Publication Date

Spring 2023

Degree Type

Thesis

Degree Name

Master of Science (MS)

Department

Meteorology and Climate Science

Advisor

Adam Kochanski

Subject Areas

Atmospheric sciences

Abstract

For coupled fire-atmosphere forecast models, such as WRF-SFIRE, the method used to initialize wildfire within the model can greatly impact meteorological, fire, and smoke extent forecasts. A well-constructed ignition process must be integrated within models to assure the atmospheric component of the model is not destabilized by excessive heat flux released during ignition and realistic fire-induced atmospheric circulation is established. Two ignition methods are tested on the 2020 Creek fire. Analyses indicate a gradual perimeter ignition with selective “masking”, fuel removal, provides the best results in terms of resolving the fire-induced vertical circulation and fire spread, and is thus preferred for use in operational forecasting. Additionally, cyclic model re-initialization leveraging fire observations is known to reduce accumulation of fire propagation errors in operational forecasting, however, integrating fire observations remains challenging. A new fire data assimilation method is described which utilizes infrared perimeters and satellite fire detections to provide critical information needed for forecast initialization. Forecasts for the 2021 Caldor fire were rerun using this new method and evaluated in terms of fire growth, surface PM2.5 concentrations, and vertical smoke extent. Analyses reveal this new method reduced error by 50% for all variables evaluated, with the greatest improvement in forecast skill observed in hours 24-48 of the forecast; thus increasing the skill of medium-range operational WRF-SFIRE forecasts and correcting previously seen overestimation of fire spread.

Download

Share

COinS