Attribution of flood impacts shows strong benefits of adaptation in Europe since 1950

doi:10.21203/rs.3.rs-5430941/v1

Download PDF

Research Article

Attribution of flood impacts shows strong benefits of adaptation in Europe since 1950

https://doi.org/10.21203/rs.3.rs-5430941/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Flood impacts in Europe are considered to be increasing, especially in connection to climate change. However, attribution of impacts to climatic and societal drivers of past floods has been limited to a selection of recent events. Here, we present an impact attribution study covering 1729 riverine, coastal and compound events that were responsible for the large majority of flood-related impacts in Europe between 1950 and 2020. We show that in most regions the magnitude of flood impacts has been regulated primarily by the opposing direct human actions. On the one hand, the population and economic value at risk have increased, exacerbated by land use change. However, it was compensated by improved risk management, manifested by better flood protection and lower vulnerability. Climate change and human alterations of river catchments were also important drivers in many regions, but ultimately less relevant for trends in total, continental-wide impacts. Overall, our study highlights the need for multidimensional impact attribution of past natural hazards. Attribution results for individual events are available on https://naturalhazards.eu/.

Hydrology

Impact attribution

climate change

catchment alteration

exposure

vulnerability

Floods are an ever-present risk to society and economy in Europe (1, 2). The general perception is that the observed increasing impacts are a consequence of climate change manifested by more extreme precipitation (3–6) and sea levels observed in recent decades (7–8). Recent decades are considered particularly rich in hydrological extremes in Europe (2), but the effects of climate change on river discharges are very diverse spatially (9, 10). Therefore, change in climate conditions alone cannot explain long-term variations in flood occurrence (11, 12), as socioeconomic change also influences the hydrological response of catchments through reservoir construction (13) and land use change (14). Further, flood risk in Europe has been exacerbated by exposure growth (1, 15–18) and urbanisation (11) in riverine and coastal floodplains, but reduced by flood adaptation and prevention measures (19–21). Attribution of individual European flood disasters has so far been limited to case studies, focussing on detecting the climate change component in precipitation (22–24) and rarely covering a large number of events (25). Only one study (16) estimated historic trends in all three risk components, but it used only modelled riverine flood impact timeseries for 1980–2010 and did not provide any information for individual countries or events. Until now, there has been a lack of consistent, detailed data on past flood events and their drivers, which is necessary for an integrated attribution of observed impacts (11, 26–28). We address the limitations of previous assessments by combining multiple pan-European datasets to determine the contribution of six distinct drivers of impacts caused by a large number of reported flood events (n = 1729) from 37 European countries that have occurred between 1950 and 2020 (Fig. 1).

Historical riverine, coastal and compound floods were obtained from the HANZE impact database that collects available documentary data on European floods (29, 30). Those floods were reconstructed utilizing hydrological and hydrodynamic modelling of historical river discharges and storm surges (30, 31), historical data on land use, population and economy (32), as well as a novel method to estimate changes in flood protection and vulnerability (33). The latter includes the effect of flood experience, governance and socio-economic development on changes in structural protection levels and inundation impacts through adaptation. Each reconstructed flood event was modelled under different climatic and socioeconomic conditions, i.e. each of the six drivers was either set to factual (historical) values, constituting the best estimate of the conditions in the year when the event has occurred, or to counterfactual values representing a stationary environment with no changes since 1950. Additionally, the uncertainty in the estimate of the historical and 1950 conditions was calculated for each driver. The likelihood that the difference between the factual and counterfactual scenario is larger than the model uncertainty is described below using the Intergovernmental Panel on Climate Change (IPCC) recommended terminology (34), i.e. virtually certain (> 99%), very likely (> 90%), likely (> 66%), about as likely as not (up to 66%).

Historical flood impacts in Europe

Analysis of the 1729 flood events in our attribution dataset shows (Fig. 2) that fatalities from flooding in Europe have declined sharply, nearly fivefold between the 1950s and the 2010s. The three main flood types (riverine, flash, and coastal) caused a similar number of fatalities, about 3000 each, but coastal flood fatalities are concentrated in the 1950s and 1960s. The number of people affected shows large temporal variability without a strong long-term trend. Slow-onset riverine floods have been responsible for the majority of people affected (76%). Direct economic losses have nearly doubled from 37 billion Euro in the 1959s to 71 billion Euro in the last decade. Yet, when expressed relative to the gross domestic product (GDP) of the study area, economic impacts have considerably decreased and are in the last decade about one-third of those in the 1950s. Overall impacts relative to the size of the population and economy have been most severe in southern (Spain, Italy) and south-eastern parts of the continent (Albania, Bosnia and Herzegovina, Romania), with very limited impacts in the northern parts of Europe. The majority of coastal and compound flood impacts occurred in Italy, the Netherlands, Germany and the United Kingdom, with a higher concentration in the 1950s and 1960s.

Drivers of flood impacts in Europe

The most important drivers of flood impacts in Europe have been exposure growth and vulnerability decline, with smaller contributions from other drivers (Fig. 3). Climate change has very likely increased population affected and economic loss (both by 8%), but there is lower confidence in the increase in fatalities (by 1%) (Fig. 3a). Climate change has likely influenced impacts for all flood types, except fatalities from coastal and compound events, which can be explained by the concentration of fatalities from those events in the first two decades (Fig. 2). The strongest increases due to climate change were found for flash floods in terms of fatalities and economic loss, and for riverine floods in terms of population affected. Catchment alteration has increased all impacts, particularly fatalities (by 7%), with population affected and economic loss increasing by 1% and 4%, respectively (Fig. 3b). For most flood and impact types, the changes due to catchment alteration are in the “as likely as not” range. However, this driver is not applicable to coastal floods and the coastal component of compound floods. The most significant changes observed here are the increase of economic loss for flash floods by 13%, which can be attributed to strong urbanization of catchments in the past 70 years, and the decline in fatalities in riverine floods by 6%, which in turn can be attributed to an increase in reservoir capacity since 1950. Also, a 37% increase in fatalities from flash floods is largely due to a single high-magnitude flood in 1962 (Supplementary Fig. S1b), which doesn’t exceed the minimum hydrological threshold in the counterfactual simulation, thus indicating zero fatalities instead of 815.

Population growth at the regional level has had a much larger effect, increasing fatalities and population affected by 15 and 22%, respectively, while GDP growth has almost quadrupled economic loss (Fig. 3c). For almost all flood and impact types, the likelihood is in the “virtually certain” range, although the increase for coastal and compound floods is much smaller than for riverine and flash floods, again due to the concentration of impacts of the former in the 1950s and 1960s. Further, the effect on population affected has reversed after the 1990s, while for fatalities it has remained stable since the 1980s, as a consequence of the significant slowdown of population growth in Europe (Supplementary Fig. S1c).

