Refine
Year of publication
Document Type
- Article (64)
- Postprint (10)
- Monograph/Edited Volume (4)
- Doctoral Thesis (1)
- Review (1)
Is part of the Bibliography
- yes (80) (remove)
Keywords
- adaptation (4)
- precipitation (4)
- uncertainty (4)
- Europe (3)
- Mekong Delta (3)
- classification (3)
- Bayesian networks (2)
- Climate variability (2)
- Costs (2)
- Damage (2)
Understanding and quantifying total economic impacts of flood events is essential for flood risk management and adaptation planning. Yet, detailed estimations of joint direct and indirect flood-induced economic impacts are rare. In this study an innovative modeling procedure for the joint assessment of short-term direct and indirect economic flood impacts is introduced. The procedure is applied to 19 economic sectors in eight federal states of Germany after the flood events in 2013. The assessment of the direct economic impacts is object-based and considers uncertainties associated with the hazard, the exposed objects and their vulnerability. The direct economic impacts are then coupled to a supply-side Input-Output-Model to estimate the indirect economic impacts. The procedure provides distributions of direct and indirect economic impacts which capture the associated uncertainties. The distributions of the direct economic impacts in the federal states are plausible when compared to reported values. The ratio between indirect and direct economic impacts shows that the sectors Manufacturing, Financial and Insurance activities suffered the most from indirect economic impacts. These ratios also indicate that indirect economic impacts can be almost as high as direct economic impacts. They differ strongly between the economic sectors indicating that the application of a single factor as a proxy for the indirect impacts of all economic sectors is not appropriate.
The Limpopo Basin in southern Africa is prone to droughts which affect the livelihood of millions of people in South Africa, Botswana, Zimbabwe and Mozambique. Seasonal drought early warning is thus vital for the whole region. In this study, the predictability of hydrological droughts during the main runoff period from December to May is assessed using statistical approaches. Three methods (multiple linear models, artificial neural networks, random forest regression trees) are compared in terms of their ability to forecast streamflow with up to 12 months of lead time. The following four main findings result from the study. 1. There are stations in the basin at which standardised streamflow is predictable with lead times up to 12 months. The results show high inter-station differences of forecast skill but reach a coefficient of determination as high as 0.73 (cross validated). 2. A large range of potential predictors is considered in this study, comprising well-established climate indices, customised teleconnection indices derived from sea surface temperatures and antecedent streamflow as a proxy of catchment conditions. El Nino and customised indices, representing sea surface temperature in the Atlantic and Indian oceans, prove to be important teleconnection predictors for the region. Antecedent streamflow is a strong predictor in small catchments (with median 42% explained variance), whereas teleconnections exert a stronger influence in large catchments. 3. Multiple linear models show the best forecast skill in this study and the greatest robustness compared to artificial neural networks and random forest regression trees, despite their capabilities to represent nonlinear relationships. 4. Employed in early warning, the models can be used to forecast a specific drought level. Even if the coefficient of determination is low, the forecast models have a skill better than a climatological forecast, which is shown by analysis of receiver operating characteristics (ROCs). Seasonal statistical forecasts in the Limpopo show promising results, and thus it is recommended to employ them as complementary to existing forecasts in order to strengthen preparedness for droughts.
The Limpopo Basin in southern Africa is prone to droughts which affect the livelihood of millions of people in South Africa, Botswana, Zimbabwe and Mozambique. Seasonal drought early warning is thus vital for the whole region. In this study, the predictability of hydrological droughts during the main runoff period from December to May is assessed using statistical approaches. Three methods (multiple linear models, artificial neural networks, random forest regression trees) are compared in terms of their ability to forecast streamflow with up to 12 months of lead time. The following four main findings result from the study.
1. There are stations in the basin at which standardised streamflow is predictable with lead times up to 12 months. The results show high inter-station differences of forecast skill but reach a coefficient of determination as high as 0.73 (cross validated).
