Annual Rainfall Forecasting Using Artificial Neural Networks with Stochastic Modeling Approach

India remains an agrarian society, with agriculture contributing around 13.7% to the national GDP and employing nearly 50% of the workforce. Rainfall plays a pivotal role in sustaining agriculture by irrigating fields and replenishing rivers as well as groundwater reserves. Consequently, understanding rainfall patterns is vital for the country’s economic growth and overall welfare. Accurate rainfall forecasting not only supports better agricultural planning but also aids in flood management and disaster preparedness. Artificial Neural Networks (ANNs) provide a promising approach for predicting monthly rainfall by capturing the cyclical nature of weather systems. This technique utilizes historical time-series data, making it less sensitive to shifts in underlying climate models, including anthropogenic climate change. In this study, ANNs are applied to forecast monthly rainfall, where rainfall data for each of the twelve months is used to predict values for the subsequent year. The neural network is trained using the gradient descent algorithm, and its performance is assessed through four key metrics: Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Deviation (MAD), and Mean Absolute Percentage Error (MAPE). Experimental results indicate that ANNs can effectively forecast monthly rainfall trends for the period 2025 to 2030.