- Forest fires are a devestating and expensive problem
- Data Science can help
- General container for all wildfire projects
Wildfires In Canada burn vast areas of forests each year. In the 2010s, the average area burned was 2.5 million hectares, the same as 6.2 million football fields. Fighting these fires cost the federal government over $10,000,000,000 (https://www.nrcan.gc.ca/our-natural-resources/forests/wildland-fires-insects-disturbances/forest-fires/13143). As wildfires increase in size and frequency, these costs will only continue to grow. Wildfires are massive polluters. Despite years of progress toward better air quality, the recent increase in Wildfires has erased decades of progress (https://www.theglobeandmail.com/canada/article-wildfires-are-impairing-air-quality-across-western-canada-reversing/). Wildfires can devastate local wildlife, causing massive loss of life to animals and ecosystems. Finally, wildfires have an immense human impact. Losing homes, property and even loved ones is awful. This repository is a holder for all my wildfire Projects. These projects aim to begin conducting further and more impactful analyses.
Currently, all data is from the Canadian National Fire Database (CNFDB https://cwfis.cfs.nrcan.gc.ca/ha/nfdb). The Canadian Government builds and maintains this database and includes access to valuable data. Please read each project's ReadME to ensure you know which data we use.
The general goal of these projects is to fix a problem or identify a challenge to tackle. For example, the alberta_wildfire_spatial_analysis project was my first attempt to understand Alberta Wildfires and see how predictable the size was. The wildfire_normalization project was my first attempt at simple spatial database construction. Finally, alberta_wildfire_forecasts was my attempt to use advanced Neural Networks to forecast fires. Each of the projects has a general goal and reason for it to be solved.
Each analysis has its own limitations and hurdles that need to be addressed and acknowledged individually. However, the most significant issue is the accuracy of historical data. Data collection has improved drastically over time. Many fires went unreported early on; therefore, correctly choosing datasets is essential for generalizable results.
Each analysis has its own limitations and hurdles that need to be addressed and acknowledged individually. However, the most significant issue is the accuracy of historical data. Data collection has improved drastically over time. Many fires went unreported early on; therefore, correctly choosing datasets is essential for generalizable results.
Thanks for reading! If you have questions or want to start working on some projects, please feel free to email me: christopher.ewanik@maine.edu