The increasing incidence of retrogressive thaw slumps (RTSs) in permafrost regions underscores the need for detailed spatial and temporal analysis using InSAR technology to monitor and predict dynamic changes in the process of RTSs. Nevertheless, current InSAR deformation forecasting methods employing deep learning strategies such as the traditional long short-term memory (LSTM) and recent transformer models encounter difficulties in effectively capturing temporal features. Moreover, they are limited in their ability to directly integrate spatial information. In this paper, an innovative deep learning approach named Spacetimeformer is proposed for predicting medium- and short-term InSAR deformation of RTSs in the Chumar River area. This method employs a transformer architecture with a spatiotemporal attention mechanism, which enhances the long-term prediction capabilities of time series models and dynamic spatial modeling. It is applicable to multivariate InSAR spatiotemporal deformation prediction problems. The findings include a list of 72 RTSs compiled based on derived InSAR deformation maps and Sentinel-2 optical images, of which 64 have an average deformation rate exceeding 10 mm/year, indicating signs of permafrost degradation. The density distribution of the displacement maps predicted by the Spacetimeformer model aligned well with the InSAR deformation maps obtained from the small baseline subset (SBAS) method, with the overall prediction deviation controlled within 20 mm. In addition, the point-scale prediction results were compared with LSTM and transformer models. This study indicates that the Spacetimeformer network achieved good results in predicting the deformation of RTSs, with a root mean square error of 1.249 mm. The Spacetimeformer method for deformation prediction with the spacetime mechanism presented in this study can serve as a general framework for multivariate deformation prediction based on InSAR results. It can also quantitatively assess the spatial deformation characteristics and deformation trends of RTSs.

The accurate detection of railway tracks from unmanned aerial vehicle (UAV) images is essential for intelligent railway inspection and the development of electronic railway maps. Traditional computer vision algorithms struggle with the complexities of high-precision track extraction due to challenges such as diverse track shapes, varying angles, and complex background information in UAV images. While deep learning neural networks have shown promise in this domain, they still face limitations in precisely extracting track line edges. To address these challenges, this paper introduces an improved NL-LinkNet network, named NL-LinkNet-SSR, designed specifically for railway track detection. The proposed NL-LinkNet-SSR integrates a Sobel edge detection module and a SimAM attention module to enhance the model’s accuracy and robustness. The Sobel edge detection module effectively captures the edge information of track lines, improving the segmentation and extraction of target edges. Meanwhile, the parameter-free SimAM attention module adaptively emphasizes significant features while suppressing irrelevant information, broadening the model’s perceptual field and improving its responsiveness to target areas. Experimental results show that the NL-LinkNet-SSR significantly outperforms the original NL-LinkNet model across multiple key metrics, including a more than 0.022 increase in accuracy, over a 4% improvement in F1-score, and a more than 3.5% rise in mean Intersection over Union (mIoU). These enhancements suggest that the improved NL-LinkNet-SSR offers a more reliable solution for railway track detection, advancing the field of intelligent railway inspection.

An earthquake lethality model was employed to assess the casualty distribution in Yunnan, Guizhou, and Sichuan provinces, taking into account the ground motion acceleration with different 50-year exceedance probabilities. When the probability is 63%, fatalities are predominantly concentrated in central and south-western Yunnan, as well as central, southern, and western Sichuan. At a 10% probability, the peaks of the casualties are observed in southern, eastern, and central Sichuan. In Yunnan (excluding the northwest and southeast regions), the casualty density exhibits unevenness, whereas Guizhou experiences relatively low casualties (except in the eastern and western mountainous areas). Xichang incurs the most substantial losses, followed by Lancang. Xundian, Songming, and Dongchuan demonstrate a high propensity for fatalities, and the risk is relatively high in the vicinity of the Longjiang and Nujiang faults. If a destructive earthquake occurs near these areas within the next 50 years, the probability of a Level-I emergency response exceeds 10%. When the ground motion acceleration doubles (especially when the exceedance probability drops to 2% in 50 year and 0.1% in a year), the predicted number of casualties remains relatively stable. However, the grid of the casualty population exhibits a higher degree of spatial concentration of casualties, and the disaster-affected area expands. There exists no linear correlation between earthquake-induced fatalities and the ground motion level. When the 50-year exceedance probability decreases from 63% to 10%, the casualty rate may increase by several dozen times.

Based on GNSS observations from the Crustal Movement Observation Network of China (CMONOC) spanning 1998–2024, this study investigates crustal deformation characteristics within the northern segment of the North–South Seismic Belt in China. The analysis incorporates the horizontal velocity field, strain rate field, sub-block strain rates, fault slip rates, and associated slip deficits. The results reveal extensional deformation across five tectonic units: the eastern Qiangtang Block, the central–eastern Bayan Har Block, the central Qilian Block, the northern Alxa Block, and the north–central Longzhong Basin. Compressional deformation is predominantly localized along the northern and southern boundaries of the Qilian Block, the western margin of the southern Qaidam Block, the eastern segment of the East Kunlun Fault, and the Hoh Xil region. Zones of elevated shear strain rate are principally concentrated in four regions: the junction of the northern Qilian Block and the Alxa Block, the southern Qaidam Block, the central–eastern Bayan Har Block, and the eastern Qiangtang Block. Historical peak values of maximum principal extensional-shear strain are identified in the northern Longzhong Basin, whereas peak compressional-shear strain occurs in the central Qilian Block–western Qaidam Block and the eastern Qiangtang Block. The analysis reveals heterogeneous fault locking along the northern Western Qinling Range. Most segments of the East Kunlun Fault exhibit slip deficits, with the exception of the Alak Lake–Tusuo Lake segment. The degree of locking is generally homogeneous, except for the Tusuo Lake and Maqu segments. Both the slip deficit and slip rate of the Liupanshan Fault exhibit an increasing trend along its strike. Regions with high seismic hazard potential are identified, primarily including: the Xiugou–Maqiu segment of the East Kunlun Fault, the Guyuan–Jingyuan segment of the Liupanshan Fault, the Hasishan–Machangshan segment of the Haiyuan Fault, the Kuban Mountain segment of the Altyn Tagh North Marginal Fault, the Yumen and Sunan segments of the Qilian North Marginal Fault, the Tianshui and Luanfeng segments of the West Qinling North Marginal Fault, the eastern Altyn Tagh Fault Zone, the central–western East Kunlun Fault Zone, the northern boundary of the Qilian Block, the northern Zhuanglanghe Fault, the southern Haiyuan Fault, the northern Liupanshan Fault, and the eastern West Qinling North Marginal Fault Zone.

This research employs the ITRF2014 framework to conduct an analysis of Eurasian GNSS data, thereby acquiring the recent velocity and strain rate fields in the Ningxia–Inner Mongolia border region, as well as in the Yinchuan and Lingwu regions of western China. It also provides in-depth insights into the relative movements, slip rates, and locking positions of active faults in these areas. Based on the elastic dislocation theory, a constraint model is constructed to perform numerical simulations of active faults. This enables the inversion of strike–slip, dip–slip, and slip deficit rates, along with the locking degree. Fault lockings are identified at the southern extremities of the northern and southern segments of the Niushoushan–Luoshan fault, the western ends of the Xiangshan–Tianjingshan fault zone, and the eastern segment of the Gancanling Fault. In Yinchuan, local locking is detected below 15 km along the southern part of the Helan Mountain eastern fault. The Bayannula Mountain Fault exhibits an uneven locking distribution, while the locking degree of the western margin fault of Zhuozi Mountain increases from the southwest to the northeast, with the fault being locked in the Etuoke Banner section.