Remote sensing monitoring of soil salinization in Hongshaquan open-pit mining area

Hongshaquan open-pit coal mine located in the desertification area of western China is one of the important coal production bases in China, and salinization is an important factor affecting the ecological environment in this area. Based on Landsat (30 m), Sentinel-2 (10 m) and UAV (0.188 m) multi-spectral images, 10 salinization indexes were calculated in the Hongshaquan mining area, and the correlation analysis was conducted between them and measured conductivity. The optimal indexes were selected according to the correlation coefficient. The index selected, the Gradient-based Structural Similarity index (GSSIM), and unitary linear regression were used to monitor the dynamic changes of soil salinization in the study area. The results show that:① the correlation coefficients between the measured soil EC and the S 3 index calculated by UAV, Landsat, and Sentinel-2 were 0.703, 0.665, and 0.723(P < 0.01), respectively. The robustness of S₃ is better than the other 9 indices. Overall, the S₃ can be employed to monitor soil salinization in the study area. ② From the spatio-temporal distribution and transfer of the S₃ index, before mining (1988-2006), the non-salinized areas were scattered in the east and north of the study area, and the heavily salinized areas were mainly distributed in the south of the study area, and the two types were relatively stable. The mildly salinized soil transferred the largest amount (about 92.37 km²), followed by moderately salinized soil (16.25 km²), especially in 2006. In the early stage of mining (2006-2010), the non-salinized area in the open pit increased, and the light, moderate, and severe saline soils were mainly transferred to non, light, and moderate salinized soil. With the intensification of mining (2010-2020), the severe salinization area increased significantly in the mine dump, and the non-salinization area showed a trend of first increasing and then decreasing in the mine pit with the mining and filling of the mine pit. Overall, the transfer amount from salinized soil to non-salinized soil (about 6.51 km²) is less than that from non-salinized soil to salinized soil (24.70 km²), and soil salinization is intensified. ③ The GSSIM monitoring results showed that the mutation areas were mainly distributed in the middle and north of the study area during 1988-2006 (before mining), and scattered around the medium variation area, while the low variation area was widely distributed in the study area. From 2006 tO₂010 (the initial stage of mining), the mutation area shifted to the north, and the low, medium, and mutation areas were dominated by reduction. From 2010 tO₂020, the mutation areas increased significantly (increased by 7.10%), and concentrated around the mine, indicating that the salinity in the mine has changed significantly. From 1988 tO₂020, low, medium, and mutation areas accounted for 25.02%, 38.52%, and 36.47%, respectively, and the mutation areas were concentrated around mines and power plants. According to the analysis of ground objects and GSSIM images, the area covered by cement and coal mine will reduce salinization, while the long-term accumulation of dump and unexploited sandstorm accumulation area may aggravate salinization. In addition, by comparing slope images, the mid-variation area of the GSSIM mutation area is corresponding to the position of the significant and insignificant change in the slope image around the mine, indicating that the GSSIM can quantitatively analyze the spatio-temporal variation rule of salinization.