Bruno Merz’s research while affiliated with Universität Potsdam and other places

Hot-wet compound events, the sequential occurrence of humid hot days followed by extreme rainfall, can cause catastrophic consequences, often exceeding the impacts of the isolated occurrence of each event. The urban-coastal microclimate is confounded by complex interactions of land-sea breeze circulations, urban effects of convection and rainfall, and horizontal advection of moisture, which can favor the hot-wet compound occurrence. We present the first observational assessment (1951-2022) of summertime hot-wet compound events across global coastal megacities. We find a significant (P < 0.001) increase in the frequency of hot-wet compound events in both hemispheres: on average,~3 events in the 1950s to 43 events in the 2020s. Cities with upward trends in the frequency of hot-wet compound events are situated < 30 km from coasts, with cities in the southern hemisphere showing faster hot-to-wet transition times (<3 days) than cities in the northern hemisphere. Further, 26 out of 29 sites show increased extreme precipitation, reaching 153%, when humid heat amplitude rises from the 50th to 90th percentiles. Understanding hot-wet compound interactions over the world's coasts is highly relevant for climate change impact assessment and informing climate adaptation. Most megacities are located in the coastal zone, with about 40% of the world's population residing within 100 km of the coast 1. Globally, coastal areas are at increased risk of flooding due to relative sea level rise, land subsidence, and altered storm intensity and frequency 2-4. In addition, the world coastlines are also hotspots of humid heat stress 5. Increased urbanization can alter the wind direction in coastal areas due to changes in the density and height of buildings, which reduces the sea breeze in the fall, whereas increased surface temperature at night in urban areas often leads to decreased land breeze 6. Observations showed a latitudinal pattern of heatwaves over the coasts with a robust increase in severity in the past decades due to increasing air temperature and reduced wind speed, often leading to slower-moving heatwave events, elevating the risk of ecosystem productivity reduction, rising energy consumption, and capacity needs 7. Further, coastal heatwaves are often accompanied by persistent high sea surface temperature , resulting in exposure to high temperature and humidity in cities close to coasts 8. Moreover, globally, coastal precipitation peaks in the boreal summer 9. The superposition of heat stress, humidity, and precipitation may lead to hot-wet compound events-the sequential occurrence of humid hot days and extreme rainfall. Such events can pose a significant threat to coastal communities and cause greater damage than the isolated occurrence of either of these extremes. For instance, a heatwave could massively increase the number of people who need medical assistance and trigger power blackouts 10. In such a vulnerable period, extreme rainfall and flooding could place additional stress on the critical infrastructure 2,11 , for instance, by interrupting traffic and water provision. During summer, temperature and precipitation are generally antic-orrelated over the interior part of the continent but positively correlated over oceans and near the coasts 12,13. Coasts are transition areas where local characteristics also affect the interplay between temperature, humidity, and precipitation. Extreme humid heatwaves often lead to high atmospheric instability and moisture convection, increasing the likelihood of precipitation extremes 14. High atmospheric instability, moisture, and frontal systems jointly mediate rainfall extremes that follow heatwaves 15. There is growing evidence of the co-occurrence of humid heatwaves and extreme precipitation in several regions from observations, such as in China 16-21 , India 14 , Australia 22 , and the USA 23 , and global climate projection scenarios 24-26. These assessments, however, are limited to either smaller spatial domain focusing on particular country 19-21 or coarse-resolution gridded observations [e.g., 0.5°spatial resolution in CRU grid-based observations in Europe 27 and China 21,28 , and 2.5°observational records from the India

Predicting unprecedented floods is essential for disaster risk reduction and climate adaptation but remains a challenge for both hydrological and deep learning models. This study evaluates three hydrological models, a Long Short-Term Memory (LSTM) network, and three hybrid models in simulating extreme floods in more than 400 catchments in Central Europe. The hybrid models integrate hydrological process variables with meteorological inputs to enhance runoff simulations. Results show that the LSTM model outperforms traditional hydrological models, while hybrid models further reduce runoff simulation errors. However, all models tend to underestimate peak discharges, with over 50 % underestimation for unprecedented floods. LSTM-based models exhibit extrapolation limits, likely due to structural and statistical constraints. To improve extrapolation to rare events, future work should integrate physical principles into deep learning, including differentiable hydrological models, physics-guided loss functions, and synthetic extreme event generation. Additionally, regional modeling approaches, such as entity-aware LSTMs, could improve predictions by leveraging spatial hydrological similarities. Combining data-driven learning with physical reasoning will be key to improving flood simulations beyond observed extremes.

