Search

Risk management has reduced vulnerability to floods and droughts globally1,2, yet their impacts are still increasing3. An improved understanding of the causes of changing impacts is therefore needed, but has been hampered by a lack of empirical data4,5. On the basis of a global dataset of 45 pairs of events that occurred within the same area, we show that risk management generally reduces the impacts of floods and droughts but faces difficulties in reducing the impacts of unprecedented events of a magnitude not previously experienced. If the second event was much more hazardous than the first, its impact was almost always higher. This is because management was not designed to deal with such extreme events: for example, they exceeded the design levels of levees and reservoirs. In two success stories, the impact of the second, more hazardous, event was lower, as a result of improved risk management governance and high investment in integrated management. The observed difficulty of managing unprecedented events is alarming, given that more extreme hydrological events are projected owing to climate change3.

Rainfall‐runoff models are commonly evaluated against statistical evaluation metrics. However, these metrics do not provide much insight into what is hydrologically wrong if a model fails to simulate observed streamflow well and they are also not applicable for ungauged catchments. Here, we propose a signature‐based hydrologic efficiency (SHE) metric by replacing the statistical components of current efficiency metrics with hydrologic signatures that can be regionalized for model evaluation in ungauged catchments. We test our new efficiency metric across 633 catchments from Great Britain. Strong correlations with Spearman rank and Pearson correlation values around 0.8 are found between our proposed metric and commonly used statistical evaluation metrics (NSE, KGE, and NP) demonstrating that the proposed SHE metric is related to existing metrics as much as these metrics are related to each other. For ungauged catchments, we regionalize the three signatures included in SHE and find that 78% of catchments have an absolute difference of SHE values between gauged and ungauged cases of less than 0.2. This difference varies depending on the quality of the regionalized bias and variance signature values. It means that the SHE metric is applicable for model evaluation in ungauged catchments if its signatures can be regionalized well. When applying the SHE metric in other domains with different hydrologic properties, modellers should carefully consider the components of the proposed metric with signatures best suited to their research domains and their regionalization potential. We propose a signature‐based hydrologic efficiency metric that can be applied in both gauged and ungauged catchmentsWe use it to assess model performances in over 600 catchments in Great Britain and compare it to established efficiency metricsModellers should replace the components of the proposed metric with signatures best suited to the hydrology of their study domains We propose a signature‐based hydrologic efficiency metric that can be applied in both gauged and ungauged catchments We use it to assess model performances in over 600 catchments in Great Britain and compare it to established efficiency metrics Modellers should replace the components of the proposed metric with signatures best suited to the hydrology of their study domains

Rainfall‐runoff models are often deficient under changing climatic conditions, yet almost no recent studies propose new or improved model structures, instead focusing on model intercomparison, input sensitivity, and/or quantification of uncertainty. This paucity of progress in model development is (in part) due to the difficulty of distinguishing which cases of model failure are truly caused by structural inadequacy. Here we propose a new framework to diagnose the salient cause of poor model performance in changing climate conditions, be it structural inadequacy, poor parameterization, or data errors. The framework can be applied to a single catchment, although larger samples of catchments are helpful to generalize and/or cross‐check results. To generate a diagnosis, multiple historic periods with contrasting climate are defined, and the limits of model robustness and flexibility are explored over each period separately and for all periods together. Numerous data‐based checks also supplement the results. Using a case study catchment from Australia, improved inference of structural failure and clearer evaluation of model structural improvements are demonstrated. This framework enables future studies to (i) identify cases where poor simulations are due to poor calibration methods or data errors, remediating these cases without recourse to structural changes; and (ii) use the remaining cases to gain greater clarity into what structural changes are needed to improve model performance in changing climate. Rainfall runoff models are tools used by hydrologists in climate change assessments to estimate how future streamflow might change in response to a given (often hypothetical) climate scenario. For example, suppose we can assume that rainfall in a particular location is going to reduce by 20% in the future. Does this mean that streamflow will also reduce by 20%? Or will it be 10% less or 40% less? Although rainfall runoff models are among the best tools available, they are often not very good at answering this question. When tested on historical multiyear droughts, they often perform poorly, and we are unsure why. One problem is that when a model fails in this task, it is difficult to know what went wrong. Perhaps there was a problem with the data, since environmental monitoring is often subject to large errors. Perhaps the problem lay not with the model itself but with the way it was trained, or calibrated, to the data. Lastly, perhaps the model itself—its mathematical equations—need to be changed. To improve our estimates, we need a method to test which cause is behind the model failure; otherwise, we might make changes where none are warranted. This paper proposes such a method, in the form of a multistep framework that can isolate the causes of model failures. By ensuring that our attention is focused in the correct direction, this framework will help us to understand and make better estimates of how river flow will be altered by a changing climate. Diagnosing model deficiency is difficult as there are many possible causes of poor simulationsSplit sample failure does not mean model structural inadequacy—further tests are requiredThis framework diagnoses the cause of poor performance and prioritizes remedial action

