Road infrastructure is facing increasing flooding risks, causing asset damage and disrupting traffic flows. Effective risk management requires integrated assessments that capture network vulnerability, disruption and recovery. While localised studies have simulated traffic disruptions, national-scale assessments have largely focused on flood exposure rather than systemic disruption analysis. We developed a modelling framework combining process-based flow model of passenger travel-to-work flows and applied it to stress-test Great Britain’s road networks against 17 historical flood events from 1953 to 2024. Results reveal significant variations between direct and indirect damage losses, with single carriageway A roads and suburban bridges emerging as critical points. Notably, indirect losses due to disruption and rerouting can be significantly higher than direct damages depending on hazard event. Early clearance and speed restriction removal are key to mitigating the overall indirect impacts. The model is generalisable and can be applied to stress-test other road networks and flood scenarios worldwide.

Accurate prediction of shoreline change is vital for effective coastal planning and management, especially under increasing climate variabilities. This study explores the applicability of deep learning (DL) techniques, particularly Long Short-Term Memory (LSTM) and Convolutional Neural Network-LSTM (CNN-LSTM) models, for shoreline forecasting at monthly to inter-annual timescales, under two modelling approaches-direct input (DI) and autoregressive (AR). All models demonstrated the ability to reproduce temporal shoreline variability, while the autoregressive DL models were performing better. Further, a noise impact assessment revealed that seasonal decomposition and noise filtering significantly enhanced the model performance. In particular, the models using 52-week data decomposition and residual noise reduction improved the model performance. The reduction of data noises also resulted in narrower ensemble prediction envelopes, indicating that ensemble candidate models behave with low diversity. The temporal data resolution analysis showed that lower data resolutions reduce the predictive performance of the model and at least fortnightly data are required to satisfactorily capture the trend of variability of the shoreline position at this beach. The use of ensemble predictions, derived from a selected subset of model trials based on their collective performance, proved beneficial by capturing diverse temporal behaviours, thereby offering a quasi-probabilistic forecast with minimal computational cost. Overall, the study underscores the potential of DL models, particularly with autoregressive architectures, for reliable and transferable shoreline change prediction. It also emphasizes the importance of data quality, resolution, and preprocessing in improving model robustness, laying the groundwork for future research into use of DL in multi-scale shoreline predictions.

Flood loss models are increasingly used in the (re)insurance sector to inform a range of financial decisions, and more broadly in research and policy analysis to understand present-day and future flood risk trends. These models simulate the interactions between flood hazard, vulnerability and exposure over large spatial domains, requiring a range of input information and modelling assumptions. Due to this high level of complexity, evaluating the impact of uncertain input data and assumptions on modelling results, and therefore the overall model “acceptability”, remains a very complex process. In this paper, we advocate for the use of global sensitivity analysis (GSA), a generic technique to analyse the propagation of multiple uncertainties through mathematical models, to improve the sensitivity testing of flood loss models and the identification of their key sources of uncertainty. We discuss key challenges in the application of GSA to large-scale flood loss models, propose pragmatic strategies to overcome these challenges, and showcase the type of insights that can be obtained by GSA through two proof-of-principle applications to a commercial model, JBA Risk Management's flood loss model, for the transboundary Rhine River basin in Europe, and Queensland in Australia.

