Going Beyond Traditional Methods: Using LSTM Networks to Predict Rainfall in Kerala

India has witnessed a notable upsurge in floods owing to shifts in global climatic patterns underlining climate change and global warming. This inorganic change is significantly evident in places like Kerala and Tamil Nadu. Kerala receives perennial rainfall throughout the year, influenced by southeast and northwest monsoon rainfall cycles, resulting in flood-induced calamities like landslides prevailing in the Thrissur and Munnar regions of Kerala. Long-Term Short Memory (LSTM) networks outperform pre-existing deep learning forecasting models as it is optimal for time series forecasting by handling non-linear spatiotemporal dynamics, and adapting to long-term dependencies in time series data while ensuring high scalability. The objective of the present work is rainfall prediction in flood-prone regions in Kerala. We propose an LSTM network to predict the monthly rainfall in Thrissur, Pathanamthitta, Munnar and Kottayam. The proposed model is evaluated using the metrics like mean absolute error and root-mean-squared error.