Assessment of the Indian summer monsoon in the WRF regional climate model

Abstract

The performance of the regional climate model, Weather Research and Forecasting in simulating the three dimensional moist and thermodynamic structure of Indian summer monsoon (ISM) during 2001–2011 is examined in this study. The model could simulate monsoon elements and convective precipitation zones over ISM region with some overestimation. Statistical analysis of sub-regional precipitation indicates that model has better skill over the monsoon core region with correlation of 0.7 and root mean square error of 2.3 mm day−1 with respect to observations. The model simulated seasonal mean vertical structures of temperature and water vapour mixing ratio (WVMR) are consistent with the Atmospheric Infrared Sounder observations. However, the core of low level jet is shifted southward in the model due to unrealistic convective heating over the lower latitudes of Indian Ocean and southern peninsular India. The tropical easterly jet is confined to 15°N in the model, which is due to the midtropospheric cold bias over the Tibetan region. The meridional asymmetric bias of sea level pressure (SLP) in model leads to weaker vertical wind shear, limiting the northward migration of maximum rain band to south of 23°N. These discrepancies have marked effects on the proper simulation of monsoon climate. The large scale spatial patterns of SLP, precipitation and winds during active and break spells are well simulated by the model. The lead-lag evolution of vertical structure of model temperature shows baroclinic structure during the active phase. It is evident from the observations that enhanced (suppressed) convection is generally preceded by a low-level moist (dry) anomaly and followed by a low-level dry (moist) anomaly. The model is inadequately representing the temporal evolution of vertical moist and thermodynamic processes. The evolution of vertical structures of temperature and WVMR is better simulated in the break phase compared to that of active phase. The evolution of cyclonic vorticity in the model is different from the observations during the convective phase. In short the model has limitations in representing convectively unstable regimes. It is anticipated that these findings will significantly contribute to the regional climate model assessment programs.