SUSCEPTIBILrrY ]p]It T)C) WESTERN BALSAM BARK BEETLE by Katherine P. Bleiker B.Sc., University of Victoria, 1995 THESIS SUBMITTED IMPARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF ivL4usrriü&(]i?:sc:iEüSK:]E in NATURAL RESOURCES MANAGEMENT © Katherine P. Bleiker, 2001 THE UNIVERSITY OF NORTHERN BRITISH COLUMBIA April 2001 All rights reserved. This work may not be reproduced in whole or in part, by photocopy or other means, without the permission of the author. 1^ 1 NmWon#! Ubf#fy of Canada BMio(h*que national# du Canada Aoqi^oNlona and BIWogmphie SarvioM Aoqua*on«#t aarvioea bMogfaphlquaa 38S W #angion Str««t OUmwON K1A0N4 395. ru9 Weaington Cmn#d» OamwON K1A0N4 Canada Irngm waeeiwwee Theaulhorhasgnnledanooexchmve hcence allowmg 6 e National lib faiy of Canada to rqxrodoce, loan, dütiiboteorscll c<^es o f dns theais m miaoAnn, paper or electrooic Annats L’auteur a accordé une Imence non exclusive permettant à la Bibhodiëque natkmale du Canada de rquoduire, prêter, disuAoer ou vœdre des oqûes de cette A&se sous lafrume demicrofrdM/Ghn,de rqirodoctionaurpagMerousarAnnat aectrmûque. The author retains ownaship of the copyright indus Aesis. Neither the Aesis nor substantial extracts from it n^y be printed or otberwise reproduced wiAout Ae author’s pennission. L’auteur conserve la propndté du droit d’auteur qui protège cette thèse. Ni la Aèse ni des extraits ndutantids de celle-ci ne doivent être mqrimés ou autrement reproduits sms sm autwisadon. 0-612-80706- CanadS APPROVAL Name: Katherine P. Bleiker Degree: Master of Science Thesis Title: SUSCEPTIBILITY OF SUBALPINE FIR TO WESTERN BALSAM BARK BEETLE Examining Committee: Chair: Dr. Deborah jP^ff Acting Dean of Graduate Studies UNBC Supervisor: Dr. B. Staffan Lindgren Associate Professor, Forestry Program Committee Member: Dr. Lorraine E. Maclauchlan Kamloops Regional Entomologist British Columbia Ministry of Forests ftee Member: Dr. Katherine J. Lewis Associate Professor, Forestry Program Committee Member: Dr. Peter MacMillan Assistant Professor, Education Program UNBC External Examirfcr: Research Scientist.,(retWd) Canadian Forest Service, Victoria, BC Date Approved: Abstract The western balsam bark beetle, Dryocoetes confusus Swaine, is the most destructive insect pest of subalpine fir, Abies lasiocarpa (Hook.) Nutt., in British Columbia (Garbutt 1992), causing scattered mortality over large areas. In order to effectively manage bark beetle populations to meet forestry objectives, it is important to understand the nature of host susceptibility. Tree characteristics were compared between a total of 22 successfully attacked, 26 unsuccessfully attacked, and 28 control (unattacked) trees at 3 sites in the interior of British Columbia. Of 12 tree characteristics measured, five showed significant differences between successfully attacked and control trees. Successfully attacked trees had a lower percent of the bole covered with constant crown (consistent branching), lower crown volume, lower radial growth in the last 5 years, were older, and produced less induced resinosis than control trees. Larger diameter trees were also more likely to be attacked than smaller diameter trees. The relationship between recent radial growth and resistance was selected for further study. A total of 26 fast- and 26 slow-growing subalpine fir trees were pheromone baited to induce attack by western balsam bark beetle at 2 sites in the interior of British Columbia. Although all baited trees were attacked, slow-growing trees were more likely to be successfully attacked than fast-growing trees. Fast-growing trees were more likely to produce resin, and in greater quantities, in response to attack than slow-growing trees. A total of 20 fast- and 20 slow-growing subalpine fir trees were inoculated with a blue-stain fungus associated with western balsam bark beetle at one site in central British Columbia. Dimensions of the light and dark areas of the resultant lesions were compared between fast- and slow-growing trees and between fungus and control treatments at 3, 7, 10, 11 17, and 41 days after inoculation. The length and width of the dark and light areas were greater in response to fungus versus control treatments at 7, 10, 17, and 41 days after inoculation. The length of the dark and light area was significantly greater in fast-growing trees than in slow-growing trees at 7, 10, and 17 days after inoculation. There was no significant difference in the size of the lesion between fast- and slow-growing trees 41 days after inoculation. This research shows that slow-growing trees are more susceptible to attack by western balsam bark beetle and have reduced defensive capabilities compared to fastgrowing trees. These findings may explain the scattered pattern of mortality observed in stands infested with western balsam bark beetle. Western balsam bark beetle’s preference for low vigour hosts and its frequent inability to induce mass attack on high vigour hosts has implications for the management of subalpine fir stands in the interior of British Columbia. Ill Table of Contents Abstract............................................................................................................................................ii Table of Contents................ iv List of Tables.................................................................................................................................. vi List of Figures..................................................................... viii Acknowledgements.......................................................................................................................... x Chapter L Introduction................................................................................................................1 Background and Objectives ..... 1 Distribution and Host of Western Balsam Bark Beetle ............................... 2 Life History of Western Balsam Bark Beetle ...............................................4 Role of Blue-Stain Fungi Associated with Bark B eetles............................. 5 Fungi Associated with Western Balsam Bark Beetle .................................. 7 Economic Impact of Western Balsam Bark Beetle in British Columbia ....8 Management of Western Balsam Bark Beetle in British Columbia ......... 10 Chapter II. Characteristics of Subalpine Fir Susceptible to Western Balsam Bark Beetle....................................................................................................................... 11 Introduction..................................................................................................... 11 Methods............................................................................................................12 Results..............................................................................................................18 Discussion........................................................ 25 Conclusions ............................................................................................. 30 Chapter m . Success of Pheromone-Induced Attack by Western Balsam Bark Beetle and the Defense Response of Fast- and Slow-Growing Subalpine Fir . .32 Introduction........................................................................... 32 Methods ........................................................................ .....33 Results ...... 37 Discussion....................................................................................................... 41 Conclusions.....................................................................................................44 Chapter IV. Defense Response of Fast- and Slow-Growing Subalpine Fir to Inoculation with a Blue-Stain Fungus............................................................ 45 Introduction.....................................................................................................45 Methods........................................................................................................... 46 Results........................................................................... 51 Discussion................................................... 61 Conclusions....................... 65 Chapter V. Conclusions and Recommendations...................................................................66 IV Literature Cited..................................................................................................................... 73 Appendix 1. Pearson correlation matrices of growth indices of successful attacks, unsuccessful attacks, and controls by site...................................... ..................... 86 Appendix 2. Pearson correlation matrices of tree characteristicsof successful attacks, unsuccessful attacks, and controls by site. ...... 87 Appendix 3. A comparison of tree characteristics between sites....................................... 89 V List of Tables Chapter H. Characteristics of Subalpine Fir Susceptible to Western Balsam Bark Beetle Table 1. Description of characteristics used for rating resin streaming....................................13 Table 2. A comparison of tree characteristics between control trees, unsuccessful attacks, and successful attacks................................................................. 22 Table 3. Number o f trees in each resin category by attack class at each site.................... 23 Table 4. Number of trees with rot present/absent by attack class at each site...................23 Table 5. Number of trees in each crown foliage category by attack class at each site..... 23 Chapter m . Success of Pheromone-Induced Attack by Western Balsam Bark Beetle and the Defense Response of Fast- and Slow-Growing Subalpine Fir Table 6. Description of unsuccessful, light, and successful attack classes.............................35 Table 7. Number (percent) of fast- and slow-growing trees with resin present/absent for Lumby and Milk River research sites, 6 to 7 weeks after initial attack..............38 Table 8. Number and attack status of fast- and slow-growing baited trees by resin production for Lumby and Milk River research sites............................. 