Linking vulnerability and impact of floods in Austria – A case study of the flood events in 2024

In September 2024, a record rainfall of up to 300 to 400 mm, or even more, fell in northeastern Austria just within five days, leading to massive floods that significantly surpassed a 100-year flood event. In the future, climate change will further increase the frequence and intensity of flooding, making the reduction of risk and damage from floods a continuing challenge. Assessing and understanding social, economic, and environmental vulnerability, alongside resilience, is therefore crucial to strengthening the adaptive and mitigation capacities of communities. Vulnerability is defined as a function of sensitivity, susceptibility, and capacity to cope and adapt. From this perspective, vulnerability describes the tendency or predisposition of exposed elements to suffer adverse effects from flooding. It is determined by physical characteristics of buildings and infrastructure, as well as social, economic, institutional, and environmental conditions that influence the capacity of individuals, households, and communities to anticipate, cope with, and recover from floods.

Knowledge of flood hazards and exposure has improved significantly in recent years. However, the assessment of vulnerability remains a major challenge. Detailed insights on municipality level are needed to evaluate and improve current protection measures for residents and mitigation strategies. Therefore, it is important to understand how vulnerability relates to flood impacts not only theoretically but also practically. In this study, we conduct a pre-event vulnerability assessment of Austrian municipalities affected by a major flood event in 2024 and evaluate if lower vulnerability correlates with a lower impact (e.g. fewer affected buildings and infrastructure, lower economic damage), and vice versa.

The exposure, susceptibility, and resilience of affected communities will be analysed to create an indicator-based vulnerability index. Based on a literature review, a set of indicators will be defined, including socio-economic (e.g. age, income), physical (e.g. proximity to rivers, elevation) and other (e.g. accessibility to health services, land use) data. The indicators are normalized and statistically weighted using machine learning techniques, such as regression analysis or random forest. The flood extent will be derived from the Copernicus Sentinel-1 Synthetic Aperture Radar (SAR) satellite data. Geospatial data will be used to obtain for example, accessibility, land use data and statistical data will be used for obtaining socio-economic or demographic information per municipality. Finally, the calculated flood vulnerability index will be evaluated by comparison with observed flood impacts, SAR-derived flood extent, as well as official flood risk maps.

Our findings will improve the understanding of the factors influencing the vulnerability of communities to floods and how vulnerability is linked to the impact of major flood events in Austria. The results can support policymakers in formulating recommendations for those responsible for flood risk management at the municipal level.