Trend analysis and forecast of rainfall over India
Student name: Ms Sumana Sarkar
Guide: Dr Ashoke Basistha
Year of completion: 2011
Host Organisation: TERI University
Abstract: Rainfall is a very important climatic variable for a country like India where the primary source of
income is agriculture, 70 percent of which is rain fed. The changing climate scenario has a major
impact on the meteorological parameters like rainfall, temperature, humidity etc, also potent to bring
changes on global scale.
In this study the major focus was on the rainfall pattern with special interest on the Pre monsoon
rainfall and its changing behaviour over India under the influence of changing climate. Through the
study the trend in the Pre monsoon rainfall over the country was analysed which is the precursor to
the monsoon rainfall.
Over the years several studies have highlighted the importance of climatic phenomenon prevailing
during the Pre monsoons over the entire climatic regime of the country. In a physiographically
diverse country like India the effect of these phenomenon on the climatic variables demand more
attention.
The current work used 61 years continuous rainfall data from the IMD for the period 1949 – 2009.
Non parametric test like Pre whitened Mann Kendall Test was used to find the trend in the seasonal
and annual rainfall. A detailed study of the Pre monsoon months was also carried out which showed
a significant rising trend ( =5%) over the stations falling in the north and north – west parts of
India and falling trend over the western coastal stations which are manifestations of rising
convective heating and land sea thermal over the nation. The magnitude of trend in the form of
percentage change from mean, for all the months was also computed using Theil and Sen's median
slope estimator. Statistical ARIMA model was fit over all the station and a forecast of all the stations
using ARMA technique was done for the period 2010 – 2029.
Keywords: Pre monsoon, Pre whitened Mann Kendall Test, Theil and Sen's median slope, Auto
Regressive Moving Average (ARIMA), Forecast.
