Yuming Wu’s research while affiliated with Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences and other places

The development of artificial intelligence makes it possible to rapidly segment landslides. However, there are still some challenges in landslide segmentation based on remote sensing images, such as low segmentation accuracy, caused by similar features, inhomogeneous features, and blurred boundaries. To address these issues, we propose a novel deep learning model called AST-UNet in this paper. This model is based on structure of SwinUNet, attaching a channel Attention and spatial intersection (CASI) module as a parallel branch of the encoder, and a spatial detail enhancement (SDE) module in the skip connection. Specifically, (1) the spatial intersection module expands the spatial attention range, alleviating noise in the image and enhances the continuity of landslides in segmentation results; (2) the channel attention module refines the spatial attention weights by feature modeling in the channel dimension, improving the model’s ability to differentiate targets that closely resemble landslides; and (3) the spatial detail enhancement module increases the accuracy for landslide boundaries by strengthening the attention of the decoder to detailed features. We use the landslide data from the area of Luding, Sichuan to conduct experiments. The comparative analyses with state-of-the-art (SOTA) models, including FCN, UNet, DeepLab V3+, TransFuse, TranUNet, and SwinUNet, prove the superiority of our AST-UNet for landslide segmentation. The generalization of our model is also verified in the experiments. The proposed AST-UNet obtains an F1-score of 90.14%, mIoU of 83.45%, foreground IoU of 70.81%, and Hausdorff distance of 3.73, respectively, on the experimental datasets.

The key to seismic landslide risk identification resides in the accurate evaluation of seismic landslide hazards. The traditional evaluation models for seismic landslide hazard seldom consider the landslide dynamic runout process, leading to an underestimation of seismic landslide hazard. Therefore, a joint Newmark–Runout model based on landslide dynamic runout is proposed. According to the evaluation results of static seismic landslide hazard, the landslide source points can be extracted, and the landslide dynamic runout process is simulated to obtain the dynamic seismic landslide hazard. Finally, the static and dynamic seismic landslide hazards are fused to obtain an optimized seismic landslide hazard. In September 2022, a strong Ms6.8 earthquake occurred in the eastern Tibetan Plateau, triggering thousands of landslides. Taking the 2022 Luding earthquake-induced landslide as a sample, the function relationship between seismic slope displacement and landslide occurrence probability is statistically modeled, which partly improves the traditional Newmark model. The optimized seismic landslide hazard evaluation of the Luding earthquake area is conducted, and then, the seismic landslide risk identification is completed by taking roads and buildings as hazard-affected bodies. The results show that the length of the roads facing very high and high seismic landslide risks are 3.36 km and 15.66 km, respectively, and the buildings on the Moxi platform near the epicenter are less vulnerable to seismic landslides. The research findings can furnish critical scientific and technological support for swift earthquake relief operations.

The eastern Himalayan syntaxis (EHS) region is char-acterized by steep topography, strong tectonic activity, and strong disaster activity, and the disaster activity has a temporal correla-tion with the change of seasonal climate factors. Different from other disaster-prone areas, the disasters in the EHS area are more destructive and more susceptible to changes in climate factors. However, previous studies on disaster susceptibility in the region mainly focused on geohazard samples, regional topography, and geological conditions and seldom considered the characteristics of disaster activity and the characteristics of seasonal tempera-ture and rainfall changes, which often led to an underestimation of disaster susceptibility in this region. Therefore, in this study, SBAS-InSAR technology and the GEE (Google Earth Engine) plat-form were used to identify 317 geological disasters and construct the spatial–temporal motion field and spatial-temporal background information of disasters in the last 5 years. Based on the random forest and optimized evaluation data model, the disaster suscepti-bility in the EHS area was quantitatively evaluated. The AUC value of the model was 0.89, indicating that the results have high reli-ability. The results showed that the high-susceptibility areas were mainly distributed in the riverbank and high-altitude areas. The SHAP model revealed that the factors of temperature, precipitation, slope, and elevation in the EHS area had a great negative influence on the disaster. The quantitative results of the influence of seasonal temperature and precipitation on disaster susceptibility showed that the area of susceptibility in summer was significantly higher than that in winter, and the area of high and very high susceptibil-ity increased by about 13%. Finally, the dynamic failure process of typical landslide and glacier disasters in highly susceptible areas was quantitatively predicted, which provides a scientific basis for disaster prevention.

