document Historic, archived Do not assume content reflects current scientific l<nowleclge, policies, or practices. Bird Populations in Logged and Unlogged Western Larch/Douglas-fir Forest^ in Nortiiwestern l\/lontan^ So ^ ^ Bret W. Tobalske Raymond C. Shearer Richard L. Hutto THE AUTHORS RESEARCH SUMMARY BRETW. TOBALSKE is a Masterof Arts student in We detected differences in population sizes of zoologyatthe University of Montana in Missoula. He 10 species of breeding birds between harvested and earned a B.S. degree in zoologyfrom Southern Illinois adjacent uncut areasfollowing awinter(1988-89) seed University at Carbondale in 1988. His research inter- tree removal on Coram Experimental Forest in north- ests are in avian behavioral and morphological ecology. western Montana. Cutting unitswere relatively small, RAYMOND C. SHEARER received a B.S. degree in and live paperbirch {Betulapapyrifera), quaking aspen {Populus tremuloides), and black cottonwood (Populus timber management in 1957 and an M.S. degree in trichocarpa) trees as well as conifer and broadleaf silviculture in 1959, both from Utah State University in Logan, and a Ph.D. degree inforest ecology in 1985 snagswere left standing within the cutting units. During the summers of 1989 and 1990, tree swallows [Tachy- fromthe Universityof Montana, Missoula. Since 1957 he has been a research silviculturist, assigned to the cinetabicoloi) were abundant in clearcuts and not Station's Silviculture of Montane and Subalpine Forest detected in untagged stands. Similarly, dark-eyed juncos {Juncohyemalis) and pine siskins {Carduelis Ecosystems research work unit atthe Forestry Sci- ences Laboratory in Missoula, MI. His primary assign- pinus) were more common within clearcut and partially cut habitats thanwithin uncut stands. Golden-crowned ment has been the study of natural and artificial regen- kinglets {Regulussatrapa), Swainson'sthrushes {Cath- eration ofwestern larch and associated species. arusustulatus), variedthrushes (Ixoreus naevius), and RICHARD L. HUTTO is a professorof biology with the Townsend'swarblers (Dendroica townsendi) were less Universityof Montana in Missoula. He earned a B.A. abundant in the cut areas. Ruby-crowned kinglets degree in zootogy from UCLA in 1971, an M.S. degree {Reguluscalendula) and fox sparrows {Passerella in biologyfrom Northern Arizona University, Flagstaff, iliaca) were least abundant in both clearcut and contigu- in 1973, and a Ph.D. degree in biologyfrom UCLA in ousold-growth stands. Chipping sparrows {Spizella 1977. His research has included studies of the ecology passerina) were common in logged stands and inter- of migratory landbirds duringthe breeding season in spersed, untagged forest, and relatively rare in western North America, and during the winterseason in contiguous old growth. western Mexico. He is currently studying how local and Foliage-foraging species and tree gleanerswere less landscape-level parameters influencethe use of postfire abundant in harvested areas, while flycatching species habitats by songbirds in the Northern Rockies. and groundforagerswere more common there. Of the nesting guilds, conifertree-nesting species were least ACKNOWLEDGMENTS abundant in clearcuts, and ground nesters were more comnrranwithin the cutover habitats. Special thanks to Sallie Hejl and Dave Turnerfor The variattan in relative abundance among bird assistance in data analysis. James Hasbrouk prepared species in logged and untagged stands points out that a preharvest survey on the study area, vital to field management of forest resources for birds should focus planning, andJack Schmidt provided extensive field on diverse community-level habitat needs. If mitigation and computersupport. Volunteerswho assisted fortree-dependent species is an objective, we recom- in variouswayswith fieldwork include Andrea mend retainingwithin cutting units broadleaftrees Stephens, Debbie Conway, Bettina Gerlemann, Gina including paperbirch, quaking aspen, and black cotton- Berkemann, Tom O'Brien, BartTobalske, and Jean wood, and snags of all tree species. Slash piles may Totebusch. Fundingforthis studywas provided by the also contribute to the suitability of logged