Analysis of black soil changes and driving forces via satellite remote sensing

Using Sentinel-2 imagery, this study examined spatiotemporal changes and drivers of black soil in Mishan City, Heilongjiang Province (2017–2024). Level-1C images were radiometrically corrected to Level-2A surface reflectance via Sen2Cor, with cloud removal using Fmask. Land use classification employed random forest (RF) and maximum likelihood (ML) methods. RF outperformed ML, achieving >90% accuracy for seven land cover types. A transfer matrix reveals cultivated land first fluctuated and grew, peaking at 3968.87 km2 (49.10%) in 2023, then shrank by 3.48% in 2024 due to urbanization. Construction land exhibited a "U-shaped" trend, surging 108.66% in 2023–2024. Woodlands and water bodies partially recovered, but wetlands/tidal flats dropped 81.9%. Drivers include population/GDP growth fueling construction expansion, secondary industry development accelerating cropland loss, and grain demand promoting intensive farming—linked to wetlands degradation. The study clarifies anthropogenic composite drivers of black soil changes, aiding cultivated land protection and sustainable development in Northeast China's black soil region.