Session NH23E-0898: Improving Wildfire Predictability via Machine Learning
Tuesday, 11 December 2018
13:40 – 18:00
Walter E Washington Convention Center – Hall A-C (Poster Hall)
Wildfire risks in a human-natural system continuum have become more concerning in the recent years, especially along wildland-urban interfaces and in densely populated and industrial areas. For example, the 2018 Mendocino Complex Fire has become the largest fire in Californian history; and the 2016 Fort McMurray wildfires scaled down Canadian oil production for two months affecting global oil. Meanwhile, wildfires remain one the least predictable perils due to both their aleatory uncertainty component mainly associated with wildfire ignition triggers, and their epistemic uncertainty component reflecting the lack of knowledge about fire fuel availability and its moisture content, physical setting, weather, and variability in climate. This interdisciplinary session covers topics related to wildfire including drivers/triggers, trends and anomalies, detection and monitoring, risk management and assessment, evaluation of socio-economic and ecological impacts. It also illustrates application of numerical modeling, machine learning, remote sensing, and laboratory or field data in this research field.