38 Table 9. Number (percent) of fast- and slow-growing baited trees classified as unsuccessful, light, or successful attacks for Lumby and Milk River research sites................................................................................................................................. 39 Chapter IV. Defense Response of Fast- and Slow-Growing Subalpine Fir to Inoculation with a Blue-Stain Fungus Table 10. Number of reisolates from control treatments that produced Ophiostoma A for fast- and slow-growing trees. .................................................................................52 Table 11. Percent of reisolates contaminated with other fungi, moulds, bacteria, and/or yeasts at each sample time by treatment group...................... 52 VI Table 12. ANOVA results for overall effects of inoculation treatment and tree growth on lengths and widths of dark and light areas.................................... 61 Table 13. Results of hypothesis testing for effects of inoculation treatment and tree growth on length and width of dark and light areas at each sample time............. 61 Vll List of Figures Chapter I. Introduction Figure 1. Area infested by western balsam bark beetle annually in BC from 1985-1995. Data compiled from Wood and Van Sickle (1986, 1987a, 1987b, 1989a, 1989b, 1990, 1991, 1993a, 1993b, 1994), and Humphreys and Van Sickle (1996)..............................................................................................................................9 Chapter H. Characteristics of Subalpine Fir Susceptible to Western Balsam Bark Beetle Figure 2. Diameter distribution of tree species at 3 sites........................................................... 19 Figure 3. Standing subalpine fir by diameter class and attack status at 2 sites. Cherry Ridge has been excluded due to low sample size........................................20 Figure 4. Mean (SE) number o f (A) 1998 successful attacks, (B) 1998unsuccessful attacks, (C) faders and red attacks, (D) subalpine fir over 12.5 cm dbh, and (E) other species over 12.5 cm dbh, occurring within a 20 m radius of control trees, unsuccessful attacks, and successful attacks at 3 sites........................ 24 Chapter IV. Defense Response of Fast- and Slow-Growing Subalpine Fir to Inoculation with a Blue-Stain Fungus Figure 5. Morphology of the induced reaction 17 days after inoculation showing point of inoculation (A), ring of darkened tissue around inoculation hole (B), light area around inoculation hole (C), dark colored area (D), light colored, translucent looking area (E), and dark staining below sapwood surface (F)........ 55 Figure 6. Comparison of the induced reaction 7 days after inoculation in response to fungus inoculation (A) and control treatment (B)............ 55 Figure 7. Comparison of mean (SE) lesion length on the sapwood surface at each sample time for fungus and control treatments on fast- and slow-growing trees for the dark and light area of the lesion........................................ 56 Figure 8. Comparison of mean (SE) lesion width on the sapwood surface at each sample time for fungus and control treatments on fast- and slow-growing trees for the dark and light area o f the lesion.................................................... 57 vm Figure 9. Comparison of mean (SE) vertical rate of lesion development between sample times for fungus and control treatments on fast- and slow-growing trees for the dark and light area of the lesion....................................... 58 Figure 10. Comparison of mean (SE) lateral rate of lesion development between sample times for fungus and control treatments on fast- and slow-growing trees for the dark and light area o f the lesion....................................................................... 59 IX Acknowledgements I sincerely thank my supervisor, Dr. B.S. Lindgren, for his support, guidance, and patience, which were exceptional during the writing process. I thank my committee; Dr. K. Lewis for discussions and use of lab facilities; Dr. P. MacMillan for statistical support; and Dr. L. Maclauchlan for providing this research opportunity, the use of lab facilities and equipment, as well as for invaluable input. I am grateful to D. Ayers for statistical assistance and appreciate the amount of time he gave to my enquiries. I thank the following for lab and/or field assistance: S. Nesbitt, K. Buxton, S. Moraes, T. Rimmer, and S. Collingridge of the Kamloops Forest Region; T. Gainer and P. Hulka of Zeidler Forest Industries; C. Gauthier and K. Lagrandeur of Bug Busters Pest Management; and J. Clarke and A. Maclsaac of UNBC. I thank Brodie for security alerts and humour in the field. I am grateful to C. Ferguson for helping me complete a long first field season under difficult conditions in short notice. I thank B. Setter of Bug Busters and H. Mueller of Zeidler for supporting the project. I am extremely grateful to A. Uzunovic of Forintek Canada Corp. for his assistance with fungi isolations and detailed explanations of lab and field procedures. I thank K. Glowa, K. Williams, and the technicians in lab 403 for their advice. I thank Seong Hwan Kim of UBC, and K. Seifert and K. Jacobs of Agriculture Canada for their efforts in fungus identification, which is in progress at the time of writing. Funding for this project was provided by the following: Science Council of BC/Forest Renewal BC #FR-96/97-829 to Dr. L. Maclauchlan, Kamloops Forest Region; Natural Sciences and Engineering Research Council of Canada OGP #0194765 to Dr. B.S. Lindgren, UNBC; Zeidler Forest Industries; and Bug Busters Pest Management. The Entomological Society of BC and UNBC provided conference travel awards. SUSCEPTIBILITY OF SUBALPINE FIR TO WESTERN BALSAM BARK BEETLE Chapter I. Introduction Background and Objectives The western balsam bark beetle, Dryocoetes confusus Swaine (Coleoptera: Scolytidae), is the most destructive pest of subalpine fir\ Abies lasiocarpa (Hook.) Nutt., in British Columbia (Garbutt 1992). Subalpine fir is becoming an increasingly important commercial species due to depletion of mature timber at low elevation. Therefore, effective strategies and tactics for the management o f western balsam bark beetle and subalpine fir are needed to ensure efficient use of the resource. Research on western balsam bark beetle has focused primarily on developing an effective pheromone bait for the beetle (Stock 1981, 1991; Stock and Borden 1983; Borden et al 1987; Stock et al 1994a, 1994b; Camacho 1993; Camacho and Borden 1994). Pheromone baits have been used experimentally and commercially to hold and concentrate beetles in selected trees (and stands) that are subsequently harvested (Stock 1991; Stock et al 1994b; Harder 1998; Jeans-Williams 1999; Phero Tech Inc. undated). One project in British Columbia is currently examining stand hazard, or stand susceptibility, based on stand structure, biogeoclimatic zone, and inventory data (Maclauchlan^, personal communication). However, little work has been done on the nature o f the relationship between western balsam bark beetle and subalpine fir. ^ Hunt (1993) differentiates Rocky Mountain subalpine fir {A. bifolia A. Murray) from subalpine fir (A. lasiocarpa). As A. bifolia has not been passed to date as a new species by the International Botanical Congress and is not widely used in current literature, it will not be used in this thesis. " Dr. Lorraine Maclauchlan, Entomologist, Kamloops Forest Region, BC Forest Service, Kamloops, BC. The ability to identify trees susceptible to attack has played a key role in managing other damaging insects. However, differences in susceptibility of subalpine fir to western balsam bark beetle has not been examined at the tree level. Understanding susceptibility at the tree level would aid in the identification of high hazard stands, as well as the development of effective stand- and landscape-level management options. For the purpose o f this research, host susceptibility or resistance was defined as the characteristics or qualities of a tree that affect its likelihood of being selected for attack and/or being damaged from attack by an insect (modified from Shore and Safranyik 1992). The objective of this research was to identify and examine characteristics of subalpine fir susceptible to western balsam bark beetle. The remaining sections of Chapter 1 review literature pertaining to this research. Chapter II identifies parameters associated with host susceptibility, and Chapters III and IV present two experimental approaches used to test the effect of one of the identified parameters on bark beetle attack and fungus inoculation, respectively. Chapter V synthesizes the results of the three studies, suggests directions for future research, and reviews potential implications of this research on the management of western western balsam bark beetle in British Columbia. Distribution of Western Balsam Bark Beetle and Silvical Characteristics of its Host Western balsam bark beetle is found throughout the range of its primary host, subalpine fir. Attacks on amabilis fir {Abies amabilis (Dougl.) Forb.) have occasionally been reported, while Engelmann spruce (Picea engelmanni Parry), white spruce (P. glauca (Moench) Voss), and lodgepole pine {Pinus contorta Dougl. var. latifolia Engehn.) are rarely attacked (Molnar 1965; Garbutt 1992). Subalpine fir grows in high-elevation western forests. Its range extends from Alaska through the Yukon