* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download The role of local atmospheric forcing on the
History of research ships wikipedia , lookup
El Niño–Southern Oscillation wikipedia , lookup
Anoxic event wikipedia , lookup
Marine pollution wikipedia , lookup
Southern Ocean wikipedia , lookup
Marine habitats wikipedia , lookup
Abyssal plain wikipedia , lookup
Future sea level wikipedia , lookup
Pacific Ocean wikipedia , lookup
Arctic Ocean wikipedia , lookup
Indian Ocean Research Group wikipedia , lookup
Global Energy and Water Cycle Experiment wikipedia , lookup
Ocean acidification wikipedia , lookup
Indian Ocean wikipedia , lookup
Effects of global warming on oceans wikipedia , lookup
Critical Depth wikipedia , lookup
Ecosystem of the North Pacific Subtropical Gyre wikipedia , lookup
The role of local atmospheric forcing on the modulation of ocean mixed layer in reanalysis and coupled single column ocean model Byju Pookkandy, Dietmar Dommenget, Nicholas Klingaman, Scott Wales, Christine Chung, Claudia Frauen, Holger Wolff*** Abstract: The role of local atmospheric forcing on ocean mixed layer depth (MLD) over the global oceans are studied using ocean data assimilation products and a single-column ocean model coupled with an atmospheric general circulation model. We focus on how the annual mean and the seasonal cycle of the MLD relate to the different forcing characteristic at different parts of the world oceans. We further analysis how anomalous variations in the monthly mean MLD relate to anomalous atmospheric forcings. In the combination of the analysis of the ocean reanalysis data and the single column ocean model we identify regions with different dominant forcings and different mean and variability characteristic of the MLD. Much of the global oceans MLD characteristics appear to be directly linked to the different atmospheric forcing characteristics at different locations. Here heating and wind input are the main driver and at some, mostly coastal, regions the atmospheric salinity forcing. The annual mean MLD is more closely relate to the annual mean wind stress and the seasonality is more closely related to the seasonality in the heating. The single column ocean model, however, also points out that the MLD characteristics over most global ocean regions, and in particular in the tropics and subtropics, cannot be maintained by local atmospheric forcings only, but are also a result of oceanic dynamics that are not simulated in a single column ocean model. Thus lateral ocean dynamics are essential in maintaining observed MLD. 1" 1. Introduction 2" The processes occurring in the upper ocean are predominantly forced by the atmosphere. 3" Heating and cooling at daily to seasonal scale, wind-stress, evaporation and precipitation are 4" the dominant physical processes that act and interact with the upper layer of the ocean. The 5" vigorous turbulent mixing near the surface by these forces results in a layer of homogeneous 6" temperature and salinity (and thus density) with depth, which is defined as the ocean mixed 7" layer depth (MLD). 8" Understanding the physics of MLD is important for climate dynamics, because the thickness 9" of the mixed layer modulates its heat capacity and hence its ability to store excess heat from 10" the atmosphere (Godfrey and Lindstrom 1989; Maykut and McPhee 1995; Swenson and 11" Hansen 1999; Peter et al. 2006; Montégut et al. 2007; Dong et al. 2007). In addition the 12" variations of the MLD, over the seasonal cycle or in anomalies from it, is one of the main 13" processes in exchanging heat between the atmosphere and the deeper oceans. Proper 14" quantification of the heat budget is important as it governs the evolution of the sea surface 15" temperature (SST) (Chen et al. 1994; Alexander et al. 2000; Dommenget and Latif 2002; Qui 16" et al. 2004; Dong et al. 2008). In a study using a simple stochastic model Dommenget and 17" Latif (2002) speculated that the low frequency (say decadal) variability in SST is an 18" integration of the heat stored in the mixed layer and the sensitivity of SST to MLD peaks at 19" midlatitudes. Therefor the seasonal and interannual variability in SST can be simulated by the 20" local air-sea interactions and proper representation of MLD variability. Mixed layer 21" dynamics is also important for the ocean biological productivity(Fasham 1995; Polovina et al. 22" 1995; Narvekar and Kumar 2006) and acts as a medium for exchange of trace gases between 23" the ocean and atmosphere (Takahashi et al. 1997; Bates 2001; Sabine et al. 2004). 24" A number of studies have examined the mixed layer dynamics and demonstrated the variation 25" in MLD at different spatial and temporal scales (Montery and Levitus, 1997; Kara et al. 26" 2003, 2000; Cronin and Kessler 2002; Montégut et al. 2004). Most of such studies were 27" limited to either a small region or at point locations where in-situ data is available and then 28" extrapolated over a wider domain using some specific techniques. Since the advent of Argo 29" data re-analysis methods has produced much more reliable source of representation of the 30" state of the ocean at global scale. It has been understood that the MLD in summer is 31" dominated by the entrainment due to wind induced mixing and surface buoyancy force in 32" winter (Alexander et al. 2000). These theories are generalised concept on MLD variability in 33" the global ocean, however, the relative role of these forcing vary in time and space. This 34" uneven distribution of MLD forcing on seasonal time scale is not known or understood in 35" detail over the global ocean. Carton et al. (2008) and Lorbacher et al. (2006) examined the 36" variability in MLD at global scale from observation, but it provides little information about 37" the Southern ocean. Numerous attempts have been made to understand the ocean surface 38" boundary layer and its sensitivity to atmospheric forcing (Adamec et al, 1984; Large and 39" Crawford, 1995; Kantha and Clayson, 1994 etc.). Most of the studies of MLD variability 40" have been limited at one location or in small ocean basin. 41" The focus of our study is to examine the mean, seasonal cycle and variability of the MLD 42" over the whole global oceans and how it relates to the different atmospheric forcings and the 43" upper ocean stratifications. The aim is to understand what are the driving mechanisms of the 44" MLD in different regions and to what extent can these be understood by the local air-sea 45" interactions. In this we will focus on open ocean regions and do not discuss sea ice regions or 46" shallow coastal ocean regions. We will therefore analyse the MLD and the atmospheric 47" forcing terms in ocean reanalysis data and in a single column ocean mixed layer model 48" coupled to an atmospheric general circulation model (GCM). The latter model simulation 49" allows us to understand to what extent the MLD characteristics can be simulated just by local 50" air-sea interactions. It also allows to diagnose the limitations of local air-sea interactions 51" assumption by analysing the model flux correction terms used in the model to maintain a 52" density profiles close to the observed mean profile. 53" The paper is organized as follows: A detailed description of the reanalysis data, coupled 54" model and the method of estimation of MLD is presented in the following section. The focus 55" of the section 3 is on the observed characteristics of MLD estimated from the reanalysis 56" observation. In section 4 we repeat most of the analysis for the single column model 57" simulation. The results are summarized and discussed in section 5. 58" 59" 2. Data, Model and Methods 60" We use reanalysis and a single column ocean model data to understand the characteristics of 61" global ocean mixed layer depth and their relation to local atmospheric forcing. Details of the 62" datasets and methods used in this study are described bellow. 