Improving the accuracy of solar radiation estimation from reanalysis datasets using surface measurements. (October 2021)

• Record Type: Journal Article
• Authors: Tahir, Zia ul Rehman; Amjad, Muhammad; Asim, Muhammad; Azhar, Muhammad; Farooq, Muhammad; Ali, Muhammad Junaid; Ahmad, Syed Uzair; Amjad, Ghulam Murtaza; Hussain, Afkar
• Is Part Of: Sustainable energy technologies and assessments. Volume 47(2021)
• Publication Date: 2021-10
• Subjects: Solar radiation -- Reanalysis datasets -- Bias correction -- MERRA-2 -- ERA-5; Renewable energy sources -- Periodicals; Energy development -- Technological innovations -- Periodicals; Electric power production -- Periodicals; Energy storage -- Periodicals; 333.79
• DOI: 10.1016/j.seta.2021.101485
• Languages: English
• ISSNs: 2213-1388

Abstract

Highlights: Evaluation of solar radiation from six reanalysis datasets is performed. The range of rMBE of hourly GHI from MERRA-2 and ERA-5 is 8–19 and 6–13 % respectively. The rMBE of KT for MERRA-2 and ERA-5 range 9–21 and 8–16 % respectively. Bias correction improves reliability of reanalysis data of solar radiation. The best bias correction model reduces rMBE and rRMSE to 100 % and up to 60 % respectively. Abstract: This paper assesses accuracy of solar radiation from different generation reanalysis datasets against ground measurements for different climatic conditions. The range of rMBE for NCEP-DOE, CFSR, MERRA, MERRA-2, ERA-Interim and ERA-5 is 4–34, 12–24, 14–29, 8–19, 5–14, 6–13% respectively, using sub-daily data. The performance of newer versions of the datasets is better than previous versions. The performance of datasets varies in different months depending upon seasonal and monthly meteorological conditions. The datasets perform better under clear-sky conditions and estimations are poor for overcast conditions. The evaluation of three datasets was performed under different sky conditions. The rMBE using hourly data under clear-sky, intermediate-sky and cloudy-sky range −3–2, 4–17 and 134–170% respectively. The performance of MERRA-2 is better than ERA-5 under clear-sky and intermediate-sky whereas ERA-5 performs better under cloudy-sky conditions for region under study. A method based on clearness index, solar zenith angle and periodicity factor is developed to correct hourly GHI estimated by datasets. Twenty models were used to correct data considering monthly variation of errors. The rMBE and rRMSE were reduced to 100% and upto 50% for months having lower clearness index and higher errors in the original data.