Impact of Forest Canopy Structure on Buoyant Plume Dynamics During Wildland Fires

TL;DR

This work addresses how forest canopy structure governs buoyant plume dynamics from surface fire sources under wind forcing. It employs PALM-LES to systematically compare six canopy configurations (NC, HC, EE, IE, G1, G2) with and without a $5000\,\mathrm{W\cdot m^{-2}}$ heat patch across geostrophic winds of $0$, $5$, and $10\ \mathrm{m\cdot s^{-1}}$, resolving mean flows and turbulence budgets. Key findings show that edges and gaps amplify turbulence (highest TKE in gap canopies) and modify plume tilt, updrafts, and pressure gradients, with buoyancy dominating the TKE production across cases; crosswind speeds reshape these effects, notably suppressing recirculation at higher winds but enhancing momentum transport near plumes at intermediate speeds. The results provide mechanistic insight for fire behavior modeling in heterogeneous forests and can inform management strategies that account for canopy-induced modulation of plume rise and spread.

Abstract

This study investigates the influence of forest canopy heterogeneity on buoyant plume dynamics resulting from surface thermal anomalies representing wildland fires, utilizing Large Eddy Simulation (LES). The Parallelized Large-Eddy Simulation Model (PALM) was employed to simulate six canopy configurations: no canopy, homogeneous canopy, external plume-edge canopy, internal plume-edge canopy, 100 m gap canopy, and 200 m gap canopy. Each configuration was analyzed with and without a static surface heat flux patch of 5000 $\mathrm{W \cdot m^{-2}}$, resulting in a resting buoyant plume. Simulations were conducted under three crosswind speeds: 0, 5, and 10 $\mathrm{m \cdot s^{-1}}$. Results show that canopy structure significantly modifies plume behavior, mean flow, and turbulent kinetic energy (TKE) budgets. Plume updraft speed and tilt varied with canopy configuration and crosswind speed. Pressure gradients associated with plume updrafts were modified based on the canopy configuration, resulting in varying crosswind speed reductions at the plume region. Strong momentum absorption was observed above the canopy for the crosswind cases, with the greatest enhancement in the gap canopies. Momentum injection from below the canopy due to the heat source was also observed, resulting in plume structure modulation based on canopy configuration. TKE was found to be the largest in the gap canopy configurations. TKE budget analysis revealed that buoyant production dominated over shear production. At the center of the heat patch, the gap canopy configurations showed enhanced buoyancy within the gap. Spatial distributions demonstrated increased shear production at the interface between the plume and crosswind.

Figures (20)

• Figure 1: (a) Schematic of LES general setup (left) and leaf area density profile used in the simulations (right). (b) Canopy configurations from the experimental campaign.
• Figure 2: Validation of LES model with field results from dias-junior_large_2015. Velocity profile obtained from the center of the domain averaged over one hour of simulation time.
• Figure 3: Velocity streamlines in the XZ plane at $y=500 \ \mathrm{m}$ for the no plume crosswind cases $\mathrm{U=5 \ m \cdot s^{-1}}$ (top row) and $\mathrm{U=10 \ m \cdot s^{-1}}$ (bottom row).
• Figure 4: Velocity streamlines in the XZ-plane at $y = 500 \ \mathrm{m}$ for the wind speed cases of $\mathrm{U = 0 \ m \cdot s^{-1}}$ (top row), $\mathrm{U = 5 \ m \cdot s^{-1}}$ (middle row), and $\mathrm{U = 10 \ m \cdot s^{-1}}$ (bottom row) for each canopy configuration with plume. The x-axis origin is located at the center of the heat patch, and both the x and z axes are normalized by the canopy height ($h$).
• Figure 5: Profiles of mean flow variables for the no-wind case $(\mathrm{U=0 \ m \cdot s^{-1}})$. Vertical profiles of $\bar{u}$ (a) and $\bar{w}$ (b) at $x/h=0$; horizontal profiles of $\bar{u}$ at $z/h=0.5$ (c) and $z/h=1$ (d); horizontal profiles of $\bar{w}$ at $z/h=1$ (e) and $z/h=5$ (f); vertical profiles of $\bar{p}$ (g); and horizontal profiles of $\bar{\theta}$ at $z/h=1$ (h).