A key challenge in flood risk analysis is the construction of hazard events that are physically plausible yet extend beyond historical observations with appropriate frequency and spatial coherence. This is commonly addressed through large simulations of synthetic weather scenarios that sample low-likelihood, high-impact events beyond the observed record. Although popular in industrial risk-based workflows, traditional statistical approaches to synthetic weather generation can be limited in their ability to represent the full range of physically plausible variability and spatial structure.Here, we demonstrate a framework that uses an AI-based weather model as a stochastic generator of event sets suitable for flood risk assessment. We adapt the huge ensembles (HENS) approach using a Spherical Fourier Neural Operator (SFNO)-based atmospheric model combined with a diagnostic precipitation model, forming a framework termed "PrecipHENS". This framework produces more than 1000 synthetic European winter seasons of precipitation and temperature at 0.25 degrees resolution, with modest computational cost (using NVIDIA Earth-2 stack, 112 GPU hours on NVIDIA L40s GPUs).Using an Elbe River case study, we evaluate PrecipHENS against risk-relevant criteria, including reproduction of present-day climatology, preservation of spatial and temporal dependence, representation of extremes, and extrapolation beyond the historical record in event space. PrecipHENS reproduces key features of precipitation and temperature climatology, preserves spatial dependence, including the decay of extremal co-occurrence with distance, and generates a substantially broader diversity of extreme precipitation events than an industry-standard conditional multivariate extreme-value benchmark. Principal component analysis of extreme precipitation fields shows that PrecipHENS spans a much broader space of storm structures than the benchmark or the historical record, indicating it is able to produce previously unseen weather rather than repetition of past patterns.To assess flood risk relevance, the AI-generated weather sequences are coupled with a hydrological model. The resulting river flow simulations are consistent with observed climatology and extreme discharge behaviour, demonstrating that meteorological realism translates into physically plausible hydrological response. Together, these results show that AI-based large-ensemble weather generation can support event set construction for flood hazard and flood risk applications. More broadly, this framework provides a pathway for expanding the physically plausible sample space in applications that require robust characterisation of extremes, including risk assessment, climate-impact analysis, and storyline development.

The temporal clustering of storms can present successive natural hazards for coastal areas in the form of extreme sea levels, storm surges and waves. Studies have investigated the prevalence of the temporal clustering of such hazards but are hindered by the rarity of the phenomena combined with short records and a lack of data availability around the coastline. This has made it difficult to determine if the levels of clustering reported were typical for the location or were being masked by natural variability or climate change over different timescales. In this study, we assess a near 500-year model simulation of extreme sea levels and storm surges forced with pre-industrial meteorological conditions to quantify the levels of temporal clustering seen from natural variability around Great Britain. We then utilise a 50-year rolling window to see how clustering statistics can change through time when dealing with time periods that are representative of the average length of a record in the United Kingdom National Tide Gauge Network. When using near 500-year timeseries, we highlight that many clustering statistics return values close to their statistical expectancies. However, when analysing discrete 50-year windows, results can vary dramatically. The percentage of years with an extreme sea level or surge exceedance at a given location at the 1 in 1-, 5-, and 10-year return level, can vary by up to ~ 33%, ~ 24%, and ~ 18%, the mean number of days between consecutive sea level or surge exceedances can vary by ~ 231, ~14,780, and ~ 17,793 days, and the extremal index can vary by ~ 0.37, ~ 0.64, and ~ 0.79, respectively. Although these results represent the best estimate of the levels of clustering to be expected under natural variability, a comparison of the longest records in the tide gauge network and their nearest model grid nodes shows a tendency for the model to underestimate the clustering statistics that are calculated from the measured data (apart from the extremal index). As such, these can be considered to represent the minimum levels of temporal clustering around Great Britain, as the potential underestimation of clustering, combined with climatic change and sea level rise, means that the temporal clustering of sea levels and storm surges are likely to be far greater over the next 500 years.

Extremely rapid rates of rise in level and discharge in a subset of flash floods ('abrupt wave front floods', AWF) are separate hazards from peak level. Such flood events are investigated for Pennine catchments in northern England using both gauged and historical information. Gauged level and flow digital records at 15-min intervals provide recent data. Historical information for 122 AWF events is extracted from a chronology of flash floods for Britain. Historical AWF events are mapped and found to occur on every major Pennine catchment; catchment descriptors are derived as a basis for assessing catchment vulnerability. We discuss the disputed origin of AWF. Using gauged data, we contrast the rising limb of AWF and 'normal' floods. We investigate time series of historical AWF, noting a puzzling peak in the late 19th century. Current rainfall and river monitoring does not provide a reliable basis for understanding AWF processes or for operational response and we suggest improvements. Similarly, current models for design flood estimation and forecasting do not generate the observed rapid increase in level in AWF floods.

