Evaluation of CMIP5 GCMs Multimodel Ensemble Precipitation over Pakistan

Abstract

This study provides an overview of the performance of the state-of-the-art global climate models of Coupled Model Intercomparison Project Phase 5 (CMIP5) in representing the present climate precipitation over Pakistan against the reanalysis dataset AgMERRA. The analysis have been done over the time period 1980-2010. The overall performance of the models are summarized by a heat map diagram based on RMSE’s w.r.t observed dataset. The results show that ENSMEAN and ENSMEDIAN models outperform individual models in all months. EC-Earth show better performance for winter months and GFDL-ESM-2M, GFDL-ESM-2G, GFDL-CM3 for summer along with MRI-CGCM3 for all months. A seasonality index has also been calculated for comparing the representation of seasonal cycle with the reanalysis dataset. MIROC5 and EC-Earth has relatively better performance over Pakistan. Box and whisker analysis show that for winter (JFM) months CMIP5 has relatively better performance as compared to the reanalysis dataset while there is difference of up to 1.5mm/day for the rest of the months. There is large variability between the models for summer months (JJAS).