Mathematical and Statistical Tools for Location Specific Weather Forecast in Medium Range up to 5 days using Numerical Weather Prediction NWP Model Output

Abstract

The main theme of the thesis is: quot Mathematical and Statistical tools for location specific weather forecast in medium range (up to 5-days) using Numerical Weather Prediction (NWP) model outputquot. First of all a brief description of need for forecasting weather in medium range, numerical weather prediction (NWP) models and interpretation and calibration techniques for getting the location specific forecast is given in the introductory chapter. The significance of the problem to the society and country is also explained. Forecasting in medium range (3-5 days) requires the use of meteorological data from all over the globe to capture the movement and genesis of weather systems in medium range. This data is used in the operational NWP models. To run the operational models super computers are used. To obtain the forecast in medium range (3-5days), the global NWP models termed as general circulation models (GCM) are used. The surface weather parameters like rainfall and Maximum/Minimum temperatures are the basic crucial parameters required for planning and applications in different sectors. The forecasts from the operational general circulation models (GCM) for these weather parameters does not show good skill when obtained for a specific location. Whatever is the skill attained, these forecasts needs to be interpreted and calibrated in order to have a desirably high skill , so as to give higher confidence to the users for using the location specific forecasts for planning and application activities. The interpretation and calibration techniques uses the basic mathematical and statistical tools like step wise linear regression and Kalman filter.