Rainfall in mountainous watersheds is characterized by strong spatial and temporal heterogeneity, which poses a great challenge to flood forecasting. Time series of rainfall and water level, as reflections of natural processes, follow fractal theory with self-similarity in their fluctuations and variations. However, the impact of fractal properties in rainfall time series on minute-scale flood forecasting has been overlooked. In this study, the rainfall series from 17 rainfall gauges and one water level station within a 339-km2 watershed were used to reveal the fractal dimensions (D) of rainfall and water level series and the impact on flood forecasting. The results show that a fractal relationship existed between rainfall and water level across time scales, with 5.33 h as the division point between short-term and medium-term time scales. Within this scale, the model showed excellent forecasting performance. The fractal dimension of rainfall time series was significantly correlated with model forecasting accuracy, for both water level and peak time. When the spatial uniformity of rainfall gauges exceeded a threshold, further improving uniformity could not enhance the fractal dimension of the rainfall time series. This study confirmed that combining rainfall fractal dimension with rainfall gauge evaluation indicators could support mountainous rainfall gauge network planning and flood prediction.

Geohazards such as collapses, landslides, and debris flows result from complex interactions between human activities and environmental conditions. However, a quantitative understanding of their coupling mechanisms remains challenging. This study developed a machine learning-based classification framework for multi-geohazard susceptibility mapping (GSM) in Zhejiang Province, China, to address this gap. The study employed XGBoost, AdaBoost, and Random Forest to construct individual models for each geohazard, using a comprehensive set of geomorphologic, geological, environmental, hydrological, and anthropogenic factors. The XGBoost model achieved Area Under the Curve (AUC) values greater than 0.9 for all geohazards, and was selected as the optimal model for GSM. Results show: (1) Topographic position index (TPI) and distance to roads are the most influential factors, with dominant roles varying by geohazard-TPI primarily controls debris flows, while collapses are more driven by road proximity. (2) Anthropogenic factors account for 15.9%-33.8% of importance across geohazards. (3) The dependence plots and heatmap of interaction values reveal the impact of human-natural factor coupling mechanisms on geohazards. The study provides a quantitative and interpretable analysis of human-natural environment coupling, offering insights for risk management and spatial planning in densely populated coastal regions under climate change.

Coastal flood exposure indicators are increasingly needed to support climate adaptation and sustainable coastal planning in small island regions, yet regional-scale assessments that translate hydrodynamic simulations into spatially explicit exposure indicators remain limited across the Caribbean. This study develops a spatially explicit coastal flood exposure framework by integrating LISFLOOD-FP simulations with settlement dynamics, land-cover change, and population distribution data. Coastal inundation was simulated under baseline (1985–2014) and SSP5-8.5 (2015–2050) 100-year extreme sea-level conditions and overlaid with land-cover, settlement, and WorldPop data to assess land-cover, settlement, and population exposure across Caribbean island systems. The results show that SSP5-8.5 generally increases coastal inundation extent relative to the baseline, although the broad regional pattern remains similar. Large islands with extensive low-lying coastal plains show greater absolute inundation, whereas several smaller islands exhibit higher proportional exposure because of limited land area and concentrated coastal settlements. Wetlands constitute the largest share of exposed terrestrial land cover, while built-up land shows increasing exposure. These findings suggest that future coastal flood exposure in the Caribbean is shaped by both increasing hazard intensity and persistent human concentration in flood-prone coastal zones. Because the modelling framework does not explicitly represent wave setup/runup, compound flooding, or coastal defences, the results are interpreted as regional-scale screening indicators rather than locally calibrated flood-risk estimates.