63" 64" a. Reanalysis data set 65" The characteristics of mixed layer presented here are, partly, based on the temperature and 66" salinity vertical profiles from German contribution to Estimating the Circulation and Climate 67" of the Ocean system (GECCO2) for the period 1948-2011. The data has a spatial resolution 68" of 1x1 degree in longitude and latitude and comprises 50 vertical levels. The synthesis used 69" model assimilation with available hydrographic and satellite data. A complete description of 70" this 71" ftp://ftp.icdc.zmaw.de/EASYInit/GECCO2/. dataset is given by Kohl (2014); and the data is available at 72" 73" b. The coupled single column ocean model (ACCESS-KPP) 74" A 50-year simulation of a coupled atmosphere-ocean single column model is used to study 75" the MLD variability over the global ocean. The atmospheric component of the coupled model 76" consists of the Australian Community Climate and Earth-System Simulator (ACCESS 77" version 1.3) atmospheric GCM, which is similar to that of the UK Met Office’s Global 78" Atmosphere version 1.0 and some modification included by the Centre for Australian 79" Weather and Climate Research (CAWCR). The horizontal resolution of 3.75º longitude by 80" 2.5º latitude (referred to as N48) and 38 levels in vertical is applied in this model setup. A 81" detailed description of the atmospheric model is given in Bi et al. (2013). 82" The ocean component of the coupled model used for this study is the single column first- 83" order nonlocal K-Profile Parameterisation (KPP) mixed layer model (Klingaman et al, 2009) 84" by Large et al. (1994). It comprises 40 vertical levels with the layer thickness increasing 85" exponentially from the surface to 1000m deep. The single column ocean model is coupled 86" globally to the atmospheric model with a spatial resolution of 3.75ºx2.5º. The atmospheric 87" model provides the near surface heat flux, wind-stress and freshwater flux (EMP) to the 88" ocean at each time step. The mixed layer model does not present processes such as horizontal 89" advection or upwelling from bellow the limited vertical domain. Mixing in the interior is 90" governed by shear instability, which is modelled as a function of local gradient Richardson 91" number. A boundary layer depth is determined at each grid point, based on the critical value 92" of the turbulent processes parameterised by a bulk Richardson number. Vertical diffusivity 93" coefficients due to turbulent shear are estimated in the diagnosed boundary layer depth. 94" Mixing is strongly enhanced in the boundary layer in both convective and wind driven 95" situations enabling boundary layer properties to penetrate well into the thermocline. Readers 96" are referred to Large et al. (1994) for detailed description of the first order nonlocal KPP 97" fundamentals. Ocean model is embedded with varying sea-ice concentration at the higher 98" latitudes. The sea ice model is a simple thermodynamical model of melting and freezing of 99" sea ice by local forcings. However, regions with sea ice will not be discussed in this study. A 100" flux correction is applied to the temperature and salinity tendency in order to reduce the 101" climate drift, which is described in the following sub-section.. 102" 103" c. Flux corrections 104" The ACCESS-KPP model computes the diffusivity profiles at each grid point and then 105" estimates the temperature and salinity profile. Since lateral ocean dynamics are absent in the 106" model, the temperature and salinity drifts away from the reference values for a longer (many 107" decades) run. So it is necessary to correct the fluxes in order to prevent climate drift. The flux 108" correction values are computed such that the mean seasonal cycle of temperature and salinity 109" closely follow the climatology from WOA09 (Antonov et al, 2010; Locarnini et al, 2010). 110" The temperature and salt flux corrections are obtained from a number of iterations, where the 111" ACCESS-KPP model temperature and salinity are free to evolve. Bias between the model 112" variables and the climatological values are then computed as the flux correction. These 113" values are then added to the tendency equation as a forcing term in the model simulation. The 114" flux corrections vary at each grid and month of year but are state independent and remain the 115" same from one year to the next. Characteristics of the flux correction terms are discussed in 116" the analysis section. 117" 118" d. Method for MLD estimation 119" A wide range of criteria exists in defining MLD from the vertical profiles of temperature, 120" salinity or density. Generally, all these criteria fall into two categories: the difference criteria 121" and the gradient criteria. In the difference criteria MLD is defined as the depth where the 122" oceanic property has changed by a critical value from a reference depth near to the surface 123" (Kara et al. 2000; Montégut et al. 2004). The gradient criteria detect the shallowest depth 124" where the vertical gradient of the oceanic tracers exceeds a given value (Brainerd and Gregg 125" 1995; Lorbacher et al. 2006). The former one largely depends on the difference value chosen 126" and produces miss-representation of MLD in the regions of weekly-stratified surface layers. 127" The gradient criteria also depend on the threshold value of the derivative, but since the 128" gradient is expected to be large at the base of the MLD, it gives a better accuracy for 129" estimation of MLD. Lorbacher et al. (2006) further developed the gradient criteria by adding 130" the standard deviation threshold to assume the shallowest extreme curvature. A better 131" estimation of MLD is done by exponential interpolation method (for thick layers) and hence 132" the Lorbacher et al. (2006) criteria is used to estimate the mixed layer from the density 133" profiles for all the data used in this study. 134" 135" 3. Observed MLD 136" We start the analysis with a first look at the seasonal mean MLDs; see Fig. 1. The seasonal 137" mean MLDs shown in Figure 1 are largely consistent with those found in previous studies 138" (Kara et al. 2003, Montégut et al. 2004 or Lorbacher et al. 2006). Fig. 2 simplifies the 139" seasonal mean MLDs to the annual mean and the relative strength of the seasonal cycle. They 140" show a number of well-known characteristics, such as the strong seasonal cycle in the extra- 141" tropical regions or the shallow MLDs in the upwelling regions along coastlines or the 142" equatorial Pacific. Overall the structure of the global oceans seasonal mean MLDs is fairly 143" complex. To simplify the following discussions and analysis it is useful to define some basic 144" MLD regimes, based on the annual mean MLD and the relative strength of the seasonal 145" cycle; see Fig. 2. We can split the world ocean into the following 3 regimes: 146" • Extra-tropical-seasonal: Regions where the seasonal cycle standard deviation is 147" larger than 60% of the annual mean MLD (Fig. 3a). These are regions with a very 148" shallow warm season MLD and a deep cold season MLD. This is mostly a band along 149" 30-40 degrees with some regions extending further to higher latitudes. The most 150" extreme seasonality is seen in the Southern Ocean in the Indian and Pacific sections, 151" which are within the Antarctic Circumpolar Current (ACC) (see also Dong et al. 152" 2008; Sallée et al. 2006). The northern North Atlantic is also a region of extreme 153" seasonality. 154" • Constant-deep: Regions with an annual mean MLD > 30m and the seasonal cycle 155" standard deviation is less than 60% of the annual mean MLD (Fig. 3b). These are 156" mostly the off equatorial tropical and subtropical regions, but are also large fractions 157" of the Southern Ocean and parts of the northern North Pacific. 158" • Constant-shallow: Regions with an annual mean MLD < 30m and the seasonal cycle 159" standard deviation is less than 60% of the annual mean MLD (Fig. 3c). These mark 160" mostly coastal regions and in particular upwelling regions along major continents and 161" equatorial regions. But also coast lines in the north-western North Atlantic and West 162" Australia. Also most of the North Indian Ocean is in this shallow regime. 163" In following analysis we focus on how these different mean MLD regimes relate to the local 164" atmospheric forcing and the stratification of the upper ocean. 