Wildfires are a major source of air pollutants in the United States, especially in the western US. In the context of global warming, wildfires are expected to be more frequent and severe, and thus of increasing concern to the public and government, especially those in fire-prone regions like western U.S. and wildland-urban interface. Improved representations of fire data through our project could help facilitate understanding of the fire occurrence pattern and help inform decision-making activities related to the mitigation and prediction of wildfire risks.
What it does
In this project, we demonstrated wildfire trends and statistics in a spatially and temporally explicit manner using a comprehensive dataset consisting of 1.88 million wildfires across the contiguous U.S. over the time period 1992–2015. Our goal is to create an interactive wildfire inventory that could be easily queried and displayed.
Accomplishments that I'm proud of
Our visual analysis of the historic wildfire records and novel interactive interface will be a unique tool to engage the viewers to better understand the trends and future projection of wildfires. The predictive power of our machine learning algorithm to identify high-risk areas of future fire activity may facilitate findings that can have important implications for fire prevention and planning.
What I learned
We found a strong correlation between the number of wildfires and the time of the year. The peak season for fires happens around the same time across different years, yet the intensity increases as a result of global warming.
What's next for Mapping wildfire dynamics in the United States
We hope to integrate a machine learning based prediction system assigning likelihood to observed trends and to demonstrate the effect that climate change and human activities have on wildfires.