Territory and British Columbia to Oregon via the Cascade Range, and to New Mexico via the Rocky Mountains (Harlow and Harrar 1958). Although subalpine fir may be present in other ecosystems, it is most common in the Engelmann Spruce-Subalpine Fir (ESSF), Montane Spruce (MS), Spruce-Boreal Spruce (SBS), and Spruce-Willow-Birch (SWB) biogeoclimatic zones in British Columbia (Meidinger and Pojar 1991). In British Columbia, subalpine fir grows in pure or mixed stands, where it is commonly associated with Engelmann spruce, white spruce, or lodgepole pine (Meidinger and Pojar 1991). Environments dominated by subalpine fir are characterized by short growing seasons, cool summers, and very cold winters, during which soils are commonly frozen (Meidinger and Pojar 1991). Climax spruce-subalpine fir stands in British Columbia are generally uneven-aged with both species present in the canopy and understorey (Bier et al. 1948). In unevenaged stands, subalpine fir trees are often suppressed at an early age and suffer from decay, which increases with age (Bier et al. 1948). In contrast, subalpine fir from evenaged stands are usually less than 120 years old, rarely show signs of early suppression, have less decay, and reach saw log size by 100 years of age (Bier et al. 1948). Subalpine fir has a pathological rotation age of approximately 130 years (Bier et al. 1948). Subalpine fir is susceptible to various root and butt rots, however, the majority of damage is caused by heart rot fungi. Red heart rot {Stereum sanguinolentum (Alb. and Schw.; Fr.) Fr.) and brown stringy heart rot (Echinodontium tinctorium (Ell. and Ev.) Ell. and Ev.) are the most common causes of decay in subalpine fir in British Columbia (Bier et al. 1948; Smith and Craig 1970). Susceptibility to these fungi increases with age, and is promoted by wounding, suppression, and shade-killed branchlets (Bier et al. 1948; Maloy and Robinson 1968; Etheridge and Craig 1970). Although western balsam bark beetle is the most destructive insect pest of subalpine fir, other insects may also cause damage. Where their distributions overlap in British Columbia, subalpine fir is defoliated by two-year cycle budworm (Choristoneura biennis Free.), eastern spruce budworm {Choristoneura fumiferana Clemens), and western blackheaded budworm (Acleris gloverana Wlshm.) (Henigman et al. 1999). Budworm damage usually results in growth reduction or stem defects, but successive years of defoliation may result in tree death. Defoliation may also predispose trees to attack by other insects. Attack by the balsam woolly adelgid (Adelges piceae Ratz.) may result in growth reduction, dead tops, and even tree death; however, it is only present in southwestern British Columbia (Henigman et al. 1999). Life History of Western Balsam Bark Beetle The western balsam bark beetle normally requires two years for the completion of its life cycle (Mathers 1931). However, Bright (1963) refers to a one-year life cycle in the western and southwestern United States, which may be due to warmer temperatures. The main flight typically commences between mid-June and mid-July when temperatures within the stand reach 15°C (Mathers 1931; Stock 1991; Hansen 1996). A second, smaller flight, which is dominated by females, occurs in August, but little flight activity takes place between flight periods (Stock 1991; Hansen 1996). The second flight may be absent or negligible at cooler sites, or during cooler summers (Hansen 1996; Gibson et al. 1997). Pioneering males, responding to host volatiles, initiate attack and then produce pheromones that attract other males and females (Stock et al. 1983). The polygamous male excavates a nuptial chamber in the phloem or cambium where it mates on average with three to four females (Bright 1963). Mated females produce an anti-aggregation pheromone, which partially inhibits the response of both sexes to the male-produced pheromone (Stock and Borden 1983). Mated females excavate radiate egg galleries in the cambium, depositing eggs in niches along gallery walls as they progress. Parent beetles overwinter in the host. The following spring females may extend their egg galleries and lay additional eggs, before both sexes emerge and establish a third brood in a new host (Mathers 1931). Brood diapause as larvae the first winter and callow adults the second winter (Mathers 1931). Brood adults emerge the following spring or summer, thus completing the life cycle in two years. Role of Blue-Stain Fungi Associated with Bark Beetles Most species of bark beetles are closely associated with blue-stain fungi. Bark beetles transport spores externally on their bodies, and in specialized microbial transport structures called mycangia. Farris (1969) made the first record of mycangia in the genus Dryocoetes, after describing an oral pouch on each mandible of both male and female western balsam bark beetles, which contained spore-like objects. However, there was no record of the contents of the mycangia being cultured and positively identified. The relationship between blue-stain fungi and bark beetles was originally thought to be one of mutualism (Whitney 1982); the fimgi gain access to an otherwise inaccessible habitat, and in return immobilize host defenses against attacking beetles by interrupting translocation. Adults and developing brood may also benefit Ifom the presence of fungi through conditioning of the host substrate, as a food source (Craighead 1928; Whitney 1982), and in the production of pheromones (Brand et al. 1976). Recent research has questioned the mutuality of the relationship between bluestain fungi and bark beetles. Although many species of blue-stain fungi associated with tree-killing bark beetles are pathogenic to their respective hosts when inoculated at sufficiently high densities (Molnar 1965; Homtvedt et al. 1983; Christiansen and Solheim 1990,1994; Solheim and Safranyik 1997; Krokene and Solheim 1998), the ability of blue-stain fungi to kill trees before attacking beetles effectively girdle the trees has been questioned (Parmeter et al. 1992; Hobson et al. 1994). Furthermore, active populations of southern pine beetle, Dendroctonus frontalis Zimm., have been discovered without their associated blue-stain fungus, O. minus (Hedgcock) H. and P. Sydow (Bridges et al. 1985). Other studies have demonstrated that while progeny of the southern pine beetle was more successful in the presence of the associated mycangial fungi complex (Barras 1973; Goldhammer et al. 1990), it was reduced in the presence of the O. minus (Barras 1970; Franklin 1970; Goldhammer et al. 1990). In contrast, brood of the mountain pine beetle, Dendroctonus ponderosae Hopk., was not inhibited by the presence of its fungal associates, O. ips (Rumbold) Nannf. or O. claverigerum (Robinson-Jefffey and Davids.) Harrington (Nevill and Safranyik 1996). In very well developed mycangia, other species of fungi may predominate over the more virulent blue-stain fungi as well as inhibit their development in the host tree (Wingfield et al. 1995). Attacking beetles introduce a wide range of microorganisms into the host tree, including numerous species of non-staining fungi, yeasts, and bacteria. The role these organisms play in beetle development and their interactions with pathogenic blue-staining fungi has not been intensively studied. The relationship between blue-stain fungi, bark beetles, and their hosts is complex and may vary between bark beetle-fungus-host systems. Fungi Associated with Western Balsam Bark Beetle A number of fungi have been found to be associated with western balsam bark beetle. The entomopathogen Beauveria bassiana (Bals.) Vuill. kills adults after they establish egg galleries (Whitney et al. 1984). Because B. bassiana does not readily colonize phloem or sapwood tissues in attacked trees, it is thought to compete poorly with other fungi that are introduced by attacking beetles (Whitney et al. 1984). A number of species of blue-stain fungi have been isolated from adult western balsam bark beetles or tissues taken from attacked subalpine fir. Ophiostoma nigrum (Davidson) de Hoog and Scheffer, O. minus, and Ceratocystis brunnea Davidson, have been isolated from perithecia (sexual reproductive structure) fruiting in insect galleries in dead or dying subalpine fir, and may be associated with western balsam bark beetle or associated insects (Davidson 1958). O. abiocarpum (Davidson) Harrington, was isolated from adult beetles, as well as from perithecia in the bark of dead trees (Davidson 1958, 1966). A species of the genus Leptographium was commonly identified in association with blue-stain fungi in subalpine fir, but no perithecial stage was definitely connected with it (Davidson 1958). O. dryocoetidis (Kendrick and Molnar) de Hoog and Scheffer has been associated with adult western balsam bark beetles and isolated from necrotic tissues in unsuccessfully attacked trees (Molnar 1965). Molnar (1965) estimated that 65% of the mortality associated with western balsam bark beetle was due to O. dryocoetidis. Unsuccessful attacks by beetles may allow for the successful introduction of the fungi into the host. Field inoculations of O. dryocoetidis into subalpine fir resulted in the formation of lesions similar to those observed in unsuccessfully attacked trees and demonstrated its pathogenicity (Molnar 1965). As a result of Molnar's work, mortality of subalpine fir attacked by western balsam bark beetle has been attributed to the beetlefungus complex (Garbutt 1992). Economic Impact of Western Balsam Bark Beetle in British Columbia Western balsam bark beetle is the only species in the genus Dryocoetes that is capable of attacking living trees and causing significant economic losses (Bright 1963). Attacks also frequently occur on windblown and felled hosts (Stock 1991; personal observation). Western balsam bark beetle typically kills individual trees or small groups of trees that are randomly dispersed throughout the infested area (Stock 1991). In British Columbia subalpine fir comprises approximately 13% (70 713 000 m^) of the volume of all products billed^ (BC Ministiy of Forest 1999). Losses due to western balsam bark beetle are difficult to estimate due to the patchy nature of mortality and difficulty in interpreting the age of attack based on the colour of red needles, which may be retained for at least 5 years (Wood and Van Sickle 1989b). In beetle-infested stands, annual levels of mortality are generally less than 6% of mature subalpine fir (Stock 1991; Unger and Stewart 1992), but cumulative losses may exceed 30% of stand volume (Wood and Van Sickle 1991). Although stand-level mortality caused by western balsam bark beetle may not be as high as other tree-killing bark beetles during epidemics, the extent and duration of the infestations result in significant mortality over time. Figure 1 shows the number of hectares infested by western balsam bark beetle annually in BC from 1985-1995. Like ^ Volume for which stumpage has been billed. Includes waste and firmwood rejects. 