165" 166" 167" 3.1 The annual mean MLD and its relation to local forcing 168" The external forces that act on the MLD are the heat, momentum and evaporation- 169" precipitation difference (EMP) fluxes. All these forces have to interact with the upper ocean 170" stratification. Strongly stratified ocean needs strong forcing to break the stratification and to 171" promote the growth of MLD. By comparing the annual mean values of the atmospheric 172" forcings and the upper ocean stratifications (Fig. 4) with the annual mean MLD (Fig. 2a) we 173" get some idea of the relative importance of the different forcings. Note that the shading in 174" Fig. 4 is coded in such a way that bluish colours indicate forcings or stratifications that would 175" support deeper mixed layers and reddish colours indicate forcings or stratifications 176" supporting shallow MLD. 177" The annual mean Net Heat Flux (NHF) in the higher latitudes (>40º) are mostly negative, 178" which implies that the ocean is losing heat (Fig 4a). Heat-lose at the surface makes the upper 179" ocean statically unstable, which tends to deepen the MLD. Upwelling regions along 180" coastlines and the equator, and tropical Indian Ocean are marked by annual mean heat gain, 181" which tends to support shallow MLD. Overall we find only a moderate match between annual 182" mean NHF forcing and MLD (spatial correlation of 0.3). Most of the constant-shallow MLD 183" regime seems to be associated with annual mean heat gain. Also much of the extra-tropical 184" regions with deeper annual mean MLD are associated with supporting NHF heat loss. 185" However, many regions with shallower than global mean MLD are associated with annual 186" mean heat loss, which does not support the observed shallow MLD (e.g. coastal regions in 187" the western North Pacific and Atlantic; southern subtropical regions in the Indian and Pacific 188" Oceans). Also regions with deeper than global mean MLD are associated with annual mean 189" heat gain, which does not support the observed deep MLD (e.g. off-equatorial tropical 190" Pacific; parts of the Southern Ocean). 191" The annual mean wind stress forcing shows a slightly better match to the annual mean MLD 192" than NHF (spatial correlation of 0.4). In particular higher latitudes regions with deep annual 193" mean MLD also are associated with strong mean wind stress and equatorial regions with 194" shallow annual mean MLD are associated with weak mean wind stress. However, we also see 195" large regions where the annual mean MLD is deep despite fairly moderate annual mean wind 196" stress (e.g. latitude bands around 30 degrees on both hemispheres) and regions where the 197" annual mean MLD is shallow despite fairly strong annual mean wind stress (e.g. subtropical 198" South Indian Ocean; Arabian Sea). We can also note that for most regions NHF and wind 199" stress have similar tendencies on the MLD, with both either supporting deeper MLD (e.g. 200" most extra-tropical regions) or supporting shallower MLD (e.g. equatorial regions). Notable 201" exceptions here are parts of the subtropical regions, parts of the Southern Ocean, northern 202" North Pacific, western North Atlantic and the Arabian Sea. 203" The Precipitation and evaporation (EMP) plays in general a minor role in mixing processes in 204" the upper ocean by changing the density structure (Dong et al, 2009), which is also 205" highlighted here by the mismatch of the annual mean EMP tendencies with the annual mean 206" MLD for most regions (spatial correlation of 0.0). The mean EMP forcing tendencies are also 207" mostly unrelated to the NHF and wind stress forcings. In some regions, however, where both 208" NHF and wind stress tendencies do not match the MLD, the EMP forcing does indeed seem 209" to match the MLD. These appear to be mostly smaller and coastal regions (e.g. coastal 210" regions of Greenland; some parts of the western boundary of the North Pacific and Atlantic). 211" Remarkable are the shallower MLD regions in the subtropical Indian and western Pacific 212" Oceans. Here all three atmospheric forcing would support deeper annual mean MLD. The 213" shallower MLD here seem to match the stronger upper ocean stratifications. Although, upper 214" ocean stratifications is partly a reflection of the atmospheric forcings, it is also a reflection of 215" oceanic process. Since the annual mean atmospheric forcings seem to mismatch the mean 216" MLD in this regions, it suggest that oceanic processes independent of the local atmospheric 217" forcing contribute to the stratification of the upper ocean and thus a shallower MLD. 218" 219" 3.2 The MLD seasonal cycle and its relation to local forcing 220" The extra-tropical-seasonal MLD regime largely represents the world ocean regions with 221" strong mean MLD seasonal cycle (Fig. 2b). These regions are fairly well matched with the 222" strength of the NHF seasonal cycle (Fig. 5a). Although the wind stress forcings in the 223" Northern Hemisphere support the extra-tropical-seasonal MLD regime, they do not show 224" strong seasonality in the Southern Hemisphere. Thus we find that the extra-tropical-seasonal 225" MLD regime is mostly a result of seasonally varying tendencies of heating and cooling at the 226" surface. This is consistent with previous findings (Monterey and Levitus, 1997; Kantha and 227" Clayson, 2000; Kara et al. 2003; Montégut et al. 2004). However, we again find some regions 228" where this simple picture is not supported. In particular the western boundaries of the North 229" Pacific and Atlantic have very strong seasonal cycle in NHF, but not in the MLD. Also the 230" very strong seasonality in the Indian and Pacific Ocean ACC regions are not directly matched 231" by the seasonality in neither of the atmospheric forcings. 232" An interesting aspect of the combined effect of annual mean forcing and the strength of the 233" seasonality in the forcings can be seen along the zonal bands in the northern midlatitudes, see 234" Fig. 6. In the zonal band at around 30oN in the North Pacific the MLD is deeper on the 235" western side and it has a stronger seasonality. Both mean and seasonality decrease to the east. 236" This is supported by the behaviour of the mean heat flux and its seasonality. In the west we 237" have an annual mean cooling and strong seasonality supporting the MLD behaviour and to 238" the east the mean cooling turns into a mean warming and the seasonality in the heat is 239" decreasing again supporting the MLD behaviour. This reflects the change in atmospheric 240" forcing from strong continental characteristics in the west (e.g. strong cooling in winter and 241" strong seasonality) and more in balance with the ocean state in the east, due to the preferred 242" westerly wind conditions that tend to transport characteristics from the west to the east. This 243" is somewhat similar in the North Atlantic, but the relationship is broken near the western 244" boundary region. 245" 246" 3.3 Anomaly variability and its relation to local forcing 247" We next consider the MLD variability beyond the seasonal cycle. The second column of 248" Fig.1 shows the standard deviation (stdv) of monthly mean anomalies MLD estimated for 249" each season. We can notice that the patterns of strong and weak MLD stdv match very well 250" the patterns of deep and shallow mean MLD (first column in Fig. 1). Thus MLD variability is 251" stronger when the mean MLD is deep and MLD variability is weak, when mean MLD is 252" shallow. This pattern even holds when we consider the coefficient of variation (CV; the ratio 253" of stdv divided by the mean); see last column in Fig. 1. The equatorial Pacific and Atlantic, 254" however, mark regions where the MLD variability is strong even though the mean MLD is 255" fairly shallow, which is consistent with the studies by Lorbacher et al. (2006). These mark 256" regions where the concept of MLD may not be so useful, as these regions are more 257" dominated by lateral ocean wave dynamics in the thermocline. It is also remarkable to note 258" that CV > 100% are observed in the spring and winter seasons in the higher latitudes of North 259" Atlantic and Pacific and in the Southern Ocean. This suggests a substantial amount of MLD 260" variability. Even for most other regions the CV values is larger than 30%, indicating 261" significant MLD variability for most regions in most seasons. Keerthi et al. (2012) have also 262" shown large CV values (>40%) at western and eastern equatorial Indian Ocean during boreal 263" summer. 