Journal of Chemical Ecology

, Volume 37, Issue 11, pp 1184–1192

Fire Injury Reduces Inducible Defenses of Lodgepole Pine against Mountain Pine Beetle

Authors

  • Erinn N. Powell
    • Department of EntomologyUniversity of Wisconsin
    • Department of Forest and Wildlife EcologyUniversity of Wisconsin
    • Department of EntomologyUniversity of Wisconsin
    • Department of Forest and Wildlife EcologyUniversity of Wisconsin
Article

DOI: 10.1007/s10886-011-0031-4

Cite this article as:
Powell, E.N. & Raffa, K.F. J Chem Ecol (2011) 37: 1184. doi:10.1007/s10886-011-0031-4

Abstract

We examined the effect of wildfire injury on lodgepole pine chemical defenses against mountain pine beetle. We compared the constitutive phloem chemistry among uninjured, lightly-, moderately-, and severely-injured trees, and the induced chemistry elicited by simulated beetle attack, among these same categories. We also compared the entry rates of caged female beetles into trees of these categories. The volatiles we studied included thirteen monoterpene hydrocarbons, four allylic monoterpene alcohols, one ester, and one phenyl propanoid, of which the monoterpene hydrocarbons always comprised 96% or more of the total. Fire injury reduced the total concentration of these compounds in the induced but not constitutive phloem tissue of lodgepole pines. Fire injury also affected the relative composition of some volatiles in both induced and constitutive phloem. For example, increased fire injury reduced 4-allylanisole, a host compound that inhibits mountain pine beetle aggregation. Increased fire injury also increased (−) α-pinene, which can serve as precursor of pheromone communication. However, it also reduced myrcene and terpinolene, which can serve as stimulants and synergists of pheromone communication. Beetle entry did not show statistical differences among fire injury categories, although there was a trend to increased entry with fire injury. These results suggest that the reduced ability of trees to mobilize induced chemical defenses is an important mechanism behind the higher incidence of attack on fire-injured trees in the field. Future studies should concentrate on whether beetles that enter fire-injured trees are more likely to elicit aggregation, based on the differences we observed in volatile composition.

Key Words

VolatilesMonoterpenesInductionLodgepole pineFireMountain pine beetlePlant defenseSemiochemicals

Introduction

Wildfire and bark beetle outbreaks are the two most important disturbance agents affecting conifer forests of western North America (Romme et al., 1986; Turner et al., 1999). Each has important effects on both natural ecosystem processes and human socioeconomic values (Kline et al., 2004; Abbott et al., 2009), and likewise is undergoing altered regimes due to anthropogenic inputs (Westerling et al., 2006; Kurz et al., 2008). These disturbance agents often have interacting effects, which can be complex (Geiszler et al., 1980; Martin and Mitchell, 1980). For example, fire injury to host trees can result in increased incidence of attack by bark beetles, but the strength of this relationship, and the extent of fire injury over which it occurs, vary with beetle species, tree species, environmental conditions, and beetle population phase (Geiszler et al., 1980, 1984; Amman and Ryan, 1991; Rasmussen et al., 1996; Ryan and Amman, 1996; Bradley and Tueller, 2001; Santoro et al., 2001; McHugh et al., 2003; Sullivan et al., 2003; Wallin et al., 2003; Elkin and Reid, 2004; Hood and Bentz, 2007; Six and Skov, 2009). Lombardero et al. (2006) found Ips pini (Say) and Ips grandicollis (Eichhoff) preferred fire-injured Pinus resinosa Aiton, suggesting such behavioral adaptations may underlie a significant ecological relationship between some bark beetles and their hosts. From a population dynamics perspective, fire-injury may provide a refuge for some bark beetles during their endemic phase, but not by itself necessarily lead to outbreaks (Powell et al., 2012).

This high variability poses substantial challenges to natural resource managers, who must make important and often contentious policy decisions involving fire and insect suppression, controlled burns, fuel accumulation to restore more natural conditions, and protection of human habitations. Understanding more about the mechanisms by which fire injury affects tree susceptibility to bark beetles could potentially address this problem.

