Sensitivity of regional photovoltaic production simulations to system orientation and spatial distribution assumptions in France

This is the authors' abstract. We don't add key points for this paper.

Accurate regional simulations of photovoltaic (PV) electricity production are essential for power system planning and climate-energy impact studies. However, they often rely on simplified assumptions regarding system orientation and spatial distribution due to limited data availability. The relative impact of these simplifications on simulation accuracy remains poorly understood. In this study, we quantify the sensitivity of regional PV production simulations to common modeling assumptions related to module tilt, azimuth orientation, and spatial distribution in France. We construct a high-resolution reference simulation based on an image-and LiDAR-derived inventory of individual PV installations combined with ERA5 meteorological data. This reference is used to evaluate a set of progressively simplified scenarios, including fixed tilt angles, idealized azimuth configurations, and a uniform spatial distribution of installed capacity. Comparison with transmission system operator (TSO) data shows that the reference accurately reproduces regional PV production, with seasonal discrepancies partly attributed to the assumption of a static PV fleet. The sensitivity analysis, conducted for two French regions, reveals that assuming a representative fixed tilt introduces minimal error (pRMSE below 2%), while uniform south-facing azimuth assumptions increase pRMSE to 4.9–6.5%. The largest degradation arises when geometric simplifications are combined with uniform spatial distribution of installed capacity, with pRMSE exceeding 11–12% annually and wintertime errors reaching 25%. These results provide practical guidance for large-scale PV modeling under data-limited conditions, quantifying which simplifications can be applied safely and which ones introduce large errors, with direct relevance for energy system planning, climate impact assessments, and regional PV simulation frameworks.