264" Fig. 7 shows the cross-correlation between monthly mean MLD and NHF anomalies for in- 265" phase and for one month lead-time for the NHF for different seasons. Negative in-phase 266" correlations dominate in all seasons and most regions, consistent with cooling (negative NHF 267" anomalies) at the surface ocean leading to increased convective mixing and, hence, a deeper 268" MLD; and vice versa for positive NHF anomalies. Equatorial regions in the central Pacific 269" and western Atlantic show, however, negligible correlation between NHF and MLD 270" anomalies. We can further note that in both hemispheres the in-phase correlation is stronger 271" in fall and winter and fairly weak in spring and summer. The one month MLD lag cross- 272" correlation to NHF is largest in the fall season and weakest in the spring season in both 273" hemispheres. This relation is consistent with the idea that deeper MLD evolves more slowly 274" and therefore has a longer lag-time relation between forcing (NHF) and MLD anomaly. 275" The wind stress forcing is in most regions positively correlated with MLD, consistent with 276" stronger than normal wind stress anomalies leading to deeper MLDs; and vice versa for 277" negative wind stress anomalies (Fig. 8). In the mid to higher latitudes this relationship tends 278" to be stronger in spring and summer. We can also note that stronger correlations are in 279" essentially two regions: the off-equatorial tropical and subtropical regions with a tendency 280" for stronger correlations in the warmer season, and in the extra-tropical higher latitudes in 281" Southern Ocean and a bit less in the northern North Pacific and Atlantic. The lag one month 282" correlation overall appears to be strongest in the subtropical regions and in the fall season in 283" the Southern Ocean. 284" Both the cross-correlations of MLD with NHF and wind stress are on a similar level, with 285" some more wide spread and clearer correlation for the NHF with MLD. Cross-correlations of 286" MLD with EMP are in general weaker than those with NHF and wind stress, but are mostly 287" positive, which is consistent with increased evaporation (saltiness forcing) leading to deeper 288" MLDs (see SFig.1). This relationship is strongest in fall and winter seasons on both 289" hemispheres. 290" 291" 4. The ACCESS-KPP model 292" We now take a look at the MLD and its relation to the atmospheric forcings in the ACCESS- 293" KPP model with a single column ocean mixed layer model. Thus in this simulation the MLD 294" is purely a result of local atmospheric forcing, vertical mixing and a background state 295" independent heat and salinity flux correction to maintain the mean density profile in the 296" ocean columns. 297" A comparison of the patterns of the seasonal mean MLD and its variability between observed 298" (Fig.1) and the ACCESS-KPP model (Fig. 9) already illustrates that the model does a very 299" good representation of the mean MLD characteristics. In the following we will discuss the 300" same MLD characteristics as discussed above for the observations and we will have a look at 301" the role of the flux corrections in maintaining the realistic upper ocean density profiles. 302" 303" 4.1 The annual MLD and its relation to local forcing 304" Figures 10 and 11 show the same atmospheric forcings and upper ocean stratification values 305" as in Fig. 4 and 5, but for the ACCESS-KPP model. First of all we can note that the annual 306" mean and seasonal cycle atmospheric forcings in the ACCESS-KPP model are very similar to 307" the observed values. This suggests that we can analysis the MLD characteristics in the 308" ACCESS-KPP model for similar atmospheric forcings conditions. Some mismatches in the 309" annual man NHF exist in the northern Indian Ocean where the observed NHF is positive, but 310" the ACCESS-KPP model values are negative. 311" The regional distribution of the annual mean MLD and the three main MLD regimes in the 312" ACCESS-KPP model are fairly similar to the observed (compare Figs. 2 and 3 with 12 and 313" 13). The ACCESS-KPP model has a very similar regional distribution of the extra-tropical- 314" seasonal MLD regime, it also simulates the constant-deep MLD regime with a subtropical 315" and southern ocean region and it simulates parts of the constant-shallow MLD regime in the 316" coastal upwelling regions. However, some significant differences between the ACCESS-KPP 317" model and the observations can be noted: first, the overall mean MLD in the ACCESS-KPP 318" model is larger than observed (63m and 41m, respectively). It misses several regions within 319" the constant-shallow MLD regime: the equatorial cold tongues, the tropical Indian Ocean and 320" the northern oceans coastal regions. The extra-tropical-seasonal MLD regime is also a bit 321" more wide spread in the northern oceans than it is observed. The ACCESS-KPP model is 322" further simulating fairly deep MLD in the eastern subtropical Pacific and subtropical Indian 323" Ocean, which is much shallower in the observations. The deeper MLD in these regions seem 324" to fit to fairly strong annual mean local atmospheric heat and wind forcings. It thus seem to 325" indicate that the observed shallow MLD in these regions does not result from the local 326" forcing, but may involve lateral ocean dynamics not represented in the single column 327" ACCESS-KPP model. 328" The relation between the atmospheric forcings and the MLD in ACCESS-KPP model is also 329" similar to the observed (compare Figs. 4 and 10). Again the wind stress forcing tends to be 330" more strongly related to the annual mean MLD than the heat flux forcing and again the EMP 331" forcing has no relation to the annual mean MLD. 332" 333" 4.2 The seasonal cycle of MLD and its relation to local forcing 334" As we said in the previous section, the model strives to produce the observed seasonal 335" changes reasonably well (compare Figures 2 and 12): it essentially captures the zonal bands 336" of strong extra-tropical-seasonal MLD regime in the midlatitudes, weaker in the tropics and 337" subtropics and also weaker in parts of the southern ocean. As also indicated in the previous 338" section the ACCESS-KPP model tends to over estimate the seasonality in the eastern North 339" Pacific and western costal area in the North Atlantic. The model is good at representing the 340" mean summer MLD, but overestimates MLD during fall and spring. This however, cannot be 341" due to stronger seasonality in the atmospheric forcings, as they appear to be similar or weaker 342" than observed. This suggests that the ACCESS-KPP model is missing some dynamics to 343" capture these features. 344" The seasonal cycle of the atmospheric forcings are well captured by the ACCESS-KPP model 345" and again the relation between the heating and wind stress forcing and the MLD are mostly 346" as observed (compare Figures 5 and 11). Thus the seasonality in the heat is the main forcing 347" for the seasonal cycle of the MLD in ACCESS-KPP model. 348" 349" 4.3 Anomaly variability and its relation to local forcing 350" The over all strength and seasonal distribution of the anomaly variability of the MLD in the 351" ACCESS-KPP model is very similar to the observed (compare Figs. 1 and 9). The stdv of 352" MLD also mostly strong where the mean MLD is deep, but it follows the mean MLD even 353" more closely. Thus the coefficient of variability is more homogenous in the ACCESS-KPP 354" model. 355" The correlation between heat flux and wind stress forcings and the MLD are similar in all 356" seasons, but are overall much stronger than observed (compare Figs. 7 and 8 with 14 and 15). 357" This could be related to both errors in the observations and to the simplified dynamics of the 358" ACCESS-KPP model. The observations include atmospheric heat fluxes, ocean observations 359" and the model simulation of the reanalysis model. The errors in these factors will lead to 360" inconsistencies between the forcings and the MLD. Subsequently we have to assume that the 361" correlation between forcings and the MLD is decreased in the observations. On the other 362" hand the ACCESS-KPP model has simplified dynamics that only include local forcings and 363" vertical mixing; neglecting all other lateral ocean dynamics. Subsequently we have to assume 364" that the correlation between local forcings and the MLD is artificially enhanced. 