Displays

Given the steady increase of economical losses and social consequences caused by extreme flood events over the last decades in many European countries, the scientific community is making an effort to better understand recent flood dynamics and their evolution in space and time.

In this context, our study considers a large dataset of annual maximum series of peak flow discharges for more than 3400 catchments across Europe. The dataset covers the period 1820-2016, with an average record length of 53 years. On the basis of this extensive dataset, our study focuses on an issue which has never been addressed at European scale: we analyse the behaviour of the specific flood of record (i.e. the largest flood observed in the time interval of interest divided by the drainage area of the corresponding catchment, hereafter also referred to as SFOR) in space and time across the European continent. In particular, we consider the spatial variability of SFOR computed for the entire observation period, and for two additional sub-samples, including observations collected in the last three decades (i.e. 1987-2016) and in the three previous ones (i.e. 1957-1986), respectively. For the selected different timespans, we then analyse the spatial variability of the year in which SFOR was observed, and the number of times in which a new record was observed at each and every gauge, also evaluating their relationship with catchment area and outlet elevation.

We also provide a continuous spatial representation of SFOR values by interpolating them at elementary catchments identified by the Joint Research Centre (JRC) of the European Commission. In particular, for each elementary catchment included in the JRC dataset, we interpolate empirical SFOR values through two different procedures: (1) a geostatistical procedure (i.e. top-kriging), and (2) a linear regression with drainage area on the basis of the SFOR values observed at the closest catchments. Both the interpolation procedures account for nesting between catchments and are applied so as to ensure for the interpolated SFOR values a monotonic decrease from upstream to downstream.

The analysis of the maps produced in our study provides useful information on the spatio-temporal evolution of flooding potential across Europe, enabling a visualization of significant changes and shifts of the flood of record occurred during the last decades. In particular, we observe that: (1) years of occurrence of SFOR values are mainly concentrated in the last thirty years (i.e. 1987-2016), especially in the area of Central Europe; (2) smaller catchments show higher sensitivity to changes in flood dynamics.

How to cite: Persiano, S., Arletti, F., Bertola, M., Parajka, J., Blöschl, G., and Castellarin, A.: Spatio-temporal mapping of floods of record across Europe, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-7753, https://doi.org/10.5194/egusphere-egu2020-7753, 2020.

Flood risk is the combination of the probability of a flood event and the potential adverse consequences for the human health, the environment, the cultural heritage and the economic activity associated with a flood event (2007/60/EC). Usually, the potential consequence of a flood with a certain probability is evaluated through the flood stage raised in the analysed area, combined with the variables characterising the type of vulnerability. Indeed, flood damages are evaluated from stage-damage curves, also called damage functions. Recently, flood social vulnerability indexes are defined taking into account the flood stage as the primary hydraulic variable. The standard approach evaluates the flood stage starting by a univariate hydrological load, corresponding to one hydrograph with a peak discharge of a certain probability and adequate durations. This correspondence is a critical issue that is originated from the approximation of the river flood flow process. A bivariate hydrological methodology for improving flood maps is proposed. A consistent number of synthetic hydrographs composes the bivariate hydrological load, with peak discharge and volume belonging to their bivariate distribution. A flood map corresponds to each hydrograph. Each flood map is a grid developing through 2d hydraulic model. A specific flood stage value corresponds to each cell of the grid. The whole set of hydrograph produces a flood stage series for each grid cell. The flood map with a certain probability, i.e. return period, results from the interpolation of the corresponding quantile values for each grid cell. The methodology is applied to a case study. The resulting map is benchmarked with a map obtained by the standard hydrological approach. The proposed methodology is based on tools that are widely known and it is replicable by the public administrations or public entities that are interesting in the hydrologic and hydraulic risk assessment.

How to cite: Caporali, E., Isola, M., and Garrote, L.: A bivariate hydrological methodology for improving flood maps, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-18593, https://doi.org/10.5194/egusphere-egu2020-18593, 2020.

