In recent decades, intensifying extreme rainfall and flood disasters underscore the importance of intelligent flood risk mitigation informed by urban drainage models (UDMs). However, high computational costs of UDMs limit their applicability to implement advanced stormwater management strategies, leading to growing interests in building efficient surrogates for existing UDMs. However, existing approaches face several key limitations including incomplete representation of rainfall-runoff-routing processes, insufficient encoding of drainage network regulation, and a lack of physical explainability. To address these gaps, this study proposes a Process-Aware Spatio-Temporal graph-neural-networks-based surrogate model, PAST, holistically simulating hydrological and hydrodynamic processes while incorporating regulation effects. PAST leverages Long Short-Term Memory (LSTM) networks for rainfall-runoff simulation and integrates Graph Attention Networks with LSTM (GAT-LSTM) for hydrodynamic routing, further introducing the adaptive connectivity layer for plausible embedding of element-level control rules. Applied to a residential area with moderate imperviousness, PAST achieves high performance with Nash-Sutcliffe Efficiency (NSE) values of 0.96 for water depth and 0.88 for total inflow simulations, substantially outperforming baseline models—particularly under regulated conditions and during extreme rainfall events. Additionally, attention-based explainability analysis reveals that PAST learns physically plausible dynamics by allocating more attention weights to nodes contributing higher inflows, thereby enhancing the transparency and diagnosability of the developed surrogate model. Compared to prior research, this work advances integrated, interpretable and efficient surrogate modeling for urban drainage systems with reliable physical mechanisms, offering a robust foundation for intelligent stormwater management.

Flooding in underground spaces has become more prevalent in recent decades due to increasing urbanization and the intensification of rainfall caused by climate change. While previous research on flood simulations has primarily focused on 1D and 2D models, limited studies have utilized 3D simulations. This study presents a comparison between 1D simulations using the SWMM model and 3D simulations using ANSYS Fluent to model flooding in an underground basement. The comparison was conducted under various flood scenarios. The results revealed a significant discrepancy in the predicted water volumes between the two models. Specifically, the 1D model overestimated the total water volume by approximately 50

Urban areas can substantially modify local hydroclimate, enhancing precipitation over and downwind of cities. Yet, the urban expansion effects on rainfall remain insufficiently understood, and a quantitative relationship between urban growth and rainfall intensification remains to be established. Using the WRF model with eight urban size scenarios for Beijing, a numerical modeling framework was implemented to investigate how changes in urban extent influence rainfall during two representative summers, one relatively wet and one relatively dry. Results show that the rainfall response exhibits an approximately linear dependence on the degree of urban expansion, with the largest impacts occurring over the city center and downwind regions. In general, rainfall increases with urban area enlargement, particularly during nighttime in relatively wet summers due to higher humidity and a more pronounced urban heat island effect. In relatively dry summers, limited moisture supply leads to smaller changes in total rainfall. Changes in hourly rainfall intensity demonstrate a contrasting pattern. Heavy rainfall intensities further intensify in response to urban expansion, while light rainfall is suppressed or remains largely unchanged. Daytime and nighttime rainfall intensity respond to urban expansion in opposite ways, with daytime intensity generally weakening and nighttime intensity strengthening as Beijing expands. These contrasting diurnal behaviors ultimately lead to a reduction in rainfall intensity during relatively dry summers and a slight increase during relatively wet summers. Overall, the results highlight the dependence of urban rainfall modification on city size and background climatic conditions.

