Site adaptation of NASA-POWER reanalysis solar radiation products for tropical climates in Ghana using bias correction for clean energy application

Abstract

Solar radiation measurements are needed to simulate a diverse range of clean energy application systems for optimization. However worldwide, whereas radiometric measurement stations are scarce and mostly sparsely distributed, satellite and reanalysis products are available, though with resolution problems. Therefore in the absence of in-situ measurements, we developed a bias-corrected long-term solar radiation database for Ghana in West Africa, using reanalysis products retrieved from NASA’s Prediction of Worldwide Energy Resource (POWER) archives available at \(0.5^\circ \times 0.5^\circ\) spatial resolution by Cumulative Distribution Function (CDF) matching technique, with reference to a synthetic solar radiation database developed from sunshine duration measurements at 22 stations, in order to site-adapt the reanalysis products. From the results, the POWER products were improved by overall minimization in mean residual difference of -1.02–1.4 ± 0.07 kWhm− 2day− 1 to -0.85–1.2 ± 0.12 kWhm− 2day− 1 after bias correction, with a 9%, 27%, and 17% increase in smaller RD ranges of -0.5–0.5, -0.2–0.2, and − 0.1–0.1 kWhm− 2day− 1 respectively. Also, the Mean Bias Error (MBE) range minimized from − 0.51–0.6 ± 0.3 kWhm− 2day− 1 to -0.35–0.2 ± 0.1 kWhm− 2day− 1 with 17 stations showing residual differences in correlation coefficients greater than zero, indicative of higher correlation after bias correction. Overall, spatiotemporal characteristics are preserved, and the improvement is in acceptable margins for engineering application.