365" As in the observations the effect of the anomalous wind stress forcings is stronger in the 366" warmer, shallower MLD seasons. The lag-1 correlation of the heat flux is also stronger in the 367" cold, deep MLD seasons as in the observations. The EMP also shows higher correlations with 368" MLD (see SFig.1 and 2) than observed, but still much weaker than the correlations with heat 369" flux and wind stress, suggesting that the EMP forcing is not as important. 370" 371" 4.4 Missing dynamics and the role of flux corrections 372" Although the ACCESS-KPP model does a fairly good representation of the seasonal mean 373" MLD and its variability by only simulating the local atmospheric forcing and the vertical 374" single column dynamics, it does indeed have substantial limitations in simulating the MLD. 375" A key aspect of the ACCESS-KPP model is that additional heat and salinity flux correction 376" terms exist on all ocean model levels. These flux corrections are essential in many regions to 377" maintain the overall stratification in the density profile of the ocean, see Fig. 16. Without the 378" heat flux corrections, in particular (salinity corrections not shown), the ocean stratification in 379" the tropics and subtropics would collapse eventually (over periods longer than 10yrs) and the 380" warm surface waters would mix all the way to the bottom layer of the KPP model (1000m 381" depth). This is indicated by the significant surface warming and subsurface cooling of the 382" heat flux correction over wide spread regions in the tropics and subtropics. This effectively 383" stabilizes the density profile and thus maintains the ocean stratifications. Since the flux 384" correction mimic the mean effect of the missing ocean dynamics, it suggests that lateral and 385" other ocean dynamics not simulated in the ACCESS-KPP model are indeed important in 386" simulating the main MLD characteristics. 387" 388" " 389" 5. Summary and Discussion 390" In this study, we described the spatial and temporal relation between the local atmospheric 391" forcing and the mean, seasonal cycle and variability of the MLD using a recent ocean 392" reanalysis product (GECCO2) and single column ocean model (KPP) coupled to an 393" atmospheric model (ACCESS1.3) over the global ocean. The aim of this study was to 394" understand what are the driving mechanisms of the MLD in different regions and to what 395" extent can these be understood by the local air-sea interactions. We focussed on the MLD 396" characteristics of the annual mean, the relative seasonal cycle strength and the variability 397" anomalous from the seasonal cycle. To further simplify the fairly complex characteristics we 398" introduced three main regimes based on the characteristics of the annual mean and the 399" relative seasonal cycle strength. 400" First of all we can summaries our findings about the annual mean and the relative seasonal 401" cycle strength of the MLD: 402" 403" • The annual mean MLD over most open ocean regions (away from sea ice or shallow coastal regions) follows the wind and heat forcing. Regions with stronger mean wind 404" forcing tend to have larger annual mean MLD and regions with annual mean net 405" heating tend to have shallower annual mean MLD. The annual mean wind forcing 406" strength appears to be most strongly related to the annual mean MLD and stronger 407" than the net heat forcing. Both of these forcings are, however, only weakly correlated 408" with the global pattern of the annual mean MLD indicating that the relationship is 409" more complex. The annual mean EMP forcing is for most regions not related to the 410" annual mean MLD suggesting it is not a dominating forcing. However, in some 411" regions where both annual mean NHF and wind stress tendencies do not match the 412" MLD, the EMP forcing does indeed seem to match the annual mean MLD (e.g. 413" coastal regions of Greenland; some parts of the western boundary of the North Pacific 414" and Atlantic). 415" • The relative seasonal cycle strength of the MLD is also mostly related to the relative 416" seasonal cycle strength of the net heating and wind stress, but here the relationship 417" with the net heating is stronger than with the wind forcing. Thus relative seasonal 418" cycle strength of the MLD mostly follows the strength of the seasonal cycle in the net 419" heating. 420" • Remarkable are a few regions in which neither of the atmospheric forcings seem to 421" follow the annual mean or seasonal cycle characteristics of the MLD. The shallower 422" MLD regions in the subtropical Indian and western Pacific Oceans does not seem to 423" relate to any of three atmospheric forcing. The western boundaries of the North 424" Pacific and Atlantic have very strong seasonal cycle in NHF, but not in the MLD. 425" Also the very strong seasonality in the Indian and Pacific Ocean ACC regions are not 426" directly matched by the seasonality in neither of the atmospheric forcings. Since the 427" atmospheric forcings seem to mismatch the MLD characteristics in these regions and 428" the single column ocean model is also not capable in simulating these characteristics, 429" it suggests that oceanic processes independent of the local atmospheric forcing 430" contribute to the stratification of the upper ocean and thus the MLD in significant 431" ways. 432" 433" Alternatively or in addition we can summarize the processes that dominate the three different 434" MLD regimes: 435" 436" • Extra-tropical-seasonal: This regime of the MLD in the midlatitudes is primarily a result of the strong seasonal cycle in the heating. The seasonal cycle in heating is 437" indeed strongest in the midlaitudes and matches the strongest relative seasonal cycle 438" strength of the MLD fairly well. This, however, is also modulated by oceanic fronts, 439" strong boundary currents and atmospheric storm track positions. 440" • Constant-deep: The regime where the MLD is fairly deep (>30m) through out the 441" year with no strong seasonal cycle mostly are in the subtropics and some high 442" latitudes. These are regions with no strong annual mean heating nor cooling, no strong 443" seasonal cycle in heating and the winds tend to be slightly below the global mean. 444" • Constant-shallow: Shallow MLD without strong seasonal cycles are mostly coastal 445" regions, equatorial regions and the tropical Indian Ocean. Some of these regions are 446" fairly well related to the local forcings, such as coastal and some equatorial upwelling 447" regions. However, much of this constant-shallow MLD regime is not well matched 448" with the local forcings and is not well simulated in the ACCESS-KPP simulation, 449" suggesting that local forcings and processes are not sufficient to explain these MLD 450" characteristics. In particular in the higher latitudes coastal regions, the central ocean 451" basin equatorial regions and in the Indian Ocean are the constant-shallow MLD 452" regimes not matched to local forcings nor processes. 453" 454" The analysis of the anomalous variability of MLD and how it relates to the local forcing can 455" be summarized to the following main points: 456" • The overall strength of the MLD variability is in general proportional to the mean 457" variability, with stronger variability when the mean MLD is deep (e.g. cold season). 458" The MLD variability appears to quite significant with values of 30% to over 100% of 459" the seasonal mean MLD values. 460" • Most of this is well captured by the ACCESS-KPP model, but in some regions this is 461" not well simulated by the model. This includes the equatorial Pacific and in higher 462" latitudes the simulation does not simulate as strong MLD variability relative to the 463" seasonal mean MLD as observed. 464" • MLD variability is mostly negatively correlated with heat flux variability and 465" positively with wind stress variability. In the mid to higher latitudes the relationship 466" to heating tends to be stronger in the cold (deep MLD) seasons and the relationship to 467" wind forcing tends to be stronger in the shallower MLD (warm) seasons. In seasons 468" with deeper mean MLD the relationship to the atmospheric forcings tends to become 469" stronger when the atmospheric forcings leads by about one month, which seem to be 470" consistent with a larger inertia of the MLD when it is deeper. 