<<

DISTRIBUTION AND POPULATION SIZE OF GNATCATCHERS ON THE , 1993-1997

JONATHAN L. ATWOOD, SOPHIA H. TSAI, CAROL A. REYNOLDS, and MICHAEL R. FUGAGLI, Manornet Center for Conservation Sciences, P.O. Box 1770, Manornet, Massachusetts 02345

The potential effects of fragmentationof natural communitiesby urban developmentare dramaticallyevident on the Palos Verdes Peninsulaof County, California.Because of the long history of agriculturaland, more recently,intensive urban developmentof the Los Angelesbasin, tracts of natural scrub habitat on the peninsulahave becomeincreasingly isolated and smallersince the mid-1900s (Fink 1966, Gales 1974). The populations of the California Gnatcatcher(F'olioptila californica) nearest that on the peninsula occur in Los Angeles County near Montebello and in Orange County near Fullerton and Newport Beach, all approximately45 km from the Palos Verdes Peninsula across the urbanized areas of . In fact, Palos Verdes exemplifies the situation that California's Natural Community ConservationPlanning (NCCP) program is intended to avoid--small,disjunct islands of natural habitat surroundedby a vast, nearly continuousurban "ocean" (Calif. Dept. Fish & Game 1993). Long-term maintenance of viable natural communitiesin such a landscape poses an especiallydifficult conservationchallenge. Through preparationof land-useplans that addressthe ecologicalrequire- ments of selected"target species" at a regionalscale, the NCCP attemptsto allow economic developmentwhile ensuring the protection of biologically viable tracts of natural habitat (Atwood and Noss 1994, Reid and Murphy 1995). Two songbirds, the California Gnatcatcher and Cactus Wren (Carnpylorhynchus brunn½icapillus),have been selected throughout much of as the primary flagship species for a test-case applicationof an NCCP to the coastalsage scrub (Calif. Dept. Fish & Game 1993, Atwood and Noss 1994, Reid and Murphy 1995). Despite the centralityof these speciesto this planningprocess, however, litfie is actually known of their behavioralecology (Atwood 1993, Rea and Weaver 1990). Consequenfiy,to obtain basic biologicaldata requiredfor effectiveconserva- tion planning, in 1993 Manomet Center for ConservationSciences began an ongoingstudy of gnatcatchersand wrens on the Palos Verdes Peninsula. Here we summarizesome of the resultsof the first five yearsof this project, especiallywith regard to annual fluctuationsin populationsize and use of natural habitats by gnatcatcherson the peninsula.

METHODS

Vegetation Analyses Discrete units of natural vegetationon the Palos Verdes Peninsulawere identifiedon plasticoverlays affixed to color aerial photographs,scale I inch

340 WesternBirds 29:340-350, 1998 CALIFORNIA GNATCATCHERS ON THE PALOS VERDES PENINSULA

-- 1200 feet, dated 8 January 1992 and 19 April 1993. The minimum size of delineated polygons was approximately 0.2 ha. We inspected each vegetation unit in the field, either from nearby vantage points through binoculars or by visits to the sites. We adjusted polygon boundaries as appropriate,and visuallyestimated the percent cover of the followingplant speciesor categoriesfor each unit: (a) , (b) Eriogonum spp. (includingE. fasciculatum and E. cinereum), (c) Salvia spp. (including $. leucophyllaand S. mellifera), (d) Opuntia littoralis, (e) , (f) Opuntia prolifera, (g) Encelia californica, (h) Elymus condensatus,(i) , (j) Foeniculum vulgate, (k) Salsola kali, (1) miscella- neous ornamentalshrubs, (m) disturbedgrassland, (n) riparian, and (o) bare ground. Vegetation data were entered in a geographic-informationsystem by means of ArcInfo software. Various classificationsystems have been applied to subassociationsof coastalsage scrub (Munz 1970, Thorne 1976, Kirkpatrickand Hutchinson 1977, Westman 1981, DeSimone and Burk 1992, Jones and Stokes Associates,Inc. 1993). Here, we used a cluster analysis to identify major vegetation groupings, then prepared a dichotomouskey reflecting these results(Table 1). Using this key, we then categorizedeach mapped polygon as one of four scrub subassociations (lemonadeberry scrub, sage- lemonadeberryscrub, sagebrushscrub, and cactusscrub) or two types of disturbed grassland (grassland-scrubecotone and grassland). Residential areas, landscaped parks, playing fields, golf courses, shopping centers, roads, and ruderal areas with >50% cover by Eucalyptus, Acacia, or other ornamental species were consideredto be developed.Polygons classified as the same subassociationand sharing a common boundary were subse- quently merged to form larger units than originallyhad been mapped in the field. Area estimatesfor vegetationpolygons, uncorrected for slope effects, were obtained through ArcInfo.

