J.ENTOMOL.SOC.BRJT.COLUMBIA 100,DECEMBER2003 27 Impact ofthe western balsam bark beetle, Dryocoetes confusus Swaine (Coleoptera: Scolytidae), at the Sicamous Creek research site, and the potential for semiochemical based management in alternative silviculture systems LORRAINE MACLAUCHLAN^ E. B.C. MINISTRY OFFORESTS, SOUTHERN INTERIORREGION, 515COLUMBIA STREET, KAMLOOPS,B.C. V2C 2T7 LEROY HARDER 23392-16™AVENUE, LANGLEY, B.C. V2Z 1K7 JOHN BORDEN H. PHEROTECH INC., 7572 PROGRESSWAY,R.R.#5,DELTA,B.C. V4G 1E9 JULIE E. BROOKS FORESTHEALTH MANAGEMENT, BOX 19, GRANTHAMSLANDING, B.C. VON 1X0 ABSTRACT Two pre-harvest baiting regimes were tested for their effect on Dryocoetes confusus in select stands at the Sicamous Creek Silviculture Systems Project. Single tree and two- tree bait treatments, in addition to a control area, were established in a grid format throughout the research area. There were significantly more new D. confusus attacks in the baited areas than in the control area. Eightypercent ofmass attacks occurred within 9 m ofsingle tree bait centres, while 75% ofmass attacks occurred withm 10 m oftwo- tree bait centres. Baiting appears to concentrate attacks into a discrete area and therefore could be used in single tree selection or patch cut systems (cuts generally less than 5 ha in size), two of the silviculture systems applied at the Sicamous Creek research area. Of 136 dead subalpine fir trees felled and examined, 105 (77%) showed clear evidence of D. confusus attack, making it the major cause of sub-alpine fir mortality at the Sicamous Creek research site. Naturally attacked trees had more advanced brood development and beetles utilized a greater percent ofthe total tree bole but had lower attack density (number ofD. confusus galleries per unit area) than was observed on baited trees. In baited trees, the higher attack densityresulted in indistinct gallery systems due to space competition ofthe brood. This suggested that there was a limited acceptable area for attack in these trees, which would not normally be A susceptible. This study concludes it is possible to reduce resident populations of confususbyvaryingthenumberandplacement ofbait trees as apre-harvesttreatment. Key words: Dryocoetes confusus, western balsam bark beetle, pheromone baiting, Abies lasiocarpa, subalpine fir ' Authortowhomall correspondence should be sent. 28 J.Entomol.soc.Brit.Columbia100,December2003 INTRODUCTION The westernbalsambarkbeetle, Dryocoetes confusus Swaine (Coleoptera: Scolytidae), is the most destructive insect pest of subalpine fir, Abies lasiocarpa (Hook.) Nutt., in BritishColumbia (Garbutt 1992; McMillin etal. 2001). Subalpine fir is also susceptible to a variety ofother disturbance agents, including other insects, root and butt rots, stem rots and windthrow (Kneeshaw and Burton 1997). Cumulative mortality due to D. confusus may reach significant levels in chronically infested stands (Garbutt and Stewart 1991), howeverD. confusus outbreak dynamics appear to be very different fromothertree-killing bark beetles. Over time, aerial overview surveys have established an average annual loss of4.2 m^ per hectare in older affected stands. D. confusus can kill many trees in a single year but usually less than 5% ofany given stand is attacked in one year (Garbutt 1992). Beetle populations can persist for many years in a stand slowly killing the entire mature and semi-mature componentofsub-alpine fir(Garbutt 1992). Subalpine fir comprises 12% oftotal timber volume (trees cut) in B.C. (B.C. Ministry of Forests 1993) and has typically been harvested in conjunction with higher valued spruce. As low elevation stands consisting ofother tree species are depleted, the number of subalpine fir sites harvested has increased. In 1990, (B.C. Ministry of Forests 1992) subalpine fir comprised 8% oftotal volume harvested in the interior ofB.C. compared to 10.9% volume in 2000-01 (B.C. Ministry of Forests 2001). As harvesting increases in subalpine fir sites, additional research is needed to develop more effective and ecologically sensitive management strategies. In 1990, the B.C. Ministry of Forests established a