Modeling the Chance Dynamics of Evapotranspiration with Some Climatic Variables as Covariates

Over the years, authors have focused on various methods of computing evapotranspiration undermining the stochastic nature of its distribution and the effect of atmospheric variables (covariates). This work employed the binary logistic regression model in modeling and predicting the chance dynamic of evapotranspiration using maximum and minimum relative-humidity, maximum and minimum air temperature, solar irradiance and wind speed as covariates. The result shows that; the model was able to classify correctly 89.4% of high evapotranspiration, 91.5% of low evapotranspiration and an overall 90.4% correct classification. The chance of high evapotranspiration occurring in Kano is higher than low evapotranspiration for a unit rise in any of the covariates except minimum relative humidity. Findings from this study clearly show that logistic regression model can predict evapotranspiration very efficiently.