE regions, annual average precipitation robustly decreases by up to about 10% in Southern sub-regions while it may increase with more than 10% in Northern Europe (figure 2(b)). When averaged over European land areas for each season, mean precipitation significantly increases in autumn (mean = 3.3%, 25% = 2.1%, 75% = 4.8%) and winter (mean = 5.3%, 25% = 5.1%, 75% = 6.5%) but changes in spring (mean = 1.7%, 25% = 0.2%, 75% = 3.1%) and summer (mean = −0.5%, 25% = −2.3%, 75% = 1.1%) do not show a clear sign (figures 2(b)–(d)). In winter, a general increase is found with maximum values in Northern Europe, especially along many coastal areas, where all models agree upon an increase of 10–15% (figure 3 middle panels). In Southern Europe the models do not agree on sign except over a few areas (Southern Italy, Greece). By contrast, in summer, the models agree on a robust decrease of precipitation in South-Central Europe of about 10–15%, together with an increase in precipitation over Scandinavia. These changes may exacerbate existing water management issues across Europe, i.e. potentially increasing water deficits in the South during the already heat and evaporation stressed summer. The only area where all models agree on the sign of change is Scandinavia (increase in both seasons) and some smaller areas in South-Eastern Europe and the West coasts of the Iberian Peninsula, France and the Southern British Isles (decrease in summer). Some climate change signal is found in sea level pressure (figure 3, right panels) in the winter season with lower pressure 6. Extremes Extreme temperature is here defined as the daily maximum temperature that is exceeded on average only once every 20 years. This is termed the return value and 20 years the return period. Changes in extreme temperature are thus, characterized through changes in the 20-year return values. Return values are estimated using the block maxima method for which the generalized extreme value (GEV) distribution describes the behavior of said maxima (Coles 2001). The approach closely follows that of Nikulin et al (2011) and further details are provided therein. 6 Environ. Res. Lett. 9 (2014) 034006 R Vautard et al Figure 4. Changes, between the 1971–2000 and the +2 ◦ C periods, in the 20-year return value for Tmin in winter (upper left), Tmax in summer (lower left), heavy precipitation in winter (top middle) and in summer (bottom middle), extreme winds (I99) in winter (top right) and in summer (bottom right). Only areas with at least 12 models agreeing on change sign are colored. Areas where at least 14 models agree on change sign are highlighted with dots (except for temperature where almost all areas satisfy this). For summer daily maximum temperatures (figure 4), the largest changes (3–4 ◦ C) are found over South-Eastern Europe and the Iberian Peninsula. In areas where this value is highest under today’s conditions (Iberian Peninsula, France, the Balkans) the 20-year return value is expected to rise well above 40 ◦ C. As increases in summer extreme heat are linked to health impacts in the form of temperature-related mortality (Baccini et al 2008) the pattern of changes projected under 2 ◦ C is likely to have important health impacts in the more vulnerable regions of Europe. Conversely, the extremes of daily minimum temperatures are reduced most notably in Northern and Eastern areas of Europe, which will have benefits in reducing current winter cold extremes and cold-related mortality as well as winter heating costs (EEA 2012), though there would also be negative impacts, such as on winter tourism and ecosystems. All temperature changes are found to be robust but the spread between models is high in Central and North-Eastern Europe. In parts of this area, notably in Northern Sweden and Finland, there are even models indicating no increase in extreme maximum temperatures. Discrepancies in this area may to some extent be related to how different RCMs treat lakes as parts of this area has a large fraction of lakes that have an impact on the regional climate (Samuelsson et al 2010). Extremes of wintertime daily minimum temperature (figure 4) undergo a large positive change in winter, ranging from 2–3 ◦ C in Central and Southern Europe to 5–8 ◦ C in Scandinavia and Russia. These changes are robust but again a large spread in model responses is found over Central