Regression analysis between mean daily intensity, rainy days and seasonal rainfall in normal, excess and deficient years: A case study

摘要

This paper is a part of research work on the study of degradation of lakes/tanks in Hunsur taluk of Mysore district, Karnataka, India. The study deals with the regression analysis of daily rainfall data for the period 1975-2014. The main objective of the study is on the variation of rainfall intensity, to establish a relationship between Rainy Days (RND), Mean Daily Intensity (MDI) versus Seasonal/Annual Rainfall. It is seen that, 3rd order polynomial curve fitting is a better fit due to the erratic fluctuation of rainfall in the region. The regression is compared with linear, logarithmic, exponential and power regression curve fitting. Least values of Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE) analysis of the residue also suggest 3rd order polynomial curve fitting as best fit. Analysis of intensity of rainfall is carried as per Indian Meteorological Department (IMD). It is observed that, very light rain and light rain intensities have significantly increased in the recent years while rather heavy and heavy rain intensities have decreased gradually. Although the average annual rainfall is normal in the region, mean daily intensity is in the decreasing trend and the number of rainy days is in the increasing trend. Hence, surface runoff may have reduced in the catchment, which is one of the factors affecting the drying of lakes/tanks in the region.