Issue 7

Smoke Gets in Your Eyes

You’re driving before dawn on a winter day. It’s bad enough to be up so early in the cold, trying to wake up. You smelled smoke when you started out; you know they’ve been burning in the national forest to reduce fuels. You start to notice some shreds of fog: before you know it, you’re inside a thick dirty cloud and can’t see a foot in front of you. The drivers of the cars ahead and behind you are equally blind, all of you driving on in a panicked faith that no one will stop too soon.

This may seem dramatic, but it happens fairly often in the South during the winter fire season, usually for only a few minutes, but sometimes for much longer. In southern Mississippi in 2000, fog and smoke from a small wildfire combined to form a “superfog” on Interstate 10 in the hours just before dawn. Visibility went down to almost zero; the inevitable pileup resulted in 5 fatalities and 24 injuries. Though the smoke in this case came from a wildfire, it could just as easily come from a fire set to improve forest health.

In the South, natural resource managers do most of their prescribed burning in the first 3 months of the year, a time when the needs of human populations and forest ecologies can come into visible—and sometimes deadly—conflict. To reduce the impact of prescribed burns on nearby human populations, SRS scientists have entered the realm of night smoke, haze—and superfog.

Where There's Fire

Prescribed burning—the setting of fires under controlled conditions—is used to treat some 6 to 8 million acres in the South each year, more than in any other part of the United States. About half the acres are burned to improve forest health, the rest for agricultural and range purposes. Southern land managers have long accepted prescribed burning as the most economical way to reduce the risk of wildfires and maintain habitat for fire-dependent plant and animal species. Unfortunately, where there is fire, there is always smoke.

As people move closer to forests, the smoke from controlled burns becomes more problematic. Smoke can cause health problems ranging from irritated eyes and throats to more serious disorders such as asthma, bronchitis, reduced lung function, and even death. At the very least, burning causes a haze that limits visibility and can contribute to poor air quality across the region.

Probably the greatest danger from smoke comes from reduced visibility on roads. Smoke on the road can be hazardous anywhere, but it poses a particular threat in the South, where prescribed burning is done during the winter rainy season when high relative humidity adds to smoke density. When you add in fragmentation from human development, some of the highest road densities in the Nation, and the erratic movement of air across a highly variable terrain, endangering early morning drivers when doing prescribed burning seems almost unavoidable.

“Most smoke-related highway accidents occur just before sunrise when temperatures are coldest and smoke entrapment is maximized,” says Gary Achtemeier, research meteorologist with the SRS Center for Forest Disturbance Science in Athens, GA. “In these conditions, weak drainage winds of even 1 mile per hour can carry smoke over 10 miles during the night—with the density of the road system in the South, there’s a good chance smoke or fog will be carried over a road.”

With fellow SRS research meteorologists Scott Goodrick and Yongqiang Liu, Achtemeier has taken on the task of producing tools managers can use to predict where smoke will drift the day they burn and on into the night and early morning of the next day. The researchers started by developing a computer program that combines high-resolution national weather data with a precise understanding of terrain to predict smoke from fires set at defined coordinates. They tackled the terrain of the Piedmont first, an area where there is significant use of prescribed burning—and where population and road networks are expanding rapidly.

Night Smoke in the Piedmont

The Piedmont, defined as the region that lies between the Coastal Plain and the Appalachian Mountains, includes parts of Virginia, North and South Carolina, Georgia, and Alabama. The terrain of the area is one of gently rolling hills and valleys bisected by numerous streams and rivers and heavily scored by various sizes of roads. “We took on a very specific task, to write a PC-based program to simulate the movement of ground-level smoke over the complex terrain of the Piedmont at night,” says Achtemeier.

Smoke can become a problem at any time during a prescribed burn, but visibility problems occur more frequently in valley bottoms at night. As night falls, air cools rapidly near the ground and wind speeds decline. Smoke begins to accumulate near the ground, especially from smoldering fuels that don’t generate much heat. This ground-hugging smoke is carried through the valleys, accumulating at low points and creating hazards where valley drainages cross roads or bridges. Figuring out exactly where smoke goes involves much more than just assuming it will travel down the terrain of the valleys.

“Under certain weather conditions in the Piedmont, smoke can get trapped in shallow layers of air near the ground at night and get carried to unexpected destinations,” says Achtemeier. “When it gets confined within valleys, smoke can be slow to disperse. When moist conditions are present—and you know how humid it is in the South—this smoke can easily turn into fog.”

Achtemeier stresses that the program he and fellow researchers developed, PB-Piedmont (PB for prescribed burn), is designed strictly to predict the movement of smoke from prescribed burning—not wildfire—under specific conditions. “PB-Piedmont is a wind and particle movement model that provides the numerical ‘eyes’ to ‘see’ where smoke trapped near the ground will go at night. It predicts movement, not concentrations, and addresses problems of complex terrain in areas where ridge and valley height differences are less than 300 feet.”

Taking into account the nature of smoke itself—a phenomenon that hasn’t been studied extensively—added complexity to the model. “We knew that smoke plumes typically diverge and split into neighboring valleys, and that smoke trapped in a valley gradually ‘bleeds’ away as air enters the valley, but we had to figure out how to model the process,” says Achtemeier. “We designed the smoke model so that particles divide into smaller particles, allowing the model to simulate the bleed out from valleys.”

To get at the subtle drift of night smoke, the smoke model was combined with an air flow model developed by Achtemeier that simulates pressure forces that move winds as slow as 4 inches a second. Add to this information about topography and landscape features such as roads, rivers, and streams, then combine it with the most powerful weather data developed so far. What you get is a model that takes up a tremendous amount of computing power and space, much more than a typical natural resource manager would have access to.

