Prof. Gollner to Present Wildfire Modeling at NASA’s Goddard Space Flight Center Monday

In 2012, over 4.3 million acres of wildland burned in the United States, including the loss of 4244 structures and 34 firefighter’s lives at a cost to the federal government of over $1.9 billion for suppression alone. Climate change and the growing movement of populations into the wildland/urban interface (WUI) is set to increase the frequency and severity of these fires, challenging researchers to develop solutions to protect property, lives and the environment. One essential tool for both for preventative planning and real-time operational needs is a fire model, providing an accurate means of predicting the spread of a wildland fire. In the early 1970’s, Richard Rothermel and his colleagues at the Missoula Fire Science Lab developed the first operational fire model, using a semi-empirical approach based upon an energy balance of the fire front. This model was calibrated with a significant number of fire spread experiments and, by incorporating properties of the fuel, moisture, wind and slope, a steady spread rate could be calculated for a wide variety of wildland fuels. While this model and its subsequent use in computational tools has been a great leap forward, many of the extreme fire behaviors observed today cannot be modeled with this simplistic steady formulation, leaving significant gaps in our prediction capability, especially during the worst fires.

In order to fill this gap, several research efforts are under way to improve current models of wildland fire spread and a means to circumvent increased accuracy of these models with real-time data. In a joint project with the US Forest Service, recent research has revealed that, despite their large size, wildland fires often spread via small-scale interactions with discrete fuels, dominated by buoyancy-enhanced boundary layer instabilities. Implications from this research towards the current approach to modeling wildland fire spread will be presented. Another avenue being pursued, through a joint project between the University of Maryland and UC San Diego, seeks to develop a cyber infrastructure for real-time sensor-driven modeling of wildland fires. Rather than relying on an increased physical accuracy of models, real-time data is used to improve predictions over time, highlighting a potential new avenue similar to weather forecasting.