Europe, where some models do not even agree on the positive sign of the change in extreme temperatures. Changes in extremes of heavy precipitation defined as the 20-year return value calculated from extreme value theory in the same way as outlined for temperature above are shown in figure 4. The ensemble mean exhibits positive changes in almost all areas both in summer and winter, with amplitude ranging from 5% to about 15%. The increase is marked over Eastern Europe and Scandinavia in summer and over Southern Europe in winter. In contrast to this study, where no overall trend in extreme summer precipitation in Southern Europe is found, Kendon et al (2008) and Maraun (2013) found a decrease for extreme summer precipitation in this region. The difference in the findings might be related to the fact that the negative trend emerges relatively late and thus might not have emerged for some ensemble members when reaching the +2 ◦ C period. It may also result from the way extremes are defined and the different sets of selected simulations, indicating a lack of robustness of a possible decrease. The increases in heavy precipitation are an important factor with respect to flood risks, thus the increases in heavy precipitation found under the +2 ◦ C scenario are likely to enhance the potential for these events. Floods are among the most important weather-related loss events in Europe and can have large economic consequences: the EEA (2010) reports total losses of over e50 billion over the past decade. The projected increase in Eastern Europe is a particular concern because this is one of the existing flood hot spots. Uncertainties remain large in the southernmost areas of Europe. Compared to the changes in seasonal means the changes in extremes are less spatially coherent and individual models exhibit patchy structures. 7 Environ. Res. Lett. 9 (2014) 034006 R Vautard et al Figure 5. Changes, between the 1971–2000 and the +2 ◦ C periods, in surface energy fluxes. Left panels: latent heat fluxes; second left panels: sensible heat fluxes; second right panels: net short-wave fluxes; right panels: net long-wave fluxes. Upper panels: MAM, Lower panels: JJA. Only areas with at least 12 models agreeing on change sign are colored. Areas where at least 14 models agree on change sign are highlighted with dots (except for temperature where almost all areas satisfy this). Windstorms are also among the most damaging extreme events in Europe (ABI 2005). Extreme winds are calculated here as the 99th percentile of the daily maximum 10-meter wind speed for each season (I99). Figure 4 (right panels) show the ensemble mean relative changes in I99 (in %) winds as simulated by the RCMs, in winter and summer. Robust increases of extreme winds of up to 10% are only seen over small scattered areas of Central/Eastern Europe in winter. Over most regions the change is generally positive but not robust. of Central Europe, have decreasing latent heat due to drought increase and soil-moisture limitations. This drying causes an increase in sensible heat in summer in this area, with large model agreement, while over Scandinavia sensible heat fluxes decrease due to a wetter and cloudier climate. In spring, sensible heat flux only increases with high model agreement in the southernmost areas. However, it is suspected that these fluxes are biased on ensemble average in the ENSEMBLES simulations. Stegehuis et al (2013) showed that several models in the ensemble have large evapotranspiration in the spring, which induces a dry, soil-moisture limited regime in summer, with large sensible heat flux accompanied by a low latent heat flux. This effect can lead to an overestimate of the ensemble mean changes and an underestimate of the inter-annual variability changes (Fischer et al 2012). This could also contribute to the nonlinear temperature bias found in this dataset by Boberg and Christensen (2012), which probably induces an overestimation of mean summertime warming in Southern Europe. Heat flux changes respond to the changes in net radiative fluxes (figure 5). In the two seasons a robust increase in short-wave radiation is found in Southern Europe and robust increase in long-wave radiation is found in Northern Europe. In summer, the short-wave radiation increase extends Northward over central Europe where drying, increase of sensible heat and temperature also occurs. This extension is consistent with the Northward propagation of drought and heat in the spring–summer transition described in Zampieri et al (2009). It is noteworthy that sea level pressure, however, does not exhibit an associated robust increase in anticyclonic weather over Mediterranean areas in summer, which could be due to compensation due to heating-induced thermal surface pressure lows. In Northern Europe, wetter and cloudier weather induces an increase in long-wave radiation. 