More local exposure drivers, such as land use change and economic restructuring, had a smaller influence, although with a high likelihood (Fig. 3d). They contributed to a reduction in fatalities (by 5%) and increases in population affected and economic loss by 1% and 6%, respectively. The largest increases in population affected and economic loss were recorded for riverine floods, with more complex pictures for flash and coastal floods, while the smallest influence has been observed, for all types of impacts, for compound events. Increased economic losses due to changes in land use and the restructuring of the economy are mainly observed since the 1980s (Supplementary Fig. S1d).

The last two drivers have been responsible, with high confidence, for a strong reduction in impacts. Flood protection, which can include a variety of structural measures, reduced impact more for flash floods (except fatalities) than for riverine floods, with the least effect on coastal and compound impacts (Fig. 3e). However, although impacts were increasingly reduced over time until the 1970s, progress seems to have stalled since the 1980s (Supplementary Fig. S1e). Vulnerability reduction has been much stronger and more uniform across flood types (Fig. 3f). In fact, it is the single most important driver of change in fatalities and population affected, both overall and for each flood type. For economic losses, where the vulnerability change has been the strongest, it is only less influential than regional GDP growth (63% reduction compared to 270% increase). The reduction in vulnerability has consistently reduced flood impacts from the 1950s to present. The strongest improvements are observed before 1990, after which the additional gains in flood impact reduction have become smaller (Supplementary Fig. S1f).

Changing flood drivers across regions

The different drivers of flood impacts show large spatial variation across Europe (Fig. 4). While the effect of climate change on flood risk is limited when aggregated at the European scale (Fig. 3a, Supplementary Fig. S1a), country-aggregated results show a clear regional divide (Fig. 4a). In north-western Europe (particularly the United Kingdom, Ireland, and Norway) climate change has considerably increased flood impacts, while in southern Europe (particularly Bulgaria, North Macedonia, Portugal, Serbia, and Spain, and Serbia) the opposite effect is observed. Lower effects of climate change are noticeable in north-eastern Europe (Finland, Latvia, and Lithuania). In between, across western and central Europe, climate change increased flood impacts to varying degrees. The effects are particularly high (with at least very likely confidence) in Hungary, the Netherlands, Poland, Romania, and Switzerland, and low (mostly likely confidence or less) for Austria, Germany, Italy, and Sweden. No clear pattern emerges for catchment alteration (Fig. 4b), with largest increases noticeable for Spain, Greece, and Belgium and decreases in Czechia and Serbia. In most countries, the confidence level is low (Supplementary Fig. S2), which can be explained by divergent trends within individual countries, especially in Spain, France and Italy (Supplementary Fig. S3).

Exposure growth has increased impacts in nearly all countries, with notable exceptions for fatalities and population affected in Germany and Bulgaria (Fig. 4c). However, at the subnational level (Supplementary Fig. S3), population trends can be divergent, with enormous differences, for example, between former West and East Germany, or between coastal and inland regions of France and Spain. The effect of land use change and economic restructuring has no clear spatial distribution, even at the sub-national level, and depends strongly on the type of impact (Fig. 4d). In contrast, flood protection levels have clearly improved more in western and southern Europe than in eastern and northern parts of the continent, where some countries experience deterioration of flood protection (Fig. 4e). The highest confidence in country-scale changes across all impacts (Supplementary Fig. S2) is estimated in the south-west: France, Italy, Spain and Portugal, with much less confidence in central and eastern Europe, where there is also strong variation within larger countries (Supplementary Fig. S3). Vulnerability has broadly declined across the continent (Fig. 4f), with only few exceptions, particularly for the population affected in parts of eastern Europe (Poland, Romania, and Bulgaria).

Our study is the first to disentangle the climatic and human drivers of flood risk across Europe, based on a large number of historical flood events spanning 71 years. It highlights the importance of improved flood protection and vulnerability reduction in mitigating flood losses since 1950. The benefit of improved structural protection is estimated as 14–21% reduction of all flood impacts (and 17–38% in the most recent decade). However, the figure would be higher if floods fully prevented by improved defences were considered, which is beyond the framework of this study, which only covers actual historical floods, i.e. instances where flood protection was inadequate. Still, the extent of those floods would have been much bigger and the impacts more severe than without the improvements to defences made after 1950. Impacts would also have been even stronger without vulnerability reduction, estimated at 39–63% for all floods and 61–75% for those between 2011 and 2020. Several factors have influenced those trends. Early warning systems were developed and improved in the last decades, enabling timely undertaking of emergency measures, including evacuation, sheltering or deployment of temporary flood barriers (35–38). Studies have indicated better societal preparedness after major floods, related to higher awareness (through flood experience or information campaigns) and improved institutional management (through enhanced organisation, planning and civil protection units) (37, 39). Individual preparedness was found to improve as well after floods through the use of various precautionary measures, e.g. building elevation, dry-proofing, or wet-proofing (21, 40). Long-term societal and economic trends also influence vulnerability. For instance, about half of the European dwelling stock was constructed after 1950, while the share of agriculture and industry in the economy has declined from almost half to about 20%, resulting in increased share of housing and services in the fixed asset structure (32). As various assets have different levels of vulnerability (41), such structural changes alone could alter economic vulnerability over time (33). Population affected has experienced the lowest reduction among the impact indicators, as many vulnerability reduction measures and factors do not prevent flooding of homes, but only result in lower probability of fatality and lower economic loss.

