Refine
Has Fulltext
- no (21)
Document Type
- Article (21) (remove)
Is part of the Bibliography
- yes (21) (remove)
Keywords
- Climate variability (2)
- trend analysis (2)
- Ahr (1)
- Ahr River (1)
- Asia (1)
- Catchment wetness (1)
- Climate prediction (1)
- Extreme value statistics (1)
- Extremwertstatistik (1)
- Flood regimes (1)
Institute
Stochastic modeling of precipitation for estimation of hydrological extremes is an important element of flood risk assessment and management. The spatially consistent estimation of rainfall fields and their temporal variability remains challenging and is addressed by various stochastic weather generators.
In this study, two types of weather generators are evaluated against observed data and benchmarked regarding their ability to simulate spatio-temporal precipitation fields in the Rhine catchment. A multi-site station-based weather generator uses an auto-regressive model and estimates the spatial correlation structure between stations. Another weather generator is raster-based and uses the nearest-neighbor resampling technique for reshuffling daily patterns while preserving the correlation structure between the observations.
Both weather generators perform well and are comparable at the point (station) scale with regards to daily mean and 99.9th percentile precipitation as well as concerning wet/dry frequencies and transition probabilities. The areal extreme precipitation at the sub-basin scale is however overestimated in the station-based weather generator due to an overestimation of the correlation structure between individual stations. The auto-regressive model tends to generate larger rainfall fields in space for extreme precipitation than observed, particularly in summer. The weather generator based on nearest-neighbor resampling reproduces the observed daily and multiday (5, 10 and 20) extreme events in a similar magnitude. Improvements in performance regarding wet frequencies and transition probabilities are recommended for both models.
Hydrodynamic interactions, i.e. the floodplain storage effects caused by inundations upstream on flood wave propagation, inundation areas, and flood damage downstream, are important but often ignored in large-scale flood risk assessments. Although new methods considering these effects sometimes emerge, they are often limited to a small or meso scale. In this study, we investigate the role of hydrodynamic interactions and floodplain storage on flood hazard and risk in the German part of the Rhine basin. To do so, we compare a new continuous 1D routing scheme within a flood risk model chain to the piece-wise routing scheme, which largely neglects floodplain storage. The results show that floodplain storage is significant, lowers water levels and discharges, and reduces risks by over 50%. Therefore, for accurate risk assessments, a system approach must be adopted, and floodplain storage and hydrodynamic interactions must carefully be considered.
Different upper tail indicators exist to characterize heavy tail phenomena, but no comparative study has been carried out so far. We evaluate the shape parameter (GEV), obesity index, Gini index and upper tail ratio (UTR) against a novel benchmark of tail heaviness - the surprise factor. Sensitivity analyses to sample size and changes in scale-to-location ratio are carried out in bootstrap experiments. The UTR replicates the surprise factor best but is most uncertain and only comparable between records of similar length. For samples with symmetric Lorenz curves, shape parameter, obesity and Gini indices provide consistent indications. For asymmetric Lorenz curves, however, the first two tend to overestimate, whereas Gini index tends to underestimate tail heaviness. We suggest the use of a combination of shape parameter, obesity and Gini index to characterize tail heaviness. These indicators should be supported with calculation of the Lorenz asymmetry coefficients and interpreted with caution.
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.
Die Hochwasserkatastrophe im Juli 2021 in Westdeutschland erfordert eine kritische Diskussion über die Abschätzung der Hochwassergefährdung, Aktualisierung von Hochwassergefahrenkarten und Kommunikation von extremen Hochwasserszenarien. In der vorliegenden Arbeit wurde die Extremwertstatistik für die jährlichen maximalen Spitzenabflüsse am Pegel Altenahr im Ahrtal mit und ohne Berücksichtigung historischer Hochwasser berechnet und verglichen. Die Schätzung der Wiederkehrperiode für das aktuelle Hochwasser mittels Generalisierter Extremwertverteilung (GEV) unter Berücksichtigung historischer Hochwasser schwankt zwischen etwa 2.600 und über 58.700 Jahren (90%-Konfidenzintervall) mit einem Median bei etwa 8.600 Jahren, wogegen die Schätzung, die nur auf der systematisch gemessenen Abflusszeitreihe von 74 Jahren basiert, theoretisch eine Wiederkehrperiode von über 100 Millionen Jahren ergeben würde. Die Berücksichtigung der historischen Hochwasser führt zu einer dramatischen Änderung der Hochwasserquan-
tile, die für eine Gefahrenkartierung zugrunde gelegt werden. Die Anpassung der GEV an die Zeitreihe mit historischen Hochwassern zeigt dennoch, dass das GEV-Modell möglicherweise die Grundgesamtheit der Hochwasser im Ahrtal nicht adäquat abbilden kann. Es könnte sich im vorliegenden Fall um eine gemischte Stichprobe handeln, in der die extremen Hochwasser im Vergleich zu kleineren Ereignissen durch besondere Prozesse hervorgerufen werden. Somit könnten die Wahrscheinlichkeiten von extremen Hochwassern deutlich größer sein, als aus dem GEV-Modell hervorgeht. Hier sollte in Zukunft die Anwendung einer prozessbasierten Mischverteilung
untersucht werden. Der Vergleich von amtlichen Gefahrenkarten zu Extremhochwassern (HQextrem) im Ahrtal mit den Überflutungsflächen vom Juli 2021
