Jump to Main Content
Assessment of mining-related seabed subsidence using GIS spatial regression methods: a case study of the Sanshandao gold mine (Laizhou, Shandong Province, China)
- Cao, Jiayuan, Ma, Fengshan, Guo, Jie, Lu, Rong, Liu, Guowei
- Environmental earth sciences 2019 v.78 no.1 pp. 26
- case studies, digital elevation models, geographic information systems, global positioning systems, gold, mining, monitoring, regression analysis, spatial variation, subsidence, China
- Land subsidence in the Sanshandao area, Laizhou, Shandong Province, China, has been a consequence of underground gold mining. This paper identifies the statistically significant mining subsidence factors, which are: (1) a digital elevation model of the surface; (2) the surface slope; (3) the slope aspect; (4) the thickness of the gold deposits; and (5) the depth of the gold deposits below the ground. The vertical displacement of the GPS monitoring in the Xishan gold mine (one of the Sanshandao gold mine) was selected as the dependent variable and five mining subsidence factors as the independent variables. Subsidence modeling was carried out in geographic information systems first with the ordinary least squares (OLS) method and then with the geographically weighted regression (GWR) method. Finally, the seabed subsidence was predicted with the geographically weighted regression model for the Xinli gold mine (another of the Sanshandao gold mine), in which the gold deposits are located under the sea. The results of the GWR analysis showed a marked improvement compared to those of the OLS analysis. The R² value of the GWR model equals 0.82, which indicates that the model captured the spatial heterogeneity of the independent variables. The accuracy of determining subsidence in the area used for validation is ± 8.5 mm with a maximum calculated subsidence of − 329.26 mm. The maximum subsidence predicted with the model for the seabed is − 63 mm with a mean subsidence of − 50 mm.