Objective Soil salinization has become a major obstacle constraining sustainable agricultural development and regional ecological balance. Given the significant spatiotemporal variations in soil salinity across different seasons, this study integrates remote sensing technology and machine learning methods based on seasonal differences in spectral characteristics, aiming to provide precise support for agricultural production decision-making and land resource management. Methods Lihe Township in the Ningxia Hui Autonomous Region was selected as the study area. A multi-source remote sensing dataset was constructed based on Sentinel-1 SAR backscatter coefficients, Sentinel-2 multispectral reflectance and derived indices (salinity indices, vegetation indices, texture features), Landsat-9 land surface temperature (LST), and SRTM topographic factors. Spectral variables were preliminarily screened using correlation analysis and the coefficient of determination (R²). The successive projections algorithm (SPA) was then applied to reduce the dimensionality of the feature space, thereby optimizing the predictor variable set. Based on this, soil salinity inversion models were developed for the bare soil period (spring) and the vegetation-covered period (autumn) using random forest (RF), support vector regression (SVR), and backpropagation neural network (BPNN), respectively. Results 1) During the bare soil period, Sentinel-2 original band reflectance and salinity indices contributed most significantly, while topographic factors and vegetation indices were the primary influencing factors for soil salinity inversion during the vegetation-covered period. 2) Comprehensive comparison of model performance revealed that the RF model performed excellently on both the training and testing sets, with R² reaching 0.92 for the bare soil period and 0.94 for the vegetation-covered period. Conclusion The combined application of multi-source remote sensing data and machine learning can achieve relatively ideal soil salinity inversion accuracy under different seasonal conditions, thereby providing a scientific basis and technical support for monitoring and managing soil salinization in the region.
JIAZ Z, TANY N, GUANX Y, et al. Saline-alkali soil formation and its remediation strategies in different regions of Ningxia: A comprehensive review[J].Journal of Irrigation and Drainage,2023,42(5):122-134.
[3]
TOMAZA, PALMAP, FIALHOS, et al. Risk assessment of irrigation-related soil salinization and sodification in Mediterranean areas[J].Water,2020,12(12):e3569.
[4]
STAVII, THEVSN, PRIORIS. Soil salinity and sodicity in drylands: A review of causes, effects, monitoring, and restoration measures[J].Frontiers in Environmental Science,2021,9:e712831.
YANGL P, RENJ, WANGY, et al. Soil salinity estimation model in Juyanze based on multi-source remote sensing data[J].Transactions of the Chinese Society for Agricultural Machinery,2022,53(11):226-235.
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.
CUIX T, JINY H, LANF, et al. Inversion of salt content in winter wheat soil based on multi-spectral images[J].Journal of Hebei Agricultural University,2025,48(1):107-115.
ZHANGZ T, WEIG F, YAOZ H, et al. Soil salt inversion model based on UAV multispectral remote sensing[J].Transactions of the Chinese Society for Agricultural Machinery,2019,50(12):151-160.
LIUX, HUY, ZHANGS W, et al. Comparison of different salinity estimation models for salinized soils on south bank of Yellow River in dalat banner[J].Transactions of the Chinese Society for Agricultural Machinery,2024,55(10):360-370.
JIANGC L, ZHAOJ Y, DINGY Y, et al. Study on soil water retrieval technology of Yellow River source based on SPA algorithm and machine learning[J].Spectroscopy and Spectral Analysis,2023,43(6):1961-1967.
YINH Y. Inversion model of soil salt content in Hetao irrigation area based on Sentinel series satellite remote sensing data[D].Yangling, Shaanxi: Northwest A&F University,2023.
[19]
YUEJ B, YANGG J, TIANQ J, et al. Estimate of winter-wheat above-ground biomass based on UAV ultrahigh-ground-resolution image textures and vegetation indices[J].ISPRS Journal of Photogrammetry and Remote Sensing,2019,150:226-244.
