Flood hazard in a changing climate

In 2008, the CEDIM-project got affiliated with the new joint research project “Flood risk in a changing climate” (“Hochwassergefahr durch Klimawandel”). The participating institutes are IMK-TRO (FZK/Karlsruhe University), IWG (Karlsruhe University), IMK-IFU (FZK) and the hydrology section of GFZ. CEDIM has now access to high competences for the areas of regional climate modeling and hydrological modeling for Eastern Germany, the low mountain range area, and the Alpine region.

The project results have been published in a research report in January 2012.

flood in germany


In the fourth assessment report of the IPCC (2007), it was stated that an increase of the frequency of extreme precipitation events was “very likely” for the near future. This means also an increase in the risk of flooding. The low spatial resolution of the models in the IPCC report leaves however room for a large variability – more so for precipitation than for other variables. This variability has been analyzed in projects such as the EU-project ENSEMBLES, and by IMK-TRO in the project “Herausforderung Klimawandel”. This projects show more details of what changes in future precipitation to expect for Germany, however with large uncertainties. Even though a general decreasing trend in summer time precipitation amount is found, there may still be regional increases in extreme precipitation. For example, the statistical analysis carried out within KLIWA shows that such tendencies are already detectable in observational data. In their studies, the hazard of increasing flood risk and damage for headwaters and small catchments (e.g. in narrow valleys) is shown. So far, there has been no systematic study addressing future changes in the frequency and intensity of extreme precipitation regarding the spatial variability and the impact on discharge and the resulting frequency of flooding. Due to the large computational resources required, there is yet to be conducted a study of the uncertainty in projected future precipitation and discharge amounts over the annual cycle. The reason for the large computational demand is due to the importance of high spatial resolution for the input fields to the hydrological impact models.

Goal and tasks

The aim of this project is to investigate possible changes in extreme precipitation and flooding for small and middle sized river catchments in Germany. The starting point is the global emission scenarios of IPCC (2007), and central goal of the project is to quantify the uncertainty due to different driving data, downscaling method, model resolution and the type of model. Data from global climate models, driven by the IPCC scenarios, will be used to perform high resolution climate simulations. The results will then be validated and used to drive water management models and the WASIM, SWIM and PRMS discharge models. This will be carried out for a control period (1971–2000) as well as for a future scenario period. Due to the complicated processes connecting snow melt, soil moisture, precipitation and discharge, the focus of the studies will be on long time series and extreme value statistics, as opposed to short event based simulations. As the major focus is on small and middle sized catchments, there will be a further focus on convective summer time events, which lead to a relatively short response time in the catchments. Regarding the future scenarios, a compromise is made between the near future (with a smaller uncertainty in the anthropogenic greenhouse gas emissions, but a weaker climate change signal) and the more distant future (with a larger uncertainty in the emissions, but a stronger climate change signal). The period of 2021– 2050 was chosen for this reason, and accounting for the planning horizon of water management related structures. Ensemble simulations with regional climate models and hydrological models will be carried out in order to assess the uncertainties. Two regional climate models (COSMO-CLM at IMK-TRO and WRF at IMK-IFU) will be used to perform high resolution (higher than 10km) project specific simulations, using different driving data and different model configurations. Additionally, openly accessible regional climate simulations can be used (e.g. REMO, CLM-CR, RegCM, HIRHAM). The statistical analysis will be carried out with a common methodology for the participants based on the procedures of the KLIWA project (so that the continuity of the methodology is secured), and with the expertize of IWG. New statistical methods, which allow the uncertainty of the simulated precipitation in a limited resolution to accounted for, will also be developed and applied so that the propagated uncertainty into the discharge can be deduced.

The catchments were chosen according to their size, type (Alpine, low mountain range), damage potential and already known climate responses. On this basis, the Ammer (Alpine), Ruhr and Mulde (low mountain ranges) catchments were chosen. Each catchment will be simulated by at least two hydrological models, using at least two different meteorological data sets (COSMO-CLM from IMK-TRO and WRF from IMK-IFU). The scheme for the hydrological model and catchment area is as follows:

Ammer   x x
Mulde x x  
Ruhr x   x

The project will be carried out in the following order:

  1. Definition of the catchment areas, set-up of the models, gathering of possibly needed data and definition of coupling regions for the regional climate models and hydrological models.
  2. Validation: climate simulations of the control period with COSMO-CLM and WRF.
  3. Validation: hydrological simulations of the control period.
  4. Validation: statistical analysis of step 2 and 3.
  5. Projections: climate simulations of the future scenario with COSMO-CLM and WRF.
  6. Projections: hydrological simulations of the future scenario period.
  7. Projections: statistical analysis of step 5 and 6.
  8. Changes between control and scenario periods.

