Performance evaluation of solar radiation equations for estimating reference evapotranspiration (ETo) in a humid tropical environment

Solar radiation (Rs) is an essential input for estimating reference crop evapotranspiration, ETo. An accurate estimate of ETo is the first step involved in determining water demand of field crops. The objective of this study was to assess the accuracy of fifteen empirical solar radiations (Rs) models and determine its effects on ETo estimates for three sites in humid tropical environment (Abakaliki, Nsukka, and Awka). Meteorological data from the archives of NASA (from 1983 to 2005) was used to derive empirical constants (calibration) for the different models at each location while data from 2006 to 2015 was used for validation. The results showed an overall improvement when comparing measured Rs with Rs determined using original constants and Rs using the new constants. After calibration, the Swartman–Ogunlade (R2 = 0.97) and Chen 2 models (RMSE = 0.665 MJ∙m−2∙day−1) performed best while Chen 1 (R2 = 0.66) and Bristow–Campbell models (RMSE = 1.58 MJ∙m−2∙day−1) performed least in estimating Rs in Abakaliki. At the Nsukka station, Swartman–Ogunlade (R2 = 0.96) and Adeala models (RMSE = 0.785 MJ∙m−2∙day−1) performed best while Hargreaves–Samani (R2 = 0.64) and Chen 1 models (RMSE = 1.96 MJ∙m−2∙day−1) performed least in estimating Rs. Chen 2 (R2 = 0.98) and Swartman–Ogunlade models (RMSE = 0.43 MJ∙m−2∙day−1) performed best while Hargreaves–Samani (R2 = 0.68) and Chen 1 models (RMSE = 1.64 MJ∙m−2∙day−1) performed least in estimating Rs in Awka. For estimating ETo, Adeala (R2 =0.98) and Swartman–Ogunlade models (RMSE = 0.064 MJ∙m−2∙day−1) performed best at the Awka station and Swartman–Ogunlade (R2 = 0.98) and Chen 2 models (RMSE = 0.43 MJ∙m−2∙day−1) performed best at Abakaliki while Angstrom–Prescott–Page (R2 = 0.96) and El-Sebaii models (RMSE = 0.0908 mm∙day−1) performed best at the Nsukka station.