Time series modelling of drought in a semi-arid region of south Gujarat, India

Gujarat is situated within a climatic spectrum ranging from arid to semi-arid, characterised by periodic and recurrent drought phenomena coupled with an enduring and consistent challenge of water scarcity. Drought events in the Gujarat region place substantial stress on its water resource infrastructure, thereby impacting not only the hydrological aspects but also exerting notable repercussions on the intricate interplay between agriculture, economics, and the broader societal domain. Despite the implementation of numerous legislations and protocols aimed at mitigating the impact of drought and ameliorating the aftermath, the state of Gujarat experiences a recurring pattern of drought, resurfacing approximately every three years and inflicting substantial disruption upon the lives of its residents. This research focuses on analysing the current drought scenarios using the standard precipitation index (SPI) and meteorological drought modelling using autoregressive integrated moving average with exogenous variable (ARIMAX) for identifying future drought conditions, which is a statistical modelling approach. Analysis of reference locations was carried out to identify the best model and the same model used for drought modelling and forecasting in the remaining locations. A meteorological drought risk map and drought frequency map were prepared using forecasting results. Model validation was done by computing RMSE and R-2 of multiple locations in the entire region.