How It Works

PB-Piedmont actually runs on weather data supplied through a high performance computer system set up by the Southern High-Resolution Modeling Consortium (SHRMC), a group of State and Federal agencies who joined together to provide the infrastructure needed to run smoke and other models. PB-Piedmont is essentially a “nowcast,” updated hourly with surface weather data as it becomes available over the Web. Computing power through the SHRMC makes it possible to run PB-Piedmont predictively out to 72 hours using MM5, the high-resolution weather data developed by a community of scientists and distributed through the National Center for Atmospheric Research.

Achtemeier realized early on that natural resource managers planning prescribed burns would have very specific needs. The model would have to be small enough to fit on a laptop, run faster than real time to make predictions—and still be powerful enough to model smoke on a fine terrain scale. In addition, the model would have to be simple enough to be run by those with no experience with meteorological modeling.

“Keeping the mathematics simple so the model can run rapidly enough to provide timely predictions is a daunting task,” says Achtemeier. “That’s why we made PB-Piedmont a simplified model designed to run for the specific weather conditions that are associated with smoke entrapment near the ground. You can’t apply it to other conditions.”

The result is a model easily installed on the user’s computer, either from a disk or downloaded from the SHRMC Web site. Two “weather grabbers” are installed with the model: every hour they go to the SHRMC Web site and grab high-resolution weather data. When a manager sits down to plan a burn, the model grabs data for the next 72 hours, plenty of time to track the movement of smoke through critical night and early morning hours—time to decide well in advance whether or not to do the burn.

The movement of smoke predicted by PB-Piedmont has been validated by aircraft video imaging from two experimental night burns, and by nearly 300 ground observations of over 30 prescribed burns. Upgraded several times, the model is now being used to plan prescribed burns by the U.S. Fish and Wildlife Service and the State of South Carolina.

Superfog Revealed

In March 1997, SRS smoke researchers conducted a 2-night experiment in the Talladega National Forest in western Alabama to validate the PB-Piedmont model. To simulate a prescribed burn, they set afire 50 bales of hay soaked in diesel fuel. Once the hay was burning, they put the fire out with water to create a moisture-laden smolder. They also set off 60 smoke bombs—all of this by the light of the full moon.

“The only way to observe an entire smoke plume moving along the ground at night is from the air,” says Achtemeier. “We knew the patterns smoke makes as it scatters from headlights. We wanted to see if we could observe the whole smoke plume by looking at the moonlight scattered from it.”

The site was selected for terrain typical of the Piedmont, safety, and absence of other light sources. The researchers flew over the test fires in a small plane mounted with a light-intensified multispectral video camera, which recorded the formation of smoke on the 2 nights of the experimental burns. Observations from the experiment were nearly identical to results predicted by PB-Piedmont. What the researchers didn’t realize at the time was that they had also recorded the formation of superfog on the first night of the experiment.

Superfog occurs when trapped smoke combines with water vapor at just the right temperature and relative humidity to produce zero visibility. Scientists had long suspected the involvement of smoke in the formation of superfog, but they hadn’t had many opportunities to observe the phenomenon, which comes on quickly late at night and dissipates just as quickly right before sunrise. Achtemeier saw fog form during the 1997 experiment, but didn’t really get the significance of it until he reviewed the video taken that night. He got excited—some would say obsessed—about a phenomenon that some scientists still doubt the existence of.

The various explanations for superfog range from “it’s just dense smoke” to the involvement of hygroscopic smoke particles that attract and bond with water molecules, leading to the formation of water droplets that scatter light. Some explanations leave out the smoke particles and attribute superfog simply to the rapid cooling of the moisture coming off smoldering logs and stumps. Achtemeier decided to take a closer look.

In 2002 and 2003, Achtemeier, with systems analyst Ken Forbus and electronics technician Tim Giddens from the SRS Athens unit, went out to 5 different prescribed burn sites, looking at over 20 individual “smokes” to see if the bulk moisture from smoldering fires alone is enough to trigger superfog. “We’ll have to look at the total moisture budget before we can make conclusions, but our preliminary findings indicate that smoldering could add enough moisture to trigger superfog,” says Achtemeier. “We did find out that on any one site you can have individual smokes that range from very dry to very moist. Though this may seem intuitive, no one has really tried to document it.”

In fall 2003, Achtemeier was out in his backyard raking and burning leaves. He just couldn’t stop thinking about smoke and superfog. When he raked over his burning pile, a dense white smoke formed that didn’t really disperse, but retained its structure while moving away, eventually breaking into patches and disappearing. Looks like superfog, he thought, so he set up an experiment, using the same instruments to measure relative humidity and temperature he used on the individual smokes out in the field. He started burning some leaves in late evening, when the temperature was relatively low. When he got the fire going, his instrument immediately registered 100 percent humidity. “Turns out I had actually produced superfog in my own backyard. I measured the visibility at less than 4 inches.”

(Photo courtesy Gary Achtemeier, USDA Forest Service)

Though Achtemeier is convinced that superfog can be caused by smoke—and a very possible result of prescribed burning—others in the scientific community remain unconvinced. But if he’s right, his models could prevent road accidents related to prescribed burning. “It may be that there are some very specific conditions in which superfog forms, and if we can isolate these conditions, we might be able to pinpoint within 72 hours the days when prescribed burns should not take place.”

For more information:

Gary Achtemeier at 706-559-4239 or gachtemeier@fs.fed.us