Trend and Change point Analysis of Extreme Temperature over India using Non-Parametric Methods and Empirical Mode Decomposition

Abstract

This paper presents the trend and change-point analyses of extreme temperature (Tmax and Tmin) datasets of annual and seasonal series of seven homogeneous regions of India for the period 1901-2007. The study used Mann-Kendall (MK) test for detection of trend, Cumulative Sum (CUSUM) test for detection of change point and Empirical Mode Decomposition (EMD) for extracting non-linear trend of all the 70-time series. The results of the MK test showed that at all the seven regions except North-West, the Tmax series of the winter season showed a significant increase at a 5 % significance level. The MK test detected a significantly increasing trend on annual and all seasonal Tmax series of West Coast (WC) and North East (NE) regions. The CUSUM test detected a change point within 1975-77 for the minimum temperature series of the East Coast (EC) region for all the seasons except monsoon, which is in agreement with the well-debated climate shift of the 1976-77 period. The test detected a change point in 1950 for the Tmax series of the winter season in all homogeneous regions except northwest (NW). The study also found that changepoint year estimated in the non-linear trend fitting by EMD may differ from that based on statistical estimations