Flood Data Research Articles

Climatology and ranking of hazardous precipitation events in the western Mediterranean area

The western Mediterranean region often suffers from the devastating effects of flooding, caused by enormous rain accumulations that sometimes resemble the values produced by tropical systems. Despite the climatic and social relevancy of this type of episodes, there are some fundamental questions that would still be difficult to answer today, for example: where within the region are more cases recorded? or, which were the most potentially dangerous episodes? In this study, we identify, and then gather and unify information from, all the daily events occurred from 1980 to 2015. Using the MESCAN high-resolution gridded rainfall dataset, events are detected and for each case, the date and affected regions are recorded. Subsequently, events are ranked according to their magnitude and classified by weather type. In addition, flood data from the FLOODHYMEX and EM-DAT databases are used to check whether the precipitation episodes resulted in flooding. All this information is collected into a publicly available single database. Results show that the highest number of events per year is recorded in the Languedoc-Roussillon region (France) and in the Valencian Community (Spain). The cases of greatest magnitude, which are associated with a larger number of floods, present a very marked seasonality, with about 80% of them occurring in September, October and November. Finally, we show that only four weather types are present in most of the days with hazardous rainfall in the western Mediterranean. The most hazardous situation is characterized by a low-pressure area at all tropospheric levels in the eastern Atlantic, forming a block pattern with an anticyclonic ridge that tends to extend from the Central Mediterranean to Central Europe. About 40% of the most extraordinary cases are associated with this configuration. As an example, the infamous Piedmont (Italy) 1994 episode, in the top 10 of the ranking, was produced by an atmospheric pattern of this type.

Can low salinity water injection be an efficient option for the Recôncavo Basin? An experimental and simulation case study

Low salinity water injection (LSWI) has been studied by several researchers as an enhanced oil recovery (EOR) method in laboratory experiments, numerical simulations, and field applications. However, its effects on oil recovery remain uncertain due to the complexity of the interactions among crude oil, brine, and the rock surface, for both sandstone and carbonate reservoirs, particularly when they contain a diversity of clay minerals. This paper investigates LSWI performance and mechanisms in sandstone rock, brine, and light crude oil with low polar compounds from a reservoir in Brazil. Core flooding experiments and reservoir simulations were carried out to understand the influence of key parameters on oil recovery. Zeta potential measurements were performed to investigate the influence of the formation water and NaCl and CaCl2 brine solutions on rock and oil surface charges. Increased oil recovery, Ca++ desorption from the rock surface, and a substantial pH increase in the effluent were noticed during LSWI. Based on the zeta potential measurement, a thicker water film was also formed on the rock and oil surface with brine dilution. A recovery factor of 64% was reached by testing LSWI from the beginning of the waterflooding (secondary mode) and this was the highest recovery of all the scenarios evaluated. Ten parameters were successfully history matched to core flood data, and optimization and sensitivity analysis showed that the concentration of Na+ and Ca++, and the water injection rate were the most important parameters in the LSWI process. Overall, this study sheds light on the LSWI mechanisms for the Brazilian oil field investigated and provides useful operational parameters for designing and achieving better LSWI performance.

Are OpenStreetMap building data useful for flood vulnerability modelling?

Abstract. Flood risk modelling aims to quantify the probability of flooding and the resulting consequences for exposed elements. The assessment of flood damage is a core task that requires the description of complex flood damage processes including the influences of flooding intensity and vulnerability characteristics. Multi-variable modelling approaches are better suited for this purpose than simple stage–damage functions. However, multi-variable flood vulnerability models require detailed input data and often have problems in predicting damage for regions other than those for which they have been developed. A transfer of vulnerability models usually results in a drop of model predictive performance. Here we investigate the questions as to whether data from the open-data source OpenStreetMap is suitable to model flood vulnerability of residential buildings and whether the underlying standardized data model is helpful for transferring models across regions. We develop a new data set by calculating numerical spatial measures for residential-building footprints and combining these variables with an empirical data set of observed flood damage. From this data set random forest regression models are learned using regional subsets and are tested for predicting flood damage in other regions. This regional split-sample validation approach reveals that the predictive performance of models based on OpenStreetMap building geometry data is comparable to alternative multi-variable models, which use comprehensive and detailed information about preparedness, socio-economic status and other aspects of residential-building vulnerability. The transfer of these models for application in other regions should include a test of model performance using independent local flood data. Including numerical spatial measures based on OpenStreetMap building footprints reduces model prediction errors (MAE – mean absolute error – by 20 % and MSE – mean squared error – by 25 %) and increases the reliability of model predictions by a factor of 1.4 in terms of the hit rate when compared to a model that uses only water depth as a predictor. This applies also when the models are transferred to other regions which have not been used for model learning. Further, our results show that using numerical spatial measures derived from OpenStreetMap building footprints does not resolve all problems of model transfer. Still, we conclude that these variables are useful proxies for flood vulnerability modelling because these data are consistent (i.e. input variables and underlying data model have the same definition, format, units, etc.) and openly accessible and thus make it easier and more cost-effective to transfer vulnerability models to other regions.

