Skip to content

Variation in estimates of the footprint of large, ground-mounted photovoltaic solar energy in the United States and its associated land-cover change across three datasets

Published Article

United States

Publication date: November 1, 2025

File format: URL

View resource

Ground‑mounted photovoltaic solar facilities (GPVs) are rapidly expanding across the United States, yet datasets describing their land footprint differ widely. This study harmonized three major datasets—two mapping panel arrays and one mapping full fencelines—to compare footprint size and associated land‑cover change. Panel‑array datasets underestimated GPV footprints by up to 34% relative to fenceline boundaries, leading to lower estimates of shrubland and scrubland conversion. Because each dataset defines “footprint” differently, they are not interchangeable and serve distinct analytical purposes. The lack of publicly available fenceline‑level data limits assessments of GPV impacts on land cover, biodiversity and other place‑based environmental considerations.

Subject Tags

  • Biodiversity
  • Renewable energy

Abstract

Solar energy contributed more than 6% of the United States’ (US) energy generation capacity in 2022, predominantly from large, ground-mounted photovoltaic solar facilities (GPVs). That proportion is expected to increase as GPV development accelerates and its land footprint expands. This study compared and harmonized three datasets describing the U.S. GPV footprint, evaluated the land-use and land-cover change associated with each, and identified their best use cases. Two publicly available datasets delineated photovoltaic panel arrays, while a third paywalled dataset mapped the broader GPV fence line. For GPVs represented in all three datasets, we harmonized the data to 2018 and calculated both footprint area and quantified the footprint area and that of the land-cover classes converted within it. Our results suggest that US GPV footprint estimates from the panel array datasets are up to 34% smaller than those from the fenceline dataset. This has implications for land-cover change estimates; conversion of the relatively undisturbed categories shrubland and scrubland was 8% higher in the fenceline dataset than in the array datasets. As such, due to differences in their definitions of “footprint” and footprint delineation methods, these datasets are not interchangeable and have specific analytical contexts where they are best suited. A lack of publicly available data describing the broader GPV footprint hinders analysis of the relationship between GPVs and other place-based phenomena, particularly considerations of land-cover change and its relationship to biodiversity conservation.

Citation

Levin, M. O., Forester, E., Kalies, E. L., Goodman, L., Hagani, J. S., Holmes, C. C., ... & Hernandez, R. R. (2025). Variation in estimates of the footprint of large, ground-mounted photovoltaic solar energy in the United States and its associated land-cover change across three datasets. Journal of Environmental Management, 394, 127634. https://doi.org/10.1016/j.jenvman.2025.127634

TNC Authors