225 g 200 1 175 2O o 150 o 125 -o 100 I I 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 Year Figure 1: Area infested by western balsam bark beetle annually in BC from 1985-1995. Data compiled from Wood and Van Sickle (1986, 1987a, 1987b, 1989a, 1989b, 1990,1991,1993a, 1993b, 1994), and Humphreys and Van Sickle (1996). most insects, western balsam bark beetle populations fluctuate, however, annual changes in the amount of area surveyed accounts for part of this variation (Wood and Van Sickle 1989b). Annual inconsistencies in survey coverage, and the recent inclusion of some chronically infested areas in the Prince Rupert region during aerial surveys, make it difficult to identify changes in infestation levels over time. In 1995/1996 western balsam bark beetle infested 128 000 ha province wide, compared to 165 129 ha infested by mountain pine, spruce, and Douglas-fir beetles combined (BC Ministry of Forests 1996). Seventy to 80% of the area infested by balsam bark beetle (Figure 1) usually occurs in the Prince Rupert region (Wood and Van Sickle 1986, 1987a, 1987b, 1989a, 1989b, 1990, 1991, 1993a, 1993b, 1994; Humphreys and Van Sickle 1996). Economic losses are compounded in this region, as subalpine fir comprises a quarter of its harvested timber volume (BC Ministiy of Forests 1999). As other regions increase harvesting in higher elevation areas, the economic impacts of western balsam bark beetle will spread throughout the province. Management of Western Balsam Bark Beetle in British Columbia Stock (1981) first provided evidence of primary and secondary attraction in western balsam bark beetle. Beetle behaviour may be manipulated at the stand level using the pheromone components exo-brevicomin, which has aggregative properties, or eWo-brevicomin, which has anti-aggregative properties (Stock and Borden 1983; Stock 1991). Further analysis o f beetle-produced volatiles has led to the development of more attractive baits using blends o f exo- and eWo-brevicomin (Camacho 1993). Field tests of the original single component bait versus the new blended bait revealed no difference in their ability to induce attack on baited trees, however, catches in pheromone-baited funnel traps were higher with the new bait (Jeans-Williams 1999). Current management tactics involve baiting blocks 1 to 2 years before harvest on a 50 m grid, using two-tree bait centers (Stock et a l 1994b; Harder 1998; Greenwood 1998). This approach is effective in concentrating new beetle attack around bait centers and reducing new attacks in a peripheral area for at least 50 m (Stock et a/. 1994b; Harder 1998; Jeans-Williams 1999). The distance from which pheromone baits may effectively draw beetles has not been determined. As western balsam bark beetle infestations often cover large areas, management strategies need to be developed at the landscape level. In order to manage at the landscape level it is necessary to understand what trees and stands are most susceptible to attack. Stock (1991) identified the need to determine the factors affecting the susceptibility of subalpine fir to western balsam bark beetle ‘before intelligent management objectives can be formulated’. 10 Chapter H. Characteristics of Subalpine Fir Susceptible to Western Balsam Bark Beetle Introduction Most species o f bark beetles inhabit downed or dying trees (Rudinsky 1962). At high population levels, some species have the ability to mass attack and kill healthy hosts (Fumiss and Carolin 1977). The specific events that trigger insect outbreaks are the subject of numerous studies, but are not well understood for most species. Epidemics of many tree-killing bark beetles have been linked to substantial increases in susceptible host material (Berryman 1972). Past studies exploring the nature of tree and stand susceptibility for major tree-killing bark beetles, such as mountain pine beetle (Dendroctonusponderosae Hopkins) (Amman 1972; Waring and Pitman 1983, 1985; Shore and Safranyik 1992), spruce beetle (Dendroctonus rufipennis Kirby) (Hard et al. 1983; Safranyik ef u f 1983), and Douglas-6r beetle (Dendroctomts jisgWot&ugae Hopkins) (Fumiss et al. 1981; Shore et al. 1999), have contributed significantly to the development of relevant management strategies and options. Western balsam bark beetle selectively kills small groups of subalpine fir at a relatively low, but constant, level each year in infested stands (Stock 1991; Unger and Stewart 1992). Although cumulative mortality may reach significant levels in chronically infested stands (Garbutt and Stewart 1991), western balsam bark beetle is less aggressive than other tree-killing bark beetles at epidemic levels. The selective and patchy distribution of mortality suggests that western bark balsam beetle may be limited by the abundance and distribution of susceptible hosts. Identifying the characteristics of 11 susceptible hosts may contribute to a broader understanding of the ecology of the beetle and aid in developing effective management practices. Trees in the genus Abies lack extensive vertical resin canals (Barman 1936) and rely on induced resinosis to defend against attack by bark beetles and fungi. Host vigour has been identified as the main factor affecting the ability of a tree to defend itself (Berryman 1982a). Tree vigour, usually indicated by measures of radial growth, may be influenced by senescence, defoliation, pathogens, as well as other factors that cause stress (Coulson 1979; Kaufman and Ryan 1986; Waring 1987; Yoder et al. 1994; Nebeker et al. 1995). The objective of this study was to identify characteristics of subalpine fir trees susceptible to western balsam bark beetle. Methods Field sampling was conducted at 3 sites in the ESSF biogeoclimatic zone of British Columbia during the summers of 1998 and 1999. The Cherry Ridge (UTM 393300 5573300) and Lumby (UTM 362300 5551700) sites were located in the Thompson-Shuswap Highlands in the Vernon Forest District. The Cherry Ridge site was in the wet cold (wc2) subzone of the ESSF, where growing season moisture deficits are rare due to late snow melt and frequent summer storms (Lloyd et al. 1990). The elevation at this site was 1 650 m with a slope of less than 10%. The Lumby site was in the very dry cold (xc) subzone of the ESSF, which receives approximately half the annual precipitation of the wc2 (Lloyd et al. 1990). The elevation at this site was 1 700 m and the slope was less than 10%. The Milk River site (UTM 5914800 655000) was located in 12 the Rocky Mountain Trench in the Robson Valley Forest District. This site was in the transition zone between the moist mild (mm) subzone of the ESSF and the moist cold (me) subzone of the interior cedar-hemlock (ICH) zone (Meidinger et al. 1988). The elevation was 1 050 m and the overall slope was less than 10%. Mature subalpine fir was the dominant species at all sites. In August and September 1998, a number of transects 10 m wide and 50 m apart were established at each site. All subalpine fir attacked in 1998 occurring in the transects were classified as either unsuccessful or successful attacks. Unsuccessful attacks were characterized by resin streaming on the bole, and very little or no ffass present. Successful attacks were characterized by the accumulation of ffass at the base, and moderate, light or no resin streaming on the bole. Control trees were selected for half of the 1998 attacked (unsuccessful and successful) trees by identifying the nearest live, unattacked subalpine fir of similar diameter (±10%). All attacked and control trees were numbered, and their location and dbh (diameter at breast height, or 1.3 m) recorded. Any attacked trees with new resin present on the bole, were given a resin rating of light, moderate, or heavy (Table 1). Table 1: Description of characteristics used for rating of resin streaming. Resin Rating Description of Resin Streaming Light Resin Light streaming or beads present, but a large part of bole lacks resin. Moderate Resin Streaming pronounced. Resin may only be present on a short length o f the bole, but covers circumference at that height. Heavy Resin Very thick streaming. Bark was obscured by resin in the area of the bole with streaming. 