* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download CSIRO_CCAM Model_Methodology_FNL
German Climate Action Plan 2050 wikipedia , lookup
2009 United Nations Climate Change Conference wikipedia , lookup
Soon and Baliunas controversy wikipedia , lookup
Heaven and Earth (book) wikipedia , lookup
Global warming hiatus wikipedia , lookup
ExxonMobil climate change controversy wikipedia , lookup
Effects of global warming on human health wikipedia , lookup
Global warming controversy wikipedia , lookup
Climatic Research Unit email controversy wikipedia , lookup
Climate resilience wikipedia , lookup
Michael E. Mann wikipedia , lookup
Numerical weather prediction wikipedia , lookup
Fred Singer wikipedia , lookup
Climate change denial wikipedia , lookup
Climate change adaptation wikipedia , lookup
Global warming wikipedia , lookup
Instrumental temperature record wikipedia , lookup
Politics of global warming wikipedia , lookup
Economics of global warming wikipedia , lookup
Climate change feedback wikipedia , lookup
Climate engineering wikipedia , lookup
Climate change in Tuvalu wikipedia , lookup
Effects of global warming wikipedia , lookup
Climatic Research Unit documents wikipedia , lookup
Citizens' Climate Lobby wikipedia , lookup
Climate governance wikipedia , lookup
Climate change and agriculture wikipedia , lookup
Atmospheric model wikipedia , lookup
Carbon Pollution Reduction Scheme wikipedia , lookup
Climate sensitivity wikipedia , lookup
Media coverage of global warming wikipedia , lookup
Climate change in the United States wikipedia , lookup
Solar radiation management wikipedia , lookup
Attribution of recent climate change wikipedia , lookup
Scientific opinion on climate change wikipedia , lookup
Effects of global warming on humans wikipedia , lookup
Public opinion on global warming wikipedia , lookup
Climate change and poverty wikipedia , lookup
Climate change, industry and society wikipedia , lookup
Surveys of scientists' views on climate change wikipedia , lookup
CSIRO OCEANS AND ATMOSPHERE High-resolution climate projections for the Philippines: Methodology Jack Katzfey 14 December 2015 GEF Grant No. TF096649 Citation Katzfey JJ (2015) High-resolution projections for the Philippines: Methodology. CSIRO, Australia. Copyright and disclaimer © 2015 CSIRO To the extent permitted by law, all rights are reserved and no part of this publication covered by copyright may be reproduced or copied in any form or by any means except with the written permission of CSIRO. Important disclaimer CSIRO advises that the information contained in this publication comprises general statements based on scientific research. The reader is advised and needs to be aware that such information may be incomplete or unable to be used in any specific situation. No reliance or actions must therefore be made on that information without seeking prior expert professional, scientific and technical advice. To the extent permitted by law, CSIRO (including its employees and consultants) excludes all liability to any person for any consequences, including but not limited to all losses, damages, costs, expenses and any other compensation, arising directly or indirectly from using this publication (in part or in whole) and any information or material contained in it. Acknowledgment GEF Grant No TF096649 of the Philippine Climate Change Adaptation Project (PhilCCAP) funded this workshop. 1 Introduction Climate change has been recognized as one of the greatest challenges facing our planet, not only for the environment, but also for economic development, with changes occurring in the physical, ecological and socioeconomic systems. Likely outcomes are changes to weather patterns and sea-level rise, with impacts on ecosystems, water resources, agriculture, forests, fisheries, industries, urban and rural settlements, energy usage, tourism and health (IPCC, 2013). The Philippines are located in South East Asia, with a tropical monsoon climate and a coastline of more than 35000 km. It is one of the most disaster-prone countries in the world, with most of the disasters related to weather and climate. Consequently, climate change and climate variability are likely to pose increasing threats to its inhabitants in the near and long-term future. Regional climate models (RCMs) have become important tools for the prediction of climate variability and change in the regions of southern and eastern Asia. In a study using the Conformal Cubic Atmospheric Model (CCAM) nested within the 2.5 degree NCEP