471" • The ACCESS-KPP model is consistent with these observed relationships between the 472" local atmospheric forcings and the MLD, but the relationships are significantly 473" stronger in the model simulations. This may to some part reflect the simplification of 474" the model to only local vertical mixing processes and neglecting other ocean 475" dynamics. Thus over estimate the role of local forcings on the MLD variability. But it 476" may also point to observational errors that lead to inconsistencies between 477" atmospheric forcings and the ocean state that artificially degrade the relationship 478" between local forcing and the MLD. 479" While overall the local forcing perspective does a fairly good representation of most of the 480" MLD characteristics in the global oceans, it does have some limitations as mentioned above. 481" This is in particular highlighted by the fact that the ACCESS-KPP model does require 482" significant flux correction in temperature (most importantly) and also in salinity. These are 483" important to maintain upper ocean stratification close to the observed. Without such terms in 484" the ACCESS-KPP model the upper ocean stratification in most subtropical regions would 485" collapse after about 10yrs and the MLD would go all the way down to the deepest layer of 486" the ACCESS-KPP model (1000m). 487" The analysis has shown that the MLD characteristics are a complex interaction between the 488" local forcings, ocean stratifications and potentially lateral ocean dynamics. Most of the 489" results discussed here result from comparisons of the overall statistics in observed MLD and 490" local forcings and in the ACCESS-KPP simulation. To further untangle the interactions and 491" the relative contribution of different forcings and different oceanic process it would require to 492" do sensitivity experiments with the ACCESS-KPP or similar model simulations, in which 493" forcings or elements of the forcings are turned ‘off’ or in which processes are altered or 494" turned ‘off’. However, this is beyond this study and is left here for future analysis. 495" 496" 497" 498" 499" Acknowledgments 500" Reference 501" 502" Alexander, M., J. Scott, and C. Deser, 2000: Processes that influence sea surface temperature 503" and ocean mixed layer depth variability in a coupled model. Journal of Geophysical 504" Research, 105, 16823, doi:10.1029/2000JC900074. 505" 506" Antonov, J. I., D. Seidov, T. P. Boyer, R. A. Locarnini, A. V. Mishonov, H. E. Garcia, O. K. 507" Baranova, M. M. Zweng, and D. R. Johnson, 2010. World Ocean Atlas 2009, Volume 2: 508" Salinity. S. Levitus, Ed. NOAA Atlas NESDIS 69, U.S. Government Printing Office, 509" Washington, D.C., 184 pp. 510" 511" Bates, N., 2001: Interannual variability of oceanic CO2 and biogeochemical properties in the 512" Western North Atlantic subtropical gyre. Deep Sea Research Part II: Topical Studies in 513" Oceanography, 48, 15071528, doi:10.1016/S0967-0645(00)00151-X. 514" 515" Bi, D. and Coauthors, 2013: The ACCESS coupled model: description, control climate and 516" evaluation. Australian Meteorological and Oceanographic Journal, 63, 41–64. 517" 518" Brainerd, KE, and MC Gregg, 1995: Surface mixed and mixing layer depths. Deep Sea 519" Research Part I: Oceanographic …,. 520" http://www.sciencedirect.com/science/article/pii/096706379500068H. 521" 522" Carton, J., S. Grodsky, and H. Liu, 2008: Variability of the Oceanic Mixed Layer, 1960– 523" 2004. Journal of Climate, 21, 10291047, doi:10.1175/2007JCLI1798.1. 524" 525" Chen, D., A. Busalacchi, and L. Rothstein, 1994: The roles of vertical mixing, solar radiation, 526" and wind stress in a model simulation of the sea surface temperature seasonal cycle in the 527" tropical Pacific Ocean. Journal of Geophysical Research: Oceans (1978–2012), 99, 20345– 528" 20359, doi:10.1029/94JC01621. 529" 530" Cronin, M., and W. Kessler, 2002: Seasonal and interannual modulation of mixed layer 531" variability at 0°, 110°W. Deep Sea Research Part I: Oceanographic Research Papers, 49, 532" 117, doi:10.1016/S0967-0637(01)00043-7. 533" 534" Dommenget, D., and M. Latif, 2002: Analysis of observed and simulated SST spectra in the 535" midlatitudes. Climate Dynamics, 19, 277–288, doi:10.1007/s00382-002-0229-9. 536" 537" Dong, S., S. Gille, and J. Sprintall, 2007: An Assessment of the Southern Ocean Mixed Layer 538" Heat Budget. J. Climate, 20, 44254442, doi:10.1175/JCLI4259.1. 539" 540" Dong, S., J. Sprintall, S. Gille, and L. Talley, 2008: Southern Ocean mixed layer depth from 541" Argo float profiles. Journal of Geophysical Research: Oceans (1978–2012), 113, 542" doi:10.1029/2006JC004051. 543" 544" Dong, S., S. Garzoli, and M. Baringer, 2009: An assessment of the seasonal mixed layer 545" salinity budget in the Southern Ocean. J. Geophys. Res., 114, doi:10.1029/2008JC005258. 546" 547" Fasham, M., 1995: Variations in the seasonal cycle of biological production in subarctic 548" oceans: A model sensitivity analysis. Deep Sea Research Part I: Oceanographic Research 549" Papers, 42, 11111149, doi:10.1016/0967-0637(95)00054-A. 550" 551" Godfrey, J., and E. Lindstrom, 1989: The heat budget of the equatorial western Pacific 552" surface mixed layer. J. Geophys. Res., 94, 8007–8017, doi:10.1029/JC094iC06p08007. 553" 554" Kantha, L. H., and C. C. Clayson, 2000: Small scale processes in geophysical fluid flows, 555" International Geophysics Series, vol. 67. Academic Press, New York, p. 157-160. 556" 557" Kara, A., P. Rochford, and H. Hurlburt, 2000: An optimal definition for ocean mixed layer 558" depth. J. Geophys. Res., 105, 16803–16821, doi:10.1029/2000JC900072. 559" 560" Kara, A., P. Rochford, and H. Hurlburt, 2003: Mixed layer depth variability over the global 561" ocean. J. Geophys. Res., 108, doi:10.1029/2000JC000736. 562" 563" Keerthi, M., M. Lengaigne, J. Vialard, C. Montégut, and P. Muraleedharan, 2012: Interannual 564" variability of the Tropical Indian Ocean mixed layer depth. Climate Dynamics, 40, 743–759, 565" doi:10.1007/s00382-012-1295-2. 566" 567" Klingaman, N. P., H. Weller, S. J. Woolnough, P. M. Inness, and J. M. Slingo, 2009: Coupled 568" simulations of the indian monsoon intraseasonal oscillation using a fine-resolution mixed- 569" layer ocean model. Proc. ECMWF Workshop on Ocean-Atmosphere Interaction (2008), 570" European Centre for Medium-Range Weather Forecasts, Reading, England, 195–2005. 571" 572" Köhl, A., 2014: Evaluation of the GECCO2 ocean synthesis: transports of volume, heat and 573" freshwater in the Atlantic. Quarterly Journal of the Royal Meteorological Society, 574" doi:10.1002/qj.2347. 575" 576" Large, W., J. McWilliams, and S. Doney, 1994: Oceanic vertical mixing: A review and a 577" model with a nonlocal boundary layer parameterization. Reviews of Geophysics, 32, 363–403, 578" doi:10.1029/94RG01872. 579" 580" Locarnini, R. A., A. V. Mishonov, J. I. Antonov, T. P. Boyer, H. E. Garcia, O. K. Baranova, 581" M. M. Zweng, and D. R. Johnson, 2010. World Ocean Atlas 2009, Volume 1: Temperature. 582" S. Levitus, Ed. NOAA Atlas NESDIS 68, U.S. Government Printing Office, Washington, 583" D.C., 184 pp. 584" 585" Lorbacher, K., D. Dommenget, P. Niiler, and A. Köhl, 2006: Ocean mixed layer depth: A 586" subsurface proxy of ocean atmosphere variability. Journal of Geophysical Research: Oceans 587" (1978–2012), 111, doi:10.1029/2003JC002157. 588" 589" Maykut, G., and M. McPhee, 1995: Solar heating of the Arctic mixed layer. J. Geophys. Res., 590" 100, 24691–24703, doi:10.1029/95JC02554. 591" 592" Montégut, C., G. Madec, A. Fischer, A. Lazar, and D. Iudicone, 2004: Mixed layer depth 593" over the global ocean: An examination of profile data and a profile based climatology. 594" Journal of Geophysical Research: Oceans (1978–2012), 109, doi:10.1029/2004JC002378. 595" 596" Montégut, C., J. Mignot, A. Lazar, and S. Cravatte, 2007: Control of salinity on the mixed 597" layer depth in the world ocean: 1. General description. Journal of Geophysical Research: 598" Oceans (1978–2012), 112, doi:10.1029/2006JC003953. 599" 600" Monterey, G., and S. Levitus, Seasonal Variability of Mixed Layer Depth for the World 601" Ocean, NOAA Atlas NESDIS, vol. 14, 100 pp., Natl. Oceanic and Atmos. Admin., Silver 602" Spring, Md., 1997. 