Source Document: Richard Taylor, John Forrester, Gunnar Dreßler, Sue Grimmond. 2015-5-31. Developing agent-based models for community resilience. Connecting indicators and interventions. Deliverable 4.4
This deliverable aims at addressing one of the overall emBRACE objectives to model societal resilience through simulation experiments in order to contribute to an understanding of community resilience from a methodological point of view. We must consider this position as a starting point and as an introductory context to this deliverable, before looking at development of modelling and the many different approaches available. This part of the work concerns computer simulation models (agent-based models), and its focus of enquiry is on resilience at the municipality, organisation, or city level. Simulation is just one of a number of modelling approaches taken in the emBRACE project.
Considering uses of modelling, several concepts are important to keep in mind. Firstly, modelling helps the investigator to explore the complexity of the situation where environment is coupled with the social system (and sub-systems to be considered, e.g. geography, community system, policy and institutional systems), and both the modelling process and model outputs can help to clarify and to communicate that understanding. Disaster management situations are often described as complex systems given their characteristic unpredictability, uncertainty, sensitivity to initial conditions and interconnectedness. Examples in the literature are given in the full report. This exploration helps to generate new knowledge; modelling is particularly useful for looking, experimentally, at possible future evolutions of the situation, using simulations.
Secondly, given that considerable complexity is represented, there are further questions about dynamics of complexity which are particularly relevant for us in emBRACE: these are to do with the actual dynamics of social complexity (e.g. cf. McLennan 2003); the interplay between social and natural sciences and engineering involved in DRR planning and responses (e.g. cf. Donaldson et al 2010); and the complexity of our responses to these complex situations (cf. Ramalingam et al 2008). These questions address untangling the factors important for how resilience changes over time. For example, in terms of social resilience: why is one community different from another and how do these differences arise and play out. In terms of individual resilience: how do people adapt differently to different types of interventions, in the long and the short run: and what is the relationship between individual [agent/actor] resilience and community-level resilience. The point is to use simulations and simulation data as an aid (where real data is often scarce), in combination with other methods, to help both researchers and practitioners, and community members themselves, in understanding dynamics, correlations among different factors, and possible causal mechanisms.
Thirdly, it offers an opportunity for integration of different types of knowledge (i.e. technical, traditional, local) and with the participation of different stakeholders, reality-checking and elicitation of preferences. In the best case it allows different actors to “play” with some representations of community resilience, on the basis of including different knowledge frames, to generate shared understandings and co-learning.
This deliverable describes the progress made towards the aim of modelling societal resilience through simulation experiments, and it should be read alongside framework document (in emBRACE Deliverable 6.6 (Jülich, Kruse and Björnsen Gurung 2014) and subsequent developments of the emBRACE framework), for understanding the overall meta-model of emBRACE. However, the ABM work also links to the framework through case studies. Thus, actual realities of each case study’s social and civil Actions; experience and Learning; and natural and social-political Resources & Capacities can be explored within the context existing there, allowing case-study-specific explorations to be carried out within an understanding provided by the generic framework.
However, the ABM method was not used heavily in the emBRACE project. The case study team working on Floods in Central Europe used the method. The case study team working on earthquakes in Turkey commented on another of the models which was found relevant – although data were not available such that direct application could be made. Therefore the full report considers two ABMs, one looking at disaster response in Germany, and another looking at disaster preparation in a more general way but connected with some aspects in the Turkish case study. Other cases considered using ABM: e.g. the London heatwaves case (see Grimmond et al 2014) and the Alpine multi-hazard study (Pedoth et al 2015). These considerations have been integrated into an outlook section in part five of this deliverable.