Conifers are defended against bark beetle attack through several integrated mechanisms. Constitutive defenses act as a first line of defense and include physical barriers such as thick bark and lignin (Franceschi et al., 2005) and chemical barriers such as terpenoids and phenolics (Martin et al., 2003). Inducible defenses are triggered by insect or pathogen invasion and include formation of necrotic tissue that confines the invader (Paine and Stephen, 1988) and increased terpene and phenolic production, which deters establishment (Cook and Hain, 1988; Raffa and Smalley, 1995; Franceschi et al., 2005). In particular, increased monoterpene concentrations, and sometimes changes in their relative proportions, have deterrent, adulticidal, and ovicidal activities against bark beetles, and inhibit growth and germination of their fungal symbionts (Coyne and Lott, 1976; Raffa et al., 2005; Keeling and Bohlmann, 2006; Seybold et al., 2006; Bonello et al., 2006; Erbilgin et al., 2007; Bohlmann and Gershenzon, 2009; Zulak and Bohlmann, 2010). The ability to undergo induced total monoterpene accumulation appears particularly well correlated with tree survival in the field (Raffa and Berryman, 1982; Boone et al., 2011).

Stressors, such as severe drought, defoliation, lightning, and root disease, can reduce tree defenses against bark beetles (Christiansen et al., 1987; Flamm et al., 1993; Klepzig et al., 1996; Lombardero et al., 2006; Wallin and Raffa, 2001; Lewis and Lindgren, 2002; Turtola et al., 2003; Raffa et al., 2005). The effects of stress caused by fire injury on tree defenses, and its subsequent effect on bark beetle colonization and reproductive success, however, are less well known. Defensive resin flow has been shown to increase with fire injury in some conifers, but only several months post-fire (Wallin et al., 2003; Lombardero et al., 2006; Knebel and Wentworth, 2007). Our knowledge of how fire affects host chemical concentration and composition is even more limited. In one study, dead, fire-injured lodgepole pine (Pinus contorta var. latifolia Douglas) (Jakubas et al., 1994) had lower total volatiles than live uninjured trees. In another, total monoterpenes decreased 1 month post-fire in red juniper (Juniperus virginiana Lawson) (Campbell and Taylor, 2007). The defenses of fire-injured lodgepole pines that remain alive and hence potentially, susceptible to bark beetle attack, have not been examined, nor have the inducible chemical defenses that follow fire injury in any system.

The mountain pine beetle (Dendroctonus ponderosae Hopkins) is of major concern throughout the western United States and Canada. It recently has caused extensive mortality in both pine forests where it historically has undergone intermittent outbreaks and at higher latitudes and elevations than were previously common or recorded, an expansion attributable to changing climate (Kurz et al., 2008; Raffa et al., 2008; Safranyik et al., 2010). The primary host of the mountain pine beetle is lodgepole pine. During colonization attempts, it appears to benefit from its fungal symbiont, Grosmannia clavigera, in degrading lodgepole pine defenses (DiGuistini et al., 2011). The mountain pine beetle also exploits some host monoterpenes as precursors, stimulants, and synergists of the aggregation pheromones that it uses to coordinate the mass attacks need to overwhelm tree resistance (Wood, 1982; Blomquist et al., 2010).

We examined the effects of fire injury on lodgepole pine chemical defenses. Specifically, we tested whether fire affects the constitutive and induced total volatile concentrations, and their composition, in lodgepole pine phloem. We also tested whether mountain pine beetles caged onto trees with varying levels of fire injury would enter at similar frequencies.

Methods and Materials

Plot Description and Classification of Fire Injury Category

The LeHardy fire began on July 30, 2008, and burned 3,777 ha of lodgepole pine and mixed lodgepole pine and spruce forest in Yellowstone National Park, U.S.A. Sampling took place 1 year post fire, in July of 2009. Trees in nine, 1 ha plots separated by at least 100 m within the burn were classified for degree of fire injury using the methods of Ryan (1982) and Powell et al., (2012). Measurements of percent basal, bole, and canopy injury, and cambium kill rating (ckr) were used to designate the injury to each tree as unburned, low, moderate, and high as follows: Unburned trees were healthy, and without fire injury. Low-injured trees had 0–10% injury on the tree base and bole, and a ckr of zero or one. Moderately-injured trees had 11–49% injury on the tree base and bole, 0–10% injury of the canopy, and a ckr of two or three. Highly-injured trees had 50–100% injury on the tree base and bole, 10–100% injury of the canopy, and a ckr of three or four. Trees were at least 15 cm diam at 1.3 m.

