Ecological Archives E094-227-A2

Brian J. Harvey, Daniel C. Donato, William H. Romme, Monica G. Turner. 2013. Influence of recent bark beetle outbreak on fire severity and postfire tree regeneration in montane Douglas-fir forests. Ecology 94:2475–2486. http://dx.doi.org/10.1890/13-0188.1

Appendix B. Information on beetle outbreak severity measurements.

Field measurements of beetle outbreak severity

Pre-fire bark beetle outbreak severity was measured by assigning trees in each plot into four distinct categories: "live at the time of fire", "killed by bark beetles prior to fire", "pre-disturbance snag", or "unknown" (Table B1). Trees categorized as "live at the time of fire" could be either dead or alive at the time of sampling. If dead, trees were categorized as "live at the time of fire" if they had no evidence of pre-fire bark beetle activity (exit holes on outer bark, galleries under bark) and were not a highly decayed or well-weathered snag that indicated death prior to both disturbances (Figure B1). These individuals accounted for 70% of trees. The "live at the time of fire" category also included live or dead trees that had obvious signs of post-fire beetle activity (boring dust [which would have been consumed by fire if beetle activity was pre-fire], resin bleeding) or had fully developed galleries but moist cambial tissue and/or any detectable level of needles in the canopy (Fig. B1) – these individuals accounted for <1% of trees. For a tree to be categorized as "killed by bark beetles prior to fire" it must have met all of the following criteria: dead at the time of sampling, presence of exit holes on the outer bark, dry cambial tissue, fully excavated (but vacated) adult and larval Dendroctonus galleries on the vascular cambium (> 50% of bole circumference or remaining visible cambium), and no needles in the canopy (Fig. B2). These individuals accounted for 24% of trees. Dead trees with well-weathered sapwood and no evidence of pre-fire bark beetle activity were categorized as "pre-disturbance snags", meaning they were most likely dead prior to the onset of the Douglas-fir beetle outbreak. These individuals accounted for 4% of trees. Severely burned dead trees with no visible cambium were categorized as "unknown," accounting for 1.5% of trees.

Douglas-fir beetles can sometimes attack fire-injured trees following light-surface fire (Parker et al. 2006), potentially leading to false-positive designation of trees as "killed by beetles prior to fire" when in fact they were killed by beetles during or after fire. Several lines of evidence show this situation to be very unlikely given our sampling protocol. First, for potential beetle attack at the time of fire, a key factor is that Douglas-fir beetle tree infestation occurs in the late spring or early summer (Schmitz and Gibson 1996). Therefore active beetle infestation in 2008 would have occurred before the Gunbarrel Fire, meaning that some burned trees would exhibit partially completed galleries and beetles charred under the bark (as we have observed in a related beetle-fire study at another site; Harvey et al., unpublished data). We saw no evidence of such partial galleries on any tree. Second, regarding potential post-fire beetle attack, for trees to be successfully attacked by beetles following fire there must be a suitably high beetle population nearby the burned forest and the fire must occur close to the time of the next beetle flight (DeNitto et al. 2000). Because the beetle outbreak had subsided >3 years before the fire (Appendix A), the local beetle population would have been low. As the Gunbarrel Fire was in late July and August, it occurred after the 2008 beetle flight and ~ 9 months before the 2009 beetle flight. Combined, these circumstances make the possibility unlikely that there was substantial post-fire beetle attack in the year following the fire. Also, any such trees would likely still retain some needles in the canopy at the time of sampling. Needles from Douglas-fir trees begin to fall one year following attack from Douglas-fir beetle and continue to fall (and thus some are retained in the canopy) for 2 or more years following attack (Schmitz and Gibson 1996, Donato et al. 2013). Therefore, we conservatively categorized dead trees as "live at the time of fire" if they exhibited beetle galleries but retained a detectable level of needles in the canopy at the time of sampling (accounting for < 1% of trees), as needles would be absent at the time of sampling if the tree were killed by bark beetles prior to fire, but present on the tree if it were killed by beetles at some point after the fire (Ken Gibson, personal communication).

Supplemental validation and cross-check of field measures of beetle outbreak severity

We performed a dendrochronological analysis to further test that trees classified as "killed by beetles prior to fire" were indeed dead prior to fire and in the gray stage. Cross-sectional wedges were collected from a random sample of 10 dead trees field-classified as “beetle-killed prior to fire” (averaging 115 visible years per wedge cross-section; range 32–425) and a master chronosequence was built from cores collected from 27 live Douglas-fir trees (averaging 109 visible years per core; range 60–201). By manual cross-dating with skeleton plotting (Speer 2012), we used consistently reliable marker years with anomalously narrow growth rings (years with region-wide severe moisture deficit: 1988, 1994, 2000, and 2001, Westerling et al. 2011) to find the best cross-dated match between all trees (live and dead). Cross-dating the 10 trees field-classified as "killed by beetles prior to fire” with the live trees indicated that all died between 2002 and 2004. The most frequent year of death was 2002 (n = 5), followed by 2003 (n = 4) and 2004 (n = 1). This matches the temporal trend in the aerial detection surveys (ADS), which show the peak of tree mortality in 2002 (Appendix A) and confirms the accuracy of our field measures of outbreak severity.