Streamflow time series are commonly derived from stage‐discharge rating curves, but the uncertainty of the rating curve and resulting streamflow series are poorly understood. While different methods to quantify uncertainty in the stage‐discharge relationship exist, there is limited understanding of how uncertainty estimates differ between methods due to different assumptions and methodological choices. We compared uncertainty estimates and stage‐discharge rating curves from seven methods at three river locations of varying hydraulic complexity. Comparison of the estimated uncertainties revealed a wide range of estimates, particularly for high and low flows. At the simplest site on the Isère River (France), full width 95% uncertainties for the different methods ranged from 3 to 17% for median flows. In contrast, uncertainties were much higher and ranged from 41 to 200% for high flows in an extrapolated section of the rating curve at the Mahurangi River (New Zealand) and 28 to 101% for low flows at the Taf River (United Kingdom), where the hydraulic control is unstable at low flows. Differences between methods result from differences in the sources of uncertainty considered, differences in the handling of the time‐varying nature of rating curves, differences in the extent of hydraulic knowledge assumed, and differences in assumptions when extrapolating rating curves above or below the observed gaugings. Ultimately, the selection of an uncertainty method requires a match between user requirements and the assumptions made by the uncertainty method. Given the significant differences in uncertainty estimates between methods, we suggest that a clear statement of uncertainty assumptions be presented alongside streamflow uncertainty estimates. Knowledge of the uncertainty in streamflow discharge measured at gauging stations is important for water management applications and scientific analysis. This paper shows that uncertainty estimates vary widely (typically up to a factor of 4) when comparing seven recently introduced estimation methods. A clear understanding of the assumptions underpinning different uncertainty estimation methods and the sources of uncertainty included in their calculations is needed when selecting a method and using and presenting its uncertainty estimates. Methods for estimating the stage‐discharge rating curve and its uncertainty were compared for stream gauges with varying hydraulic complexityUncertainty estimates varied widely at high and low flows for the different methodsCareful description of the assumptions behind uncertainty methods is needed

Inflow discharge and outflow stage estimates for hydraulic flood models are generally derived from river gauge data. Uncertainties in the measured inflow data and the neglect of rainfall‐runoff contributions to the modeled domain downstream of the gauging locations can have a significant impact on these estimated “whole reach” inflows and consequently on flood predictions. In this study, a method to incorporate rating curve uncertainty and local rainfall‐runoff dynamics into the predictions of a reach‐scale flood model is proposed. The methodology is applied to the July 2007 floods of the River Severn in UK. Discharge uncertainty bounds are generated applying a nonparametric local weighted regression approach to stage‐discharge measurements for two gauging stations. Measured rainfall downstream from these locations is used as input to a series of subcatchment regional hydrological model to quantify additional local inflows along the main channel. A regional simplified‐physics hydraulic model is then applied to combine these contributions and generate an ensemble of discharge and water elevation time series at the boundaries of a local‐scale high complexity hydraulic model. Finally, the effect of these rainfall dynamics and uncertain boundary conditions are evaluated on the local‐scale model. Accurate prediction of the flood peak was obtained with the proposed method, which was only possible by resolving the additional complexity of the extreme rainfall contributions over the modeled area. The findings highlight the importance of estimating boundary condition uncertainty and local rainfall contributions for accurate prediction of river flows and inundation at regional scales. A method to incorporate local rainfall dynamics and inflow discharge and outflow stage uncertainty into flood predictions was developedA computationally efficient three‐level, regional‐regional‐local, hydrologic‐hydraulic‐hydraulic model was applied and testedIt was necessary to simulate local rainfall‐runoff to capture the magnitude and timing of the flood peak stage for the selected event