Measurements were taken for additional variables that potentially could affect tree defenses. Tree-level variables included diameter at 1.3 m, tree height, presence and rating of dwarf mistletoe (Arceuthobium americanum Nuttall Ex. Engelmann) by using the method of Hawksworth (1979), and presence of Cronartium spp. stem rust by using the method of Geils and Jacobi (1984). We deliberately selected healthy trees without attacks by stem-boring insects; in a few cases, we found evidence of old attacks or new attempts between initial designation and final sampling, so this was also built into our model. Stand level variables included aspect, slope, and elevation.

Effects of Fire Injury on Constitutive and Induced Volatile Compounds

Two pieces of phloem (approximately 5 cm long by 2 cm wide) were removed with a razor blade from 50 lodgepole pines from each fire injury category, sealed in plastic tubes, transported from the field in dry ice-cooled Styrofoam coolers, and frozen at −20°C until analysis. One phloem piece each was removed from the north and south aspects of the tree, or if the tree was moderately- or highly- fire-injured, they were removed from the burned or unburned portions of the tree. These served as the constitutive samples.

We elicited induced responses on the same trees by challenge inoculations that mimicked the combined entry of the mountain pine beetle and its major fungal symbiont, Grosmannia clavigera. We isolated G. clavigera by rolling live beetle larvae from lodgepole pine phloem onto malt extract agar, isolating and culturing at 25°C for 7 d by using the method of Wallin and Raffa (1999). A pure culture of G. clavigera was isolated and identified by its morphology by using microscopy and descriptions from Upadhyay (1981). Simulated attacks of 1 cm agar-fungi plugs were administered at the xylem/phloem interface by using sterile techniques. The 2 cm bark-phloem plug was replaced immediately to the site of introduction. Two inoculations were administered per tree, on the north and south aspects, or the burned or unburned portions for moderately- and highly-injured trees. We simulated attack at the same time when constitutive phloem was sampled. We sampled the reaction zone 7 d after simulated attack and stored the samples as above.

Compounds were extracted from finely chopped constitutive and induced phloem samples after 24 hr of agitation in 1 ml of hexane as described in Raffa and Smalley (1995). This solution was filtered with glass wool. The sample vials were rinsed twice with 250 μl of hexane, which resulted in a final volume of 1.5 ml volatile-hexane solution. One μl of 0.1% isobutylbenzene was added to each sample as an internal standard. This internal standard has been commonly used in previous, including recent, analyses of these compounds (Robert et al., 2010; Boone et al., 2011), thereby facilitating between-study comparisons. However, some variation could arise from different response factors to any common internal standard (Raffa and Steffeck, 1988), and likely increases when non-monoterpene hydrocarbons are included. The remaining phloem was dried at 25°C for 1 wk and weighed.

Samples were analyzed with a Shimadzu 17 gas chromatograph with an Agilent Technologies cyclodex chiral column: 30 m long, 0.25 mm internal diam, and 0.20 μm film thickness with Helium as the carrier gas. Each analysis began at an initial temperature of 60°C for 10 min, followed by an increase in 5°C per min until 160°C. The total run for each sample lasted 69 min. Thirty pure volatile standards from Sigma-Aldrich for those commonly found in pines also were run using the above described method. We compared the retention times of these standards with the peaks on the chromatogram produced after the run of each sample. Milligrams of each compound were calculated by integrating its area under the curve and the internal standard, and we multiplied this value by the density of the internal standard.

