节水灌溉Issue(5)：72-78,7.DOI:10.12396/jsgg.2025312
基于多源遥感数据的宁夏河东沙地土壤水分反演研究
Study on Regional Scale Soil Moisture Inversion in Sandy Land in the East of Yellow River in Ningxia Based on Multi-source Remote Sensing Data
摘要
Abstract
In order to accurately understand the distribution characteristics of soil moisture in the sandy land east of the Yellow River in Ningxia,this study explores inversion models and methods for soil moisture content in sandy land,which provides theoretical support and decision-making reference for research on the spatial distribution of soil moisture.Using microwave remote sensing data from Sentinel-1 SAR combined with multispectral data from Sentinel-2,the soil moisture content at different depths(5,10 and 20 cm)in the sandy land east of the Yellow River in Ningxia was inverted by integrating the Water Cloud Model and the Random Forest algorithm.The results showed:①Correlation analysis between different environmental variables and soil moisture content across layers revealed significant differences.Band reflectance(B11,B12)showed a significantly negative correlation with soil moisture at all depths.The normalized difference vegetation index was positively correlated with soil moisture at 5 cm and 10 cm,while the backscattering coefficient(VV+VH)exhibited a significantly negative correlation with soil moisture at 5,10 and 20 cm.②Analysis of model evaluation metrics showed that the coefficient of determination between measured and validated values for soil moisture across the three layers ranged from 0.75 to 0.85,with root mean square errors between 0.02%and 0.04%,indicating that the Random Forest model achieves high accuracy in soil moisture inversion.③ The soil water content in the study area was generally low,and the overall soil water content was less than 3.00%.The inversion results of soil water content show strong consistency with the actual measured values.By integrating Sentinel-1/2 remote sensing data with vegetation indices and backscattering coefficients,the method demonstrates high accuracy in soil moisture inversion for this area and can effectively capture the spatial distribution characteristics of regional moisture.关键词
土壤水分/水云模型/随机森林RF/遥感反演/宁夏河东沙地Key words
soil moisture/water cloud model/random forest RF/remote sensing inversion/sandy land in the East of Yellow River in Ningxia分类
农业科技引用本文复制引用
张维福,展秀丽,马思怡,张呈春,马晓霞..基于多源遥感数据的宁夏河东沙地土壤水分反演研究[J].节水灌溉,2026,(5):72-78,7.基金项目
国家自然科学基金项目"宁夏河东沙地水分在土壤-植被系统演变中的驱动作用"(42161013) （42161013）
宁夏自然科学基金项目(2024AAC03078). （2024AAC03078）
