Abstract

Aim/Background
Accurate estimation of reference evapotranspiration (ETo) is crucial for effective irrigation scheduling and water resource management. However, the universal application of the Penman-Monteith FAO-56 method is often limited by insufficient or low-quality meteorological data. This study addresses these limitations by developing and evaluating a fuzzy logic-based model for estimating ETo in the Ogun-Osun River Basin (OORB).

Methods
A 35-year dataset of four meteorological variables (temperature, relative humidity, wind speed, and net radiation) was utilized. The actual ETo was computed using the Penman-Monteith FAO-56 method. Data preprocessing and wrangling were conducted using Python programming. A fuzzy inference system was implemented, employing membership functions and a rule-based system to model the complex relationships between meteorological variables and ETo. Model performance was evaluated using statistical metrics, including root-mean-square error (RMSE), sum-of-square error (SSE), mean-percentage error (MPE), and coefficient of determination (R²).

Results
The sensitivity analysis revealed that relative humidity was the most influential parameter (0.99 ± 0.01), followed by temperature (0.98 ± 0.02), net radiation (0.55 ± 0.1), and wind speed (0.003 ± 0.001). The fuzzy logic model demonstrated strong predictive performance, with R² values ranging from 0.9526 to 0.9775, RMSE values between 0.05167 mm/day and 0.1261 mm/day, MPE between –0.0039% and 0.00175%, and SSE ranging from 0.2136 mm²/day² to 1.2728 mm²/day², indicating low prediction errors.

Conclusion
The findings validate the effectiveness of fuzzy logic in climate modeling, particularly for estimating ETo under conditions of data scarcity. The developed fuzzy logic model provides a reliable alternative to traditional methods and can be applied in similar climatic regions to enhance water resource management and irrigation planning.

Key words: reference evapotranspiration, machine learning, climate change, FAO56-PM, modelling