Zum Hauptinhalt springen

Project Q11 by Seth Bryant (GFZ Potsdam): Quantifying the contributions of hazard, exposure and vulnerability changes to flood damage trends

Timescale: Oct. 2021 – Sept. 2024


Prof. Bruno Merz, GFZ Potsdam & University of Potsdam

PD Heidi Kreibich, GFZ Potsdam & HU Berlin


Worldwide, flood damages have increased substantially over the last decades. However, it is unclear to which extent the different components of risk have contributed to this increase. Normalization of damage, i.e. accounting for exposure changes, for river floods and other weather-related hazards tends to remove the observed upward trends, suggesting that growth in economic values and population is the main driver of increasing damage. However, such data-based normalization studies do not allow to quantify the effects of changes in hazard and vulnerability on changes in damage, and it is probable that counteracting effects are masked. Methods are needed to be able to attribute changes in damage to the different risk components.   

Objectives and Methods

The objective of this PhD thesis is to

  1. gain generic understanding about the possibilities and limitations for quantifying the contributions of risk components to damage trends, and
  2. attribute the changes in river flood damage for Germany for 1951-2015 to the important drivers.

The first part of the PhD work will investigate typical situations for river flooding (for Germany) in a virtual, simplified world. Typical changes in hazard (climate, land use, reservoirs, river training), changes in exposure (land use, assets), and changes in vulnerability (awareness, preparedness, emergency management) will be combined and the effects on flood damage will be analysed (Metin, et al. 2018). This simplified model will be developed in a Bayesian framework and the reliability of the separation of the different components will be quantified, similar to the flood trend attribution framework of (Viglione, et al. 2016). The emergence timescale for the detection of changing drivers in damage data will be analysed (Crompton, et al. 2011). This is particularly important, since damage time series are highly volatile and uncertain.

Based on the findings of the first part, an attempt will be made to analyse the contributions of the most important drivers to changes in flood damages in Germany for 1951-2015. This analysis will make use of existing models and data bases. The Regional Flood Model implemented for Germany (Falter, et al. 2016) will be run in a ‘constant mode’, representing the situation of 2015. In addition, the most important drivers will be implemented in different ‘change modes’. The different ‘change modes’ will be set up to investigate specific hypotheses, such as ‘effects of climate change are masked by increasing flood protection’. A simulation-based reanalysis of river flood damage for Germany (limited to private households and companies) will be derived and will be compared to existing damage time series (Paprotny, et al. 2018). The comparison between the ‘constant mode’ and the different ‘change modes’ reanalysis runs will help to understand which drivers have dominated changes in flood damage, and whether substantial superposition and masking effects exist.

Expected results

Attribution of temporal changes in disaster damage is a weakly developed scientific field. Both the Bayesian analysis of typical situations and the analysis for Germany will provide methodological progress and novel insights in the possibilities and limitations of attributing observed damage changes to the underlying drivers.

Dedicated Regional Cluster(s): Germany

Links to former and current PhD-projects: Q3 (first cohort, Jonas Laudan), Q4 (first cohort, Tobias Sieg), Q1 (first cohort, Dadiyorto Wendi), Q5 (second cohort, Abhirup Banerjee), Q6 (second cohort, Matthias Kemter), Q7 (second cohort, Lukas Schoppa) and P7 (second cohort, Lisa Berghäuser).


Crompton, Ryan, P., Jr. Pielke, Roger, A. , and K. John McAneney (2011): Emergence timescales for detection of anthropogenic climate change in US tropical cyclone loss data. Environmental Research Letters 6(1):014003.

Falter, D., et al. (2016): Continuous, large-scale simulation model for flood risk assessments: proof-of-concept. Journal of Flood Risk Management 9(1):3-21.

Metin, A. D., et al. (2018): How do changes along the risk chain affect flood risk? Nat. Hazards Earth Syst. Sci. 18(11):3089-3108.

Paprotny, Dominik, et al. (2018): Trends in flood losses in Europe over the past 150 years. Nature Communications 9(1):1985.

Viglione, A., et al. (2016): Attribution of regional flood changes based on scaling fingerprints. Water Resour Res 52(7):5322-5340.