Extreme flood events can occur due to manifold combinations of different generating factors. A differentiation into flood types helps to distinguish between the main runoff generating processes and the shape of the flood wave. However, the genesis of extreme flood events cannot always be explained by the flood type only. In a first step, flood peak and flood volume are classified to determine their extremity by a robust classification based on moments. Extreme cases of runoff generating processes like the amount of event precipitation, runoff coefficient and antecedent soil moisture are detected by their deviation from the population distribution. With this, we then analyse significant coherences between the drivers of extreme runoff generating processes and the extreme flood characteristics. It turns out, that the different flood types show very different coherences between these two factors. Moreover, many extreme peaks cannot be explained by either of these factors. Instead, the spatial and temporal distribution of precipitation plays the most important role, especially for floods caused by short and medium rain. In a second step, these two factors are included in the coherence analyses, where significant dependencies of the extremity of the flood peak on these are detected. The approach is applied to several basins in Germany and Austria, including alpine, mountainous and flatland catchments. For these, significant spatial differences in the coherences occur. In the alpine catchments e.g. the soil moisture has much more impact on the extremity of floods than for flatland catchments.

How to cite: Schumann, A., Fischer, S., and Bühler, P.: Hydrological causes of extreme flood events and dependence on flood type, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-1525, https://doi.org/10.5194/egusphere-egu2020-1525, 2020.

Hydrologists are challenged to estimate extreme discharges from catchments with data of poor temporal and spatial resolution. Floods are complex processes derived from catchment responses to various meteorological inputs, commonly summarized under one distribution function, representing the cumulative effect of all triggering events (Merz & Blöschl, 2003). A better understanding of driving precipitation inputs, catchment properties and a-priori conditions are required to characterize flood mechanisms and to determine shape, volume and peak of the extreme discharges. This research focuses on the estimation of floods. The study area is the northwestern Switzerland with small to medium catchments (0.5 to 200 km2), with low concentration times and a highly variable response to the meteorological input in terms of associated peak discharges and volumes.

We use a random forest algorithm to evaluate similar catchment reactions at the occurrence of a flood. We consider catchment descriptors and event specific characteristics for the training of the model. The flood hydrograph serves as the training target variable in order to describe the catchment response. Our regionalization method suggest that the meteorological input of a catchment, specifically the temporal entropy of precipitation, is the most significant parameter for clustering catchment reactions and should, therefore, be consider for such a task. This model has the potential of identifying donor catchments for estimating extreme discharge at the ungauged catchments, using the floods similarities derived by the random forest.

References:

Merz, R., and G. Blöschl, A process typology of regional floods, Water Resour. Res., 39(12), 1340, doi:10.1029/2002WR001952, 2003.

How to cite: Pavia Santolamazza, D., Lebrenz, H., and Bárdossy, A.: Random forest algorithm as a regionalization model of flood-mechanisms, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-5102, https://doi.org/10.5194/egusphere-egu2020-5102, 2020.

Investigation of extreme events and their spatial extent is one of the crucial tasks in the hydrology. Most hydrological models are not able to accurately simulate the upper quantile of the discharge time series. In this research, a new method for determining the spatiotemporal similarity of extreme floods was developed. The maximum absolute difference among CDFs was combined with an agglomerative hierarchical cluster, and a new tree based on clustering distribution properties was done. Initially, the continuous discharge time series of 46 gauges in the Neckar catchment were examined. Then, the two most prominent events of each year were selected. Subsequently, the empirical cumulative distribution functions of each point, based on selected peaks, were calculated; and the probability of occurrence of each event was determined. The pairwise similarity of CDFs, and consequently, the absolute deviation between them were computed. Thereupon, the hierarchical cluster tree based on the matrix of maximum differences was performed by employing a distinct distance method. At the final step, the cluster tree divided the basin into three major clusters, which contain some sub-catchments. The results illustrated a non-particular pattern for flood occurrences in the Ward linkage map. However, the Average linkage in the clustering showed that the catchment has a more or less homogeneous behavior with some small independent parts. Each separate category revealed a different response concerning the highest flooding mechanism, which the hydrological modeling has to take into account.

How to cite: Modiri, E. and Bárdossy, A.: Spatio-temporal determination of the similarity of extreme floods in the Neckar catchment , EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-10008, https://doi.org/10.5194/egusphere-egu2020-10008, 2020.