A large-scale 2005–2012 flood map record derived from ENVISAT-ASAR data: United Kingdom as a test case

Synthetic Aperture Radars (SAR) are adequate sensors for mapping water bodies from space as they can be used to acquire data of equal quality both day and night and practically regardless of weather conditions. Furthermore, the global coverage of SAR data provides an opportunity to generate global scale flood records that are essential for improving our understanding of flood risks worldwide and of how these risks are changing over time. In this study, we introduce an automatic change-detection based method that allows global-scale flood records to be generated using the readily and freely available ENVISAT-ASAR data collection. It consists of the following three steps: (i) flood image identification; (ii) reference image selection; (iii) floodwater detection. As a test case, this study uses all available ENVISAT-ASAR images from eight different orbital tracks that were acquired over the United Kingdom over the period 2005–2012. Due to a lack of large-scale ground truth data, the evaluation of the results is carried out using different data sources. First, subsets of the flood maps over the Severn River basin are evaluated using a flood extent map that was manually digitized from very high-resolution aerial imagery. According to our results, the overall accuracy of both flood maps' subsets is higher than 85% while the user accuracy of the flood class is above 88%. Next, for the regions and images without available ground truth data, a visual inspection is carried out using simulations generated by the hydraulic model LISFLOOD-FP, as well as LANDSAT 7 ETM+ images obtained with a 30 m spatial resolution. Meanwhile, by comparing the acquisition dates of identified flood SAR images, the LISFLOOD-FP model results and optical data, a good agreement has been found. The experimental results over the United Kingdom indicate that the proposed method has strong potential for the generation of a global flood data record from the ENVISAT-ASAR archive.

A comprehensive flood event specification and inventory: 1930–2020 Turkey case study

Flooding is one of the most frequent natural disasters that have significant impact on communities in terms of loss of life, direct and indirect economic losses, and disruption of daily life. Decision makers often depend on flood data inventories to make more informed decisions on the development of flood mitigation plans to protect flood prone communities. A comprehensive inventory that covers multiple aspects of a flood event is critical to identify vulnerable regions, historical trends, and mitigate possible flood impacts. This study proposes an integrated flood data specification to support multi-stakeholder use cases, community-based sustainable domain specific maintenance, and crowdsourced data collection and management. The specification is designed based on comprehensive review of existing global and national repositories, scientific studies and needs and requirements of stakeholders. The specification is designed to include metadata on environmental, economic, and demographic impact, hydraulic, hydrologic, and meteorological features, and detailed location information of a flood event. As a case study, a flood event inventory was compiled for Turkey between 1930 and 2020 using existing national and global data sources and digitized media archives. A total of 2101 flood events with 64 data attributes have been collected over the period of 90 years. An initial statistical analysis of the inventory is also presented for assessment of the seasonal and regional characteristics of flooding in Turkey.

Sub-2 mm depth of interaction localization in PET detectors with prismatoid light guide arrays and single-ended readout using convolutional neural networks.

Depth of interaction (DOI) readout in PET imaging has been researched in efforts to mitigate parallax error, which would enable the development of small diameter, high-resolution PET scanners. However, DOI PET has not yet been commercialized due to the lack of practical, cost-effective, and data efficient DOI readout methods. The rationale for this study was to develop a supervised machine learning algorithm for DOI estimation in PET that can be trained and deployed on unique sets of crystals. Depth collimated flood data was experimentally acquired using a Na-22 source with a depth-encoding single-ended readout Prism-PET module consisting of lutetium yttrium orthosilicate (LYSO) crystals coupled 4-to-1 to 3×3 silicon photomultiplier (SiPM) pixels on one end and a segmented prismatoid light guide array on the other end. A convolutional neural network (CNN) was trained to perform DOI estimation on data from center, edge and corner crystals in the Prism-PET module using (a) all non-zero readout pixels and (b) only the 4 highest readout signals per event. CNN testing was performed on data from crystals not included in CNN training. An average DOI resolution of 1.84mm full width at half maximum (FWHM) across all crystals was achieved when using all readout signals per event with the CNN compared to 3.04mm FWHM DOI resolution using classical estimation. When using only the 4 highest signals per event, an average DOI resolution of 1.92mm FWHM was achieved, representing only a 4% dropoff in CNN performance compared to using all non-zero pixels per event. Our CNN-based DOI estimation algorithm provides the best reported DOI resolution in a single-ended readout module and can be readily deployed on crystals not used for model training.

