Prediction of Soil Salinization in Arid Regions During Wet and Dry Seasons Based on Spectro‐Polarimetric Features and Machine Learning

ABSTRACT Soil salinization is one of the main causes of soil degradation and ecosystem deterioration in arid regions, posing a serious threat to ecological environments and agricultural security. Understanding the factors influencing soil salinization is crucial for soil management and improvement. However, the sensitivity of soil salinization to seasonal changes has not been thoroughly studied in arid regions. Therefore, this study focuses on the Yanqi Basin, where 129 soil samples were collected (wet season of 51, dry season of 78) for laboratory analysis to determine the soil saturated extract conductivity (EC e ). Soil salinity feature variables were extracted from Sentinel‐1 radar remote sensing data, Sentinel‐2 optical remote sensing data, and digital elevation models (DEM). The Boruta algorithm was used to select feature variables, and the optimal feature variables were combined with Random Forest (RF), Support Vector Machine (SVM), and Extreme Gradient Boosting (XGBoost) models to construct prediction models. The results indicate: (1) Red‐edge spectral features (RE) can effectively predict soil salinization. In addition, the variables most correlated with EC e are elevation (DEM) and river network baseline (CNBL), mainly because the terrain in the study area is higher in the northwest and lower in the southeast, with flat farmland in the central region, where the movement of water and salt is significantly influenced by the terrain. (2) The RF model is the best prediction model in this study, with R 2 = 0.78, effectively revealing the spatial distribution of soil salinity during both the wet and dry seasons. (3) The degree of salinization in the wet season is significantly higher than in the dry season due to the combined effects of higher precipitation, lower vegetation cover, evaporation, and salt migration. (4) During both the dry and wet seasons, salinized soil is mainly concentrated along the shores of Bosten Lake, the Kaidu River, and Huangshui Ditch, while light salinization is distributed in the Gobi Desert areas. This study provides scientific evidence for the management and improvement of soil salinization caused by seasonal changes in arid regions.

Authors

Institutions

Publication Details

Journal
Land Degradation and Development
Published
2025-05-22
DOI
https://doi.org/10.1002/ldr.5635
Citations
2
Primary Topic
Soil Geostatistics and Mapping
Type
article
Field-Weighted Citation Impact
1.60

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Prediction of Soil Salinization in Arid Regions During Wet and Dry Seasons Based on Spectro‐Polarimetric Features and Machine Learning

Mireguli Ainiwaer, Xayida Subi, Jiaqi Zhang, Aizemaitijiang Maimaitituersun et al.
2 citations
Land Degradation and Development
Soil Geostatistics and Mapping
1.60
article

Prediction of Soil Salinization in Arid Regions During Wet and Dry Seasons Based on Spectro‐Polarimetric Features and Machine Learning

Mireguli Ainiwaer, Xayida Subi, Jiaqi Zhang, Aizemaitijiang Maimaitituersun, Xiaobing Wang
article en
2 citations

Abstract

ABSTRACT Soil salinization is one of the main causes of soil degradation and ecosystem deterioration in arid regions, posing a serious threat to ecological environments and agricultural security. Understanding the factors influencing soil salinization is crucial for soil management and improvement. However, the sensitivity of soil salinization to seasonal changes has not been thoroughly studied in arid regions. Therefore, this study focuses on the Yanqi Basin, where 129 soil samples were collected (wet season of 51, dry season of 78) for laboratory analysis to determine the soil saturated extract conductivity (EC e ). Soil salinity feature variables were extracted from Sentinel‐1 radar remote sensing data, Sentinel‐2 optical remote sensing data, and digital elevation models (DEM). The Boruta algorithm was used to select feature variables, and the optimal feature variables were combined with Random Forest (RF), Support Vector Machine (SVM), and Extreme Gradient Boosting (XGBoost) models to construct prediction models. The results indicate: (1) Red‐edge spectral features (RE) can effectively predict soil salinization. In addition, the variables most correlated with EC e are elevation (DEM) and river network baseline (CNBL), mainly because the terrain in the study area is higher in the northwest and lower in the southeast, with flat farmland in the central region, where the movement of water and salt is significantly influenced by the terrain. (2) The RF model is the best prediction model in this study, with R 2 = 0.78, effectively revealing the spatial distribution of soil salinity during both the wet and dry seasons. (3) The degree of salinization in the wet season is significantly higher than in the dry season due to the combined effects of higher precipitation, lower vegetation cover, evaporation, and salt migration. (4) During both the dry and wet seasons, salinized soil is mainly concentrated along the shores of Bosten Lake, the Kaidu River, and Huangshui Ditch, while light salinization is distributed in the Gobi Desert areas. This study provides scientific evidence for the management and improvement of soil salinization caused by seasonal changes in arid regions.

Land Degradation and DevelopmentVol. 36(12)
Xinjiang Normal University (CN), Hohai University (CN)
National Natural Science Foundation of China, Xinjiang Normal University
Openalex Percentile: Top 18%
Soil Geostatistics and Mapping
1.60
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.