← Quentin Paletta

SolarBench: A Global Solar Energy Nowcasting Benchmark

arXiv preprint  ·  September 5, 2026 arXiv:2609.06187

SolarBench sky and satellite image dataset preview

[arXiv]

Authors

Yuhao Nie, Stephen Campbell, Quentin Paletta, Liwenbo Zhang, Tao Jing, Samer Chaaraoui, Jonathan Giezendanner, Andea Scott, Tao Sun, Cong Feng, Max Aragon, Jacques Camier, Adam Jensen, Florian Kotthoff, Yuexing Yang, Yang Ming, Mengying Li, Stefanie Meilinger, Yupeng Wu, Adam Brandt, Sherrie Wang

Abstract

SolarBench is an open global benchmark for image-based solar nowcasting. It combines more than six million sky and satellite images from 11 diverse sites spanning a decade, together with irradiance or PV output and auxiliary atmospheric data. The benchmark evaluates representative forecasting models and reveals a gap between average forecasting accuracy and the ability to capture rapid solar fluctuations, shows how predictability varies across different cloud conditions, and demonstrates adaptation methods for new photovoltaic systems — alongside an accompanying toolbox for reproducible research.

arXiv  /  PDF  /  BibTeX


Page maintained by Quentin Paletta. Design adapted from Jon Barron.