Delta regions worldwide face escalating coastal flood risks driven by the compound effects of sealevel rise (SLR) and vertical land motion (VLM). Existing studies often analyze these hazards separately and rely heavily on simplified static inundation models, limiting the accuracy of flood impact assessments and neglecting dynamic socioeconomic factors. This study develops an integrated framework combining high-resolution VLM monitoring (SBAS-InSAR), dynamic hydrodynamic modeling (LISFLOOD-FP), and socioeconomic projections (Shared Socioeconomic Pathways: SSP1-2.6, SSP2-4.5, SSP5-8.5) for comprehensive flood impact evaluation in three globally significant deltas: the Ganges-Brahmaputra-Meghna (GBM), Mississippi, and Yangtze. Results highlight severe and spatially variable subsidence rates-most notably in the GBM Delta (-8.98 mm/year), followed by the Mississippi (-2.93 mm/year) and Yangtze (-1.60 mm/year)- with human activities likely playing an important role in driving surface deformation. Projected flood scenarios (2050 and 2080) indicate significant increases in inundation extents and exposed populations and economic assets, particularly under combined SLR + VLM scenarios. The Yangtze Delta shows the highest economic exposure (up to approximately 1 trillion USD), whereas the GBM Delta exhibits the greatest demographic vulnerability, potentially affecting approximately 20 million individuals. The relative contributions analysis emphasizes an increasing dominance of SLR over time, especially under high-emission scenarios. These findings underscore the critical importance of tailored, region-specific adaptation strategies including resilient infrastructure, nature-based solutions, and adaptive spatial planning.

We present a 1970–2020 dataset of daily maximum coastal water levels reconstructed for 23 tide gauges along China’s coast. The product combines storm-surge residuals predicted with an Informer-based deep learning workflow (benchmarked against LSTM, CNN-LSTM, and ConvLSTM) with astronomical tides estimated by UTide from historical observations. Predictors are drawn from ERA5 reanalysis and multi-source tide-gauge records are used for training and validation. For each station, the model with best validation skill generates residuals combined with tidal harmonics to form daily maxima. Across stations, the reconstruction attains a mean correlation coefficient of 0.81 and RMSE of 11.7 cm for daily maxima; for events above the 95th percentile, the mean correlation is 0.68 and RMSE is typically below 20 cm. The release includes metadata, data splits, and skill metrics for transparency and reuse. This dataset enables spatiotemporal analyses of extreme coastal water levels and coastal hazard mitigation in regions with sparse observations. Daily maxima are computed as the sum of the maximum tide and maximum surge. This serves as an upper bound, as the peaks of tide and surge rarely coincide. Using hourly data, we estimate a mean non-coincidence bias of 14.9 cm (14.8

—Rapid identification of building damage distribution during the critical rescue phase is essential for efficient emergency response. However, most existing studies focus on earthquake-like structural hazards, limiting their generalization across different disaster scenarios. In addition, current methods often struggle to effectively integrate cross-temporal features from pre- and post-disaster imagery, while effective links between pixel-level classification and regional decision-making information remain limited. To address these issues, this study proposes an improved deep learning framework, NBDANet, developed based on BDANet for building damage assessment. NBDANet adopts a two-stage architecture with Spatial Pyramid Pooling (SPP) modules embedded in both stages to enhance multi-scale contextual representation. The first stage performs building localization, while the second stage introduces a Dual-input Triplet Attention (DTA) module to improve feature interaction between pre- and post-disaster images and enhance damage classification accuracy. The model is trained and evaluated on the xBD dataset and further applied to the Derna case using high-resolution GeoEye-1 imagery. Validation is supplemented by Sentinel-derived Normalized Difference Built-up Index (NDBI) and Damage Proxy (DP) maps, confirming spatial consistency. In addition, Global Human Settlement Layer (GHSL) data are integrated to estimate affected populations and delineate rescue-priority zones. Results show that 22% of buildings were damaged, including approximately 20% of all buildings classified as Major damage or Destroyed. Estimated affected populations in Major damage and Destroyed zones were approximately 5,200 and 2,650, respectively. The proposed approach demonstrates strong performance in building localization and damage classification, providing a scalable solution for multi-source disaster assessment and recovery planning.

