With global climate change and accelerating urbanization, urban flood is becoming more frequent worldwide. Understanding the urban vulnerability is crucial for making decisions on urban flood control. This study uses urban flood susceptibility (UFS) as an indicator, and comprehensively applies three machine learning models, XGBoost, CatBoost and LightGBM, in the Kangyi area of Ordos City. Combined with the Shapley Additive explanations method, the driving mechanism and spatial heterogeneity of flood susceptibility was explored in the study area. The results show: (1) Model performance comparison: All three models have high accuracy, with XGBoost performing well in overall classification (OA = 0.96) and CatBoost performing well in distinguishing flood/non-flood samples (AUC = 0.85). (2) Multi-model adaptability assessment: The proposed “model-factor-space” framework highlights the sensitivity of XGBoost to urbanization indicators, the ability of CatBoost to capture natural geographical elements, and the efficiency of LightGBM in analyzing terrain thresholds. (3) Dynamic thresholds and synergies: Impervious surface density (ISD) is the most critical factor, and when ISD > 0.2, the risk of flooding will continue to increase by 60%. Comprehensive analysis with spatial heterogeneity shows that high-risk areas are mainly affected by ISD, road density (> 10,000 m/km2) and low altitude (< 1300m) in urban built-up areas, while low-to-medium risk areas are sensitive to vegetation coverage (> 7,000) and Dis2Water bodies (> 1,500m). (4) Hierarchical governance strategy: A three-level spatial governance strategy is proposed: in the core area, priority is given to ISD control (< 0.2) and pipe network upgrades; in the transitional area, slope interception and ecological restoration are combined; and in the potential risk area, a multi-scale monitoring and early warning system is established for multi-scale monitoring.

Velocity changes ($dv/v$) during and after earthquakes are important indicators for understanding the earthquake-induced mechanical damage evolution of rock slopes. However, studying slope responses associated with various seismic loading still remains challenging due to limited in situ observations. In this article, we conduct a 20 min temporal resolution monitoring of $dv/v$ at the frequency band between 2 and 20 Hz by applying ambient noise interferometry on the Pubugou rock slope in Southwest China. We observe an instantaneous ∼0.41% $dv/v$ drop on the slope caused by the 2021 M 7.4 Maduo earthquake at a distance of 780 kM, following a characterized logarithmic recovery process of ∼31.39 h towards its pre-earthquake state. Moreover, the $dv/v$ in five narrow frequency bands show a similar drop and subsequently increased recovery times associated with the deceased frequencies due to the long-distant earthquake shaking. We discuss two possible mechanisms related to the heterogeneous rock slope excited by the long-distant earthquake at a low frequency. The study motivates the damage assessment of the rock slope using in situ $dv/v$ and furthers the understanding of subsurface geological risks under diverse seismic loadings.

As global climate change intensifies, the global atmospheric circulation process is undergoing significant changes, and the local water vapor pattern has also changed. This study takes the Wei River Basin as the research area. Firstly, an evaluation index system for extreme precipitation was established, and the time-series characteristics of the magnitude, frequency, and duration of extreme precipitation were analyzed. Statistical methods were used to analyze the non-consistency in time-series changes in extreme precipitation indicators. Using spatial heterogeneity analysis methods, the spatial variation differences in extreme precipitation in the Wei River Basin were identified. This study selected the El Niño-Southern Oscillation (ENSO) index, global land-ocean temperature index (LOTI), and land surface temperature (LST) index to quantitatively evaluate the impact of climate change on regional extreme precipitation and analyzed the correlation between temperature and extreme precipitation, identifying the key driving factors of extreme precipitation changes. The conclusions of this study are as follows: (1) The southern region of the Wei River Basin experiences more frequent and intense precipitation events, while the northern region experiences relatively few. (2) From 1981 to 2021, the intensity, frequency, and duration of precipitation events in the Wei River Basin gradually increased, with the most significant increase in extreme precipitation in the Guanzhong Plain. (3) Global climate change has an important impact on precipitation events in the Wei River Basin. The increase in the ENSO, LOTI, and LST indices may indicate an increase in the probability of drought and flood events in the Wei River Basin. The relationships between extreme precipitation and temperature present a peak structure. This conclusion is helpful to better understand the impact of climate change on extreme precipitation in the Wei River Basin and provides some support for the response to extreme meteorological events under the background of future climate change.

Climate change intensifies Maximum Check Flood (MCF) occurrence frequency, threating the cascade reservoir systems in the river basin. However, few studies have quantified the breach risk and casualties in short-distance cascade floods formed by upstream large dams and downstream small dams. This study developed the MDMHP system, integrating Monte Carlo simulation, the DB-IWHR model, the Muskingum model, the HEC-RAS model, and the PG-TCANet-LTD surrogate model, preserving the physical causal chain from inflow to downstream response. Parameter quantification was performed through Sobol analysis, with Monte Carlo sampling applied to highly sensitive parameters; the DB-IWHR, Muskingum, and HEC-RAS models executed full-process physical simulations; the PG-TCANet-LTD surrogate model was trained to enable batch two-dimensional inundation simulations, achieving 93.6% consistency in flood prediction and increasing single two-dimensional simulation speed by 3000 times; spatial grid-based statistics were used to assess death risk probabilities under three warning scenarios. Applied to the Changhe Dam-Huangjinping cascade flood, the downstream breach probability after upstream overtopping reached 99.7%; expected casualties decreased from 5,816 without warning to 2,500 with a 0.5-hour warning, and further dropped to 541 with a 1-hour warning, representing a reduction of over 90% compared to no warning. This framework addresses quantification challenges and provides support for for the risk management of cascading dam-break floods in cascade reservoir systems.

While various methods exist for assessing urban rainwater and flood resources, there is a lack of targeted evaluation for the rainwater harvesting potential of areas equipped with sponge city facilities. This study employs the Yield Before Spillage (YBS) principle to design rainwater collection tanks for sponge facilities under different design return periods, conducting a specialized assessment of the rainwater resource potential in built-up sponge facility areas within the “Dongsheng–Kangbashi–Ejin Horo Banner” urban cluster. The results indicate that the collection potential follows the patterns of “wet year > normal year > dry year” and “Ejin Horo Banner > Kangbashi District > Dongsheng District.” A rainwater collection tank designed for a 5-year return period (p = 5a) is more applicable to the study area. The sponge facilities in the study area achieve an annual runoff volume control rate exceeding 85%, effectively alleviating drainage pressure. The conclusions demonstrate that the YBS method can effectively assess the rainwater and flood resources of sponge facilities in arid regions. Tanks designed for the three different return periods all meet the rainwater retention requirements of sponge cities across various hydrological years. In arid areas, tanks designed for lower return periods are sufficient for harnessing rainwater collection potential, offering lower costs.