A cross-sensor spectral index for soil sealing detection with improved impervious-bare soil discrimination

Rapid urbanization has led to widespread soil sealing, an often irreversible process that reduces soil permeability and disrupts key ecosystem functions, necessitating consistent monitoring. While medium-resolution satellite missions such as Sentinel-2 and Landsat provide consistent global observations, distinguishing sealed surfaces from exposed bare soils remains challenging due to spectral similarities, particularly in heterogeneous urban and peri-urban environments. This study introduces a novel index, the Sealed Urban Index (SUI). The SUI is designed to improve the discrimination of sealed surfaces using spectrally homologous bands available in Sentinel-2 and Landsat 8–9 imagery and computationally efficient across different sensors. The index was evaluated using annual median composites derived from Sentinel-2 and Landsat surface reflectance data for the period 2017 to 2024. The analysis was implemented in the Google Earth Engine platform across representative regions of Europe, North America, South America, and North Africa. Classification performance was assessed using overall classification accuracy, precision, recall, the harmonic mean of precision and recall, and disagreement metrics that quantify differences in class quantities and spatial allocation. Results demonstrate stable multi-year performance for both sensors, with overall accuracy ranging from approximately 80–83% for Landsat and 82–86% for Sentinel-2. Low quantity disagreement indicates limited bias in the estimated extent of sealed surfaces. In contrast, higher allocation disagreement suggests that residual errors are mainly associated with spatial boundary placement in heterogeneous landscapes. Sentinel-2 outperformed Landsat likely due to its finer spatial resolution. Ultimately, the index provides a transferable and computationally efficient approach for large-scale soil sealing monitoring, supporting global assessments of impervious surface dynamics.