Interpreting the socio-technical interactions within a wind damage-artificial neural network model for community resilience

Integrated graph measures reveal survival likelihood for buildings in wildfire events

Wildfire events have resulted in unprecedented social and economic losses worldwide in the last few years. Most studies on reducing wildfire risk to communities focused on modeling wildfire behavior in the wildland to aid in developing fuel reduction and fire suppression strategies. However, minimizing losses in communities and managing risk requires a holistic approach to understanding wildfire behavior that fully integrates the wildland's characteristics and the built environment's features. This complete integration is particularly critical for intermixed communities where the wildland and the built environment coalesce. Community-level wildfire behavior that captures the interaction between the wildland and the built environment, which is necessary for predicting structural damage, has not received sufficient attention. Predicting damage to the built environment is essential in understanding and developing fire mitigation strategies to make communities more resilient to wildfire events. In this study, we use integrated concepts from graph theory to establish a relative vulnerability metric capable of quantifying the survival likelihood of individual buildings within a wildfire-affected region. We test the framework by emulating the damage observed in the historic 2018 Camp Fire and the 2020 Glass Fire. We propose two formulations based on graph centralities to evaluate the vulnerability of buildings relative to each other. We then utilize the relative vulnerability values to determine the damage state of individual buildings. Based on a one-to-one comparison of the calculated and observed damages, the maximum predicted building survival accuracy for the two formulations ranged from [Formula: see text] for the historical wildfires tested. From the results, we observe that the modified random walk formulation can better identify nodes that lie at the extremes on the vulnerability scale. In contrast, the modified degree formulation provides better predictions for nodes with mid-range vulnerability values.

Update article: applicability of artificial neural networks to integrate socio-technical drivers of buildings recovery following extreme wind events

In a companion article, previously published in Royal Society Open Science, the authors used graph theory to evaluate artificial neural network models for potential social and building variables interactions contributing to building wind damage. The results promisingly highlighted the importance of social variables in modelling damage as opposed to the traditional approach of solely considering the physical characteristics of a building. Within this update article, the same methods are used to evaluate two different artificial neural networks for modelling building repair and/or rebuild (recovery) time. By contrast to the damage models, the recovery models (RMs) consider (A) primarily social variables and then (B) introduce structural variables. These two models are then evaluated using centrality and shortest path concepts of graph theory as well as validated against data from the 2011 Joplin tornado. The results of this analysis do not show the same distinctions as were found in the analysis of the damage models from the companion article. The overarching lack of discernible and consistent differences in the RMs suggests that social variables that drive damage are not necessarily contributors to recovery. The differences also serve to reinforce that machine learning methods are best used when the contributing variables are already well understood.