Floods are among the costliest and deadliest natural hazards. Flood risk assessments are required to better manage risk associated with floods. Nowadays, numerous flood risk models are available at various scales, from catchment to regional or even global scale. These models estimate risk (usually expressed in terms of the probability of flood loss) as the product of the hazard, exposure and vulnerability. Flood risk models are affected by numerous uncertainties that propagate through the model and contribute to the final uncertainty in risk estimates. Knowing which uncertainty sources mostly control risk estimates is essential to guide efforts for model improvement, as well as to help risk managers make better decisions. Past efforts to quantify and attribute the output uncertainty of risk models have reached conflicting conclusions. This may be because these studies used different risk models and different uncertainty and sensitivity analysis approaches; or, that they were conducted at relatively small (catchment and/or city) scale, in places with different climatic, hydrological, and socio-economic characteristics. In this project, we investigate dominant uncertainties of a flood risk model across a much larger scale, namely the entire Rhine River basin, and explore whether dominant uncertainties at specific places can be linked to their physical or socio-economic characteristics. In particular, we analyse two model outputs: the Average Annual Losses (AAL) and Loss Exceedance Curves (LECs). For each output, we first identify the dominant input uncertainties (among uncertainty in the flood depth estimates, vulnerability curves and exposure dataset) in each spatial unit of the modelled domain; and second, we link those dominant input uncertainties to the characteristics of the spatial units. We find that uncertainties in the vulnerability component dominate the AAL. The dominant uncertainties for the LECs change with the return period of loss, with vulnerability becoming increasingly important with increasing return period. Topography (flat versus steep terrains), degree of urbanization and economic value of the buildings are key characteristics for determining how dominant uncertainties change spatially within our study domain.

Floods are extreme natural hazards often with disastrous impacts on the economy and society. Flood risk assessments are required to better manage risk associated with floods. Nowadays, numerous flood risk models are available at various scales, from catchment to regional or even global scale. They involve a complex modelling chain that estimates risk as the product of probability of occurrence of an event (hazard) with its footprint (exposure) and the consequences over society and economy (vulnerability). Each component of this chain contains uncertainties, that propagate and contribute to the uncertainty in the model outputs. Much effort has been made to quantify such output uncertainty and attribute it to the various uncertainty sources in the modelling chain. However, the key drivers of uncertainty in flood risk estimates are still unclear because previous studies have reached conflicting conclusions. Two things could possibly explain these ambiguous outcomes. First, these studies were implemented with different models and with different data, as well as different assumptions for the uncertainty and sensitivity analysis. Second, the studies were conducted at catchment and/or city scale with limited variability of physical and socio-economic characteristics within a study region, but with potentially large differences across study regions. In this project, we study the question of uncertainty quantification and attribution at much larger scale, namely the heterogeneous region of the Rhine River basin. In this way, we can identify spatial patterns of dominant input uncertainties and link them to characteristics, e.g. physical, socio-economic, in the different sub-basins. To this end, we use an industry flood risk model (catastrophe model) provided by JBA Risk Management which is capable of simulating flood risk across such a large region. Our ultimate goal is to provide evidence of how the importance of uncertainties varies across places with different climatic, hydrologic and socio-economic characteristics.

There has been rapid progress in the development of capabilities to analyse infrastructure networks on very large scales, up to global scales. This is enabled by the growing availability of geospatial data products with global coverage and computational capabilities, which enable processing of these datasets and analytics on large-scale. Global analyses of the risks from climatic hazards to infrastructure networks serve several important purposes: * Quantified risk estimates in future climate scenarios contribute to the overall picture of the scale of climate risks worldwide, which helps to motivate climate mitigation and adaptation. * Geospatial analysis of hotspots of infrastructure vulnerability helps to target adaptation actions. * Cost-benefit analysis of adaptation enables the prioritization of scarce adaptation resources. * Quantified climate risk analysis is increasingly required for financial disclosure of physical climate risks by infrastructure investors. There are inevitable limitations to global-scale analyses, but they enable cross-country comparisons, and the monitoring of changing risks and national infrastructure resilience. Global analyses also provide a convenient starting point for national analyses and a motivation to collect better data to inform national-scale decisions. Here we present recent developments in capability for global-scale climate risk analysis to infrastructure networks. The analysis combines (i) global-scale probabilistic hazard layers (including floods, hurricanes and coastal storm surges); (ii) infrastructure asset and network exposure, for energy, transport and telecommunications networks (iii) analysis of the people and economic activities that are dependent upon these networks. This quantified risk analysis framework has been efficiently implemented for global-scale computations, yielding new results on the scale of climate-related risks. Analysis of resource flows on networks and their connection to infrastructure users is enabling calculation of the numbers of people and economic activity that may be disrupted in catastrophic events. A recent development has been in the introduction of probabilistic event sets for hurricanes and flooding, which enables accurate estimation of the impacts from spatially extensive extreme events. The research is being made available as part of the Global Resilience Index Initiative https://www.cgfi.ac.uk/global-resilience-index-initiative/ and as an open source toolset and interface for geospatial visualisation.

