The aim of this study was to develop a wildfire risk model capable of determining wildfire risk in topographically and vegetatively complex region while taking into account both wildfire vulnerability and susceptibility. The model was a sucess at combining vulnerability and susceptibility factors into a final risk model, with the final model being useful at defining high problem areas for wildfires. A highly accurate and more polished version of this model could be useful to wildfire managers when trying to determine areas that require the most attention when it comes to forest management and surveillance for wildfires. 

     The model contains many limitations, with the amount of data available to include in our processing of the final risk model being one of them. The lack of data for percent dead vegetation required us to remove the dead biomass factor from our model. This could have affected the overall accuracy of the model. Further limitations were that the raster cell sizes of the input data required our risk model to be scaled at a coarser resolution then we had initially anticipated. Climate data sets (max annual temperature and precipitation) were only available in cell sizes of 833 meters. Other raster data sets typically consisted of 100-meter cell sizes, making the climate data set our largest cell size by far. This created a coarser suceptability and risk output overall, which decreased the overall precision of the model. Furthermore, the climate data was interpolated, introducing an additional potential for error in areas where there was little to no data collection. All these factors combined contributed to a model that was useful, but far from perfect when trying to asses wildfire susceptability and in turn wildfire risk.

     The model accuracy test revealed approximately 60% accuracy. Although this is lower than we had anticipated, this result does show some promise for future studies in wildfire risk assessment and model design. We hope that this model serves as a basis for future designs, ultimately highlighting the lack of ideal data required to create proficient wildfire models. Higher resolutions and more complete data collections for fuel conditions, such as dead biomass accumulations, would be invaluable for future wildfire models. More accurate factor weighting could also result in a more accurate model overall, but further statistical analyses of the model results is needed to determine this.