A Procedure for Combining Improved Correlated Sampling Methods and a Resampling Strategy to Generate a Multi-Site Conditioned Streamflow Process

In this study, a new method is developed to generate multi-site streamflow series. First, a correlated sampling method based on spectral decomposition is employed to simulate a large number of individual years of multi-site streamflow series. Then, combined with historical design flood data, flood characteristic series considering historical floods are generated. Finally, a subset of the streamflow series corresponding to these flood characteristics is selected from the pool of multi-site simulated streamflow series. The proposed model is applied to multi-site streamflow simulations in the upper Yangtze River basin. Experimental results show that the simulated data not only ensure the fundamental statistical characteristics of the observed data, but also provide good preservation of high order self-dependent, cross-dependent at finer and coarser time scales with any marginal distributions and any correlation structures. In addition, the simulated extreme situations are more reliable for risk and vulnerability analysis of multi-reservoir joint operation systems due to the consideration of the historical design flood data.

A GIS-Based Artificial Neural Network Model for Flood Susceptibility Assessment

This article presents a geographic information system (GIS)-based artificial neural network (GANN) model for flood susceptibility assessment of Keelung City, Taiwan. Various factors, including elevation, slope angle, slope aspect, flow accumulation, flow direction, topographic wetness index (TWI), drainage density, rainfall, and normalized difference vegetation index, were generated using a digital elevation model and LANDSAT 8 imagery. Historical flood data from 2015 to 2019, including 307 flood events, were adopted for a comparison of flood susceptibility. Using these factors, the GANN model, based on the back-propagation neural network (BPNN), was employed to provide flood susceptibility. The validation results indicate that a satisfactory result, with a correlation coefficient of 0.814, was obtained. A comparison of the GANN model with those from the SOBEK model was conducted. The comparative results demonstrated that the proposed method can provide good accuracy in predicting flood susceptibility. The results of flood susceptibility are categorized into five classes: Very low, low, moderate, high, and very high, with coverage areas of 60.5%, 27.4%, 8.6%, 2.5%, and 1%, respectively. The results demonstrate that nearly 3.5% of the study area, including the core district of the city and an exceedingly populated area including the financial center of the city, can be categorized as high to very high flood susceptibility zones.

Evaluation and validation of flood hazard zoning using Analytical Hierarchy Process and GIS: A case study of Lam River basin (Vietnam)

In recent years, the Lam river basin had suffered various forms of natural disasters such as floods, inundations, windstorms, tornadoes, etc. Among all these, the flood has proved to be the greatest threat to the people and the socio-economic development in the basin. Moreover, it is very frequent as compared to other natural disasters. In view of the fact that such disastrous floods are still occurring in the basin, it becomes a necessity to determine the causes and analyze the components affecting flood. This is important in order to develop an early flood warning system and thus minimize the negative impact of flood in the Lam river basin on the people and the facilities. In this paper, the Analytical Hierarchy Process (AHP) analysis method integrated with GIS technology is used to map flood risk zones in the Lam river basin. The parameters used for the analysis are the main causes affecting the floods. In addition to the 5 most commonly used factors such as slope, rainfall, land cover, soil, and drainage density, this study also includes a new factor - relative slope length to compute a more rigorous and reliable model. The results were compared with the two more methods of flood hazard zoning in the same study area: the method of the main flood caused factor analysis and the method of inheriting, data analyzing, and processing. The results were also validated by the historical flood data of three years 2010, 2013, and 2016.