The landslide dynamics model is one of the methods for evaluating landslide motion processes, contributing to disaster prevention and mitigation. With the advancement of science and technology, GIS has become the mainstream platform for landslide simulation. However, the three-dimensional movement of landslides is intricate, leading to a lack of methods for three-dimensional landslide numerical simulation on GIS platforms. In this paper, we propose a three-dimensional, two-phase landslide dynamics model. Through the proposed solution, three-dimensional modeling and numerical simulation of landslides can be achieved on GIS platforms. Simultaneously, drawing inspiration from the SPH kernel functions, we visualize the results of the three-dimensional model on the GIS platform. Simulation of the Yigong landslide demonstrates that our solution can realize three-dimensional landslide simulation on the GIS platform. Our model adeptly captures numerous details in the landslide motion process. However, constrained by the inherent limitations of the three-dimensional model, the model results are susceptible to numerical oscillations and diffusion, with the accuracy of the model being controlled by grid partitioning.

Ice–snow freezing may disrupt the growth condition and structure of forest vegetation, increasing combustible loads and thus triggering forest fires. China’s subtropical regions are rich in forest resources, but are often disturbed by ice–snow freezing, especially due to climate change. Clarifying the responsive areas and times of forest fires to ice-snow freezing in this region is of vital importance for local forest fire management. In this study, meteorological data from 2001 to 2019 were used to extract the precipitation and its duration during the freezing period in order to analyze the freezing condition of forest vegetation in subtropical China. To improve the accuracy of identifying forest fires, we extracted forest fire information year-by-year and month-by-month based on the moderate resolution imaging spectroradiometer (MODIS) active fire data (MOD14A2) using the enhanced vegetation index (EVI), and analyzed the forest fire clustering characteristics in the region using the Moran’s Index. Then, correlation analysis between forest fires and freezing precipitation was utilized to explore the responsive areas and periods of forest fires caused by ice–snow freezing. Our analysis shows the following: (1) during the period of 2001–2019, the ice–snow freezing of forest vegetation was more serious in Hunan, Jiangxi, Hubei, and Anhui provinces; (2) forest fires in subtropical China have shown a significant downward trend since 2008 and their degree of clustering has been reduced from 0.44 to 0.29; (3) forest fires in Hunan, Jiangxi, and Fujian provinces are greatly affected by ice–snow freezing, and their correlation coefficients are as high as 0.25, 0.25, and 0.32, respectively; and (4) heavy ice–snow freezing can increase forest combustibles and affect forest fire behavior in February and March. This research is valuable for forest fire management in subtropical China and could also provide a reference for other regions.

Large paleolandslides are developed in the upper reaches of Jinsha River, which seriously threaten the safety of nearby residents and engineering facilities. It is important to study the movement characteristics of these landslides. In this work, we inspect the deformation characteristics of a rapid landsliding area along the Jinsha River by using multi-temporal remote sensing, and analyzed its future development. Surface deformations and damage features between January 2016 and October 2020 were obtained using multi-temporal InSAR and multi-temporal correlations of optical images, respectively. Deformation and failure signs obtained from the field investigation were highly consistent. Results showed that cumulative deformation of the landsliding area is more than 50 cm, and the landsliding area is undergoing an accelerated deformation stage. The external rainfall condition is an important factor controlling the deformation. The increase of rainfall will accelerate the deformation of slope. The geological conditions of the slope itself affect the deformation of landslide. Due to fault development and groundwater enrichment, slopes are more likely to slide along weak structural plane. The Jinsha River continuously scours the concave bank of the slope, causing local collapses and forming local free surfaces. Numerical simulation results show that once the landsliding area fails, the landslide body may form a 4 km long dammed lake, and the water level could rise about 200 m.

