«Donald L. Martinez a, Gerald Ryszka b, James J. Worrall c, Roy A. Maskc, Tom Eager c and Leanne Egeland c a b North and South Zones, Timber ...»
United States Renewable Resources
Department of Technical Report
Deterioration of Fire-Killed Trees in the Rocky Mountain Region
Donald L. Martinez a, Gerald Ryszka b, James J. Worrall c,
Roy A. Maskc, Tom Eager c and Leanne Egeland c
North and South Zones, Timber Management, Rocky Mountain Region, USDA Forest Service
Gunnison Service Center, Forest Health Protection, Rocky Mountain Region, USDA Forest Service Abstract: We assessed deterioration of 851 fire-killed trees over six years following the wildfires of summer 2002 (sampling in 2002-2007). Five tree species and nine wildfires were studied in seven national forests of the Rocky Mountain Region. One year after the fires, subalpine fir had lost about 40% of volume to firerelated defects and Engelmann spruce about 20%; loss in other species was negligible. The early loss was due primarily to checking, as suggested by the steep declines in log density due to rapid drying of wood and the absence of sap rot after one year. After two years, volume loss of subalpine fir and Engelmann spruce was about 60%. That of ponderosa pine was highly variable, about 10% in two southern fires but over 60% in fires in the Black Hills. Sap rot became an increasing component of defect during this time.
Wood borer galleries continued to increase through the fourth year. Sapwood stain (not considered a defect) was only abundant in lodgepole and ponderosa pines, where it reached a maximum in two years.
Although data on internal, pre-existing decay (heart rot) were variable, evidence suggests it did not expand significantly in the dead trees. By three years after the fires, all species on all fires were substantially affected and volume losses averaged over 50%. Nineteen species of wood-boring insects were identified from emergence traps. Thirty-eight species of fungi decaying the trees were identified from cultures, including one new record for Colorado, but these represent a small portion of the decay community in these trees.
Key words: Wildfire, tree deterioration, decomposition, wood decay, borers, fungi.
Following the unprecedented 2002 fire season in the Rocky Mountain Region (Anonymous 2003), there was demand from Forest Service and timber industry representatives for information on the rate and type of deterioration of fire-killed trees. Most of this interest was related to marketability and salvage of timber, but there was also interest in the hazard of trees in developed areas as well as the ecological processes of decay and deterioration of the coarse woody debris.
A literature search showed that there was information for some of our tree species, but it was highly variable and almost all from other regions. Therefore, we initiated a project to gather the needed information and began implementation in fall of 2002 within months of the fires. The goal was to provide answers to these questions that will be useful after future fires.
1.1 Snag longevity Although we did not measure fall rates of snags in this study, it is often of interest following a large-scale mortality event, so we review some relevant studies here. Studies of snag longevity following bark beetle outbreaks and fires are included.
Ponderosa pine. One of the largest and longest studies of snag longevity was by Keen (1955).
In the northern Sierra Nevada, various surveys marked and numbered beetle-killed snags of ponderosa pine annually from 1919 to 1949. The rate of fall was very low for the first five years (at the end of which 85% were still standing), very rapid between 5 and 15 yr, then tapered off to a very low rate. After 25 years, 10% of snags were still standing. Small trees with high proportion of sapwood fell faster than the average. Loam soils had a more rapid fall rate than the drier pumice soils of the area. Kimmey (1955) noted that many ponderosa pine killed by fire break off at ground level or up to 50 ft high during the 4th and 5th year.
Beetle-killed ponderosa pine ranging from 7-22″ DBH was studied in the Front Range of Colorado (Schmid et al. 1985). No trees fell within two years of death. Thereafter, an annual fall rate of 3-4% was observed, but the rate soon increased to as high as 17%. When winds exceeded 75 mph, more snags fell. On one site, 70% of trees were down within six years, but for other sites this took ten years. 90-100% of trees fell to the east, further supporting the role of strong winds in many fall events. Most trees broke between the soil line and two feet above. Beetle-killed ponderosa pine snags fell more quickly on the Black Hills National Forest in South Dakota (Schmid et al., 2009).
Only one of 277 monitored snags broke within two years, but after five years, 76%, 91% and 95% of snags were broken in three stands. The most common point of failure was in the first two feet above ground, but nearly 25% broke at 25-35 feet high.