zeigt eine deutliche Diskrepanz in den betroffenen Gebieten und die Notwendigkeit, die Grundlagen zur Erstellung der Extremszenarien zu überdenken. Die hydrodynamisch-numerischen Simulationen von 1.000-jährlichen Hochwassern (HQ1000) unter Berücksichtigung historischer Ereignisse und des größten historischen Hochwassers 1804 können die Gefährdung des Juli-Hochwassers 2021 deutlich besser widerspiegeln, wenngleich auch diese beiden Szenarien die Überflutungsflächen unterschätzen. Besondere Effekte wie die Verklausung von Brücken und die geomorphologischen Änderungen im Flussschlauch führten zu noch größeren Überflutungs- flächen im Juli 2021, als die Simulationsergebnisse zeigten. Basierend auf dieser Analyse wird eine einheitliche Festlegung von HQextrem bei Hochwassergefahrenkartierungen in Deutschland vorgeschlagen, die sich an höheren Hochwasserquantilen im Bereich von HQ1000 orientiert. Zusätzlich sollen simulationsbasierte Rekonstruktionen von den größten verlässlich dokumentierten historischen Hochwassern und/oder synthetische Worst-Case-Szenarien in den Hochwassergefahrenkarten gesondert dargestellt werden. Damit wird ein wichtiger Beitrag geleistet, um die potenziell betroffene Bevölkerung und das Katastrophenmanagement vor Überraschungen durch sehr seltene und extreme Hochwasser in Zukunft besser zu schützen.
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.
From precipitation to damage
(2018)
Flood risk assessments for large river basins often involve piecing together smaller-scale assessments leading to erroneous risk statements. We describe a coupled model chain for quantifying flood risk at the scale of 100,000 km(2). It consists of a catchment model, a 1D-2D river network model, and a loss model. We introduce the model chain and present two applications. The first application for the Elbe River basin with an area of 66,000 km(2) demonstrates that it is feasible to simulate the complete risk chain for large river basins in a continuous simulation mode with high temporal and spatial resolution. In the second application, RFM is coupled to a multisite weather generator and applied to the Mulde catchment with an area of 6,000 km(2). This approach is able to provide a very long time series of spatially heterogeneous patterns of precipitation, discharge, inundation, and damage. These patterns respect the spatial correlation of the different processes and are suitable to derive large-scale risk estimates. We discuss how the RFM approach can be transferred to the continental scale.
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.
This study refines the method for calibrating a glacio-hydrological model based on Hydrograph Partitioning Curves (HPCs), and evaluates its value in comparison to multidata set optimization approaches which use glacier mass balance, satellite snow cover images, and discharge. The HPCs are extracted from the observed flow hydrograph using catchment precipitation and temperature gradients. They indicate the periods when the various runoff processes, such as glacier melt or snow melt, dominate the basin hydrograph. The annual cumulative curve of the difference between average daily temperature and melt threshold temperature over the basin, as well as the annual cumulative curve of average daily snowfall on the glacierized areas are used to identify the starting and end dates of snow and glacier ablation periods. Model parameters characterizing different runoff processes are calibrated on different HPCs in a stepwise and iterative way. Results show that the HPC-based method (1) delivers model-internal consistency comparably to the tri-data set calibration method; (2) improves the stability of calibrated parameter values across various calibration periods; and (3) estimates the contributions of runoff components similarly to the tri-data set calibration method. Our findings indicate the potential of the HPC-based approach as an alternative for hydrological model calibration in glacierized basins where other calibration data sets than discharge are often not available or very costly to obtain.
Despite its societal relevance, the question whether fluctuations in flood occurrence or magnitude are coherent in space has hardly been addressed in quantitative terms. We investigate this question for Germany by analysing fluctuations in annual maximum series (AMS) values at 68 discharge gauges for the common time period 1932-2005. We find remarkable spatial coherence across Germany given its different flood regimes. For example, there is a tendency that flood-rich/-poor years in sub-catchments of the Rhine basin, which are dominated by winter floods, coincide with flood-rich/-poor years in the southern sub-catchments of the Danube basin, which have their dominant flood season in summer. Our findings indicate that coherence is caused rather by persistence in catchment wetness than by persistent periods of higher/lower event precipitation. Further, we propose to differentiate between event-type and non-event-type coherence. There are quite a number of hydrological years with considerable nonevent-type coherence, i.e. AMS values of the 68 gauges are spread out through the year but in the same magnitude range. Years with extreme flooding tend to be of event-type and non-coherent, i.e. there is at least one precipitation event that affects many catchments to various degree. Although spatial coherence is a remarkable phenomenon, and large-scale flooding across Germany can lead to severe situations, extreme magnitudes across the whole country within one event or within one year were not observed in the investigated period. (C) 2018 Elsevier B.V. All rights reserved.