Effects of Fire Injury on Mountain Pine Beetle Entry

Ten separate trees from each of the four fire injury categories were selected for a beetle entry bioassay that was conducted under no-choice conditions. Beetles were captured using ten 12-funnel flight traps (Lindgren, 1983) that were baited with trans-verbenol, myrcene, and exo-brevicomin from Contech Enterprises Inc. Captured beetles were kept in a refrigerator for no longer than 1 wk prior to bioassay. Only vigorous, active female beetles were used. A 13 cm2 square of aluminum screening was secured to the north or unburned portion of the tree bole at 1.3 m high. Five females were introduced into each cage. Needles and lichen from lodgepole pine branches were provided to give the beetles traction on a natural surface. Beetle entry and mortality were tabulated after 24 hr. For each tree, we noted diameter at 1.3 m, height, and potential sources of stress to incorporate conditions other than fire injury category that may affect beetle entry in our analysis, as described previously. The trees included in this experiment were of similar size and condition as those analyzed for chemical content.

Statistical Analyses

Concentrations of volatiles were calculated on a per dry weight of phloem basis. Concentrations from the north and south, or burned and unburned, aspects of the bole were averaged. We did not observe any difference between burned or unburned bole portions for either treatment, and we only observed differences between the north and south aspects of the bole for constitutive, unburned samples, and induced low injury samples (Supplementary Materials: Table 1). Concentrations were square root transformed to achieve normality. Normality and homogeneous variance were tested using quantile-quantile plots and Bartlett tests, respectively. We performed a two-way ANOVA using the statistical program R v2.7.2 (R Core Development Team, 2008) to identify potential differences of total volatile amounts among fire injury categories, constitutive or induced treatment, and the interaction between category and treatment. We used ANOVA to test for differences across fire injury categories within constitutive and induced phloem, and then performed t-tests with a Bonferroni correction to test for pair-wise differences. Separate two-way ANOVAs tested for potential differences of total concentrations between fire injury category and bole aspect, and between constitutive or induced treatment and bole aspect. We constructed a single generalized linear model with a square-root transformation to evaluate other potentially important pathogenic and environmental variables on total concentrations. We used a backward elimination procedure with Akaike’s information criteria (AIC) (Akaike, 1973), and selection of variables with α ≤ 0.05, to identify the best model for analyzing total volatile concentrations.

Individual volatile percentages in each sample were calculated and compared among fire-injury categories and constitutive and induced treatments. Normality was not achieved after an arcsin square-root transformation, so tests were performed using the non-parametric Kruskal-Wallis rank-sum test.

Beetle entry into trees of various fire-injury categories was analyzed using a Chi-square analysis. We constructed a single generalized linear model with a logit transformation to evaluate variables other than fire injury on no-choice, beetle entry. We used a backward eliminated procedure with AIC (Akaike, 1973), and selection of variables with α ≤ 0.05, to identify the best model for no-choice, beetle entry analysis.

Results

Effects of Fire Injury on Constitutive and Induced Volatiles

We detected 13 monoterpene hydrocarbons in the tissue of the constitutive samples, 12 monoterpene hydrocarbons in the induced samples, 4 allylic monoterpene alcohols, 1 ester and 1 phenyl propanoid, of which the monoterpene hydrocarbons comprised at least 96% in all induction and fire injury categories (Tables 1 and 2). Hereafter, our use of ‘total volatiles’ refers to these compounds. Total volatile concentrations (mg/g dry phloem) were affected by fire injury category, induction treatment, and the interaction between category and treatment (category: F(3, 391) = 8.65, P < 0.001, treatment: F(1,391) = 67.9, P < 0.001, interaction: F(3,391) = 4.73, P = 0.003). Total concentrations of volatiles within constitutive phloem tissue did not vary with fire injury (F(3,196) = 1.04, P = 0.376). In contrast, the concentrations of induced volatiles decreased by nearly half with the occurrence of any fire injury, regardless of severity (Fig. 1, F(3,195) = 18.1, P < 0.001). This pattern did not vary among the various measurements of fire injury.
Table 1

Effects of fire injury on percentages of individual monoterpenes and additional volatiles in constitutive phloem tissue of lodgepole pine, by Kruskal-Wallis rank sum