reanalyses, Nguyen and McGregor (2009) demonstrated that the model could simulate the main features of the Asian monsoon, a major influence on weather and climate in the Philippines, including seasonal shifts of the precipitation throughout the year. To assist the Philippines in its efforts to better understand the impacts of climate change and prioritise its adaptation measures, the detailed climate change projections at 10 km resolution produced for the High-resolution Climate Projections for Vietnam (HCPV) project (Katzfey et al., 20141) were extracted for the Philippine region. The stretched-grid of the downscaling model used in the HCPV project allowed for the extraction of the projections over the Philippines (see section 2.3 for more description of the downscaling process and grid). To address the inherent uncertainty in future climate change projections, a range of global climate models (GCMs) and emission scenarios were used for the downscaled simulations, as well as analysis techniques such as ensemble statistics to ensure that a broad range of plausible changes to the climate was evaluated. The aims of this project were to: Incorporate new climate science information released by the Intergovernmental Panel on Climate Change. Improve understanding of potential future climate changes in the region with reduced uncertainty. Integrate past and current research for a more complete assessment of the potential effects of climate change. Produce regional climate change projections to help identify the people and sectors at risk at the local level, where most impacts are felt. Further climate science research and build capacity in the Philippines. Provide information necessary for appropriate planning and investment to adapt to climate change. Data analysed for the project were based on output from six of the latest available GCMs from the Coupled Model Intercomparison Project Phase 5 (CMIP5) that were selected on the basis of their ability to realistically capture current climate and climate features such as El Niño-Southern Oscillation (ENSO) that have a large influence on the region. The global data was dynamically downscaled using a stretched-grid RCM to produce high-resolution simulations for current and future climate. Simulations were performed for historical (1970-2005) and future (to 2099) time periods using two representative concentration pathways: RCP4.5 and RCP 8.5 (Meinshausen et al. 2011). 1 Project data and reports can be accessed at www.vnclimate.vn. The HPCV project was funded by Australia’s Department of Foreign Affairs and Trade (DFAT) and carried out by partners from the Institute of Meteorology, Hydrology and Climate Change (IMHEN) and the Hanoi University of Science – Vietnam National University (HUS) in Vietnam and the Commonwealth Scientific and Industrial Research Organisation (CSIRO) in Australia. Section 2 gives details of the downscaling methodology for the ensemble simulations of future climate for the Philippines. Section 2.1 gives describes the procedures used for selection of the six GCMs to be dynamically downscaled to finer resolution. Section 2.2 describes the methodology used to correct the biases of the GCM Sea Surface Temperatures (SSTs) before they are used to drive the RCMs, followed by a description of the two-step downscaling procedure in Section 2.3. A summary is presented in Section 3. 2 Methodology This section describes the methods used to provide high-resolution climate projection information for the Philippines. GCMs provide the best available tools for simulating large-scale future climates based on various greenhouse gas and aerosol emission scenarios, since they are able to couple atmosphere and ocean systems and incorporate their complex linked interactions over the entire Earth system. However, their resolution (approximately 100-200 km) is too coarse to capture regional impacts of climate change, especially in areas of complex topography, coastline and land use, and for areas that are smaller than the grid box size of the models, such as islands, where local effects and land-sea interactions are of great importance. For this reason, the GCM data was dynamically downscaled to higher resolution using CCAM. The two approaches to downscale GCM data are dynamical or statistical. Dynamical downscaling, sometimes called regional climate modelling, uses an atmospheric model forced with inputs from the global model to simulate the regional climate. Statistical downscaling typically applies relationships between