Estimating Day-Ahead Solar Forecast Errors From Site-Specific Variability

Lauret, Philippe, La Salle, Josselin Le Gal, David, Mathieu

ISES Solar World Congress 2025 · Fortaleza, Brazil · 2025-11-03
Published by International Solar Energy Society (ISES)
DOI: 10.18086/swc.2025.04.21

Abstract

This work introduces a straightforward statistical model to estimate the Root Mean Square Error (RMSE) of day-ahead solar irradiance forecasts using only a site-specific solar variability metric. The proposed method leverages a global dataset of 60 quality-controlled ground stations and employs a linear regression between RMSE and the standard deviation of hourly changes in the clear-sky index. This results in an efficient predeployment tool that helps practitioners assess forecast errors at a site before implementing a forecasting method. The model includes also a 95% uncertainty interval, providing valuable insight into the bounds within which future forecast errors are likely to fall. Validation on seven independent sites from the SURFRAD network confirms that the proposed model can reasonably predict RMSE forecast errors based on a site’s nominal variability.

Keywords

Solar forecast error, Solar variability, Root Mean Square Error (RMSE), Forecast verification

← All proceedings papers