Previous studies looking at individual drivers at shorter time scales have observed some different trends than presented here. In particular, the IPCC’s latest assessment report (7) indicates “low confidence” that climate change has influenced riverine and coastal floods in Europe during 1970–2020. By contrast, we detected at least likely contributions of climate change to at least some types of flood impacts for all flood types (Fig. 3a) and almost all countries (Supplementary Fig. S2a). We also found a spatial pattern of climate change contribution (Supplementary Fig. S3a) that is broadly similar to a major study (5) based on hydrological observations for 1960–2010. There, increasing annual maxima of discharges in north-western Europe (attributed to increasing rainfall and soil moisture) is contrasted by a decline in southern and eastern Europe, attributed to decreasing rainfall and increasing evaporation in the south, and decreasing and earlier snowmelt in the east. However, we find that several parts of southern and eastern Europe (e. g. Croatia, Greece, Hungary, Poland, Slovenia, and southern Italy) have experienced higher impacts as a result of climate change, in contrast to downward riverine discharge trends in Blöschl et al. (5). Both limited observational data for some countries and the fact that we translate river discharge into impacts could explain the difference. For instance, discharge trends in most of Poland are driven by the reduction of snowmelt, but reported flood impacts were mostly caused by precipitation-driven summertime discharges (29, 42). Sauer et al. (16), who analysed riverine economic losses during 1980–2010, found a statistically insignificant downward trend in extreme discharge in Europe, in contrast to our finding of a virtually certain increase, albeit over a longer time period (Fig. 3a). This could be at least partially explained by a larger domain in their study, incorporating more countries with snowmelt-driven floods, which have declined due to a warmer climate, and also a much shorter time frame and a very different modelling approach. Although their study also found strong exposure growth, the vulnerability reduction was shown as much less important than in our results. Rentschler et al. (18) showed a slight increase in relative flood hazard in Europe between 1985 and 2019 due to land use change and uneven population growth, which is broadly similar to our study, but with many differences at national and regional level (Fig. 4d). Finally, Mazzoleni et al. (11) also highlighted the strong influence of both climate change and catchment alteration on floods along major European rivers, but without the spatial variation observed in our study.

Finally, our results on the multi-factor attribution of floods during 1950–2020 put into context more recent floods, the attribution of which has been limited to meteorological drivers only (24, 25, 43). In 2021, western and central Europe was heavily affected (44–46), likely making it the costliest year as a share of GDP since 1966 and the most fatal since 1973. On the other hand, 2022 has been one of the least impactful years since 1950, while 2023, despite major impacts in Greece, Italy and Slovenia (44, 47, 48) was likely only average in terms of economic loss relative to GDP. This is because exposure growth will have contributed strongly to the absolute value of losses, creating the impression of exceptional magnitude of impacts, which is less significant in relative terms. Still, the observed vulnerability reduction over past decades also shows that the impacts of those events would likely have been much higher if they had occurred earlier.

Records of historical floods. Data on historical floods are taken primarily from the HANZE (Historical Analysis of Natural HaZards in Europe) database, v2.1 (29). The dataset was constructed with more than 800 sources of information ranging from news media through government reports to scientific papers. It records the date, location (by subnational regions from Paprotny and Mengel (32), based on European Union’s NUTS level 3 classification, v2010 (49)), type and the impact of 2521 floods that caused significant impacts between 1870 and 2020. Of these, 2037 events occurred within this study’s timeframe (1950–2020), and 1504 events were detected in the factual flood catalogue (30), described below. A reported event was considered matched with modelled floods if both hydrological and impact thresholds defined in the catalogue were reached in at least one NUTS3 region that was impacted by a reported event within the same timeframe as that event. Not all events had all impacts statistics available (fatalities, population affected, economic loss). In such situations, impacts were gap-filled with model estimates made in Paprotny et al. (33), also described below. In addition to the HANZE database, 237 floods were identified in Paprotny et al. (30) as significant, based on qualitative or partially quantitative data. Of these, 225 were included in the attribution study, with impacts estimated in Paprotny et al. (30) to complete the dataset of 1729 events. 12 events were excluded as they represented a phase of a flood event from the HANZE database, which was indicated in the model as a separate event, but whose impacts were already included as part of the “main” model event connected with HANZE. In total, our attribution dataset includes an estimated 83% of fatalities, 96% of the population affected and 95% of direct economic losses from flooding in Europe between 1950 and 2020 (Supplementary Table S 2). Due to the resolution of meteorological data and hydrological models, it was not possible to include, in particular, many small or rapid flash floods.

Factual flood data. Each historical event in the study was reconstructed using a set of connected, pan-European models of hazard, exposure and vulnerability. We highlight the main points of the methodology, but refer to the underlying publications on the individual components for technical details.

Climate change (riverine and flash floods). The basis of the riverine flood catalogue is the Hydrological European ReAnalysis (HERA) dataset (31), in which river discharge for Europe was simulated for the period 1950–2020 using a state-of-the-art hydrological model LISFLOOD (50), with spatial resolution of 1 arc minute (1.8 km at the equator) and temporal resolution of six hours. The model was forced with climate reanalysis data (ERA5-Land (51)), bias-corrected and downscaled to the model resolution with weather observations using the ISIMIP3BASD method (52). Extreme river discharge events along river segments with an upstream catchment area of at least 100 km² were identified in the dataset and merged in space and time. Then, based on maximum river discharge at each grid cell, they were combined with a stack of flood hazard maps for different discharge scenarios (53, 54) to derive footprint and water depths of each event. Only regions with sufficient potentially flooded area, and events with sufficient impact potential (based on population and asset value maps³²) were included in the final catalogue of 11,205 riverine floods (30).

Climate change (coastal and compound floods). Similarly to river discharge, the coastal flood catalogue was constructed by combining a set of extreme sea level events with coastal flood maps. To obtain sea levels, a Delft3D-based simulation of storm surges along European coasts (55) was carried out, driven by ERA5 reanalysis data (56). Hourly storm surge heights were combined with data on tidal elevations, significant wave height (converted to wave run-up), long-term sea level rise, mean dynamic topography of the ocean, and glacial isostatic adjustment to estimate the total water level. Based on those water levels, flood hazard maps were computed according to the methodology of Vousdoukas et al. (57) to derive footprints and water depths. The final number of potential coastal events in the catalogue is 2436. In addition, where potential riverine and coastal floods co-occurred in the same NUTS3 region, a separate ‘compound’ event was added to the catalogue (1058 events in total).

Catchment alteration. The hydrological model to derive river discharges for the factual flood catalogue included changes in socio-economic input maps for every year of simulation (31). This accounted for new reservoirs (based on the year of construction), land use changes (rice, other irrigated land, forest, sealed surfaces, open water, and other, i.e. non-irrigated agriculture, non-forest natural and pervious artificial), and water demand changes.

Population and economic growth. This driver represents the changing population and gross domestic product (GDP) at the level of NUTS3 regions. A harmonised dataset of historical statistics, based on almost 400 different data sources, was created as part of the HANZE database (32) spanning the period 1870–2020.

Land use and economic structure. Historical statistics in HANZE were applied to an exposure model that downscales NUTS3-region level changes in land use, population, GDP and fixed asset value to a 100 m grid. The modelling approach combines high-resolution predictor maps for changes in exposure distribution with rule-based and statistical methods (32). The dataset can capture historical changes in the spatial redistribution of exposure due to population and economic growth, land use change (especially through urbanization and building of new infrastructure), change in the structure of the economy (GDP and asset composition) and change in asset-to-GDP ratio per economic sector.