Seven years after wildfire in northern Arizona, 41% of ponderosa pine snags had fallen (Chambers & Mast 2005). In a chronosequence of severe fires in northern Arizona (Passovoy & Fulé 2006), fallen snags increased with time. In fires 3-4 yr old, 14.6% of snags were down, but that increased to 60.8% in fires 8-9 yr old.
Ponderosa pine was followed up to 11 yr after two wildfires in western Idaho (Russell et al.
2006). Snag dynamics were similar in the two fires; 55-60% of snags were down after 9 yr and 85% were down after 11 yr. Estimated half-life (50% of snags down) was 7-8 yr in salvage-logged plots and 9-10 yr in unlogged plots. The logging effect was due largely to a reduction in average snag size (smaller snags fall sooner) and partly to a reduction in density.
Ponderosa pine killed by an escaped, prescribed fire was studied in west-central Oregon (Dahms 1949). Snags were first tallied 10 yr after the fire and they were followed for 12 additional years. Ten years after the fire about 55% of the snags were down. At the end of the study, at 22 yr, 78% were down. The curve was nearly flat, with very low fall rate, by that point. Smaller snags tended to fall early. Unlike Douglas-fir, which reportedly tend to gradually break down from the top, the pine snags tended to fall by uprooting.
Ponderosa pine killed by prescribed fire was also studied by Harrington (1996), on the San Juan National Forest. Fall rates were higher than in the California and Oregon studies. Five years after death, up to 50% of trees were down (in the spring/summer burn treatment). Nine years after death, 62-78% of trees were down. There was not a significant difference in fall rate among tree sizes, but all trees were small, probably accounting for the unusually high fall rates.
On the east side of the Cascades in Washington, Everett et al. (1999) studied a chronosequence of fires and estimated that small ponderosa pine snags (23 cm DBH) had a half-life of only about 7 yr. Too few larger ponderosa pine snags were available for regression analysis, but it appeared they lasted longer.
Deterioration of fire-killed trees 4 Lodgepole pine. Lodgepole pine killed by fire was studied in Montana (Lyon 1977). Very few snags fell in the first two years. Thereafter, snags 3″ DBH fell at 8.4% yr-1. Ten years after the fire, the percentage of snags down was: 3−8″, 47%; and 8−12″, 62%. There was high variability among transects. Slope and snag density did not influence fall rate.
Hinds et al. (1965) included some lodgepole pine in a study of beetle-killed spruce in Colorado.
Trees averaged 12″ DBH and fell at an average annual rate under 2% yr-1 for 17 yr. It was estimated that 45% of the stems would be down in 20 yr.
In the study by Everett et al. (1999; see ponderosa pine above), small lodgepole pine snags (23 cm DBH) had a half-life of only about 11 yr and medium snags (23-41 cm) about 15 yr. No large snags were available for study.
In dry sites in central Oregon, 602 previously tagged trees were monitored annually after a mountain pine beetle outbreak (Mitchell & Preisler 1998). Some plots had been thinned from below before the outbreak, and suffered lower mortality (20%) than the unthinned plots (43%).
Snags began falling 3-5 years after death. Half were down in 8-9 yr, and 90% were down in 12-14 yr. Snags in thinned stands fell slightly faster than those in unthinned stands. After a lag phase, annual fall rates were about 10% per year in unthinned plots and 13% in thinned plots (these authors apparently calculated their fall rates each year on the basis of the original number of trees, rather than the number remaining the year before). Small trees fell slightly faster than large trees in thinned stands, but not in unthinned stands. These are among the fastest falling snags of all the studies reviewed here.
In northeastern Oregon, Harvey (1986) also measured fall rate of beetle-killed lodgepole pine.
Of 427 trees, only one fell in the first 5 yr. By 10 yr, however, 25% had fallen.
In British Columbia, stands resulting from a 1979 outbreak of mountain pine beetle were assessed 25 years later (Anonymous 2007). Among 14 plots, the snags still standing (divided by those plus the ones on the ground) ranged from 0 to 80%, with an average of approximately 45%.
A more recent study in British Columbia investigated a chronosequence of lodgepole pine killed by mountain pine beetle (Lewis & Thompson 2009). Snag fall was negligible until 6 yr after mortality, and most rapid beginning at 8 yr. Fall rate was highly variable among plots. Snag failure was due primarily to sap rot near the soil line. DBH had no significant relationship to fall rate.
