Sensitivity of convective and land surface parameterization in the simulation of contrasting monsoons over CORDEX-South Asia domain using RegCM-4.4.5.5

Abstract

An effort for testing the capability and suitability of different cumulus and land surface parameterization schemes in simulating the Indian summer monsoon (ISM) and associated features during contrasting monsoon years using a regional climate model framework is presented. The latest Regional Climate Model system (hereafter RegCM, Vn4.4.5.5) developed and managed by the International Centre for Theoretical Physics (ICTP) has been used to downscale the ERA-Interim reanalysis dataset for three different ISM years, namely, normal (1990), deficit (1987), and excess (1988). Three different convection schemes, namely, Grell (GFC), Emanuel (MIT), and mixed type (i.e., Grell over land and Emanuel over ocean (MIX)), have been used for the simulation of ISM and its associated features. Moreover, three different land surface schemes, namely, the Biosphere–Atmosphere Transfer Scheme or BATS (control), subgrid disaggregation of BATS (SUB-BATS), and Community Land Model 4.5 (CLM4.5), have also been tested for their performance. With these combinations, 27 different suites of experiments were simulated at a horizontal resolution of 50 km over the COordinated Regional Climate Downscaling Experiment (CORDEX)-South Asia domain. The validation of experiments was performed against the India Meteorological Department (IMD) and Climatic Research Unit (CRU) daily dataset for precipitation and near surface air temperature, respectively. In light of different performance metrics, it is found that the near surface air temperature and precipitation in the seasonal-scale simulations are well represented in space for most of the experiments with inherent biases. Among all the experiments, the SUB-BATS scheme in association with GFC cumulus scheme is found to perform consistently among all the years in simulating the ISM precipitation. In most of the cases for the simulation of precipitation, the SUB-BATS seems to outperform the control and CLM4.5 experiments. The domination of dry bias in CLM4.5 experiments is attributed to the improper representation of mean sea level pressure, weaker ISM circulation, and higher atmospheric stability, thus inhibiting the convective activity. Compared to other convection schemes, the MIT scheme is more suitable in representing the features of ISM simulation while using CLM4.5 as a land surface scheme. However, the GFC scheme was found suitable for the simulation of different precipitation years while using SUB-BATS as a land surface model. Such simulations portray reduced biases and improved spatial patterns compared to other schemes. Additionally, CLM4.5 experiments display more utility for the simulation of near surface air temperature. A more comprehensive process-based approach is envisaged to investigate the performance of CLM4.5 in the simulation of ISM for climate-scale simulations.