The study of flood risks requires a better understanding of the hydrological response of catchments, by identifying the drivers responsible for their variability, such as seasonal and regional rainfall patterns, initial catchment conditions and geology. Many catchments are not conservative, mainly due to Interbasin Groundwater Flows (IGF), that limits the application of traditional water balances methods on a regional scale. The role of karst areas on IGF is highly suspected to promote particular hydrological processes that change the annual water budgets as well as flood event dynamics.

The aim of our work is to assess the impact of IGF in karst and non-karst catchments of medium size (100-500 km²), on annual water budgets and flood dynamics. To this end, we developed a two-step methodology, applied on 120 elementary catchment in France, for which daily rainfall and runoff time series of several decades were available.

First, the traditional annual water budget method of L’vovich was adapted to non-conservative catchments, including an explicit term of IGF, as well as hydrograph decomposition. Results show that IGF occurrence is linked to the presence of karst areas, and that it affects both flood and baseflow components, sometimes in a very significant way. Second, a flood event analysis was conducted using a hydrograph characterization, including the analysis of lateral losses and gains on reaches delimited by 2 stations. The variability of these parameters was then studied as a function of seasonal and regional rainfall patterns, initial catchment conditions, and geology. Results show that geology (with the presence of karst areas) affect all parameters (flood shape and lateral exchanges), while rainfall pattern and initial catchment conditions mainly influence the flood dynamics.

Globally, our results show that, in addition to classical drivers (rainfall & initial catchment conditions), the spatial variability of flood pattern and dynamics is highly influenced by geology and notably karst areas. This study brings ways to improve the efficiency of hydrological models, by including IGF as a specific process. Results are also interesting in terms of extension to ungauged basins, as IGF occurrence is linked to the occurrence of karst areas.

How to cite: Le Mesnil, M., Charlier, J.-B., Moussa, R., Caballero, Y., and Dörfliger, N.: Control of karst areas on space-time dynamics of floods, by combining annual and event-based analyses, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-19519, https://doi.org/10.5194/egusphere-egu2020-19519, 2020.

When rivers flood, surrounding rivers often flood at the same time. When large precipitation events occur, floods do not always occur. Here we explore the drivers and synchronicity of river flooding. Using flood data from thousands of European and US rivers, we demonstrate that the flood synchrony scale—the distance over which multiple rivers flood near synchronously—far exceeds the size of individual drainage basins and varies regionally by more than an order of magnitude. Regions of large flood synchrony scales are mostly uncorrelated with regions of large precipitation synchrony scales; across most of Europe and the US few floods are caused by the biggest rainfall peaks. Instead, most floods are caused by the concurrence of heavy precipitation with high antecedent soil moisture. Risk finance, flood forecasting, and interpretations of flood trends can benefit from accounting for what drives flooding and how flood risks extend beyond the borders of individual drainage basins.

How to cite: Berghuijs, W. and Kirchner, J.: Synchronicity and drivers of river flooding , EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-9625, https://doi.org/10.5194/egusphere-egu2020-9625, 2020.

The superposition of two or more flood events at confluences is an important factor in the genesis of extreme floods. With extreme floods in 2002 and 2013 in Germany, analyses of the genesis of those floods becomes a more vital part in the understanding of their mechanisms. While the superposition is not the main driver of extremity, it still influences the peak significantly, depending on the arrival of the peaks from upstream floods as well as the shape and steepness of the floods. Hereby, the degree of superposition depends much on the flood type: for steep and short waves, the probability of overlapping peaks is low but has a high chance to result in an extraordinary flood or the overlapping of long flood waves has a high probability with a smaller chance to produce a flood event with an extreme peak after the point of confluence. In order to quantify the effects of superposition for extreme events, confluences of tributaries in the Mulde river basin in the east of Germany were analysed based on hourly discharge data. For these events, the range of best- and worst case scenarios was analysed based on sensible shifts of the routing. The travel times and therefore the arrivals of flood events at the downstream gauge were evaluated from the data as well as with a theoretical approach calculated by the mean slope of the stream as a static and the peak discharge of a flood event as the dynamic component. It is shown how the different combinations of arrival times in the downstream gauge may result in long events with damped peaks or a maximization of the flood peaks by overlaying. With the developed methodology, the observed peak can be compared with an ensemble of possible flood events and their peaks, caused by different superposition scenarios. This leads to an extended range of high empirical quantiles for flood statistics, with impact on the selection of most appropriated distribution functions.