silviculture systems project at Sicamous Creek near Salmon Arm, B.C., to address ecosystem responses to a wide range ofdisturbance levels created by harvesting. The Sicamous Creek site is located within the Engelmann Spruce- Subalpine Fir wet, cold subzone (ESSFwc2) (Lloyd et al, 1990), which is the largest of the seven ESSF subzones in the Kamloops Forest Region. This study was established at the Sicamous Creek research site to test how two baiting techniques couldbe used to manageD. confusus under differentharvesting regimes. Baiting trees with semiochemicals as a pre-harvest containment and concentration tactic is a well established pest management methodology for other bark beetles such as the mountain pine beetle, Dendroctonusponderosae Hopkins, (Borden 1990; Maclauchlan and Brooks 1999) but has not been developed for D. confusus. Therefore, a trial was developed to test baiting systems that could be used in a single tree selection and patch cutting harvest scenarios. Our objectives were to: assess past infestations ofD. confusus at the Sicamous Creek Silviculture Systems Project; test the efficacy of pre-harvest baiting systems for D. confusus in different silviculture systems; and determine if pre-harvest baiting could concentrate more beetles forremoval atharvestthanno baiting. METHODS Pre- andpost-harvestlevels ofD. confusus The Sicamous Creek research site is dominatedby subalpine fir and Engelmann spruce {Picea engelmanni Parry ex Engelm.). The harvest regimes, each on 30 ha, were: 1) control/no removal; 2) single tree selection in which 33% ofthe volume was removed over a 30 ha area by cutting every fifth tree using faller's choice; 3) 0.1 ha patch cuts; 4) 1 ha patch cuts; and, 5) 10 ha clearcut. Each of the harvest regimes removed 33% of the volume. The area was harvested without taking into account the presence or impact ofD. confusus, even thoughthere was significant mortalitythroughoutthe area. Aerial photographs (1:5,000) ofthe Sicamous Creek silviculture systems project forthe years 1993-1995 were used to map the location of red trees over a 200 ha area. Using J.ENTOMOL.SOC.BRJT.COLUMBIA100,DECEMBER2003 29 1993 photographs, groups ofred trees were identified and mapped on an acetate overlay. For 1994, groups ofred trees adjacent to the original clusters were identified and mapped. Each cluster was coded in relation to the eventual harvest regime conducted in the winter of1994-1995. The red trees mapped fromthe 1995 photographs were used to compare the relative efficacy ofthe five cutting regimes in removing D. confusus. The uncut strips of trees between the 0.1 ha and 1.0 ha patch cuts were located on the photographs and assessed forredtrees. Infestation characteristics Prior to adult emergence in the spring of 1995, 15 red, 31 grey and 90 older snags, were felled and examined for evidence of D. confusus activity. The following characteristics were measured or noted in the red and grey trees: diameter at breast height (d.b.h.) of the bole and distance from the ground for upper and lower limits of attack; resinosis typical ofD. confusus attack; exit holes; and the presence of live adults, pupae and larvae. Because snags were usually quite degraded, only the presence or absence ofD. confusus galleries, exit holes and associated resinosis were recorded and measured inthose trees. Pheromone baitingtrial In June 1995, a baiting trial was established at Sicamous Creek. Two treatments and a control area were laid out. The aggregationpheromone (+)-exo-brevicomin (released at 0.3 m mg/24 hrs release rate) was used (Phero Tech Inc.). Baits were stapled at 1.5 on the north side oflarge subalpine fir. In the single tree bait treatment, established in the single m m tree selection area, bait lines were 50 apart, with baited trees at 25 intervals in a grid pattern. In the two-tree bait treatment, established in the 0.1 ha patch cut area, bait lines m m were placed 33 apart, withbaits affixed every 66 along the bait line, on two adjacent, large subalpine firs at eachpoint. Baits on adjacent lines were offset by 33 m. In total, 86 single and 82 paired trees were baited. No baits were used