Numerical study of western disturbances over Western Himalaya using mesoscale model

Abstract

A non-hydrostatic version of the Penn State University - National Center for Atmospheric Research (PSU-NCAR), US, Mesoscale Model (MM5) is used to simulate the characteristic features of the Western Disturbances (WDs) occurred over the Indian region during winter. In the present study sensitivity eight experiments are carried out by using two planetary boundary layer schemes, viz., Blackadar and Hong-Pan, and four convection parameterization schemes, viz., Kuo, Grell, Kain-Fristch and Betts- Miller, with 60 km horizontal model resolution. And also the role of horizontal model resolution and topography is studied by carrying out six experiments based on two factors: horizontal model resolution of 30 km, 60 km and 90 km with assumed no topography and normal topography. For this study two active WDs are chosen which yielded extensive precipitation over western Himalayas. WD from 18 to 21 January 1997 is chosen for study one and WD from 20 to 25 January 1999 is chosen for experiment two. National Center for Environmental Prediction – National Center for Atmospheric Research (NCEP-NCAR), US, reanalyzed data is used for initial and boundary conditions. It is found that the performance of combination of the Hong-Pan and Betts-Miller as planetary boundary layer and cloud convection parameterization schemes respectively is best compared to the other combinations of schemes used in this study. The model physics could able to simulate sea level pressure better with this combination as compared to the combinations with other schemes. Further, WD simulations with assumed no topography shows lesser amount of precipitation compared to WD simulations with normal topography. When normal topography is included, intense localized of precipitation was observed along the Himalayan range. Model integrations of precipitation fields are found close to the corresponding verification analysis. Sensitivity studies of precipitation field shows that finer domain (30 km) of the model simulation gives better results.