7. Surface energy budget Surface weather changes are influenced both by larger scale dynamical changes and by local changes in the physics of the vertical column above the surface. These latter changes are due to clouds, atmospheric composition, turbulence, soil moisture and temperature as well as land use properties. All these processes influence the surface energy budget (SEB), characterized by radiative and heat fluxes. Changes in the SEB are particularly important for driving changes in summer temperature and precipitation, which involve amplifying feedback processes (Seneviratne et al 2010, Fischer et al 2007). They also have impacts on drought changes, river discharge, water and energy resources, with important economic implications. We consider net short- and long-wave fluxes, sensible and latent upward fluxes, only in spring (MAM) and summer (JJA) in order to analyze the evolution of heat fluxes across the growing season. Spatial patterns of latent and sensible heat flux changes are shown in figures 5(a)–(d). In spring, almost all models agree on an increase of latent heat except in southernmost areas and in other smaller areas such as North-Western coastal areas (this is also the case in fall and winter). This distribution is consistent with the energy-limited nature of evapotranspiration in Europe in spring (Teuling et al 2009). In summer, the latent heat increase is restricted to Northern areas while Southern areas, including large parts 8 Environ. Res. Lett. 9 (2014) 034006 R Vautard et al 8. Conclusions Many of the changes—in terms of the sign and magnitude of the change, and perhaps more importantly the spatial location and distributional pattern across Europe—will act to exacerbate existing and future impacts. For example, there is higher relative warming and greater relative increase in heat extremes in Southern Europe in summertime, which will drive heat and temperature-related impacts such as cooling costs and heat-related mortality. Similarly, there are higher relative (and more robust) signals for increased precipitation and heavy precipitation events in Eastern Europe along existing flood risk corridors, but lower projected summer rainfall in the Mediterranean, which will increase pressure on water resource management. There are some exceptions (e.g. higher winter warming in the north, with the benefit of reduced winter mortality and winter heating demand, though there would also be negative impacts on winter tourism and ecosystems in these regions). However, the general findings are that the distributional patterns of change across Europe are likely to drive geographically specific negative impacts. This is of policy relevance: even if the 2 ◦ C goal is achieved, Europe will experience impacts, and these are likely to exacerbate existing climate vulnerability. Further work on identifying key hotspots, potential impacts and advancing carefully planned adaptation is therefore needed. While it does not qualitatively affect the robust patterns of climate changes for temperature and precipitation, the bias correction may in some areas slightly modify the amplitude of temperature changes (e.g. Dosio et al 2012). This was deduced from additional experiments (not reported here). Precipitation changes appear less sensitive to bias correction. We have identified changes in European regional climate associated with a +2 ◦ C global warming relative to pre-industrial climate. The +2 ◦ C period was characterized using 30-year periods of an ensemble of global climate simulations of the SRES Scenario A1B, downscaled at a 25 km resolution by an ensemble of regional climate models (RCMs), simulations carried out in the framework of the FP6 ENSEMBLES project. The robustness of these changes has been quantified by measuring model agreement on the sign of the change and by assessing the statistical significance