Publication Date:
2021-06-27
Description:
Private precaution is an important component in contemporary flood risk management and climate adaptation. However, quantitative knowledge about vulnerability reduction via private precautionary measures is scarce and their effects are hardly considered in loss modeling and risk assessments. However, this is a prerequisite to enable temporally dynamic flood damage and risk modeling, and thus the evaluation of risk management and adaptation strategies. To quantify the average reduction in vulnerability of residential buildings via private precaution empirical vulnerability data (n = 948) is used. Households with and without precautionary measures undertaken before the flood event are classified into treatment and nontreatment groups and matched. Postmatching regression is used to quantify the treatment effect. Additionally, we test state‐of‐the‐art flood loss models regarding their capability to capture this difference in vulnerability. The estimated average treatment effect of implementing private precaution is between 11 and 15 thousand EUR per household, confirming the significant effectiveness of private precautionary measures in reducing flood vulnerability. From all tested flood loss models, the expert Bayesian network‐based model BN‐FLEMOps and the rule‐based loss model FLEMOps perform best in capturing the difference in vulnerability due to private precaution. Thus, the use of such loss models is suggested for flood risk assessments to effectively support evaluations and decision making for adaptable flood risk management.
Description:
Plain Language Summary:
Private precautionary measures such as adapted building use, sealing basements and purchasing flood barriers reduce flood damage to residential buildings. Using an empirical dataset consisting of 948 flooded households in Germany, we estimate that the average loss reducing effect of implementing private precautionary measures is 11‐15 thousand EUR per household. This is approximately equal to 27% of the average building loss suffered by the flooded households (48000 EUR). Despite this significant risk mitigation effect, these precautionary measures are hardly considered in flood risk assessment modelling. This results in biased flood loss predictions being used for evaluating risk management strategies. Hence, we compare state‐of‐the‐art flood loss models in respect to their ability to account for building loss reduction due to private precaution. From all tested flood loss models, the expert Bayesian Network based model BN‐FLEMOps and the rule‐based loss model FLEMOps are best able to capture the damage reducing effect of private precaution. These models can be valuable for evaluating adaptable flood risk management strategies.
Description:
Key Points:
Private precaution significantly reduces the flood vulnerability of private households as shown by robust empirical matching methods
State‐of‐the‐art flood damage models differ strongly based on their ability to capture differences in vulnerability of private households
Methodology applied and validated using an extensive object‐level flood damage data set from Germany
Description:
European Union
http://dx.doi.org/10.13039/100011102
Keywords:
333.91
;
flood loss
;
average treatment effect
;
matching methods
;
loss models
;
risk analysis
;
adaptation
Type:
article