2. A large range of potential predictors is considered in this study, comprising well-established climate indices, customised teleconnection indices derived from sea surface temperatures and antecedent streamflow as a proxy of catchment conditions. El Nino and customised indices, representing sea surface temperature in the Atlantic and Indian oceans, prove to be important teleconnection predictors for the region. Antecedent streamflow is a strong predictor in small catchments (with median 42% explained variance), whereas teleconnections exert a stronger influence in large catchments.
3. Multiple linear models show the best forecast skill in this study and the greatest robustness compared to artificial neural networks and random forest regression trees, despite their capabilities to represent nonlinear relationships.
4. Employed in early warning, the models can be used to forecast a specific drought level. Even if the coefficient of determination is low, the forecast models have a skill better than a climatological forecast, which is shown by analysis of receiver operating characteristics (ROCs). Seasonal statistical forecasts in the Limpopo show promising results, and thus it is recommended to employ them as complementary to existing forecasts in order to strengthen preparedness for droughts.
We investigate the usefulness of complex flood damage models for predicting relative damage to residential buildings in a spatial and temporal transfer context. We apply eight different flood damage models to predict relative building damage for five historic flood events in two different regions of Germany. Model complexity is measured in terms of the number of explanatory variables which varies from 1 variable up to 10 variables which are singled out from 28 candidate variables. Model validation is based on empirical damage data, whereas observation uncertainty is taken into consideration. The comparison of model predictive performance shows that additional explanatory variables besides the water depth improve the predictive capability in a spatial and temporal transfer context, i.e., when the models are transferred to different regions and different flood events. Concerning the trade-off between predictive capability and reliability the model structure seem more important than the number of explanatory variables. Among the models considered, the reliability of Bayesian network-based predictions in space-time transfer is larger than for the remaining models, and the uncertainties associated with damage predictions are reflected more completely.
Private precaution is an important component in contemporary flood risk management and climate adaptation. However, quantitative knowledge about vulnerability reduction via private precautionary measures is scarce and their effects are hardly considered in loss modeling and risk assessments. However, this is a prerequisite to enable temporally dynamic flood damage and risk modeling, and thus the evaluation of risk management and adaptation strategies. To quantify the average reduction in vulnerability of residential buildings via private precaution empirical vulnerability data (n = 948) is used. Households with and without precautionary measures undertaken before the flood event are classified into treatment and nontreatment groups and matched. Postmatching regression is used to quantify the treatment effect. Additionally, we test state-of-the-art flood loss models regarding their capability to capture this difference in vulnerability. The estimated average treatment effect of implementing private precaution is between 11 and 15 thousand EUR per household, confirming the significant effectiveness of private precautionary measures in reducing flood vulnerability. From all tested flood loss models, the expert Bayesian network-based model BN-FLEMOps and the rule-based loss model FLEMOps perform best in capturing the difference in vulnerability due to private precaution. Thus, the use of such loss models is suggested for flood risk assessments to effectively support evaluations and decision making for adaptable flood risk management.
Hierarchical Bayesian Approach for Modeling Spatiotemporal Variability in Flood Damage Processes
(2019)
Flood damage processes are complex and vary between events and regions. State-of-the-art flood loss models are often developed on the basis of empirical damage data from specific case studies and do not perform well when spatially and temporally transferred. This is due to the fact that such localized models often cover only a small set of possible damage processes from one event and a region. On the other hand, a single generalized model covering multiple events and different regions ignores the variability in damage processes across regions and events due to variables that are not explicitly accounted for individual households. We implement a hierarchical Bayesian approach to parameterize widely used depth-damage functions resulting in a hierarchical (multilevel) Bayesian model (HBM) for flood loss estimation that accounts for spatiotemporal heterogeneity in damage processes. We test and prove the hypothesis that, in transfer scenarios, HBMs are superior compared to generalized and localized regression models. In order to improve loss predictions for regions and events for which no empirical damage data are available, we use variables pertaining to specific region- and event-characteristics representing commonly available expert knowledge as group-level predictors within the HBM.