13 In 1998 a subset o f the attacked trees and their respective controls were selected for falling and intensive sampling at each site. To ensure attacked trees selected for sampling were not attacked as a result of a spillover effect from adjacent trees, all attacked trees selected for falling were located a minimum distance of 15 m apart, with at least 2 unattacked mature subalpine fir trees between them. After falling, trees were reclassified as unsuccessful or successful attacks based on the following criteria: unsuccessful attacks had no live adults present, or had a limited number of live adults stmggling in restricted pitchy galleries with no live larvae present; successful attacks had live adults and larvae present. After this reclassification, the exclusion of some misclassified 1998 unsuccessful attacks (there was resin streaming on the bole, but another insect was found mining the phloem), and the exclusion of some misclassified 1998 successftil attacks (a few were determined to be 1997 successful attacks that were just starting to fade), sample sizes were as follows: 7 unsuccessful attacks, 6 successful attacks, and 8 control trees were sampled at Cherry Ridge; 11 unsuccessful attacks, 10 successful attacks, and 12 control trees were sampled at Lumby; and 8 unsuccessful attacks, 6 successful attacks, and 8 control trees were sampled at Milk River. Sampling o f Tree Characteristics Sample measurements on each tree included dbh (cm), height (m), mean crown width (m), mean phloem thickness (mm), height to start of live crown (m), and height to start of constant crown (m). (Because the lower branches on subalpine fir are often sparse or limited to one side of the bole, the distance to the start of consistent live branching around the bole was measured and termed 'height to start of constant crown'). 14 The percent of bole with live crown (CRTTL) and percent of bole with constant live crown (CRCNT) were calculated using total tree height and the latter two measurements respectively. Mean crown width (AVCRWD) was calculated from two measurements taken at 90-degree angles to each other. Crown volume (CRVOL) was calculated based on a cone using the following formula: (7i(crown width/2 )^)(tree height-height to start of constant crown)/3. Mean phloem thickness (AVPHLM) was derived from measurements taken on the east and west aspects at 1 m, 4 m, 8 m, 12 m, and 16 m. Crown foliage was classified to one of the following conditions: ( 1 ) sparse, noticeable defoliation or a very thin looking crown; (2) medium, average looking foliage for that site; or (3) heavy, above average foliage in terms of density or lushness of needles. Stem disks were taken at 0.5 m from each felled tree, except at the Milk River site where disks were re-cut in 1999 at 1.3 m because flared and uneven growth on some of the original disks obscured the growth pattern. Two radii per disk were selected for measurement using the method outlined by Chapman and Meyer (1949). The longest diameter, from cambium to cambium, but not necessarily crossing the pith, and the diameter perpendicular to it and dissecting the pith, were measured. The sum of the diameters was divided by four, yielding the average radius of the disk, which usually occurred at only two places on the disk. Where the average radii occurred more than twice, two well-spaced radii were selected. The following growth indices were calculated using the data averaged from the two radii: cumulative growth for the last 5 years (CUM5), cumulative growth for the last 10 years (CUM 10), 5 year periodic growth rate (PGR5) (growth over the last 5 years divided by growth over the 5 years previous to that), 10 year periodic growth rate (PGRIO) (growth over the last 10 years divided by 15 growth over the 10 years previous to that), 5 by 45 year periodic growth rate (PGR5/45) (growth over the last 5 years divided by growth over the 45 years previous to that), basal area increment in the last 5 years (BAI5), and five year growth standardized by dbh (dbh divided by growth over the last 5 years). Age and canopy age (CNAGE), at 0.5 m or 1.3 m, were recorded from one of the radii. Canopy age was taken from the time when the tree showed evidence o f sustained release. In cases where there was no significant sustained increase in ring width over the radii, canopy age was the same as age. Disks were measured using Windendro software (Regent Instmments Inc., Quebec City, QC, Canada) and a Hewlett-Packard ScanJet 4c/T scanner (Hewlett-Packard Ltd., Palo Alto, CA, USA). The presence/absence o f heart rot for each tree was recorded from the sanded disks because incipient rot was often missed on rough-cut stumps in the field. A tally was made of all trees greater than 12.5 cm dbh within a 20 m radius of the stump by tree species. Each tree was classed according to one of the following conditions: (1) 1998 unsuccessful attack; (2) 1998 successful attack; (3) 1997 successful attack (fader), bright or dull red attack; (4) grey attack; (5) live, unattacked subalpine fir; or (6 ) live, species other than subalpine fir. Aaw/ DfOTMgfrr S'wrvgy In 1999, a number of strip plots, 20 m by 5 m, were systematically located along the established transects at each site. Species and dbh of every tree greater than 1.3 m in height occurring in the plot were recorded. Subalpine fir trees were classified according 16 to one o f the following conditions: (1) nnattacked, (2) 1998 unsuccessful attack; (3) 1998 successful attack; or (4) pre-1998 successful attack (fader, red or grey attack). Data from the strip plots were graphed and used to determine the species composition and diameter distribution at each site. Site differences in the dbh of attacked trees were tested using a one-factor analysis of variance (ANOVA). Pearson correlation matrices were used to examine the associations among growth indices and tree characteristics. The continuous variables dbh, height, age, CNAGE, CUM5, CRTTL, CRCNT, AVCRWD, CRVOL, and AVPHLM were analyzed for differences between attack classes (unsuccessful, successful, and control) and sites using a three-factor nested ANOVA. Attack class and site were entered in the model as fixed factors, and individual trees were nested under attack class and site and entered as a random factor. Variables were visually assessed for normality using histograms. A logarithmic transformation was applied to CRVOL prior to analysis to correct for non-normality. Due to small sample sizes frequency tables were used to identify potential relationships between attack class and the following: resin production rating, presence/absence of rot, and crown foliage rating. Data from the 20 m radius plots were examined using descriptive statistics as transformations failed to normalize most of the variables. The level of significance was set at 0.05 for all statistical tests. Significant ANOVAs were followed by Bonferroni’s multiple comparisons test. Data were analyzed using SYSTAT 9.0 (SPSS Inc., Chicago, IL, USA). 17 Results jDfa/Mgfgr The diameter survey measured 64 trees at Cherry Ridge, 354 trees at Lumby, and 468 at Milk River. Subalpine fir was the dominant tree species at each site, followed by hybrid interior spruce, Picea glauca x engelmannii Engelm. (Figure 2). Western hemlock, Tsuga heterophylla (Raf.) Sarg., was only present at the Milk River site. Mean diameter (±SE) of attacked trees was significantly different between sites, with the largest trees at Cherry Ridge (44.4 cm (±3.3)), followed by Milk River (31.2 cm (±1.0)), and Lumby (22.3 cm (±1.0)) (Fa, 103 = 39.086, P <0.001, Bonferroni MCP P < 0.001). The majority of attacks occurred in the top three diameter classes at Lumby and Milk River (Figure 3). (Due to a small sample size. Cherry Ridge was excluded from Figure 3). Western balsam bark beetle contributed to the mortality of over three-quarters of the subalpine fir in the top two diameter classes at Lumby, and half the subalpine fir in three of the larger diameter classes at Milk River (Figure 3). Tree Characteristics Following dissection of the felled trees, 2 of the 8 successful attacks were reclassified as unsuccessful attacks at Cherry Ridge, 2 of the 12 successful attacks were reclassified as unsuccessful attacks at Lumby, and 3 of the 9 successful attacks were reclassified as unsuccessful attacks at Milk River. The new classifications were used for all analyses. Tree growth indices were well correlated with each other (Appendix 1). Of the growth indices calculated, CUM5 was selected for further analyses because of its high correlation to the other growth variables, and ease of measurement. 18 (A) Cherry Ridge Subalpine fir Spruce Hemlock (B) Lumby xn I (w 0 O 1 ■ (C) Milk River 0.01-10.0 10.01-20.0 20.01-30.0 30.01-40.0 40.01-50.0 50.01-60.0 Diameter Classes (cm) Measured at 1.3 m Figure 2; Diameter distribution of tree species at 3 sites. 19 100 1998 Unsuccessful Attacks 1998 Successful Attacks Pre-1998 Successful Attacks Unattacked Subalpine Fir (A) Lumby 75 - 50 25 U I o (w 0 n 0 1 100 (B) Milk River Ph 75 50 25 JZL 0.01-10.0 10.01-20.0 20.01-30.0 30.01-40.0 40.01-50.0 50.01-60.0 Diameter Classes (cm) Measured at 1.3 m Figure 3: Standing su b a^ in efrly diameter class and attack status at 2 sites. Cherry Ridge has been excluded due to low sample size. * only 2 trees were in this dWi class. 