Trend analysis of rainfall in four meteorological subdivisions of southern India using nonparametric methods and discrete wavelet transforms

摘要

In recent times, trend analysis and change point detection in hydroclimatic variables receiving significant attention due to climate change and its socioeconomic consequences. In this study, long-term trends of rainfall in four subdivisions of southern India namely Kerala, Tamil Nadu, North Interior (NI) Karnataka and Telangana regions are analysed using linear regression, nonparametric Mann-Kendall (MK) test and Sen's slope estimator methods. Trend analysis of annual rainfall time series shows an increasing trend in three subdivisions - Tamil Nadu, NI Karnataka and Telangana, and a decreasing trend in Kerala subdivision. Further the sequential change in trend of annual and seasonal rainfalls in the four subdivisions is conducted using sequential MK (SQMK) method. The SQMK analysis shows an early divergence of progressive and retrograde modes of post-monsoon rainfall of Kerala and winter rainfall of Telangana subdivisions. Further it is observed that among different seasonal rainfalls, the post-monsoon rainfall of Kerala subdivision shows a statistically significant increase in the recent past. Then the trend analysis based on discrete wavelet transform (DWT) in conjunction with SQMK method is performed on the post-monsoon rainfall time series of Kerala subdivision, and the results show that there is a dominancy of short-term periodicity of less than a decade in the subdivision.