Large-scale flood risk assessments are crucial for decision making, especially with respect to new flood defense schemes, adaptation planning and estimating insurance premiums. We apply the process-based Regional Flood Model (RFM) to simulate a 5000-year flood event catalog for all major catchments in Germany and derive risk curves based on the losses per economic sector. The RFM uses a continuous process simulation including a multisite, multivariate weather generator, a hydrological model considering heterogeneous catchment processes, a coupled 1D-2D hydrodynamic model considering dike overtopping and hinterland storage, spatially explicit sector-wise exposure data and empirical multi-variable loss models calibrated for Germany. For all components, uncertainties in the data and models are estimated. We estimate the median Expected Annual Damage (EAD) and Value at Risk at 99.5% confidence for Germany to be euro0.529 bn and euro8.865 bn, respectively. The commercial sector dominates by making about 60% of the total risk, followed by the residential sector. The agriculture sector gets affected by small return period floods and only contributes to less than 3% to the total risk. The overall EAD is comparable to other large-scale estimates. However, the estimation of losses for specific return periods is substantially improved. The spatial consistency of the risk estimates avoids the large overestimation of losses for rare events that is common in other large-scale assessments with homogeneous return periods. Thus, the process-based, spatially consistent flood risk assessment by RFM is an important step forward and will serve as a benchmark for future German-wide flood risk assessments.
Pluvial flood risk is mostly excluded in urban flood risk assessment. However, the risk of pluvial flooding is a growing challenge with a projected increase of extreme rainstorms compounding with an ongoing global urbanization. Considered as a flood type with minimal impacts when rainfall rates exceed the capacity of urban drainage systems, the aftermath of rainfall-triggered flooding during Hurricane Harvey and other events show the urgent need to assess the risk of pluvial flooding. Due to the local extent and small-scale variations, the quantification of pluvial flood risk requires risk assessments on high spatial resolutions. While flood hazard and exposure information is becoming increasingly accurate, the estimation of losses is still a poorly understood component of pluvial flood risk quantification. We use a new probabilistic multivariable modeling approach to estimate pluvial flood losses of individual buildings, explicitly accounting for the associated uncertainties. Except for the water depth as the common most important predictor, we identified the drivers for having loss or not and for the degree of loss to be different. Applying this approach to estimate and validate building structure losses during Hurricane Harvey using a property level data set, we find that the reliability and dispersion of predictive loss distributions vary widely depending on the model and aggregation level of property level loss estimates. Our results show that the use of multivariable zero-inflated beta models reduce the 90% prediction intervalsfor Hurricane Harvey building structure loss estimates on average by 78% (totalling U.S.$3.8 billion) compared to commonly used models.
Residential assets, comprising buildings and household contents, are a major source of direct flood losses. Existing damage models are mostly deterministic and limited to particular countries or flood types. Here, we compile building-level losses from Germany, Italy and the Netherlands covering a wide range of fluvial and pluvial flood events. Utilizing a Bayesian network (BN) for continuous variables, we find that relative losses (i.e. loss relative to exposure) to building structure and its contents could be estimated with five variables: water depth, flow velocity, event return period, building usable floor space area and regional disposable income per capita. The model's ability to predict flood losses is validated for the 11 flood events contained in the sample. Predictions for the German and Italian fluvial floods were better than for pluvial floods or the 1993 Meuse river flood. Further, a case study of a 2010 coastal flood in France is used to test the BN model's performance for a type of flood not included in the survey dataset. Overall, the BN model achieved better results than any of 10 alternative damage models for reproducing average losses for the 2010 flood. An additional case study of a 2013 fluvial flood has also shown good performance of the model. The study shows that data from many flood events can be combined to derive most important factors driving flood losses across regions and time, and that resulting damage models could be applied in an open data framework.