How to cite: Bühler, P., Fischer, S., and Schumann, A.: Maximisation of flood events by superposition at confluences, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-1552, https://doi.org/10.5194/egusphere-egu2020-1552, 2020.

In 2016 and 2018, severe convective rainfall events caused several flash floods in small mesoscale catchments in eastern Luxembourg. While the runoff coefficients of these events remained rather common, the high proportion of the generated flood peaks was very unusual - requiring further research into the dominating runoff generating processes (e.g. flow paths type, activation/connectivity, antecedent conditions).

Here we to intend to explore and quantify, based on a longitudinal monitoring of discharge, the contrasted hydrological responses of nested catchments having multiple geological and pedological discrepancies. Our study area is the elongated Ernz Blanche catchment (102 km², approximatively 22.5 km long, 4.5 km wide), located in eastern Luxembourg (Europe). This mesoscale catchment is representative of the physiographic diversity of the country. Its upstream part is almost equally split between marly terrain (middle Keuper) and Luxembourg sandstone outcrops. The downstream part of the catchment mainly consists of deeply cut Luxembourg sandstone, alternating with marly plateaus (Lias). We have installed in early 2019 six stream-gauges along the 27.5 km long Ernz Blanche River. In addition, we have dispatched four rain-gauges and soil moisture sensors across the catchment to measure precipitations and soil water content, respectively.

Our first year of observations shows a spatially homogeneous response of the catchment during the winter period, with the specific discharge values observed at the six stream-gauges being highly correlated. During the summer flood events, the hydrological responses between the upper and downstream parts of the catchment are clearly distinct. More specifically, the downstream part generates two-peaked flood hydrographs – the first peak consisting of a flashy and non-attenuated response to precipitation, while the second peak clearly relates to the total precipitation amount. Interestingly, we have observed this pattern even for moderate events – 12 mm of incident rainfall and 3.2 mm.15min^-1 of rainfall intensity being sufficient to produce a double-peak hydrograph. We conjecture that this dual hydrological response of the downstream part of the catchment is caused by either (i) the very dry antecedent weather conditions during the summer 2019 impacting soil hydraulic properties or (ii) the generally lower rainfall intensities observed in winter (< 1.8 mm.15min^-1), causing the initial flashy response to be either limited or totally insignificant.

Our preliminary conclusion on the hydrological behavior of the Ernz Blanche catchment (based on response times, runoff coefficients and hydrograph separations) suggests a clear distinction in hydrological response between the upstream luxembourg sandstone outcrops and marly terrain and the downstream marly plateaus and deeply cut sandstone valleys. Hydrological responses differ between the upstream part of the catchment, where homogeneous and damped flood responses prevail throughout the seasons, and the downstream part, where a threshold behaviour dominates (between summer and winter). These findings will contribute to improve the design of conceptual flow processing models. This is an important milestone and prerequisite for any subsequent development and transposition of a suitable flood forecasting model, adapted to the large physiographic diversity of Luxembourg.

How to cite: Douinot, A., Meisch, C., Bastian, C., Meyer, J., and Pfister, L.: Flood patterns in a catchment with mixed bedrock geology: characterization of diverging flood shape responses using longitudinal discharge monitoring, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-2320, https://doi.org/10.5194/egusphere-egu2020-2320, 2020.

Under the background of climate change and rapid urbanization，the risk of urban flood is increasing across the globe. To alleviate the urban flooding problems, the sponge city strategy has been proposed in China. The urban flood control system based on sponge city is gradually formed, which is an integrated system composed of green and grey infrastructure. However, mechanism of the corresponding flood control function and corresponding quantitative assessment of flood control capacity of the integrated green and grey infrastructure is relatively lacking. Based on polit sponge cities in China, this study summarized and put forward the construction mode of urban inundation control system of sponge city, including source control system, stormwater pipe network system, over-standard stormwater storage and drainage system, etc., identified the mechanism of urban flood control functions of urban flood control system, including detaining, releasing, peak rate cutting, peak rate delaying and discharging the stormwater runoff. Furthermore, a gauss-function based approach for quantitative flood control capacity assessment of integrated green and grey infrastructure was established. This study builds the relationship between the gauss function and mechanism of urban flood control capacity, according to the mathematical meaning of parameters of the gauss function. It provides a new method for urban flood control capacity assessment of the integrated green and grey infrastructure.