habitatfor Forest Service, U.S. Department of Agriculture. A Five certain bird species, and future research should attempt Valleys Audubon Society 1990 Field Biology Research to quantifythis factor. Awardwas grantedto the lead authorto assistwith aspectsofthe study related to red-naped sapsuckers. The use oftrade orfirm names in thispublication is forreaderinformation anddoes not imply endorsementbythe U.S. DepartmentofAgriculture ofanyproductorsen/ice. Intermountain Research Station 324 25th Street Ogden, UT84401 Bird Populations in Logged and Unlogged Western Larch/ Douglas-fir Forest in Northwestern Montana Bret W. Tobalske Raymond C. Shearer Richard L. Hutto INTRODUCTION about28 miles (45 km) east ofKalispell, MT, on the Flathead National Forest (fig. 1). It includes both The influence ofloggingon breeding-season bird the Terrace Hill sale area (THSA) and the adjacent populations hasbeen documented for a variety of Coram Research Natural Area (CRNA) (fig. 2). forest types in the Western United States: mixed conifer (Franzreb and Ohmart 1978; Mannan and Description Meslow 1984; Verner and Larson 1989), coastal Douglas-fir {Pseudotsuga menziesii) (Hagar 1960; The site occupies a portion ofthe watershed Meslow 1978), interior Douglas-fir (Medin 1985; drained by the South Fork ofAbbot Creek. The Medin and Booth 1989), giant sequoia {Sequoia- THSAranges in elevation from 3,360 to 4,500R dendrongiganteum) (Kilgore 1971), lodgepole pine (1,021 to 1,372 m), lies mostly southwest ofAbbot (Pinus contorta) (Austin and Perry 1979), and Creek, and is on northeast-facing slopes. The ponderosa pine {Pinusponderosa) (Szaro and Balda CRNA ranges in elevation from 3,500 to 4,458 ft 1979). (1,067 to 1,359 m), lies northeast ofAbbot Creek, Medin (1985) provided a summary ofbird species' and is mostly on southwest-facing slopes. Slopes responses to various logging methods based on ofboth areas range from 5 to 35 percent. Annual several ofthe cited studies. Although general precipitation averages 34 inches (86 cm). Tempera- patterns ofresponse were evidentfor many species, ture averages 60 °F (16 °C) from May through certain species, includingyellow-rumped warbler August (Hungerford and Schlieter 1984). (scientific names ofbird species censused in the The study area represents a serai stage ofthe study area are in the appendix) and hairy wood- subalpine fir/queencup beadlily {Abies lasiocarpal pecker were reported to increase, decrease, or show Clintonia uniflora)habitat type classifiedby Pfister no response following clearcutting. This suggests and others (1977). Western larch and Douglas-fir that studies ofbird responses to loggingmay be {Pseudotsuga menziesii var.glauca) are the domi- spatially or temporally specific. Thus, a need exists nant trees on the CEF. The old-growth larch is for other investigations ofthe effects ofloggingon composed mostly of300-year-old trees and occasion- birds, especially in unstudiedforesttypes. ally of500-year-old veterans. Other trees include Recently we studied the influence ofloggingon Engelmann spruce {Picea engelmannii), subalpine a wide range ofbirds in the Coram Experimental fir, lodgepole pine {Pinuscontorta), western white Forest, located within the western larch (Larix pine {Pinus monticola), paper birch {Betula occidentalis) forest cover type (Eyre 1980). This papyrifera), quaking aspen {Populus tremuloides), report describes 2 years ofobserved differences in andblack cottonwood {Populus trichocarpa). Domi- the avian assemblage followingloggingand site- nant shrubs include Rocky Mountain maple {Acer preparation treatments. glabrum) and fool's huckleberry {Menziesia ferruginea). Queencup beadlily and pinegrass STUDYAREA {Calamagrostis rubescens) are the primaryherbs. This study was located within the southern quarter ofthe Coram Experimental Forest (CEF), 1 Figure 1—Location ofstudy areawithin Coram Experimental Forest, northwestern Montana. logging road — Figure 2 Posttreatment stand cxanditions and location of bird-count/ vegetation sampling pointson theTerrace Hill sale areaand Coram Research Natural Area, Coram Experimental Forest, 1989 and 1990. 