Parcel-level remote sensing evaluation framework for cropland ecological quality in arid oasis regions: A case study of Luopu County, Xinjiang

Cropland ecological quality evaluation in arid oasis regions faces three challenges: the disconnect between evaluation units and parcel management, the neglect of core-edge ecological gradients by homogeneous methods, and the difficulty in quantifying human management responses. This study constructs a new evaluation framework integrating functional zoning and spatial variable weighting based on remote sensing. Using cropland parcels as the basic evaluation unit, the study divides the area into core irrigation zones, ecological transition zones, and desert edge zones based on geospatial features. Subsequently, a Pressure-State-Response (PSR) evaluation system is built, strengthening indicators related to human management dimensions. Finally, a spatial variable weight model is used to assign differentiated criterion-layer weights to the three functional zones, enabling dynamic evaluation weights adjustment with spatial function. The case in Luopu County, Xinjiang, shows that the Pearson correlation coefficient between the results of this framework and the remote sensing ecological index (RSEI) is 0.867, higher than the 0.858 of the uniform weight model and significantly higher than the 0.814 of the pixel-level method. These results indicate that parcel-level valuation suppresses noise and aligns better with management boundaries than pixel-level valuation, while the spatially variable weighting model responds more sensitively to the dominant driving mechanisms of different functional zones. This study shifts the evaluation unit from pixels to parcels, realizes the differentiated diagnosis of ecological gradients within oases, and provides an operable methodological framework for the refined monitoring and targeted management of cropland ecological quality in arid regions.