20 Appendix 3 lists the specific differences in tree characteristics between sites. Generally, Lumby had smaller and older trees with thinner phloem than Cheny Ridge or Milk River (Appendix 3). However, mean values of 5-year recent radial growth (CUM5) did not differ between sites (Appendix 3). Of the tree characteristics analyzed, age, CUM5, CRCNT, and CRVOL showed significant differences between attack classes (Table 2). Successful attacks were significantly older than control trees, had significantly lower growth over the last 5 years, had a lower percent of their bole covered in constant crown, and had a smaller crown volume (Table 2). Although unsuccessful attacks did not differ significantly from successful attacks or controls in terms of age, 5-year cumulative growth, and percent of bole covered in constant crown, the means were in between those of suecessful attacks and controls (Table 2). O f the tree characteristics that differed significantly between attack classes, CRCNT and CRVOL were significantly correlated at all 3 sites and likely overlap in their measurement of some phenomena (Appendix 2). CUM5 and CRCNT were moderately correlated, but only at one site (Appendix 2). The remaining tree characteristics, dbh, height, CNAGE, CRTTL, AVCRWD, and AVPHLM showed no significant differences between attack classes (Table 2). Although resin streaming occurred on the majority of attacked trees, unsuccessful attacks generally produced more resin than successful attacks (Table 3). The majority of unsuccessful attacks produced moderate quantities of resin, while the majority of successful attacks trees produced only light quantities of resin (Table 3). The number of trees with rot present on disks appeared to vary with attack class at the Cherry Ridge and Lumby sites (Table 4). One out of 8, and 1 out of 12 control trees 21 had rot present at Cherry Ridge and Lumby respectively, while approximately half of the trees in the successful and unsuccessful attack classes had rot present (Table 4). At Milk River, there did not appear to be a difference in the j&equency of rot between attack classes. Based on field observations of the incidence of fruiting bodies and the type of rot on the disks, the m^ority o f rot was probably brown stringy trunk rot, EcAiwxfbntfWM tinctorium (Ell. and Ev.) Ell. and Ev. Red ring rot, Phellinus pini (Thore: Fr.) A. Ames, and red heart rot, Stereum sanguinolentum (Albertini and Schein: Fr.) Fr., were likely present to a lesser degree. No pattern was clearly discernable between crown foliage rating and attack class (Table 5). At Cherry Ridge successful attacks appeared to have sparser crowns than unsuccessful attacks and controls. Milk River showed a similar pattern, although not as distinct, and at Lumby there was little variation in the crown foliage ratings (Table 5). Table 2: A comparison of tree characteristics between control trees, unsuccessful attacks, and successful attacks. Controls Trees Unsuccessful Attacks Successful Attacks n Mean (SE) n Mean (SE) n Mean (S E ) Diameter*^ (cm) 28 33.7(1.9)0 26 34.7(1.7)0 22 32.7(1.8)0 Height (m) 28 24.4 (0.8)o 26 24.4 (0.6)o 22 24.2 (0.9)0 25 i 19 196.0 ; 12.3 IX 2(3.4 ('2.1)6 26 117.9 (3.7)0 23 116.7 (3.3)0 20 121.5 (2.4)0 5 year growth (mm) 27 ! 4.5(0.;,;./ 26 3.7 (0.3)06 22 2.8 (0.4)6 % bole with live crown 28 69.0 (2.6)0 26 68.4 (2.4)0 22 62.4 (2.4)0 'ill hole \'i!h const: ni crov n 2k 50.0 (2.0k/ 26 54.6 ( ! .Sloh Mean crown width (m) 28 3.0 (0.1)0 26 3.1 (0.2)o 22 2.8 (0.2)o Crown volume (m ’) 28 36.3 (4.1)0 26 36.6 (1.5)0 22 27.0 (3.7)6 Mean phloem (mm) 27 5.4 (0.3)o 26 5.4 (0.3)o 22 5.7 (0.3)o Canony Age 48.7(1.9)A Means within each variable followed by the same letter are not significantly different ANOVA, P > 0.05. Significant ANOVAs were followed by Bonferroni MCP, significant i f f < 0.05. Crown volume was transformed to a logarithm to correct for non-normality. ^ Diameter o f control trees was not expected to be significantly different from unsuccessful or successful attacks because the former were selected to be o f similar diameter to attacked trees. 22 Table 3: Number of trees in each resin category by attack class at each site. Lumby Cherry Ridge Milk River No' LR MR HR No LR MR HR No LR MR HR Unsucc 0 2 4 1 0 2 7 2 0 4 4 0 Succ. 0 6 0 0 2 5 3 0 0 5 1 0 * No, no resin beads or streaming; LR, light resin; MR, moderate resin; HR, heavy resin. Table 4: Number of trees with rot present/absent by attack class at each site. Cherry Ridge' Milk River^ Lumby* Rot No Rot Rot No Rot Rot No Rot Controls 1 7 1 11 4 4 Unsucc. 4 3 6 5 4 4 Succ. 3 3 3 7 2 4 disks cut at 0.5 m ' disks cut at 1.3 m Table 5: Number of trees in each crown foliage category by attack class at each site. Cherry Ridge Lumby Milk River Sprs' Mdm Thck Sprs Mdm Thck Sprs Mdm Thck Controls 1 7 0 0 9 3 0 7 1 Unsucc. 1 4 2 1 7 3 2 5 1 Succ. 3 3 0 1 8 1 3 2 1 Sprs, sparse; Mdm, medium; and Thck, thick. Twenty Metre Radius Plots Figure 4 shows the mean (±SE) number of 1998 successful attacks, 1998 unsuccessful attacks, faders and red attacks, subalpine fir over 12.5 cm dbh, and other species over 12.5 cm dbh within 20 m of controls, unsuccessful attacks, and successful attacks. No clear pattern emerged between attack classes, except there were three times more successful attacks within a 20 m radius of successful attacks compared to unsuccessful attacks or control trees at Lumby (Figure 4). Density of stems over 12.5 cm dbh was higher at Lumby than Cherry Ridge or Milk River (Figure 4). 23 .......... (B) 1998 Unsuccessful Attacks (A) 1998 Successful Attacks i Z z S' e § QO i T i f 4 1 T f Control Unsucc. Succ. Control (C) Faders and Red Attacks Unsucc. (D) Subalpine Fir Over 12.5 cm dbh ! T :z % 12 Control Unsucc. Succ. Succ. 1 ; I 1 ; Control Unsucc. Succ. # ■ A Cherry Ridge Lumby Milk River (E) Other Species Over 12.5 cm dbh Control Unsucc. Succ. Figure 4: Mean (SE) number of (A) 1998 successful attacks, (B) 1998 unsuccessful attacks, (C) faders and red attacks, (D) subalpine fir over 12.5 cm dbh, and (E) other species over 12.5 cm dbh, occurring within a 20 m radius of control trees, unsuccessful attacks, and successful attacks at 3 sites. 24 Discussion Susceptibility of subalpine fir to attack by western balsam bark beetle was associated with tree diameter, age, recent radial growth, induced resinosis, the proportion of the bole with constant crown, and crown volume. Large diameter is a common characteristic of trees susceptible to bark beetle attack. The preference o f western balsam bark beetle for larger hosts observed in this study was consistent with research on Douglas-fir, spruce, and mountain pine beetles (Cole and Amman 1969; Baker and Kemperman 1974; Fumiss et al. 1981; Shore et al. 1999). Because diameter was well correlated with phloem thickness (Appendix 2; Amman 1972; Cole and Cahill 1976), larger diameter trees may provide a more suitable habitat and food source for attacking adults and developing offspring. Past studies on tree-killing bark beetles have identified phloem thickness as a major factor in determining attack, gallery, egg, and brood adult densities (Amman 1972; Cole 1973; Amman and Pace 1976; Cole and Cahill 1976; Haack et al. 1987b). Failure to find differences in phloem thickness between attacked and unattacked (control) trees of similar diameter suggests that attacking beetles do not differentiate between trees of the same diameter based on phloem thickness. Phloem thickness was more likely related to host suitability, not susceptibility, although it may partially explain the preference of beetles for larger diameter trees. Western balsam bark beetle attacked trees from a wide range of diameters (9-55 cm dbh), however, small-diameter attacked trees were usually located next to larger, massattacked trees (personal observations). Western balsam bark beetles attracted to the larger diameter, mass-attacks may have landed on the smaller, adjacent trees as a result of increasing levels of anti-aggregation pheromones produced by mated females (Stock 1991). 25 Generally, the size o f tree attacked depended on the diameters available in the stand. Western balsam bark beetle consistently attacked trees from the 3 to 4 largest diameter classes at each site. Thus, a large diameter, highly susceptible tree at one site, may be only a medium sized, less susceptible tree at another site. This indicates that factors other than diameter also contribute to the susceptibility of subalpine fir to western balsam bark beetle. Tree diameter and age are usually correlated, with the oldest trees in a stand being the largest. Although Abies may subsist in the understorey for an extended period of time, there was a moderate correlation between diameter and age at two of the three sites (Appendix 2). Therefore, the increase in the susceptibility of subalpine fir to attack by western balsam bark beetle with tree age may be due in part to the effects of diameter. However, given that there was no significant difference in the mean diameters of unsuccessful and successful attacks (control trees were selected to be of similar diameter to attacked trees), the increase in susceptibility with age may be related to the effects of senescence. Senescence has been associated with a decline in host vigour, although the nature of the relationship has not been resolved (Coulson 1979; Kaufman and Ryan 1986; Yoder et al. 1994). Studies by Fumiss et al. (1981) and Shore et al. (1999) have shown that Douglas-fir beetle, which prefers downed or weakened hosts, preferentially attacks older trees. The decline in host vigour, as indicated by recent radial growth, from control trees to unsuccessful attacks to successful attacks, suggests that western balsam bark beetle may not be able to overcome the defenses of more vigourous hosts. In addition, beetles may be able to recognize low vigour hosts and preferentially select them for attack. Other studies have also associated susceptibility to bark beetle attack with host vigour. Hard (1985) found that the mean annual increment for the last 5 years was lower in both unsuccessfully and 26 successfully attacked spruce than unattacked spruce. Susceptibility o f ponderosa pine, Pinus poWerofa Laws., and lodgepole pine to attack by mountain pine beetle increased with reductions in stem growth efficiency (i.e. amount of stem wood produced per unit leaf area), although beetles were less selective in areas immediately surrounding killed trees (Larsson et al. 1983; Waring and Pitman 1985). Reduced radial growth may increase the susceptibility of Douglas-fir to attack by Douglas-fir beetle, although study results have been contradictory. Lessard and Schmid (1990) found