<p>Rather than being isolated events, natural hazards often occur simultaneously or successively, resulting in compounding and cascading impacts. For instance, droughts and heatwaves often occur together (i.e. compound hazards) and trigger secondary hazards such as wildfires (i.e. cascading hazards). Furthermore, their impacts compound (i.e. compound impacts) and propagate through socio-economical systems (i.e. cascading impacts).</p><p>To move from cascading hazards towards cascading impacts, the use of qualitative tools such as narratives, storylines, and cognitive maps have emerged. Still, to predict how the impacts cascade across and within societies, quantitative methods are required.</p><p>Here, a new methodology for quantifying and visualizing drought compound and cascading impacts is presented using the case of the 2018/19 drought in Germany. The use of network inference and data mining tools is proposed to unravel patterns in an existing drought impact dataset (de Brito et al. 2020). Based on a co-occurrence analysis, the strength of compound impact patterns was quantified. Moreover, the most common cascading paths were identified through sequential pattern mining.</p><p>Results demonstrate that the occurrence of compound and cascading drought impacts follow a pattern and do not happen by chance. &#160;Indeed, statistically significant co-occurrence associations outnumbered randomly distributed ones (91.1% versus 8.9%). This has important implications for impact mitigation, suggesting that the understanding of past patterns can help in the prediction of future consequences. Based on this information, efforts can be directed to reduce the initiation of impact interaction networks. Moreover, the visualizations used can support the communication regarding impacts interactions, facilitating a knowledge-driven response by those involved in drought risk management.</p><p>The tools used here can be applied to other hazards. The obtained results can serve help to develop complex models for understanding causalities between drought consequences. They can, for instance, support the development of system dynamics and agent-based models. Hence, instead of using qualitative perceptions, the causal equations would be data-driven. We expect that this work will encourage a more holistic approach to natural hazards impact research.</p><p>&#160;</p><p>de Brito, M.M., Kuhlicke, C., Marx, A. (2020) Near-real-time drought impact assessment: A text mining approach on the 2018/19 drought in Germany. Environmental Research Letters.&#160;doi:10.1088/1748-9326/aba4ca</p>
SUstaiNability: a science communication website on environmental research
