Each year, natural hazards globally cause thousands of deaths, affect millions of people, and destruct billions of dollars’ worth of property. Even though Sweden is currently relatively spared from large natural hazard, faster warming in the Arctic makes climate change more pronounced here. This will lead to increased risk of several extreme weather events, altering the frequency and magnitude of what are commonly called natural hazards, such as landslides and wildfires. 

Natural hazards can interact with each other, forming multiple or compound natural hazards. The interactions are context-dependent, which means that knowledge about them should be obtained through context-specific interaction analyses. Despite the indications from climate change research about increased frequency and magnitude for a range of natural hazards in Sweden, there are no previous compiled interaction analyses done for the Swedish environmental conditions. Furthermore, natural hazards and the interactions between them present challenges to decision making and planning withing emergency response systems. In Sweden, the emergency response system is decentralized and complex, involving many actors with statutory responsibility to contribute in response to different natural hazards. As the hazards are expected to be more frequent and larger, it will likely require more coordination and put more pressure on decision making and planning in the response system. This calls for increased attention on natural hazard interactions in a Swedish context, and how the knowledge about them can be utilized to enhance natural hazard preparedness and response. The purpose of this thesis is to investigate how the understanding and modeling of natural hazard interactions can be used to improve natural hazard preparedness and response, with a focus on decision support for emergency response systems. Divided into three research questions, the research investigates: natural hazard interactions in a Swedish context (RQ1), decision support needs in the Swedish emergency response system (RQ2), and computational resource planning tools (RQ3). 

A case study was conducted to identify natural hazard interactions relevant in a Swedish environmental context by applying a qualitative multi-method approach featuring a literature review, a review of agency documentation, and an expert workshop. The main result of the study was a natural hazard interaction framework including 115 interactions classified according to four different interaction types. 

An interview study with a supplementary document analysis was conducted to identify decision support needs to manage natural hazards in the Swedish emergency response system. Through an analysis of collected data using activity theory, eight needs for decision support were identified. 

Based on the identified interactions and decision support needs, two further studies were conducted. The first was a mixed-methods case study, in which an optimization-based decision support tool was developed to determine preparedness locations for aerial fire-suppressing resources. The development adopted a participatory approach by involving the intended end users throughout the process. The evaluation of results from real-time tests and validation meetings indicated that the tool had potential in terms of its relevance, perceived utility, performance, and usefulness. The second study was another mixed-methods case study evaluating the potential of an optimization model for the allocation of flood protection resources. The formulation of the model and the input data were informed by interviews with actors recently involved in flood responses, flood projection data, and an expert review session. The result from the model was then compared with a base case scenario. The model showed potential in terms of its performance and usefulness. 

This dissertation contributes to the natural hazard preparedness and response literature by showcasing how to integrate natural hazard interactions in decision support tools for emergency resource planning. The constructed interaction framework provides a foundation for understanding natural hazard interactions, especially in a Swedish environmental context. Furthermore, the dissertation highlights the importance of anchoring both problems and solutions by involving professionals, experts, and intended end-users in the addressed context. 

Our ability to cope with future hazards and events caused by climate change depends on our emergency preparedness and response capabilities. Inevitably this requires increased attention and further research. More hazard interaction analyses should be conducted on different spatial scales and from different perspectives. More research is needed to further refine the decision support tools presented in this dissertation. One such example is to expand the tool for location of aerial fire-suppressing resources in terms of more input data, such as vegetation types and critical infrastructure.