in the control area. A chi square analysis was usedto compare the number ofgreen (live) trees to red (attacked) trees inthe three treatmentregimes. In September 1995, a 100% ground assessment of all subalpine fir in the three study areas was conducted. Using Stock's (1991) criteria for "attack classes" in Table 1 a stem map was produced ofthe baited, attacked, mass attacked, red and grey trees. The d.b.h. of these trees was measured, and the number of snags in each area was counted. In each m treatment, 10 randomly placed 15 radius circular plots were established, to discern infestation characteristics. In each circular plot the d.b.h., species and tree class were recorded foreach tree with minimum9 cmd.b.h. Chi square analysis was used to compare the d.b.h. frequency distribution ofred, greyand snags to unattacked trees. Comparison ofinsectdevelopmenton baitedandnaturallyattackedsubalpinefir In late August 1996, 10 baited mass attacked trees in the single tree bait treatment area and 13 new mass attacked trees outside the study area were felled. Beginning at the stump (cut end oftree), gallery systems were dissected in 10x30 cm bark sections every 1.5 m along the bole. For each sample, the number of gallery systems and the occurrence of associated species were recorded. Within each gallery system, the presence or absence of D. confusus life stages and resin was recorded, and the length of each egg gallery was measured. Female D. confusus constructs the egg gallery away from the nuptial chamber where she mates and deposits eggs along the sides ofthese galleries. The upperbole ofthe tree was examined for secondary scolytids and other associated insects. These scolytids are often referred to as secondarybark beetles as they do not typically kill trees but occupy trees infestedby other tree killing species ofthe Scolytidae. Height limits for conspicuous resin flow was also noted. Foliage colour change was rated using a six point rating system (Table 2). 1 30 J.ENTOMOL.SOC.BRIT.COLUMBIA100,DECEMBER2003 Table 1 Tree classifications assigned to subalpine firs attacked by D. confusus. These "Attack Classes" were developedby Stock(1991) andmodifiedbyL. Harder. A'H'ool/' 1*1000 i-^CodipilUll attacked Streams ofresinonbole (presumedunsuccessfullyattacked) mass attacked fi-ass and possibly resin on bole (presumed successful intense colonizationoftree) red red foliage present (represents old attack fi-om which new mature beetles emerge) grey needles mostly gone, but fine twigs present and bark generally intact (nobeetles remaining inbark) snag a long dead tree; minimum height 2 m and d.b.h. 12 cm, with bark loose orabsentand fine twigs gone Table 2 Foliage colour classes used to classify colour changes in subalpine fir trees one year after mass attackbyD. confusus. ColourClass Description 0 No colour change noticeable 1 Redneedles on some tree limbs, usually on lowerbole Foliage on less than Vz ofthe tree limbs starting to turnred, usuallyon 2 lowerbole 3 Halfthe foliage turnedred 4 Most ofthe foliage turnedred, some faded green left 5 Foliage completelyred Baited and naturally mass attacked trees were compared using a chi square test based on the number oftrees containing different life stages ofyoung brood. Samples without gallery systems, and gallery systems without brood, were classed as failed gallery systems and compared to other characteristics of attack by using a chi square test. Analysis of variance comparing naturally attacked trees and baited trees were done on gallery length, resin flow, number ofegg galleries, and the totalbrood gallery length. RESULTS AND DISCUSSION Pre- andpost-harvestlevels ofD. confusus The mapped number ofred trees decreased from 1993 to 1995 in both undisturbed and harvested areas (Table 3) indicating an overall decline in the Dryocoetes population. Because clearcut treatments (1 ha and 10 ha) remove all trees in an area, all D. confusus attacks were also removed in the harvested areas. Fewerred trees were observed in the 0. ha and single tree selection cut areas than inthe 1.0 ha patch cut area (Table 3). The buffer strips between the 1.0 ha patch cuts were