Evaluation of bias free rainfall forecasts and Kalman filtered temperature forecasts of T-80 model over Indian monsoon region

Abstract

A forecasting system for obtaining objective medium range location specific forecast of surface weather elements is evolved at National Centre for Medium Range Weather Forecasting (NCMRWF). The basic information used for this is the output from   the general circulation models (GCMs) T-80/T-254 operational at NCMRWF. The most essential component of the system is Direct Model Output (DMO) forecast. This is explained in brief.  Direct Model Output (DMO) forecast is obtained from the predicted surface weather elements from the GCM. The two important weather parameters considered in detail are rainfall and temperature. Both the weather parameters  have biases. While the bias from the rainfall is reduced by adopting bias removal technique based upon  threshold values for rainfall and for removing bias from temperature forecast a two parameter Kalman filter is applied. The techniques used for getting bias free forecast are explained in detail. Finally an evaluation of the forecast skill for the  Kalman filtered temperature forecast and  bias free rainfall forecast during monsoon 2007 is presented.