large-scale variables and the local climate, derived from the current climate, to downscale from the global models. Using dynamical downscaling, all variables are consistent with each other. With statistical downscaling, the relationship between variables may not be preserved and the approach assumes that the statistical relationships used do not change in the future. In this project only the dynamical downscaling approach is used. In order to capture the range of possible futures, multiple GCMs are downscaled to provide ensemble climate simulations. This attempts to address the variations in climate simulations due to variations in each model’s internal dynamics and parameterisations and helps to capture a broader range of plausible climate futures. The process of producing the high-resolution simulations for the Philippines is outlined below. All the global and regional climate simulations were driven by observed changes in greenhouse gases and aerosols until 2005 and then a range of possible scenarios were used until 2099. Some GCM simulations included direct and indirect effects of aerosols, some included ozone depletion, and some included volcanic aerosols and solar forcing (IPCC, 2013). 2.1 GCM selection for downscaling Downscaling the results of all available CMIP5 GCMs is computationally too expensive at present. However, to address the inherent uncertainty in climate projects due to differing model physics and parameterisations, as well as assumptions made about climate and human behaviour and to capture a plausible range of the climate projections, ensemble simulations are needed. For this project, data from a subset of six different GCMs provide a compromise between computational costs and the provision of a range of different climate change signals that arise from modelto-model variations. To determine which of the 24 GCMs available from CMIP5 at the time of this study to downscale, their performance at simulating current climate was ranked, based on three criteria: Ability to capture observed spatial patterns and trends of atmospheric variables such as mean sea level pressure, temperature and precipitation; Ability to simulate oceanic features such as ENSO, which have a large impact on climate variability; Preference for models that capture the most probable/realistic range of changes in climate variables such as temperature and precipitation, to address the uncertainty inherent in climate modelling. Since only the change signal is used in these simulations, it is desirable to have different amounts and patterns of warming in the SSTs from the GCMs chosen. A sample of the warming patterns projected by the GCMs chosen for the end of the century with RCP 8.5 is presented in Figure. 1. The six GCMs chosen to be downscaled to fine resolution for this project, along with an assessment of their strengths and limitations, can be found in Table 1. 2.2 Correction of GCM SSTs GCMs simulate the global climate reasonably well, but still have biases and do not simulate inter-annual variability in the atmospheric and oceanic system (e.g. ENSO) realistically (IPCC, 2013). To improve the dynamical downscaling results, the SSTs from the GCMs can be corrected before being used by RCMs. CCAM, which requires only input of SSTs and sea ice concentrations (SIC) from the host GCM can use this technique. Limited area models cannot easily use this approach since they require lateral boundary atmospheric data from GCMs. The deficiencies or biases of GCMs, due to differing model configurations and physics, if not corrected before downscaling can cause unrealistic behaviour of the RCM simulations and thereby affect the reliability of the climate projections. In previous downscaling simulations with CCAM, a simple correction of the climatological monthly means was applied to GCM SSTs (Katzfey et al., 2009; Nguyen et al., 2012). This method preserved the inter-annual (year-toyear) variation of GCM SSTs, and therefore any errors in variability of SSTs from GCMs, such as those related to ENSO, were also imposed upon the downscaled simulations. Therefore, a newly-developed method that corrects both the mean and interannual variability of monthly GCM SSTs was applied in this project. As a result, the regional model will have SSTs with no climatological bias for the current climate and with more realistic amplitude and spatial pattern of the interannual variability. Consequently, while the climate change signals of the GCMs are preserved in the downscaled simulations, the