undisturbed by harvesting therefore, there was a similar level of attack there as in the undisturbed control areas. In the other two treatments, the 0.1 ha patch cut plus buffer strip and the single tree selection, there was so little area between the cuts that more attacked trees were removed at the time ofharvest. Any dead trees within the narrow buffer strips, many of which were infested with Dryocoetes, were removed at the time ofharvest. Thus, despite the lack of a conscious effort to manage forDryocoetes, much ofthe residentbeetle population was removed from these areas when harvested. J.ENTOMOL.SOC.BRIT.COLUMBIA100,DECEMBER2003 31 Table3 Numbers ofred subalpine fir per hectare in undisturbed and treated areas before and after treatmentas seeninthree consecutive years ofaerialphotographs. No. redtrees perha Sample pre-treatment post-treatment Location ofredtrees Size(ha) 1993 1994' 1995 Undisturbedcontrol area 108 7.4 6.2 4.5 Within 10 ha clearcut 10 5.3 7.9 0 Within 1.0hapatch cuts 9 10.1 9.7 0 Inbuffer stripbetween 1.0 ha 30 8.6 7.8 4.5 patch cut In0.1 hapatch cutandbuffer 18 5.5 8.9 1.3 strip In single tree selection area 21 4.0 3.6 0.7 ^ Roadright-of-way cut throughresearch area in 1994. Table 4 Evidence of past attack by D. confusus in felled red, grey and snag subalpine fir. The characteristics assessed included D. confusus brood (eggs, larvae, pupae), adult beetles, galleries, exitholes made by emergingbeetles and resin flow on the bole ofthe tree caused by attacking beetles. The number of trees having all of the above-mentioned characteristics was also summarized. % subalpine fir with characteristic Characteristic assessed Red(n=15) Grey(n=31) Snag (n=90f D. confusus brood 27 3 0 D. confusus adults 13 3 0 Galleries 100 90 70 Exitholes 80 87 63 Resin flow 93 94 56 Characteristics combined 100 97 76 ' Fewer snags were assessed for exit holes (n=87) and resin flow (n=72) than for other characteristics due to deterioration and loss ofbark. Infestation characteristics All 15 felled red trees and 30 of 31 grey trees showed evidence ofpast attack by D. confusus (Table 4). Twelve of 15 red trees had exit holes, while only 4 of 15 hadjuvenile life stages, and two had adults. There was less evidence ofbeetle attack in snags due to deterioration and loss ofbark. This is strong evidence that most ofthe dead subalpine fir inthe studyarea hadbeenattackedbyD. confusus. Attacked subalpine fir can retain their red foliage for a number of years prior to shedding needles and being termed grey. The examination ofred and grey trees revealed that few D. confusus adults were still present (Table 4), suggesting that adult beetles leave red treesbefore treesbecome grey. Successful completion of development by D. confusus occurred primarily along the lowerportion ofthe bole. In general, exit holes occurred within the upper and lower limits of the gallery systems (Figures 1, 2). In 32 trees that had visible resmosis, resin flow usually overlapped the exit hole zone and in 28 trees, extended above the exit holes a few metres. There was less variation in the lower height limit for gallery systems and exit 32 J.ENTOMOL.SOC.BRIT.COLUMBIA100,DECEMBER2003 holes than in the upper height Hmit. Exit holes associated with otherbeetles in the Family Scolytidae occurred throughout the resin flow zone extending past it in both directions. Secondary scolytids found in the lower bole were identified as Pityokteines minutus, and those inthe upperbole asPityophthorussp. The narrow variance inheightofthe lowerlimit ofD. confusus galleries andexitholes, and the broader variance in diameter (Figure 2) suggest that height was of greater importance than diameter in limiting D. confusus occupation at the lower end ofthe bole. Poorgallery developmentbetween 1 and 2 mmaybe relatedto cooler nighttime summer temperatures close to the ground, typical in the ESSF (Famden 1994). Gallery systems close to the ground at or below the lower limit ofexit holes had short egg galleries. Bark in this area ofthe bole is often wet, encouraging the growth ofdecay fungi that overgrow D. confusus galleries. Incontrast, the upperlimit was