Analysis and forecasting of rainfall trends in semi-arid Bundelkhand region of Madhya Pradesh, India: using statistical methods

The present study is mainly concerned with one of the crucial climatic variables, i.e., precipitation for analyzing the trend of rainfall in five districts (Chhatarpur, Damoh, Panna, Sagar, Tikamgarh) of the Bundelkhand region (a semi-arid region) of Madhya Pradesh. In this region, agriculture is the single most important activity in sustaining livelihoods, and though several irrigation schemes are being operated, the quality of irrigation services is poor and most of the cultivated land is still dependent on rainfall. Hence, examining the temporal variation of rainfall is vital and it will help to assess climatic-induced changes and suggest feasible adaptation strategies. Statistical analysis techniques like the Mann–Kendall test, Sen’s slope estimator, the MGCTI “Bertin matrix” and climate extreme indices (CDD, R95p and RX1Day) were used to estimate the seasonal, monthly and annual rainfall trend and its variability for the period from 1951 to 2018. The Autoregressive Integrated Moving Average Model (ARIMA) has been used for the forecasting of annual rainfall in this study area from 2019 to 2050. The Results show that there is no significant trend in the annual rainfall pattern, though in every case it is negative (Sen slope). The summer and winter season of each district had above 75% of CV, highlighting the extremely high variability of precipitation. The Regional index showed that the second half of the study period was more drought prone than the previous one, indicating a clear symptom of climate change. The annual decrease in rainfall from the ARIMA model (2019–2050) for the region is 7.60 mm per year, though spatial variation is also observed from ARIMA predictions.