土壤湿度作为表征矿区地表环境的重要指标，准确监测荒漠化露天矿的土壤湿度对矿区生态保护和可持续发展具有关键作用。以典型荒漠化露天矿区-准东红沙泉矿区及周边区域为研究对象，基于Landsat 8 OLI影像，利用基于植被信息的归一化植被指数NDVI、比值植被指数RVI、条件植被指数VCI，基于热红外波段的表观热惯量ATI和条件温度指数TCI以及基于反射率特征空间的垂直干旱指数PDI、土壤湿度监测指数SMMI反演红沙泉矿区的土壤湿度，结合实测土壤湿度数据验证得知上述指数的反演结果与实测样点的土壤湿度不符，即传统的土壤湿度指数不能单独作为研究区土壤湿度的指数因子。鉴于以上指数存在的不足和研究区特殊的生态环境，提出新裸岩石砾地指数NGLI，并结合SMMI、NGLI和NDVI通过线性加权构建一种适用于荒漠化露天矿区的土壤湿度模型NMSMI，最后评估其精度和适用性。结果表明：① 构建的裸岩石砾地指数NGLI能够有效提取研究区裸岩石砾地和露天采坑的土壤湿度，NGLI与实测土壤湿度呈显著负相关（R2=0.245,P<0.01），验证点的均方根误差RMSE为0.009；② 基于熵权法和线性加权法，利用SMMI、NDVI、NGLI三种指数构建新型修正土壤湿度监测指数NMSMI，经验证NMSMI与实测土壤湿度呈显著负相关（R2=0.517,P<0.01），验证点的均方根误差RMSE为0.004，且反演结果在空间分布上与实测土壤湿度相符。通过实测数据与各传统指数及NMSMI指数的对比与验证，说明NMSMI指数适用于特殊荒漠化露天矿区的土壤湿度监测。

Soil moisture is an important indicator that characterizes the surface environment of mining areas. Accurate monitoring of soil moisture in desertified open-pit mines plays a key role in ecological protection and sustainable development of mining areas. Taking the typical desertification open-pit mining area-Zhundong Hongshaquan Mining Area and surrounding areas as the research object,based on Landsat 8 OLI image,the normalized vegetation index (NDVI),the ratio vegetation index (RVI),the vegetation conditional  index (VCI),the apparent thermal inertia (ATI) and the Conditional Temperature Index (TCI) based on the thermal infrared band,the perpendicular drought index (PDI) based on the reflectance characteristic space and the soil moisture monitoring index (SMMI) based on the vegetation information were used to retrieve the soil moisture in desertification mining areas. The inversion results are not consistent with the soil moisture of the measured sample points,that is,the traditional soil moisture index cannot be used as an exponential factor of the soil moisture in the study area alone. In view of the shortcomings of the above indexes and the special ecological environment of the study area,the new bare rock gravel index NGLI is proposed,and combined with SMMI,NGLI and NDVI,a soil moisture model NMSMI,was constructed to evaluate its accuracy and applicability,and finally evaluate its accuracy and applicability. The results show that: ① The constructed NGLI (new gravel land index,NGLI) can effectively extract the soil moisture of bare gravel land and the open pit in the study area,and the NGLI is negatively correlated with the measured soil moisture (R2=0.245,P<0.01),the root mean square error RMSE of the verification point is 0.009; ② Based on the linear weighting method,a new modified soil moisture index (NMSMI) was constructed by using SMMI,NDVI,and NGLI. It was verified that the index was negatively correlated with the measured soil moisture (R2=0.517,P<0.01),the root mean square error RMSE of the verification point is 0.004,and the inversion results were consistent with the measured soil moisture in spatial distribution. By comparing and verifying the measured data with the traditional indexes and NMSMI index,it is shown that NMSMI index is suitable for soil moisture monitoring in special desertification open-pit mining areas.