A Machine Learning-Based Approach to Assess and Predict Drought Events: A Case of Rajasthan, India

The work focuses on understanding and analyzing drought events in the region of Rajasthan, India, using drought indices and machine learning models. The Standard Precipitation Index (SPI) and Standardized Precipitation and Evapotranspiration Index (SPEI) are calculated to assess drought severity, duration, and intensity. The study area includes two stations, Jodhpur and Banswara, with different climatic characteristics. Various hydro-climatic factors such as precipitation, temperature, humidity, and solar radiation are considered to analyze their correlation with drought indices. The machine learning models, long short-term memory (LSTM) networks, and random forest (RF) are employed for drought prediction. The results indicate that LSTM performs better than RF in short-term predictions, with lower root mean square error (RMSE) values. The study provides valuable insights into drought monitoring, characteristics, and prediction, contributing to effective drought mitigation strategies in the region.