a higher 5-year periodic growth rate in susceptible trees, but the stands were recovering from defoliation stress. Shore et al. (1999) found infested Douglas-fir trees had lower growth over the last 10 years than uninfested trees (not statistically significant), but a study by Fumiss et al. (1981) found the opposite. The differences in results may be due to combining unsuccessful and successful attacks into one category, the long time frame over which radial growth was measured, differences in site or stand parameters, or variations in bark beetle attack behaviour. Faster-growing trees were less susceptible to attack by western balsam bark beetle than slower-growing trees, which may be due to the higher quantities of induced resinosis that were observed on fast-growing trees. Induced resinosis is responsible for repelling attacks by adult beetles, inhibiting the establishment of blue-stain fungi associated with bark beetles, deterring oviposition, and increasing brood mortality (Christiansen 1985; Berryman and Ashraf 1970; Berryman 1969; Reid et a l 1967). Qualitative aspects of the induced response, e.g., relative amounts of toxic compounds or viscosity, may also play an important role in defending the host against attacking organisms (Nagy et al. 2000; Rohde et a l 1996; Lewinsohn et ai. 1993; Rafïa et oZ. 1985; Rafla and Berryman 1983b; Bordasch and Berryman 1977; Wong and Berryman 1977). However, this was not examined in the current 27 study. It is generally accepted that host vigour affects the induced response o f a host (Berryman 1982a), although the nature o f the relationship, and o f the induced response itself^ is not fully understood. Although subalpine fir lacks extensive primary resin canals, it has cortical resin blisters (Barman 1936). The role of resin blisters in host defense may be limited because attacking beetles generally enter the tree in bark cracks or under flakes, thereby avoiding the blisters (personal observations). The fir engraver, Scolytus ventralis LeConte, was rarely observed puncturing resin blisters on grand fir, Abies grandis (Douglas) Lindley (Berryman 1969) or white fir, A. concolor (Gord. and Glend.) Lindl.ex Hildebr. (Ferrell 1983). Furthermore, resin from cortical blisters on grand fir contained only trace amounts of mildly repellent compounds to the fir engraver and its associated fungus, Trichosporium symbioticum Wright, whereas wound resin contained higher quantities of higher-repellency compounds (Russell and Berryman 1976; Bordasch and Berryman 1977). The location of resin blisters in the outer cortex may also limit their defensive role in larger diameter trees that have formed secondary bark (Ferrell 1983). Pathogens, such as root diseases and stem rusts, may affect tree vigour (Oren et al. 1985; Lewis 1997; Mallett and Volney 1999). Past studies have demonstrated associations between trees weakened by root diseases or stem rusts, and susceptibility to bark beetle attack (Kulhavy et al. 1980; James and Goheen 1981; Livingston et al. 1983). Although roots were not excavated during this study, root diseases were likely not frequent on subalpine fir at these sites (Smith and Craig 1970). However, decay caused by heart rot fungi was common. The natural role of heart rot fungi is to eontribute to the mortality of old, suppressed, or unhealthy trees in the forest (Manion 1991). Vigourous trees may be infected. 28 but are often able to suppress or compartmentalize decay and maintain their structural integrity. Although extensive decay in the heartwood could reduce the structural integrity of the host, the sapwood may be unaffected and remain fully functional; the effect of heart rot on tree vigour is unknown. Susceptibility of subalpine fir to heart rot increases with diameter, age, and suppressed growth (Maloy and Robinson 1968; Smith and Craig 1970; Etheridge and Craig 1976; Hunt and Etheridge 1995). Therefore trees most likely to have heart rot were also highly susceptible to western balsam bark beetle due to their physiological characteristics. Reduced photosynthetic ability may affect the induced response, which depends on the efficient translocation of current photosynthate to the invasion site (Christiansen and Ericsson 1986; Miller and Berryman 1986). Defoliation and pruning may increase susceptibility to bark beetles or their symbiotes (Wright et al. 1979; Miller and Berryman 1986; Christiansen and Fjone 1993). Failure to find a relationship between crown rating and resistance of subalpine fir may have been due to the small sample size, the gross classification scheme used to rate foliage, and the lack of variation in crown ratings between trees (most were moderate). However, the tendency of successfully attacked trees to have a lower proportion of the bole covered with constant crown and smaller crown volume suggests that crown size may be related to resistance. Increased susceptibility of trees with a lower proportion of the bole with constant crown could also be due to higher rates of successful landings on such trees. Branches may inhibit incoming beetles from locating and landing on the bole. Although beetles may land lower on the bole and walk up the tree before attacking, the relatively high flight path of western balsam bark beetle as shown by Stock (1991), suggests that this beetle may also land on the upper bole. 29 High stand density, which may lead to reduced radial growth from competition, may be an ideal condition for the occurrence of a number of tree-killing bark beetles (Coulson et al. 1974; Schenk et al. 1977; Fumiss et al. 1981; Shore and Safranyik 1992). However, in this study no diflerence was detected in the density o f stems within 20 m o f the sampled trees. This discrepancy could be due to differences in spatial scales, as density is usually calculated at the stand level and considered to increase susceptibility at that level. The higher number of mass attacked trees around successful attacks at Lumby may be due to secondary attraction. Dispersing beetles drawn to the mass attack by aggregation pheromones produced by male beetles may land on adjacent trees, repelled by the anti-aggregation pheromones produced by mated females (Stock and Borden 1983). Another explanation may be that susceptible hosts may occur in small groups (Stock 1991). However, this relationship was not apparent at the other two sites, or for faders and red attacks. Because of the patchy distribution of western balsam bark beetle (Stock 1991), and beetle mortality associated with flight length, a greater number of red attacks were expected in the vicinity of successful attacks. However, the beetle’s preference for temporary or scattered habitats, e.g., downed or stressed hosts, may increase its inclination to disperse greater distances (Atkins 1966). The large plot radius or the rare occurrence of 1998 successful attacks, 1998 unsuccessful attacks, and faders and red attacks makes relationships hard to identify. Conclusions The preference of western balsam bark beetle for low vigour hosts is shared by most other species of hark beetles (Rudinsky 1962). Stressed and downed trees may emit volatiles that enable insects to locate weakened hosts. Furthermore, weakened hosts may have lowered defense systems (White 1969), which would increase the likelihood of successful 30 brood production. Epidemics of tree-killing bark beetles are often linked to events causing a temporary increase in susceptible host material. Assuming that events causing an increase in susceptible subalpine fir have occurred in the past, there have been few reports of western balsam bark beetle mortality levels approaching those caused by other tree-killing bark beetles during epidemics. Future research on the nature of the defense response of low and high vigour hosts may help explain western balsam bark beetle’s apparent limitation to stressed or downed hosts. Stand management practices that increase host vigour could be used to reduce mortality caused by western balsam bark beetle. The factors that affect susceptibility identified in this study should also be considered in the development of a hazard rating model for western balsam bark beetle. This study did not exhaust the list of factors that may affect host susceptibility; other important factors may include macro- and microsite characteristics, genetic variation, and the chemical composition of the induced response. 