Description: Central Europe has experienced a severe drought almost every April for the last 14 years consecutively, driven by record high temperatures, low flows, high evapotranspiration, and high soil moisture deficit. The dynamic of this recent and recurrent mid-spring dryness is not yet understood. Here we show that the period 2007â€``2020 was characterized by a reduction of ~50% of the usual April rainfall amount over large areas in central Europe. The precipitation deficit and the record high temperatures were triggered by a multiyear recurrent high-pressure system centered over the North Sea and northern Germany and a decline in the temperature gradient between the Arctic region and the mid-latitudes, which diverted the Atlantic storm tracks northward. From a long-term perspective, the precipitation, temperature, and soil moisture anomalies observed over the last 14 years have reached the highest amplitudes over the observational record. Our study provides an in-depth analysis of the hydroclimate extremes in central Europe over the last 140 years and their atmospheric drivers, enabling us to increase our dynamical understating of long-term dry periods, which is vital to enhance forecasting and mitigation of such events.

Description: The objective of this study is to validate actual evapotranspiration (Aet) simulation (1982 – 2011) over three agro-climatic regions (Sahel, Savannah, and Guinea) within Nigeria. For this investigation, the mesoscale hydrologic model (mHM) was calibrated for streamflow simulation using CHIRPS and ERA5 rainfall datasets in three basins (Jamaare, Hadejia, and Kaduna) in Nigeria. Model parameter sets obtained from these simulations were each used to set up mHM for Aet simulation over Nigeria and the three regions. Model results were compared against estimates from Aet-based products (GLEAM3.5a and FLUXNET). deseasonalized monthly Aet time series plots for all agro-climatic regions and the whole Nigeria domain in comparison to both gridded datasets (GLEAM3.5a and FLUXNET) gave satisfactory correlation scores, especially with the GLEAM3.5a. In the Savannah region, all parameter sets gave satisfactory correlation values (r 〉 0.5) except that from the Kaduna basin forced with ERA5. Results in the Sahel for all model setups are acceptable (r 〉 0.5) but poorly correlated (r 〈 0.5) with both observed datasets in the Guinea region except when using parameters from Jamaare and Hadejia driven with ERA5 and when using model default parameters. For the whole Nigeria domain, simulated Aet results showed satisfactory performance ( r 〉 0.5) mostly in comparison with the GLEAM3.5a, both for all chirps-driven datasets and for the model default setup. This study demonstrates the potential of using the GLEAM3.5a product for water resources modeling in data-scarce locations such as Nigeria.

Description: The disastrous consequences of the July 2021 flood in Western Europe have again demonstrated that current flood risk management is too strongly focused on design events. For instance, the 100-year flood is often used as the target safety level for flood defense, and events beyond such design scenarios are neglected. This disregard of ‘High-Impact / Low-Probability’ (HILP) events would not be advisable in a stationary system, but is even more inappropriate given the widespread climatic, environmental and socio-economic changes. We discuss methods to develop HILP flood scenarios, such as downward counterfactuals and perfect storms. Taking the Ahr catchment as an example, which experienced massive destruction and more than 130 fatalities during the 2021 flood, we demonstrate how a flood risk model chain can be used to develop HILP scenarios. The model chain, consisting of hydrological, hydrodynamic and damage models, is driven by a stochastic weather generator, which generates a very long time series of synthetic weather. In combination with modifying processes along the model chain, for instance, assuming failure of early warning, this setup allows (1) understanding how HILP events could evolve, and (2) generating a large range of flood scenarios beyond design events.

Description: Extreme floods are evoked by a variety of factors, including high precipitation volume and/or intensities and specific pre-event conditions in the catchments. Due to that, each extreme regional flood is characterized by event-specific spatio-temporal variability in flood-inducing characteristics. Thus, the flood conditions vary along the course of large rivers and could be higher/lower, in particular after large confluences.In this study, we analyse event characteristics of extreme floods along the Danube river basin. In order to extend the limited number of extreme flood events in historical data, we apply the mesoscale Hydrological Model mHM for the region. mHM is calibrated and validated to a large set of about a hundred gauges in Germany with a focus on peak flow. The model is fed by the stochastic regional weather generator RWG which allows the generation of long synthetic weather data. Through simulations of these two combined models, a set of 60,000 years of synthetic stream flow data series is derived at several gauges. In the next step, a large number of extreme flood events with different characteristics are derived. This event-rich dataset is then used for the analysis to enhance the understanding of large regional floods and to trace the flood characteristics back to the meteorological and catchment event conditions in the catchment.

