Lv et al. (2025) Analysis of driving factors of soil salinity in Southern Xinjiang irrigation areas under dry-sowing and wet-emerging conditions

Identification

• Journal: Agricultural Water Management
• Year: 2025
• Date: 2025-12-01
• Authors: Tingbo Lv, Shaozhong Kang, Yi‐Fan Liu, Menghan Bian, Ling Tong, Wenhao Li
• DOI: 10.1016/j.agwat.2025.109962

Research Groups

• College of Water Resources and Civil Engineering, China Agricultural University, Beijing, China
• College of Water Conservancy & Architectural Engineering, Shihezi University, Shihezi, Xinjiang, China
• State Key Laboratory of Efficient Utilization of Agricultural Water Resources, Beijing, China

Short Summary

This study applied Multiscale Geographically Weighted Regression (MGWR) for the first time to quantitatively analyze the spatially heterogeneous driving factors of topsoil salinity (0–30 cm) in the Xiaohaizi Irrigation District under dry-sowing and wet-emerging conditions, revealing groundwater salinity as the dominant positive driver and proposing zone-specific management strategies.

Objective

• To determine the spatial patterns of surface salinity gain zones and inhibition zones under the Dry-Sowing and Wet-Emerging (DSWE) model.
• To identify which factors exert what type of influence on salinity, and in which specific areas.
• To develop a "zone–factor" precision regulation strategy based on the underlying spatial mechanisms.

Study Configuration

• Spatial Scale: Xiaohaizi Irrigation District, Xinjiang, China (78°56′–79°35′ E, 39°39′–40°43′ N), covering an area with 87 high-density sampling sites based on a 5 × 5 km grid, focusing on topsoil (0–30 cm) salinity.
• Temporal Scale: Soil moisture and salinity data were collected in March 2023, representing pre-sowing conditions.

Methodology and Data

• Models used: Multiscale Geographically Weighted Regression (MGWR), Ordinary Kriging (OK), Ordinary Least Squares (OLS).
• Data sources:
  • Field measurements: 261 layered soil samples (0–30 cm, 30–60 cm, 60–100 cm) from 87 sites for soil salinity and moisture content; GPS for coordinates and elevation.
  • Remote sensing: Landsat-8 OLI imagery for fractional vegetation coverage.
  • Geospatial data: 30 m Digital Elevation Model (DEM) for elevation and slope.
  • Monitoring wells: Groundwater depth and groundwater salinity.

Main Results

• MGWR significantly outperformed global OLS regression, achieving localized R² values ranging from 0.52 to 0.91, compared to OLS R² of 0.41, effectively capturing the spatial variability of soil salinity.
• Groundwater salinity was identified as the dominant positive driver, with a mean coefficient of 0.399 and a peak of 0.861 in the northwestern part of the district (bandwidth: 2.1 km), directly delineating high-risk zones for salt accumulation.
• Elevation and groundwater depth exhibited the strongest negative effects in the northeastern low-lying areas (coefficients: −0.164 and −0.179, respectively), suppressing salt accumulation.
• Slope showed a weak positive effect in the southern gentle-slope region, suggesting that under shallow irrigation in DSWE systems, ponding and secondary evaporation may amplify topographic influence on salinization.
• Fractional vegetation coverage significantly suppressed salinity only in low-coverage areas, while its effect was diluted in high-coverage zones, indicating a threshold-type regulation.
• OK was the most accurate interpolator (independent validation RMSE: 0.234 g kg⁻¹, MAE: 0.385 g kg⁻¹), revealing that slightly and moderately saline soils dominate the surface layer (53.7 % and 9.9 % of the area, respectively), with severe salinization concentrated in the northwestern 50th Regiment.
• A "salinity accumulation triangle" (high mineralization, shallow groundwater depth, gentle-slope ponding) and a "salinity suppression triangle" (elevation, drainage, vegetation cover) were identified, explaining the spatial heterogeneity.
• A three-zone precision regulation strategy was proposed: implementing well drainage and freshwater irrigation to dilute groundwater in the northwest (target < 3 g L⁻¹ groundwater salinity), intensifying subsurface drainage to maintain groundwater depth ≥ 0.8 m in the northeast, and optimizing the lower irrigation limit under DSWE in the southern gentle-slope zone to reduce ponding duration.

Contributions

• First-time introduction and application of Multiscale Geographically Weighted Regression (MGWR) to quantitatively disentangle the spatially varying impacts of topography, groundwater, soil properties, and vegetation on topsoil salinity under Dry-Sowing and Wet-Emerging (DSWE) conditions.
• Provides a methodological paradigm for spatially explicit salinity management in arid oasis regions, moving beyond empirical regulation to precision management.
• Offers a novel application case of MGWR in agricultural hydrological process analysis, demonstrating its capability in identifying multi-scale driving mechanisms.
• Develops a "zone–factor" precision regulation strategy, directly translating spatially driven mechanisms into actionable production management practices for sustainable salinity control.

Funding

• Support Plan for Key Industrial Innovation and Development in Southern Xinjiang Corps (2022DB024).

Citation

@article{Lv2025Analysis,
  author = {Lv, Tingbo and Kang, Shaozhong and Liu, Yi‐Fan and Bian, Menghan and Tong, Ling and Li, Wenhao},
  title = {Analysis of driving factors of soil salinity in Southern Xinjiang irrigation areas under dry-sowing and wet-emerging conditions},
  journal = {Agricultural Water Management},
  year = {2025},
  doi = {10.1016/j.agwat.2025.109962},
  url = {https://doi.org/10.1016/j.agwat.2025.109962}
}