location of the ENSO variability, particularly over the tropical Pacific, is more realistic. However, it should be noted that the frequency of the ENSO variability simulated by the GCMs was not adjusted. The GCM SSTs are corrected to match the Optimum Interpolation SST dataset Version 2 (Reynolds et al., 2007), which contains daily SST and SIC data on a 0.25 longitude by 0.25 latitude grid for the period 1982 (September) to 2011. The dataset is based on measurements conducted by NOAA’s polar orbiting Advanced Very High Resolution Radiometer (AVHRR) meteorological satellites2 in combination with buoy data and ship measurements. Since the GCM data are available as monthly averages, both SST and SIC daily data are averaged monthly for the computation of the bias correction. Additionally, prior to correction the model SSTs and SIC are interpolated to match the observation grid. In the bias correction method, the following steps are conducted for each month of the year separately (see Katzfey et al., 2009, Katzfey et al., 2014 and Hoffmann et al., 2015 for more details of the bias correction procedure): 1. The SST data are de-trended for 30-year backward running trends. 2. The standard deviation (SD), which is a measure of the year-to-year variability of the de-trended 30year period, is calculated. 3. The monthly anomalies are corrected using the ratio of the observed SD to the model SD. 4. Monthly biases are then calculated and subtracted, giving the bias- and variance-corrected SSTs. Problems due to differences between SIC distribution in the GCMs and in the observations are avoided in the bias correction by reducing the correction near the ice edges as a function of SIC. In addition, the correction is linearly reduced from the equator to 50N and 50S. This means that the full variance correction is applied at the equator, but no variance correction is applied north of 50N or south of 50S. This step is necessary to avoid problems arising from the possible shift of strong ocean currents such as the Gulf Stream in a future climate. When omitting this step, the change signal can be quite different between corrected and uncorrected SSTs for those regions. As an example, Figure 2 shows the uncorrected and corrected July SSTs from the ACCESS1.0 GCM as well as the observed SSTs in the tropical Pacific for the period 1982-2011. As intended, the GCM SSTs are very close to the observed SSTs after the correction. Also the SD is very similar to the observations, especially close to the equator. 2.3 Dynamical downscaling methodology A range of possible methodologies can be used to downscale global climate information (see for example McGregor, 1997 and Katzfey, 2013). In this study, one regional model was used to dynamically downscale global climate model 2 See http://noaasis.noaa.gov/NOAASIS/ml/avhrr.html for more information on AVHRR satellites. information. The following section describes the method used to produce the dynamically-downscaled simulations. The process had two steps: 1. Global simulations with initial input from six CMIP5 GCMs (see section 2.1 for details of the selection process) were completed using CCAM with a uniform grid. 2. These global simulations were dynamically downscaled to fine resolution, using the RCM CCAM with variableresolution in order to produce simulations of the current climate or future climate at regional scale. The process is known as dynamical downscaling, because the equations on which the model is based explicitly simulate atmospheric dynamical and thermodynamical processes. This is in contrast to other methods, such as statistical ones, that are based on statistical relationships between large-scale and small-scale variables. CCAM, a global model, requires only SST and SIC data from the GCMs as inputs to drive the model. The resolutions of the GCM and RCM grids, emission scenarios used, and time periods for the various simulations varied due to technical and computer resource constraints. For details of the set-up of the downscaled simulations, see the following sections and Error! Reference source not found.. 