On 15 November 2024, a large-scale coastal flooding event hit China coastal areas, highlighting the urgency of understanding tide-driven flood risks. Yet this type of tide-driven flooding has received little research attention across the China coastline. Here we aim to raise awareness of high-tide flooding (HTF) along the China coastline by quantifying the future potential of frequent HTF occurrence. Using a likelihood-based proxy for HTF flooding threshold and modeled water levels, we quantify how the combined effects of sea level rise (SLR) and amplified astronomical tides drive increases in future HTF exceedance frequency. Our results indicate a marked increase in HTF exceedance frequency in the coming decades, driven by SLR and the ascending phase of the nodal tidal cycle. Specifically, under the Shared Socioeconomic Pathway 2-4.5 (SSP2-4.5) scenario, HTF exceedance events are projected to occur twice weekly on average along China northern coast by the 2050s, significantly more frequently than along the southern coast. Chronic HTF exceedance will first emerge in Bohai Bay under the SSP2-4.5 scenario in the mid-2040s. We also demonstrate that future tidal amplification will not only increase the number of HTF exceedance days but also advance the onset of chronic HTF. Our results highlight the substantial projected rise in HTF occurrence potential along China coastline, emphasizing the need for greater attention to HTF in this region.

Compound flooding arising from concurrent extreme precipitation and storm surge poses severe risks to coastal regions. Using a 40-year dataset from 28 estuaries along China's coast, we identified 337, 208, and 193 compound events in the northern, eastern, and southern coastal China, respectively, revealing strong spatial contrasts in event frequency. Tail dependence between precipitation and storm surge is strongest in the eastern and southern regions with Kendall's τ roughly 0.15–0.25 and weakest in the north with τ between −0.10 and 0.15. These contrasts substantially modify compound flood likelihood. Joint return periods (JRPs) are much shorter when dependence is accounted for than when tail independence is assumed, particularly in regions with stronger dependence. For example, at a representative eastern coastal site, a concurrent 1-in-5-year precipitation and surge event yields a JRP of about 13 years when dependence is considered, compared with 25 years under independence, indicating that neglecting positive dependence underestimates compound flood risk. However, this reduction is far less pronounced for the same combination in the northern region where dependence is weak. Cyclone-type contributions help explain these regional contrasts. Tropical cyclones (TCs) generate the most extreme compound events in the east and south and strongly enhance tail dependence, such that removing TC events markedly lengthens JRPs. In the north, extratropical cyclones (ETCs) dominate event frequency but exhibit weaker synchrony between extremes. Removing these frequent but weakly dependent events therefore shortens JRPs. These findings underscore the need to incorporate regional dependence structures and cyclone-type influences into coastal flood risk assessments.

Rising future extreme sea levels (ESL) increase coastal flooding, putting people and assets at risk in low-elevation coastal zones. Coastal ecosystems such as mangroves and coral reefs can attenuate waves and storm surge, thereby reducing flood impacts along tropical and subtropical coastlines. Despite earlier studies of global flood risk, the benefits provided by the simultaneous presence of mangroves and coral reefs, under climate change, have not been quantified. By using process-based inundation modelling with and without these coastal ecosystems, and accounting for population and asset changes under future socio-economic conditions, this study estimates how the spatial distribution of these ecosystems reduces coastal flooding for a baseline period (1980–2014), for 2030 and for 2050 under the SSP5-8.5 scenario. When mangroves and reefs are included, our results show a reduction in global expected annual flood damage by USD 1.7 billion (12.9\%) in the baseline period and by USD 2.7 billion (13.9\%) by 2050. The presence of ecosystems reduces the mid-century annual expected flooded area with 7811 \unit{\square\kilo\meter} and reduces the mid-century annual affected assets and population by USD 45.8 billion and 1.6 million people per year, respectively. In relative terms, we find that benefits resulting from the presence of ecosystems are concentrated in countries with high climate vulnerability and low adaptation readiness, implying that ecosystem degradation would raise future flood risk most where adaptive capacity is lowest. Therefore, our results suggest that mangrove preservation/rehabilitation and reef conservation directly support adaptation to climate change.

In coastal areas, projected increases in the amplitude of the sea level seasonal cycle will lead to the intensification of coastal flood hazards. Conversely, projected decreases in the amplitude may act to suppress flood hazards in some regions, partially offsetting the impact of sea level rise. Simulations of the sea level seasonal cycle by the Coupled Model Intercomparison Project Phase 6 (CMIP6) models suggest changes up to ±30% are projected under a high emission scenario. We demonstrate that a currently unaccounted increase of 5–10 cm in the sea level seasonal cycle will increase the frequency of extremes from once in 20 years to once in 10 years. However, the sea level seasonal cycle amplitudes in CMIP6 models are underestimated compared to seasonal variability observed by satellite altimetry for almost all coastlines, suggesting that future changes might be underestimated, leading to a large impact on extreme sea levels in coastal areas. Including the changes in the future sea level seasonal cycle will advance future projections of sea level extremes and contribute to the crucial decisions about adaptation in coastal areas.

