LAUSR

Abstract The economics of photovoltaic systems are closely associated with the geographical locations of the sites. To identify appropriate land for photovoltaic systems, previous studies have mainly used geographic-information-system-based solar energy models capable of estimating spatial and temporal availability of solar energy. The models often require handling large and computationally challenging datasets to cover vast areas. In this study, we proposed and assessed the concept of a two-stage approach that sequentially searches and prioritizes suitable sites of photovoltaic panels using low- and high-resolution maps to reduce computational burdens and increase estimation accuracy. Resolutions at raster-cell sizes of 30 × 30 m and 5 × 5 m were selected to study the South Korean national highway network. Selected candidate sites in the 10th-percentile for low- and high-resolution cases produced 1,097 and 721 clusters with calculated solar energy potentials of 19,510,115 and 4,186,935 MWh, respectively. Approximately 90% of the edge-based distances between two adjacent potential sites for photovoltaic panels from the low- and high-resolution maps were found to be within 240 m. This result shows that although the low-resolution maps may not provide precise locations for prospective solar photovoltaic sites, they may help to identify potentially suitable areas for further investigation.