2.3.1 CONFORMAL CUBIC ATMOSPHERIC MODEL (CCAM) CCAM is a variable-resolution global atmospheric model that has been developed at CSIRO (McGregor 2005b; McGregor and Dix 2001, 2008). It includes a fairly comprehensive set of physical parameterisations to represent subgrid scale atmospheric processes that are not directly simulated by the model. The updated GFDL parameterisations for long-wave and short-wave radiation (Schwarzkopf and Ramaswamy 1999; Freidenreich and Ramaswamy, 1999) are employed, with interactive cloud distributions determined by the liquid- and ice-water scheme of Rotstayn (1997). The simulations also include the scheme of Rotstayn and Lohmann (2002) for the direct and indirect effects of sulphate aerosols. The model employs a stability-dependent boundary-layer scheme based on Monin–Obukhov similarity theory (McGregor et al., 1993). The CABLE biosphere-atmosphere exchange model is included, as described by Kowalczyk et al. (2006), having six layers for soil temperatures, six layers for soil moisture (solving the Richards equation), and three layers for snow. The cumulus convection scheme uses mass-flux closure as described by McGregor (2003), and includes downdrafts and detrainment. CCAM also includes a simple parameterisation to enhance SSTs under conditions of low wind speed and large downward solar radiation, affecting the calculation of surface fluxes. The dynamical formulation of CCAM includes a number of distinctive features. The model is non-hydrostatic, with two-time-level semi-implicit time differencing. It employs semi-Lagrangian horizontal advection with bi-cubic horizontal interpolation (McGregor, 1993; McGregor, 1996), in conjunction with total-variation-diminishing vertical advection. The grid is un-staggered, but the winds are transformed reversibly to/from C-staggered locations before/after the gravity wave calculations, providing improved dispersion characteristics (McGregor, 2005a). Threedimensional Cartesian representation is used during the calculation of departure points, and also for the advection or diffusion of vector quantities. CCAM may be employed in quasi-uniform mode or in stretched mode by utilising the Schmidt (1977) transformation. Further details of the model dynamical formulation are provided by McGregor (2005b). Step 1: CCAM at 50 km resolution For these simulations, CCAM was first set up on a C192 grid (with six panels each of 192 x 192 grid points) having a quasi-uniform horizontal resolution of about 50 km over the whole globe (Fig. 3, top) and 27 vertical levels. It was run for 130 model years (1970–2099) forced by the bias- and variance-corrected SSTs (as described in Section 2.2) from each of the six selected CMIP5 GCMs. From 1970 to 2005, historical values of greenhouse gases and aerosols were used. From 2006 to 2099, two sets of simulation were completed using the greenhouse gases and aerosol emissions specified by both the RCP 4.5 (moderate) and RCP 8.5 (high) emission scenarios. Step 2: CCAM at 10 km resolution The outputs from Step 1 were further downscaled using CCAM to 10 km resolution over Vietnam, utilizing the C96 stretched grid shown in Fig. 3 (lower plot) and Fig. 4. The land-sea mask, terrain and soil type are also shown in Fig. 4. To provide a further degree of consistency with the host CCAM simulation, a scale-selective digital filter (Thatcher and McGregor, 2009) was applied every 6 hours to replace selected broad-scale (with length-scale of about 8000 km) fields of the high-resolution CCAM simulation with the corresponding fields of the 50 km CCAM simulation. The filter was applied to the MSLP, moisture, temperature, and wind components above pressure-sigma level 0.9 (about 1 km above the surface). The model output was saved four times per day at 00, 06, 12 and 18 GMT. These data have been post-processed by interpolating them onto a 0.5 grid for the 50 km simulations, and a 0.1 grid for the 10 km simulations for easier interpretation. Many prognostic and diagnostic fields are available for impact assessment studies at local to regional scales. 3 Conclusions Six CMIP5 GCMs were dynamically downscaled using the stretched-grid RCM CCAM to about 25 km resolution over the Philippines from 1970-2099. Two representational concentration pathways were used: RCP 4.5 (moderate) and RCP8.5 (high). Results from these high-resolution downscaled simulations were used to produce the ensemble climate projections presented in Katzfey, 2015. The downscaling methodology used in this study is different to that typically used by other downscaling groups in that no atmospheric information from the GCMs was used. Instead, the SSTs from the GCMs were corrected so that the monthly mean SSTs match the observed SSTs. In addition, the interannual variance of SSTs in the GCMs was corrected to match the observed variance. These corrected SSTs and the SIC were then used to drive CCAM globally with an even resolution at 50 km