Predicting changes in urban pluvial flood hazards under climate warming is crucial for risk mitigation and disaster management. A key challenge in simulating future urban flood hazards is the scarcity of high-resolution rainfall projections, particularly at the sub-daily and kilometer scales required for hydrodynamic modeling. We present a cascading process-informed framework that requires minimal observed climatic data, enabling scenario analysis even in data-scarce cities. This framework consists of a distribution‐based spatial quantile mapping (DSQM) method to morph observed rainfall fields conditioned on temperature changes, a stochastic storm transposition (SST) method to account for the spatial variability of urban rainfall, and a rain‐on‐grid hydrodynamic model (AUTOSHED) for efficient simulation of urban pluvial floods at high spatio-temporal resolution. The framework allows the generation of stochastic rainfall fields under different rainfall return levels and regional warming levels. It supports the quantification of changes in future urban flood statistics with detailed hazard maps of inundation depth, duration, and flow velocity. We select the metropolitan area of Beijing (300 km2) as a case study area and utilize gridded hourly and 1 km rainfall data to simulate flood evolution at 5 min and 5 m resolution under regional warming levels of 1, 3, and 5 °C relative to the period 1998–2019. Our results show that with rising temperatures, regional storms tend to become more intense but smaller in spatial extent, which may in turn drive increased local flood depth, accelerated flow velocity, and deeper inundation, collectively elevating pluvial flood risk. Specifically, mean rainfall intensity increases by 6 %, 11 %, and 20 % (respectively with the warming levels), peak flood depth exhibits a nonlinear increase of 4 %, 7 %, and 8 %, due to the complex interactions of reduced storm area, increased storm intensities, and rainfall spatial variability. The proposed DSQM-SST-AUTOSHED framework offers a data-driven, physically grounded, and efficient approach to assess urban flood risk under regional warming. It only requires observed rainfall fields and temperature datasets, which are readily accessible from public sources, making the approach easily extendable to other cities.

Abstract. Hybrid hydrological models integrating embedded neural networks (ENNs) have demonstrated strong potential for improving streamflow prediction. However, how uncertainties induced by ENNs propagate through hydrological processes and internal system states remains poorly understood. This limits a mechanistic understanding of the stability and reliability of hybrid hydrological systems. This study therefore moves beyond performance evaluation and focuses on the internal uncertainty propagation structure of fully coupled hybrid hydrological systems. We investigate the internal uncertainty propagation mechanisms of hybrid hydrological models with different levels of ENN embedding across 31 cold-region basins. An ensemble-based framework driven by stochastic optimization was used to quantify uncertainty across hydrological fluxes and internal state variables. Multiple complementary metrics were employed to characterize uncertainty magnitude, ensemble consistency, coverage and spread. Results show that ENNs improve streamflow simulation performance but alter the internal uncertainty structure of hydrological systems. Uncertainty amplification is primarily localized within ENN-replaced hydrological processes, while propagation to physically based unreplaced processes remains limited. Instead, internal state variables with memory effects act as key “uncertainty reservoirs”, where uncertainty accumulates over time. This reveals a structured rather than uniform pattern of uncertainty propagation in hybrid hydrological systems. Furthermore, increased ensemble coverage is accompanied by wider uncertainty bands, indicating a trade-off between predictive reliability and sharpness induced by stochastic optimization. Overall, this study provides new insights into how ENNs reshape uncertainty propagation pathways in hydrological models. The results highlight the importance of jointly considering predictive performance, uncertainty structure and internal system stability in hydrological modeling.

Study region Beijing-Tianjin-Hebei (BTH) region Study focus Understanding urban flood risk-resilience relationships facilitates the shift from defense-oriented to resilience-informed risk management, which is essential for building resilient cities. However, existing studies primarily treat risk and resilience as independent dimensions and overlook their coupling effects, limiting identifications of areas with urgent needs and greater potential for coordinated enhancement of resilience and risk mitigation under differentiated management. This often leads to an overemphasis on high-risk-low-resilience areas while neglecting other priority areas. Building on conventional analytical frameworks that examine spatial patterns, numerical values, and spatial correlations, this study advances resilience-informed risk management by explicitly incorporating risk-resilience coupling effects. The improved approach is applied to the BTH region, a flood-prone urban system facing increasing threats from extreme events, over the period 2010–2020. Priority areas are identified and implications for resilience-informed risk management are discussed. New hydrological insights for the region Results indicate that the BTH region has been characterized by low risk but low resilience. Targeted measures should be prioritized in areas with low coupling coordination but a relatively high degree of coupling. These areas are extended from being limited to high-risk-low-resilience areas to encompass those within Medium-Low, High-Very Low, High-Low, High-Medium, High-High, Very High-Very Low, Very High-Low, Very High-Medium, and Very High-High risk-resilience classes. Routine risk management considering coupling effects can reveal potential vulnerability under extreme events.