2 TREATMENTS The THSA was initially harvested from 1942 to 1944, leavingthree to five seed trees over 12 inches (31 cm) d.b.h. per acre (0.4 ha) (Fredeking 1953). The seed trees and trees that were unmerchantable in the 1940's were removed in the winter of1988-89 within partial-cut units, and more extensive over- story removal occurred in the clearcut units. Based on guidelines for the management ofcavity-nesting birds by McClelland and Frissell (1975), snags ofall species oftrees, along with living paperbirch, quak- ing aspen, and black cottonwood, were retained in the cuttingunits duringthe recent harvest (figs. 3 and 4). The THSA was 650 acres (263 ha) and included clearcut and partially cut units alongwith unlogged forest (fig. 1). There were five clearcuts ranging from 14 to 35 acres (6 to 14ha). The largestpartial- cutunit was 70acres (28 ha), and eightothers ranged between 5 and 40 acres (2 and 16 ha). Experimental controls were provided by the unlogged forest of 330acres (134 ha) within theTHSAand the adjacent 837-acre (339-ha) CRNA. Slash piling and scarification ofthe units was accomplished with a bulldozer in 1989. Most ofthe — slash piles were burned duringthe autumn of1989, Figure3 Evidenceofwoodpeckerforaging and sites with slopes greater than 20 percent were surrounds awildlifetree markeron awestern larch snag. The snag was retained forwildlife broadcastburned in September 1990. usewithin aclearcut on the Terrace Hill sale area. — Figure 4 Clearcut stand ontheTerrace Hillsale area, with western larch snags and living paperbirch retained forwildlife use. 3 METHODS 273 yards (250 m), a census plot was centered at the destination. Subsequent plots were selected in the Tobalske collected field dataon bird populations same manner with a newbearingfor all 10 points. and vegetation characteristics for this study. Our Bearings that would carry a plot within 109 yards statistical analysis was selected to quantify differ- (100 m) ofthe boundary ofthe natural area were ences among stand conditions in the study area. ignoredby returningto the origin and taking anew bearing. Design Vegetation was sampled once peryear during July 1989 and 1990 at the 10 points in each stand Duringthe 1989 and 1990 breeding seasons, bird condition that were selected forbird censusing populations were censused with fixed-point counts (table 1). Samplingfollowed the guidelines ofthe 109 yards (100 m) in radius. Species detections OcularMethod in Hahn and Jensen (1987) using were by sightor sound, includingbirds in flight over the General Plot Data and Ocular Plant Species the plot. Duringbothyears, each point was visited Data forms. Each plothad a 37-ft (11-m) radius, on three evenly spaced daysbetween June 1 and with one plot at the center ofeach bird-count point. July 7. Counts were donebetween one-halfhour Snags were included in all tree measurements. after sunrise and 10 a.m., for a period of 10 minutes Tree basal area was estimated with a "Relascope" each. prism, and average dominant tree d.b.h. was deter- Count points were located with a coordinate grid mined through diameter-tape measurements overlay placed on figure 2. After scaling distances offive trees, or the maximum number lefl standing between lines on the grid to be 218 yards (200 m) iffewer than five. Percentage ofcover was mea- apart on the ground and numbering the points of sured for trees, shrubs, graminoids, and forbs, using intersection sequentially, a random-numbers table horizontal estimates ofthe vertical projection of was read to select 10 points within each ofthree each life form. Tree and shrub cover was further stand conditions: clearcut, partial-cut, and unlogged sampled at three height categories: pole and larger, forest. Points were selected so that the radius ofthe sapling, and seedlingtree classes, and tall, mid, and plot would not intersect another stand condition. low shrub classes. These points defined the census plots on the ground; each was marked at the center with a survey flag. Because the CRNA (fig. 2) was difficult to map Analysis accurately, a different method was used there to The mean count per point for each bird species select census points. Beginning at the southwest censused in 1989 and 1990 was calculated