Objective This study aims to achieve efficient monitoring and dynamic assessment of soil salinity in typical arid irrigation areas. Methods The typical Yihe irrigation area in northwestern China was taken as the study area. Various spectral indices were constructed using Sentinel-2 multispectral imagery, and surface soil salinity data were simultaneously collected during the bare soil period. Through correlation analysis, salinity-sensitive spectral indices were selected. Then, inversion models were established using four machine learning algorithms—support vector machine (SVM), random forest (RF), LightGBM, and XGBoost—and their accuracy was validated. Results The spectral indices effectively suppressed background noise interference and significantly enhanced the correlation with measured soil salinity. The inversion accuracy of the four models ranked as SVM>XGBoost>LightGBM>RF. Among them, the SVM model performed the best (R²>0.80 for both training and validation sets; MAE=0.85 and 0.71, respectively). Multi-temporal inversion results based on the optimal SVM model indicated that from 2022 to 2023, the area of non-saline and mildly saline soil in the Yihe irrigation area dominated. Highly saline areas were scattered and exhibited minimal interannual variations, indicating a stable overall distribution pattern. Conclusion This study constructs a high-precision soil salinity inversion model applicable to the Yihe Irrigation Area, enabling dynamic multi-temporal monitoring. It provides a scientific basis for the early identification of abnormal salinity areas, zonal management, and differentiated treatment, offering significant reference value for regional sustainable agricultural development and ecological restoration in arid regions.
选取决定系数(coefficient of determination,R²)、均方根误差(root mean square error,RMSE)和平均绝对误差(mean absolute error,MAE)作为主要精度评价指标。其中,R²越接近1,表明模型拟合效果越好;RMSE越接近0,说明估测误差越小,预测精度越高;MAE越小,模型预测效果越优[21]。
LUB J, TIANS C, ZUOZ, et al. Review and prospect on sustainable utilization of salinized land[J].Journal of Ningxia University (Natural Science Edition),2023,44(1):79-88.
[3]
曹丹.基于水量平衡的黄河三角洲典型农田水盐调控研究[D].西安:长安大学,2022.
[4]
CAOD. Water and salt regulation in typical farmland of the yellow river delta based on water balance[D].Xi′an: Changan University,2022.
GUOJ L, MAY G, PANH, et al. Comparative analysis of satellite monitoring of soil salinization in Ebinur Lake during spring and summer[J].Arid Land Geography,2025,48(12):2143-2157.
BAY L, ZHANGZ T, XIEP L, et al. Inverting soil salinity of farmland in Xinjiang by integrating Sentinel-1/2 and environmental variables[J].Transactions of the Chinese Society of Agricultural Engineering,2024,40(16):171-179.
GENGM Q, TANGX Y, LIX C, et al. Research progress and prospect of saline soil monitoring based on remote sensing technology[J].World Science and Technology Research and Development,2025,47(1):60-81.
[11]
NURMEMETI, AILIY, XIANGY, et al. A three-dimensional feature space model for soil salinity inversion in arid oases: Polarimetric SAR and multispectral data synergy[J].Agronomy,2025,15(7):e1590.
WANGJ, YANGP N, WANGY P, et al. Mapping soil salinity in the Weigan River basin using Sentinel-2 imagery and machine learning[J].Journal of Irrigation and Drainage,2025,44(10):93-102.
LIZ, SUW Z, LIX G, et al. Modeling and validation of soil salinity content based on optimized hyperspectral feature parameters[J].Xinjiang Agricultural Sciences,2021,58(12):2342-2352.
[16]
张佘淑.西北干旱区土壤盐分遥感监测变量优选与模型构建研究[D].兰州:西北师范大学,2024.
[17]
ZHANGS S. Variable selection and model construction for remote sensing monitoring of soil salinity in arid regions of northwest China[D].Lanzhou: Northwest Normal University,2024.
[18]
宋阿维.青铜峡灌区土壤盐渍化演变过程及驱动机制[D].南京:南京信息工程大学,2024.
[19]
SONGA W. Evolution process and driving mechanisms of soil salinization in the Qingtongxia irrigation district[D].Nanjing: Nanjing University of Information Science and Technology,2024.
DONGY X. Inversion of soil salinity in the Hetao irrigation district based on the synergy of UAV and Sentinel-2A remote sensing data[D].Yangling, Shaanxi: Northwest A&F University, 2025.
YANY, SHIH B, MIAOQ F, et al. Water and salt transport pattern and balance analysis among different land classes in Yellow River irrigation area based on dry drainage salt control model[J].Transactions of the Chinese Society for Agricultural Machinery,2024,55(10):346-359.
[24]
ZHAOR Y. Pouring vitality and reconstructing memory:Baosteel industrial heritage landscape renovation design study[J].E3S Web of Conferences,2024,536:e01026.
[25]
OLIVAM, MAFFIAA, MARRAF, et al. The complex impacts of fire on soil ecosystems:Insights from the 2021 Aspromonte National Park wildfire[J].Journal of Forestry Research,2025,36(1):e68.
HUANGH Y, DINGQ D, ZHANGJ H, et al. Inversion of soil salinity and pH in farmland of the Hetao Plain based on Sentinel-2 and explainable machine learning[J].Chinese Journal of Applied Ecology,2025,36(8):2407-2419.
GAOR, FENGW J, ZHANGJ Y, et al. Inversion of winter wheat leaf area index based on Sentinel-2 data and machine learning [J].Shandong Agricultural Sciences, 2025,57(12):145-152.
CHENY Y, XUY K, GUOY Y, et al. Dynamic monitoring of Spartina alterniflora along Jiangsu coastal based on multi-source temporal remote sensing and deep learning model[J].Resources and Environment in the Yangtze Basin,2025,34(10):2210-2221.
CHENY, SHENX J, ZHOUB, et al. Monitoring method of soil salinization in Yinbei irrigation area based on remote sensing images and random forest algorithm[J].Acta Agriculturae Universitatis Jiangxiensis (Natural Sciences Edition),2025,47(3):803-816.
WANGY X, QUZ Y, BAIY Y, et al. Soil salt inversion of typical improvement demonstration area of south bank of the Yellow River based on Sentinel-2 images[J].Transactions of the Chinese Society for Agricultural Machinery,2024,55(4):290-299.
[36]
尹承深.基于多源遥感的河套灌区农田土壤水盐协同反演[D].呼和浩特:内蒙古农业大学,2024.
[37]
YINC S. Synergistic inversion of soil water and salinity in farmland of the Hetao irrigation district based on multi-source remote sensing[D].Hohhot: Inner Mongolia Agricultural University,2024.
HUANGY, LIX Y, GAOH J. Soil salinity inversion with Sentinel-1/2 multi-source remote sensing data[J].Transactions of the Chinese Society for Agricultural Machinery,2025,56(11):640-650.
HEJ B, KONGF P, PEIY D, et al. Remote sensing extraction model of saline-alkali land information based on recursive feature elimination and Bayesian optimization[J].Hydrogeology and Engineering Geology,2026,53(2):246-256.
[44]
JIAP P, ZHANGJ H, LIANGY N, et al. The inversion of arid-coastal cultivated soil salinity using explainable machine learning and Sentinel-2[J].Ecological Indicators,2024,166:e112364.