Peri-urban vegetation influences urban hydroclimates, yet its role in shaping urban precipitation remains understudied due to binary urban-non-urban framings, the limited representation of peri-urban landscapes in models and datasets, and a predominant focus on intra-urban areas. Here we integrate satellite-derived vegetation trends with an evapotranspiration model and an atmospheric moisture-tracking model to quantify how peri-urban vegetation change affects urban precipitation within 1,029 cities worldwide. We identify a spatially coupled hydroclimatic mechanism in which vegetation-driven shifts in peri-urban evapotranspiration modulate urban precipitation via atmospheric moisture transfer. Although these changes contribute only 1.9% of annual urban precipitation, they account for 18.3% of its long-term increase, indicating a disproportionate and systematic influence on urban hydroclimate trajectories. We further find that this coupling strengthens in cities with more abundant surrounding vegetation, wind-aligned greening and lower background humidity. Our findings clarify how peri-urban land-atmosphere interactions regulate urban climates and highlight the need to integrate peri-urban ecosystems into climate-resilience planning.

Underground space floods are a complex type of calamity mainly caused by increasing rainfall and temperature, along with urbanization and population growth necessitating the creation of underground spaces such as subway stations, underground parking lots, underground garages, etc. The causes above have contributed to the frequent and severe appearance of floods in underground spaces in recent decades. The situation is complicated, so further study and research are still needed. This work applied numerical simulation using Fluent software to investigate, analyze, and compare the impact of altering the magnitude and type of inlet velocities on the underground space's flooding characteristics. Two types of velocities, fixed and transient velocities, were involved in the investigation.

This work presents an efficient graph-reconstruction-based approach for generating physical sewer models from incomplete information, addressing the challenge of representing the sewer drainage effect in urban pluvial flood simulations. The approach utilizes graph-based topological analysis and hydraulic design constraints to derive gravitational flow directions and nodal invert elevations in decentralized sewer networks with multiple outfalls. By incorporating linearized programming formulation to solve reconstruction problems, this approach can achieve high computational efficiency, making it suitable for application to city-scale sewer networks with thousands of nodes and links. Tested in Yinchuan, China, the approach integrates with a 1D/2D coupled hydrologic-hydrodynamic model and accurately reproduces maximum inundation depths (R2=0.95) when the complete network layout and regulated facilities are available. Simplifications, such as the adoption of road-based layouts and the omission of regulated facilities, can degrade simulation performance for extreme rainfall events compared to calibrated equifinal methods. However, design rainfall analysis demonstrates that the physical reconstruction approach can reliably outperform equifinal methods, achieving reduced variation and higher accuracy in simulating inundation areas. However, proper configuration of regulated facilities and network connectivity remains crucial, particularly for simulating local inundation during extreme rainfall. Thus, it is recommended that the proposed algorithm be integrated with targeted field investigations to further improve urban pluvial flood simulation performance in data-scarce regions.