Flood protection levels. In the factual dataset, the locations where flood defences were insufficient due to the return period of the event are identified as NUTS3 regions recorded as impacted in the HANZE database. Other NUTS3 regions with potentially significant flooding identified in the modelled flood catalogue were considered to be protected from flooding.

Flood vulnerability. For each event, within the historical impact zone, a set of four vine-copula models was applied to estimate fatalities, population affected and economic loss. The models were built using data on historical, reported impacts in HANZE converted to relative impacts using potential impact in the modelled flood catalogue (30). A set of hydrological and socioeconomic predictors were derived to predict the changes in flood vulnerability in space and time, and then applied to all events in this study. Where available, reported impacts were used, while any gaps in the data were filled with the model predictions. Two vine-copula models were derived to estimate fatalities, one to derive the probability of any fatalities (as a large share of events had no fatalities) and the other to derive the magnitude of fatalities. However, in most cases (all 1504 events from the HANZE database) the occurrence or non-occurrence of fatalities was known, hence only the second model was applied to estimate the magnitude of impacts. Otherwise, for the remaining 225 events the estimated fatalities integrate the probability of fatality occurrence and estimated magnitude. As one of the parameters in the vulnerability models is the total relative impact of the event, if under a certain counterfactual scenario (climate change, catchment alteration, flood protection level) additional regions are affected beyond the factual ones, the vulnerability estimate under that scenario is used only for those additional regions, and not for the factual impact zone, to maintain a consistent estimate of vulnerability.

Counterfactual flood data. Each event was modelled with counterfactual scenarios, each representing the climatic and socioeconomic conditions of the year 1950. Every combination of factual and counterfactual scenarios (64 in total) were computed, but here we present the scenario in which only the driver in question is set to counterfactual, while all other drivers remain factual. Modelled impacts under all scenarios were scaled by a factor corresponding to the difference between modelled and reported impacts, if such reported data were available in HANZE. However, this adjustment was done only for those regions that were reported as impacted in HANZE. Additional regions impacted under counterfactual scenarios were not adjusted. Also, if no fatalities were reported for the event, the counterfactual model estimates were also not adjusted.

Climate change (riverine and flash floods). River discharge was converted to counterfactual timeseries based on the LISFLOOD run under static 1950 socioeconomic conditions (see ‘Catchment alteration’ below), in which variation of discharge is only due to changes in climate forcing. The detrending of extreme discharges was carried out using the transformed-stationary extreme value analysis (tsEVA) approach (58). The method decouples the detection of non-stationary patterns from the fitting of the extreme value distribution, assuming the generalized Pareto distribution (GPD). It was applied here considering: (a) discharges above 95th percentile (in 6-hourly resolution), (b) minimum distance between flood peaks of 3 days, (c) a moving 30-year time window and (d) the seasonal (monthly) cycle of discharges. The method converted extreme discharges at a given return period (assessed at the time of the event under historical climate) to the discharge for the same return period in 1950.

Climate change (coastal floods). Three components of extreme sea levels were converted to counterfactual. Hourly storm surge height and significant wave height were calculated for 1950 using the tsEVA approach, as for river discharge. However, parameters of the transformation were modified by using a 99th percentile (due to higher temporal resolution and shorter duration of coastal events) and not using seasonal changes (as it is less relevant here and allows more confident estimation of return periods). Additionally, long-term sea level rise after 1950 was removed from the series, with data for 1950–1999 from a high-resolution reconstruction of sea levels (59) and data for 2000–2020 from satellite altimetry (60, 61). The tidal cycle from the FES2014 model (62) and glacial isostatic adjustment from the ICE-6G_C model (63, 64) were preserved in the counterfactual.

Catchment alteration. The riverine discharge simulation was repeated with fixed socioeconomic input maps (land use, reservoirs, water demand) for the year 1950. In case of water demand, the intra-annual (monthly) cycle for the domestic and energy sectors was preserved as it varies depending on the temperature distribution with the year. The time series was scaled proportionally so that the annual totals were reduced to 1950 level.

Population and economic growth, land use and economic structure. Exposure data at regional (NUTS3) and gridded levels for the 1950 timestep were used in the impact calculations.

Flood protection levels. Factual and counterfactual protection level, defined as the probability of the hydrological event causing socioeconomic impacts, was modelled both at the event level and for each NUTS3 region constituting the event’s impact zone. Two vine-copula models were used, based on historical data on flood impact occurrence and non-occurrence within the modelled flood catalogue, and a set of hydrological and socioeconomic predictors (33). An event was considered prevented by defences in the counterfactual scenario, if the mean counterfactual model prediction was ‘no flooding’, while the factual prediction was correctly identified as ‘flooding’. At the NUTS3 level, counterfactual predictions of impacts were only considered for regions that were correctly classified by the model under factual conditions as ‘flooding’ or ‘no flooding’. This is a relatively conservative assumption, but increases the confidence in the results.

Flood vulnerability. The vulnerability models were applied under 1950 socioeconomic conditions and flood experience. For floods with zero reported fatalities, they were possible under counterfactual conditions, if the mean counterfactual prediction of the vine-copula model was ‘fatalities’, but the factual prediction was correctly identified as ‘no fatality’. For floods with reported fatalities, the reverse approach was applied: only floods with correct factual model predictions were allowed to transition to a ‘no fatality’ counterfactual. As for flood protection, this is a relatively conservative assumption, but increases the confidence in the results.

Uncertainty calculation. An uncertainty distribution of each driver of change was estimated using a method appropriate for the underlying modelling approach. When aggregating uncertainties of many floods, e.g. per flood type or country, we assume that the uncertainty of each flood is independent of all others. Consequently, the uncertainty of individual events is higher than for any aggregate (Supplementary Fig. S2).

Climate change, catchment alteration. We assume that the uncertainty in extreme sea levels and river discharge can be represented by a normal distribution with a location parameter of zero. The scale parameter of the distribution is assumed to be equal to the mean absolute error between modelled and observed extreme sea levels and river discharge during events recorded in the factual modelled flood catalogue (30). The mean absolute errors from the validation were averaged per country. Countries with very few data points or no validation data at all were combined with neighbouring countries with more data. River discharge in this analysis was normalized per km² of upstream area. Then, a transformation factor of sea level or discharge into impacts was estimated per NUTS3 region using available flood hazard maps and exposure maps for 2020. In this way, the uncertainty distribution of sea level or discharge can be sampled and then converted into a percent deviation of flood impacts from the model’s deterministic estimate.

