Multimodel Ensemble Prediction Systems for Short and Medium Range Weather Forecasting

Abstract

The main objective of this thesis is to improve the weather forecast skill over India in the newlineshort and medium range using Multi-Model Ensemble (MME) techniques. The problems newlineaddressed in this thesis are: (i) Deterministic Numerical Weather Prediction (NWP) newlinemodel bias estimation and calibration (ii) Development of MME Prediction Systems for newlinequantitative rainfall forecast. newlineWeather significantly affects the health, safety, and economic well-being of everyone. newlineWeather forecast in medium range is extremely useful, as 3-5 days lead time is very newlineessential for taking precautionary measures to save crops from bad weather in agriculture newlineand managing resources in other sectors. Operational forecasting of quantitative rainfall newlineremains a challenging task. Producing accurate, unbiased and user specific forecast has newlinebecome a crucial challenge for meteorologist. Increasing dependence on automated newlineweather forecast from NWP models highlight the individual model bias for various newlineweather events. The MME techniques discussed in this thesis address the weakness and newlinebias of various NWP models for different weather events to provide reliable forecast for newlineshort and medium range time scales. newlineThe challenge of producing reliable and informative weather forecasts of future state is at newlinethe forefront of research in meteorology. This has generated many fascinating areas of newlineresearch in atmospheric and oceanic sciences. For day to day operational weather newlineforecasts, one has to depend on NWP model outputs. The prediction skill of NWP newlinemodels is still not adequate to satisfactorily address detailed aspects of Indian summer newlinemonsoon. This is because of incomplete understanding of the monsoon mechanism of rainfall and some inherent limitations of deterministic NWP models. The deterministic newlinemodel forecast is based on some initial conditions, which neglect small scale effects and newlinealso approximate complicated physical processes and interactions....