How to cite: Wang, J., Liu, J., Wang, H., and Mei, C.: A gauss-function based approach for flood control capacity assessment of integrated green and grey infrastructure, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-2352, https://doi.org/10.5194/egusphere-egu2020-2352, 2020.

For derived flood frequency analyses, stochastic rainfall models can be linked with rainfall-runoff models to improve the accuracy of design flood estimations when the length of observed rainfall and runoff data is not sufficient. In the past, when using stochastic rainfall time series for hydrological modelling purposes, catchment rainfall for use in hydrological modelling was calculated from the multiple point rainfall time series. As an alternative to this approach, it will be tested whether catchment rainfall can be modelled directly, negating the drawbacks (and need) encountered in generating spatially consistent time series. An Alternating Renewal rainfall model (ARM) will be used to generate multiple point and lumped catchment rainfall time series in hourly resolution. The generated rainfall time series will be used to drive the rainfall-runoff model HBV-IWW with an hourly time step for mesoscale catchments in Germany. Validation will be performed by comparing modelled runoff regarding runoff and flood statistics using stochastically generated lumped catchment rainfall versus multiple point rainfall. It would be advantageous if the results based on catchment rainfall are comparable to those using multiple point rainfall, so catchment rainfall could be generated directly with the stochastic rainfall models. Extremes at the catchment scale may also be better represented if catchment rainfall is generated directly.

How to cite: Thiele, L.-B., Pidoto, R., and Haberlandt, U.: Comparing runoff statistics simulated with a hydrological model utilizing stochastically generated lumped catchment rainfall and multiple point rainfall, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-7435, https://doi.org/10.5194/egusphere-egu2020-7435, 2020.

Heavy rainfall events and the high variability of their occurrence have a significant effect on the urban water cycle and are commonly thought to increase in the future. The increasing risk of urban flash floods is a problem jointly faced by the urban infrastructure, water networks and -systems. A better understanding of the diurnal and seasonal precipitation cycles of short-duration heavy rainfall events is therefore required. This study presents the diurnal and seasonal distribution of those events (10-minute and one-hour) in Germany and puts them into a spatial context. Precipitation data from 22 weather stations of the German Weather Service were statistically examined for the period 2000 - 2018. In addition, the spatial and temporal distribution patterns were compared to spatiotemporal patterns of various controlling factors. Three diurnal distribution patterns can be identified: 1) a homogeneous distribution of events over a maximum period of 24 hours in the S-SW, 2) a non-uniform grouping of events in the morning and afternoon predominantly in the NE and 3) an occurrence of heavy rainfall events in the afternoon in a much shorter time interval in the North. These patterns are not necessarily identical for both event durations and suggest different forms and degrees of drivers. From a seasonal perspective, events of both durations occur exclusively between May and September, with the majority occurring in July and August. Temporal distributions can mainly be explained by controlling factors such as sunshine duration and intensity of radiation whereas spatial differences are also linked to geographical altitudes and typical, summery large-scale weather conditions with the main wind direction from the SW.

How to cite: Haacke, N. and Paton, E. N.: Spatiotemporal patterns of short-duration heavy precipitation in Germany, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-9391, https://doi.org/10.5194/egusphere-egu2020-9391, 2020.

High mountain ranges are characterised by steep slopes and high precipitation rates, making Alpine catchments prone to frequent flood events. Fast runoff during heavy rainfall events, sometimes in combination with snow melt events, can cause severe damages in residential areas. Flood retention mainly depends on retention properties of the headwater catchment area and its interaction with the occurring flood regime. However, due to their special characteristics, Alpine catchments are ideal candidates for storage power plants as well. Currently, around 70 storage power plants are operating in Austria. Their large artificial reservoirs alter the flood retention properties in the upper catchment by potentially providing a higher flood peak attenuation, which of course depends on the available storage volume at the time of flooding. Since it already has been reported that climate change driven processes will increase flood intensity and frequency in Austria, it is of particular interest to understand how hydropower reservoirs alter flood dynamics and if they systematically could be used for flood retention in the future.