QIZ S, FUX, ZHANGH R, et al. Remote sensing inversion and spatial distribution characteristics of soil salinity in the Yellow River irrigation area of the Yellow River delta[J].Journal of University of Jinan (Science and Technology),2025,39(3):339-345.
[22]
WANGN, PENGJ, XUEJ, et al. A framework for determining the total salt content of soil profiles using time-series Sentinel-2 images and a random forest-temporal convolution network[J].Geoderma,2022,409:e115656.
ZHUW X, SUNZ G, LIB B, et al. Analysis of spatial heterogeneity for soil attributes and spectral indices-based diagnosis of coastal saline-alkaline farmland stress using UAV remote sensing[J].Journal of Geo-Information Science,2021,23(3):536-549.
HONGG J, XIEJ B, ZHANGL, et al. Monitoring soil salinization of cotton fields in the Aral Reclamation Area using multispectral imaging[J].Arid Zone Research,2024,41(5):894-904.
GAOL, WANGX F, GUX F, et al. Exploring the influence of soil types underneath the canopy in winter wheat leaf area index remote estimating[J].Chinese Journal of Plant Ecology,2017,41(12):1273-1288.
ZHAOQ Z, DINGJ L, HANL J, et al. Exploring the application of MODIS and Landsat spatiotemporal fusion images in soil salinization: A case of Ugan River-Kuqa River Delta Oasis[J].Arid Land Geography,2022,45(4):1155-1164.
DINGB B. A method for estimating soil salinity using unmanned aerial vehicle multispectral remote sensing based on deep learning[D].Yangling, Shaanxi: Northwest A&F University,2024.
[33]
ZHANGQ Q, LIL, SUNR Z, et al. Retrieval of the soil salinity from sentinel-1 dual-polarized SAR data based on deep neural network regression[J].IEEE Geoscience and Remote Sensing Letters,2020,19:e4006905.
[34]
WANGJ J, WUF, SHANGJ L, et al. Saline soil moisture mapping using Sentinel-1A synthetic aperture radar data and machine learning algorithms in humid region of China's east coast[J].Catena,2022,213:e106189.
[35]
WANGN, XUEJ, PENGJ, et al. Integrating remote sensing and landscape characteristics to estimate soil salinity using machine learning methods: A case study from southern Xinjiang, China[J].Remote Sensing,2020,12(24):e4118.
NINGK X, SUNC X, WANH W, et al. Research on the inversion method of desert grassland fractional vegetation cover based on collaborative UAV-satellite remote sensing[J].Remote Sensing Technology and Application,2024,39(5):1054-1063.
[38]
HUANGY, ZHAOL. Review on landslide susceptibility mapping using support vector machines[J].Catena,2018,165:520-529.
LUQ, SUK X, ZHANGJ W, et al. Structural optimization of simple catenary based on BP neural network and genetic algorithm[J].Journal of Mechanical Engineering,2024,60(12):313-320.
Office of the Leading Group for the Third National Soil Census, CouncilState. Technical specifications for the calibration of soil type names in the Third National Soil Census (Revised Edition)[EB/OL].(2023-02-27).
[43]
CHENGG, DINGH, YANGJ, et al. Crop type classification with combined spectral, texture, and radar features of time-series Sentinel-1 and Sentinel-2 data[J].International Journal of Remote Sensing,2023,44(4):1215-1237.
[44]
LIJ, ZHANGT T, SHAOY, et al. Comparing machine learning algorithms for soil salinity mapping using topographic factors and sentinel-1/2 data: A case study in the Yellow River Delta of China[J].Remote Sensing,2023,15(9):e2332.
[45]
CALLOWJ N, HIPSEYM R, VOGWILLR I J. Surface water as a cause of land degradation from dryland salinity[J].Hydrology and Earth System Sciences,2020,24(2):717-734.
[46]
WANGP, ZHANGN C. Decision tree classification algorithm for non-equilibrium data set based on random forests[J].Journal of Intelligent and Fuzzy Systems,2020,39(2):1639-1648.