Accurate PM2.5 concentration prediction is crucial for protecting public health and improving air quality. As a popular deep learning model, deep belief network (DBN) for PM2.5 concentration prediction has received increasing attention due to its effectiveness. However, the DBN structure parameters that have a significant impact on prediction accuracy and computation time are hard to be determined. To address this issue, a modified grey wolf optimization (MGWO) algorithm is proposed to optimize the DBN structure parameters containing number of hidden nodes, learning rate, and momentum coefficient. The methodology modifies the basic grey wolf optimization (GWO) algorithm using the nonlinear convergence and position update strategies, and then utilizes the training error of the DBN to calculate the fitness function of the MGWO algorithm. Through the multiple iterations, the optimal structure parameters are obtained, and a suitable predictor is finally generated. The proposed prediction model is validated on a real application case. Compared with the other prediction models, experimental results show that the proposed model has a simpler structure but higher prediction accuracy.

Two significant uncertainties that are crucial for landslide susceptibility prediction modeling are attribute interval numbers (AIN) division of continuous landslide impact factors in frequency ratio analysis and various susceptibility prediction models. Five continuous landslide impact factor interval attribute classifications (4, 8, 12, 16, 20) and three data-driven models (deep belief networks (DBN), random forest (RF), and neural network (back propagation (BP)) were used for a total of fifteen different scenarios of landslide susceptibility prediction studies in order to investigate the effects of these two factors on modeling and perform a landslide susceptibility index uncertainty analysis (including precision evaluation and statistical law). The findings indicate that: (1) The results demonstrate that for the same model, as the interval attribute value rises from 4 to 8 and finally to 20, the forecast accuracy of landslide susceptibility initially increases gradually, then progressively grows until stable. (2) The DBN model, followed by the RF and BP models, provides the highest prediction accuracy for the same interval attribute value. (3) AIN = 20 and DBN models have the highest prediction accuracy under 15 combined conditions, while AIN = 4 and BP models have the lowest. The accuracy and efficiency of landslide susceptibility modeling are higher when the AIN = 8 and DBN models are combined. (4) The landslide susceptibility index uncertainty predicted by the deeper learning model and the bigger interval attribute value is comparatively low, which is more in line with the real landslide probability distribution features. The conditions that the environmental component attribute interval is divided into eight parts and DBN models are used allow for the efficient and accurate construction of the landslide susceptibility prediction model.

In order to characterize the spatial–temporal properties of postseismic slip motions associated with the 2015 Illapel earthquake, the daily position time series of 13 GNSS sites situated at the near-field region are utilized. Firstly, a scheme of postseismic signal extraction and modeling is introduced, which can effectively extract the postseismic signal with consideration of background tectonic movement. Based on the extracted postseismic signal, the spatial–temporal distribution of afterslip is inverted under the layered medium model. Compared with coseismic slip distribution, the afterslip is extended to both deep and two sides, and two peak slip patches are formed on the north and south sides. The afterslip is mainly cumulated at the depth of 10–50 km, and the maximum slip reaches 1.46 m, which is situated at latitude of − 30.50°, longitude of − 71.78°, and depth of 18.94 m. Moreover, the postseismic slip during the time period of 0–30 days after this earthquake is the largest, and the maximum of fault slip and corresponding slip rate reaches 0.62 m and 20.6 mm/day. Whereas, the maximum of fault slip rate during the time period of 180–365 days is only around 1 mm/day. The spatial–temporal evolution of postseismic slip motions suggests that large postseismic slip mainly occurs in the early stage after this earthquake, and the fault tend to be stable as time goes on. Meanwhile, the Coulomb stress change demonstrate that the postseismic slip motions after the Illapel earthquake may be triggered by the stress increase in the deep region induced by coseismic rupture.

In this paper, we perform a comprehensive analysis of contemporary three-dimensional crustal deformations over the Tibetan Plateau. Considering that the coverage of continuous GNSS sites in the Tibetan Plateau is sparse, a newly designed method that mainly contains Spatial Structure Function (SSF) construction and Median Spatial Filtering (MSF) is adopted to conduct GNSS imaging of point-wise velocities, which can well reveal the spatial pattern of vertical crustal motions. The result illustrates that the Himalayan belt bordering Nepal appears significant uplift at the rates of ~3.5 mm/yr, while the low-altitude regions of Nepal and Bhutan near the Tibetan Plateau are undergoing subsidence. The result suggests that the subduction of the Indian plate is the driving force of the uplift and subsidence in the Himalayan belt and its adjacent regions. Similarly, the thrusting of the Tarim Basin is the main factor of the slight uplift and subsidence in the Tianshan Mountains and Tarim Basin, respectively. In addition, we estimate the strain rate changes over the Tibetan Plateau using high-resolution GNSS horizontal velocities. The result indicates that the Himalayan belt and southeastern Tibetan Plateau have accumulated a large amount of strain rate due to the Indian-Eurasian plate collision and blockage of the South China block, respectively.