Smoke plume from fire Lagrangian simulation: dependence on drag coefficient and resolution

Abstract

Plume rising from wildfire due to the buoyancy generated by the heat released from the fire is a crucial phenomenon to model for a correct description of smoke dispersion within the atmosphere. During the rise, the plume experiences the drag of external air which limits the rising itself. In this work, we investigate the dependence of the hybrid Eulerian-Lagrangian plume rise scheme embedded in the Lagrangian stochastic particle model SPRAY-WEB on the drag coefficient and on the horizontal resolution of the plume rise grid. We test four different drag coefficient models depending on the Reynolds number of the cells as well as a constant drag coefficient. As for the horizontal resolution, we use three different horizontal cell sizes: 200 m, 400 m, and 600 m, namely roughly a quarter, a half, and three-quarters of the source size. We compare the simulation results with the observations taken during a field experiment performed in Idaho organized by the US Environmental Protection Agency, where they collected lidar data and aircraft CO concentration measurements. We found that the drag coefficient influences mainly the plume near the source, where the drag role is more important due to the higher vertical velocities. It has also turned out that the best cell-to-source size ratio for our purpose is one to two.