Bridges are critical-infrastructure components of road and rail transport networks. A large number of these critical assets cross or are adjacent to waterways and flood-plains and are therefore exposed to flood actions such as scour, hydrodynamic loading, and inundation, all of which are exacerbated by debris accumulations. These stressors are widely recognized as responsible for the vast majority of bridge failures around the world, and they are expected to be exacerbated due to climate change. While efforts have been made to increase the robustness of bridges to the flood hazard, many scientific and technical gaps remain. These gaps were explored during an expert workshop that took place in April 2021 with the participation of academics, consultants, and decision makers operating mainly in the United Kingdom and specializing in the fields of bridge risk assessment and management and flood resilience. The objective of the workshop was to identify and prioritize the most urgent and significant impediments to bridge flood resilience. In particular, the following issues, established at different levels and scales of bridge flood resilience, were identified and analysed in depth: (i) characterization of the effects of floods on different bridge typologies, (ii) uncertainties in formulae for scour depth assessment, (iii) evaluation of consequences of damage, (iv) recovery process after flood damage, (v) decision-making under uncertainty for flood-critical bridges, and (vi) use of event forecasting and monitoring data for increasing the reliability of bridge flood risk estimations. These issues are discussed in this paper to inform other researchers and stakeholders worldwide, guide the directions of future research in the field, and influence policies for risk mitigation and rapid response to flood warnings, ultimately increasing bridge resilience.

The temporal clustering of storms presents consecutive storm surge and wave hazards that can lead to amplified flood and erosional damages; thus, clustering is important for coastal stakeholders to consider. We analyse the prevalence of storm clustering around the UK coastline by examining the temporal and spatial characteristics of storm surge, wave height, and high still sea level exceedances at the 1 in 1- and 5-year return levels. First, at the interannual timescale, we show that there are periods of high/low exceedance counts on national and regional scales. Elevated annual counts of exceedances with smaller magnitudes can occur without a respective signal of higher-magnitude exceedances. Secondly, at the intra-annual timescale, we show that high proportions of exceedances are clustering over short timescales. Storm surge, wave height and still sea level exceedances occurring < 50 days after the prior exceedance at a given site account for between ~ 35–44% and ~ 15–22% of all exceedances at the 1 in 1- and 5-year return levels, respectively. Still sea levels have the highest proportion of exceedances clustered in quick succession, with ~ 25% of 1 in 1-year exceedances occurring < 2 days after the previous at the same site. Spatially, for UK storm surges and still sea levels, the North Sea has the lowest proportion of clustering, whereas the North Atlantic and Bristol Channel have the highest. For English wave records, the highest proportions of clustering are found in the North Sea for exceedances of a lower magnitude and the English Channel for exceedances of a higher magnitude. These findings illuminate the prevalence of the clustering of coastal hazards around the UK—helping coastal stakeholders evaluate the threat of surges, waves, and sea levels clustering over short periods.

A ‘roadmap’ for the future of UK flood hydrology over the next 25 years has been published, based on a wide-ranging and inclusive co-creation process involving more than 270 individuals and 50 organisations from different sectors and disciplines. This paper highlights key features of the roadmap and its development as a community-owned initiative. The roadmap's relationship with hydrological research and practice is discussed, as is its context within the wider flood risk management innovation landscape, including funding. While the paper has a focus on UK flood hydrology, reflecting the scope of the roadmap, it is also considered in the context of advances in hydrology internationally.

