**1. Introduction**

Wildland fires cover a broad range of spatiotemporal scales and are shaped by the complex interaction of fuel, terrain, and meteorological conditions. While scientific understanding of wildland fires and associated smoke plumes are central to the successful mitigation of negative air-quality impacts, the complex and highly dynamic nature of fires presents a challenge for modeling. Existing smoke plume prediction models span a vast range of complexity from simple empirical relations to the more recent coupled fire–atmosphere numerical approaches. Often the choice of model is dictated by the context of its application, subject to the trade-off between fidelity and timely execution.

Large eddy simulation (LES) is a method that uses computational fluid dynamics at a very fine spatial and temporal resolution to simulate a wide range of scales of atmospheric motions down to the size of large turbulent eddies. The Weather Research and Forecasting Model, combined with a semi-empirical fire-spread algorithm (WRF-SFIRE), allows two-way coupling between an LES and a fire behavior model [1–3]. Several studies have examined the ability of WRF-SFIRE to capture the ground-spread behavior of a fire line, near-surface temperatures and winds [1,4,5]. Large-scale simulations of two real fires were carried out by Kochanski et al. [6], comparing modeled plume tops with satellite data. To the authors' best knowledge, very limited consideration has been given to assessing the ability of WRF-SFIRE to simulate wildfire smoke plume dynamics, and vertical rise and distribution of emissions on a local scale. This is the central motivation for this study.

As noted by Mallia et al. [7], there is a general lack of research focusing on modeling the vertical distribution of smoke emissions as a result of wildfires. This knowledge gap can, in part, be explained by the difficulty of constraining potential sources of error in both inputs and models themselves. Until recently, evaluation of coupled fire–atmosphere models required a combination of studies, as no dataset was complete enough to rigorously constrain the problem [1]. Comprehensive field observations were needed to better our understanding of the interactions between fuels, fire behavior and meteorology.

In response, the Prescribed Fire Combustion and Atmospheric Dynamics Research Experiment (RxCADRE) was designed to address this critical research need [8]. The project brought together researchers from a wide range of disciplines to collect data on fuel, meteorology, fire behavior, energy, smoke emissions and fire effects. Simultaneous measurement of multiple fire aspects on the same prescribed burns provided a detailed model evaluation dataset, while also capturing the effects of fire–atmosphere coupling [9]. Data from this comprehensive experiment offers a unique opportunity to assess the accuracy WRF-SFIRE simulated plume rise and dynamics.

In addition, the modeling work we present may help inform future observational studies by identifying key aspects of experimental design. Note that the focus of this work is the evaluation of the LES ability to capture the atmospheric response to a simulated fire of known bulk properties, rather than the fire behavior itself. Effectively, the work aims to validate the relationship between the simulated surface forcing due to a fire and the resultant turbulent convection.

The findings are likely to be of interest for atmospheric and air-quality modelers, as detailed measurements of wildfire smoke plumes are scarce. "Synthetic" plume data from an LES would provide researchers with an alternative resource for validating their models. Therefore, the broad goal of this work is to assess the utility of WRF-SFIRE for improving plume rise and dispersion parameterizations.
