Evaluation of wintertime extreme temperature indices simulated for East Asia from a quantitative approach

Abstract:

Climate extremes, one of the seven WCRP grand challenges, have been projected to occur more frequently or severely with the continuing global warming. To obtain a confidence level of the projections, evaluation of the extremes simulated with state-of-the-art climate models is required in parallel with the projections. On the global scale, the assessments on CMIP3 and CMIP5 models for various climate extreme indices have been conducted while on a regional scale, analyses on extremes have been focused on heavy rainfall or drought in different regions. This study evaluates the extreme temperature indices in January from 1979 to 2005 simulated with 25 CMIP5 and 15 CMIP6 models against ERA-interim reanalysis over East Asia (20-60°N, 90-160°E). For the evaluation, we applied skill-score metrics and relative error among the model results. Overall, the model performance for the spatial variability and that for the temporal variability appears to be indifferent, showing a high correlation coefficient greater than 0.9 for the climatological mean spatial patterns. As expected, the best model among the CMIP6 models shows a better performance than that among the CMIP5 models in all indices averaged over East Asia. Although other results may change depending on the other CMIP6 models that will become available sooner or later, this result will remain the same.