Solar photovoltaic (PV) is favored by the market because of its clean and renewable characteristics. There are abundant solar resources in the tropical regions of China. It is important and necessary to carry out comprehensive analysis of rooftop PV projects for tropical regions for scientific policy-makings. Here, we select Nanning as a case study to analyze the optimal options for PV installation on different roof types and estimate the electricity generation potential of rooftop PVs and its additional returns. Our analysis shows that: 1) the annual optimal azimuth and tilt angle in Nanning are 245° and 32.5°, respectively; 2) the tilt angle in southwest orientation has more space for adjustment, while that in the opposite orientation should remain horizontal; 3) for flat roofs, being fixed at the annual optimal angles is practical, while for gable roofs, the east-west direction is favorable; 4) the total potential of rooftop PV projects in Nanning can reach 19.99 TWh/year, resolving 76.1% of the social electricity demand. This research is valuable for rooftop PV installation and optimization in the tropical regions of China, which also could provide reference for other regions.

China is one of an increasing number of countries in the world that is suffering from frequent and severe natural disasters, which cause serious loss of life. The Chinese government has set up a special financial fund for natural disaster mitigation and reduction. Therefore, based on the financial expenditure data and disaster losses data obtained from ministries of emergency management and the China Statistical Yearbook, we analyzed the spatio-temporal distribution of natural disaster losses at the economic zonal scale during 2010–2020, and then evaluated the efficiency of disaster mitigation and reduction using a DEA model. The results showed that the natural disaster losses decreased significantly in most provinces from 2010 to 2020. The distribution of precipitation is extremely uneven (more in the southeast and less in the northwest). Moreover, the Central and Western Economic Zones are the most earthquake-prone regions in China, especially Xinjiang, Tibet, Sichuan, Yunnan and Gansu. Among all natural disasters, floods were the leading natural disasters, causing the most severe losses in China on the national scale. Furthermore, the cities with higher comprehensive efficiency, mean the ratio between the effects and funding on disaster mitigation and reduction, were either economically developed or geographically large and sparsely populated. Finally, we used an exponential regression equation model to explore the relationship between financial input and direct economic losses caused by natural disasters in 2019 and 2020; we found that there is a negative correlation between the financial investment and the direct economic losses. In conclusion, it is necessary to improve the technology of natural disaster mitigation and reduction and to adjust the scale of investment according to the actual situation of each region and the different disasters in China. This paper aims to provide relevant experience and basis for China’s comprehensive disaster mitigation and reduction work.

Land reclamation in the Yan'an New District (YND) on the Chinese Loess Plateau is one of the largest earthworks projects in the world, involving the excavation of loess from ridges and the deposition of the material in adjacent valleys, flattening an area of more than 78 km². It can take multiple years for the landscape to adjust to the new topography after the earthworks are completed, with subsidence in the fill areas and uplift in the areas of excavation. Understanding the pattern and extent of this differential vertical movement has great importance for ensuring the structural integrity of any infrastructure built on the site. We therefore studied the spatial deformation field of the YND from January 2015 to December 2018 using satellite synthetic aperture radar interferometry (InSAR) technology. Persistent scatterers (PS) and distributed scatterers (DS) were combined in a two-tier network to overcome problems with temporal decorrelation of the InSAR phase signals. The derived deformation field was validated with in-situ ground leveling and GNSS measurements, and interpreted using a forward geophysical model. The results indicate that our approach provides a more detailed understanding of the deformation than conventional PSI and SBAS methods. Subsidence of up to 87 mm/yr occurred over about 4.4 km² of the fill region, and uplift of up to 26 mm/yr in the excavated areas. There is a strong linear relationship between these displacements and prior topographic elevation change. The primary cause of subsidence in fill areas is compaction of remoulded loess. Instantaneous elastic and gradual poroelastic deformation are the main causes of uplift in excavated areas.