9-Impact of Satellite Derived Land Surface Data in Regional Climate Simulations Over India.

Abstract

It is well established that mesoscale models have a pronounced skill in reproducing regional climate as compared to global models. The objective of this study is to assess the impact of improved land surface representation on the simulation of Indian monsoon climate 1998 in a mesoscale model. The Fifth-Generation National Centre for Atmospheric Research (NCAR) / Penn State Mesoscale Model (MM5) is used for the present study. To see the impact of satellite derived land surface data such as vegetation type (VT), vegetation fraction (VF), and surface soil moisture (SM) on modeled meteorological fields during the July 1998 episode of Indian summer monsoon, MM5 simulations are conducted in two way nesting mode using Noah land surface scheme. We have performed five experiments (EXP1, EXP2, EXP3, EXP4 and EXP5) using different land surface data. In EXP1 we used the VF and VT of United States Geological Survey (USGS) and SM from National Centre for Environmental Prediction (NCEP). The EXP2 is with VT derived from SPOT satellite and other land surface parameters same as EXP . In EXP3 we used VT and VF derived from SPOT and SM same as EXP1. EXP4 uses SM derived from Spectral Sensor Microwave Imager (SSMI) and other surface parameters same as EXP 1 . EXP5 is with VT and VR from SPOT and SM from SSMI. The results of different simulations are compared with one another and with NCEP reanalysis and India Meteorological Department (IMD) observations. Results show that even though none of the experiment simulates the diurnal cycle of 2m-air temperature correctly, EXP5 shows improved diurnal cycle, particularly over western and southern part of India. All the experiments fail in simulating the monthly climatology of LHF. RMS error in simulated 2m-specific humidity is comparatively less in EXP5 for most of the regions. Simulated rainfall by EXP5 lies closer to observed rainfall, particularly over central and southern part of India. In genera 1, incorporation of satellite data in the MM5 model initial condition improved the model simulations of monsoon regional climate, but significant improvement is not observed while different satellite data are incorporated individually.