Population and economic growth. We assume that the uncertainty of population and GDP can be represented by a normal distribution with a location parameter of zero. The scale parameter of the distribution was assumed to be 2% for population and 3% for GDP per capita in 2020. The value of the parameter for population remains stable for all years, while for GDP it increases by 0.1 pp. per year up to 10% in 1950. The value for population is based on scarce post-enumeration evaluations of population census in developed countries (65–67), which indicate that error of those is usually less than 2%. The 3% value for GDP in 2020 is equivalent to the average revision of historical GDP figures in European countries when transitioning from the ESA 1995 to the ESA 2010 system of national accounts (based on current and archived Eurostat data (32)). The 10% figure is roughly equivalent to the revision of GDP figures for European countries in 1950 between the 2010 and 2023 iterations of the Maddison Project Database of historical GDP estimates (68). It also reflects the lower availability of subnational GDP data in early decades compared to almost complete data in the recent two decades.

Land use and economic structure. The uncertainty was sampled from the uncertainty bounds (nearly uniformly-distributed) of the HANZE exposure model (32). It represents the uncertainty of changes in population distribution due to urbanization, modelled with copulas, and the uncertainty of land-use transitions modelled with a Bayesian Network.

Flood protection levels, flood vulnerability. The uncertainty was sampled from the probabilistic output of the vine-copula models (33). For flood protection levels, the change in impacts due to uncertainty of impacts at the NUTS3 level was assessed using potential impact with depth-damage functions without applying the vine-copula vulnerability models. This was done to avoid reapplying the vine-copula model for each sample of possible flood footprints, which would be too computationally demanding.

Data and materials availability: Results of this analysis for individual events and aggregated to subnational regions can be downloaded from Zenodo (doi:10.5281/zenodo.13149828) and also viewed on https://naturalhazards.eu/. Code to reproduce the analysis in this paper is publicly available on Zenodo (doi:10.5281/zenodo.13150853). The necessary input data for the code are also available on Zenodo (doi:10.5281/zenodo.13150102).

Funding: This research has been supported by the Deutsche Forschungsgemeinschaft (DFG) through project “Decomposition of flood losses by environmental and economic drivers” (FloodDrivers), grant no. 449175973.

Author contributions:

Conceptualization: DP, MM

Methodology: DP, MM, MV

Software: DP, ST, AB

Validation: DP

Formal analysis: DP, ST, AB, MV

Investigation: DP

Data Curation: DP

Writing - Original Draft: DP

Writing - Review & Editing: DP, LF, HK, BM

Visualization: DP

Supervision: MM, KF

Project administration: DP, MM

Funding acquisition: DP

D. Paprotny, A. Sebastian, O. Morales Nápoles, S.N. Jonkman, Trends in flood losses in Europe over the past 150 years. Nat. Commun.9, 1985 (2018).
G. Blöschl, A. Kiss, A. Viglione, M. Barriendos, O. Böhm, R. Brázdil, D. Coeur, G. Demarée, M. C. Llasat, N. Macdonald, D. Retsö, L. Roald, P. Schmocker-Fackel, I. Amorim, M. Bělínová, G. Benito, C. Bertolin, D. Camuffo, D. Cornel, R. Doktor, L. Elleder, S. Enzi, J. C. Garcia, R. Glaser, J. Hall, K. Haslinger, M. Hofstätter, J. Komma, D. Limanówka, D. Lun, A. Panin, J. Parajka, H. Petrić, F. S. Rodrigo, C. Rohr, J. Schönbein, L. Schulte, L. P. Silva, W. H. J. Toonen, P. Valent, J. Waser, O. Wetter, Current European flood-rich period exceptional compared with past 500 years. Nature583, 560–566 (2020).
P. J. Ward, J. C. J. H. Aerts, M. Alberton, W. J. Wouter Botzen, S. Fatica, L. Feyen, R.J.T. Klein, S. Muis, J. Mysiak, D. Paprotny, J. Sillmann, K. van Ginkel, “Large-scale flooding” in European Climate Risk Assessment (Publications Office of the European Union, 2024), pp. 223-236.
H. J. Fowler, G. Lenderink, A. F. Prein, S. Westra, R. P. Allan, N. Ban, R. Barbero, P. Berg, S. Blenkinsop, H. X. Do, S. Guerreiro, J. O. Haerter, E. J. Kendon, E. Lewis, C. Schaer, A. Sharma, G. Villarini, C. Wasko, X. Zhang, Anthropogenic intensification of short-duration rainfall extremes. Nat. Rev. Earth Environ.2, 107–122 (2021).
H. Alifu, Y. Hirabayashi, Y. Imada, H. Shiogama, Enhancement of river flooding due to global warming. Sci. Rep.12, 20687 (2022).
L. Tarasova, D. Lun, R. Merz, G. Blöschl, S. Basso, M. Bertola, A. Miniussi, O. Rakovec, L. Samaniego, S. Thober, R. Kumar, Shifts in flood generation processes exacerbate regional flood anomalies in Europe. Commun. Earth Environ.4, 49 (2023).
Intergovernmental Panel on Climate Change (IPCC), “Europe” in Climate Change 2022: Impacts, Adaptation, and Vulnerability. Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (Cambridge Univ. Press, 2022), pp. 1817-1927.
T. Frederikse, F. Landerer, L. Caron, S. Adhikari, D. Parkes, V. W. Humphrey, S. Dangendorf, P. Hogarth, L. Zanna, L. Cheng, Y.-H. Wu, The causes of sea-level rise since 1900. Nature584, 393–397 (2020).
G. Blöschl, J. Hall, A. Viglione, R. A. P. Perdigão, J. Parajka, B. Merz, D. Lun, B. Arheimer, G. T. Aronica, A. Bilibashi, M. Bohac, O. Bonacci, M. Borga, I. Čanjevac, A. Castellarin, G. B. Chirico, P. Claps, N. Frolova, D. Ganora, L. Gorbachova, A. Gül, J. Hannaford, S. Harrigan, M. Kireeva, A. Kiss, T. R. Kjeldsen, S. Kohnová, J. J. Koskela, O. Ledvinka, N. Macdonald, M. Mavrova-Guirguinova, L. Mediero, R. Merz, P. Molnar, A. Montanari, C. Murphy, M. Osuch, V. Ovcharuk, I. Radevski, J. L. Salinas, E. Sauquet, M. Šraj, J. Szolgay, E. Volpi, D. Wilson, K. Zaimi, N. Živković, Changing climate both increases and decreases European river floods. Nature573, 108–111 (2019).
B. Merz, G. Blöschl, S. Vorogushyn, F. Dottori, J. C. J. H. Aerts, P. Bates, M. Bertola, M. Kemter, H. Kreibich, U. Lall, E. Macdonald, Causes, impacts and patterns of disastrous river floods. Nat. Rev. Earth Environ.2, 592–609 (2021).
M. Mazzoleni, F. Dottori, H. L. Cloke, G. Di Baldassarre, Deciphering human influence on annual maximum flood extent at the global level. Commun. Earth Environ.3, 262 (2022).
Z. W. Kundzewicz, S. Kanae, S. I. Seneviratne, J. Handmer, N. Nicholls, P. Peduzzi, R. Mechler, L. M. Bouwer, N. Arnell, K. Mach, R. Muir-Wood, G. R. Brakenridge, W. Kron, G. Benito, Y. Honda, K. Takahashi, B. Sherstyukov, Flood risk and climate change: global and regional perspectives. Hydrolog. Sci. J.59, 1–28 (2014).
J. Boulange, N. Hanasaki, D. Yamazaki, Y. Pokhrel, Role of dams in reducing global flood exposure under climate change. Nat. Commun.12, 417 (2021).
A. Rajib, Q. Zheng, C. R. Lane, H. E. Golden, J. R. Christensen, I. I. Isibor, K. Johnson, Human alterations of the global floodplains 1992–2019. Sci. Data10, 499 (2023).
J. I. Barredo, Normalised flood losses in Europe: 1970–2006. Nat. Hazards Earth Syst. Sci.9, 97–104 (2009).
I.J., Sauer, R. Reese, C. Otto, T. Geiger, S. N. Willner, B. P. Guillod, D. N. Bresch, K. Frieler, Climate signals in river flood damages emerge under sound regional disaggregation. Nat. Commun. 12, 2128 (2021).
B. Tellman, J. A. Sullivan, C. Kuhn, A. J. Kettner, C. S. Doyle, G. R. Brakenridge, T. A. Erickson, D. A. Slayback, Satellite imaging reveals increased proportion of population exposed to floods. Nature596, 80–86 (2021).
J. Rentschler, P. Avner, M. Marconcini, R. Su, E. Strano, M. Vousdoukas, S. Hallegatte, Global evidence of rapid urban growth in flood zones since 1985. Nature622, 87–92 (2023).
Intergovernmental Panel on Climate Change (IPCC), Special Report on Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation (Cambridge Univ. Press, 2012).
B. Jongman, H. C. Winsemius, J. C. J. H. Aerts, E. Coughlan de Perez, M. K. van Aalst, W. Kron, P. J. Ward, Declining vulnerability to river floods and the global benefits of adaptation. Proc. Natl. Acad. Sci. U.S.A.112, E2271–E2280 (2015).
H. Kreibich, P. Bubeck, M. van Vliet, H. de Moel, A review of damage-reducing measures to manage fluvial flood risks in a changing climate. Mitig. Adapt. Strat. Gl.20, 967–989 (2015).
F. E. L. Otto, K. van der Wiel, G. J. van Oldenborgh, S. Philip, S. F. Kew, P. Uhe, H. Cullen, Climate change increases the probability of heavy rains in Northern England/Southern Scotland like those of storm Desmond—A real-time event attribution revisited. Environ. Res. Lett.13, 24006 (2018).
S. Y. Philip, S. F. Kew, G. J. van Oldenborgh, E. Aalbers, R. Vautard, F. Otto, K. Haustein, F. Habets, R. Singh, Validation of a Rapid Attribution of the May/June 2016 Flood-Inducing Precipitation in France to Climate Change. J. Hydrometeorol. 19, 1881-1898 (2018).