Flood events can be expressed by a variety of characteristics such as flood magnitude and extent, event duration or incurred loss. Flood estimation and management may benefit from understanding how the different flood characteristics relate to the hydrological catchment conditions preceding the event and to the meteorological conditions throughout the event. In this study, we therefore propose a methodology to investigate the hydro-meteorological controls on different flood characteristics, based on the simulation of the complete flood risk chain from the flood triggering precipitation event, through runoff generation in the catchment, flood routing and possible inundation in the river system and floodplains to flood loss. Conditional cumulative distribution functions and regression tree analysis delineate the seasonal varying flood processes and indicate that the effect of the hydrological pre-conditions, i.e. soil moisture patterns, and of the meteorological conditions, i.e. weather patterns, depends on the considered flood characteristic. The methodology is exemplified for the Elbe catchment. In this catchment, the length of the build-up period, the event duration and the number of gauges undergoing at least a 10-year flood are governed by weather patterns. The affected length and the number of gauges undergoing at least a 2-year flood are however governed by soil moisture patterns. In case of flood severity and loss, the controlling factor is less pronounced. Severity is slightly governed by soil moisture patterns whereas loss is slightly governed by weather patterns. The study highlights that flood magnitude and extent arise from different flood generation processes and concludes that soil moisture patterns as well as weather patterns are not only beneficial to inform on possible flood occurrence but also on the involved flood processes and resulting flood characteristics.
The hydrological load causing flood hazard is in many instances not only determined by peak discharge, but is a multidimensional problem. While the methodology for multivariate frequency analysis is well established, the estimation of the associated uncertainty is rarely studied. In this paper, a method is developed to quantify the different sources of uncertainty for a bivariate flood frequency analysis. The method is exemplarily developed for the Mekong Delta (MD), one of the largest and most densely populated river deltas worldwide. Floods in the MD are the basis for the livelihoods of the local population, but they are also the major hazard. This hazard has, however, not been studied within the frame of a probabilistic flood hazard analysis. The nature of the floods in the MD suggests a bivariate approach, because the societal flood severity is determined by both peak discharge and flood volume. The uncertainty caused by selection of statistical models and parameter estimation procedures are analyzed by applying different models and methods. For the quantification of the sampling uncertainty two bootstrapping methods were applied. The developed bootstrapping-based uncertainty estimation method shows that large uncertainties are associated with the estimation of bivariate flood quantiles. This uncertainty is much larger than the model selection and fitting uncertainty. Given the rather long data series of 88 years, it is concluded that bivariate flood frequency analysis is expected to carry significant uncertainty and that the quantification and reduction of uncertainty merit greater attention. But despite this uncertainty the proposed approach has certainly major advantages compared to a univariate approach, because (a) it reflects the two essential aspects of floods in this region, (b) the uncertainties are inherent for every bivariate frequency analysis in hydrology due to the general limited length of observations and can hardly be avoided, and (c) a framework for the quantification of the uncertainties is given, which can be used and interpreted in the hazard assessment. In addition it is shown by a parametric bootstrapping experiment how longer observation time series can reduce the sampling uncertainty. Based on this finding it is concluded that bivariate frequency analyses in hydrology would greatly benefit from discharge time series augmented by proxy or historical data, or by causal hydrologic expansion of time series. (C) 2015 Elsevier B.V. All rights reserved.