Compound

Unburned

Low

Moderate

High

χ2

df

P

(−) α-Pinene

2.55 ± 0.157

3.60 ± 0.140

3.58 ± 0.231

5.25 ± 0.370

60.8

3

<0.001

(+) α-Pinene

0.720 ± <0.100

1.30 ± 0.726

1.05 ± 0.109

1.57 ± 0.155

30.5

3

<0.001

Myrcene

3.96 ± 0.470

3.76 ± 0.335

2.94 ± 0.351

1.73 ± 0.229

41.1

3

<0.001

(−) Camphene

<0.100 ± <0.100

<0.100 ± <0.100

<0.100 ± <0.100

<0.100 ± <0.100

2.69

3

0.442

(+) Camphene

<0.100 ± <0.100

<0.100 ± <0.100

<0.100 ± <0.100

<0.100 ± <0.100

9.94

3

0.019

(+) 3-Carene

0

<0.100 ± <0.100

0

0

3.04

3

0.385

(+) β-Pinene

12.0 ± 1.07

7.48 ± 0.749

7.68 ± 0.982

10.9 ± 1.14

22.1

3

<0.001

α-Terpinene

<0.100 ± <0.100

<0.100 ± <0.100

0

<0.100 ± <0.100

2.00

3

0.573

(−) β-Pinene

10.9 ± 0.796

14.2 ± 0.757

13.8 ± 1.03

17.4 ± 1.27

16.4

3

<0.001

(+) Limonene

<0.100 ± <0.100

0.156 ± <0.100

0.275 ± <0.100

0.782 ± 0.118

55.2

3

<0.001

β-Phellandrene

64.5 ± 1.29

64.3 ± 1.25

64.4 ± 1.55

56.7 ± 1.72

14.6

3

0.002

Terpinolene

1.90 ± 0.224

1.61 ± 0.215

1.50 ± 0.310

0.403 ± <0.100

38.4

3

<0.001

Linalool

<0.100 ± <0.100

<0.100 ± <0.100

<0.100 ± <0.100

<0.100 ± <0.100

5.15

3

0.161

4-Allylanisole

1.05 ± 0.178

0.551 ± <0.100

0.371 ± <0.100

0.236 ± <0.100

28.2

3

<0.001

α-Terpineol

0

<0.100 ± <0.100

<0.100 ± <0.100

0.365 ± 0.198

8.45

3

0.038

Borneol

<0.100 ± <0.100

<0.100 ± <0.100

0.127 ± <0.100

<0.100 ± <0.100

3.92

3

0.270

Geraniol

<0.100 ± <0.100

<0.100 ± <0.100

0.201 ± 0.115

<0.100 ± <0.100

3.88

3

0.275

(−) trans-Caryophyllene

2.16 ± 0.404

1.77 ± 0.261

2.77 ± 1.09

<0.100 ± <0.100

61.4

3

<0.001

Bornyl Acetate

<0.100 ± <0.100

0

0

0

2.96

3

0.398

Table 2

Effects of fire injury on percentages of individual monoterpenes and additional volatiles in induced phloem tissue of lodgepole pine, by Kruskal-Wallis rank sum

