Applications of Neural Network in Long Range Weather Forecasting

Long-range weather prediction over high-resolution geographical regions like districts or subdivisions is very complicated. it has been found that these statistical models are not satisfactory and it is impractical to study the highly non-linear relationships between climate and its predictors over the districts or subdivisions. We developed ANN model for long-range weather forecasting of smaller Indian region like districts and subdivision and spatial interpolation of mean climate variable. It has been observed that internal dynamics of meteorological time series data are easily identified by these models and have provided excellent results especially in pattern recognition and prediction. The capability of ANN model in spatial interpolation of mean climate variable of 102 raingauge stations of Chhattisgarh state has also been identified. The model in interpolation has produced excellent results, enforcing strong relation between dependent (climate variable), and independent variables (latitude, longitude, altitude). An application entitled AGRIMETCast has been developed deploying the outcome of the research. The entire development process and outcomes are presented in this book.

Autorentext

Associate Professor,Bhilai Institute of Technology,Durg,C.G.,India. Karmakar did B.Sc. MCA., PhD in Computer & Information Technology.He has published 60 papers in National/International Conferences/Journals. He has published 4 books. He is also IEEE reviewer, Editor, Subject Expert, PhD supervisor.Currently Four PhD scholars are working under him.