resolution. The 50 km global simulation data were then used to spectrally nudge the atmosphere of a stretched-grid version of CCAM at high resolution. Simulations were performed from 19702005 using historical values of greenhouse gases and aerosol emissions. From 2005 to 2099, greenhouse gases and aerosol emissions as specified by both the moderate (RCP 4.5) and high RCP 8.5) representational concentration pathways. The downscaling approach used in this study has several advantages. First, by correcting the biases in the SSTs from the GCMs for the current climate, the downscaled simulations will provide a more realistic current climate. Secondly, by using a stretched grid RCM, issues related to the specification of lateral boundaries was avoided. Finally, by using a range of GCMs and two emission scenarios (RCPs), an ensemble of projections were produced in order to provide a range of possible futures. As with all climate projection, the uncertainty related to future greenhouse and aerosol concentrations is an important concern. Changes in technology, legislation, and human behaviour may have large impacts on future climate.. Since large ensemble projections capture a broader range of climate change signals, using only a limited set (six) of GCMs to downscale may limit the range of possible futures in the simulations. Use of only one RCM, with its inherent biases, may also limit the full range of possible futures to be considered. For this reason, the results of this study must be interpreted with understanding of the assumptions made in producing them. References Freidenreich, S.M., and V. Ramaswamy, 1999: A new multiple-band solar radiative parameterization for general circulation models. J. Geophys. Res., 104, 31,389–31,409 Hoffmann, P., J.J. Katzfey, J.L. McGregor, and M. Thatcher, 2015: The derivation of downscaling SSTs corrected for both bias and variance. Geophys. Res. Letters in preparation IPCC, 2013: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 1535 pp Katzfey, J.J., J.L. McGregor, K.C. Nguyen, and M. Thatcher, 2009: Dynamical downscaling techniques: Impacts on regional climate change signals. In MODSIM09 Int. Congress on Modelling and Simulation, www.mssanz.org.au/modsim09 13:2377-2383. ______, 2013: Chapter 15: Regional climate modelling for the energy sector. In A. Troccoli et al. (eds.), Weather matters for energy, DOI: 10.1007/978-1-4614-9221-4_15. ______, J.L. McGregor, and R. Suppiah, 2014: High-resolution climate projections for Vietnam: Technical report. CSIRO, Australia. 352 pp. http://www.hpsc.csiro.au/users/kat024/final%20summaries/VN_TechnicalReport_266pp_WEB.pdf. ______, 2015: Climate Scenarios for the Philippine Climate Change Adaptation Project (PhilCCAP). CSIRO, Australia. 110 pp. Kowalczyk, E.A., Y.P. Wang, R.M. Law, H.L. Davies, J.L. McGregor, and G. Abramowitz, 2006: The CSIRO Atmosphere Biosphere Land Exchange (CABLE) model for use in climate models and as an offline model. CSIRO Marine and Atmospheric Research Paper 13, 37pp. McGregor, J.L., 1993: Economical determination of departure points for semi-Lagrangian models. Mon. Wea. Rev., 121, 221-230. ______, 1996: Semi-Lagrangian advection on conformal-cubic grids. Mon. Wea. Rev., 124, 1311-1322. ______, 1997: Regional climate modelling. Meteor. Atmos. Phys., 63, 105-117. ______, 2003: A new convection scheme using a simple closure. In Current issues in the parameterization of convection, BMRC Research Report 93, 33–36. ______, 2005a: Geostrophic adjustment for reversibly staggered grids. Mon. Wea. Rev., 133, 1119-1128. ______, 2005b: C-CAM: Geometric aspects and dynamical formulation [electronic publication]. CSIRO Atmospheric Research Tech Paper 70, 43pp. ______, and M.R. Dix, 2001: The CSIRO conformal-cubic atmospheric GCM. In IUTAM Symposium on advances in mathematical modelling of atmosphere and ocean dynamics, P.F. Hodnett (Ed), Kluwer, Dordrecht, 197-202. ______, and M.R. Dix, 2008: An updated description of the Conformal-Cubic Atmospheric Model. In: High resolution simulation of the atmosphere and ocean, K. Hamilton and W. Ohfuchi (Eds), Springer, 51–76. ______, H.B. Gordon, I.G. Watterson, M.R. Dix and L.D Rotstayn, 1993: The CSIRO 9-level atmospheric general circulation model. CSIRO Div. Atmospheric Research Tech Paper 26, 89pp. Meinshausen, Smith, Calvin, Daniel, Lamarque, Matsumoto, Montzka, Raper, Riahi, Thomson, Velders, Van Vuuren. 