Compound

Unburned

Low

Moderate

High

χ2

df

P

(−) α-Pinene

3.21 ± 0.800

3.78 ± 0.140

4.29 ± 0.280

5.57 ± 0.432

46.9

3

<0.001

(+) α-Pinene

1.69 ± <0.100

1.69 ± <0.100

1.54 ± 0.139

1.66 ± 0.177

1.71

3

0.634

Myrcene

2.58 ± 0.131

2.95 ± 0.194

2.48 ± 0.321

1.66 ± 0.264

52.1

3

<0.001

(−) Camphene

<0.100 ± <0.100

0.153 ± <0.100

<0.100 ± <0.100

0

22.7

3

<0.001

(+) Camphene

<0.100 ± <0.100

0.145 ± <0.100

<0.100 ± <0.100

0

24.1

3

<0.001

(+) β-Pinene

14.7 ± 1.09

13.5 ± 1.00

9.92 ± 1.09

11.6 ± 1.28

20.3

3

<0.001

(−) β-Pinene

9.23 ± 0.638

11.8 ± 0.940

14.0 ± 1.07

18.3 ± 1.39

24.9

3

<0.001

(+) Limonene

<0.100 ± <0.100

0.288 ± <0.100

0.598 ± 0.126

0.600 ± 0.144

24.3

3

<0.001

β-Phellandrene

65.5 ± 1.03

63.0 ± 0.956

63.2 ± 1.62

54.3 ± 2.19

19.4

3

<0.001

Terpinolene

1.32 ± 0.107

1.09 ± 0.112

0.662 ± 0.136

0.133 ± <0.100

111

3

<0.001

Linalool

0

<0.100 ± <0.100

<0.100 ± <0.100

0

43.0

4

<0.001

4-Allylanisole

0.676 ± <0.100

0.538 ± <0.100

0.212 ± <0.100

0.218 ± <0.100

81.7

3

<0.001

α-Terpineol

<0.100 ± <0.100

0.116 ± <0.100

0.158 ± 0.118

0

5.68

3

0.128

Borneol

0

<0.100 ± <0.100

<0.100 ± <0.100

0.132 ± 0.104

13.5

3

0.004

Geraniol

<0.100 ± <0.100

<0.100 ± <0.100

<0.100 ± <0.100

0

6.70

3

0.082

(−) trans-Caryophyllene

0.679 ± 0.226

0.697 ± 0.129

0.605 ± 0.184

<0.100 ± <0.100

40.8

3

< 0.001

Bornyl Acetate

<0.100 ± <0.100

<0.100 ± <0.100

<0.100 ± <0.100

<0.100 ± <0.100

0.878

3

0.831

trans-β-Ocimene

0

0

0

<0.100 ± <0.100

3.21

3

0.361

https://static-content.springer.com/image/art%3A10.1007%2Fs10886-011-0031-4/MediaObjects/10886_2011_31_Fig1_HTML.gif
Fig. 1

Effect of wildfire injury on concentrations of total monoterpenes and additional volatiles in lodgepole pine phloem. a. Constitutive tissue (F(3,196) = 1.04, P = 0.376). b. Induced tissue (F(3,195) = 9.59, P < 0.001). Means with different letters are statistically different at P < 0.05

The linear regression model incorporating all tree- and stand-level factors that yielded the best fit was R2 = 0.263, F(11,387) = 12.6, P < 0.001, and included fire injury category, induction treatment, the interaction between category and treatment, tree height, hillside slope steepness, and UTM-easting and UTM-northing plot position (Table 3). In addition to the effects of fire injury and induction, total volatile concentrations decreased with tree height, slope steepness, and more eastern and northern plot positions (Supplementary Material: Table 2).
Table 3

Effects of fire injury category, induction by fungal symbionts of mountain pine beetle, and additional tree- and site-level variables on total content of monoterpenes and additional volatiles in lodgepole pine phloem

Variable

Estimate (x)

Std. Error

P

βo (intercept; unburned category and constitutive treatment)

2.27 × 105

6.98 × 104

0.001

β1 (low category)

−17.6

6.69

0.009

β2 (moderate category)

−34.4

9.63

<0.001

β4 (high category)

−20.0

8.21

0.015

β5 (induced treatment)

35.6

5.87

<0.001

β6 (tree height)

−0.783

0.302

0.010

β7 (hillside slope steepness)

−0.827

0.294

0.005

β8 (UTM – easting)

−0.022

0.008

0.009

β9 (UTM – northing)

−0.044

0.013

0.001

β10 (low category, induced treatment interaction)

−29.6

8.30

<0.001

β11 (moderate category, induced treatment interaction)

−16.5

8.32

0.048

β12 (high category, induced treatment interaction)

−24.3

8.30

0.004

Plot locations (UTM zone 12). 1: 548328 easting, 4938992 northing; 2: 548380 easting, 4938563 northing; 3: 548432 easting, 4938674 northing; 4: 548502 easting, 4938749 northing; 5: 548215 easting, 4938187 northing; 6: 548311 easting, 4938616 northing; 7: 549591 easting, 4937979 northing; 8: 548312 easting, 4938455 northing; 9: 548367 easting, 4938992 northing

