An attempt at simulating the spread of forest fires using GIS data for elevation, wind, and fuel.
This project is maintained by maxcdietrich
We came into this problem knowing very little about the challenges we would face along the way. We found the acquisition of data and the development of the model to be two separable tasks when figuring out how to split up the model development within our team, which ended up being productive.
Both model development and data acquisition strategies benefitted from peer reviews in class. The map that is the building block of the simulation is cached after its first build to increase the overall efficiency of the program. We also cache information about the relationship between map cells so we don’t have to collect that information on every run through. In terms of our data acquisition, our first guess was to parse a PDF image to extract information about vegetation. However, we found that in general finding the file representing raw data provided much more accurate information.
The intensity and impact of forest fires, especially in the western U.S., have continued to increase over the last several years. According to National Geographic, the average number of fires in the U.S. since 2000 is double the average seen in the 1990’s. On top of that, 2015 saw 10 million acres of land burn due to wildfires.
We don’t have enough faith in the accuracy of this model to suggest it should be used as a tool for those fighting fires. However, it can illustrate the parameters that are most important to the spreading of fires and serve as an educational tool.