'ral•le ! Key to and Grassland Categories Used in Vegetation Mapping on the Palos Verdes Peninsulaa

A Percent cover ofAc + Er + Sa + Op + En + Rh < 30%...... B A' Percent cover ofAc + Er + Sa + Op + En + Rh_> 30%C...... B Percent cover of Ac + Er + Sa + Op + En + Rh <15%...... Grassland B' Percent cover of Ac + Er + Sa + Op + En + Rh _>15% ...... Grassland-scrubecotone C Percent cover of Rh >_percent cover of Ac + Er + Sa + Op + En ...... Lemonadeberryscrub C' Percent cover of Rh < percentcover of Ac + Er + Sa + Op + En ...... D D Percent cover of Sa + Rh _>percent cover of Ac + Er + Op + En ...... Sage-lemonadeberryscrub D' Percent cover of Sa + Rh < percentcover of Ac + Er + Op + En ...... E E Percent cover of Op _>25% ...... Cactus scrub E' Percent cover of Op < 25%...... Sagebrush scrub aAc, Arternisia californica; Er, Eriogonurn spp.; Sa, Salvia spp.; Op, Opuntia spp.; Rh, Rhus integrifolia; En, Encelia californica.

341 CALIFORNIA GNATCATCHERS ON THE PALOS VERDES PENINSULA

PopulationSurveys All major areas of natural habitat on the Palos VerdesPeninsula were surveyedfor breedingCalifornia Gnatcatchers from Februaryto June of each year of the study(1993-1997). Surveyswere generallyconducted before 11:00 and after 16:00, in weather deemedacceptable in terms of wind and temperature.Taped recordings of gnatcatchercalls were usedto elicitresponses. In areaswhere adjacent territories of unbandedbirds posed potentialconfusion over the numberof pairsactually present, teams of two to four biologistsrevisited the site to observeall birdsin questionsimulta- neously.Because of the limitedamount of suitablehabitat and our popula- tion estimates'being based on observationsof uniquelybanded birds, the locationsof simultaneouslyactive nests, or simultaneousobservations of unbandedbirds, we feel confidentin the accuracyof our results.There was no substantialchange in surveyeffort or techniqueduring the study.We recordedthe locationsof breedingpairs on 7.5-minuteUSGS topographic mapsthat were enlarged to a scaleof approximately1 inch-- 500 feet,then digitizedwith ArcInfo.

RESULTS

Habitat Distribution We locatedapproximately 642 ha of coastalsage scrubon the Palos VerdesPeninsula (Figure 1), of which170 ha (26%) consistedof fragments of lessthan 5 ha each. Only one area of scrubexceeded 100 ha; this single tractrepresented approximately 17% of the totalcoastal sage scrub remain- ing on the peninsula. Sagebrushscrub represented 46% of the total coastalsage scrub on the Palos Verdes Peninsula(Table 2). The least common of the four major subassociationswas cactusscrub, which made up only 6%. Approximately 198 ha of grassland-scrubecotone, where coastal sage scrub plants repre- sented15-29% of the vegetationcover, and 829 ha of grasslandoccurred on the peninsula. Approximately59% of coastalsage scrub remaining on the PalosVerdes Peninsulawas locatedwithin the jurisdictionof the city of RanchoPalos Verdes(Table 2). Evengreater percentages of the total remainingamounts of sagebrushscrub (69%) and cactusscrub (81%) occurredwithin Rancho Palos Verdescity limits. Scrub in the citiesof PalosVerdes Estates and RollingHills was classified predominanfiy as lemonadeberryscrub or sage- lemonadeberryscrub (Table 2).

