This study presents an integrated framework combining a real probability-based approach with automated machine learning (AutoML) to quantitatively assess the landslide risk of the southern segment of the Beijing–Guangzhou Railway under heavy rainfall. A total of 6660 landslides triggered by a single rainfall event were analyzed, incorporating twelve key influencing factors, including accumulated rainfall, topography, and geological conditions. A sample database was constructed based on the real probability criterion, and three models, namely XGBoost, Random Forest (RF), and Extra Trees (ET), were efficiently optimized using the FLAML framework. Among them, XGBoost yielded the best performance, achieving the highest AUC and demonstrating the strongest spatial agreement between predicted high-risk zones and observed landslide occurrences. Feature importance analysis identified accumulated rainfall, strata, and elevation as the dominant contributing factors. The results reveal that the overall risk of the railway section during this event was primarily classified as extremely low or very low. Notably, areas experiencing intense precipitation still pose significant risks to railway operations due to potential rapid-onset landslides. This research provides a data-driven, scalable tool for quantitative disaster risk assessment and offers a foundation for the development of a dynamic early warning platform by incorporating real-time rainfall forecasts.
Urban pluvial floods are attracting growing concern due to rising intense precipitation and increasing consequences, particularly for underground transit networks. Accurate risk assessment is critical to efficient urban pluvial flood management. This paper describes an integrated methodology to evaluate the potential impact and risk of pluvial flooding on major metro systems in the Guangdong-Hong Kong-Macao Greater Bay Area (GBA). We use a high-resolution hydrodynamic model (FloodMap-HydroInundation2D) coupled with complex network analysis to simulate surface inundation and assess topological vulnerability under 10-, 50-, and 100-year rainfall return periods. Results suggest that the pluvial flood hazard increases non-linearly with escalating extreme rainfall, with Shenzhen and Hong Kong, China demonstrating the most pronounced hazard levels. Spatially, medium- and high-risk stations are predominantly concentrated in the central urban districts. As the flood hazard intensifies, the number of at-risk stations increases significantly, with the Guangzhou-Foshan Metro systems and the Shenzhen Metro system expanding most drastically. Network performance metrics of all metro systems deteriorate as the flood hazard escalates; the global network efficiency of the Guangzhou-Foshan Metro systems remains the lowest among all networks. The tested approach provides detailed flood information for the strategic management of urban underground transit networks. It can be widely applied to other megacities to support sustainable climate adaptation.
Hydrothermal geothermal resources are abundant and widely distributed across China, with karst thermal water emerging as a key geothermal medium. The Luxi area, situated within the North China Craton karst belt, is a critical area for hydrothermal resource. However, its geothermal regime and controlling factors remain poorly understood. This study analyzed data from 26 newly drilled geothermal wells and thermal conductivity measurements from 46 Ordovician carbonate rocks to characterize the geothermal regime within different structural units and guide future exploration. Findings revealed significant spatial variations in geothermal characteristics: the QH field, near recharge zones, is dominated by thermal convection, with a mean temperature of 47.4°C, geothermal gradient of 23.55°C·km−1, and heat flow of 60.3 mW·m−2. Conversely, the FX field, influenced by deep faults, exhibits higher values (mean temperature: 62.8°C, gradient: 25.20°C·km−1, heat flow: 66.9 mW·m−2). The JC field, located in the central Heze uplift, presents intermediate values. Fault zones play a pivotal role in heat accumulation, particularly in shallow layers (< 3000 m), while deeper layers show exponential temperature variations along fault margins. Groundwater facilitates vertical heat convection, strongly correlating with the thickness of overlying clastic rock layers (R2 = 0.772, P = < 0.0001). Tight clastic rocks, such as Upper Paleozoic mudstone, act as effective seals for Ordovician karst reservoirs. This study enhances understanding of geothermal characteristics in Luxi area, providing theoretical guidance for resource exploration.
This study employs the Soil and Water Assessment Tool (SWAT) to investigate the dynamics of runoff and sediment in the Fuhe River Basin of Poyang Lake. After calibration and validation, the model’s coefficient of determination (R2) for both runoff and sediment yield exceeded 0.9, indicating high model accuracy. During the study period from 2001 to 2010, the cropland area in the Fuhe River Basin decreased by 195.9 km2, while forest and urban land areas increased by 105.4 km2 and 86.1 km2, respectively. By inputting the multi-year LULC data from 2001 to 2010 into the SWAT model, we assessed the impact of LULC changes on the multi-year water and sediment yields of the Fuhe River Basin. The results showed annual differences in water and sediment yields, both below 10 mm and 2 t/ha2, respectively. At the sub-basin scale, LULC changes had a significant impact on water and sediment yields. Under the 2010 baseline landuse scenario, the simulated runoff and sediment yields were 837.11 m3/s and 4.32 × 106 t, respectively. Compared to the baseline scenario, the reforestation scenario resulted in reductions in water and sediment yields by −1.0
The comprehensive investigation of sediment transport during flood events offers valuable insights into the hydrological and erosion processes of watersheds. It also plays a crucial role in flood disaster prevention and control. In this study, we focused on the northern earth-rocky mountainous areas of China and employed K-medoids clustering to classify a total of 261 flood events spanning from 1959 to 2021 into four distinct types. By comparing the sediment transport characteristics of different flood types and periods, as well as analyzing sediment source distribution using SSC- Q hysteresis loops, we obtained the following results. 1) The study period witnessed a notable decrease in the annual number of flood events and sediment yield, with reductions of 48.09% and 34.01%, respectively. The suspended sediment concentration during flood events exhibited a substantial decline of 76.54% compared to the baseline period. 2) In the Mihe River Basin, the majority of sediment yield could be attributed to flood events classified as Types A and B. These flood types were characterized by short duration, high peak flow, and substantial runoff depth. Among them, Types A were significantly greater than Types B in terms of runoff and sediment transport. 3) The hysteresis loops observed in the Mihe River Basin predominantly displayed a figure-eight and clockwise pattern, indicating potential sediment sources within the river channel and banks. Addressing sediment challenges and ensuring sustainable watershed management practices require connecting these loop characteristics to future management initiatives, given abundant sediment sources near the Huangshan hydrological station.
