shouting fire in a crowded landscape
by Christina Hulbe
We're heading into the "home stretch" of fire season here in the west but it's been an intense year across the United States. 2007 got off to a typical enough start but fire activity jumped dramatically in July (NOAA graphic), especially in western Montana and central Idaho.
According to recent analyses of fire data (Westerling et al., Science, 2006 pdf) Western U.S. fire activity increased dramatically in the mid-1980's. Fires now occur more frequently and last longer than they did earlier in the observational period (starting in 1970), and the length of the fire season has increased. Along with the increase in fire activity has come an increase in spending on fire management (well in excess of $1 billion dollars a year). The questions you should be asking are "are the last few years unusual" and "is this just the result of past fire prevention efforts?"
Fire is essential to the function of many ecologic communities (examples: Sequoia; rangeland; lessons from the 1987 Silver fire complex in Oregon). Species (including humans) living in fire-dependent ecosystems are adapted to tolerate and benefit from the effects of fire. The fire regime of a particular ecosystem is defined by typical patterns of fuel consumption (ground fires burn through soil organic matter; surface fires burn through dry or dead plant material; crown fires burn through the tops of shrubs and trees), intensity, severity, frequency, and the season in which fires tend to occur (dryness and lightning are important). Fires impact both biotic and abiotic components of ecosystems, altering soil temperature, moisture, physical properties, and chemistry.
Fire promotes heterogeneity in ecosystems, an attribute exploited by humans for millennia. Today, we think of using fire in "controlled burns" as either a land clearing tool or for management of the biomass that fuels unwanted fires. Our ancestors used fire as a means to ensure their food supplies by improving soils and driving species diversity. Examples abound in the anthropological literature and recently, academics have been making the case that we modern folk would do well to learn from the dwindling indigenous knowledge base before it is gone forever (Australian Aborigines, abstract ; Brazilian Kraho abstract; and a nice paper about Amazonian terra mulata and terra prieta at pacific views).
Unintentional and non-anthropogenic fires are of course what we care about in the context of increased fire incidence and related property damage. The impact of fire in a region depends on environmental factors such as topography, climate, vegetation type, health of plants in the system (insects, drought), and past fire suppression (which increases the biomass stored in the system and available to burn). When trying to understand changes in fire regime over time, both climate and land management strategies are of interest.
land management v. climate
The effects of land use on ecosystems, accumulated biomass, and fire depend on the undisturbed fire regime for a given ecosystem type. The impact of fire prevention would be most profound in forests that previously experienced frequent, low-intensity fires (such as ponderosa pine in the southwestern U.S.) and of least importance in forests that would otherwise experience infrequent, high-intensity fires (northern lodgepole pine or spruce-fir).
The effects of climate change, either a change in variability (amplified wet/dry cycles) or drought frequency, would be felt uniformly across forest types. Tree ring records of drought and fire history show a clear correlation between fire and drought in the summer fire season (article in the High Country News)and a lesser correlation with relatively moist conditions in years prior to the drought.
Climate change may also affect vulnerability to insect infestation, and thus forest health. The relationship among climate, insect outbreaks, and fire is not well understood but forest paleoecologists are working on it.
Depending on one's political sensibilities, it may be attractive to favor either land management or climate change as a cause for the dramatic increase in fire events since 1986 (and the political landscape grows only more complicated from here). The two may in fact be complimentary in some regions. For example, the undisturbed ponderosa forest experiences relatively frequent, low intensity and dead-biomass clearing fires. Where land managers have chosen to suppress fire in this ecosystem, they have perhaps primed the pump, allowing biomass to accumulate at the same time that climate change is increasing the likelihood of fire.
Overall, ecologists see climate as the leading culprit in the recent increase in fire incidence. On decadal and longer time scales, climate (the mean state and variability) shapes the character of regional vegetation by modifying species composition and biomass continuity, and through this, the regional fire regime. On shorter time scales, climate variability affects the moisture content, and thus flammability, of forest vegetation.
