A 25-year assessment of soil salinity in an arid oasis: Spatiotemporal validation and SHAP attribution
- 1Xinjiang Key Laboratory of Hydraulic Engineering Security and Water Disasters Prevention, Urumqi 830052, China; College of Hydraulic and Civil Engineering, Xinjiang Agricultural University, Urumqi 830052, China
- 2Xinjiang Key Laboratory of Hydraulic Engineering Security and Water Disasters Prevention, Urumqi 830052, China; College of Hydraulic and Civil Engineering, Xinjiang Agricultural University, Urumqi 830052, China; Corresponding author at: College of Hydraulic and Civil Engineering, Xinjiang Agricultural University, Urumqi 830052, China.
- 3Xinjiang Key Laboratory of Hydraulic Engineering Security and Water Disasters Prevention, Urumqi 830052, China; College of Hydraulic and Civil Engineering, Xinjiang Agricultural University, Urumqi 830052, China; Xinjiang Water Conservancy Development Investment (Group) Co., Ltd., Urumqi 830000, China
Abstract
Long-term soil salinity mapping in arid oases is vulnerable to spatial dependence and interannual shifts, yet transfer to unsampled areas and years is rarely tested explicitly. We combined 1107 quality-controlled topsoil observations from the Yanqi Basin, China (2000, 2005, 2015, 2020, and 2025), with 30 environmental covariates. Ten algorithms and five input strategies were compared using site-grouped five-fold, 15-km spatial-block five-fold, and leave-one-year-out (LOYO) validation. Extremely Randomized Trees (ERT) provided the most stable overall performance (R 2 = 0.602, 0.597, and 0.501), whereas Transformer performed best under LOYO (R 2 = 0.523), showing that algorithm rankings depended on the transfer scenario. PCC, VIP, GRA, and GBM feature selection produced no consistent gains; ERT fitted with all 30 predictors was therefore used for mapping. Within the common valid support, the area with SSC ≥ 6 g kg −1 increased from 23.65% in 2000 to 36.41% in 2005, declined to 22.40% in 2015, and rose to 33.88% in 2025, indicating an expansion–relief–reintensification pattern. ERT-SHAP identified SI3, SI1, and Int1 as the leading predictive indicators, while groundwater depth added little independent information after the other covariates were included. ERT systematically underestimated SSC ≥ 10 g kg −1 (spatial-block bias, −4.426 g kg −1 ), requiring caution at the upper end of the observed range. The framework supports reproducible regional comparison and potential hotspot screening by combining spatiotemporal transfer tests, high-value diagnostics, attribution, and long-term mapping.
Related
- Long-term ozone formation sensitivity in China: spatiotemporal evolution and machine learning attribution
- An integrated framework for agroclimatic, salinity, and groundwater assessment in the Jizzakh Region, Uzbekistan
- Bay Assessment Model: A Python-based Salinity Projection Tool for Florida Bay, USA
- Explainable spatiotemporal Graph Neural Networks: Multi-City PM2.5 forecasting in Bangladesh with geographic graph topology validation
- Structural diagnosis and evolution of urban resilience in dryland oasis urban agglomerations from an eco-immunological perspective
BibTeX
@article{li202625year,
title = {{A 25-year assessment of soil salinity in an arid oasis: Spatiotemporal validation and SHAP attribution}},
author = {Zhipeng Li and Pengnian Yang and Youyang Luo and Fanghong Han and Yongpeng Wang and Long Zhou},
journal = {Ecological Indicators},
year = {2026},
month = oct,
doi = {10.1016/j.ecolind.2026.115475},
url = {https://doi.org/10.1016/j.ecolind.2026.115475},
note = {CC BY},
}
RIS
TY - JOUR
TI - A 25-year assessment of soil salinity in an arid oasis: Spatiotemporal validation and SHAP attribution
AU - Zhipeng Li
AU - Pengnian Yang
AU - Youyang Luo
AU - Fanghong Han
AU - Yongpeng Wang
AU - Long Zhou
PY - 2026
DA - 2026/10/23
JO - Ecological Indicators
DO - 10.1016/j.ecolind.2026.115475
UR - https://doi.org/10.1016/j.ecolind.2026.115475
AB - Long-term soil salinity mapping in arid oases is vulnerable to spatial dependence and interannual shifts, yet transfer to unsampled areas and years is rarely tested explicitly. We combined 1107 quality-controlled topsoil observations from the Yanqi Basin, China (2000, 2005, 2015, 2020, and 2025), with 30 environmental covariates. Ten algorithms and five input strategies were compared using site-grouped five-fold, 15-km spatial-block five-fold, and leave-one-year-out (LOYO) validation. Extremely Randomized Trees (ERT) provided the most stable overall performance (R 2 = 0.602, 0.597, and 0.501), whereas Transformer performed best under LOYO (R 2 = 0.523), showing that algorithm rankings depended on the transfer scenario. PCC, VIP, GRA, and GBM feature selection produced no consistent gains; ERT fitted with all 30 predictors was therefore used for mapping. Within the common valid support, the area with SSC ≥ 6 g kg −1 increased from 23.65% in 2000 to 36.41% in 2005, declined to 22.40% in 2015, and rose to 33.88% in 2025, indicating an expansion–relief–reintensification pattern. ERT-SHAP identified SI3, SI1, and Int1 as the leading predictive indicators, while groundwater depth added little independent information after the other covariates were included. ERT systematically underestimated SSC ≥ 10 g kg −1 (spatial-block bias, −4.426 g kg −1 ), requiring caution at the upper end of the observed range. The framework supports reproducible regional comparison and potential hotspot screening by combining spatiotemporal transfer tests, high-value diagnostics, attribution, and long-term mapping.
SN - 1470-160X
SN - 1872-7034
ER -