Study region: The coastal region of Tianjin, China. Study focus: Due to the sea-level rise under climate change and the increase in extreme weather events, this study aims to simulate the spatiotemporal variation process of coastal flooding using deep learning surrogate models. Four models-U-Net, CNN-LSTM, ConvLSTM, and CNN-Transformer-were developed and evaluated. The training dataset was generated using the LISFLOOD-FP hydrodynamic model under extreme sea-level scenarios. These models were applied to predict the spatiotemporal dynamics of inundation extent and water depth in Tianjin. Model performances were compared based on prediction accuracy and efficiency. New hydrological insights for the region: All models achieved high accuracy using only DEM, land cover, coastline, and sea level time series as inputs. The U-Net model showed the best performance (MAE = 0.0125 m, RMSE = 0.0486 m, R2 = 0.9935), with 98.52 % classification accuracy for flood severity. This study highlights that deep learning models offer high computational efficiency, strong predictive accuracy, and a streamlined modeling process, enabling rapid largescale spatiotemporal simulations once trained. Furthermore, the model exhibits strong generalization capability across different extreme sea-level scenarios. This study not only provides a novel data-driven approach for simulating coastal flooding dynamics but also offers valuable insights into the flood risk response in low-lying coastal cities under climate-induced sea-level rise, supporting early warning systems and adaptive flood management strategies.

River deltas are critical socio-economic and ecological regions but face heightened flood risks due to climate change and urbanization. Taking the Ganges-Brahmaputra-Meghna (GBM) River Delta, the Mississippi River Delta, and the Yangtze River Delta as case studies, this research aims to reveal the characteristics and formation mechanisms of human adaptation to flood risks across different deltaic regions. Through integrating hydrodynamic modeling, spatiotemporal analysis, and multi-source datasets, this study systematically investigates flood adaptation characteristics across three major deltas based on a newly developed comprehensive framework of Human-Flood Distance (HFD) and resilience. The results show that: spatially, while these three deltas exhibit varying degrees of inundation extent, each faces unique flood vulnerability challenges; temporally, the GBM River Delta exhibits stabilized population growth and HFD recovery after initial contraction, the Mississippi River Delta shows significant fluctuations in both population and HFD, while the Yangtze River Delta demonstrates continuous population growth with steady HFD increase; in terms of adaptation mechanisms, resilience assessment indicates that the Mississippi River Delta demonstrates the highest resilience, primarily driven by recovery capacity, the Yangtze River Delta shows limited but structurally supported resilience, while the GBM River Delta exhibits negative indices due to multiple constraints. These findings emphasize the importance of developing context-specific flood risk management strategies and provide feasible flood prevention solutions for policymakers, particularly in formulating adaptation strategies that comprehensively consider both flood safety and sustainable development.

Study region: This study targets 76 counties in Zhejiang Province, China, frequently affected by rainstorm disasters. Study focus: Rainstorm-induced losses arise from complex interactions between extreme weather and human-environment systems. To assess these losses, we propose a framework with five machine learning models—MLP, Random Forest, CatBoost, LightGBM, and XGBoost. Using 461 disaster records from 2001 to 2019, a multi-dimensional indicator system was developed, covering hazard, exposure, vulnerability, and environmental factors. Model performance was evaluated using accuracy, F1 score, ROC-AUC, and Cohen’s Kappa. Feature importance and SHAP analysis identified key drivers and explained behavior. New hydrological insights for the region: This study advances the understanding of compound mechanisms underlying rainstorm-related economic losses in Zhejiang Province. Among the five models, XGBoost exhibited relatively better performance than the other models in assessing rainstorm disaster losses. Short-duration extreme rainfall emerged as the dominant hazard driver, while socioeconomic indicators, particularly population density and per capita GDP—amplified disaster severity under high-exposure conditions. SHAP analysis revealed nonlinear threshold effects: losses escalate rapidly when extreme rainfall coincides with densely built environments. Meanwhile, environmental factors such as terrain slope and vegetation cover played a buffering role in mitigating low to moderate losses. These findings highlight the value of interpretable machine learning models in capturing the spatial heterogeneity and compound nature of rainstorm disaster impacts, providing scientific support for targeted risk zoning and mitigation planning.