Sedimentation in the floodplains of the Mekong Delta, Vietnam. Part I: suspended sediment dynamics
(2014)
Suspended sediment is the primary source for a sustainable agro-ecosystem in the Mekong Delta by providing nutrient input for the subsequent cropping season. In addition, the suspended sediment concentration (SSC) plays an important role in the erosion and deposition processes in the Delta; that is, it influences the morphologic development and may counteract the deltaic subsidence and sea level rise. Despite this importance, little is known about the dynamics of suspended sediment in the floodplains of the Mekong Delta. In particular, quantitative analyses are lacking mainly because of data scarcity with respect to the inundation processes in the floodplains. In 2008, therefore, a comprehensive in situ system to monitor the dynamics of suspended sediment in a study area located in the Plain of Reeds was established, aiming at the characterization and quantification of suspended sediment dynamics in the deeply inundated parts of the Vietnamese part of the Mekong Delta. The monitoring system was equipped with seven water quality-monitoring stations. They have a robust design and autonomous power supply suitable for operation on inundated floodplains, enabling the collection of reliable data over a long period of time with a high temporal resolution. The data analysis shows that the general seasonal dynamics of suspended sediment transport in the Delta is controlled by two main mechanisms: the flood wave of the Mekong River and the tidal backwater influences from the coast. In the channel network, SSC decreases exponentially with distance from the Mekong River. The anthropogenic influence on SSC could also be identified for two periods: at the start of the floodplain inundation and at the end of the flood period, when subsequent paddy rice crops are prepared. Based on the results, we recommend an operation scheme for the sluice gates, which intends to distribute the sediment and thus the nutrients equally over the floodplain.
Sedimentation in the floodplains of the Mekong Delta, Vietnam Part II: deposition and erosion
(2014)
Deposition and erosion play a key role in the determination of the sediment budget of a river basin, as well as for floodplain sedimentation. Floodplain sedimentation, in turn, is a relevant factor for the design of flood protection measures, productivity of agro-ecosystems, and for ecological rehabilitation plans. In the Mekong Delta, erosion and deposition are important factors for geomorphological processes like the compensation of deltaic subsidence as well as for agricultural productivity. Floodplain deposition is also counteracting the increasing climate change induced hazard by sea level rise in the delta. Despite this importance, a sediment database of the Mekong Delta is lacking, and the knowledge about erosion and deposition processes is limited. In the Vietnamese part of the Delta, the annually flooded natural floodplains have been replaced by a dense system of channels, dikes, paddy fields, and aquaculture ponds, resulting in floodplain compartments protected by ring dikes. The agricultural productivity depends on the sediment and associated nutrient input to the floodplains by the annual floods. However, no quantitative information regarding their sediment trapping efficiency has been reported yet. The present study investigates deposition and erosion based on intensive field measurements in three consecutive years (2008, 2009, and 2010). Optical backscatter sensors are used in combination with sediment traps for interpreting deposition and erosion processes in different locations. In our study area, the mean calculated deposition rate is 6.86kg/m(2) (approximate to 6mm/year). The key parameters for calculating erosion and deposition are estimated, i.e. the critical bed shear stress for deposition and erosion and the surface constant erosion rate. The bulk of the floodplain sediment deposition is found to occur during the initial stage of floodplain inundation. This finding has direct implications on the operation of sluice gates in order to optimize sediment input and distribution in the floodplains.
Groundwater transit time is an essential hydrologic metric for groundwater resources management. However, especially in tropical environments, studies on the transit time distribution (TTD) of groundwater infiltration and its corresponding mean transit time (mTT) have been extremely limited due to data sparsity. In this study, we primarily use stable isotopes to examine the TTDs and their mTTs of both vertical and horizontal infiltration at a riverbank infiltration area in the Vietnamese Mekong Delta (VMD), representative of the tropical climate in Asian monsoon regions. Precipitation, river water, groundwater, and local ponding surface water were sampled for 3 to 9 years and analysed for stable isotopes (delta O-18 and delta H-2), providing a unique data set of stable isotope records for a tropical region. We quantified the contribution that the two sources contributed to the local shallow groundwater by a novel concept of two-component lumped parameter models (LPMs) that are solved using delta O-18 records. The study illustrates that two-component LPMs, in conjunction with hydrological and isotopic measurements, are able to identify subsurface flow conditions and water mixing at riverbank infiltration systems. However, the predictive skill and the reliability of the models decrease for locations farther from the river, where recharge by precipitation dominates, and a low-permeable aquitard layer above the highly permeable aquifer is present. This specific setting impairs the identifiability of model parameters. For river infiltration, short mTTs (<40 weeks) were determined for sites closer to the river (<200 m), whereas for the precipitation infiltration, the mTTs were longer (>80 weeks) and independent of the distance to the river. The results not only enhance the understanding of the groundwater recharge dynamics in the VMD but also suggest that the highly complex mechanisms of surface-groundwater interaction can be conceptualized by exploiting two-component LPMs in general. The model concept could thus be a powerful tool for better understanding both the hydrological functioning of mixing processes and the movement of different water components in riverbank infiltration systems.
