Estimating Long-Term Fractional Cover of Oil and Gas Well Pads in the Permian Basin Using Landsat Time Series (2000–2023)

The Permian Basin of the United States has undergone continued oil and gas (O&G) infrastructure expansion over the past century, yet long-term, spatially explicit records of this development remain limited. In this study, we developed a regression-based remote sensing method to estimate annual fractional cover of O&G well pads at 30 m resolution from 2000 to 2023 across the Permian Basin. A very-high-resolution (0.6 m) classification map derived from 2018 National Agriculture Imagery Program (NAIP) imagery was used to generate 30 m pixel-level fractional cover reference data, which served as training data for a Random Forest regression model. Predictor variables included seasonal metrics (May–October) derived from Landsat surface reflectance time series, along with topographic features (elevation, slope, aspect), to capture intra-annual surface variability. The trained model was applied to the Landsat archive to produce annual 30 m fractional cover estimates, constrained to the 2018 mapped O&G footprints. This approach enables robust temporal tracking within established development zones while minimizing commission errors outside of active O&G extraction areas. The results reveal strong spatiotemporal variability in O&G land-use development over the past two decades, coinciding with the widespread adoption of horizontal drilling and hydraulic fracturing. This study demonstrates a robust framework for long-term monitoring of O&G land use change and provides a foundation for assessing ecosystem and economic impacts of fossil energy extraction in arid and semi-arid landscapes such as the Permian Basin.

Li, H., Song, X.P. and Ma, Y. (2025), Estimating Long-Term Fractional Cover of Oil and Gas Well Pads in the Permian Basin Using Landsat Time Series (2000–2023). AGU Fall Meeting, Abstracts B51E-0689, December 14-19, New Orleans, Louisiana, USA.