HR: 08:40h
AN: A51F-05    [Abstracts]
TI: Present-day climatic simulations run with two GCMs: a comparative evaluation against ERA- 40 reanalysis data
AU: * Marras, S
EM: simone.marras@bsc.es
AF: Barcelona Supercomputing Center - Centro Nacional de Supercomputacion (BSC-CNS), Earth Sciences Division, C/ Jordi Girona, 29, Barcelona, 08034, Spain
AU: Jimenez, P
EM: pedro.jimenez@bsc.es
AF: Barcelona Supercomputing Center - Centro Nacional de Supercomputacion (BSC-CNS), Earth Sciences Division, C/ Jordi Girona, 29, Barcelona, 08034, Spain
AU: Jorba, O
EM: oriol.jorba@bsc.es
AF: Barcelona Supercomputing Center - Centro Nacional de Supercomputacion (BSC-CNS), Earth Sciences Division, C/ Jordi Girona, 29, Barcelona, 08034, Spain
AU: Perez, C
EM: carlos.perez@bsc.es
AF: Barcelona Supercomputing Center - Centro Nacional de Supercomputacion (BSC-CNS), Earth Sciences Division, C/ Jordi Girona, 29, Barcelona, 08034, Spain
AU: Baldasano, J M
EM: jose.baldasano@bsc.es
AF: Barcelona Supercomputing Center - Centro Nacional de Supercomputacion (BSC-CNS), Earth Sciences Division, C/ Jordi Girona, 29, Barcelona, 08034, Spain
AU: Baldasano, J M
EM: jose.baldasano@bsc.es
AF: Environmental Modelling Laboratory, Technical University of Catalonia, Diagonal, 647, Barcelona, 08028, Spain
AB: Global circulation models (GCMs) are the best currentently available tools to describe the complexity of atmospheric variability and climate evolution on a global scale. However, the large number of existing models shows a wide spectrum of approaches and results, as shown by the intercomparison project of the Intergovernmental Panel on Climate Change (IPCC); or, e.g. Garcia-Herrera et al. (2006), Schmidt at al. (2006), Hansen et al. (2007) and Garcia et al. (2007), among others. Therefore, this work compares side-by-side results from the ModelE GCM version of the NASA Goddard Institute for Space Studies (GISS) at 2º x 2.5º horizontal resolution and 20 vertical layers and the NCAR Whole-Atmosphere Community Climate Model (WACCM) at 2º x 2.5º resolution against reanalysis products of the European Centre for Medium-Range Weather Forecasts (ERA-40) available at 2.5º x 2.5º horizontal resolution and 23 pressure levels (obtained by the ECMWF three-dimensional assimilation system based on satellite, radiosondes and other conventional observations). Both GISS ModelE and WACCM were implemented in a parallel high- performance computing infrastructure, the Marenostrum supercomputer. Model outputs are available for different decades (GISS ModelE simulations were run from 1950 to 2050 while results from WACCM cover the period 1950-2003), but comparison focuses on the period 1957-2002, conditioned by the availability of the ERA-40 re- analysis. The main aim of this study is the definition of the degree of reliability of two specific models for their use in climate prediction and to analyze their seasonal behavior in several regions. Statistical comparisons are performed for global and regional averaged distribution maps of sea level pressure, 2m-temperature, geopotential heights and precipitation at several time scales. The Root Mean Square Error (RMSE) and bias with respect to ERA-40 data have also been estimated and diagrammed for the years 1957- 2002 evolution. Averaged values are then analyzed for different seasons and regions. Moreover, through the use of Taylor diagrams, we quantify and discuss the different performances for each model simulated patterns in terms of standard deviation, centered RMSE and their correlations. Discrepancies between reanalysis and models emerge widely through the study, especially for total cloud cover and precipitation. The global behavior of both models is accurate when compared to ERA-40 re-analysis in terms of low bias and RMSE for 2m-temperature computed by both models, and very high correlations appear (greater than 0.95). Given that the WACCM and the GISS simulations were run at specified sea surface temperatures (SST), in the oceanic regions this result must be expected. However, in terms of sea-level pressure larger differences emerge, with RMSE as large as 0.45 (normalized value) and correlation coefficients in the range between 0.8 and 0.9. In extremely complex topography regions such as the Himalaya range, large differences in error and tendencies appear for pressure and geopotential heights for both models. Despite such problems, a detailed analysis of the results coming shows that both GISS ModelE and WACCM results are within the range of error estimates found in the revised scientific literature.
DE: 0550 Model verification and validation
DE: 1616 Climate variability (1635, 3305, 3309, 4215, 4513)
DE: 3309 Climatology (1616, 1620, 3305, 4215, 8408)
DE: 3337 Global climate models (1626, 4928)
SC: Atmospheric Sciences [A]
MN: 2007 Fall Meeting