2011. The RCP Greenhouse Gas Concentrations and their Extensions from 1765 to 2500. Climatic Change 109:1-2, p213-241. DOI 10.1007/s10584-011-0156-z. Nguyen K.C., J.J. Katzfey, and J.L. McGregor, 2012: Global 60 km simulations with CCAM: evaluation over the tropics. Clim. Dyn., 39, 637-654, 10.1007/s00382-011-1197-8. Reynolds, R.W., T.M. Smith, C. Liu, D.B. Chelton, K.S. Casey, and M.G. Schlax, 2007: Daily high-resolution blended analyses for sea surface temperature. J. Climate, 20, 5473-5496. Rotstayn, L.D., 1997: A physically based scheme for the treatment of stratiform clouds and precipitation in largescale models. I: Description and evaluation of the microphysical processes. Quart. J. R. Meteor. Soc., 123, 1227–1282. ______, and U. Lohmann, 2002: Simulation of the tropospheric sulfur cycle in a global model with a physically based cloud scheme. J. Geophys. Res., 107, doi:10.1029/2002JD002128. Schmidt, F, 1977: Variable fine mesh in spectral global model. Beitr. Phys. Atmos., 50, 211–217. Schwarzkopf, M.D., and V. Ramaswamy, 1999: Radiative effects of CH4, N2O, halocarbons and the foreignbroadened H2O continuum: A GCM experiment. J. Geophys. Res., 104, 9467–9488. Thatcher, M., and J.L. McGregor, 2009: Using a scale-selective filter for dynamical downscaling with the conformal cubic atmospheric model. Mon. Wea. Rev., 137, 1742-1752. Table 1. Summary of the strengths and limitations of the six GCMs selected for downscaling. See Katzfey et al., 2014 for more discussion of the selection process. GCM CCSM4 CNRM-CM5 Strengths - Good agreement with precipitation and temperature observations over South East Asia - ENSO pattern well reproduced - Limitations - Observed temperature trends poorly reproduced - Less realistic SST pattern in the tropical Pacific ENSO pattern well reproduced - Good agreement with precipitation observations over South East Asia Observed trends poorly reproduced over South East Asia - Too few ENSO events - Good agreement with observations globally and over Asia NorESM1-M - ENSO pattern and tropical Pacific SSTs well reproduced - Poor agreement with precipitation patterns over South East Asia ACCESS1.0 - SSTs in the Pacific well reproduced - - Observed temperature trends well reproduced Poor agreement with precipitation patterns over South East Asia - Good agreement with observations globally - ENSO pattern and SSTs in the Pacific well reproduced - ENSO variability not well reproduced - Good agreement with temperature observations over South East Asia - Good agreement with precipitation observations over South East Asia - Poor agreement with temperature patterns over South East Asia MPI-ESM-LR GFDL-CM3 Table 2. Details of the CCAM simulations used in this project, including resolution, number of levels, and years simulated. Model Resolution/ vertical levels CCAM50 50 km/L27 GCM data used Input data CNRM-CM5 Sea ice and variance and bias-corrected SSTs CCSM4 ACCESS1.0 NorESM1-M (IC: Initial condition) Years simulated Emission scenarios 1970-2099 Historical for 1970-2005 RCP 4.5 and 8.5 for 20062099 IC: 01 Jan 1979 ERA-Interim MPI-ESM-LR GFDL-CM3 CCAM10 10 km/L27 CNRM-CM5 CCSM4 ACCESS1.0 NorESM1-M MPI-ESM-LR GFDL-CM3 CCAM 50 km 1970-2099 Historical for 1970-2005 RCP 4.5 and 8.5 for 20062099 Figure. 1. SST changes in the tropical Pacific (°C) from the best-performing models for the months July-August by 2071-2099 compared with 1971-2000, based on the RCP 8.5 emission scenario. Starred GCMs are the ones chosen to downscale in this study. Figure 2. Long-term mean (top panel) and standard deviation (bottom panel) of July SSTs (°C) for the period 1982-2011 from (a, d; left) uncorrected ACCESS1.0 results, (b, e; middle) corrected ACCESS1.0 results and (c, f; right) observed data from the Optimum Interpolation SST dataset version 2 (Reynolds et al., 2007). Figure 3. CCAM grids used for the 50 km global (C192 grid; top) and 10 km (C96 grid; bottom) downscaled simulations over Vietnam (plotting every 2nd grid point). Figure 4. CCAM high-resolution model grid resolution (upper left), land-sea mask (upper right), terrain (lower left) and soil type index (lower right). CONTACT US FOR FURTHER INFORMATION t 1300 363 400 +61 3 9545 2176 e [email protected] w www.csiro.au Oceans and Atmosphere Jack Katzfey t +61 3 9239-4562 e [email protected] w http://people.csiro.au/K/J/Jack-Katzfey.aspx YOUR CSIRO Australia is founding its future on science and innovation. Its national science agency, CSIRO, is a powerhouse of ideas, technologies and skills for building prosperity, growth, health and sustainability. It serves governments, industries, business and communities across the nation.