For attributing hydrological changes to anthropogenic climate change, catchment models are driven by climate model output. A widespread approach to bridge the spatial gap between global climate and hydrological catchment models is to use a weather generator conditioned on weather patterns (WPs). This approach assumes that changes in local climate are characterized by between-type changes of patterns. In this study we test this assumption by analyzing a previously developed WP classification for the Rhine basin, which is based on dynamic and thermodynamic variables. We quantify changes in pattern characteristics and associated climatic properties. The amount of between- and within-type changes is investigated by comparing observed trends to trends resulting solely from WP occurrence. To overcome uncertainties in trend detection resulting from the selected time period, all possible periods in 1901-2010 with a minimum length of 31 years are analyzed. Increasing frequency is found for some patterns associated with high precipitation, although the trend sign highly depends on the considered period. Trends and interannual variations of WP frequencies are related to the long-term variability of large-scale circulation modes. Long-term WP internal warming is evident for summer patterns and enhanced warming for spring/autumn patterns since the 1970s. Observed trends in temperature and partly in precipitation are mainly associated with frequency changes of specific WPs, but some amount of within-type changes remains. The classification can be used for downscaling of past changes considering this limitation, but the inclusion of thermodynamic variables into the classification impedes the downscaling of future climate projections.
To understand past flood changes in the Rhine catchment and in particular the role of anthropogenic climate change in extreme flows, an attribution study relying on a proper GCM (general circulation model) downscaling is needed. A downscaling based on conditioning a stochastic weather generator on weather patterns is a promising approach. This approach assumes a strong link between weather patterns and local climate, and sufficient GCM skill in reproducing weather pattern climatology. These presuppositions are unprecedentedly evaluated here using 111 years of daily climate data from 490 stations in the Rhine basin and comprehensively testing the number of classification parameters and GCM weather pattern characteristics. A classification based on a combination of mean sea level pressure, temperature, and humidity from the ERA20C reanalysis of atmospheric fields over central Europe with 40 weather types was found to be the most appropriate for stratifying six local climate variables. The corresponding skill is quite diverse though, ranging from good for radiation to poor for precipitation. Especially for the latter it was apparent that pressure fields alone cannot sufficiently stratify local variability. To test the skill of the latest generation of GCMs from the CMIP5 ensemble in reproducing the frequency, seasonality, and persistence of the derived weather patterns, output from 15 GCMs is evaluated. Most GCMs are able to capture these characteristics well, but some models showed consistent deviations in all three evaluation criteria and should be excluded from further attribution analysis.
To understand past flood changes in the Rhine catchment and in particular the role of anthropogenic climate change in extreme flows, an attribution study relying on a proper GCM (general circulation model) downscaling is needed. A downscaling based on conditioning a stochastic weather generator on weather patterns is a promising approach. This approach assumes a strong link between weather patterns and local climate, and sufficient GCM skill in reproducing weather pattern climatology. These presuppositions are unprecedentedly evaluated here using 111 years of daily climate data from 490 stations in the Rhine basin and comprehensively testing the number of classification parameters and GCM weather pattern characteristics. A classification based on a combination of mean sea level pressure, temperature, and humidity from the ERA20C reanalysis of atmospheric fields over central Europe with 40 weather types was found to be the most appropriate for stratifying six local climate variables. The corresponding skill is quite diverse though, ranging from good for radiation to poor for precipitation. Especially for the latter it was apparent that pressure fields alone cannot sufficiently stratify local variability. To test the skill of the latest generation of GCMs from the CMIP5 ensemble in reproducing the frequency, seasonality, and persistence of the derived weather patterns, output from 15 GCMs is evaluated. Most GCMs are able to capture these characteristics well, but some models showed consistent deviations in all three evaluation criteria and should be excluded from further attribution analysis.
