Computer simulations to reveal what cannot be measured

Three-dimensional simulation of wildfire plumes in a nutshell

One of the tasks of Wageningen University & Research (WUR) in the EWED project is to generate 3D computer simulations of wildfires that are observed in the field. For this, we use the MicroHH modelopens in new tab, a computer code designed for the simulation of turbulent flows in the atmosphere.

So how does this work? MicroHH can be viewed as a three-dimensional box, with the land surface at the bottom, and its top at a height that is high enough to contain a large wildfire plume. The box is about 10 km wide in the horizontal direction. The model solves the equations that describe atmospheric flow, temperature, and humidity. In addition, we release smoke at the location of the fire plume. We prescribe the initial wind, temperature and moisture in the atmosphere in line with observations, and add boundary conditions (friction, heat and moisture release) in line with the landscape and the daily weather of the location that we study. We represent fires as a very large source of heat and moisture, and release smoke at this specific location.  As the animation below shows, a fire plume will develop.

The nice thing about the simulation of turbulent flows is that all the detail in the simulation is automatically generated by the model, just by prescribing the weather and the fire. The simulation can even produce pyrocumulus clouds, as the animation demonstrates.

Application to the Santa Coloma de Queralt wildfire

Within the EWED project, WUR PhD student Tristan Roelofs has been using this simulation setup to study the wildfire plume dynamics of the Santa Coloma de Queralt wildfire in detail (information and data on this wildfire will be available in the soon-to-be-launched EWED Wildfire Data Portal). His findings are explained in detail in a recently published preprintopens in new tab, and a short summary is explained here.

The first goal was to compare the simulation against the sounding observation. This turned out to be rather challenging as the plume sounding was taken at a moment when the fire plume had already lost some of its intensity displayed earlier in the day (and earlier in our simulation). Nonetheless, we have been able to extract the main structure of the plume in our simulation. In our simulations (d, bottom in the plot), we observe the same structure as in the photo from the field (c, top in the plot), with most notably the downward-moving part. This part is often labelled the "nose" of the plume.

Our simulation revealed many aspects of the three-dimensional structure of the plumes that are hard to observe in the field. Most notably, we discovered that the nose of the plume is part of a complex three-dimensional wind circulation that results from the fire and that causes a strong inward flux into the fire moving against the wind in the head of the fire. This structure, as shown in the plot below, contrasts earlier observations and conceptual descriptions of wildfire plumes. While earlier studies show the convective plume (1) and the descending and accelerating flow upwind, the downwind circulation is often not observed.