Abstract. Coastal flooding and damage can result from compound extremes of wind and precipitation that elevate sea level anomalies. However, the global patterns and impacts of such conditions are poorly understood. Here we analyze observational and model data to reveal a positive correlation between wind and precipitation extremes across most of the global coastline, especially at higher latitudes. We also show that these variables exhibit stronger dependence on higher quantiles, indicating more frequent and severe compound conditions. Moreover, we demonstrate that sea level anomalies are enhanced during compound conditions compared to normal conditions, implying increased coastal flooding risk. We project that both the intensity and frequency of compound conditions will rise in 2020–2100 compared to 1940–2014 under two emission scenarios, with larger changes at high latitudes. Our findings highlight the need for assessing and managing the risks and impacts of compound extremes on coastal communities and infrastructure.

Marine aquaculture plays a significant role in China’s economic development, accounting for nearly one-third of the aquaculture industry. Tropical Cyclone (TC)-induced extreme waves are one of the primary factors that destabilize the structures of aquaculture net cages, resulting in substantial economic losses. However, current research on quantitative risk assessment in marine aquaculture is limited. To fill this gap, we took Northern East China Sea (NECS) as the study area to examine the potential impact of tropical cyclone-induced extreme waves on marine aquaculture. First, we simulated TC-induced extreme waves between 1979 and 2018 by a tightly coupled ADvanced CIRCulation (ADCIRC) model and Simulating Waves Nearshore numerical model, and calculated the probability of occurrence and return period of the hazard. Subsequently, by constructing the failure probability of net cage structures under different significant wave heights, we established a physical vulnerability function relating wave height to failure probability. Using the developed physical vulnerability curves, we assessed the risk faced by offshore marine surface aquaculture under extreme typhoon waves with different return periods, and calculated the expected loss for marine aquaculture. The research results reveal that the hazard of extreme typhoon waves exhibited a spatial pattern of higher occurrences in the vicinity of Qinhuangdao, the Shandong Peninsula, and the northern Jiangsu region compared to other coastal regions, and the risk of marine aquaculture is high in the southern part of Liaoning Province, the eastern part of Shandong Province, and the northeastern part of Jiangsu Province. It is crucial to enhance the capacity for disaster response, reduce potential losses, and improve the ability of marine aquaculture to withstand TC-induced extreme waves in these areas.

Study region: Shanghai, China Study focus: This paper proposes a comprehensive framework for quantifying storm surge floods in coastal cities by incorporating the influences of both climate change and urbanization. The framework achieves a physically process-based numerical simulation of storm surge-induced flood hazards due to tropical cyclones in coastal cities by coupling the fast flood inundation model (SFINCS) and the land use change model (GeoSOS-FLUS), along with the numerical nested model for storm surges (Delft 3D Flow & Wave). Using a 1000-year tropical cyclone simulated by the STORM model as an example, this study analyzes and maps coastal flood impacts under the moderate climate scenario (SSPs245) and high emission scenario (SSPs585), and also evaluates the impact of land use changes on these scenarios. New hydrological insights for the region: Taking Shanghai, China as an example, the results show that by 2100, urban land use changes will lead to an increase in the extent of 1000-year TC flooding areas by 4.91-34.00 %, underestimating the inundation area of storm surges if future urban land use changes are not considered. Additionally, our predictions indicate the vulnerability of Chongming island and Changxing island to the impacts of climate change, despite the protective role of coastal embankments considered in the tropical cyclone storm surge simulation. The results of this study represent an important contribution to a better understanding of how future urban land use changes will affect storm surge flooding risks in and around Shanghai. The proposed methodology can be applied to coastal areas worldwide that are vulnerable to tropical cyclones, aiding in the formulation of hazard mitigation policies to alleviate flood impacts in these regions.

