Tracking Wildfire from Space
With conditions just right, fire growth can be explosive, searing everything in its path. Monitoring efforts are extremely challenging and dangerous—expensive piloted aircraft or autonomous drones battling intense winds with no guarantee of safety or success. Space-based fleets provide an alternative, with periodic observations at the global scale, allowing for trusted, routine monitoring of fire growth and intensity across our planet.
In collaboration with colleagues at UC Irvine, I’m a contributing developer of the Fire Events Data Suite, an extensible algorithm capable of automatically monitoring fire growth worldwide. I also serve as the science lead for connecting the complex ways that fire behavior and burn severity are intertwined, allowing for the prediction of the post fire conditions relevant to flash flooding and debris flows—before the fire is even extinguished.
Our fire tracking work has garnered international attention, assisting with disaster mapping in Chile, Greece, and Italy, as well as domestically in Lahaina and Los Angeles. It’s additionally featured in the Smithsonian Museum of Natural History.
See how the algorithm works below:
Previous Work
A Global Assessment of Post-Fire Debris Flows
The world is getting hotter. Wildfires are burning more frequently and for longer. Long after they scorch our planet’s surface, the dirt and rock left behind are easily mobilized by rainfall. On steep slopes, rain carries these remains and incorporates them into turbulent debris flows with the power to take away lives and destroy infrastructure downslope. As part of the Landslide Hazards Group - a subset of the Hydrological Sciences Laboratory at NASA’s Goddard Space Flight Center - I’ve worked to develop a global model for assessing the risk of these debris flows in near real time. You can read about this work here.
Landslide Prediction
What if we could forecast the next landslide? It’s not easy, but science is getting ever closer to doing so.
I’ve previously worked to couple physical and hydrological models to predict hazardous conditions on hillslopes, where I utilized machine learning to assess the probability of a landslide. This information is crucial for residents and governments alike.
My efforts were part of a multi-million dollar grant involving several research and governmental agencies, led by Nobel Laureate, Robert Lempert.
Read more about it here.
The hazards of flooding
With the advent of more extreme storms around our planet, there is a need to provide early warning to communities near flood-prone rivers.
I’ve previously partnered with other researchers at NASA to help implement a deep-learning based tool that takes in real-time river stage and forecasted rainfall to anticipate extreme flood events several hours in advance. This gives community leaders the resources to issues warnings accordingly.