Improving the National Solar Radiation Data Base using PSM V4

作者：Manajit Sengupta, Yu Xie, Aron Habte, Brandon Benton, Paul Edwards, Jaemo Yang, Michael P. Foster, Andrew K. Heidinger · 年份：2025 · DOI：10.1109/pvsc59419.2025.11133215 · 被引用次数：1 · 研究领域：Solar Radiation and Photovoltaics

The National Solar Radiation Database (NSRDB) provides comprehensive global solar resource data at a high temporal and spatial resolution. The NSRDB employs satellite-based solar modeling to retrieve cloud properties and subsequently compute solar radiation. The other input parameters—including aerosol optical properties, precipitable water vapor, surface albedo, temperature, and pressure—are also employed by the model. The NSRDB data were recently updated using physical solar model (PSM) version 4, which includes an improved gap-filling algorithm, based on machine-learning techniques, for the missing cloud properties, corrections of solar position calculations, an enhanced algorithm for computing land-surface albedo over snow or ice surfaces, and a physics-based model for computing DNI. The updated PSM was used to generate solar radiation data for 2021-2023, including global horizontal irradiance (GHI), direct normal irradiance (DNI), and diffuse horizontal irradiance (DHI). This model was also employed to reprocess all previous NSRDB data from 1998-2020. The NSRDB was validated using high-quality measurements from surface sites, including NOAA’s the Surface Radiation Budget (SURFRAD). According to data for 2019-2023, the NSRDB processed by the latest PSM has mean bias error (MBE) within 5% and 8%, respectively, for GHI and DNI. This study provides users with the latest NSRDB information and outlines plans for ongoing development and updates.