In this study the influence of a storage power plant on flood dynamics is shown for an example in the central Austrian Alps. The chain of analysed reservoirs is situated in the headwaters of the river Salzach, a Danube tributary. Based on observed runoff, the retention potential is analysed by comparing the possible natural flood event and the retained flood event in the catchment influenced by the storage power plant. Then its possible impact on the flood hazard downstream is investigated until the tributary drains into the Danube. 

This contribution is part of the interdisciplinary research project “Policy Coordination in Flood Risk Management” (PoCo-FLOOD), which is funded by the Earth System Sciences program of the Austrian Academy of Sciences.

How to cite: Lebiedzinski, K., Fürst, J., Schulz, K., and Herrnegger, M.: Possible impacts of a hydropower reservoir on the flood hazard of an Alpine valley, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-10355, https://doi.org/10.5194/egusphere-egu2020-10355, 2020.

Evidence of past flood regime changes in Europe has been shown by several local and regional trend detection studies. These studies typically analysed changes in the mean or median flood. In this work, we investigate regional trends in the 2-year flood and in the 100-year flood. Additionally, it is of interest to investigate the effect of catchment scale on the changes in time of the selected flood quantiles. We analyse 2370 flood records, selected from a newly available pan-European flood database, with record lengths of at least 40 years over the period 1960-2010 and catchment areas ranging from 5 to 100 000 km². In order to estimate the regional trend in flood quantiles, a non-stationary regional flood frequency approach is used, consisting of a regional Gumbel distribution whose parameters are allowed to vary with time and with catchment area. A Bayesian Markov Chain Monte Carlo (MCMC) approach is used for parameter estimation. With a spatial moving window approach, regional trends of the selected flood quantiles, and the related uncertainties, are estimated and compared across Europe, for hypothetical catchment sizes ranging from 10 to 100 000 km². Distinctive patterns of flood regime change are identified for large regions across Europe which depend on flood magnitudes and catchment areas. The resulting trends in flood magnitudes are positive (with the exception of very large catchments) in Atlantic catchments, where the magnitude of trends decreases with increasing catchment size and for bigger return periods. In Mediterranean catchments the regional trends are negative, with small floods experiencing a stronger decrease than large floods. In Eastern European catchments, the regional trends are clearly negative, with larger magnitudes (in absolute value) for larger catchments; they do not appear to vary substantially with the return period.

How to cite: Bertola, M., Viglione, A., Lun, D., Hall, J., and Blöschl, G.: Regional trends in flood quantiles across Europe between 1960 and 2010, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-10395, https://doi.org/10.5194/egusphere-egu2020-10395, 2020.

In the last decade, floods on the rivers of Russia have become one of the most terrifying natural disasters. Among the catastrophic events, historical flood in Krymsk (2012), Amur River basin (2013), Veliky Ustyug (2016), floods in the Voronezh and Volgograd Region (2018) and Irkutsk and Novgorod Region (2019) can be called.

Floods on the rivers of the Russian Plain are divided into three main genetic types: rain, snowmelt and mixed. There is also a classification by seasons in which they can be observed. The seasonality of the flood peaks passage depends on the geographic location of the catchment and it’s local features. For most of the rivers of Central Russia, it was traditionally believed that occasional floods are mainly observed in the summer-autumn low flow period. In the summer, they are most often associated with intensive rainfall, and in the fall, with prolonged and drizzling rains. The influence of climate change on the processes of runoff formation has led to a transformation of the conditions for the occurrence of flood peaks and the need to rethink traditional ideas.

In this work, we analyzed the daily discharge time-series and highlighted flood peaks at 60 hydrological stations located in different natural zones of the European territory of Russia. Occasional flood peaks were divided into 5 classes, taking into account the time of their formation and genesis: a) thaw peaks during the winter low flow period, b) mixed peaks during the winter low flow period, c) mixed peaks during the rise of the main seasonal (snowmelt) wave, d) rain peaks during the decline of the main seasonal (snowmelt) wave, e) rain peaks during the summer-autumn low flow period.

The total number of peaks, the maximum peak discharge and its unit discharge rate, the beginning, end and duration of the flood peak, the total runoff volume of the flood, the relative stability of the low-flow period were estimated.