Python program for spatial reduction and reconstruction method in flood inundation modelling

Fast and accurate modelling of flood inundation has gained increasing attention in recent years. One approach gaining popularity recently is the development of emulation models using data driven methods, such as artificial neural networks. These emulation models are often developed to model flood depth for each grid cell in the modelling domain in order to maintain accurate spatial representation of the flood inundation surface. This leads to redundancy in modelling, as well as difficulties in achieving good model performance across floodplains where there are limited data available. In this paper, a spatial reduction and reconstruction (SRR) method is developed to (1) identify representative locations within the model domain where water levels can be used to represent flood inundation surface using deep learning models; and (2) reconstruct the flood inundation surface based on water levels simulated at these representative locations. The SRR method is part of the SRR-Deep-Learning framework for flood inundation modelling and therefore, it needs to be used together with data driven models. The SRR method is programmed using the Python programming language and is freely available from https://github.com/yuerongz/SRR-method.•The SRR method identifies locations which are representative of flood inundation behavior in surrounding areas.•The representative locations selected following the SRR method have sufficient flood data for developing emulation models.•Flood inundation surfaces can be reconstructed using the SRR method with a detection rate of above 99%.

Flash flood risk management modeling in indian cities using IoT based reinforcement learning

Each year, flash floods in India have affected life, infrastructure and the economy of the country and there is a need for a systemic model of real-time flash flood management. So, to mitigate these losses, we proposed a flash flood management model focused on reinforcement learning. Based on their severity, the flash flood data is collected and rewards are distributed and this data is compared to the data collected from smart IoT devices deployed in the region impacted by the flood. We allocate the state-based reward values once the comparison is completed. The evacuation of flash flood water through the contour of the flash flood is carried out. This bypass has gates at different points that mean that, depending on the incentives, the gates are opened to remove flash flood water. The proposed solution was evaluated to be a faster, more effective and more accurate real-time flash flood management method.

Estimation of Hourly Flood Hydrograph from Daily Flows Using Artificial Neural Network and Flow Disaggregation Technique

Flood data on a high temporal scale are required for the design of hydraulic structures, flood risk assessment, flood protection, and reservoir operations. Such flood data are typically generated using rainfall-runoff models through an accurate calibration process. The data also can be estimated using a simple relationship between the daily and the sub-daily flow records as an alternative to rainfall–runoff modelling. In this study, we propose an approach combining an artificial neural network (ANN) model for peak flow estimation and the steepness index unit volume flood hydrograph (SIUVFH) method for sub-daily flow disaggregation to generate hydrographs on an hourly time scale. The SIUVFH method is based on the strong relationship between the flood peak and the steepness index, which is defined as the difference between the daily flood peak and daily flow several days before the peak; it is also used for selecting a reference unit volume flood hydrograph to be scaled to obtain the sub-daily flood hydrograph. In this study, to improve the applicability of the SIUVFH method for locations with a weak relationship between the flood peak and steepness index, the ANN-based flood peak estimation was used as an additional indicator to determine a reference unit volume flood hydrograph. To apply the proposed method, ANN models for estimating the peak flows from the mean daily flows during peak and adjacent days were constructed for the studied dam sites. The optimal ANN structures were determined through Monte Carlo cross-validation. The results showed a good performance with statistical measurements of relative root mean square errors of 0.155–0.224, 0.208–0.301, and 0.244–0.382 for the training, validation, and testing datasets, respectively. An application of the combined use of the ANN-based peak estimation and the SIUVFH-based flow disaggregation revealed that the disaggregated hourly flows satisfactorily matched the observed flood hydrograph.

Political / Power Structure and Vulnerability to Natural Disaster in North Western Pakistan

The current study tries to explore the link between the power structure and vulnerability to natural disasters like flash floods in Chitral, situated in the northwest of Pakistan. Flash floods are common in northern areas of Pakistan because of the steep nature of the terrain, melting of glaciers, and torrential rain in summer. In the Chitral district, its effects are adverse because of dwelling patterns and scarcity of safe locations for living. Ethnographic research tool, participant observation, in-depth interviews, informal talking, were used to understand the relationship between the power structure/ political structure in the district and vulnerability to flash floods and its aftermaths. Descriptive analysis technique has been used to analyze the ethnographic data. In princely state lower status people used to be compensated through land. In the locale of research, situated in upper Chitral, it has been observed that most of the low-status families are living in marginal locations making them vulnerable to flash floods. 2013 and 2015 flash floods data in the village of Reshun shows that majority of the affected families belong to low-status groups. Mainly it is the vulnerable location where these families are living. These families were not only affected more than upper-status people living in the village but the post-disaster situation was also worse for them in comparison to other parts of the village.