GnatcatcherPopulation Size and Distribution We found 26 to 56 breedingpairs of the CaliforniaGnatcatcher during the 1993-1997 nestingseasons (Figure 2). The populationwas greatest (56 pairs)in 1994. From 1994 to 1995 the numberof breedingpairs declined 54%, then increased50% from 1995 to 1996. None of thesepopulation fluctuationswas evidentlyrelated to changesin habitatavailability, as the amountof coastalsage scrub on the peninsuladid not varysignificantly from

342 CALIFORNIA GNATCATCHERS ON THE PALOS VERDES PENINSULA

0 1000 2000

ß SagebrushScrub, Cactus Scrub ß Sage-LernonadeberryScrub, Lemonacleberry Scrub •iD Grassland,Grassland-Scrub ecotone [] Developecl

Figure1. Distributionof coastalsage scrub, grassland-scrub ecotone, and grasslandon the Palos VerdesPeninsula, 1993-1997. Blacklines, coastline and cityboundaries.

1993 to 1997. Most gnatcatcherson the peninsulawere located in the city of Rancho Palos Verdes (Table 3). In all years of our study the majority of pairs were located in polygons classifiedas sagebrushscrub (Table 3; Figure 3). In the absenceof detailed territory mapping, however, plotting of locations must be somewhat arbi- trary and therefore unlikelyto fully reflect the habitat as the birds perceive it. We suggestthat habitat preferencesare better demonstratedby the broader landscapewhere pairs are found (Figure 3). For example, during 1997, 14 pairs were located in areas classified as "grassland"or "grassland-scrub ecotone," raising questionsregarding precisionof the mapped locations of these pairs and/or the vegetation classificationof these polygons (Table 3). Habitat analysisat a larger scale, however, showed that virtually all of these pairs were located in areas containing large amounts of sagebrushscrub, grassland, and grassland-scrubecotone (Figure 3). On the Palos Verdes

343 CALIFORNIA GNATCATCHERS ON THE PALOS VERDES PENINSULA

Talkie 2 Extent (ha) of PrincipalCoastal Sage Scruband GrasslandHabitats on the Palos Verdes Peninsula, 1993-1997

Patch size Jurisdictionø

Total Vegetationtype area Mean SD n Max RPV PVE RH RHE SP

Sagebrushscrub 298 2.79 2.90 107 22.4 205 27 33 14 19 Cactusscrub 37 1.06 0.76 35 2.9 30 5 2 0 0 Sage-lemonadeberry scrub 101 3.05 4.25 33 24.5 46 0 49 6 0 Lemonadeberry scrub 206 4.91 6.23 42 26.7 97 42 61 7 0 Grassland-scrub ecotone 198 2.28 2.42 87 15.9 135 25 24 2 12 Grassland 829 6.74 12.77 123 69.3 579 57 41 58 94

øRPV.Rancho Palos Verdes; PVE, PalosVerdes Estates; RH, RollingHills; RHE, RollingHills Estates; SP, San Pedro.

Peninsula,as well as elsewherein southernCalifornia (Atwood unpubl. data), CaliforniaGnatcatchers are more likelyto occurin landscapeswhere coastalsage scrubis patchilydistributed within a grasslandmatrix than in continuoustracts of scrubas often occurnear . Some apparentlysuitable habitat on the peninsulawas unoccupiedby breedinggnatcatchers throughout the study.We neverfound breeding pairs in

6O

50-

40-

30-

2O I I I I I 1993 1994 1995 1996 1997

YEAR Figure2. Annualvariation in numberof breedingpairs of CaliforniaGnatcatchers on the Palos Vetdes Peninsula, 1993-1997.