Flood risk is impacted by a range of physical and socio-economic processes. Hence, the quantification of flood risk ideally considers the complete flood risk chain, from atmospheric processes through catchment and river system processes to damage mechanisms in the affected areas. Although it is generally accepted that a multitude of changes along the risk chain can occur and impact flood risk, there is a lack of knowledge of how and to what extent changes in influencing factors propagate through the chain and finally affect flood risk. To fill this gap, we present a comprehensive sensitivity analysis which considers changes in all risk components, i.e. changes in climate, catchment, river system, land use, assets, and vulnerability. The application of this framework to the mesoscale Mulde catchment in Germany shows that flood risk can vary dramatically as a consequence of plausible change scenarios. It further reveals that components that have not received much attention, such as changes in dike systems or in vulnerability, may outweigh changes in often investigated components, such as climate. Although the specific results are conditional on the case study area and the selected assumptions, they emphasize the need for a broader consideration of potential drivers of change in a comprehensive way. Hence, our approach contributes to a better understanding of how the different risk components influence the overall flood risk.
Flood risk assessments are typically based on scenarios which assume homogeneous return periods of flood peaks throughout the catchment. This assumption is unrealistic for real flood events and may bias risk estimates for specific return periods. We investigate how three assumptions about the spatial dependence affect risk estimates: (i) spatially homogeneous scenarios (complete dependence), (ii) spatially heterogeneous scenarios (modelled dependence) and (iii) spatially heterogeneous but uncorrelated scenarios (complete independence). To this end, the model chain RFM (regional flood model) is applied to the Elbe catchment in Germany, accounting for the spatio-temporal dynamics of all flood generation processes, from the rainfall through catchment and river system processes to damage mechanisms. Different assumptions about the spatial dependence do not influence the expected annual damage (EAD); however, they bias the risk curve, i.e. the cumulative distribution function of damage. The widespread assumption of complete dependence strongly overestimates flood damage of the order of 100% for return periods larger than approximately 200 years. On the other hand, for small and medium floods with return periods smaller than approximately 50 years, damage is underestimated. The overestimation aggravates when risk is estimated for larger areas. This study demonstrates the importance of representing the spatial dependence of flood peaks and damage for risk assessments.
Flood risk is impacted by a range of physical and socio-economic processes. Hence, the quantification of flood risk ideally considers the complete flood risk chain, from atmospheric processes through catchment and river system processes to damage mechanisms in the affected areas. Although it is generally accepted that a multitude of changes along the risk chain can occur and impact flood risk, there is a lack of knowledge of how and to what extent changes in influencing factors propagate through the chain and finally affect flood risk. To fill this gap, we present a comprehensive sensitivity analysis which considers changes in all risk components, i.e. changes in climate, catchment, river system, land use, assets, and vulnerability. The application of this framework to the mesoscale Mulde catchment in Germany shows that flood risk can vary dramatically as a consequence of plausible change scenarios. It further reveals that components that have not received much attention, such as changes in dike systems or in vulnerability, may outweigh changes in often investigated components, such as climate. Although the specific results are conditional on the case study area and the selected assumptions, they emphasize the need for a broader consideration of potential drivers of change in a comprehensive way. Hence, our approach contributes to a better understanding of how the different risk components influence the overall flood risk.