The rising sea levels and the increasing frequency of extreme coastal disasters, exacerbated by rapid urbanization in coastal areas, have led to more frequent and severe coastal flooding events. Yet, large-scale coastal flooding hydrodynamic modeling in mainland China remains limited, particularly due to inadequate consideration of driving factors, with most relying on the bathtub method. This study, considering the existing coastal protection standards, utilized the LISFLOOD-FP two-dimensional hydrodynamic model to simulate the spatial distribution of 100-year return period coastal flooding inundation under both the baseline and SSP5-8.5 2050 scenarios with higher accuracy and a spatial resolution of 90 m. In addition to climate change, we also conducted an in-depth analysis from the perspectives of urban expansion and coastal reclamation. The study finds that under the backdrop of rapid urbanization, settlement areas (1985-2015) and coastal reclamation zones (1990-2019) have expanded rapidly into flood-prone areas, especially in the 21st century, where the growth rate has greatly outpaced that of non-exposed areas. The area of settlement located in coastal flood hazard zones has increased 6.5 times, and the area of reclamation zones in coastal flood hazard zones has increased 26.3 times. Additionally, this study highlights the relative contributions of climate change and urban expansion to the future (SSP5-8.5 2050) coastal inundation risks in cities, providing valuable insights for the sustainable management of coastal flood hazards in China.

In the context of climate change, the combined effects of coastal land subsidence and sea level rise exacerbate coastal flood risks by altering relative sea levels. This study leverages high-resolution land subsidence rate data obtained from Interferometric Synthetic Aperture Radar (InSAR) and employs the LISFLOOD-FP two-dimensional hydrodynamic model to simulate coastal flooding for 43 coastal mega-cities globally. Our findings indicate that, when considering subsidence, over 76% of these cities experience an expansion in inundation areas under both Baseline and SSP5-8.5 scenarios. Furthermore, we conduct a quantitative assessment of the relative contributions of land subsidence and climate change to coastal flood inundation, identifying 19 cities where land subsidence plays a dominant role. Moreover, the impact of urban expansion on coastal flood risk cannot be underestimated, particularly in coastal cities that experience rapid urbanization and extensive coastal reclamation activities. By incorporating annual data on the expansion of settlements, reclaimed coastal areas, and urban built-up areas, we evaluate the dynamic changes in coastal flood exposure and uncover a long-term trend of increasing potential impacts of coastal flooding in mainland China's coastal regions, which is at a continental scale. Specifically, the area of settlements located in coastal flood hazard zones has grown to 6.5 times its original size, while the area of reclaimed land within these zones has expanded to 26.3 times its original extent. The insights from this study provide a valuable reference for sustainable development strategies and measures to address the escalating coastal flood hazards in coastal cities worldwide.

Timely detection of the spatial distribution of building damage in the immediate aftermath of a disaster is essential for guiding emergency response and recovery strategies. In January 2025, a large-scale wildfire struck the Los Angeles metropolitan area, with Altadena as one of the most severely affected regions. A timely and comprehensive understanding of the spatial distribution of building damage is essential for guiding rescue and resource allocation. In this study, we adopted the MambaBDA framework, which is built upon Mamba, a recently proposed state space architecture in the computer vision domain, and tailored it for spatio-temporal modeling of disaster impacts. The model was trained on the publicly available xBD dataset and subsequently applied to evaluate wildfire-induced building damage in Altadena, with pre- and post-disaster data acquired from WorldView-3 imagery captured during the 2025 Los Angeles wildfire. The workflow consisted of building localization and damage grading, followed by optimization to improve boundary accuracy and conversion to individual building-level assessments. Results show that about 28% of the buildings in Altadena suffered Major or Destroyed levels of damage. Population impact analysis, based on GHSL data, estimated approximately 3241 residents living in Major damage zones and 31,975 in Destroyed zones. These findings highlight the applicability of MambaBDA to wildfire scenarios, demonstrating its capability for efficient and transferable building damage assessment. The proposed approach provides timely information to support post-disaster rescue and recovery decision-making.