by corner ofthe natural area stand, arandom number averaging data for the threeyearly visits to each was generated with a pocket calculator to determine point, giving 10 average counts for each bird a compass bearing. Afterfollowingthis bearingfor — Table 1 Estimates ofvegetation components within each ofthefour stand conditions atthe Terrace Hill study area and adjacent Coram Research Natural Area, Coram Experimental Forest, in 1989 and 1990 Stand condition Clear- Partial- Unlogged Natural Proba- Vegetation component cut^ cut^ forest area bility2 Tree basal area (ft^/acre) 8* 35* 104^ 112^ 0.001 Tree d.b.h. ofdominanttree (inches) 4* 7* I4B 15^ .001 Tree total cover (percent) 3* 24B 59c 63° .001 Pole and larger (percent) 1* 18^ 44c 48° .001 Sapling (percent) 1* 5* 13^ 13^ .001 Seedling (percent 1 2 2 2 1.000 Shrubtotal cover (percent) 13* 379 47B 44B .006 TalP (percent) 1* gAB IOAB 13^ .016 Mid3(percent) 10 22 25 18 .192 Low^(percent) 2* gAB 12^ .016 Graminoid total cover (percent) 9 11 6 4 1.000 Forbtotal cover (percent) 22 18 35 42 .032 ^Treesnagsofall speciesandliving paperbirchtreeswereleftstandingwherepossible. ^Lowprobabilityvalues indicatesignificantlydifferentmeansamong standconditions, Bonferroniadjustedtocontrolfor experimentwiseerror(SPSS Inc. 1983);superscriptlettersgroupsimilarmeansforeach vegetation component. ^Tall shrub, >10ft(3m); midshrub, 2to 10ft(0.6to3m);lowshrub, <2ft(0.6 m) (HahnandJensen 1987). 4 species within each stand condition andyear. The analysis (table 3). Statistical significance in all mean ofeach ofthese 10 averages was used to test tests was P < 0.05, employingthe Bonferroni for differences in relative abundance ofbird species adjustmentto control for experimentwise error rate among stand conditions andyears. Samphngveg- (SPSS Inc. 1983). We used two-way analysis of etation components at the census points once per variance to testfor significant interaction and main yeargenerated 10 sample values for the selected effects for a given species' abundance by stand variables within each stand condition. The mean condition andyear (SPSS Inc. 1983). Parametric ofeach set of10 samples was compared to the analysis wasjustified on the basis ofnormal prob- means for the other stands during each year. ability plots for the common species within the All bird species seen less than eight times during sample. Although count data is heavily right- both years' censusingover all points were arbitrar- skewed, the analysis ofvariance procedure is robust ily excluded from species-level statistical analysis for comparison ofsamples with similar distributions (table 2). These species were included in the guild (Sokal and Rohlf 1981). — Table 2 Mean numberof birds counted percensus point within thefourstand conditions atthe Terrace Hill study area and adjacent Coram Research Natural Area, Coram Experimental Forest, in 1989 and 1990 Stand condition Foraging Nesting Clear- Partial- Unlogged Natural Proba- Species^ guild^ guild^ cut* cut* forest Area bility' Killdppr GF GR 0.15 VV/?diUiiAy'Oq QO\WAI/iIfIt FC SC 07 n 1 nnn TD PC 97 1L/ H^iirv wnnHnprkpr TD PC .17 13 .19 .33 1 nnn GF PC .1o .1 1 nnn Olix/fi-^iripH fl\/r;?trhAr FC CB .17 07 .02 224 \A/0^tprn wonH-nPWPP FC CB 07 03 .05 1 nnn Trpo QW^IInuu FC SC 62* 07^ 0^ yj Steller'sjay FF CB .03 .05 .13 .05 1.000 Common raven GF CB .11 .03 .10 .22 1.000 Black-capped chickadee TG SC .30 .27 .40 .57 1.000 l^ountain chickadee TG SC .08 .17 .28 .28 1.000 Red-breasted nuthatch TG SC .30 .42 .58 .53 1.000 Brown creeper TG SC .03 .02 .12 1.000 Winterwren GF GR .13 .08 .17 1.000 Golden-crowned kinglet FF CT 0* 45AB 47AB .73B .016 Ruby-crowned kinglet FF CT .13* .52^ _40*B .17* .016 Mountain bluebird GF SC .13 .03 1.000 Townsend's solitaire FF GR .13 .25 .02 .192 Swainson'sthrush FF CB .15* 1.05^ 1.05B I.23B .002 American robin GF BT .58 .30 .22 .05 .096 Variedthrush GF BT .03* .05* .10* .38^ .016 Solitary vireo FF CB .05 .10 .07 .18 1.000 Warbling vireo FF BT ,13 .20 .18 .18 1.000 Yellow-rumped warbler FF CT .28 .35 .28 .25 1.000 Townsend'swarbler FF CT .43* 1.00^ 1.22^ 1.50^ .002 MacGillivray's warbler FF BT .02 .12 .13 .22 .864 Western tanager FF CT .10 .20 .10 .896 Chipping