Accurate estimation of surface precipitation with high spatial and temporal resolution is crucial for disaster weather detection and decision-making regarding water resources management. Polarimetric weather radar is an important instrument for quantitative precipitation estimation (QPE). Conventional parametric approaches, such as the radar reflectivity (Z) and rain rate (R) relations, cannot fully represent the spatial and temporal variability of clouds and precipitation due to parameterization errors and dependence on raindrop size distribution (DSD). Furthermore, these relations estimate rainfall on a grid-by-grid basis, preventing the incorporation of spatial information into precipitation estimation. In recent years, machine learning has made rapid advancements in non-linear fitting and feature extracting. Since 2020, multiple studies constructed MLP or CNN-based QPE models that used polarimetric radar observations to retrieve precipitation. These researches have consistently demonstrated that machine learning algorithms perform better than traditional parametric methods in different regions and climatic conditions(Chen & Chandrasekar, 2021; Li et al., 2023; Osborne et al., 2023; Tian et al., 2020; Zhang et al., 2021; Zhou et al., 2023). The aforementioned studies have highlighted the immense potential of deep learning for radar QPE, but they are based on S-band radar data. Because X-band radar has a shorter wavelength, the electromagnetic scattering characteristics of hydrometeors differ from those of S-band radar, especially for specific differential phase (kdp), which is closely related to rainfall. Furthermore, X-band radars have different spatial resolutions from S-band radars, which indicates that directly applying a model trained with S-band radar data to X-band radar data may introduce biases. Therefore, we develop a CNN-based QPE model using polarimetric measurements from X-band radars and compare its performance against traditional parametric methods. The input data for the CNN model is a matrix with dimensions (6, 9, 9). The matrix is composed of two matrices of size (3, 9, 9), which is the polarimetric measurements from the two lowest scan elevation angles and 9*9 surrounding range gates. This allows the input data to capture the spatial and physical characteristics of the precipitation field. The results reveal that the CNN-based model not only enhances the accuracy of radar QPE with a diminished bias but also provides a more precise depiction of the spatial distribution of precipitation in comparison to conventional methods.

The Tibetan Plateau is the headwaters of several major river basins, but uncertainties exist in the estimated contributions of glacial melt and groundwater to runoff. We present a new tracer-aided glacio-hydrological model constrained by multiple datasets for five major river basins of the Tibetan Plateau. We show that the contribution of glacier melt to the annual runoff is less than 5% in all the five basins at the outlets-much less than previous estimates. Our secondary finding is that the partitioning between surface runoff and groundwater flow varied greatly across the watersheds, with groundwater runoff contributing 35-75% of the annual runoff. The contribution of glacier melt has a strong spatial variability and scale dependency, but the population heavily dependent on it is limited, so a potential significant decrease in water resources due to glacier shrinkage is not a problem that should raise public worries in the Tibetan Plateau.

Implementing the 3-Dimensional Variational (3DVar) data assimilation technique using high-density automatic weather station (AWS) observations substantially improves the precipitation simulation and forecast capabilities in the Weather Research and Forecasting (WRF) model. Given the impact of spatial distribution and quantity of observation data on assimilation effectiveness, there is a growing need to assimilate the most efficient amount of observation data to improve the precipitation forecast accuracy, especially in the context of the proliferation of data from diverse sources. This study investigates the impacts of spatial density of assimilated data on enhancing model predictions, focusing on a squall line event in Beijing on 2 August 2017 which has approximately 2400 AWSs in the simulation domain. Seven experiment groups assimilating varying proportions of AWS data (3.125, 6.25, 12.5, 25, 50, 75, and 100 percent of total AWSs) were conducted, comprising 10 experiments per group. The results were then compared with the experiment without data assimilation (CTRL) and the observations. Results show that while the WRF model roughly captured the evolution of this event, it overestimated the precipitation amount with significant deviations in precipitation locations. A general positive correlation was observed between the spatial density of assimilated data and the enhancement in model performance. However, there is a notable threshold beyond which additional data ceases to enhance forecast accuracy. The model performs best when the ratio of the number of assimilated AWSs to the model simulated area reaches 1/40 km-2. Moreover, significant variations in improvement effects across experiments within the same group indicate the substantial impact of spatial distribution of assimilated AWSs on forecast outcomes. This study provides a reference for devising more efficient and cost-effective data assimilation strategies in numerical weather prediction.

