Floods in Germany are becoming increasingly costly. Early warning systems should primarily protect human lives, but also reduce financial losses. Warning systems for floods along rivers in central Germany have been installed as early as 1889. Floods still cause high damage to society and individuals. Scientists from the German Research Centre for Geosciences and the University of Potsdam have now quantitatively investigated which factors are decisive in respect to early warning when it comes to reducing financial losses of floods. Their results can be translated in explicit recommendations for action. The data set on which the researchers were able to draw was unique. It contained detailed information on how flood warnings were received by those affected, what information they contained and what monetary damage was ultimately incurred. An interview with PD Dr. Heidi Kreibich on the recent publication.

PD Dr. Kreibich, what are the basic components of an early warning system for floods?

A flood early warning system is based on knowledge about flood risk and consists of the interacting components of monitoring and forecasting, dissemination of warnings and communication, and the response capacity of the affected society. Different organisations and stakeholders, including the public, must be able to contribute to these components and act accordingly.

What factors have a decisive influence on the resulting financial losses in the event of a flood event?

Important factors are the warning time, which determines how much time there is for emergency measures. Then it is important to know the flood intensity, in particular water levels and flow velocities. Also important is the efficiency of the disaster management and the ability of those affected to take effective emergency measures.

Your current work is based on an extensive data set that contains many details on the damage processes during floods. How did you collect this data base?

Starting with the extreme flood of 2002, we conducted surveys of affected people after all major damaging flood events in Germany to collect details about the damage processes. The standardised questionnaires for all survey campaigns contained about 180 questions. Aspects covered by our questionnaire are the flood hazard, i.e. for example the flooding depth and duration as well as the flow velocity, then we also included questions about the people's experience with floods and their flood awareness. In addition, the questionnaire deals with early warning itself, specific emergency and precautionary measures, but also with building and socio-economic characteristics. Of course, we also used it to record the building and household damage that occurred. Thus, we were able to base our analysis on a unique dataset containing 4,468 claims from six flood events (2002 -2013).

In your current work, you analysed what kind of early warnings the affected people received in the event of a flood, what knowledge they had and how this affected the financial losses of household assets. How did you undertake your analysis?

We quantified the average effect of different flood early warning situations in terms of reducing building and household losses by applying a method called Propensity Score Matching. This method removes confounding effects of other influences on damage, so we were able to determine the causal damage-reducing effect of the following variants: First, people received an early warning with a warning time of at least one hour; second, people received an early warning with a warning time of at least one hour that contained particularly helpful information; or third, people received an early warning with a warning time of at least one hour and additionally stated that they knew what to do.

In a second step, we used a regression model to analyse which factors were associated with people being more likely to know what to do when they received a flood warning.

Among other things, you assessed the quality of the information that the affected people received before the flood. What is really high-quality information for those affected by a flood disaster? What is important?

Clear information about the specific area at risk and the time of occurrence of the flood is very important. Information about dyke or dam breaches, diversions and road closures, as well as necessary evacuations are also important. If there is a warning with sufficient lead time, behavioural tips and recommendations for self-protection are particularly helpful in reducing financial damage. These include, for example, advice to move inventory to higher floors, to lock windows and doors, to use sandbags, or to relocate motor vehicles to higher ground.

What is the most important result of your scientific investigation?

We show that a significant reduction in monetary damages is only achieved if those affected know what to do when they receive a timely flood warning. Thus, our study provides quantitative evidence for the demands of early warning experts to supplement warnings with helpful information and to improve emergency communication. Furthermore, we show that in addition to being supported by helpful warning information, people who have taken precautionary measures and have flood experience are more likely to know what to do when they receive a flood warning than others. Effective risk communication, training and support for private preparedness are therefore doubly helpful; damage reduction through precautionary measures is joined by damage reduction through more effective emergency response.  

Which private emergency measures are particularly effective?

Our research shows that it is potentially easier to reduce household damage than building damage. There are more emergency measures available that limit damage to household contents than those that mitigate damage to buildings, as expensive items can be moved to higher floors, for example. However, protecting the building, for example, if water penetration into the building can be effectively prevented by bulkheads on windows and doors, saves more money on average.

Some private precautionary measures are particularly effective in the event of a flood. Which measures are these?

The long-term precautionary measures of adapted use, such as low-value use of the basement, and adapted interior design, such as using tiles instead of parquet flooring, are particularly effective and can reduce damage by up to 50 percent on average. 

Original study: Kreibich, H, Hudson, P, Merz, B (2021): Knowing What to Do Substantially Improves the Effectiveness of Flood Early Warning. Bulletin of the American Meteorological Society, 102 (7). DOI: doi.org/10.1175/BAMS-D-20-0262.1

Scientific contact:
Priv. Doz. Dr. Heidi Kreibich
Group Leader
Hydrology
Telegrafenberg
14473 Potsdam
Phone: +49 331 288-1550
Email: heidi.kreibich@gfz-potsdam.de

Media contact:
Dipl.-Geogr. Josef Zens
Head Public Relations
Telegrafenberg
14473 Potsdam
Tel.: +49 331 288-1040
E-Mail: josef.zens@gfz-potsdam.de