J. S. Tradowsky, S. Y. Philip, F. Kreienkamp, S. F. Kew, P. Lorenz, J. Arrighi, T. Bettmann, S. Caluwaerts, S. C. Chan, L. De Cruz, H. de Vries, N. Demuth, A. Ferrone, E. M. Fischer, H. J. Fowler, K. Goergen, D. Heinrich, Y. Henrichs, F. Kaspar, G. Lenderink, E. Nilson, F. E. L. Otto, F. Ragone, S. I. Seneviratne, R. K. Singh, A. Skålevåg, P. Termonia, L. Thalheimer, M. van Aalst, J. Van den Bergh, H. Van de Vyver, S. Vannitsem, G. Jan van Oldenborgh, B. Van Schaeybroeck, R. Vautard, D. Vonk, N. Wanders, Attribution of the heavy rainfall events leading to severe flooding in Western Europe during July 2021. Clim. Change176, 90 (2023).
B. Merz, V. D. Nguyen, B. Guse, L. Han, X. Guan, O. Rakovec, L. Samaniego, B. Ahrens, S. Vorogushyn, Spatial counterfactuals to explore disastrous flooding. Environ. Res. Lett.19, 044022 (2024).
H. Kreibich, V. Blauhut, J. C. J. H. Aerts, L. M. Bouwer, H. A. J. Van Lanen, A. Mejia, M. Mens, A. F. Van Loon, How to improve attribution of changes in drought and flood impacts. Hydrolog. Sci. J.64, 1–18 (2019).
P. Scussolini, L. N. Luu, S. Philip, W. R. Berghuijs, D. Eilander, J. C. J. H. Aerts, S. F. Kew, G. J. van Oldenborgh, W. H. J. Toonen, J. Volkholz, D. Coumou, Challenges in the attribution of river flood events. WIREs Clim. Change 15, e874 (2024).
K. Frieler, J. Volkholz, S. Lange, J. Schewe, M. Mengel, M. del Rocío Rivas López, C. Otto, C. P. O. Reyer, D. N. Karger, J. T. Malle, S. Treu, C. Menz, J. L. Blanchard, C. S. Harrison, C. M. Petrik, T. D. Eddy, K. Ortega-Cisneros, C. Novaglio, Y. Rousseau, R. A. Watson, C. Stock, X. Liu, R. Heneghan, D. Tittensor, O. Maury, M. Büchner, T. Vogt, T. Wang, F. Sun, I. J. Sauer, J. Koch, I. Vanderkelen, J. Jägermeyr, C. Müller, S. Rabin, J. Klar, I. D. Vega del Valle, G. Lasslop, S. Chadburn, E. Burke, A. Gallego-Sala, N. Smith, J. Chang, S. Hantson, C. Burton, A. Gädeke, F. Li, S. N. Gosling, H. Müller Schmied, F. Hattermann, J. Wang, F. Yao, T. Hickler, R. Marcé, D. Pierson, W. Thiery, D. Mercado-Bettín, R. Ladwig, A. I. Ayala-Zamora, M. Forrest, M. Bechtold, Scenario setup and forcing data for impact model evaluation and impact attribution within the third round of the Inter-Sectoral Model Intercomparison Project (ISIMIP3a). Geosci. Model Dev.17, 1–51 (2024).
D. Paprotny, P. Terefenko, J. Śledziowski, HANZE v2.1: an improved database of flood impacts in Europe from 1870 to 2020. Earth Syst. Sci. Data16, 5145-5170 (2024).
D. Paprotny, B. Rhein, M. I. Vousdoukas, P. Terefenko, F. Dottori, S. Treu, J., Śledziowski, L. Feyen, H. Kreibich, Merging modelled and reported flood impacts in Europe in a combined flood event catalogue, 1950-2020. Hydrol. Earth Syst. Sci.28, 3983–4010 (2024).
A. Tilloy, D. Paprotny, S. Grimaldi, G. Gomes, A. Bianchi, S. Lange, H. Beck, L. Feyen, HERA: a high-resolution pan-European hydrological reanalysis (1950-2020). Preprint at https://dx.doi.org/10.5194/essd-2024-41 (2024).
D. Paprotny, M. Mengel, Population, land use and economic exposure estimates for Europe at 100 m resolution from 1870 to 2020. Sci. Data10, 372 (2023).
D. Paprotny, C. M. P. 't Hart, O. Morales-Napoles, Evolution of flood protection levels and flood vulnerability in Europe since 1950 estimated with vine-copula models. Preprint at https://doi.org/10.21203/rs.3.rs-4213746/v1 (2024).
M. D. Mastrandrea, C. B. Field, T. F. Stocker, O. Edenhofer, K. L. Ebi, D. J. Frame, H. Held, E. Kriegler, K. J. Mach, P. R. Matschoss, G.-K. Plattner, G. W. Yohe, F. W. Zwiers, Guidance Note for Lead Authors of the IPCC Fifth Assessment Report on Consistent Treatment of Uncertainties (IPCC, 2010).
K. T. Lendering, S. N. Jonkman, M. Kok, Effectiveness of emergency measures. J. Flood Risk Manage.9, 320-334 (2016).
G. Formetta, L. Feyen, Empirical evidence of declining global vulnerability to climate-related hazards. Glob. Environ. Change 57, 101920 (2019).
H. Kreibich, A. F. Van Loon, K. Schröter, P. J. Ward, M. Mazzoleni, N. Sairam, G. W. Abeshu, S. Agafonova, A. AghaKouchak, H. Aksoy, C. Alvarez-Garreton, B. Aznar, L. Balkhi, M. H. Barendrecht, S. Biancamaria, L. Bos-Burgering, C. Bradley, Y. Budiyono, W. Buytaert, L. Capewell, H. Carlson, Y. Cavus, A. Couasnon, G. Coxon, I. Daliakopoulos, M. C. de Ruiter, C. Delus, M. Erfurt, G. Esposito, D. François, F. Frappart, J. Freer, N. Frolova, A. K. Gain, M. Grillakis, J. O. Grima, D. A. Guzmán, L. S. Huning, M. Ionita, M. Kharlamov, D. N. Khoi, N. Kieboom, M. Kireeva, A. Koutroulis, W. Lavado-Casimiro, H.-Y. Li, M. C. Llasat, D. Macdonald, J. Mård, H. Mathew-Richards, A. McKenzie, A. Mejia, E. M. Mendiondo, M. Mens, S. Mobini, G. S. Mohor, V. Nagavciuc, T. Ngo-Duc, T. T. N. Huynh, P. T. T. Nhi, O. Petrucci, H. Q. Nguyen, P. Quintana-Seguí, S. Razavi, E. Ridolfi, J. Riegel, M. S. Sadik, E. Savelli, A. Sazonov, S. Sharma, J. Sörensen, F. A. A. Souza, K. Stahl, M. Steinhausen, M. Stoelzle, W. Szalińska, Q. Tang, F. Tian, T. Tokarczyk, C. Tovar, T. V. T. Tran, M. H. J. Van Huijgevoort, M. T. H. van Vliet, S. Vorogushyn, T. Wagener, Y. Wang, D. E. Wendt, E. Wickham, L. Yang, M. Zambrano-Bigiarini, G. Blöschl, G. Di Baldassarre, The challenge of unprecedented floods and droughts in risk management. Nature608, 80–86 (2022).
S. N. Jonkman, A. Curran, L. M. Bouwer, Floods have become less deadly: an analysis of global flood fatalities 1975–2022. Nat. Hazards120, 6327–6342 (2024).
H. Kreibich, G. Di Baldassarre, S. Vorogushyn, J. C. J. H. Aerts, H. Apel, G. T. Aronica, K. Arnbjerg-Nielsen, L. M. Bouwer, P. Bubeck, T. Caloiero, D. T. Chinh, M. Cortès, A. K. Gain, V. Giampá, C. Kuhlicke, Z. W. Kundzewicz, M. C. Llasat, J. Mård, P. Matczak, M. Mazzoleni, D. Molinari, N. V. Dung, O. Petrucci, K. Schröter, K. Slager, A. H. Thieken, P. J. Ward, B. Merz, Adaptation to flood risk: Results of international paired flood event studies. Earths Future5, 953–965 (2017).
J. C. J. H. Aerts, A review of cost estimates for flood adaptation. Water 10, 1646 (2018).
J. Huizinga, H. de Moel, W. Szewczyk, Global Flood Depth-Damage Functions. Methodology and the Database with Guidelines (Publications Office of the European Union, 2017).