Hydrodynamic Analysis of a Stormwater System, under Data Scarcity, for Decision-Making Process: The Duran Case Study (Ecuador)

Urban flooding is a major problem in many coastal cities around the world, mainly caused by factors such as poor urban planning, outdated sewer capacity or high frequent extreme events. In developing countries such as Ecuador, lack of monitoring, financial constraints and absence of proper policies exacerbate flooding problems. The use of the Stormwater Management Model (SWMM) helped to identify zones that are flooded even with a 2-year precipitation event at Duran city. Since there are no flood monitoring stations across the study area, model calibration was performed against unofficial records (people complaints, media photographs, and water marks on walls). Model results fairly agreed with those unofficial record. However, a validation process is not achievable yet due to lack of enough flood data even in flood prone areas. To further increase city resilience, authorities and stakeholders should engage in climate actions to raise flood hazard-risk awareness, monitoring hydro-meteorological factors, complement drainage operations with nature-based solutions, and improve hydrodynamic modelling to develop a flood early warning system for the city.

Study on the influence of vegetation change on runoff generation mechanism in the Loess Plateau, China

Abstract In recent years, the amount of water and sediment in the Yellow River Basin has dropped drastically. This paper selected 125 rainfall and flood data points from 1965 to 2015, combined hydrological methods and mathematical statistics to analyze the hydrological factors and runoff generation mechanism, and combined the underlying surface conditions of the Gushanchuan Basin. The characteristics of change revealed the temporal and spatial variation characteristics and related factors of the runoff generation mechanism in the basin. The results showed that the Gushanchuan Basin is still dominated by HOF runoff, but the runoff generation mechanism has also changed with changes in the underlying surface, which are reflected in increased runoff components, the reduced proportion of HOF runoff, and the increased proportion of saturation-excess overland flow (SOF) runoff and mixed runoff. We analyzed the variation law of underlying surface in the basin, which indicated that the increase in the forest grass area was the main factor affecting changes in the watershed runoff generation mechanism. This research will enable a deeper understanding of the runoff generation mechanism of the main soil erosion areas in the Loess Plateau, and reveal variations in the runoff generation mechanism in the Yellow River.

Regional Flood Frequency Analysis using Dimensionless Index Flood Method

Hydrologic designs require accurate estimation of quartiles of extreme floods. But in many developing regions, records of flood data are seldom available. A model framework using the dimensionless index flood for the transfer of Flood Frequency Curve (FFC) among stream gauging sites in a hydrologically homogeneous region is proposed. Key elements of the model framework include: (1) confirmation of the homogeneity of the region; (2) estimation of index flood-basin area relation; (3) derivation of the regional flood frequency curve (RFFC) and deduction of FFC of an ungauged catchment as a product of index flood and dimensionless RFFC. As an application, 1983 to 2004 annual extreme flood from six selected gauging sites located in Anambra-Imo River basin of southeast Nigeria, were used to demonstrate that the developed index flood model: , overestimated flood quartiles in an ungauged site of the basin. It is recommended that, for wider application, the model results can be improved by the availability and use of over 100 years length of flood data spatially distributed at critical locations of the watershed. Doi: 10.28991/cej-2020-03091627 Full Text: PDF

Factors influencing the success of information systems in flood early warning and response systems context

Flash flood is a natural disaster that often occurs after heavy rain, and it is getting more common nowadays. The flood early warning and response system (FEWRS) can be installed to minimize the level of damage and the number of casualties due to flood by providing accurate and reliable flood data. Unfortunately, the existing number of studies detailing on the factors affecting the efficiency of FEWRS in flood disaster is quite limited. The above issue is addressed in the current work, which involves conducting a comprehensive literature review on the factors that drive the effectiveness of information systems (IS) in FEWRS. The current analysis was based on the Wymer and Regan's standards. From the 66 factors identified from the previous studies on IS adoption, the most significant factors affecting the effectiveness of FEWRS are: system quality, information quality, user satisfaction, service quality, use, perceived usefulness, intention to use,net benefits, perceived ease of use,compatibility, user experience, relative advantage, complexity, perceivedrisks, educational quality, and confirmation, these factors can be constructed to the success model to address the effectiveness of FEWRS in disaster management.