In this study, we utilise Artificial Neural Network (ANN) models to estimate the 100- and 1500-year return levels for around 900,000 ungauged catchments in the contiguous USA. The models were trained and validated using 4,079 gauges and several selected catchment descriptors out of a total of 25 available. The study area was split into 15 regions, which represent major watersheds. ANN models were developed for each region and evaluated by calculating several performance metrics such as root-mean-squared error (RMSE), coefficient of determination (R2) and absolute percent error. The availability of a large dataset of gauges made it possible to test different model architectures and assess the regional performance of the models. The results indicate that ANN models with only one hidden layer are sufficient to describe the relationship between flood quantiles and catchment descriptors. The regional performance depends on climate type as models perform worse in arid and humid continental climates. Overall, the study suggests that ANN models are particularly applicable for predicting ungauged flood quantiles across a large geographic area. The paper presents recommendations about future application of ANN in regional flood frequency analysis.

The theory that forms the basis of TOPMODEL (a topography-based hydrological model) was first outlined by Mike Kirkby some 45 years ago. This paper recalls some of the early developments, the rejection of the first journal paper, the early days of digital terrain analysis, model calibration and validation, the various criticisms of the simplifying assumptions, and the relaxation of those assumptions in the dynamic forms of TOPMODEL. A final section addresses the question of what might be done now in seeking a simple, parametrically parsimonious model of hillslope and small catchment processes if we were starting again.

The risk of surface water flooding (SWF) in England is already high and its frequency and severity is projected to increase in the future. SWF generally occurs due to intense, highly localised rainfall, which is challenging to forecast with sufficient accuracy to take proactive action ahead of flood events. Being able to manage the risk effectively lies in improved rainfall and flood forecast products, better communication of uncertainty and building the capacity of local responders. This study utilises state‐of‐the‐art high‐resolution ensemble rainfall forecasts and hydraulic modelling tools alongside a novel post‐processing method to develop and trial new SWF forecast products within an incident workshop attended by forecast producers and regional forecast users. Twenty‐two of 24 workshop participants reported that the new information would be useful to their organisation but more product development and training in its interpretation is required. Specific recommendations to improve SWF forecast provision include increased support for local government through a single government organisation responsible for SWF, making more use of existing static SWF mapping in a real‐time context and employing the process of user‐based consultation, as outlined in this study, to guide the future development of future SWF forecast information and processes.

Digital technology is having a major impact on many areas of society, and there is equal opportunity for impact on science. This is particularly true in the environmental sciences as we seek to understand the complexities of the natural environment under climate change. This perspective presents the outcomes of a summit in this area, a unique cross-disciplinary gathering bringing together environmental scientists, data scientists, computer scientists, social scientists, and representatives of the creative arts. The key output of this workshop is an agreed vision in the form of a framework and associated roadmap, captured in the Windermere Accord. This accord envisions a new kind of environmental science underpinned by unprecedented amounts of data, with technological advances leading to breakthroughs in taming uncertainty and complexity, and also supporting openness, transparency, and reproducibility in science. The perspective also includes a call to build an international community working in this important area.

Floods are extreme hydro-meteorological hazards that pose significant risks to the economy and society. Reducing the risk associated with floods and better adapting to them is a daunting task because flood risk dynamics are influenced by different factors. Flood risk is usually defined as the product of three components: hazard, exposure and vulnerability. Global Flood Risk Models (GFRM) represent the underlying physical hazard, the exposure of people, properties or other assets to the hazard, and the losses that may occur following a flood event. Consequently, they are used by governmental agencies, risk reduction organisations, global investors and the (re)insurance industry to help manage the societal and financial risks associated with floods. GFRMs are subject to many sources of uncertainty, including uncertainty in processes representation, model parameters and input data; however, the relative importance of these different sources is poorly understood. Currently, no evidence exists on which uncertain input factor mostly control the final uncertainty in predicted losses in different places and circumstances. In this project, we use JBA’s (a leading flood risk modelling company) Global Flood Model and Open Exposure Data (OED) to develop an appropriate methodological approach to analyse the sensitivity of loss predictions in a structured way. This is particularly challenging as input uncertainties exhibit complex spatially distributed and spatially-structured (correlated) patterns. We apply the methodology to the Rhine river basin, covering regions with different physical and socio-economic characteristics. We pursue the following objectives; (1) Identify and quantify the various sources of uncertainty e.g. associated to rainfall data, extraction of flood events sets, defence database, vulnerability curves, exposure portfolios (2) Analyse their relative importance on flood losses predictions across places along the river (3) Understand which of them are most important at each place. We aim to produce scientifically robust evidence about the importance of different sources of uncertainty across places with different climate, hydrology and socio-economic characteristics and try to address questions related to exposure and vulnerability dynamics, flood losses modelling and adaptation strategies. Such evidence base will help prioritise efforts for uncertainty reduction of the case study model, as well as other flood risk models used by (re)insurers and government agencies, ultimately contributing to more informed decisions for flood risk mitigation.

