Project 2 - Atmospheric drivers of extreme floods

Bodo Ahrens, Cristina Primo Ramos and Amelie Krug
Institute for Atmospheric and Environmental Sciences, Goethe University Frankfurt, Frankfurt am Main

Extreme precipitation is the main atmospheric driver of extreme floods. Processes leading to heavy precipitation are controlled by preceding and ongoing weather. They are amplified by local factors such as frontal activity, convective and orographic processes. Furthermore, climate variability controls the weather at synoptic scales by modifying the frequencies of weather types or typical cyclone paths. Therefore, the prerequisites of extreme floods are not only influenced on a local scale but also on a large scale. Hence, the study of atmospheric factors driving to extreme flood events implies the analyses of multi-processes and multiple-scales.

Our research will focus on these aspects. The aim of the SPATE project is to provide a multi-scale and multiprocess point of view, based on robust statistics, to better understand which atmospheric factors lead to extreme flood events, what are possible future flood extremes and what is their predictability. In particular, we will focus on three aspects of the extreme floods events:

leading large-scale atmospheric conditions, involved small scales processes and temporal variability.

1) Large-scale atmospheric conditions leading to extreme floods events:

We would like to identify typical circulation weather types and water source regions involved in extreme flood events. Of special interest are multiplicative events like a Vb cyclone in an already wet season, clusters of cyclone events, or heavy rain over melting snow events. This study will be based on the statistics of these factors obtained from a new centennial data set.

2) Small-scale processes:

We want to know which processes enhance precipitation and its regional and local impact. Therefore, we want to identify which processes change the air mass and fronts on their path to the flood events. It is important to identify the local effects of orography and land surface characteristics (i.e. soil moisture, roughness) on precipitation processes. With convection permitting hindcasts, the relative importance of frontal, orographic and convective precipitation shall be investigated. Afterwards, new simulations will be run, amplifying those factors leading to extreme precipitation.


Fig. 1: Extreme-flood events are controlled by multiple processes on a large (left) and regional (right) scale. Photo from Nora Leps.

3) Temporal variability:

We want to understand and quantify, how and why the large- and local-scale atmospheric factors have changed in the past. For this aim, we want to identify the role of the low-frequency atmospheric variability, like the wintertime northern annular mode (NAM) and the North Atlantic Oscillation (NAO), and their influence on the weather types which lead to extreme floods. Ultimately, the goal is to simulate future extreme flood-producing precipitation events under climate change scenarios and to quantify their predictability.

To provide robust statistics to give answer to the previous three points, we need long meteorological data sets. We will compare observations and atmospheric state. Regarding observations, precipitation data over Germany and Austria will be combined into a consistent and highresolution dataset for the last decades. The atmospheric state over Central Europe will be obtained from dynamically downscaled (≈ 12km) centennial re-analyses using an atmospheric-ocean coupled model. To analyse the temporal component, we will provide with time series of extremeness indices.


Fig. 2 Sketch of the atmospheric-ocean coupled model.