CLIMATIC VARIABILITY STUDIES OVER PARTS OF INDIA WITH FOCUS ON HYDRO-CLIMATIC VARIABLES

Abstract

India is a developing country facing tremendous challenges to sustain its fast growing economic growth with the threat of global warming. Global warming or climate change may affect the livelihood of the most vulnerable people living in the country by changing the quality and distribution of India’s natural resources. The assessment of climate related changes is most important across India. This can be examined by analysing the hydro-climatic datasets. Therefore, the present work has been taken up to investigate the changes in hydro-climatic variables over parts of the country. The objectives of the present study can be broadly categorised as follows: (i) Analysis of recent temperature and rainfall datasets to understand the climate induced changes over parts of India. (ii) Frequency analysis of extreme rainfall and streamflow series considering non-stationarity. The analysis of temperature datasets cover entire India, while the rainfall analysis covers only east, west, north and central zones of India. Trend and homogeneity in annual and seasonal temperature data of 125 stations for 1941 to 2012 period has been investigated. The Mann-Kendall trend detection test, Theil and Sen’s trend slope estimator, Cumulative deviation test, Standard normal homogeneity test and Wilcoxon Rank-Sum test are used for identification of trends and homogeneity in the datasets. The annual average, maximum and minimum temperatures show a rising trend at the rate of 0.44ºC/100 years, 0.51ºC/100 years and 0.19ºC/100 years respectively. All the seasonal temperature variables show a rising trend. Only the minimum temperature of the monsoon season shows the falling trend at the rate of 0.05ºC/100 years. The monsoon maximum temperature depicts a maximum increase at the rate of 0.80ºC/100 years. The homogeneity analysis shows break years around 1972, 1974 and 1977 for the annual average, maximum and minimum temperatures respectively for India. The seasonal analysis shows that the minimum temperatures of winter and monsoon season with the maximum temperature of the post-monsoon season are homogeneous in nature. The magnitude of a trend in all annual and seasonal temperature variables changed significantly iv after the break year. The results of the regional analysis also show similar trends as observed in the case of entire India. Rainfall is one of the most important factors of Indian economy. The estimation of spatio-temporal trends of rainfall on a regional basis will help in understanding the global impact of climatic systems over the region. The Von Neumann ratio, Hurst’s coefficient, Mann-Kendall trend and Spearman’s rank correlation tests are used for identification of short-term dependence, long-term dependence, trends and the correlation between rainfall of 148 stations with climate change and climate variability indices. Only one-fourth of the station rainfall and rainy day datasets exhibit short-term and long-term dependences on annual and seasonal basis. Most parts of the study area exhibit a decrease of 0 to 200 mm rainfall over the 1951 to 2007 period in annual and monsoon season. In non-monsoon season, the decrease is 0 to 100 mm. The number of rainy days on annual and seasonal basis also show a decrease of 0 to 10 days during the analysis period. Global average temperature and NINO3.4 SST anomaly have indicated good correlation with rainfall indices over the analysis period. Trends in multi-datasets of gridded rainfall also follow the similar pattern. The CRU and GPCC gridded datasets in general do not follow the spatial patterns shown by other gridded datasets. Frequency analysis of annual maximum rainfall (AMR) data across India and the streamflow series of Narmada river basin has been carried out considering both stationarity and non-stationarity assumptions under recent climatic conditions using GAMLSS model. The annual maximum rainfall datasets of 239 stations, which are well distributed across India are used to see the effect of climate change (global average temperature, GAT), climate variability (NINO and IOD) and local temperature change (RAT) on rainfall quantiles of 25, 50 and 100 years return periods. The results of frequency analysis show that the Gumbel, Lognormal and Generalised Gamma are the most useful distributions for the non-stationarity case. The high values of climate indices, GAT and RAT generally provide higher peaks for annual maximum rainfall and vice versa. Frequency analysis of annual maximum discharge of six stations of Narmada river basin of India is carried out under stationary and non-stationary conditions using external covariates, like, GAT, NINO, IOD and reservoir index (RI). The results exhibit that in comparison with the non-stationary model; the stationary model either overestimates or underestimates the design flood quantiles. The non-stationary model can also capture the effect of natural v variability of any flood process. The results of non-stationary model confirm that the climate variability and construction of reservoirs have a sublime effect on flood process of Narmada river basin. The Hoshangabad station shows a significant influence of reservoir in annual maximum discharge compared to the other stations. The non-stationarity in annual maximum discharge is mainly due to the influence of reservoirs for three out of six stations, while other three stations are showing non-stationarity due to the effect of natural climate variability across the Narmada river basin.