TITLE:
Prediction of Monthly Rainfall in Togo Using Artificial Neural Networks (ANN)
AUTHORS:
Lamboni Batablinlè, Lawson Latévi, Serge Dimitri Bazyomo, Ablam L. Afanou, Lare Yendoubé, Djibib Zakari, Magolmeena Bannaa, Lawin Agnidé Emmanuel
KEYWORDS:
Artificial Neural Network (ANN), Monthly Rainfall, Rainfall Projection, Togo, Seasonal Variability
JOURNAL NAME:
Open Journal of Applied Sciences,
Vol.15 No.9,
September
24,
2025
ABSTRACT: This study evaluates the performance of a feed-forward Artificial Neural Network (ANN) model in simulating historical monthly rainfall in Togo (1981-1985, 1986-1990, 1991-1996, and 1996-2000) and projecting rainfall patterns for 2026-2030. The model reliably captures the seasonal progression and north-south gradient, with low RMSE and high R2 values in most periods. Historical results reveal early-season overestimation in northern and central regions (+15% to +45%), mid-season peak biases in southern and central zones (+25% to +50%), and late-season underestimation in the north (up to −75%). Projected rainfall for 2026-2030 indicates a relatively dry scenario in northern Togo, an irregular minor rainy season in the south, and notable spatial variations across regions. These findings demonstrate the model’s capacity to reproduce both historical and future rainfall dynamics, while highlighting uncertainties related to ITCZ positioning, coastal monsoon influence, and localized convective events, supporting its application in regional rainfall simulation, water resource management, and climate adaptation planning.