Probabilities

Probably
EMBYR is a probabilistic model, meaning that each run produces slightly different results. And that is an advantage when modeling something that cannot be completely quantified, Hargrove adds. One of the key determinants of fire spread is firebrands, the burning sticks that are carried in updrafts to ignite nearby trees. Little is known about how far firebrands can spread, he says, so it's impossible to create a model that specifies where they will land. Yet since a model cannot ignore firebrands and concentrate on the easier question of how fast a fire spreads to adjacent trees, the best approach is to estimate the probability of spreading by various mechanisms.

Here's one example of a burn scenario developed by EMBYR. The different colors show simulated severity of the burn, as reflected by heat output per unit time.

Click here to see an EMBYR movie [1.2 MB].

The submitted manuscript has been authored by a contractor of the U.S. Government under contract No. DE-AC05-96OR22464. Accordingly, the U.S. Government retains a nonexclusive, royalty-free license to publish or reproduce the published form of this contribution, or allow others to do so, for U.S. Government purposes.

The model produces results that look reasonable to fire managers, Hargrove says, who know that small changes in conditions can lead to large changes in outcome. "Fire bosses basically think this way. They can easily imagine that if conditions had been slightly different today, things would go quite differently tomorrow."

Weather or not...
Another uncertain element in the fire equation is weather. Let's say tomorrow's weather forecast includes a 20 percent chance of rain. In EMBYR, Hargrove says, "You could run 100 simulations, with rain in 20 of them. Then, you would produce a combined, composite risk map for tomorrow."

Oddly enough, EMBYR was not developed to help fire bosses predict where to deploy their resources, but rather to help anticipate the effect of changing climate on forest fires. The researchers did this by running the model for 1,000 years under dry, normal and wet climates, with intriguing results:

In a dry climate, fires were smaller but more common.
In a wet climate, they were larger but less common.

Strangely, the total acreage burned did not vary a great deal in the three simulations. "The climate change simulations were pretty interesting," says ecologist Monica Turner, who also worked on the project. "Essentially, the risk of very large fires increases under a cooler wet climate, which is counterintuitive. Because drought years still occur, the fires become less frequent, but they're larger. Under warmer drier conditions, fires are more frequent and smaller."

Here's another take on the effort to model forest fires.

A spy-in-the-sky can see Earth burning. How?

There are 1 2 3 4 5 6 7 8 9 10 11 documents. (Glossary | Bibliography)