This study reveals the spatiotemporal variability of event characteristics and their controls from 290,743 rainfall‐runoff events across 431 Great Britain catchments. Metrics characterizing event runoff coefficient, event timescale and normalized peak discharge and their mean, variability and seasonality statistics are extracted to quantify event characteristics. Using random forest models, we find climate attributes are most influential in controlling the spatial pattern, with land cover, geology and soil properties also playing a critical role especially for explaining the variability of event runoff response. Rainfall characteristics (particularly rainfall depth) exhibit dominant controls on the temporal dynamic at event scale, while catchment initial state (antecedent soil moisture) determines its seasonal variations. Clear regional patterns of event characteristics are observed in Great Britain, while the influences of geology and anthropogenic activities on local differences are detected. Extremely low runoff coefficient values (<0.05) but with high variability are typical in Chalk catchments in Southeast region. Flashy events with short event timescale (<10 hr) and less seasonality of runoff response are observed in urban catchments. A large reservoir close to a gauge leads to the reduction of peak flows, yet the deviation is relatively small in summer. The impact of surface water abstractions is not obvious in our data set, whereas the impact of groundwater abstractions largely reflects the geology. This study provides the basis for regionalization studies by revealing the (dis)similarities of catchment runoff response and emphasizes the need for further research into the role of geology and human activities on rainfall‐runoff characteristics. Climate, land cover, and geology are key controls on the spatial pattern of rainfall‐runoff event characteristics across Great BritainRainfall depth and intensity control runoff response at event scale, while the catchment initial state controls its seasonal variabilityRunoff response often follows a clear regional pattern, yet local geology, reservoirs, and urbanization can cause anomalous behavior Climate, land cover, and geology are key controls on the spatial pattern of rainfall‐runoff event characteristics across Great Britain Rainfall depth and intensity control runoff response at event scale, while the catchment initial state controls its seasonal variability Runoff response often follows a clear regional pattern, yet local geology, reservoirs, and urbanization can cause anomalous behavior

Reliable information about hydrological behavior is needed for water-resource management and scientific investigations. Hydrological signatures quantify catchment behavior as index values, and can be predicted for ungauged catchments using a regionalization procedure. The prediction reliability is affected by data uncertainties for the gauged catchments used in prediction and by uncertainties in the regionalization procedure. We quantified signature uncertainty stemming from discharge data uncertainty for 43 UK catchments and propagated these uncertainties in signature regionalization, while accounting for regionalization uncertainty with a weighted-pooling-group approach. Discharge uncertainty was estimated using Monte Carlo sampling of multiple feasible rating curves. For each sampled rating curve, a discharge time series was calculated and used in deriving the gauged signature uncertainty distribution. We found that the gauged uncertainty varied with signature type, local measurement conditions and catchment behavior, with the highest uncertainties (median relative uncertainty ±30–40% across all catchments) for signatures measuring high- and low-flow magnitude and dynamics. Our regionalization method allowed assessing the role and relative magnitudes of the gauged and regionalized uncertainty sources in shaping the signature uncertainty distributions predicted for catchments treated as ungauged. We found that (1) if the gauged uncertainties were neglected there was a clear risk of overconditioning the regionalization inference, e.g., by attributing catchment differences resulting from gauged uncertainty to differences in catchment behavior, and (2) uncertainty in the regionalization results was lower for signatures measuring flow distribution (e.g., mean flow) than flow dynamics (e.g., autocorrelation), and for average flows (and then high flows) compared to low flows. We quantify impact of data uncertainty on signatures and their regionalizationMedian signature uncertainty ±10–40%, and highly variable across catchmentsNeglecting gauging uncertainty causes overconditioning of regionalization

The derived distribution method is a promising approach for flood estimation in ungauged catchments. In this approach, event runoff coefficient (ERC) is often adopted to behave as the runoff generation component since the behavior of ERC is closely related to flood generation mechanisms and its distribution reflects catchment climatological and landscape characteristics. However, there is a lack of understanding of how to transfer the information on ERC characteristics from gauged to ungauged catchments at catchment scale. Hence, here, we propose a generalized framework to estimate the probability distribution of ERC for ungauged catchments based on knowledge of the spatial and temporal controls on ERC from gauged catchments. Key components of the framework include cluster analysis based on ERC characteristics, linking clusters with catchment attributes, and constructing typical probability distribution of ERC conditioned on its temporal indicators. A total of 290,743 rainfall‐runoff events observed in 431 GB catchments during the period 1990–2014 have been employed to verify the framework. Good estimations are obtained with the median value of coefficient of determination (R2) reaching 0.85 across all test catchments. The results indicate that similar pre‐event catchment conditions may cause distinct runoff response in different catchments, thus predicting the correct spatial cluster is crucial to the estimation accuracy. This work sheds light on constructing a stochastic generator model of ERC according to its spatial pattern and temporal dynamic, facilitating a new alternative for flood estimation method in ungauged catchments. The framework has been verified in 431 Great Britain catchments achieving good estimation with a median value of R2reaching 0.85Similar pre‐event conditions may result in various runoff responses across catchments, spatial clustering is crucial to estimation accuracyThis framework facilitates a new alternative for flood estimation in ungauged catchments The framework has been verified in 431 Great Britain catchments achieving good estimation with a median value of R2reaching 0.85 Similar pre‐event conditions may result in various runoff responses across catchments, spatial clustering is crucial to estimation accuracy This framework facilitates a new alternative for flood estimation in ungauged catchments