Description: The Central Highlands of Vietnam, located in an area susceptible to extreme weather conditions brought on by El Niño, has endured severe droughts during the dry seasons, leading to significant environmental and socio-economic consequences in recent decades. This study, conducted as part of the Drought-ADAPT project (https://www.bmbf-client.de/en/projects/drought-adapt), aims to assess the regional water resources with a focus on both surface water and groundwater for the present period and future projection scenarios. Therefore, the grid-based mesoscale Hydrological Model was implemented for the region encompassing two major river basins Sesan and Seprok, tributaries of the Mekong. The model, mHM-2S, that was driven by daily weather data from either ERA5 reanalysis dataset or the regional climate model REMO, considered various dominant hydrological processes and the impact of multiple large hydroelectric reservoirs in the area. The calibration of mHM-2S was performed with a multi-objective approach, fitting not only observed stream flow at several gauges but also aligning groundwater storage information derived from the GRACE/G3P gravity-based groundwater product. The refined mHM-2S was then used to estimate the regional water resources in terms of means and variations of seasonal discharge, total and groundwater storages under future climate forcings, land-use changes as well as various reservoir operational plans.

Description: The July 2021 devastating floods in Central Europe, particularly in the river catchments in Western Germany, have resulted in a large number of death toll and vast economic damage. The BMBF-funded joint project KAHR (https://hochwasser-kahr.de) deals with the effects of this flood and develops scientific knowledge to assist the reconstruction process in flood-prone area in Western Germany, with a focus on the fast-reacting catchments of Ahr, Erft, and Rur. Therefore, small-scale flood modeling with high spatial and temporal resolution is employed to understand past floods and to develop future flood management strategies in these regions. Here, we apply the mesoscale hydrological model mHM at hourly timesteps and around 1 km spatial resolution for the three flood-prone catchments. We are able to accurately capture the dynamics of the extreme flood events for the recent period including the flood in 2021. To assess the present and future flood risk, a regional weather generator and a disaggregation procedure are applied to generate 10,000 years of synthetic hourly meteorological data. These data are used to force the mHM model to simulate the long time series of river discharge. Major flood events are extracted from this synthetic dataset to investigate the spatial-temporal characteristics of extreme flood events and associated flood risks under future climate conditions.

Description: This dataset comprises event peak flows, representing extreme floods at 516 stations in Germany. The data generation process involves several key steps. Initially, observed rainfall events associated with 10 historical flood disasters from 1950 to 2021 are undergone spatial shifts. These shifts involve three distances (20, 50, and 100 km) and eight directions (North, Northeast, East, Southeast, South, Southwest, West, Northwest), resulting in 24 counterfactual precipitation events. Including the factual (no shift) event, a total of 25 distinct shifting events are considered. Subsequently, these shifted fields are used as atmospheric forcing for a mesoscale hydrological model (mHM) set up and calibrated for the entire Germany. The model produces daily stream flows across its domain, from which the event peak flows are derived. This dataset is expected to provide a valuable resource for analyzing and modeling the dynamics extreme flood events in Germany.

Description: Flood-prone people and decision-makers are often unwilling to discuss and prepare for exceptional events, as such events are hard to perceive and out of experience for most people. Once an exceptional flood occurs, affected people and decision-makers are able to learn from this event. However, this learning is often focussed narrowly on the specific disaster experienced, thus missing an opportunity to explore and prepare for even more severe, or different, events. We propose spatial counterfactual floods as a means to motivate society to discuss exceptional events and suitable risk management strategies. We generate a set of extreme floods across Germany by shifting observed rainfall events in space and then propagating these shifted fields through a flood model. We argue that the storm tracks that caused past floods could have developed several tens of km away from the actual tracks. The set of spatial counterfactual floods generated contains events which are more than twice as severe as the most disastrous flood since 1950 in Germany. Moreover, regions that have been spared from havoc in the past should not feel safe, as they could have been badly hit as well. We propose spatial counterfactuals as a suitable approach to overcome society's unwillingness to think about and prepare for exceptional floods expected to occur more frequently in a warmer world.