To make sense of the recent trend toward larger and longer fires, we need to place it in a climate context (does fire correlate with a specific climate parameter?) and to study as long a time period as possible (do very large fires of the last two decades stand out?). In a paper published last year in the journal Science, Tony Westerling and colleagues at U.C. Merced, U. Arizona, Scripps, and the U.S. Geological Survey tackled the first part (pdf of technical paper; podcast; news story) and a few research groups, mainly at universities in the western U.S., are working on the latter (for example, U. Montana, U. New Mexico, and U. Arizona).
Westerling and his colleagues compiled a time series of large (more than 400 hectares) forest wildfires from 1970 to 2003 on federally managed lands in the western U.S., for which data are fairly complete (fire area, elevation, discovery and control dates, timber damage, and so on). Fire frequency was four times larger in the period from 1986 to 2003 than it was in the earlier part of the record. The annual average total area burned increased by about six and a half times after 1986. The research group also found that interannual fire variability is strongly associated with regional spring and summer temperature such that there are more fires in hotter years than in cooler years (66% of the variance is explained by this one variable). The length of the fire season also increased in the mid-1980's, on average by 78 days, as did the duration of individual fires.
Western spring and summer temperatures were on average 0.87 degrees Celsius warmer over the interval 1987 to 2003 than for 1970 to 1986. Indeed, the years 1987 to 2003 were the warmest since the start of the western instrumental record in 1895. So what is it about warmer years that drives wildfire?
The recent increase in fire frequency is concentrated between 1680 and 2590 meters elevation. At these elevations, fire events are episodic and strongly associated with the timing of spring snowmelt (which itself is correlated with temperature variations). The earlier the start of snowmelt, the more intense the fire season. Because winter snow accumulation dominates water supply in western mountains, earlier snowmelt can result in an earlier, longer dry season. A longer dry season provides more opportunities for large fires by drying soils and vegetation and by increasing the time over which ignition is possible. Warmer spring and summer temperatures also, or course, affect summer drought.
The next thing we need to investigate is the history of western fire. Are the recent large fires unusual? How does fire vary over time in ecosytems not managed by people? This is where the paleoclimatologists and paleoecologists come into the picture. Fire history over past millennia is difficult to trace, due to the nature of fire itself, but several approaches have been followed with good results.
Fire scars on living and dead trees is a logical place to start and have the advantage of direct comparison between climate (tree ring characteristics) and fire. This approach is limited though because in most places, 500 years is old for a tree and the record is biased toward fires that were not severe enough to actually kill the trees.
Another approach is to auger sediment cores from lakes to look for bits of charcoal that fell into the lake as fire raged around it. Such records can extend back a few thousand years but can be as patchy as fires themselves (if a regional fire left unscathed the alpine basin in which the pond you sample lies, you'll never know the fire happened).
Yet another approach is to track fire indirectly, through its effects upon the landscape, primarily erosion (pdf of technical report on one such study). In mountain environments, major fires are often followed by increased slope erosion, debris flows, and floods. These effects can be traced in sediment deposits and changes in sedimentation over time can then be used as a proxy for past fire frequency and magnitude.
Taken together, these sorts of studies make two important points for this discussion. First, recent dramatic fire events are not unprecedented. This is important because it indicates that large, intense fires are not only the result of modern fire suppression but can occur in the "undisturbed" forest as well. Aggressive thinning of managed forests is no guarantee of fewer fires. Second, there is a clear correlation between shifts to relatively warm and dry conditions and increases in fire intensity. This point is important if we want to make projections about changes to fire hazards due to global warming.
adding fuel to the fire
What all of the studies referenced here show is that warmer and drier means increased fire frequency and fire intensity, regardless of land management strategy. Global warming effects such as enhanced drought and earlier melting of western snowpack are likely to increase fire frequency and magnitude in the future.
Wildfire suppression is not the only way we add fuel to the fire. We're building houses too. As our population grows and people build houses deeper and deeper into formerly wild landscapes (especially out here in the west), the load of flammable material on the landscape increases and of coursse, the likelihood of property damage due to fire increases as well.