Accurate and fine-grained precipitation nowcasting holds paramount importance for weather-dependent decision-making and is facing escalating expectations and challenges. While researchers have notably advanced precipitation nowcasting using deep learning (DL) models with larger sizes and more complicated structures, there is scarce research exploring the potential improvement from employing radar data with higher spatial resolutions-a hundred-meter scale rather than a kilometer scale. To evaluate the improvement of higherresolution data, two U-Net architecture-based models, one larger and another smaller, were designed and trained with radar data at different spatial resolutions-1000 m, 500 m, and 100 m. Their effectiveness was examined by comparison to two baseline models. The models trained with diverse resolutions of data underwent comparative evaluation through two specific precipitation cases. The results unveil a positive correlation between the precipitation nowcasting performance and the spatial resolution of radar data. Models trained with higher-resolution data demonstrate superior forecasting accuracy, reduced bias, and enhanced spatial alignment between predictions and observations. Higher-resolution data empowers DL models to capture boundaries and local-scale patterns of convective systems more accurately, thereby improving the performance in precipitation nowcasting. More importantly, the comparison indicates that to further promote the performance of DL-based precipitation nowcasting, improving data resolution is more efficient than expanding model size. The use of high-resolution data diminishes computational and development costs by concurrently reducing the size of DL models, underscoring pragmatic benefits for related services. Given limited resources, employing higherresolution data is recommended for priority consideration over larger-size models.

The Tibetan Plateau (TP) is widely known as the 'Asian Water Tower', due to its role in providing fresh water to downstream Asian countries. Based on the runoff data of large river basins on the TP, the weighted average proportion of the TP runoff is approximately 18% (ranging from 6% to 49%) for all the rivers. We argue that the name 'Water Tower' is an inappropriate and misleading perception of the TP, and such misperception would influence policy-making processes and diplomatic activities. We therefore call for correcting the misunderstanding and an ensuring accurate understanding of the TP and its role in water supply for downstream countries. We propose using the term "Towering Asian Spring" instead of "Asian Water Tower" to better illustrate the role of the TP in water supply: while it serves as the source of several major rivers in Asia, its contribution to the overall water supply is relatively limited.

In the face of escalating urban pluvial floods exacerbated by climate change, conventional roof systems fall short of effectively managing precipitation extremes. This paper introduces a smart predictive solution: the Smart Internal Drainage Roof (SIDR) system, which leverages forecasted data to enhance the mitigation of pluvial floods in Central Business District (CBD) areas. Unlike traditional approaches, SIDRs utilize a synergistic combination of Rule-based Control (RBC) and Model Predictive Control (MPC) algorithms, tailored to optimize the operational efficiency of both grey and green roofs. Within the examined 1.3 km2 area in Beijing, China, SIDRs, covering 11% of the site, decreased total flooded areas by 30%-50% and eliminated 60%-100% of high-risk zones during three actual events. Moreover, SIDRs streamlined outflow processes without extending discharge time and reduced flood duration at a high-risk underpass by more than half. The SIDR's distinct features, including a high control resolution of 5 min, integration with existing waterproofs, and advanced 2D dynamic runoff visualization, position it as a scalable and cost-efficient upgrade in urban flood resilience strategies.

Flooding in underground spaces, such as subway stations, underground malls, and garages, has increased due to intensified rainfall, urbanization, and population growth. Traditional 2D simulations often overlook crucial vertical flow variations, especially in steep transitions like stairs and ramps. The current study aims to investigate the flood dynamics in large underground geometries by taking a parking lot in Beijing, China, as a study case. The model overcomes the limitations of previous simulations by adapting a full 3D mesh-based simulation with reasonable computational cost. Unlike earlier studies, this model employs a high temporal resolution transient inflow at the inlet to the underground space. Simulation scenarios consider different return periods (5, 20, and 100 years) and inlet water depths, providing an analysis of their impact on flood status in the underground structure. The model generates high spatial–temporal results, enabling precise detection of flood-prone locations, evacuation times, and suggested mitigation techniques. The results recommend evacuating from hazard areas before the 10th minute during extreme flood events. Additionally, the study estimates a 40% increase in flood hazards for scenarios with direct connections between levels. Overall, the study highlights the importance of 3D simulations for accurate risk assessment.