of the change. The main characteristics of the changes in Europe expected for a +2 ◦ C global warming relative to a reference period of 1971–2000 are: (1) Europe generally experiences higher warming than the global average, i.e. it will experience more than 2 ◦ C of warming even if the 2 ◦ C goal is achieved. There is also a strong distributional pattern of warming across Europe (and thus different countries). A warming over all European regions is found, with slightly weaker amplitude than the global warming over North-Western Europe but a more intense warming (up to +3 ◦ C) in Northern and Eastern Europe in Winter and in Southern Europe in Summer. (2) A robust increase of precipitation over Central and Northern Europe in winter and only over Northern Europe in summer, while precipitation decreases in Central/Southern Europe in summer, with changes reaching 20%. (3) A marked trend with an increased amplitude of up to more than 4 ◦ C in the 20-year return value of the summer daily maximum and an even larger warming (up to more than 6 ◦ C) over Scandinavia for extreme cold daily minima in winter. (4) A robust increase in heavy precipitation everywhere and in all seasons, except Southern Europe in summer, with amplitudes in the range 0–20%. (5) A modest and marginally robust increase in extreme winds in parts of Central Europe in winter, while in summer wind extremes changes are not robust. (6) Sensible and latent heat flux changes have a strong seasonality with increasing (almost everywhere) evapotranspiration in spring, while it decreases in Southern/Central Europe; sensible heat fluxes exhibit an opposite pattern with an even higher amplitude in Southern/Central Europe; In summer fewer clouds in this area allow more intense net radiation input in this area. However models may overestimate evapotranspiration in spring, leading to an exaggerated drying and sensible heat flux increase in summer. (7) The analysis also led us to conclude that a +2 ◦ C change is, on average, approximately equivalent to a change for the 2031–2060 period in the A1B scenario. Choosing the time period reflecting a global mean change of +2 ◦ C, however, reduces the spread in the results for temperature. A similar reduction in spread is not seen for other variables. Acknowledgments The research leading to these results has received funding from the European Union Seventh Framework Programme (FP7/2007–2013) under the project: IMPACT2C: Quantifying projected impacts under 2 ◦ C warming, grant agreement no. 282746. The ENSEMBLES data used in this work was funded by the EU FP6 Integrated Project ENSEMBLES (Contract number 505539) whose support is gratefully acknowledged. References Association of British Insurers 2005 Financial Risks of Climate Change (London: Association of British Insurers) www.abi.org.uk/climatechange Baccini M et al 2008 Heat effects on mortality in 15 European cities Epidemiology 19 711–9 Boberg F, Berg P, Thejll P, Gutowski W J and Christensen J H 2010 Improved confidence in climate change projections of precipitation further evaluated using daily statistics from ENSEMBLES models Clim. Dyn. 35 1509–20 Boberg F and Christensen J H 2012 Overestimation of Mediterranean summer temperature projections due to model deficiencies Nature Clim. Change 2 433–6 CEC 2005 Winning the Battle Against Global Climate Change, Brussels, 9.2.2005. COM(2005)35 Final (Brussels: Commission of the European Communities) 9 Environ. Res. Lett. 9 (2014) 034006 R Vautard et al IPCC (1995) IPCC Second Assessment: Climate Change 1995. A Report of the Intergovernmental Panel on Climate Change. Full Report Published by the IPCCC www.ipcc.ch/pdf/climate-chang es-1995/ipcc-2nd-assessment/2nd-assessment-en.pdf (Accessed December 2013) IPCC 2001 Climate Change 2001: Synthesis Report. A Contribution of Working Groups I, II, and III to the Third Assessment Report of the Integovernmental Panel on Climate Change ed R T Watson and the Core Writing Team (Cambridge: Cambridge University Press) 398p IPCC 2007 Summary for policymakers Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change ed M L Parry, O F Canziani, J P Palutikof, P J van der Linden and C E Hanson (Cambridge: Cambridge University Press) pp 7–22 Jacob D and Podzun R 2010 Global warming below 2 ◦ C relative to pre-industrial level: how might climate look like in Europe Nova Acta