Simulation of present-day precipitation over India using a regional climate model

Abstract

The objective of the present paper is to examine the capability of the regional climate model version 3 (RegCM3) to simulate the annual as well as seasonal precipitation variability over the Indian subcontinent. RegCM3 has been run at 40 km horizontal resolution for the period of 1982–2006 continuously and model results were compared to the observed precipitation datasets of India Meteorological Department (IMD) and CPC Merged Analysis of Precipitation (CMAP). Model evaluation has been done using different statistical methods like mean bias error (MBE), root mean square error, mean percentage error (MPE) and studied the spatial pattern of annual and seasonal variability and trend. Daily precipitation data at 1° × 1° grids of IMD have been used to study observed climatological means (both annual and seasonal), regression trends, interannual and intraseasonal variability over India from 1951 to 2007. The spatial distribution of annual precipitation shows a decreasing trend over west coast of India, central India, hilly region of India and an increasing trend is found over the northwest India, peninsular India and northeast India. The temporal distribution of daily precipitation shows highest rainfall of 18 mm/day in mid July (in composite flood cases only) and 12 mm/day during August (in composite drought cases only). The RegCM3 simulated annual and seasonal precipitation variability is close to the observed IMD and CMAP over all India (AI). During winter and pre-monsoon season, the model has overestimated the mean precipitation while underestimated in summer and post-monsoon season. Overall, annual precipitation showed the deficiency of −22.44 % compared to IMD and −1.41 % compared to CMAP over India. To understand the possible cause of annual and seasonal precipitation biases over India and its six homogeneous regions, the vertical difference (model mines National Centre for Environmental Prediction; NCEP) fields of water vapor mixing ratio (WVMR) and air temperature from surface to top level (1σ–18σ level) have been studied in detail. The model produced more moisture at the lower level during winter (JF) and pre-monsoon (MAM) season and reduced WVMR at the same levels during summer (JJAS) and post (OND) monsoon over India. High moisture content at the lower level simulated in the model can explain the increased precipitation during JF and MAM season and vice versa during JJAS and OND. Similar results were found over the homogeneous regions of India. This analysis indicates that the RegCM3 has more transport of vertical WVMR from surface to mid layer of the troposphere compared to NCEP.