Mapping Soil Organic Carbon in Floodplain Farmland: Implications of Effective Range of Environmental Variables

Accurately mapping soil organic carbon (SOC) is conducive to evaluating storage and quality. However, the high spatial heterogeneity of SOC caused by river-related factors agricultural management brings challenges digital in floodplain farmland. Moreover, current studies focus on non-linear relationship between covariates, but ignore effective range environmental variables SOC, which prevents revelation differentiation mechanism. Using 375 samples collected from Jiangchang Town near Han River, we aim determine main controlling reveal variables, obtain map using gradient boosting decision tree (GBDT) model partial dependence plots. Linear regression was used as a reference. Results showed that GBDT outperformed linear regression. results show distance river most important factor, confirming importance River pattern. The plots indicate all have their ranges, when values are extremely or low, they do not respond changes SOC. Specifically, influential ranges rivers, irrigation canals, rural settlements were within 4000, 200, 50 m, respectively. peak obtained with clay (≥31%), total nitrogen (≥1.18 g/kg), potassium contents (≥11.1 it remained steady these covariates further increased. These highlight revealing provides data support for understanding pattern farmland, achieving sequestration farmland precision agriculture. plot fitting mapping.