characterizedbywide variation in Figure 1. Upper and lower height limits for the majority ofD. confusus galleries, exit holes, and resin flow observed on felled red and grey subalpine fir. J.ENTOMOL.SOC.BRIT.COLUMBIA 100,DECEMBER2003 33 Galleries Lowerlimit llL 1.1 Galleries Upperlimit III .ililill.ii Exitholes Lowerlimit Exit holes Upperlimit .11 ilill.l Resin Flow Lowerlimit Resin Flow Upperlimit il 1 . , llllllli 11, loioinioiominio , , , c\iTtcbcboc\iTtcbocio CO CO CNJ CNJ -- •<- Diameter(cm) Height(m) Figure 2. Frequencydistributions forupper and lower limits forD. confusus galleries, exit holes andresin flow, based onbole diameterandheight. height, indicating a weak influence. The upper limit for resin flow, however, was tightly clustered around 17.5 cm diameter (Figure 2), indicating a possible influence related to diameter that limits D. confusus attack. The 17.5 cm upper limit diameter peak for resin flow was greater than the average 10 cm upper limit for attack on trees that were 34 J.ENTOMOL.SOC.BRIT.COLUMBIA 100,DECEMBER2003 previously assessed by Stock (1991). This difference may indicate a different diameter limitpreferencebyD. confusus onstanding trees versus downedtrees. A large percentage of subalpine fir basal area consisted of dead trees, totaling 31%, 28% and 25% for the single tree bait treatment, two-tree bait treatment and the control area, respectively, at the time of baiting. These proportions are similar to those documented by Unger and Stewart (1993) and Stock (1991). Parrish (1997) determined that losses take place over a long period in subalpine fir stands, with some existing snags having been dead for over 45 years. This slower, but continuous tree mortality affects stand structure very differently than the devastation caused by mountain pine beetle to maturepine stands. D. confusus is likely the major mortality causing agent for standing dead trees at the Sicamous Creek research area. The d.b.h. distribution of red and grey trees contained more trees in the larger d.b.h. range (>20 cm d.b.h.), ranging in size between 19 and 49 cm, while the d.b.h. distribution of snags contained on average smaller trees, between 9 and 39 cm, similar to unattacked trees. Direct observation also confirmed D. confusus activity in all felled red trees, 97% ofthe grey trees and 76% ofsnags examined (Table 4). The greater percentage ofsmall diameter trees among snags (Figure 3), suggests that the smaller trees were killed by Armillaria ostoyae, a common root disease of conifers. Merler (1997) found that A. ostoyae killed mostly subdominant balsam and spruce at this site. Pheromone baitingtrial Baiting trial assessment. The ratio ofmass attacked to red attacked trees in the single tree and two-tree bait treatments was similar, and was significantly higherthan inthe Unattacked RedN=7^3 P<0.0001 GreyN=55 f^O.0001 90 190 290 390 490 to to to to + 189 289 389 489 dbh(cm) Figure 3. Frequency distribution by diameter class of unattacked, red, grey and snag subalpine fir. The DBH distribution ofred, grey and snag trees were significantly different fromnon-attacked trees (chi square analysis). J.ENTOMOL.SOC.BRIT.COLUMBIA100,DECEMBER2003 35 control area (Table 5), despite differences in the numbers ofred trees per hectare among the three areas in the pheromone baiting experiment. Trees in the two-tree bait treatment area were more frequently mass attacked than those in the single tree baiting area (Table 5). Table 5 Comparison ofnumbers and ratios ofgrey, red and newly mass attacked subalpine fir in baitedand control areas. Numberofaffected subalpine fir/ha Mass Mass attacked: Red: Treatment area attacked Red Grey Red' Grey' Control 4.4 18.5 15.5 0.24a 1.19a Single treebaited 4.9 8.7 23.6 0.56b 0.37b Two-treebaited 14.2 27.3 19.6 0.52b 1.40c ' Proportions followedbythe same letter are not significantly different, x^, P<0.001 Single-baited trees were consistently mass attacked. Spillover attack (attack on trees m directly adjacent to a baited tree, resulting from the bait treatment) was highest out to 1 away from the bait centres (Figure 4), and decreased at greater distances. Eighty percent m