Spatial parameter fields are required to model hydrological processes across diverse landscapes. Transfer functions are often used to relate parameters to spatial catchment attributes, introducing large uncertainties. Quantifying these uncertainties remains a key challenge for large‐scale modeling. This paper extends the multiscale parameter regionalization (MPR) technique to consider parameter uncertainties. We evaluate this method of producing nationally consistent parameter fields, which maintain a constant relationship between model parameters and catchment attributes, across 437 catchments in Great Britain (GB). By sampling multiple transfer function parameters, we produce thousands of possible model parameter fields which are constrained within an uncertainty framework. This is compared to spatially homogeneous parameter sets constrained for individual catchments. The nationally consistent MPR parameter fields perform well (KGE* > 0.75) across 60% of catchments. Performance is similar or better than catchment‐constrained parameters (KGE* drop < 0.1) across 82% of catchments. Advantages of our national parameter fields include (a) improved representation of flows within catchments, (b) more robust performance between calibration and evaluation periods, and (c) spatial parameter fields reflecting hydrologically meaningful variation in catchment characteristics. By including uncertainties, we show that hydrographs produced using MPR have smaller uncertainty bounds which are better able to encompass flows. As the first application of MPR to both the DECIPHeR modeling framework and GB, we developed transfer functions and identified key catchment attributes to constrain model parameters, which are transferrable to other models alongside the addition of uncertainty. Methodologies presented here are informative for future regionalization efforts in GB and elsewhere. By including uncertainties in multiscale parameter regionalization (MPR), we generate and evaluate ensembles of national model parameter fieldsRegionalized parameter fields result in more consistent performance and hydrographs with smaller uncertainty bounds than Monte‐Carlo constrained parametersThe method of incorporating uncertainties, transfer functions, and selection of catchment attributes can be applied in future regionalization studies By including uncertainties in multiscale parameter regionalization (MPR), we generate and evaluate ensembles of national model parameter fields Regionalized parameter fields result in more consistent performance and hydrographs with smaller uncertainty bounds than Monte‐Carlo constrained parameters The method of incorporating uncertainties, transfer functions, and selection of catchment attributes can be applied in future regionalization studies

This study reports a new and significantly enhanced analysis of US flood hazard at 30 m spatial resolution. Specific improvements include updated hydrography data, new methods to determine channel depth, more rigorous flood frequency analysis, output downscaling to property tract level, and inclusion of the impact of local interventions in the flooding system. For the first time, we consider pluvial, fluvial, and coastal flood hazards within the same framework and provide projections for both current (rather than historic average) conditions and for future time periods centered on 2035 and 2050 under the RCP4.5 emissions pathway. Validation against high‐quality local models and the entire catalog of FEMA 1% annual probability flood maps yielded Critical Success Index values in the range 0.69–0.82. Significant improvements over a previous pluvial/fluvial model version are shown for high‐frequency events and coastal zones, along with minor improvements in areas where model performance was already good. The result is the first comprehensive and consistent national‐scale analysis of flood hazard for the conterminous US for both current and future conditions. Even though we consider a stabilization emissions scenario and a near‐future time horizon, we project clear patterns of changing flood hazard (3σchanges in 100 years inundated area of −3.8 to +16% at 1° scale), that are significant when considered as a proportion of the land area where human use is possible or in terms of the currently protected land area where the standard of flood defense protection may become compromised by this time. We develop a method to estimate past, present, and future flood risk for all properties in the conterminous United States whether affected by river, coastal or rainfall flooding. The analysis accounts for variability within environmental factors including changes in sea level rise, hurricane intensity and landfall locations, precipitation patterns, and river discharge. We show that even for a conservative climate change trajectory we can expect locally significant changes in the land area at risk from floods by 2050, and by this time defenses protecting 2,200 km2of land may be compromised. The complete dataset has been made available via a website (https://floodfactor.com/) created by the First Street Foundation in order to increase public awareness of the threat posed by flooding to safety and livelihoods. First complete high‐resolution flood hazard analysis of conterminous US flood risk from all major sources (fluvial, pluvial, and coastal)In validation tests the model achieved Critical Success Index scores of 0.69–0.82, similar to many local custom‐built 2D modelsBy 2050, flood hazard increases for the Eastern seaboard and Western states, but decreases or changes little for the center and South‐West First complete high‐resolution flood hazard analysis of conterminous US flood risk from all major sources (fluvial, pluvial, and coastal) In validation tests the model achieved Critical Success Index scores of 0.69–0.82, similar to many local custom‐built 2D models By 2050, flood hazard increases for the Eastern seaboard and Western states, but decreases or changes little for the center and South‐West

