Performance of an ensemble of CORDEX-SA simulations in representing maximum and minimum temperature over the Himalayan region

Abstract

The performance of an ensemble of 10 regional climate model (RCM) experiments in simulating the seasonal climatology of maximum and minimum near-surface temperature (Tmax and Tmin, respectively) over the Himalayan region is studied. These simulations are carried under Coordinated Regional Climate Downscaling Experiments-South Asia (hereafter, CORDEX-SA) project. The purpose of the study is to evaluate through various statistical methods, the ability of models to consistently simulate the observed Tmax and Tmin across the ensemble, space and seasons for the present climate (1971–2005). The performance varies from one model to other as well as individually with a season and region specific response. Due to the fine resolution feature of RCMs, the simulations capture the distribution of temperature over the Himalayan region very well which is exhibited by their very high correlation (> 0.8) with observation in all the seasons. The magnitude of temperature, however, is highly underestimated in a form of a large cold bias which is consistently found across all the experiments and region. Nonetheless, in the case of Tmin, warm biases also exist, which is evident in eastern parts of study region. An amplification of cold bias with elevation is noticed in the distribution of Tmin as well as Tmax. An identical spatial pattern of temperature and its bias is found in experiments that involve a particular RCM (in the cases of CCAM and RegCM) but forced with different global climate models (hereafter, GCMs). The simulated rate of warming is greater than that of the observation in all the seasons with winter warming rate being the greatest.