Climate impacts on photovoltaic (PV) energy production are commonly assessed using global climate model output from the Coupled Model Intercomparison Project (CMIP) with coarse spatial resolutions of 100 to 200 km and daily-mean output. Recently, kilometer-scale global climate models have emerged with resolutions of a few kilometers and sub-hourly output, potentially offering added value for PV assessments. We evaluate this potential by quantifying how spatial and temporal data resolution affects regional PV power potential (PVpot). Using climate model output at 12 km horizontal resolution and 15-minute frequency, we calculate PVpot and systematically coarse-grain the climate data to resolutions typical of CMIP models. We show that errors in daily PVpot are primarily driven by temporal rather than spatial averaging. Daily and half-daily climate data overestimate PVpot due to insufficient representation of the diurnal cycle of solar irradiance, with relative errors of up to 10%. In contrast, 3-hourly or finer temporal resolution reduces errors to below 1%. Spatial averaging from 12 to 192 km introduces negligible errors and minimally affects the identification of low-PVpot days. We conclude that high spatial resolution alone provides little added value for regional and continental PV assessments, provided that the temporal resolution adequately captures the diurnal cycle. • Regional daily PVpot is insensitive to kilometer-scale spatial resolution. • Spatial averaging from 12 to 192 km introduces negligible errors. • Temporal resolution dominates errors. • Resolving the diurnal cycle with 3-hourly data reduces errors to near zero. • Low-PVpot days are robustly captured also at coarse spatial resolution.
Haslehner et al. (Fri,) studied this question.