ISSN 2756-3251
Advanced Journal of Environmental Science and Technology | Vol. 17 (7), pp. 1494-1508, July, 2026
Research Article
Satellite-Coupled Deep Learning Models Map Urban Soil Sealing and Subsurface Stormwater Infiltration Capacity Across Rapidly Expanding Megacities
Chidinma Okafor¹*, Lucas Meyer², Camila Gomez³
¹Department of Geography and Environmental Management, University of Lagos, Lagos, Nigeria
²Institute of Cartography and Geoinformation, ETH Zurich, Zurich, Switzerland
³Department of Civil and Environmental Engineering, University of Los Andes, Bogota, Colombia
Corresponding Author: Chidinma Okafor — [email protected]
Abstract
Unplanned impervious surface expansion in rapidly urbanizing megacities overwhelms aging drainage networks and intensifies flash flooding during extreme tropical rainstorms. We fused Sentinel-1 synthetic aperture radar (SAR) backscatter with Sentinel-2 multispectral imagery inside a U-Net deep learning architecture to map sub-pixel soil sealing across five coastal megacities. Validated against 1,200 field double-ring infiltrometer measurements, the model predicted saturated hydraulic conductivity and surface runoff coefficients with an R² of 0.89 at 10-meter resolution. Scenario simulations identified priority green-infrastructure retrofit zones capable of reducing peak urban flood volumes by 34%.
Keywords: urban soil sealing; stormwater infiltration; satellite remote sensing; deep learning; urban flood mitigation; impervious surfaces
Manuscript Timeline: Received 20 May, 2026; Revised 17 June, 2026; Accepted 28 June, 2026; Published 30 June, 2026.
Citation: Okafor, C., Meyer, L., and Gomez, C. (2026). Satellite-Coupled Deep Learning Models Map Urban Soil Sealing and Subsurface Stormwater Infiltration Capacity Across Rapidly Expanding Megacities. Advanced Journal of Environmental Science and Technology, 17(6), pp. 1494-1508.
Subscribe to read the full article: https://www.internationalscholarsjournals.org/subscribe-to-read