603" 604" Narvekar, J., and S. Kumar, 2006: Seasonal variability of the mixed layer in the central Bay 605" of Bengal and associated changes in nutrients and chlorophyll. Deep Sea Research Part I: 606" Oceanographic Research Papers, 53, 820835, doi:10.1016/j.dsr.2006.01.012. 607" 608" Peter, A., M. Hénaff, Y. Penhoat, C. Menkes, F. Marin, J. Vialard, G. Caniaux, and A. Lazar, 609" 2006: A model study of the seasonal mixed layer heat budget in the equatorial Atlantic. J. 610" Geophys. Res., 111, doi:10.1029/2005JC003157. 611" 612" Polovina, J., G. Mitchum, and G. Evans, 1995: Decadal and basin-scale variation in mixed 613" layer depth and the impact on biological production in the Central and North Pacific, 1960- 614" 88. Deep Sea Research Part I: Oceanographic Research Papers, 42, 1701–1716, 615" doi:10.1016/0967-0637(95)00075-H. 616" 617" Qui, B., S. Chen, and P. Hacker, 2004: Synoptic-Scale Air–Sea Flux Forcing in the Western 618" North Pacific: Observations and Their Impact on SST and the Mixed Layer. Journal of 619" Physical Oceanography, 34, 2148–2159, doi:10.1175/1520- 620" 0485(2004)034<2148:SAFFIT>2.0.CO;2. 621" 622" Sabine, C. L. and Coauthors, 2004: The oceanic sink for anthropogenic CO2. Science, 305, 623" 367–371, doi:10.1126/science.1097403. 624" 625" Sallée, J.-B., N. Wienders, K. Speer, and R. Morrow, 2006: Formation of subantarctic mode 626" water in the southeastern Indian Ocean. Ocean Dynamics, 56, 525542, doi:10.1007/s10236- 627" 005-0054-x. 628" 629" Swenson, M., and D. Hansen, 1999: Tropical Pacific Ocean Mixed Layer Heat Budget: The 630" Pacific Cold Tongue. J. Phys. Oceanogr., 29, 6981, doi:10.1175/1520- 631" 0485(1999)029<0069:TPOMLH>2.0.CO;2. 632" 633" Takahashi, T., R. Feely, R. Weiss, R. Wanninkhof, D. Chipman, S. Sutherland, and T. 634" Takahashi, 1997: Global air-sea flux of CO2: An estimate based on measurements of sea–air 635" pCO2!difference. Proceedings of the National Academy of Sciences, 94, 8292–8299, 636" doi:10.1073/pnas.94.16.8292. 637" Figure Captions:- 638" Fig" 1:" Seasonal" statistics" of" mixed" layer" depth" from GECCO2 averaged over January- 639" March (a,b,c), April-June (d,e,f), July-September (g,h,i) and October-December (j,k,l) . Left 640" side panel of the figure shows the long-term monthly mean averaged over the seasonal 641" months of the year (a,d,g,j). The standard deviations of MLD in each month for the years 642" 1948-2011 averaged then to the seasonal months are shown in the middle panel (b,e,h,k). On 643" the right panel (c,f,i,l) we plotted the standard deviation relative to the corresponding 644" monthly mean MLD (i.e., the Coefficient of Variance). 645" 646" Fig"2:"(a) Annual mean MLD estimated from GECCO2 density profiles, and (b) Seasonal 647" amplitude (computed as the standard deviation of monthly climatology) of MLD relative to 648" annual mean MLD." 649" 650" Fig"3:"Different"MLD"regimes"derived"from"GECCO2"dataset."(a)"ExtraLtropicalLseasonal," 651" shows"the"regions"where"seasonal"amplitude"change"is">60%"of"the"annual"mean"MLD." 652" (b)"ConstantLseep,"where"MLD">30m"and"seasonal"change"is"<60%."(c)"ConstantL 653" shallow,"shows"the"regions"with"<60%"relative"seasonal"change"with"MLD"<30m." 654" 655" Fig"4:"Annual mean fields from GECCO2. (a) net heat flux (NHF); positive/negative sign 656" indicate the ocean is gaining/losing heat, (b) wind-stress, (c) density gradient between surface 657" and 100m depth, and (d) Evaporation minus Precipitation (E-P). Colour code is in such a way 658" that blue shading represents favourable for deep MLD and red shading for shallow mixed 659" layers. Spatial correlation (r) between the forcing variable and annual mean MLD (Fig 2a) is 660" shown at top-left corner of the corresponding figure. " 661" 662" Fig" 5: Seasonal amplitude of GECCO2 variables (computed as the standard deviation of 663" monthly climatology). (a) net heat flux (NHF), (b) wind-stress, (c) density gradient between 664" surface and 100m depth, and (d) Evaporation minus Precipitation (E-P). Spatial correlation 665" (r) between seasonal change for forcing variable and MLD (relative to annual mean, in Fig 666" 2b) is shown at top-left corner of the corresponding figure. 667" 668" Fig"6:"The annual mean MLD, net heat flux (NHF) and wind-stress (from GECCO2) 669" averaged along (a) 30º-32ºN (North Pacific) and (b) 33º-35ºN (North Atlantic). The 670" positive/negative heat flux (NHF) values show the ocean is losing/gaining heat. The colour 671" shading is the corresponding seasonal amplitude change at each grid point; light-red for NHF, 672" light-blue for MLD and light-yellow is for wind-stress." 673" 674" Fig"7:"Cross"correlation"of"Net"heat"flux"(NHF)"and"MLD"anomalies"in"different"seasons" 675" from"GECCO2."Left"panel"shows"concurrent"correlations"and"the"right"panel"represents" 676" corresponding"oneLmonth"lag"correlations,"where"MLD"lags"behind"oneLmonth"of"the" 677" forcing."Figure"aLb"shows"the"lag"correlation"during"JanuaryLMarch"(JFM),"cLd" 678" represents"AprilLJune"(AMJ),"eLf"shows"the"lag"correlation"during"JulyL"September"(JAS)," 679" and"gLh"corresponds"to"OctoberLNovember"(OND)."" 680" 681" Fig 8: Cross correlation of wind-stress (TAU) and MLD anomalies in different seasons from 682" GECCO2. Left panel shows concurrent correlations and the right panel represents 683" corresponding one-month lag correlations, where MLD lags behind one-month of the forcing. 684" Figure"aLb"shows"the"lag"correlation"during"JanuaryLMarch"(JFM),"cLd"represents"AprilL 685" June"(AMJ),"eLf"shows"the"lag"correlation"during"JulyL"September"(JAS),"and"gLh" 686" corresponds"to"OctoberLNovember"(OND)."" 687" 688" Fig 9:"Seasonal"statistics"of"mixed"layer"depth,"same"as"in"Fig"1"but"for ACCESS-KPP 689" coupled model. 690" 691" Fig"10:"Annual mean fields from ACCESS-KPP coupled model. (a) net heat flux (NHF); 692" positive/negative sign indicate the ocean is gaining/losing heat, (b) wind-stress, (c) density 693" gradient between surface and 100m depth, and (d) Evaporation minus Precipitation (E-P). 694" Colour code is in such a way that blue shading represents favourable for deep MLD and red 695" shading for shallow mixed layers. Spatial correlation (r) between the forcing variable and 696" annual mean MLD (Fig 12a) is shown at top-left corner of the corresponding figure. 697" 698" Fig 11: Seasonal amplitude of ACCESS-KPP forcing variables (computed as the standard 699" deviation of monthly climatology). (a) net heat flux (NHF), (b) wind-stress, (c) density 700" gradient between surface and 100m depth, and (d) Evaporation minus Precipitation (E-P). 701" Spatial correlation (r) between seasonal change for forcing variable and MLD (relative to 702" annual mean, in Fig 12b) is shown at top-left corner of the corresponding figure. 703" " 704" Fig"12:"(a) Annual mean MLD estimated from ACCESS-KPP simulated density profiles, and 705" (b) Seasonal amplitude of MLD relative to annual mean MLD (same as in Fig 2). 706" 707" Fig"13:"Different"MLD"regimes"derived"from"ACCESSLKPP"simulation."(a)"ExtraLtropicalL 708" seasonal,"shows"the"regions"where"seasonal"amplitude"change"is">60%"of"the"annual" 709" mean"MLD."(b)"ConstantLseep,"where"MLD">30m"and"seasonal"change"is"<60%."(c)" 710" ConstantLshallow,"shows"the"regions"with"<60%"relative"seasonal"change"with"MLD" 711" <30m." 712" 713" Fig"14:"Cross"correlation"of"Net"heat"flux"(NHF)"and"MLD"anomalies"in"different"seasons" 714" from"ACCESSLKPP"model."Left"panel"shows"concurrent"correlations"and"the"right"panel" 715" represents"corresponding"oneLmonth"lag"correlations,"where"MLD"lags"oneLmonth"of" 716" the"forcing."Figure"aLb"shows"the"lag"correlation"during"JanuaryLMarch"(JFM),"cLd" 717" represents"AprilLJune"(AMJ),"eLf"shows"the"lag"correlation"during"JulyL"September"(JAS)," 718" and"gLh"corresponds"to"OctoberLNovember"(OND)."" 719" 720" Fig 15: Cross correlation of wind-stress (TAU) and MLD anomalies in different seasons 721" from"ACCESSLKPP"model. Left panel shows concurrent correlations and the right panel 722" represents corresponding one-month lag correlations, where MLD lags one-month of the 723" forcing. Figure"aLb"shows"the"lag"correlation"during"JanuaryLMarch"(JFM),"cLd"represents" 724" AprilLJune"(AMJ),"eLf"shows"the"lag"correlation"during"JulyL"September"(JAS),"and"gLh" 725" corresponds"to"OctoberLNovember"(OND)." 