July 24, 2009

The Angora Fire: A series of unfortunate fuel treatment effectiveness tests

Article reviewed: Effects of fuel treatments on fire severity in an area of wildland-urban interface, Angora Fire, Lake Tahoe Basin, California

By H.D. Safford, D.A. Schmidt, and C.H. Carlson. Published in the journal Forest Ecology and Management, Volume 258 pp. 773-787

The plot line: Before the 2007 Angora fire burned through an area near Lake Tahoe, CA, a number of different treatments designed to reduce fire hazard had been implemented. The fire then became an unplanned test of the effectiveness of those treatments. In general, the treatments were very effective at reducing the intensity of the fire. Most treatments reduced surface fuels (litter and debris on the ground) and also reduced the density of standing trees. Where treatments did not treat surface fuels or where they did not reduce tree density of mature trees, the treatments were less effective. The authors point out the importance of extensive fuel treatments AND targeted intensive treatments by landowners surrounding their homes.   

Relevant quote: “Our results show that, in general, fuel treatments within and around the perimeter of the Angora Fire substantially changed fire behavior and subsequent fire effects to forest vegetation.”

Relevance to landowners:

Landowners like me who live in a wildland-urban interface (i.e. “out in the woods”) have an intense interest in how their neighbors are managing their land. In my case, I have federal land on one side and private land on the other side of my property. Neither landowner has treated their land to reduce fire intensity. When a fire does occur, it may burn down my house even though I have treated my own surrounding forest (not as thoroughly as I’d like to, I admit). Likewise, even if the adjacent landowners have treated their land, my home has a high probability of burning down if I haven’t treated the land immediately surrounding my home. There were two especially important points from this article with relevance for landowners like me:

  1.  We need both extensive fuel treatments surrounding wildland-urban interface areas AND intensive targeted treatments immediately surrounding structures.
  2.  It was not enough to do JUST surface fuel treatments (i.e. piling and burning surface litter and debris) or JUST thinning treatments (removing live trees). Treatments were effective where BOTH of these treatments were done.

The authors calculated that the extensive treatments in this case would have had to have been about ¼ mile wide to have been wide enough for suppression crews to control the fire. This was under severe fire weather conditions, but also in a scenario that had very rapid response times from fire fighters. The ¼ mile width should probably be considered a minimum width that may need to be wider in steep areas or where response times are greater.

Relevance to managers:

The most interesting result (in my humble opinion) seemed to be the importance of live tree density as a major factor in fire intensity. Where live tree densities were higher (including larger trees), fire intensity was greater. Commercial thinning of trees was important, as long as surface fuels were also treated. Typically, fuels managers rank treatment of surface fuels as the biggest bang for your buck when it comes to fuel treatments. But in this case, they were necessary but not sufficient. There needed to also be a reduction of live tree density.

Another interesting result was the differential morality between fir and pine species. It is known that fir is less resistant to fire, but the magnitude of the species difference in this case was huge. Fir were way less resistant, even where treatments that preferentially reduced fir relative to pine had previously occurred.

The treatments in this case consisted of a series of multiple treatments (commercial thin, then precommercial hand thinning, then piling and burning) that took multiple years. A commercial thin with a biomass harvest could have done all three treatments at one time. This would have potentially avoided the situation where debris piles were not yet burned at the time of the wildfire. The local infrastructure for utilizing biomass material, however, has to be in place to make such an operation feasible. Perhaps the Angora fire can serve as an example of the need for more biomass utilization infrastructure.

Critique (I always have one, no matter how good the article is) for the pedants:

A statistician could have a good time tearing this apart. There are potential biases (treatments were probably targeted in areas of higher hazard to begin with) and sampling problems related to non-independence of samples. However, I think the authors make a reasonable defense that observational studies like these are nonetheless important. We can not feasibly conduct wildfire experiments in the traditional sense. Instead we rely on statistics, but not statistical testing, to make inferences (kind of like we know that Ken Griffey Jr. is a good baseball player, not because we ran experiments, but because of his statistics observed over many seasons). The more fire seasons of statistics we have, the more convinced we can be that the conclusions are correct. Because we now have quite a few observations of wildfires where they have burned into areas that were previously treated, we can make some pretty strong inferences even though they are not true experiments.