Hazardous WildFires| Automated Detection of Hazards
Awards & Nominations
Hazardous WildFires has received the following awards and nominations. Way to go!
Automated Detection of Hazards
Hazardous WildFires - AFTERFIRE
Summary
After the flames on a wildfire are extinguished, there remains immediate dangers to the human environment. Floods, landslides, and soil instability are all major hazards in post-fire areas. To help the forest restore optimally, and protect vulnerable assets in fire-adjacent areas, our project focuses on increasing the efficiency of burned forest hazard mapping. We accomplish this by combining multiple data sources with machine learning to develop a system which will be used to mitigate post-fire hazards.
How We Addressed This Challenge
During this challenge, we focused on what happens after large scale wildfires. A wildfire can have a unique profile, both in its intensity, fuel source, and landscape type. Because of this, post-wildfire areas have a range of fire-initiated properties. Severe soil erosion and lithification can occur after a severe fire burns through trees and understory vegetation. Shrubs, grasses, trees, and the litter layer stabilize the soil, and stems and leaves slow the water to give it time to percolate into the soil profile. Fire can destroy this soil protection, making the area susceptible to rapid degradation and flash flooding. The image below shows what happens to specific elements in the soil at certain soil temperatures.
Our project solves this problem by combining multiple data sources together with machine learning tools in order to estimate soil quality and landscape instability after a fire, providing a crucial resource to landowners, insurance companies, state management agencies, and utilities. The information provided by our tool can be used for forest regeneration, slope stabilization, research, and damage mitigation, as well as policymaking.
The tool will bring together GIS data, satellite data, fire profile data (such as intensity, duration, and pre-fire forest type), and weather prediction data. The machine learning aspect of the project will attempt to look into past fire areas with post-fire damages, and learn what characteristics of the data in those areas contributed to the landscape degradation.
We envision the tool’s ideal use case being that of a municipality learning of an unstable hillside caused by a wildfire that could threaten their community with flooding. The response taken from our tool’s conclusion is to institute a rapid forest management plan that sees a quick effort to plant fast growing and stabilizing plant species on exactly the areas that our tool addressed as hazardous. This policy averts the future instability and leads to a healthy transition back to stable forest. Without our tool, the community might understand the danger posed by flooding or landslides, but might not have the resources to address the entire area. With our tool, they could focus their efforts on only the most necessary areas.
How We Developed This Project
Upon in-depth research, we came across the noteworthy effects of post wildfires in the forests. What served as an inspiration to us was the detrimental effects on the soil after the wildfire has occurred, which could not only cause landslides but even contribute to flooding. We based our approach by analyzing images of an area contributed to wildfires, with respect to both day and night, to investigate whether the post-fire soil had different diurnal temperature properties that would help pinpoint the hazard areas. It brought us further insight into the proportion of damage that might have caused the soil factors.With respect to the dataset, we used NASA MODIS Mod21 and with the help of ArcGIS, we carried out raster processing to the images in order to attain the desired format that can help us analyze the post-wildfire effects on soil.
How We Used Space Agency Data in This Project
To test our first data integration, which was to use diurnal land surface temperature fluctuations, we used the NASA earthdata viewer to choose a MODIS MOD21 tile which covered known burn areas. We selected a cloud free image for both day and night on the same day. With the two examples of the tile, we performed a raster calculation to view the magnitude of diurnal cycles. Our theory posited that heavily scarred soil which had become lithified and impermeable would have a higher magnitude of temperature cycling. We were able to achieve success in raster processing, and we will investigate further to develop robust data on the relationship between lithified soils and temperature fluctuations.
In further development of the process, we will investigate the inclusion of meteosat, SAR, and higher resolution optical payloads.
Project Demo
Data & Resources
Image credits:
NASA Visible Earth
https://visibleearth.nasa.gov/
Data Sources:
Nasa Earth Data portal
https://urs.earthdata.nasa.gov/
Research Articles:
Cannon, S. H., & DeGraff, J. (2009). The increasing wildfire and post-fire debris-flow threat in western USA, and implications for consequences of climate change. In Landslides–disaster risk reduction (pp. 177-190). Springer, Berlin, Heidelberg.
Gluck, M. J., & Rempel, R. S. (1996). Structural characteristics of post-wildfire and clearcut landscapes. In Global to Local: Ecological Land Classification (pp. 435-450). Springer, Dordrecht.
Krasko, V., & Rebennack, S. (2017). Two-stage stochastic mixed-integer nonlinear programming model for post-wildfire debris flow hazard management: Mitigation and emergency evacuation. European Journal of Operational Research, 263(1), 265-282.
Massman, W. J., Frank, J. M., & Mooney, S. J. (2010). Advancing investigation and physical modeling of first-order fire effects on soils. Fire Ecology, 6(1), 36.
Neary, D. G., Ryan, K. C., & DeBano, L. F. (2005). Wildland fire in ecosystems: effects of fire on soils and water. Gen. Tech. Rep. RMRS-GTR-42-vol. 4. Ogden, UT: US Department of Agriculture, Forest Service, Rocky Mountain Research Station. 250 p., 42.
Peppin, D., Fulé, P. Z., Sieg, C. H., Beyers, J. L., & Hunter, M. E. (2010). Post-wildfire seeding in forests of the western United States: an evidence-based review. Forest Ecology and Management, 260(5), 573-586.
Shakesby, R. A. (2011). Post-wildfire soil erosion in the Mediterranean: review and future research directions. Earth-Science Reviews, 105(3-4), 71-100.
Van Leeuwen, W. J., Casady, G. M., Neary, D. G., Bautista, S., Alloza, J. A., Carmel, Y., ... & Orr, B. J. (2010). Monitoring post-wildfire vegetation response with remotely sensed time-series data in Spain, USA and Israel. International Journal of Wildland Fire, 19(1), 75-93.