These steps requires a large work effort, and will take up most of the project time. Therefore the project will strive to at least make the first risk and damage assessments, and then go into this deeper if time and working force is available. Simple flood models could e.g. be used by diploma students to make some first assessments.


Berg, P., Panitz, H.-J., Schädler, G., Feldmann, H., Kottmeier, Ch.   Modelling Regional Climate Change in Germany, in: W.E. Nagel et al. (eds.).   High Performance Computing in Science and Engineering‚10, doi:10.1007/978-3-642-15748-6_34.    
Berg, P., Wagner, S., Kunstmann, H., Schädler, G.   High resolution RCM simulations for Germany: Part I – validation.   submitted to Clim. Dyn.    
Wagner, S., Berg, P., Schädler, G., Kunstmann, H.   High resolution RCM simulations for Germany: Part II – projected climate changes.   submitted to Clim. Dyn.    
Berg, P., Duethmann, D., Liebert, J., Wagner, S.   Uncertainty aspects of changes in flood hazard for medium size river catchments for the near future.   Deutsches Komitee Katastrophenvorsorge e.V. (DKKV), 11. Forum Katastrophenvorsorge.    
Liebert, J., Berg, P., Düthmann, D., Ihringer, J., Kunstmann, H., Merz, B., Ott, I., Schädler, G., Wagner, S.   Wie ändern sich die Charakteristika von Howchwasserereignissen unter dem Klimawandel und mit welchem Unsicherheiten sind solche Aussagen behaftet?   Conference Publication acqua alta 2011 Hamburg, International Conference on Climate Impact, Flood Protection and Hydraulic Engineering, TuTech Verlag, ISBN 978-3-941492-38-7.    
Wagner, S., Berg, P., Düthmann, D., Liebert, J., Ott, I.,Kunstmann, H.   High resolution regional climate simulations for hydrological impact studies in Germany - Hydrological change: Regional hydrological behaviour under transient climate and land use conditions   European Geosciences Union, General Assembly 2011, 3.-8. April 2011, Vienna, Austria, Geophysical Research Abstracts, Vol. 13, EGU 2011-10858.    
Berg, P., Düthmann, D., Feldmann, H., Liebert, J., Wagner, S.   Assessing uncertainties in observations and RCM bias correction, submitted to International Journal of Climatology.   International Journal of Climatology    
Berg, P., Panitz, H.-J., Schädler, G., Feldmann, H., Kottmeier, Ch.   Downscaling climate simulations for use in hydrological modelling of medium-sized river catchments, Editor M. Resch et al.   High performance computing on vector systems 2010, doi:10.1007/978-3-642-11851-7_12.    
Haerter, J.O., Berg, P., Hagemann, S.   Heavy rain intensity distributions on varying time scales and at different temperatures   J. Geophys. Res., vol 115, D17102, doi:10.1029/2009JD013384    
Berg, P., Düthmann, D., Ihringer, J., Kunstmann, H., Liebert, J., Merz, B., Schädler, G., Wagner, S.   The CEDIM-project Flood risk in a changing climate   Geophysical Research Abstracts, Vol. 12, EGU 2010-10789    
Berg, P., Düthmann, D., Ihringer J., Kunstmann, H., Liebert, J., Merz, B., Schädler, G., Wagner, S.   Bias correction of RCM data for use in hydrological catchment modelling   Geophysical Research Abstracts, Vol. 12, EGU 2010-10850    
Berg, P. Haerter, J.O., Hagemann, S.   Changes in extreme precipitation and their dependence on temporal resolution and precipitation classification   Geophysical Research Abstracts, Vol. 12, EGU 2010-10899    
Berg, P., Panitz, H.-J., Schädler, G., and Feldmann, H.   Ensemble-simulations with regional climate models   12th Teraflop workshop 15. March 2010, Stuttgart, Germany    
Haerter, J.O., Berg, P., Hagemann, S.   Is there a timescale where the Clausius-Clapeyron relation describes precipitation rate changes?   Geophysical Research Abstracts, Vol. 12, EGU 2010-10760    
Wagner, S., Berg, P., Düthmann, D., Ihringer, J., Kunstmann, H., Liebert, J., Merz, B., Schädler, G., Werhahn, J.   High resolution regional climate modeling for flood hazard impact study in Germany   AGU Fall Meeting, 13.-17. December 2010, San Francisco