Validation of WUI-NITY compared to real-world data

Abstract: This thesis compares the wildfire evacuation simulation programme, WUI-NITY, against real-world data by analysing and comparing three wildfire events of the past. The wildfire events were all limited to be within the state of California and was further analysed based on the available quality of data derived of the chosen database. The goal with this thesis was to setup simulations based on the evidence found from the wildfire events with the aim of establishing benchmark values and to address further improvement recommendations towards simulating real-world events. This was performed by doing a total of 24 simulations, eight of each case whereof there was a total of two varying factors, route choice and response curve, meaning there was four route choices and four derived response curves for each wildfire case. The four route choices available was fastest, shortest, random and route depending on evacuation group. For each route choice it was simulated a default response curve which was compared against a derived response curve. Each simulation was compared to the equivalent real-world case to establish the wanted benchmark values and to address further improvement recommendations. For all simulations, no matter which case was studied, it was found that it must always be derived a response curve when replicating a real-world event. Another finding was that even with a proper response curve it may still not mirror the real event in the sense that the population and the number of vehicles used may not match the one of the real-world. This effects the traffic flow and therefore also the time of traffic events to happen (e.g congestions). As to future improvement recommendations it is of interest to implement devices to measure critical outputs such as traffic flow, congestions, and number of vehicles.