Spruce. Engelmann spruce snags killed by spruce beetle were studied in Utah (Mielke 1950) and in Colorado (Hinds et al. 1965). They fell much more slowly than the pines already reviewed.
In the Dixie National Forest, Utah, snags were tallied 25 yr after the outbreak. Only 16% of the snags appeared to have fallen during that period. Fall rate was not influenced by aspect, slope, or soil type. Annual fall rates were 0.92% for snags 3-7″ DBH and 0.69% for those 8″. Decay in the roots and/or butt was considered the primary factor in causing 77% of snags to fall.
In the Colorado study, snag fall (termed “windthrow”) was reported on the basis of volume.
One would expect that to be lower than a tree basis since fall rate is generally lower in larger size classes, but the authors reported that the percentage fall rate was essentially similar by the two bases. About 8% were down 10 yr after peak mortality, and 28% after 20 yr. This was much higher than in the Utah study, apparently because of the high incidence of root and butt rot in the trees when they were killed. Decay at the base, including butt rot, was thought to substantially contribute to fall of 96% of the snags.
In the eastside Cascades study of Everett et al. (1999), Engelmann spruce snags had an exponential decay curve that varied by size. They estimated that medium-sized snags (23-41 cm DBH) had a half-life of about 23 yr, while larger snags had a half-life of 20 yr. However, the curve for large snags flattened later, so it was estimated that after 80 yr, up to 30% of large spruce snags would remain standing, while the medium class would retain only 15-20% for that period.
Norway spruce logs on the ground in Norway were studied to determine date of death by crossdating ring chronologies of the fallen log (Storaunet & Rolstad 2002). This study was not based on Deterioration of fire-killed trees 5 a single mortality event or cause. Date of fall was determined either by dating wounds inflicted during the fall on neighboring trees, or by aging trees that had established on the fallen log. The frequency of time from death to fall had a reverse J-shaped distribution, with a mean of 22 yr and a range of 0-91 yr. Decomposition was more closely related to time since fall than to time since death.
Douglas-fir. Russell et al. (2006) included Douglas-fir in their study (see above under ponderosa pine). Snags fell at a lower rate than for ponderosa pine. There appeared to be some discrepancy in actual fall data between the text and a figure, but 10% or 18% of snags fell in 9 yr in the unlogged area, and 15% or 19% fell in the partly salvage-logged area in 11 yr. Estimated halflife of snags was 15-16 yr in unlogged plots and 12-13 yr in salvage-logged plots, which is difficult to reconcile with either of the sets of actual data.
In the chronosequence of fires on the east side of the Cascades (Everett et al. 1999), small Douglas-fir snags (23 cm DBH) had a half-life of only 9-10 yr and were gone by 45 yr. Medium snags (23-41 cm) were 50% gone in about 15 yr. However, larger snags had a half-life of about 35 yr, and after 80 yr 40% still remained standing. These large Douglas-fir were the most persistent snags of the study, which included five species.
Subalpine fir. Subalpine fir fall rates were studied by decay class in British Columbia (Huggard 1999). Trees were grouped into decay classes so that fall rates could be based on features recognizable by forest workers. Range of time since death for each class was determined by crossdating increment cores. Fall rate was estimated by the number of snags and ages of the class, assuming that the same number of trees become snags each year. Annual fall rate increased with decay class (Table 1). Calculations suggest that 20% of snags fall by 27 years and 50% by 53 years.
Because the fall rate of the first snag class was not higher than that of live trees (Table 1), Huggard recommended that they be disregarded as hazards to forest workers, and that they not be felled unless they have obvious structural defects.
Table 1. Annual fall rates of subalpine fir in British Columbia by class (Huggard 1999).
Class Time since death (yr) Annual fall rate (%) ± standard error Live trees 0.5 ±0.1 Fine branches present, bark intact 0-19 0.2 ±0.9 Fine branches few, bark cracked 12-35+ 0.7 ±0.9 Fine branches none, ≤50% bark lost 16-66+ 3.7 ±0.9 Fine branches none, 50% bark lost 40-90 6.0 ±1.7 In the east Cascades chronosequence of fires (Everett et al. 1999), small subalpine fir (23 cm DBH) had a half-life of about 11 yr. Medium snags had a half-life about 25 yr. Insufficient large snags were available for regression.