Precipitation nowcasting is a crucial element in current weather service systems. Data-driven methods have proven highly advantageous, due to their flexibility in utilizing detailed initial hydrometeor observations, and their capability to approximate meteorological dynamics effectively given sufficient training data. However, current data-driven methods often encounter severe approximation/optimization errors, rendering their predictions and associated uncertainty estimates unreliable. Here a probabilistic diffusion model-based precipitation nowcasting methodology is introduced, overcoming the notorious blurriness and mode collapse issues in existing practices. Diffusion models learn a sequential of neural networks to reverse a pre-defined diffusion process that generates the probability distribution of future precipitation fields. The precipitation nowcasting based on diffusion model results in a 3.7% improvement in continuous ranked probability score compared to state-of-the-art generative adversarial model-based method. Critically, diffusion model significantly enhance the reliability of forecast uncertainty estimates, evidenced in a 68% gain of spread-skill ratio skill. As a result, diffusion model provides more reliable probabilistic precipitation nowcasting, showing the potential to better support weather-related decision makings.

Despite the implications of winter precipitation for socioeconomic activities and transportation services, the influence of cities on winter precipitation is less studied compared to that on summer precipitation. Here we investigated the statistical relations between precipitation, temperature, and impervious surface fraction in 12 major cities across the contiguous United States. The results showed negative correlations between snowfall intensity and impervious surface fraction. The correlations depend on latitude and the distance to complex terrain features (water bodies or topography), with stronger correlations for inland cities than coastal/lakeside cities. We further selected Kansas City for modeling analyses based on the Weather Research and Forecasting model. Simulation results indicated that the heating effect of urban land occurs in the near-surface atmosphere during the precipitation period, leading to changes of different hydrometers and an overall tendency of reducing snowfall but increasing rainfall.

On August 8th, 2022, an extreme rainfall event (the 88ER) occurred over South Korea's metropolitan area and resulted in immense losses of human lives and properties. Previous study has attributed the rainfall event to the intersection of warm and cold air induced by a Northeast China Cold Vortex (NCCV) and the persistently northward displacement of the West Pacific Subtropical High (WPSH). However, in addition to dynamic drivers, understanding the moisture transport of the 88ER is likewise crucial for developing effective strategies to prevent rainstorm disasters. In this study, based on the output from a WRF model, the primary moisture sources and transport pathways of the 88ER are investigated in a Lagrangian view. The Yellow Sea and East China Sea (YSECS) are identified as the most significant moisture source region (84.42%), followed by South Korea (KR), the eastern China (EC) and Democratic People's Republic of Korea (DPRK), which contribute 12.52%, 1.52% and 1.43% of the released moisture, respectively. Furthermore, to assess the sensitivity of moisture fluxes and heavy rainfall to the sea surface temperature (SST) anomalies in the YSECS, an additional WRF model experiment is conducted in which the SST anomalies are replaced by the average SST over the past 30 years. It is found that the SST anomalies in the YSECS cause differences in atmospheric circulation, and therefore exert a strong influence on moisture transport. The SST anomalies finally enhance the moisture contribution of the YSECS by 1.72%, but decrease that over KR, EC and DPRK by 1.03%, 0.35% and 0.33%, respectively. This study aims to explore the primary moisture sources of an extreme rainfall event that occurred over South Korea in August 2022 from a Lagrangian perspective. Meanwhile, the sensitivity of moisture transport to the warm sea surface temperature in the the Yellow Sea and East China Sea are investigated. image