31 Chapter HI. Success of Pheromone-Imduced Attack by Western Balsam Bark Beetle and the Defense Response of Fast- and Slow-Growing Subalpine Fir Introduction Tree-killing bark beetles must overcome the defenses of live hosts in order to successfully colonize trees and produce brood. Plants possess two general types of defense systems to combat attacking organisms: a constitutive defense system, also known as passive, primary or preformed resistance; and an induced defense system, also known as active, secondary or induced resistance, or dynamic wound response (Klement and Goodman 1967; Berryman 1972; Merrill 1992; Paine et al. 1997). The constitutive defense system of a conifer consists of an extensive system of resin-filled canals that exist in the tree prior to attack. Attacking beetles may sever the preformed canals and be deterred, drowned, or physically expelled by a short-lived flow of resin. Extensive resin canals are present in many of the Pinaceae, including trees in the genera Larix, Pseudotsuga, Picea, and Pinus (Barman 1936). However, trees in the genera Abiej', Tsuga, and Cedrus lack extensive preformed resin canals, and rely mainly on the induced response for defense (Barman 1936; Berryman 1972). The induced defense response is a rapid reaction by a plant to mechanical wounding or invasion by organisms, e.g., insects, pathogens, viruses, or bacteria (Klement and Goodman 1967; Cates and Alexander 1982; Merrill 1992). The induced reaction involves immediate changes in the function, division, and differentiation of cells adjacent to the attack site (Rohde et al. 1996; Franceschi et al. 2000). These changes lead to cell necrosis (hypersensitivity), decreased sugar concentrations, increased monoterpene production, formation o f traumatic resin ducts, induced resinosis, and wound periderm formation (Reid et 32 al. 1967; Berryman 1969, 1972; Shrimpton 1973; Wong and Berryman 1977; Raffa and Benyman 1983b, 1985; Lewisohn gf aZ. 1993; Rohde gf aZ. 1996; Nagy gZaZ. 2000). The invading organism(s) may or may not be able to survive and overcome these obstacles. The induced defense response may be the most important faetor in determining host resistance, even in trees that have a preformed defense system (Reid et al. 1967; Berryman 1972; Raffa and Berryman 1982a; Langstrdm et al. 1992). The ability of a tree to defend itself is often associated with host vigour, as well as the density and vimlence of attacks (Berryman 1972, 1982). In Chapter II susceptibility of subalpine fir to western balsam bark beetle was related to cumulative radial growth in the last 5 years. This study further examines that relationship by using pheromone baits to induce attack on fast- and slow-growing trees. The specific objectives of this study were: 1) to compare the relative attack success of western balsam bark beetle in fast- and slowgrowing trees; and 2) to compare the rate and extent of the induced defense response of fast- and slowgrowing trees to attack by western balsam bark beetle. Methods S'ZZg D g g ^ c r ^ Z Z o n a w Z T k g g % Z g c f t o n Fifty-two trees were selected for pheromone baiting at two sites in the Engelmarm spruce-subalpine fir (ESSF) biogeoclimatic zone in the interior of British Columbia. Ten fast- and 11 slow-growing trees were selected at a site (UTM 5911200 654100) in the Milk River drainage in the Robson Valley Forest District. The Milk River site was located at 1 350 m elevation on a south-facing 35% slope in the moist mild (mm) subzone of the ESSF. Sixteen fast- and 15 slow-growing trees were selected at a site (UTM 362300 5551500) south 33 of Lumby, in the Vernon Forest District. The Lumby site was located at 1 700 m elevation on a slope of less than 10% in the very dry cold (xc) subzone of the ESSF. Transects 20 m wide, spaced at intervals of 50 m or 100 m at Lumby and Milk River, respectively, were established. Two cores were taken from mid to large diameter stems along the transects and used to calculate mean cumulative growth for the last 5 years (CUM5). Cores were taken perpendicular to the slope at the Milk River site, and from east and west aspects at the Lumby site, where hill slope was negligible. The mean CUM5 of control trees, unsuccessful attacks, and successful attacks identified in Chapter II (Table 2) were used to classify trees into fast- and slow-growing categories. Trees were classified as slow-growing if the mean CUM5 was 2.8 mm or less and fast-growing if the mean CUM5 was equal to or greater than 4.0 mm. Trees with a mean CUM5 between 2.8 and 4.0 mm were excluded from the study. A minimum distance of 15 m was left between fast- and slow-growing trees selected for baiting to avoid saturating the stand with pheromones. At Milk River, selected trees ranged in age from 85 to 189 years at breast height and in diameter from 26.5 to 56.9 cm at breast height. At Lumby, selected trees ranged in age from 93 to 266 years at breast height, and in diameter from 20.8 to 37.2 cm at breast height. Tree Baiting and Data Collection Selected trees were baited in late May and mid June of 1999 at Milk River and Lumby respectively, prior to the main beetle flight. The standard (±)-cxo-brevicomin bait for western balsam bark beetle (Phero Tech Inc., Delta, BC, Canada) was stapled to the north side of the bole of each tree at 2.0 m height. Trees at the Milk River site were assessed for resin production approximately 3, 5, and 6 weeks after the main flight. At each observation time resin production was rated as 34 either none, light, moderate or heavy (Table 1). Trees at the Lumby site were assessed for resin production at approximately 7 weeks after the main flight. Trees were felled approximately 6 or 7 weeks after the main flight at each site, and the following information was recorded for each tree: dbh, tree height, height of lowest and highest unsuccessful and/or successful attack if present, and attack class (Table 6). Five 400 cm^ (20 X 20 cm) bark samples were taken at random heights and aspects (excluding the aspect that the tree lay on) in the attack zone (between the lowest and highest attack regardless of success). For each bark sample the following information was recorded if present; number of unsuccessful attacks, length and width of each lesion associated with an unsuccessful attack, number of successful attacks (gallery systems), number of female galleries in each gallery system, and the mean female gallery length per gallery system. Data from the bark samples were collated and the following variables summarizing attack were constructed for each tree; • • • • • • • • density of unsuccessful attacks in the attack zone; density of successful attacks in the attack zone; density of total attacks (unsuccessful and successful) in the attack zone; proportion o f tree’s total height successfully attacked; mean length of lesion; mean width of lesion; mean number of female galleries per gallery system; and mean female gallery length per gallery system. Table 6; Description of unsuccessful, light, and successful attack classes. Attack Status Description of Attack Unsuccessful No live adults or larvae present in phloem. Lesions from unsuccessful attacks present Light Live adults and larvae present. Phloem discolored around galleries, but lines of periderm present, and live phloem between attacked areas. Some lesions from unsuccessful attacks may also be present. Successful Live adults and larvae present. Phloem highly discolored throughout attack zone. 35 ANOVA procedures were used to test for differences in basic tree characteristics between sites and tree growth categories. The proportions of fast- and slow-growing trees that produced resin were compared using Fisher’s exact test for small sample sizes. The frequency of occurrences in tree growth categories, resin ratings, and attack classes were not statistically compared due to very small sample sizes. Variables summarizing attack were analyzed for relationships with site and tree growth using ANOVA. Many of the attack variables were not normally distributed and their distributions varied between cells. Various logarithmic, square root, and arcsine transformations failed to normalize the data because of the variety of distributions found across cells. Although a nonparametric two-factor analysis of variance exists (Zar 1999), the option of a second grouping variable was not available in a number of common statistical packages. Therefore, data from the two sites were pooled and a nonparametric one-factor ANOVA (Mann-Whitney test) was conducted for each attack variable using tree growth as the grouping variable. The tests were repeated on the attack variables using site as the grouping variable and pooling the data from fast- and slowgrowing trees at each site. Results from the nonparametric tests were compared to results from parametric two-factor ANOVAs conducted on each attack variable using site and tree growth as factors. Statistical significance did not differ between the two methods and the Pvalues produced by the nonparametric tests were usually within 0.002 of the P-values produced by the corresponding parametric tests. Because the results of the two methods were similar, the two-factor design of the experiment, and the robustness of parametric ANOVA to violations of normality (Tabachnick and Fidell 1996), the results presented for all statistical analysis are for parametric tests performed on non-transformed data. 36 The level o f significance was set at 0.05 for all statistieal tests. Data were analyzed using SYSTAT 9.0 (SPSS Inc., Chicago, IL, USA). Results Mean (±SE) dbh of baited trees was significantly greater at the Milk River site than at the Lumby site (38.5 (±1.6) cm and 26.0 (±0.8) cm, respectively) (F i/g= 55.181, P < 0.001). Mean dbh of fast- and slow-growing trees did not differ significantly within sites (Fi _48 = 0.308, P = 0.582). Height of baited trees did not differ significantly between fast- and slowgrowing trees (Fi,4 g = 0.340, P = 0.563) or between sites (Fi ,4 8 ~ 1.295, P = 0.261). Beetle pressure, a subjective measure based on the relative abundance of red-attacked trees observed at each site, was described as high at the Lumby site and moderate at the Milk River site. Tfb&t De/èn&g orwf Tree GrowtA Beetles attacked all 52 baited trees. Six to 7 weeks after the main beetle flight, 29 of the 52 baited trees had produced resin. At the Lumby site, seven weeks after initial attack the proportions o f fast- versus slow-growing trees that had produced resin were not significantly different (Fisher’s exact test, P = 0.479) (Table 7). At Milk River, a significantly greater proportion of fast-growing trees had produced resin by seven weeks after initial attack than slow-growing trees (Fisher’s exact test, T*= 0.008) (Table 7). Table 7: Number (percent) o f fast- and slow-growing trees with resin present/absent for Lumby and Milk River research sites, 6 to 7 weeks after initial attack. Lumby Milk River Fast Slow Fast Slow Resin present 10 (62.5) 7 (47) 9 (90) 3 (27) Resin absent 6 (37.5) 8 (53) 1 0.05. Significant ANOVAs were followed by Bonferroni MCP, significant if P < 0.05. Crown volume was transformed to a logarithm to correct for non-normality. 89