On average, the number of flood peaks in the rivers of the study area varies from 1 to 8 events per year. The greatest number of flood peaks is characteristic of the foothills of the Caucasus and the rivers of the Kola Peninsula, as well as the most western regions - the upper reaches of the Seversky Donets, Dnieper, and Western Dvina. The maximum unit discharges of rain floods on average is from 5 to 50 and more and thaw from 2 to 20 l/s*km². The spatial pattern shows that higher unit discharges are typical for the windward western slopes of the hills, and relatively low ones are observed on the leeward, eastern slopes. In general, unit discharge rats increase from southwest to east, northeast.

In recent decades, the seasonality of flood peaks has changed significantly, they began to be observed in almost any period of the year, the number of events in the pre-flood period increased, as well as in the autumn period, at the time of transition to negative air temperatures.

The study was supported by the Russian Science Foundation grant No.19-77-10032

How to cite: Kireeva, M., Rets, E., Natalia, F., and Artem, G.: Occasional floods on the Russian Plain: types, frequency and conditions for the origin in the face of changing climate, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-11417, https://doi.org/10.5194/egusphere-egu2020-11417, 2020.

Climate change will modify magnitude and timing of floods in the future. Consequently, new methodologies to quantify the impact of climate change on floods at the catchment scale are required. This study aims to quantify the changes in peak flow quantiles expected in the future, by using the climate change projections of the Fifth Report (AR5) of the IPCC, supplied by the EURO-CORDEX programme. Four catchments located in the Duero River Basin in northwestern Spain are considered as case studies.

First, biases in precipitation and temperature climate projections have been corrected by using the available observations in the control period (1971-2004) in the four catchments. Second, the hydrological response in the four catchments has been simulated with the continuous simulation model HBV. The model has been calibrated in the four catchments. Time series of soil moisture content in the catchment were obtained, identifying the initial moisture content in the day of occurrence of the annual maximum rainfalls. Third, an event model has been used to simulate flood response to the annual maximum rainfalls, considering the initial soil moisture content supplied by the HVB model. The results of the event model provides a better characterization of the catchment flood response than the continuous HBV model.

The methodology has been applied in the control period (1971-2004), for validation purposes. Then, the methodology has been applied to the future period (2011-2095), to obtain the expected changes in peak flow quantiles, as a consequence of climate change. The combined use of the results of the continuous hydrological simulation with the HBV model with the event model improves the results provided by either the HBV model or the event model independently. The proposed methodology allows a better characterisation of the catchment flood response to a given precipitation event, while also considering the expected variation in the antecedent moisture content in the catchment in the future, as a consequence of expected changes in temperature and precipitation regimes. The application of the proposed methodology to the case studies has shown that climate change will increase peak flow quantiles in the future, in three of the four catchments.

How to cite: Soriano Martín, E., Mediero Orduña, L., and Garijo Sarria, C.: Quantifying the impact of climate change on floods by using both continuous and event-based hydrological modelling, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-11466, https://doi.org/10.5194/egusphere-egu2020-11466, 2020.

Flood is one of the most frequent disasters which dangerously impacts societies and economies worldwide. Floodplain management and hydraulic risk analysis based on design flood estimation are essential tools to reduce damages and save human lives. Flood Frequency Analysis (FFA) has been classically used to derive design river discharge estimates, however, the scarce availability of discharge observations, especially in small catchments (<150 km2), makes its application not always possible. In addition, with the projections foreseen by the International Panel on Climate Change (IPCC) the use of FFA might lead to incorrect estimates of design river discharge as FFA is based on the concept of stationarity. Generally, long rainfall and temperature time series are much more available than discharge observations but their temporal coverage is often not sufficient for carrying out FFA via a hydrological simulation.

To handle these drawbacks, the combination of a stochastic generation of rainfall and temperature time series, Regional Circulation Model (RCM) projections and continuous hydrological models provides a reliable tool for obtaining long river discharge time series to implement FFA. However, design flood estimations can be significantly uncertain due to several factors such as 1) the specific model structure, parameterizations and processes representation, 2) the catchment hydrology and 3) the specific climate change scenario.

