A New Method for Terrain Representation in Diagnostic Wind Models: Applications for Prescribed Fire Modeling

Prescribed fires are low-risk method of reducing intensity of wildfires. Local wind patterns are strongly influenced in the presence of complex terrain. Hence accurate and robust terrain representations are needed. We present a new method for complex terrain representations that meet fast and accurate operational requirements to create a rigorous, repeatable, and science-based approach. Two challenges in prescribed fire modeling are in representing complex topography at a ‘well-enough’ resolution that balances the multiscale phenomena of the problem while allowing for fast simulation to solution runtime as well as generating starting winds for fire and smoke propagation for use in QUIC-Fire simulations. The key products highlighted here are representations of topography with use of primitive shapes (geomorphons), and early work on generating sample winds for training data using FIRETEC for a physics-informed machine learning method.