Soil organic matter (SOM) monitoring using remote sensing is critical for effective land resource management and environmental protection.  Mapping SOM in areas where saline and black soils are intertwined, with complex soil types and significant environmental variability, remains a challenging task.  This study integrated prior knowledge and classified Jilin Province, China, into saline-alkali and black soil areas.  All available Sentinel-2 images from 2019 to 2023 during the bare soil period (April to July) were collected, and the images were categorized into three time windows: day of year (DOY) 90–120, DOY 120–150, and DOY 150–180.  The potentials of these time windows, spectral indices (salinity index and vegetation moisture index), environmental variables (topography and climate), and local regression models for SOM mapping in the saline-black soil transition areas were then systematically evaluated.  The results revealed four key findings: (1) the optimal time window for SOM mapping in both the saline-alkali area and black soil area was DOY 90–120; (2) including the salinity index improved SOM mapping accuracy in the saline-alkali area but reduced it in the black soil area, whereas the vegetation moisture index enhanced accuracy in both areas; (3) incorporating environmental variables improved the SOM mapping accuracy in all areas, with topographic variables being more influential in the black soil area and climatic variables being more significant in the saline-alkali area; and (4) local regression models based on the saline-alkali area and black soil area outperformed the global regression model in terms of SOM mapping accuracy, although they exhibited higher uncertainty.  This study demonstrates that the integration of prior knowledge and multi-temporal remote sensing images significantly enhance SOM mapping accuracy in areas where saline and black soils intersect, thus providing a scientific foundation for the precise management and protection of areas with different soil types.

Soil organic matter (SOM) is a core indicator of soil fertility and ecosystem function. However, in regions where Mollisol and non-Mollisol coexist, high-precision spatial mapping faces significant challenges due to pronounced terrain heterogeneity and redundancy in high-dimensional covariates. This study proposes a “remote sensing zoning-feature selection optimization-random forest (RSZ-FSO-RF)” framework. By integrating Landsat-8 multi-temporal imagery from 2014-2023 with topographic and climatic factors, and leveraging the Google Earth Engine (GEE) platform, it achieves high-precision remote sensing zoning of Mollisol and non-Mollisol areas (overall accuracy: 92.13%, Kappa coefficient: 0.70). Subsequently, local Random Forest (RF) regression models were established within each zone for SOM prediction, with predictive variables optimized using Recursive Feature Elimination (RFE). Results demonstrate that compared to FAO-zone-based modeling, the RSZ-FSO-RF framework significantly enhances prediction accuracy (R²=0.619, RMSE=6.849 g kg^-1). And further feature optimization continued to enhance model performance (R²=0.627, RMSE=6.781 g kg^-1). Notably, optimal predictor combinations varied significantly across zones, with SOM spatial variability generally higher in non-Mollisol areas than in Mollisol regions. By organically integrating remote sensing zoning with feature selection, this framework effectively mitigates covariate redundancy while accounting for local heterogeneity, significantly enhancing the accuracy and stability of high-resolution SOM mapping. Furthermore, this study provides scientific basis and decision support for soil resource management and sustainable agricultural development under complex topographic conditions.