sparrow GF BT .68* .72* .52*B .I7B .016 Fox sparrow GF GR .18* 32AB .40*8 0* .016 Dark-eyed junco GF GR 1.27* .80* .25^ .13^ .002 Pine siskin FF CB .78* .57*B .17^ .08^ .032 Total No. species 32 32 29 30 28 25 'Scientificnamesofbirdspeciesarelistedin appendix. ^Foraging guild: FF =foliageforager; FC =flycatcher; TD=treedriller;TG= treegleaner; GF=groundforager(DiemandZeveloff 1980). ^Nestingguild:CT=conifertree; CB=coniferorbroadleaftree; BT= bushorsmall tree; PC = primarycavity;SC =secondarycavity;GR= ground(DiemandZeveloff 1980). ^Tree snagsofall speciesandliving paperbirchtreeswereleftstandingwhere possible. ^Lowprobabilityvalues indicatesignificantlydifferentmeansamong standconditions, Bonferroni adjustedtocontrol forexperimentwiseerror; superscriptlettersgroup similarmeansforeachspecies(SPSS Inc. 1983). 5 — Table 3 Mean numberof birds, byguilds, counted percensuspointwithinthefourstand conditions atthe Terrace Hill studyarea and adjacent Coram Research Natural Area, Coram Experimental Forest, in 1989 and 1990 Stand condition Numberof Clear- Partial- Unlogged Natural Proba- Gulld^ species cut^ cut^ forest area bility^ Foraging Foliageforager 23 2.4* 5.0^ 4.5^ 4.7B 0.001 Flycatcher 5 1.0* .3^ .1^ 0^ .001 Tree driller 4 .7 .5 .5 .9 .544 Tree gleaner 5 .7* gAB I.3BC 1.5c .003 Groundforager 14 3.4* 2.7*^ 1.8^ I.4C .001 Total 51 esting Conifertree 8 1.0* 2.6^ 2.5B 2.7B .001 . Coniferorbroadleaftree 10 1.4 2.0 1.6 1.8 1.000 Bush or small tree 10 1.5 1.6 1.2 1.0 1.000 Primary cavity 5 .9 .6 .6 1.1 .416 Secondary cavity 9 1.6 1.0 1.3 1.6 1.000 Ground 9 1.7* 1.5*B QBC .3c .001 Total 51 'Guildsadaptedfrom DiemandZeveloff(1980). ^ree snagsofall speciesandliving paperbirch leftstandingwhere possible. ^Lowprobabilityvaluesindicatesignificantlydifferentmeansamong standconditions, Bonferroni adjustedtocontrolfor experimentwiseerror; superscriptlettersgroupsimilarmeansforeachguild (SPSS Inc. 1983). The same statistical analysis was used on guild in the CRNA, while chipping sparrows were com- populations (table 3) and vegetation structure mon in all areas except the CRNA. (table 1). Several other species did not exhibit significant differences in abundance, but were less abundant RESULTS in clearcuts than elsewhere. These include brown creeper, winter wren, and MacGillivray's warbler The differences amongthe stand conditions in the (table 2). Likewise, species that were more abun- study area are described forbird populations and dantin harvested stands but did not exhibit signifi- vegetation components. cant differences include oHve-sided flycatcher, Townsend's solitaire, and American robin. Interest- Birds ingly, robins were least abundant in the CRNA and relatively common in the unlogged forestbetween Tobalske observed 51 bird species during harvestunits on the THSA. censusing ofthe study area (appendix), but only When species were grouped into foragingguilds, 32 species were sufficiently abundant to include in following Diem and Zeveloff(1980), foliage foragers the statistical analysis (table 2). Pine siskins and were least abundant in clearcuts. Similarly, tree varied thrushes exhibited a significant interaction gleaners were less abundant in clearcuts and partial effectbetween stand condition and year. Significant cuts than in unharvested areas. In contrast, fly- differences in abundance occurred among stand catchers and ground foragers were more abundant conditions for 10 species ofbirds. Tree swallows in harvested habitat, particularly clearcuts. Among were frequently detected in clearcuts, butvirtually the nestingguilds, conifer tree nesters were least absent elsewhere. Species that were less abundant abundant in clearcuts and ground nesters were in clearcut or partial-cut areas were golden-crowned most abundant in cutover sites. Ground nesters kinglet, Swainson's thrush, varied thrush, and and ground foragers had a significant interaction Townsend's warbler. Conversely, dark-eyedjuncos effectbetween stand conditions andyears because and pine siskins were more abundantin harvested ofa marked increase in abundance in harvested habitat than elsewhere. Ruby-crowned kinglets and areas in 1990. fox sparrows were least abundant in clearcuts and 6