Leopoldina, NF 71–6 Jacob D et al 2013 EURO-CORDEX: new high-resolution climate change projections for European impact research Reg. Environ. Change 1–16 doi:10.1007/s10113-013-0499-2 Kendon E J, Rowell D P, Jones R G and Buonomo E 2008 Robustness of future changes in local precipitation extremes J. Clim. 21 4280–97 Kjellström E, Nikulin G, Hansson U, Strandberg G and Ullerstig A 2011 21st century changes in the European climate: uncertainties derived from an ensemble of regional climate model simulations Tellus 63 24–40 Kjellström E, Thejll P, Rummukainen M, Christensen J H, Boberg F, Christensen O and Band Fox Maule C 2013 Emerging regional climate change signals for Europe under varying large-scale circulation conditions Clim. Res. 56 103–19 Kriegler E, Hall J W, Held H, Dawson R and Schellnhuber H J 2009 Imprecise probability assessment of tipping points in the climate system Proc. Natl Acad. Sci. USA 106 5041–6 Lenton T M, Held H, Kriegler E, Hall J W, Lucht W, Rahmstorf S and Schellnhuber H J 2008 Tipping elements in the Earth’s climate system Proc. Natl Acad. Sci. USA 105 1786–93 Levermann A et al 2012 Potential climatic transitions with profound impact on Europe: Review of the current state of six ‘tipping elements of the climate system’ Clim. Change 110 845–78 Maraun D 2012 Nonstationarities of regional climate model biases in European seasonal mean temperature and precipitation sums Geophys. Res. Lett. 39 L06706 Maraun D 2013 When will trends in European mean and heavy daily precipitation emerge? Environ. Res. Lett. 8 014004 Maraun D et al 2010 Precipitation downscaling under climate change: recent developments to bridge the gap between dynamical models and the end user Rev. Geophys. 48 RG3003 Meehl G A, Covey C, Delworth T, Latif M, McAvaney B, Mitchell J F B, Stouffer R J and Taylor K E 2007 The WCRP CMIP3 multimodel dataset: a new era in climate change research Bull. Am. Meteorol. Soc. 88 1383–94 Nakićenović N et al 2000 Emission scenarios, A Special Report of Working Group III of the Intergovernmental Panel on Climate Change (New York: Cambridge University Press) p 599 Nikulin G, Kjellstrom E, Hansson U, Strandberg G and Ullerstig A 2011 Evaluation and future projections of temperature, precipitation and wind extremes over Europe in an ensemble of regional climate simulations Tellus A 63 14–55 Salas y Mélia D, Chauvin F, Déqué M, Douville H, Guérémy J F, Marquet P, Planton S, Royer J F and Tyteca S 2005 Description and validation of CNRM-CM3 global coupled codel CNRM CEC 2007 Limiting Global Climate Change to 2 Degrees Celsius The Way ahead for 2020 and beyond COM(2007)2 (Brussels: Commission of the European Communities) CEU 1996 Council of the European Union 1939th Council Meeting (Luxembourg, June 1996) CEU 2004 Spring Council Meeting of the European Union, 2004 Christensen J H, Boberg F, Christensen O B and Lucas-Picher P 2008 On the need for bias correction of regional climate change projections of temperature and precipitation Geophys. Res. Lett. 35 L20709 Christensen J H, Carter T R, Rummukainen M and Amanatidis G 2007 Evaluating the performance and utility of regional climate models: the PRUDENCE project Clim. Change 81 1–6 Coles S 2001 An Introduction to Statistical Modeling of Extreme Values (London: Springer) Collins M, Booth B B, Bhaskaran B, Harris G R, Murphy J M, Sexton D M H and Webb M J 2011 Climate model errors, feedbacks and forcings: a comparison of perturbed physics and multi-model ensembles Clim. Dyn. 36 1737–66 Déqué M 2007 Frequency of precipitation and temperature extremes over France in an anthropogenic scenario: model results and statistical correction according to observed values Glob. Planet. Change 57 16–26 Déqué M, Somot S, Sanchez-Gomez E, Goodess C M, Jacob D, Lenderink G and Christensen O B 2012 The spread amongst ENSEMBLES regional scenarios: regional climate models, driving general circulation models and interannual variability Clim. Dyn. 38 951–64 Dosio A, Paruolo P and Rojas R 2012 Bias correction of the ENSEMBLES high resolution climate change projections for use by impact models: Analysis of the climate change signal J. Geophys. Res.-Atmos. 