The extraction of underground mineral sources has a significant negative impact on the local environment, results in land surface subsidence. As far as subsidence monitoring technology is concerned, leveling is the most accurate. However, leveling can only obtain discrete point data but not the whole area information of the subsidence basin. In this study, Differential Interferometric Synthetic Aperture Radar (D-InSAR) combined with Unmanned Aerial Vehicle (UAV) technology is used to study the subsidence characteristics of the whole working panel. In this analysis, the Huainan mining area is tested as a research area; measured data are compared to the elevation accuracy of the Digital Surface Model (DSM) data, which can be considered for the subsequent works of the mining area. Based on the subsidence affected area, the ground object-type information is recorded to provide basic information for the ecological restoration work after mining so that the data before and after mining can be obtained synchronously. Finally, the differential interference results and Digital Orthophoto Map (DOM) data are combined to assess the spatiotemporal evolution of working panel subsidence and its influence on surface features. The main novelty of the proposed work is combining UAV and D-InSAR to get more accurate analysis of mining subsidence. It can be done using the proposed method.

Ground surface monitoring (GSM) points collect information for mining surface subsidence monitoring and environmental governance. However, GSM points submerge in high groundwater mining areas, preventing the collection of monitoring data. The application of machine learning (ML) algorithms to subsidence prediction ignores the uncertainty and irregularity in subsidence changes. Thus, an innovative GSM point information prediction model, which improves the multikernel support vector machine (GA-MK-SVM) using chaos residual theory commonly used for capturing GSM point information, is proposed. The mean relative errors (MREs) between the predicted and observed results of GA-SVM and GA-MK-SVM were 8.2% and 6.1% during active periods, respectively. The GA-MK-SVM also performed better than the GA-SVM during stable periods. The residual error accumulates as the ML algorithms progress, resulting in imprecise predictions of the GSM points. Thus, the GA-MK-SVM model was improved using chaotic theory (Chaos-GA-MK-SVM), with MREs of 5.0% and 0.9% during the active and stable periods, respectively. The accuracy of the proposed model was improved by 1.1% and 3.2% compared with the unimproved GA-MK-SVM, respectively. The proposed approach provides practical GSM point information for mining subsidence studies and governance in high groundwater mines.

Atmospheric water vapor is an essential source of information that predicts global climate change, rainfall, and disaster-natured weather. It is also a vital source of error for Earth observation systems, such as the global navigation satellite system (GNSS). The Zenith Tropospheric Delay (ZTD) plays a crucial role in applications, such as atmospheric water vapor inversion and GNSS precision positioning. ZTD has specific temporal and spatial variation characteristics. Real-time ZTD modeling is widely used in modern society. The conventional back propagation (BP) neural network model has issues, such as local, optimal, and long short-term memory (LSTM) model needs, which help by relying on long historical data. A regional/single station ZTD combination prediction model with high precision, efficiency, and suitability for online modeling was proposed. The model, called KR-RBF-LSTM, is based on the machine learning algorithms of radial basis function (RBF) neural network, assisted by the K-means cluster algorithm (K-RBF) and LSTM of real-time parameter updating (R-LSTM). An online updating mechanism is adopted to improve the modeling efficiency of the traditional LSTM. Taking the ZTD data (5 min sampling interval) of 13 international GNSS service stations in southern California in the United States for 90 consecutive days, K-RBF, R-LSTM, and KR-RBF-LSTM were used for regions, single stations, and a combination of ZTD prediction models regarding research, respectively. Real-time/near real-time prediction results show that the root-mean-square error (RMSE), mean absolute error (MAE), coefficient of determination (R2), and training time consumption (TTC) of the K-RBF model with 13 station data are 8.35 mm, 6.89 mm, 0.61, and 4.78 s, respectively. The accuracy and efficiency of the KR-RBF-LSTM model are improved compared with those of the conventional BP model. The RMSE, MAE, R2, and TTC of the R-LSTM model with WHC1 station data are 6.74 mm, 5.92 mm, 0.98, and 0.18 s, which improved by 67.43%, 66.42%, 63.33%, and 97.70% compared with those of the LSTM model. The comparison experiments of different historical observation data in 24 groups show that the real-time update model has strong applicability and accuracy for the time prediction of small sample data. The RMSE and MAE of KR-RBF-LSTM with 13 station data are 4.37 mm and 3.64 mm, which improved by 47.70% and 47.20% compared to K-RBF and by 28.48% and 31.29% compared to R-LSTM, respectively. The changes in the temporospatial features of ZTD are considered, as well, in the combination model.