The statistical distributions of observed flood peaks often show heavy-tailed behaviour, meaning that extreme floods are more likely to occur than for distributions with an exponentially receding tail. Falsely assuming light-tailed behaviour can lead to an underestimation of extreme floods. Robust estimation of the tail is often hindered due to the limited length of time series. Therefore, a better understanding of the processes controlling the tail behaviour is required. Here, we analyse how the spatial variability of rainfall and runoff generation affects the flood peak tail behaviour in catchments of various sizes. This is done using a model chain consisting of a stochastic weather generator, a conceptual rainfall-runoff model, and a river routing routine. For a large synthetic catchment, long time series of daily rainfall with varying tail behaviours and varying degrees of spatial variability are generated and used as input for the rainfall-runoff model. In this model, the spatial variability and mean depth of a sub-surface storage capacity are varied, affecting how locally or widely saturation excess runoff is triggered. Tail behaviour is characterized by the shape parameter of the generalized extreme value (GEV) distribution. Our analysis shows that smaller catchments tend to have heavier tails than larger catchments. For large catchments especially, the GEV shape parameter of flood peak distributions was found to decrease with increasing spatial rainfall variability. This is most likely linked to attenuating effects in large catchments. No clear effect of the spatial variability of the runoff generation on the tail behaviour was found.

Flood risk models provide important information for disaster planning through estimating flood damage to exposed assets, such as houses. At large scales, computational constraints or data coarseness often lead modelers to aggregate asset data using a single statistic (e.g., the mean) prior to applying non‐linear damage functions. This practice of aggregating inputs to nonlinear functions introduces error and is known as Jensen's inequality; however, the impact of this practice on flood risk models has so far not been investigated. With a Germany‐wide approach, we isolate and compute the error resulting from aggregating four typical concave damage functions under 12 scenarios for flood magnitude and aggregation size. In line with Jensen's 1906 proof, all scenarios result in an overestimate, with the most extreme scenario of a 1 km aggregation for the 500‐year flood risk map yielding a country‐wide average bias of 1.19. Further, we show this bias varies across regions, with one region yielding a bias of 1.58 for this scenario. This work applies Jensen's 1906 proof in a new context to demonstrate that all flood damage models with concave functions will introduce a positive bias when aggregating and that this bias can be significant.

As climatic extremity intensifies, a fundamental rethink is needed to promote the sustainable use of freshwater resources. Both floods and droughts, including water scarcity, are exacerbating declines in river biodiversity and ecosystem services, with consequences for both people and nature. Although this is a global challenge, densely populated regions such as Europe, East Asia and North-America, as well as the regions most affected by climate change, are particularly vulnerable. To date mitigation measures have mainly focused on individual, local-scale targets, often neglecting hydrological connectivity within catchments and interactions among hydrology, biodiversity, climate change and human wellbeing. A comprehensive approach is needed to improve water infiltration, retention and groundwater recharge, thereby mitigating the impacts of heavy rainfall and floods as well as droughts and water scarcity. We propose a holistic catchment-scale framework that combines mitigation measures including conventional civil engineering methods, nature-based solutions and biodiversity conservation actions. This framework integrates legislation, substantial funding and a governance structure that transcends administrative and discipline boundaries, enabling coordinated actions across multiple spatial and temporal scales. It necessitates the collaboration of local and regional stakeholders including local people with scientists and practitioners. A holistic vision for the sustainable management of freshwater resources could have synergistic effects that support biodiversity and mitigate climate change within functional ecosystems that deliver benefits to people.

Flood hazard and risk assessments (FHRAs) and their underlying models form the basis of decisions regarding flood mitigation and climate adaptation measures and are thus imperative for safeguarding communities against the devastating consequences of flood events. In this perspective paper, we discuss how FHRAs should be validated to be fit for purpose in order to optimally support decision-making. We argue that current validation approaches focus on technical issues, with insufficient consideration of the context in which decisions are made. To address this issue, we propose a novel validation framework for FHRAs, structured in a three-level hierarchy: process based, outcome based, and impact based. Our framework adds crucial dimensions to current validation approaches, such as the need to understand the possible impacts on society when the assessment has large errors. It further emphasizes the essential role of stakeholder participation, objectivity, and verifiability in assessing flood hazard and risk. Using the example of flood emergency management, we discuss how the proposed framework can be implemented. Although we have developed the framework for flooding, our ideas are also applicable to assessing risk caused by other types of natural hazards.

We present a novel non-stationary regional weather generator (nsRWG) based on an auto-regressive process and marginal distributions conditioned on climate variables. We use large-scale circulation patterns as a latent variable and regional daily mean temperature as a covariate for marginal precipitation distributions to account for dynamic and thermodynamic changes in the atmosphere, respectively. Circulation patterns are classified using ERA5 reanalysis mean sea level pressure fields. We set up the nsRWG for the central European region using data from the E-OBS dataset, covering major river basins in Germany and riparian countries. The nsRWG is meticulously evaluated, showing good results in reproducing at-site and spatial characteristics of precipitation and temperature. Using time series of circulation patterns and the regional daily mean temperature derived from general circulation models (GCMs), we inform the nsRWG about the projected future climate. In this approach, we utilize GCM output variables, such as pressure and temperature, which are typically more accurately simulated by GCMs than precipitation. In an exemplary application, the nsRWG statistically downscales precipitation from nine selected models from the Coupled Model Intercomparison Project Phase 6 (CMIP6), generating long synthetic but spatially and temporally consistent weather series. The results suggest an increase in extreme precipitation over the German basins, aligning with previous regional analyses. The nsRWG offers a key benefit for hydrological impact studies by providing long-term (thousands of years) consistent synthetic weather data indispensable for the robust estimation of probability changes in hydrologic extremes such as floods.