Fine-Tuning Extreme Rainfall Predictions – A Machine Learning Approach Hardcover

This book presents primarily the results of in-house research of the authors in the fields of numerical weather prediction, extreme events, and machine learning. The focus of the book is on improving the accuracy of short-range predictions for extreme rainfall events in India, such as monsoons and cyclones, which can cause significant damage. The accuracy of any numerical weather prediction model depends on both the representation of atmospheric physics and the initial conditions. This book traces a step-by-step approach to calibrating model parameters that critically influence the outcomes using machine learning and enhancing initial conditions through data assimilation. The book also provides a beginnerfriendly introduction to the basics of numerical weather prediction. It delves into advanced techniques, such as Sobol sensitivity analysis, adaptive surrogate modeling-based calibration, and hybrid ensemble-variational data assimilation. Python codes for the algorithms are included in the text, making it a valuable resource for researchers in the field. The methodology proposed in this book can be adapted to other meteorological events and related fields like turbulence modeling, hydrology, agricultural meteorology, and, with modifications, to any general problem in science and engineering concerning large-scale computations of boundary and initial value problems. Contents 1 Introduction: Extreme Rainfall Events in India 2 Review of Literature: Prediction of Extreme Rainfall Events in India 3 A Brief Overview of the Weather Research and Forecasting (WRF) Model 4 An Overview of Events Simulated, Methodology, and Evaluation Metrics 5 Identification of Sensitive Model Parameters for Heavy Rainfall Events During Indian Monsoon 6 Calibration of Sensitive Model Parameters for Heavy Rainfall Events During Indian Monsoon 7 Impact of Data Assimilation on Parameter-Calibrated Model for Heavy Rainfall Events During Indian Monsoon 8 Identification of Sensitive Paramete.