Spatial estimation of large-scale soil salinity using enhanced inverse distance weighting method and identifying its driving factors

Spatially regional estimates of soil salinity and identifying their main driving factors are critical for effective land management. However, the understanding of soil salinity variation remains limited due to challenges in using conventional spatial interpolation methods with the highly variable soil salinity data in regional scales. The unrealistic "bull's-eye" effect is a common problem when using the inverse distance weighting method (IDW) for interpolation. Improvements like considering anisotropy and multi-parameter co-optimization have been proposed, while their application in regional soil salinity prediction that owing strong variability has not been investigated. To address these gaps, we conducted the case study with a sample size of 6045, across all the 15 major oasis irrigation districts with an area of 2.1 × 107 hectares in southern Xinjiang, and then the spatial variability of soil salinity was investigated by using an improved inverse distance weighting method with multi-parameter co-optimization (named as AIDW). Compared with the results of the ordinary kriging, local polynomial interpolation and inverse distance weighting methods, the accuracy of AIDW ranked first with the root mean square error (RMSE) from 0.59 to 1.48 g/kg and determination coefficient (R²) from 0.91 to 0.99. The "bull's-eye" effect has been greatly reduced, with the average contour roundness index decreasing from 0.64 to 0.68 in IDW to 0.42–0.49 in AIDW. The results of AIDW were then used to generate soil salinity maps for the whole region, along with an analysis of the driving factors. It showed that an estimated 1160.4 × 104∼1329.5 × 104 hectares of soils in southern Xinjiang were saline, with a clear clustering pattern indicating significantly higher soil salt content in the northwest oases than in the southeastern regions. Vertically, soil salt content followed the pattern: 0–10 cm > 10–25 cm > 25–50 cm, with moderate spatial variability that decreased with increasing depth. The major driving factors include climatic factors, hydrogeology, soil texture and topography as well as human activities. Overall, this study provides crucial insights and guidelines for the management of soil salinization in arid regions.