Multiple Linear Regression Model Was Selected Based on Best Fit: \( {\mathbf{y}} = {\beta_0} + {\beta_1}{{\mathbf{x}}_{{\mathbf{1}}}} + \ldots + {\beta_k}{{\mathbf{x}}_{{\mathbf{k}}}}.\;{R^2} = 0.{263},\;{F_{{(11,387)}}} = {12}.{6},\;P < 0.00{1} \)

Effects of Fire Injury on Constitutive and Induced Volatile Composition

Fire injury influenced the proportions of several volatiles in constitutive phloem tissue (Table 1). Concentrations (mg/g dry phloem) are listed in Supplementary Material: Tables 3 and 4. Proportions of (−) α-pinene, (+) α-pinene, (−) β-pinene, (+) limonene, and α-terpineol increased with fire injury by 3%, 1%, 7%, 0.6%, and 0.3%, respectively. Proportions of myrcene, β-phellandrene, terpinolene, 4-allylanisole, and (−) trans-caryophyllene decreased with fire injury, by 2%, 7%, 1%, 0.8%, and 2%, respectively. Two monoterpenes showed a non-linear response to fire injury: (+) 3-Carene was detected only in the low fire injury category, and (+) β-pinene was highest in the unburned category and lowest in the low fire injury category. (−) Camphene, (+) camphene, α-terpinene, linalool, borneol, geraniol, and bornyl acetate showed no change in percentage with fire injury.

Fire injury also affected the percentages of individual compounds within induced phloem tissue (Table 2). Proportions of the monoterpenes (−) α-pinene, (−) β-pinene, and (+) limonene increased with fire injury, by 2%, 9%, and 0.6%, respectively. Proportions of myrcene, β-phellandrene, 4-allylanisole, and (−) trans-caryophyllene decreased with fire injury by 1%, 11%, 0.4%, and 0.6%, respectively. (+) Camphene was highest in the low injury category but not detected in the high injury category. (−) Camphene likewise was not detected in the high injury category. (+) β-Pinene was highest in the unburned category and lowest in the moderate category. Linalool was detected only in the low and moderate injury categories. Terpinolene was highest in the high fire injury category and lowest in the moderate category. Borneol was not detected in uninjured trees, and geraniol was not detected in the highly-injured trees. Percentages of (+) α-pinene, α-terpineol, bornyl acetate, and trans-β-ocimene in induced tissue were not affected by fire injury.

Effects of Fire Injury on Mountain Pine Beetle Entry

There was no statistical difference among fire injury categories in the percentage of trees entered by caged female mountain pine beetles (Fig. 2a, χ2 = 3.51, df = 3, P = 0.320), nor in the percentage of beetles that entered these trees (Fig. 2b, χ2 = 4.58, df = 3, P > 0.250). The overall trend, however, was of a parabolic relationship, with highest entry into lightly injured trees. Incorporating other variables did not yield statistically significant explanatory variables of mountain pine beetle entry (Table 4, R2 = 0.012, F(3,36) = 1.15, P = 0.341).
https://static-content.springer.com/image/art%3A10.1007%2Fs10886-011-0031-4/MediaObjects/10886_2011_31_Fig2_HTML.gif
Fig. 2

Entry of caged female mountain pine beetles (MPB) into lodgepole pines that had experienced varying degrees of wildfire injury. a. Average percentage of lodgepole pines entered by mountain pine beetles (χ2 = 3.51, df = 3, P = 0.320). b. Average percentage of mountain pine beetles entered per tree (χ2 = 4.58, df = 3, P > 0.250)

Table 4

Linear regression model, with logit transformation, relating incidence of mountain pine beetle entry to fire injury category

Variable

Estimate (x)

Std. Error

P

βo (intercept; unburned category)

0.500

0.159

0.003

β1 (low category)

0.200

0.225

0.380

β2 (moderate category)

−0.100

0.225

0.659

β3 (high category)

−0.200

0.225

0.380

Model Was Selected Based on Best Fit by R2. These Results Are Fitted to Equation, \( { \ln }\left( {\mu /{1} - \mu } \right) = {\beta_0} + {\beta_1}{{\mathbf{x}}_{{\mathbf{1}}}} + \ldots + {\beta_k}{{\mathbf{x}}_{{\mathbf{k}}}}.\;{R^2} = 0.0{12},\;{F_{{(3,36)}}} = {1}.{15},\;P = 0.{341} \)

