Optimization of Farmland Management Zoning in the Black Soil Region: A Climate Adaptability Assessment Considering Crop Growth Response and Topographic Characteristics

Management zoning underpins precision agriculture, but conventional approaches relying on single-date imagery fail to capture inter-annual crop environment variations. Under increasing extreme climate events, single-year zoning exhibits limited adaptability across differing hydro-climatic conditions. This study evaluated the climate adaptability of 29 feature combinations—incorporating Sentinel-2 multispectral, PCA, NDVI, and DEM data—for dryland management zoning under dry, wet, and fused scenarios at Youyi Farm in the Black Soil Region. A Heterogeneous Spatial Attention Network (HSAN) and K-means clustering were implemented, using the coefficient of variation (CV) to evaluate stability and adaptability. Results showed that the HSAN outperformed K-means under multi-source fusion, achieving CVs of 11.303–14.774% versus 14.823–16.011%. Both methods confirmed that incorporating multi-period NDVI data was the dominant factor, achieving a 34.850–53.701% relative CV reduction compared to the outside-zone baseline. Conversely, DEM contributions were limited, whereas PCA enhanced stability, and multi-period fusion consistently outperformed single-year data. Crucially, the multi-period fusion framework exhibited prominent spatial heterogeneity, enhancing zoning ecological consistency and applicability across extreme climate years. Overall, multi-period fusion demonstrates stronger robustness under complex climatic conditions, providing a scientific basis for climate-adaptive management in the Black Soil Region.