Sensitivity of land surface parameterization on Regional Spectral Model forecasts

Experiments were carried out to study the impact of different land surface schemes on a Regional Spectral Model (RSM) forecasts. RSM is based on the perturbation method of NCEP, where the dependent variables are the differences between the regional and global model fields called 'perturbations'. The perturbation method ensures the use of global model values as the base fields all over the domain and predicts the mesoscale features embedded in the base field forecasts. The first version of RSM has a land surface scheme with a single layer of soil moisture, which is the same as the operational global model with which it is nested. The second version of RSM has a land surface scheme with two layers of soil moisture and a more complex treatrnent of evaporation. The model was integrated for five days nested with the operational global spectral model during August 2001. The RSM with 2-layer soil moisture scheme was found to have slightly less easterly bias over north India. However, the two-layer scheme showed higher evaporation and precipitation over Andhra Pradesh region. Additionally, major differences were also observed in all the components of the surface energy balance over the same region.