Discussion

These results indicate that fire injury reduced the induced, but not constitutive quantities of total monoterpenes and other less abundant volatiles available to defend against attack by mountain pine beetles and its associated microorganisms. Further, only a small amount of injury was needed to exert this effect. This reduction in the ability to undergo induced accumulation of monoterpenes and additional volatiles at an attack site corresponds with increased likelihood of successful attack by mountain pine beetle in lodgepole pines injured by wildfire in this same region (Powell et al., 2012). Induced accumulation of monoterpenes likewise has been shown to be compromised by a broad range of biotic and abiotic stress agents in multiple conifer-bark beetle systems (Reviewed in Raffa et al., 2005). Similarly, a recent study by Boone et al. (2011) identified induced monoterpene accumulation as the best overall predictor of lodgepole pine defense against mountain pine beetle.

Fire injury also affected the proportions of individual compounds within trees. Among the volatiles affected in constitutive, induced, or both phloem tissues, (−) α-pinene, (+) limonene, myrcene, 4-allylanisole, and terpinolene, are known to affect several aspects of bark beetle colonization, such as entry (Coyne and Lott, 1976), survival (Raffa and Berryman, 1983), and pheromone communication (Wood, 1982; Hayes and Strom, 1994; Erbilgin et al., 2007; Borden et al., 2008; Blomquist et al., 2010). 4-Allylanisole inhibits mountain pine beetle aggregation (Hayes and Strom, 1994), so its reduction in both constitutive and induced phloem tissue with fire injury potentially could create an environment more conducive for mountain pine beetle success. Likewise, the increase in (−) α-pinene with fire injury could benefit the beetle, as this compound is a precursor of mountain pine beetle aggregation pheromone during biosynthesis (Wood, 1982). However, the relative reductions in myrcene and terpinolene with fire injury could have the opposite effect, as both of these compounds synergize the attractiveness of the mountain pine beetle’s aggregation pheromone (Borden et al., 2008). Overall, the compositional changes in individual volatiles caused by fire injury appeared potentially more important in constitutive, rather than induced phloem tissue. We currently lack sufficient information with which to characterize these proportionate changes as programmed tree reactions to specific agents that predispose them to bark beetle attack, generalized stress reactions, beetle adaptations to chemical changes that accompany stress, idiosyncratic events, or various combinations thereof. For example, Boone et al. (2011) saw some of these same changes, such as increased α-pinene and reduced β-phellandrene, in lodgepole pine following infestation by lower stem beetles, but other reactions differed.

Our caged beetle entry results showed no significant difference among fire injury categories, but identified a slight trend of more entry in the low injury category. This pattern is interesting because several studies have shown the incidence of bark beetle attack is often higher in moderately burned than in either unburned or severely burned trees (Elkin and Reid, 2004; Powell et al., 2012). We cannot discount the possibility that larger sample sizes would have generated a statistically significant relationship, and so additional research is necessary. Additionally, future research should evaluate the relative ability of entered beetles in healthy vs. fire-injured trees to elicit attraction in the field.

Acknowledgements

This study was funded by the National Science Foundation DEB-0816541, McIntire Stennis WIS0469, Joint Fire Sciences Program 06-2-1-20, and the University of Wisconsin College of Agricultural and Life Sciences. Permission to conduct research within Yellowstone National Park was kindly provided by the National Park Service, especially Cindy Hendrix. Lodging and support during the field season was provided by the University of Wyoming and the University of Wyoming, National Park Service Research Center. Field and laboratory assistance was provided by Christopher Pennings, Andrew Long, and Hillary Thompson. We thank Joerg Bohlmann, Univ. British Columbia and Celia Boone, Univ. Northern British Columbia for analytical chemistry guidance. Helpful reviews by two anonymous referees improved this manuscript and are much appreciated.

Supplementary material

10886_2011_31_MOESM1_ESM.doc (116 kb)
ESM 1(DOC 116 kb)

Copyright information

© Springer Science+Business Media, LLC 2011