Use of observed temperature statistics in rankingCMIP5model performance over the Western Himalayan Region of India

Abstract

ABSTRACTAssessing warming over the Western Himalayan Region (WHR) of India is challenging due to its limited station data availability and poor data quality. The missing values in the station data were replaced using the Multiple Imputation Chained Equation technique. Finally, 16 stations having continuous records during 1969–2009 were considered as the ‘reference stations’ for assessing the warming/cooling trends in addition to evaluate the Coupled Model Intercomparison, phase 5 (CMIP5), Global Circulation Model (GCM). Station data indicates winter (DJF) warming is higher and rapid (1.41 °C) than the other seasons and less warming was observed in the post‐monsoon (0.31 °C) season. Overall mean annual warming overWHRis ∼0.84 °C during 1969–2009. The performance of 34CMIP5models was evaluated based on three different criteria namely (1) mean seasonal cycle, (2) temporal trends and (3) spatial correlation between simulated and observed signals for common available period of 1969–2003 over the study area. Models are provided a final rank on the basis of the cumulative rank obtained in each of three approaches.CMCC‐CM,GISS‐E2‐HandMIROC 5are three top‐ranked models whileMIROC‐ESM,MIROC‐ESM‐CHEMand bcc‐csm1‐1 are three bottom‐ranked models over theWHR. The study also extended to judge whether the selected top‐ranked models perform well through two alternative data sources namely European Reanalysis (ERA)‐interim and Climate Research Unit (CRU), which have not used in the process of model evaluation. The spatial patterns of top‐rankedGCMare similar to the spatial pattern obtained throughERA‐interim andCRUwhile zoomed in toWHRbut bottom‐ranked models fail to reproduce such spatial patterns indicating the top‐rankedGCMswould offer more reliability for projecting future climate overWHR.