CAELUS, a novel classification algorithm that relies on various thresholds to separate all possible sky conditions into six classes, is presented in Ruiz-Arias and Gueymard (2023, doi: 10.1016/j.solener.2023.111895). This dataset was used to develop, validate and benchmark CAELUS. It is made up by 1-min quality-assured observations of global horizontal irradiance (GHI) and diffuse horizontal irradiance at 54 stations of the Baseline Surface Radiation Network (BSRN) archive, which is publicly available (see download instructions in https://bsrn.awi.de/data). The dataset includes 5 years of data per station, except in two of them (Petrolina, Brazil, and Solar Village, Saudi Arabia), combined with other variables that are required to run CAELUS, namely: solar zenith angle (sza), extraterrestrial horizontal solar irradiance (eth), clear-sky GHI (ghics) and GHI in a clean and dry atmosphere (ghicda). In addition, the dataset also provides the sky classification obtained with CAELUS. Further details about CAELUS and the dataset compilation is available in Ruiz-Arias and Gueymard (2023, doi: 10.1016/j.solener.2023.111895). A Python implementation of CAELUS is available in https://github.com/jararias/caelus.

This work was supported by the project PID2019-107455RB-C21 funded by MCIN/AEI/ 10.13039/501100011033, the project UMA20-FEDERJA-134 jointly funded by the FEDER 2014-2020 Operative Program and the Consejería de Economía, Conocimiento, Empresas y Universidad of the Junta de Andalucía, and by Solargis s.r.o. through the collaboration agreement 2021-124 with the University of Málaga. The authors would like to thank the scientists and personnel in charge of the BSRN stations for acquiring, processing and kindly sharing their datasets, which have been central to this study. Moreover, the authors acknowledge the scientists and personnel of the Global Modelling and Assimilation Office at NASA Goddard Space Flight Center who provided the MERRA‑2 atmospheric data that were advantageously used to calculate the clear-sky solar irradiance.