Description: Ab Weihnachten 2023 kam es zu einer Hochwassersituation vor allem im Bereich der Flüsse Ems, Weser und Elbe, welche bis in den Januar 2024 hinein zu einer angespannten Lage vor allem in weiten Teilen Nordwestdeutschlands führte. Dieses Winterhochwasser wurde ausgelöst durch großräumige Dauerniederschläge vor allem in Norddeutschland über Weihnachten 2023 (19.-25.12.2023) und verstärkt durch nachfolgende, wenn auch schwächere Niederschlagsereignisse bis Anfang Januar 2024, welche darüber hinaus auf bereits gesättigte Böden trafen. Die Ungewöhnlichkeit des Weihnachtsniederschlagsereignisses bestand in seiner großen räumlichen Ausdehnung und langen zeitlichen Andauer von sieben Tagen. Es hing zusammen mit einer Wetterlage (charakterisiert durch ein ausgedehntes Tiefdrucksystem mit Zentrum über Südskandinavien), welche an sich nicht außergewöhnlich war, jedoch extrem lange andauerte. Das spezifische räumliche Muster dieses einwöchigen Niederschlagsereignisses war auch in der Vergangenheit sowohl mit ergiebigen Dauerniederschlägen als auch mit der zuvor erwähnten Wetterlage assoziiert. Die nachfolgenden Niederschlagsereignisse Ende Dezember und Anfang Januar waren für sich betrachtet wesentlich schwächer. Deren zeitliches Zusammenspiel mit dem vorherigen Weihnachtsereignis führte jedoch dazu, dass gebietsweise über lange Zeiträume von bis zu zweieinhalb Wochen extreme mittlere Niederschlagsintensitäten auftraten. Das Hochwasser als Auswirkung der Niederschläge war ebenfalls im Wesentlichen durch seine große räumliche Ausdehnung charakterisiert, nur vereinzelt wurden extreme Flusspegelstände gemessen. Unter allen Hochwassern in Deutschland seit 1955 (für welche die Abflüsse über zwei Wochen eine Wiederkehrzeit von mindestens 10 Jahren aufwiesen) rangiert die räumliche Ausdehnung des Weihnachtshochwassers 2023 mit gut 100.000 km² auf Platz 9. Die räumliche Ausdehnung ist hierbei nicht die Überflutungsfläche, sondern die Fläche, in der die Hochwasserabflüsse einen bestimmten Schwellenwert überschreiten. Die Überflutungsfläche selbst erreichte eine Ausdehnung von ca. 1000 km² und betraf mehr als 40 Landkreise, vor allem in Niedersachsen und Bremen, aber auch in Teilen Hessens und Nordrhein-Westfalens. Betroffen waren dabei, je nach Abschätzung, 18.000 bis 30.000 Personen, rund 2000 Gebäude, 4,6 km² bebaute Fläche und 470 km Straßen. Die im Dezember 2023 gemessene Monatsniederschlagssumme von 164 mm (über einem besonders von den Niederschlägen betroffenen Gebiet in Niedersachsen: 51,5°N - 53,5°N, 8,0°O - 11,0°O) tritt in den Wintermonaten im heutigen Klima durchschnittlich nur ca. alle 120 Jahre auf. Eine Attributionsstudie des Deutschen Wetterdienstes (DWD) zeigt, dass sich die Wahrscheinlichkeit für ein Ereignis dieser Intensität aufgrund der bisherigen Klimaerwärmung von 1,2°C (seit etwa 1900) um den Faktor 1,8 (Ergebnisspanne: 0,1 bis 140) erhöht hat, und dass sich diese Wahrscheinlichkeit im Falle eines 2°C wärmeren Klimas, d.h. einer zusätzlichen Erwärmung um weitere 0,8°C, nochmals erhöhen wird. Die Studie zeigt jedoch auch, dass diese Abschätzungen mit großen Unsicherheiten verbunden sind. Dennoch sind diese Abschätzungen konsistent mit den Ergebnissen verschiedener Studien, welche sowohl eine Zunahme des mittleren Winterniederschlags zeigen als auch eine Intensivierung extremer Niederschlagsereignisse im nördlichen Mitteleuropa.

Description: Despite considerable advances in flood forecasting during recent decades, state-of-the-art, operational flood early warning systems (FEWS) need to be equipped with near-real-time inundation and impact forecasts and their associated uncertainties. High-resolution, impact-based flood forecasts provide insightful information for better-informed decisions and tailored emergency actions. Valuable information can now be provided to local authorities for risk-based decision-making by utilising high-resolution lead-timemaps and potential impacts to buildings and infrastructures. Here, we demonstrate a comprehensive flood lain inundation hindcast of the 2021 European Summer Flood illustrating these possibilities for better disaster preparedness, offering a 17-hour lead time for informed and advisable actions.