Change Point Analysis of Air Temperature in India

Abstract

Change point analysis was performed on air temperature at different pressure levels in the Indian subcontinent to identify the time at which a major change in trend, if any, has occurred. Pettit test, a nonparametric test to identify change points in a time series was used for this purpose. It tests, the null hypothesis that the variable follows one or more distributions that have the same location parameter against the alternate hypothesis that a change point exists. The significance of the change point is determined and if it is greater than the considered level of 90% confidence, then, the change point is considered to be significant. The test was performed on the surface temperature data of the Indian subcontinent for the period 1949–2014, obtained from the NCEP/NCAR reanalysis data set at a resolution of 2.5°. The results of the test for the dry period indicate that the southern, northern and northeastern parts of India exhibited a significant change point in the nineteen seventies. During the wet season and the southwest monsoon season, a significant change was observed in the southern, central and eastern parts of India in the last decade. Analysis of the annual mean temperature revealed that a significant change point occurred in South India in the last decade.