How much spatial resolution do we need to model a local flood event? Benchmark testing based on UAV data from Biga River (Turkey)

Digital elevation models (DEMs) are important in flood modelling as they present surface information and obstacles giving the direction of water during a flood event. All hydraulic modelling software needs both channel and floodplain elevation data to simulate the water flow from upstream to downstream. Based on the accuracy and representative capacity of the surface, modelling results can be evaluated in terms of hazard and risk analysis. Developments in remote sensing and GIS technology provide many sources to freely obtain DEM data of different resolution and accuracy. However, access to and the processing of high-resolution data associated with large areas is difficult and time-consuming. This study, therefore, focuses on determining the availability of different resolutions and DEM sources in flood modelling by testing unmanned aerial vehicle (UAV) data. For this, freely available DEMs such as ALOS DEMs (12.5 and 30 m), EU-DEM (25 m), ASTER GDEM (30 m), SRTM (30 and 90 m), MERIT DEM (90 m) and TANDEM-X (90 m) were chosen to investigate part of Biga River (Turkey) which features a natural channel. Then, 1D flood modelling was examined using different DEM resolutions, Manning’s values and flood data of various frequencies (10, 100 and 1000 years). Results were tested using high-resolution DEM data (5 cm) gathered from a UAV. It was concluded that EU-DEM is the most suitable dataset in terms of flood extent, depth and velocity. However, depending on the boundary characteristics of the data, ALOS 12.5 DEM and MERIT DEM (90 m) data are the most suitable in their extent, depth and velocity for high-resolution and low-resolution flood modelling worldwide extent.

Spatio-temporal analysis of flood data from South Carolina

To investigate the relationship between flood gage height and precipitation in South Carolina from 2012 to 2016, we built a conditional autoregressive (CAR) model using a Bayesian hierarchical framework. This approach allows the modelling of the main spatio-temporal properties of water height dynamics over multiple locations, accounting for the effect of river network, geomorphology, and forcing rainfall. In this respect, a proximity matrix based on watershed information was used to capture the spatial structure of gage height measurements in and around South Carolina. The temporal structure was handled by a first-order autoregressive term in the model. Several covariates, including the elevation of the sites and effects of seasonality, were examined, along with daily rainfall amount. A non-normal error structure was used to account for the heavy-tailed distribution of maximum gage heights. The proposed model captured some key features of the flood process such as seasonality and a stronger association between precipitation and flooding during summer season. The model is able to forecast short term flood gage height which is crucial for informed emergency decision. As a byproduct, we also developed a Python library to retrieve and handle environmental data provided by some main agencies in the United States. This library can be of general usefulness for studies requiring rainfall, flow, and geomorphological information over specific areas of the conterminous US.

New flood frequency estimates for the largest river in Norway based on the combination of short and long time series

Abstract. The Glomma River is the largest in Norway, with a catchment area of 154 450 km2. People living near the shores of this river are frequently exposed to destructive floods that impair local cities and communities. Unfortunately, design flood predictions are hampered by uncertainty since the standard flood records are much shorter than the requested return period and the climate is also expected to change in the coming decades. Here we combine systematic historical and paleo information in an effort to improve flood frequency analysis and better understand potential linkages to both climate and non-climatic forcing. Specifically, we (i) compile historical flood data from the existing literature, (ii) produce high-resolution X-ray fluorescence (XRF), magnetic susceptibility (MS), and computed tomography (CT) scanning data from a sediment core covering the last 10 300 years, and (iii) integrate these data sets in order to better estimate design floods and assess non-stationarities. Based on observations from Lake Flyginnsjøen, receiving sediments from Glomma only when it reaches a certain threshold, we can estimate flood frequency in a moving window of 50 years across millennia revealing that past flood frequency is non-stationary on different timescales. We observe that periods with increased flood activity (4000–2000 years ago and <1000 years ago) correspond broadly to intervals with lower than average summer temperatures and glacier growth, whereas intervals with higher than average summer temperatures and receding glaciers overlap with periods of reduced numbers of floods (10 000 to 4000 years ago and 2200 to 1000 years ago). The flood frequency shows significant non-stationarities within periods with increased flood activity, as was the case for the 18th century, including the 1789 CE (“Stor-Ofsen”) flood, the largest on record for the last 10 300 years at this site. Using the identified non-stationarities in the paleoflood record allowed us to estimate non-stationary design floods. In particular, we found that the design flood was 23 % higher during the 18th century than today and that long-term trends in flood variability are intrinsically linked to the availability of snow in late spring linking climate change to adjustments in flood frequency.