117 D17110 Eden J M, Widmann M, Grawe D and Rast S 2012 Skill, correction, and downscaling of GCM-simulated precipitation J. Clim. 25 3970–84 EEA 2010 Mapping the impacts of natural hazards and technological accidents in Europe. An overview of the last decade EEA Technical Report No 13/2010. ISSN 1725-2237 EEA 2012 Climate change, impacts and vulnerability in Europe 2012 An indicator-based report (Copenhagen: EEA), 2012. www.eea.europa.eu/publications/climate-impacts-and-vulnerabili ty-2012 (Accessed December 2013) Fischer E M, Rajczak J and Schär C 2012 Changes in European summer temperature variability revisited Geophys. Res. Lett. 39 L19702 Fischer E M and Schär C 2010 Consistent geographical patterns of change in high-impact European heat waves Nature Clim. Change 3 398–403 Fischer E M, Seneviratne S I, Luthi D and Schär C 2007 Contribution of land-atmosphere coupling to recent European summer heat waves Geophys. Res. Lett. 34 L06707 Fox-Rabinovitz M, Cote J, Dugas B, Deque M, McGregor J L and Belochitski A 2008 Stretched-grid model intercomparison project: decadal regional climate simulations with enhanced variable and uniform-resolution GCMs Meteorol. Atmos. Phys. 100 159–77 Gudmundsson L, Bremnes J B, Haugen J E and Engen-Skaugen T 2012 Technical note: downscaling RCM precipitation to the station scale using statistical transformations—a comparison of methods Hydrol. Earth Syst. Sci. 16 3383–90 Haylock M R, Hofstra N, Klein Tank A M G, Klok E J, Jones P D and New M 2008 A European daily high-resolution gridded data set of surface temperature and precipitation for 1950–2006 J. Geophys. Res. 113 D20119 10 Environ. Res. Lett. 9 (2014) 034006 R Vautard et al Working Note 103 (www.cnrm.meteo.fr/scenario2004/paper cm3.pdf) Samuelsson P, Kourzeneva E and Mironov D 2010 The impact of lakes on the European climate as simulated by a regional climate model Boreal Environ. Res. 15 113–29 Seneviratne S I, Corti T, Davin E L, Hirschi M, Jaeger E B, Lehner I, Orlowsky B and Teuling A J 2010 Investigating soil moisture-climate interactions in a changing climate: a review Earth-Sci. Rev. 99 125–61 Smith J B et al 2008 Assessing dangerous climate change through an update of the Intergovernmental Panel on Climate Change (IPCC) reasons for concern. 2008 Proc. Natl Acad. Sci. USA Early Edition. doi:10.1073/pnas.0812355106 Stegehuis A I, Vautard R, Ciais P, Teuling A J, Jung M and Yiou P 2013 Summer temperatures in Europe and land heat fluxes in observation-based data and regional climate model simulations Clim. Dyn. 41 455–77 Taylor K E, Stouffer R J and Meehl G A 2012 An overview of CMIP5 and the experiment design Bull. Am. Meteorol. Soc. 93 485–98 Tebaldi C, Arblaster J M and Knutti R 2011 Mapping model agreement on future climate projections Geophys. Res. Lett. 38 L23701 Teuling A J et al 2009 A regional perspective on trends in continental evaporation Geophys. Res. Lett. 36 L02404 Themeßl M J, Gobiet A and Heinrich G 2012 Empirical-statistical downscaling and error correction of regional climate models and its impact on the climate change signal Clim. Change 112 449–68 Themeßl M J, Gobiet A and Leuprecht A 2011 Empirical-statistical downscaling and error correction of daily precipitation from regional climate models Int. J. Climatol. 31 1530–44 UNFCCC 1992 United Nations Framework Convention on Climate Change: Full Text of the Convention http://unfccc.int/essential ba ckground/convention/background/items/1353.php (Accessed December 2013) UNFCCC 2010 The Cancun Agreements. United Nations Framework Convention on Climate Change http://unfccc.int/meet ings/cancun nov 2010/meeting/6266.php (Accessed December 2013) van der Linden P and Mitchell J B (ed) 2009 ENSEMBLES: Climate Change and its Impacts: Summary of Research and Results From the ENSEMBLES Project (Exeter: Met Office Hadley Centre) 160pp WBGU 1995 Scenario zur Ableitung CO2 -Reduktionsziele und Umsetzungsstrategien—Stellungnahme zur ersten Vertragsstaatenkonferenz der Klimarahmenkonvention in Berlin (Scenario to Deduct CO2 Reduction Targets und Implementation Strategies–Position for the First Conference of the Parties of the Framework Convention on Climate in Berlin), Wissenschaftlicher Beirat der Bundesregierung Globale Umweltveränderungen, Dortmund Wilcke R A I, Mendlik T and Gobiet A 2013 Multi-variable error correction of regional climate models Clim. Change 120 871–87 Zampieri M, D’Andrea F, Vautard R, Ciais P, de Noblet-Ducoudre N and Yiou P 2009 Hot European summers and the role of soil moisture in the propagation of Mediterranean drought J. Clim. 22 4747–58 11