Testing the wildland urban interface fire evacuation tool WUI-NITY: A case study of a rural community

Abstract: This thesis makes use of a freely available platform based on multi-physics simulations, named WUI-NITY, to model WUI fire evacuation. This simulation platform includes three sub-models addressing fire spread, pedestrian movement and traffic movement allowing the representation of their interaction to produce outputs useful for dynamic vulnerability assessment. The goal of the thesis is to test the WUI-NITY platform for a set of WUI fire evacuation scenarios in a rural community by taking into consideration the evacuation component of the model. In addition, the sensitivity of the model results to the input values assigned are investigated. A default scenario is first constructed based on an evacuation drill conducted by the Roxborough Park (Colorado, USA) WUI community. Five variables were selected to test the sensitivity of the platform to model inputs. The values were changed from the default scenario to generate 15 different scenarios adopting the one at a time (OAT) sensitivity analysis approach. The variables considered are total population, response time of the agents, number of available goals as exits, shelter capacity and the activation of a lane reversal order. The simulations run in WUI-NITY were able to capture the expected impact of each variable. For most of the scenarios, the total evacuation time differs by a maximum of 5 minutes. However, for the scenarios with longer response time and the unavailability of Goal F as an exit, the total evacuation time differs by 11 to 40 minutes. The scenarios under consideration included limited congestion on the road. Results from the scenarios considering as variable the available goals and shelter capacity were highly influenced by the road network data. As the evacuation time does not differ much, the evolving condition during the evacuation is of more interest to illustrate how many people are left in vulnerable condition. This is possible to derive from the simulations run in WUI-NITY.