Development and Application of Enhanced Grey Forecasting Model for Prediction of Annual Rainfall in Zamfara State, Nigeria

Forecasting agricultural production in Nigeria is challenging due to limited data and the inherent volatility of climatic factors like rainfall. Hence, the objective of this paper was to investigate the Development and application of enhanced grey forecasting model for prediction annual rainfall in Zamfara State, Nigeria using standard procedures involving the modification. The improved GM (1,1) model is further refined by integrating an Artificial Neural Network (ANN) to correct forecast residuals, resulting in a hybrid Grey Artificial Neural Network (GANN) model. The models were trained on ten years of historical rainfall data. Results demonstrated high predictive accuracy, with the GM (1,1) model achieving a Mean Absolute Percentage Error (MAPE) of 7.49% and the hybrid GANN model achieving a superior MAPE of 6.86%. Both models fall within the "high" accuracy grade according to standard metrics. The study concludes by providing rainfall forecasts for Zamfara State from 2021 to 2030, establishing the proposed models as effective and practical tools for climate-driven agricultural planning in data-scarce environments.