ofmass attacks occurred within 9 ofsingle tree bait centres. m In the paired bait treatments, mass attacks peaked within 3 to 4 ofthe bait centre, apparently corresponding to half the distance between two bait trees (Figure 4). Cumulative mass attacks increased to 60% of total mass attacks within 5 m of the bait centre. At 10 m, 75% ofall cumulative attacks hadbeenrecorded. The aggregation pheromone (±)-exo-brevicomin used in the baiting trial successfully concentrated D. confusus attacks on and around bait centres (Figure 4). However, it was unclear whether the higher ratios in the baited vs. control areas were caused by retaining dispersing beetles within the baited areas, attracting beetles into those areas (Table 5) or spreading the same number of beetles among more trees. Baits used for spruce beetle seemtohave a limit of25 meters efficacy (Shore etal. 1990), and Gray and Borden (1989) found that the influence ofpheromone baiting for mountain pine beetle extended up to 75 m fromgrid-baited stands. Stock et al. (1994) observed consistently higher mass attacked m to red ratios within baited blocks than in 50 wide buffer strips surrounding the blocks. m This suggests at least a 50 range ofinfluence onD. confusus. Comparison ofinsectdevelopment in baitedandnaturallyattackedsubalpinefir. There was great variation in the utilization ofboth baited and naturally attacked trees by western balsam bark beetle, ranging from trees with few gallery systems and little brood to those having long galleries occupying a large proportion of the bole with advanced brood development. Naturally mass attacked trees had more advanced brood development (Table 6) and a greaterpercentage ofthebole was occupied (Harder 1998). The average meters ofegg galleries and average density ofegg galleries were less in the naturally mass attacked trees than in the mass attacked, baited trees (Figure 5). Mean egg gallery length however was not different between natural and baited trees (Fig. 5) indicating more egg galleries constructed in baited trees. Some trees with low attack densityhadhighnumbers ofD. confusus parentadults pergallery system (up to 10). These gallery systems were stained black, evidence of the fungus Ophiostoma dryocoetidis. Westernbalsambarkbeetle is closely associated with this pathogenic fungus (Garbutt 1992; Bleiker et al. 2003). Initial beetle attacks may be pitched out and O. dryocoetidis introduced, which in turn facilitates successful subsequent attack by the 36 J.ENTOMOL.SOC.BRIT.COLUMBIA100,DECEMBER2003 1 treebait centres 2 treebaitcentres 80 80 Attack - 60 + 60 40 40 - 20 20 0 123456789 0 l"l^F'F'l' 10 1 3 5 7 9 11 13 15 T3 O 80 80 Mass 60 60 - Attack 40 40 20 20 - 0 0 1 2 3 4 5 7 8 9 1 1 3 5 7 9 11 13 15 80 80 All Attack 60 - 60 (attackand 40 -- 40 -f mass attack 20 1 20 A combined) 0 123456789 0 13 10 5 7 11 13 15 Distance frombait centres (m) Figure 4. Distribution of attacked trees (% trees) at one bait and two bait tree centres. Solidbars indicate % attack at Imintervals. Clearbars indicate cumulative attack. beetle. Coalescing lesions caused by the fungus may also girdle and kill the trees without any furtherbeetle activity (Garbutt 1992). Egg galleries were so close together inbaited trees that the centre of the gallery system was completely excavated, and the nuptial chamber and brood galleries were no longer distinguishable. This high density ofgallery construction and subsequent brood activity make certain characteristics of the gallery system obscure. It is possible then, that D. confusus was confined to a limited area in some baited trees because the tree would not have been susceptible to attack had it not beenbaited (Bleiker etal 2003) In 1995, colonizationby secondary scolytids was limited to Crypturgus borealis andP. minutus. The year following mass attack, P. minutus was present in much larger numbers in halfofthe baited trees examined (Harder 1998). Rhizophagus dimiatus and C. borealis were found primarily in trees with advanced western balsam bark beetle brood development (Table 6). R. dimiatus is a bark beetle predator and may use D. confusus pheromones to locate itsprey.