Water scarcity occurs when water demand exceeds natural water availability over a range of spatial and temporal scales. Though meteorological and hydrological droughts have been analyzed over large spatial scales, the impacts of water scarcity have typically been addressed at a catchment scale. Here we explore how droughts and water scarcity interact over a larger and more complex spatial domain, by combining climate, hydrological, and water resource system models at a national scale across England and Wales. This approach is essential in a highly connected and heterogeneous region like England and Wales, where we represent 80 different catchments; 70 different water resource zones; 16 water utility companies; and the water supply for over 50 million people. We find that if a reservoir's storage is in its first percentile (i.e., the volume that is exceeded 99% of the time), then there is, on average, a 40% chance that reservoirs in neighboring catchments will also be at or below their first percentile storage volume. The coincidence of low reservoir storage decays relatively quickly, stabilizing after about 100–150 km, implying that if interbasin transfers are to be provided to enhance drought resilience, they will need to be at least this length. Based on a large ensemble of future climate simulations, we show that extreme droughts in precipitation, streamflow, and reservoir storage volume are projected to worsen in every catchment. The probability of a year with water use restrictions doubles by 2050 and is four times worse by 2100. There is high spatial coincidence between drought indices (precipitation, runoff, and reservoir volume indicators) for the United KingdomWhen reservoir storage is low precipitation/runoff drought indices are also low, but the best indicators to use vary from region to regionUnder future climates all drought indices get worse in every catchment, with risk of water use restrictions doubling by 2050

Global flood models integrate flood maps of constant probability in space, ignoring the correlation between sites and thus potentially misestimating the risk posed by extreme events. Stochastic flood models alleviate this issue through the simulation of flood events with a realistic spatial structure, yet their proliferation at large scales has historically been inhibited by data quality and computer availability. In this paper, we show, for the first time, the efficacy of modeled river discharge reanalyses in the characterization of flood spatial dependence in the absence of a dense stream gauge network. While global hydrological models may show poor correspondence with absolute observed river flows, we find that the rate at which they can simulate the joint occurrence of relative flow exceedances at two given locations is broadly similar to when a gauge‐based statistical model is used. Evidenced over the United States, flood events simulated using observed gauge data from the U.S. Geological Survey versus those generated using modeled streamflows have similar (i) distributions of site‐to‐site correlation strength, (ii) relationships between event size and return period, and, importantly, (iii) loss distributions when incorporated into a continental‐scale flood risk model. Extremal dependence is generally quantified less accurately on larger rivers, in arid climates, in mountainous terrain, and for the rarest high‐magnitude events. However, local‐scale errors are shown to broadly cancel each other out when combined, producing an unbiased flood spatial dependence model. These findings suggest that building accurate stochastic flood models worldwide may no longer be a distant aspiration. Global flood risk is commonly estimated through flood inundation maps with a defined probability of occurrence. These flood simulations have a key drawback in that they fail to capture the spatial patterns exhibited during real flood events, instead modeling the same probability of flooding on every river at once. Solutions which rely on networks of gauged river flow observations will necessarily break down in the majority of the world's regions which lack such a resource. In this paper, we use historic river flows simulated by global rainfall‐runoff models (rather than observed flows) into a statistical model which captures the spatial correlation of flow extremes. If we examine the relative flow exceedance probabilities from these hydrological models rather than the volumetric flow values, flood events are generated which exhibit similar characteristics to those when gauged flow observations are used. Crucially, the simulation‐ and observation‐generated flood events produce near‐identical losses to buildings in the United States. The implications of this are that true stochastic flood risk models, which account for spatial dependence, can proliferate globally via the generation of realistic flood event sets from hydrological models. Large‐scale flood hazard models typically neglect to represent the spatial dependence of real flood eventsRelative flow exceedances simulated by the fusion of global hydrological with statistical models reproduce gauge‐driven flood event setsAt the continental scale, key characteristics of a flood risk model are indistinguishable when driven with observed versus modeled flow data