There is a no lack of significant open questions in the field of hydrology. How will hydrological connectivity between freshwater bodies be altered by future human alterations to the hydrological cycle? Where does water go when it rains? Or what is the future space–time variability of flood and drought events? However, the answers to these questions will vary with location due to the specific and often poorly understood local boundary conditions and system properties that control the functional behaviour of a catchment or any other hydrologic control volume. We suggest that an open, shared and evolving perceptual model of a region's hydrology is critical to tailor our science questions, as it would be for any other study domain from the plot to the continental scale. In this opinion piece, we begin to discuss the elements of and point out some knowledge gaps in the perceptual model of the terrestrial water cycle of Great Britain. We discuss six major knowledge gaps and propose four key ways to reduce them. While the specific knowledge gaps in our perceptual model do not necessarily transfer to other places, we believe that the development of such perceptual models should be at the core of the debate for all hydrologic communities, and we encourage others to have a similar debate for their hydrologic domain.

With so many things around us continuously producing and processing data, be it mobile phones, or sensors attached to devices, or satellites sitting thousands of kilometres above our heads, data is becoming increasingly heterogeneous. Scientists are inevitably faced with data challenges, coined as the 4 V's of data - volume, variety, velocity and veracity. In this paper, we address the issue of data variety. The task of integrating and querying such heterogeneous data is further compounded if the data is in unstructured form. We hence propose an approach using Semantic Web and Natural Language Processing techniques to resolve the heterogeneity arising in data formats, bring together structured and unstructured data and provide a unified data model to query from disparate data sets.

The Q-natural flood management project has co-developed with the Environment Agency 18 monitored micro-catchments (similar to 1 km(2)) in Cumbria, UK installing calibrated flumes aimed at quantifying the potential shift in observed flows resulting from a range of nature-based-solutions installed by local organizations. The small-scale reduces the influence of variability characterizing larger catchments that would otherwise mask any such shifts, which we attempt to relate to a shift in model parameters. This paper demonstrates an approach to applying donor-parameter-shifts obtained from modelling two of the paired micro-catchments to a much larger scale, in order to understand the potential for improved distributed modelling of nature-based solutions in the form of additional tree-planting. The models include a rainfall-runoff model, Dynamic Topmodel, and a 2D hydrodynamic model, JFlow, permitting analysis of changes in hillslope processes and channel hydrodynamics resulting from a range of distributed measures designed to emulate natural hydrological processes that evaporate, store or infiltrate flows. We report on attempts to detect shift in hydrological response using one of the paired-micro-catchment moorland versus forestry sites in Lorton using Dynamic Topmodel. A donor-parameter-shift approach is used in a hypothetical experiment to represent new woodland in a much larger catchment, although testing all combinations of spatial planting strategies, responses to multiple-extremes, failure-modes and changes to synchronization becomes intractable to support good decision making. We argue that the problem can be re-framed to use donor-parameter-shifts at multi-local-scale catchments above communities known to be at risk, commensurate with most of the evidence of NbS impacts being effective at the small scale (ca. 10 km(2)). This might lead to more effective modelling to help catchment managers prioritize those communities-at-risk where there is more evidence that NbS might be effective.