Prediction and Trends of Rainfall Variability over Bangladesh

Abstract

Rainfall is one of the most common natural disasters in Bangladesh which rigorously affect agro-based economy and people’s livelihood in almost every year. The main objective of this study is to examine the variation, prediction and trend of rainfall in Bangladesh. The data for this study have been extracted from the Bangladesh Meteorological Department (BMD). Data used in this study were collected from 31 rain gauge stations located in different parts of the country for a period of 40 years (1975-2014). Linear regression model is used to understand the variation, trend and prediction of rainfall for annual and various climatic seasons such as pre-monsoon, monsoon, post-monsoon and winter. We also estimated mean rainfall with standard deviation of pre-monsoon, monsoon, post-monsoon and winter. Finding reveals that, the trends of mean rainfall of annual, pre-monsoon and winter have decreased, whereas rainfall remained unchanged in monsoon season and has increased in post-monsoon. Data predicts lesser rainfall in the period 1975, 1989, 1992, 1994, 2004, 2009, 2012, 2013 and 2014 years. These results indicate lesser precipitation in future over Bangladesh. The predicted rainfall amount from the best fitted model was compared with the observed data. The predicted values show reasonably good result. Thus the model can be used for future rainfall prediction. It is expected that this long term prediction will help the decision makers in efficient scheduling of flood prediction, urban planning, and rainwater harvesting and crop management. Classification of rainfalls in a systematic way is therefore critical in order to take necessary actions toward drought mitigation and sustainable development.