Selection of GCMs for the projection of spatial distribution of heat waves in Pakistan

Performance of 31 General Circulation Models (GCMs) of Coupled Model Intercomparison Project Phase 5 (CMIP5) was assessed according to their ability to reconstruct the different properties of heat waves (HWs); HW frequency, HW duration and HW index estimated using Princeton Global Meteorological Forcing (PGF) daily temperature data for the period 1961 to 2005 in order to generate an ensemble for the projection of HWs in Pakistan. The GCMs were selected based on three criteria: (1) ability to replicate the decadal variability in HW properties, (2) ability to reconstruct the spatial distribution of HW properties based on Taylor skill score, (3) replicate the annual time series of HW properties based on standard statistical indices and compromise programming. Results revealed four GCMs: CCSM4, CESM1(BGC), CMCC-CM and NorESM1-M are the most suitable for the projection of HWs over Pakistan. Projection of HWs using the selected GCMs revealed increase in the frequency and severity of HWs in most parts of Pakistan for both the radiative concentration pathway (RCP4.5 and RCP8.5) scenarios used in the study. The frequency of HWs was projected to increase up to 12 events per year while the duration was projected to increase up to 100 days in a year during 2060 to 2099 for the highest emission scenario. Overall, the HWs were projected to be more frequent and longer duration in the east and the southern coastal regions.