G. Blöschl, J. Hall, J. Parajka, R. A. P. Perdigão, B. Merz, B. Arheimer, G. T. Aronica, A. Bilibashi, O. Bonacci, M. Borga, I. Čanjevac, A. Castellarin, G. B. Chirico, P. Claps, K. Fiala, N. Frolova, L. Gorbachova, A. Gül, J. Hannaford, S. Harrigan, M. Kireeva, A. Kiss, T. R. Kjeldsen, S. Kohnová, J. J. Koskela, O. Ledvinka, N. Macdonald, M. Mavrova-Guirguinova, L. Mediero, R. Merz, P. Molnar, A. Montanari, C. Murphy, M. Osuch, V. Ovcharuk, I. Radevski, M. Rogger, J. L. Salinas, E. Sauquet, M. Šraj, J. Szolgay, A. Viglione, E. Volpi, D. Wilson, K. Zaimi, N. Živković, Changing climate shifts timing of European floods. Science357, 588–590 (2017).
World Weather Attribution. Limited net role for climate change in heavy spring rainfall in Emilia-Romagna, https://www.worldweatherattribution.org/limited-net-role-for-climate-change-in-heavy-spring-rainfall-in-emilia-romagna/ (2023).
EM-DAT. Public EM-DAT, https://public.emdat.be/ (2024).
European Commission. Beneficiary States of the EU Solidarity Fund Interventions since 2002 - Only Natural Disasters, https://ec.europa.eu/regional_policy/sources/funding/solidarity-fund/interventions_natural_disaster_2024.pdf (2024).
Eurostat. Climate related economic losses by type of event - values at constant 2022 prices, https://ec.europa.eu/eurostat/databrowser/view/cli_iad_loss/default/table?lang=en&category=cli.cli_iad (2024).
Swiss Re. sigma 01/2024: Natural catastrophes in 2023, https://www.swissre.com/institute/research/sigma-research/sigma-2024-01.html (2024).
Munich Re. Record thunderstorm losses and deadly earthquakes: the natural disasters of 2023, https://www.munichre.com/en/company/media-relations/media-information-and-corporate-news/media-information/2024/natural-disaster-figures-2023.html (2024).
Eurostat. Regions in the European Union 2011. Nomenclature of Territorial Units for Statistics NUTS 2010/EU-27 (Publications Office of the European Union, 2011)
P. Burek, J. van der Knijff, A. De Roo, LISFLOOD - Distributed Water Balance and Flood Simulation Model - Revised User Manual 2013 (Publications Office of the European Union, 2013).
Muñoz-Sabater, E. Dutra, A. Agustí-Panareda, C. Albergel, G. Arduini, G. Balsamo, S. Boussetta, M. Choulga, S. Harrigan, H. Hersbach, B. Martens, D. G. Miralles, M. Piles, N. J. Rodríguez-Fernández, E. Zsoter, C. Buontempo, J.-N. Thépaut, ERA5-Land: A state-of-the-art global reanalysis dataset for land applications. Earth Syst. Sci. Data13, 4349–4383 (2021).
S. Lange, Trend-preserving bias adjustment and statistical downscaling with ISIMIP3BASD (v1.0). Geosci. Model Dev. 12, 3055–3070 (2019).
D. Paprotny, O. Morales-Nápoles, S.N. Jonkman, Efficient pan-European river flood hazard modelling through a combination of statistical and physical models. Nat. Hazards Earth Syst. Sci. 17, 1267–1283 (2017).
F. Dottori, L. Alfieri, A. Bianchi, J. Skoien, P. Salamon, A new dataset of river flood hazard maps for Europe and the Mediterranean Basin. Earth Syst. Sci. Data14, 1549–1569 (2022).
D. Paprotny, O. Morales Nápoles, G. Nikulin, Extreme sea levels under present and future climate: a pan-European database. E3S Web Conf. 7, 02001 (2016).
H. Hersbach, B. Bell, P. Berrisford, S. Hirahara, A. Horányi, J. Muñoz-Sabater, J. Nicolas, C. Peubey, R. Radu, D. Schepers, A. Simmons, C. Soci, S. Abdalla, X. Abellan, G. Balsamo, P. Bechtold, G. Biavati, J. Bidlot, M. Bonavita, G. De Chiara, P. Dahlgren, D. Dee, M. Diamantakis, R. Dragani, J. Flemming, R. Forbes, M. Fuentes, A. Geer, L. Haimberger, S. Healy, R. J. Hogan, E. Hólm, M. Janisková, S. Keeley, P. Laloyaux, P. Lopez, C. Lupu, G. Radnoti, P. de Rosnay, I. Rozum, F. Vamborg, S. Villaume, J.-N. Thépaut, The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc.146, 1999–2049 (2020).
M. I. Vousdoukas, E. Voukouvalas, L. Mentaschi, F. Dottori, A. Giardino, D. Bouziotas, A. Bianchi, P. Salamon, L. Feyen, Developments in large-scale coastal flood hazard mapping. Nat. Hazards Earth Syst. Sci.16, 1841–1853 (2016).
L. Mentaschi, M. Vousdoukas, E. Voukouvalas, L. Sartini, L. Feyen, G. Besio, L. Alfieri, The transformed-stationary approach: a generic and simplified methodology for non-stationary extreme value analysis. Hydrol. Earth Syst. Sci.20, 3527–3547 (2016).
S. Treu, S. Muis, S. Dangendorf, T. Wahl, J. Oelsmann, S. Heinicke, K. Frieler, M. Mengel Reconstruction of hourly coastal water levels and counterfactuals without sea level rise for impact attribution. Earth Syst. Sci. Data16, 1121–1136 (2024).
M.-I. Pujol, Y. Faugère, G. Taburet, S. Dupuy, C. Pelloquin, M. Ablain, N. Picot, DUACS DT2014: the new multi-mission altimeter data set reprocessed over 20 years. Ocean Sci.12, 1067–1090 (2016).
G. Taburet, A. Sanchez-Roman, M. Ballarotta, M.-I. Pujol, J.-F. Legeais, F. Fournier, Y. Faugere, G. Dibarboure, DUACS DT2018: 25 years of reprocessed sea level altimetry products. Ocean Sci.15, 1207–1224 (2019).
F. H. Lyard, D. J. Allain, M. Cancet, L. Carrère, N. Picot, FES2014 global ocean tide atlas: design and performance. Ocean Sci.17,615–649 (2021).
D. F. Argus, W. R. Peltier, R. Drummond, R., A. W. Moore, The Antarctica component of postglacial rebound model ICE-6G_C (VM5a) based upon GPS positioning, exposure age dating of ice thicknesses, and relative sea level histories. Geophys. J. Int.198, 537-563 (2014).
W. R. Peltier, D. F. Argus, R. Drummond, Space geodesy constrains ice-age terminal deglaciation: The global ICE-6G_C (VM5a) model. J. Geophys. Res. Solid Earth120, 450-487 (2015).
Bureau of the Census. World Population 1983. Recent Demographic Estimates for the Countries and Regions of the World (U.S. Department of Commerce, 1983).
Cyprus Statistical Service. Population - Publications, https://www.cystat.gov.cy/en/PublicationList?s=46 (2020).
United Nations Statistics Division. Demographic Yearbook System, https://unstats.un.org/unsd/demographic-social/products/dyb/#statistics (2024).
J. Bolt, J. L. van Zanden, Maddison style estimates of the evolution of the world economy: A new 2023 update. J. Econ. Surv., 1–41 (2024).

The authors declare no competing interests.

SupplementaryInformation.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Attribution of flood impacts shows strong benefits of adaptation in Europe since 1950

Status:

Version 1

Abstract

Figures

Introduction

Results

Historical flood impacts in Europe

Drivers of flood impacts in Europe

Changing flood drivers across regions

Discussion

Materials and Methods

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1