Finally, to cross-check the sensitivity of our analysis, results, and conclusions to any potential error in field classification of beetle-killed trees, we repeated all statistical analyses using measures of beetle outbreak severity from pre-fire ADS data, providing a second (albeit less quantitative and precise) measure of outbreak severity recorded independently from our field measures. ADS data can underestimate the magnitude of outbreak severity (Meddens et al. 2012) and are known to contain spatial error (Johnson and Ross 2008), two important reasons why they are not relied on as a primary source of information for fine-scale beetle outbreak severity. However, for purposes of our sensitivity analysis, we assume these errors to contain no bias that would make relative comparisons across space (e.g., plot to plot) problematic for a valid cross-check of our findings based on field data. In our sensitivity analysis, we substituted the cumulative number of trees per hectare that were beetle-killed between 1994 and 2008 on the ADS maps for our field measure, and then repeated all correlation and regression analyses with each fire-severity response variable. There was no qualitative difference between these results and our original analysis (i.e., all effects remained the same; there was no beetle effect on fire severity before or after accounting for burning conditions and slope position). This cross-check using ADS data as substitute of our field measures of beetle outbreak severity corroborates our findings and conclusions from our main analysis using our detailed quantitative field data.


Table B1. Evidence and criteria used to classify each tree in a plot into one of four categories: "live at the time of fire", "killed by bark beetles prior to fire", "pre-disturbance snag", or "unknown."

Field classification

Likely time of death

Likely cause of death

Evidence

Percent of all trees sampled

 

 

 

 

 

"Pre-disturbance snag"

Prior to both disturbances

Unknown

  • dead at time of sampling
  • highly weathered/decayed sapwood
  • most branches and bark missing
  • no evidence of bark beetle activity (pre- or post-fire)

4%

 

 

 

 

 

"Killed by bark beetles prior to fire"

Pre-fire

Bark beetle attack

  • dead at time of sampling
  • Dendroctonus exit holes on the outer bark
  • dry cambial tissue
  • fully excavated (but vacated) adult and larval Dendroctonus galleries on the vascular cambium (> 50% of bole circumference or remaining visible cambium)
  • no needles in the canopy

Relevant references: Schmitz and Gibson 1996; Donato et al. 2013; Ken Gibson, pers. comm.

24%

 

 

 

 

 

"Live at time of fire"

Gunbarrel Fire

Fire

  • dead at time of sampling
  • charred bark, branches, or outer sapwood
  • no evidence of bark beetle activity (no exit holes on outer bark, no galleries under bark)
  • not a highly-decayed or well-weathered snag

63%

 

 

 

 

 

"Live at time of fire"

Beetle attack at time of fire

Fire

  • dead at time of sampling
  • partially completed galleries with adult beetles charred under the bark

Relevant references: Schmitz and Gibson 1996; Harvey et al., unpublished data

0%

 

 

 

 

 

"Live at time of fire"

Beetle attack post-fire

Bark beetle attack

  • alive or dead at the time of sampling
  • clear signs of post-fire beetle activity (boring dust [which would have been consumed by fire], resin bleeding) or fully developed galleries but moist cambial tissue and/or any detectable level of needles in the canopy (which would still be present given needle-drop period of 2–3 yrs)

Relevant references: DeNitto et al. 2000; Schmitz and Gibson 1996; Donato et al. 2013; Ken Gibson, personal communication

<1%

 

 

 

 

 

"Live at time of fire"

Currently live

n/a

  • alive at the time of sampling
  • green foliage, no sign of Dendroctonus beetle activity

7.0%

 

 

 

 

 

"Unknown"

unknown

unknown

  • none available - excessive charring, etc.

1.5%

 

 

 

 

 

 

FigB1

Fig. B1. Photographs of general bole condition (a) and crown condition (b and c) of example trees that were dead at the time of sampling and classified as "live at the time of fire." Photo credits: B. J. Harvey.


 

FigB2

Fig. B2. Photographs of general bole condition (a and b), and bole and crown condition (c–d) of example trees that were classified as "killed by bark beetles prior to fire." Photo credits: B. J. Harvey.


 

Literature cited

DeNitto, G., Cramer, B., Gibson, K., Lockman, B., McConnell, T., Stipe, L., Sturdevant, N., and Taylor, J. 2000. Survivability and deterioration of fire-injured trees in the Northern Rocky Mountains: a review of the literature. USDA Forest Service Northern Region Forest Health Protection Report 2000-13. USDA, Missoula, Montana, USA.

Donato, D. C., B. J. Harvey, W. H. Romme, M. Simard, and M. G. Turner. 2013. Bark beetle effects on fuel profiles across a range of stand structures in Douglas-fir forests of Greater Yellowstone. Ecological Applications 23:3–20.

Johnson, E., and J. Ross. 2008. Quantifying error in aerial survey data. Australian Forestry 71:216–222.

Meddens, A. J. H., J. A. Hicke, and C. A. Ferguson. 2012. Spatiotemporal patterns of observed bark beetle-caused tree mortality in British Columbia and the western United States. Ecological Applications 22:1876–1891.

Parker, T. J., K. M. Clancy, and R. L. Mathiasen. 2006. Interactions among fire, insects and pathogens in coniferous forests of the interior western United States and Canada. Agricultural and Forest Entomology 8:167–189.

Schmitz, R., and K. Gibson. 1996. Douglas-Fir Beetle. USDA Forest Service, Forest insect and disease leaflet 5, 7.

Speer, J. H. 2012. Fundamentals of tree ring research. University of Arizona Press, Tuscon, Arizona, USA.

Westerling, A. L., M. G. Turner, E. A. H. Smithwick, W. H. Romme, and M. G. Ryan. 2011. Continued warming could transform Greater Yellowstone fire regimes by mid-21st century. Proceedings of the National Academy of Sciences 108:13165–13170.


[Back to E094-227]