The primary objective of this study is to explore the sensitivity of the design river discharge estimates to the hydrological model complexity and parameterization. For this, three continuous hydrological distributed models named the Modello Idrologico SemiDistribuito in continuo (MISDc), the Soil & Water Assessment Tool (SWAT) and GEOFrame NewAGE model are forced with long timeseries of rainfall and temperature obtained via the Neyman-Scott rectangular pulse model (NSRP) for stochastic rainfall generation, and the fractionally differenced ARIMA model (FARIMA) for stochastic temperature generation. A secondary objective is to understand the impact of climate change and the catchment hydrology on the design river discharge estimates via the use of different RCM projections.

The study is carried in the Upper Nera catchment in Central Italy which was impacted by the recent 2016 earthquake and for which is necessary to identify hydraulic risk mitigation measures and adaptation for a forward planning in the floodplain areas where new settlements will be rebuilt.

Preliminary results suggest the high dependency of the design river discharge estimates to the chosen hydrological model and a different response of the sub-catchments to the climate change scenario.

How to cite: Azimi, S., Barbetta, S., Moramarco, T., Tarpanelli, A., Camici, S., Formetta, G., Rigon, R., and Massari, C.: On the impact of the hydrological model and catchment hydrology on the design flood estimation in a small catchment in Central Italy affected by the recent 2016 earthquake events, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-13571, https://doi.org/10.5194/egusphere-egu2020-13571, 2020.

The severity of floods is not only affected by the physiogeographic characteristics and the meteorological conditions of the catchment, but also by the river network. If a flood occurs at the same time in tributary and main river, the tributary flood wave can amplify the flood wave in the main river. To investigate the impact of flood wave superposition, the 6-10 largest floods in the four main German river basins (Danube, Elbe, Rhine, Weser) are analyzed. The flood waves are tracked along the river course. Flood magnitude and flood timing are analyzed at each triple point. A triple point consists of the hydrological stations in the tributary and in the main river (upstream and downstream of the confluence). The return periods are calculated separately at each triple point for all three hydrological stations. In addition, changes in the return periods along a river course are analyzed for each flood event. The flood magnitudes and their return periods are compared with the spatiotemporal precipitation distributions and other influencing factors. The results show that the contribution of the different confluences to the flood severity at the main river is event-specific. Partly, the return period is only high at the lower parts of the river basin, partly a high return period in the upper parts of the river basin does not lead to a high return period downstream.

How to cite: Guse, B., Wietzke, L., Ullrich, S., Merz, B., and Vorogushyn, S.: Impact of river confluences on return periods of large floods, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-19448, https://doi.org/10.5194/egusphere-egu2020-19448, 2020.

Human activities can strongly influence the capacity of ecosystems to provide flood regulating ecosystem services (ES). Therefore, the effects of land use alteration, population migration and urbanization are key aspects to be considered when dealing with flood management. This study aims at analyzing the spatio‑temporal dynamics of flood regulating ES to support watershed management planning. The spatial explicit analysis of flood regulating ES is carried out with SWAT - Soil and Water Assessment Tool, using daily meteorological data between 2000 and 2014. Two indicators are elaborated in order to evaluate the retention capacity of each land use setting and to map the ES supply. Demand quantification is obtained from the information derived by the existing flood management plans (i.e. PAI-Piano per l’Assetto Idrogeologico and PGRA-Piano di Gestione del Rischio Alluvioni) which contain the identification and the perimeter of hydraulic hazard classes. Supply and demand data are then merged in order to obtain budget maps of flood regulating ES and their evolution from 1960 up to 2012 (1960, 1990, 2000 and 2012). The results show that both the demand and the supply of ecosystem services change during the time. With the increasing urbanization, the demand values have grown in the Arno floodplains, where residential, industrial and commercial zones are located. At the same time, land use changes (e.g. intensive agriculture) have caused negative effects on water regulation supply. This work shows the advantages of assessing flood regulating ES to improve flood regulation in the Arno river basin and provide a sound base of knowledge to identify floods prevention and mitigation measures.

How to cite: Mori, S., Pacetti, T., Brandimarte, L., and Caporali, E.: Spatio-temporal dynamics of flood regulating ecosystem services in the Arno river basin, Italy, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-20148, https://doi.org/10.5194/egusphere-egu2020-20148, 2020.