726" " 727" Fig"16:"Annual"mean"Heat"flux"correction,"(a)"surface"to"50m"depth"and"(b)"200"to"300m" 728" depth,"applied"to"the"temperature"tendency"equation"in"the"model."Flux"correction"is" 729" applied"in"order"to"reduce"the"drift"in"temperature"values"for"long"simulation."" 730" 731" ! ! ! ! ! Fig! 1:! Seasonal! statistics! of! mixed! layer! depth! from GECCO2 averaged over JanuaryMarch (a,b,c), April-June (d,e,f), July-September (g,h,i) and October-December (j,k,l) . Left side panel of the figure shows the long-term monthly mean averaged over the seasonal months of the year (a,d,g,j). The standard deviations of MLD in each month for the years 1948-2011 averaged then to the seasonal months are shown in the middle panel (b,e,h,k). On the right panel (c,f,i,l) we plotted the standard deviation relative to the corresponding monthly mean MLD (i.e., the Coefficient of Variance). ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! Fig!2:!(a) Annual mean MLD estimated from GECCO2 density profiles, and (b) Seasonal amplitude (computed as the standard deviation of monthly climatology) of MLD relative to annual mean MLD ! ! ! Fig!3:!Different!MLD!regimes!derived!from!GECCO2!dataset.!(a)!ExtraEtropicalEseasonal,! shows!the!regions!where!seasonal!amplitude!change!is!>60%!of!the!annual!mean!MLD.! (b)!ConstantEseep,!where!MLD!>30m!and!seasonal!change!is!<60%.!(c)!ConstantE shallow,!shows!the!regions!with!<60%!relative!seasonal!change!with!MLD!<30m.! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! Fig!4:!Annual mean fields from GECCO2. (a) net heat flux (NHF); positive/negative sign indicate the ocean is gaining/losing heat, (b) wind-stress, (c) density gradient between surface and 100m depth, and (d) Evaporation minus Precipitation (E-P). Colour code is in such a way that blue shading represents favourable for deep MLD and red shading for shallow mixed layers. Spatial correlation (r) between the forcing variable and annual mean MLD (Fig 2a) is shown at top-left corner of the corresponding figure. ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! Fig!5: Seasonal amplitude of GECCO2 variables (computed as the standard deviation of monthly climatology). (a) net heat flux (NHF), (b) wind-stress, (c) density gradient between surface and 100m depth, and (d) Evaporation minus Precipitation (E-P). Spatial correlation (r) between seasonal change for forcing variable and MLD (relative to annual mean, in Fig 2b) is shown at top-left corner of the corresponding figure. ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! (a) North Pacific ! ! ! ! ! ! ! ! ! ! ! ! ! (b) North Atlantic ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! Fig!6:!The annual mean MLD, net heat flux (NHF) and wind-stress (from GECCO2) averaged along (a) 30º-32ºN (North Pacific) and (b) 33º-35ºN (North Atlantic). The positive/negative heat flux (NHF) values show the ocean is losing/gaining heat. The colour shading is the corresponding seasonal amplitude change at each grid point; light-red for NHF, light-blue for MLD and light-yellow is for wind-stress.! ! ! ! ! ! ! ! ! ! ! ! Fig!7:!Cross!correlation!of!Net!heat!flux!(NHF)!and!MLD!anomalies!in!different!seasons! from!GECCO2.!Left!panel!shows!concurrent!correlations!and!the!right!panel!represents! corresponding!oneEmonth!lag!correlations,!where!MLD!lags!behind!oneEmonth!of!the! forcing.!Figure!aEb!shows!the!lag!correlation!during!JanuaryEMarch!(JFM),!cEd!represents! AprilEJune!(AMJ),!eEf!shows!the!lag!correlation!during!JulyE!September!(JAS),!and!gEh! corresponds!to!OctoberENovember!(OND).!! ! ! ! ! ! ! ! ! ! ! Fig 8: Cross correlation of wind-stress (TAU) and MLD anomalies in different seasons from GECCO2. Left panel shows concurrent correlations and the right panel represents corresponding one-month lag correlations, where MLD lags behind one-month of the forcing. Figure!aEb!shows!the!lag!correlation!during!JanuaryEMarch!(JFM),!cEd!represents!AprilE June!(AMJ),!eEf!shows!the!lag!correlation!during!JulyE!September!(JAS),!and!gEh! corresponds!to!OctoberENovember!(OND).!! ! ! ! ! ! ! ! Fig 9:! Seasonal! statistics! of! mixed! layer! depth,! same! as! in! Fig! 1! but! for ACCESS-KPP coupled model. ! ! ! ! ! ! ! ! ! ! ! ! Fig!10:!Annual mean fields from ACCESS-KPP coupled model. (a) net heat flux (NHF); positive/negative sign indicate the ocean is gaining/losing heat, (b) wind-stress, (c) density gradient between surface and 100m depth, and (d) Evaporation minus Precipitation (E-P). Colour code is in such a way that blue shading represents favourable for deep MLD and red shading for shallow mixed layers. Spatial correlation (r) between the forcing variable and annual mean MLD (Fig 12a) is shown at top-left corner of the corresponding figure. ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! Fig 11: Seasonal amplitude of ACCESS-KPP forcing variables (computed as the standard deviation of monthly climatology). (a) net heat flux (NHF), (b) wind-stress, (c) density gradient between surface and 100m depth, and (d) Evaporation minus Precipitation (E-P). Spatial correlation (r) between seasonal change for forcing variable and MLD (relative to annual mean, in Fig 12b) is shown at top-left corner of the corresponding figure. ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! Fig!12:!(a) Annual mean MLD estimated from ACCESS-KPP simulated density profiles, and (b) Seasonal amplitude of MLD relative to annual mean MLD (same as in Fig 2). ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! Fig!13:!Different!MLD!regimes!derived!from!ACCESSEKPP!simulation.!(a)!ExtraEtropicalE seasonal,!shows!the!regions!where!seasonal!amplitude!change!is!>60%!of!the!annual! mean!MLD.!(b)!ConstantEseep,!where!MLD!>30m!and!seasonal!change!is!<60%.!(c)! ConstantEshallow,!shows!the!regions!with!<60%!relative!seasonal!change!with!MLD! <30m.! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! Fig!14:!Cross!correlation!of!Net!heat!flux!(NHF)!and!MLD!anomalies!in!different!seasons! from!ACCESSEKPP!model.!Left!panel!shows!concurrent!correlations!and!the!right!panel! represents!corresponding!oneEmonth!lag!correlations,!where!MLD!lags!oneEmonth!of!the! forcing.!Figure!aEb!shows!the!lag!correlation!during!JanuaryEMarch!(JFM),!cEd!represents! AprilEJune!(AMJ),!eEf!shows!the!lag!correlation!during!JulyE!September!(JAS),!and!gEh! corresponds!to!OctoberENovember!(OND).!! ! ! ! ! ! ! Fig 15: Cross correlation of wind-stress (TAU) and MLD anomalies in different seasons from! ACCESSEKPP!model. Left panel shows concurrent correlations and the right panel represents corresponding one-month lag correlations, where MLD lags one-month of the forcing. Figure! aEb!shows!the!lag!correlation!during!JanuaryEMarch!(JFM),!cEd!represents!AprilEJune! (AMJ),!eEf!shows!the!lag!correlation!during!JulyE!September!(JAS),!and!gEh!corresponds! to!OctoberENovember!(OND).!! ! ! ! ! ! ! ! ! ! ! ! Fig!16:!Annual!mean!Heat!flux!correction,!(a)!surface!to!50m!depth!and!(b)!200!to!300m! depth,!applied!to!the!temperature!tendency!equation!in!the!model.!Flux!correction!is! applied!in!order!to!reduce!the!drift!in!temperature!values!for!long!simulation.!! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! Supplementary,figures, ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! SFig!1:!Cross!correlation!of!EMP!and!MLD!anomalies!in!different!seasons!from!GECCO2.! Left!panel!shows!concurrent!correlations!and!the!right!panel!represents!corresponding! oneEmonth!lag!correlations,!where!MLD!lags!behind!oneEmonth!of!the!forcing.!Figure!aEb! shows!the!lag!correlation!during!JanuaryEMarch!(JFM),!cEd!represents!AprilEJune!(AMJ),! eEf!shows!the!lag!correlation!during!JulyE!September!(JAS),!and!gEh!corresponds!to! OctoberENovember!(OND).! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! SFig!2:!Cross!correlation!of!EMP!and!MLD!anomalies!in!different!seasons!from!ACCESSE KPP!model.!Left!panel!shows!concurrent!correlations!and!the!right!panel!represents! corresponding!oneEmonth!lag!correlations,!where!MLD!lags!behind!oneEmonth!of!the! forcing.!Figure!aEb!shows!the!lag!correlation!during!JanuaryEMarch!(JFM),!cEd!represents! AprilEJune!(AMJ),!eEf!shows!the!lag!correlation!during!JulyE!September!(JAS),!and!gEh! corresponds!to!OctoberENovember!(OND).!