344 CALIFORNIA GNATCATCHERS ON THE PALOS VERDES PENINSULA

Table 3 Distribution of California Gnatcatchers on the Palos Verdes Peninsulaby VegetationType and Civil Jurisdiction

Vegetationtype a Jurisdictionb

Year SS CS SL LS GS GR RPV PVE RH RHE SP

1993 34 9 2 2 2 2 42 3 1 0 5 1994 37 9 1 1 7 1 46 3 1 1 5 1995 15 4 2 1 2 2 24 1 1 0 0 1996 28 1 0 0 6 4 38 1 0 0 0 1997 15 5 3 1 8 6 35 1 1 0 1

aSS, sagebrushscrub; CS, cactus scrub; SL, sage-lemonadeberryscrub; LS, lemonadeberry scrub; GS, grassland-scrubecotone; GR, grassland. bRPV,Rancho Palos Verdes; PVE, PalosVerdes Estates; RH, RollingHills; RHE, RollingHills Estates; SP, San Pedro. approximately427 ha of coastal sage scrub vegetation,including approxi- mately 159 ha that were classifiedas sagebrushscrub (representing53% of the total amount of this subassociation).Patch size effectively predicted presenceor absenceof gnatcatchersin polygonsclassified as sagebrushscrub (P < 0.001, logisticregression) (Figure 4). Patchesof this subassociationwith gnatcatcherswere significanfiylarger than unoccupiedareas (occupied:mean 4.07 ha, SD 3.88, n = 39, range 0.34-22.4 ha; unoccupied:mean 2.05 ha, SD 1.81, n = 68, range 0.18-10.48 ha; P = 0.001, Mann-WhitneyU test). Patch size showedno significantrelationship with gnatcatcheroccupancy for polygonsdelineated as cactusscrub, sage-lemonadeberry scrub, or lemonade- berry scrub (P > 0.10, logisticregression) (Figure 4). Of 34 patches of coastalsage scrub (all subassociationscombined) used by breedinggnatcatchers during the study,7 (21%) supportedpairs during all 5 years, 3 (9%) during4 of the years, 5 (15%) during 3 of the years, 9 (26%) during 2 of the years, and 10 (29%) during only a single breedingseason. Patch size was a successfulpredictor of the consistencyof occupancyby gnatcatchers (P = 0.016, logistic regression). There was a significant differencein area among patchesoccupied for one to five years (P = 0.032, Kruskal-Wallistest), with patches used for four or five years being substan- tially larger than patchesoccupied one, two, or three years (Figure5). During each year of the study,a large portion of the peninsula'sentire gnatcatcher population was concentrated in Agua Amarga Canyon, the vicinity of Edward's Canyon (McCarrell's Canyon), and near Klondike Canyon (Figure6). These three localities,which includedonly about 33% of the area of coastal sage scrub (all subassociations)on the peninsula, supportedfrom 41% to 58% (mean 46%) of the annual breedingpopulation from 1993 to 1997.

DISCUSSION

Shaffer (1981), Gilpin and Soula (1986), and Hanski (1991) summarized major forces that may contributeto populationextinction. "Deterministic"

345 CALIFORNIA GNATCATCHERS ON THE PALOS VERDES PENINSULA

60% [] AVAILABLENATURAL VEGETATION 50% - m 1997 PAIRLOCATIONS m VEGETATIONTYPE WITHIN 40% - 100 rn RADIUS CIRCLE

30% - T 20% - T lO%

0% ss GR GS CS SR LS

Habitat Fi•Jre 3. Dist•bution of California Gnatcatcherpairs relative to habitat availabilityon the Palos Verdes Peninsula.Cross-hatched bars, veõetation types as percentaõe of total natural habitats. Gray bars, percentaõeof 1997 pairs in each veõetationtype. Black bars, mean value of each veõetationt•pe based on its abundancewithin 100-m radius of each pair's location in 1997. Error bars, 2 standarderrors. $$, saõebrush scrub; GR, õrassland;G$, õrassland-scrubecotone; CS, cactus scrub; $R, saõe- lemonadeberryscrub; L$, lemonadeherW scrub.

extinctionsmay result from the removalof an essentialresource, such as a specific habitat type on which an organism is dependent. "Stochastic" extinctionsmay result from normal, random perturbations,such as demo- graphic, genetic, and environmentalvariations, as well as catastrophes.As noted by Gilpin and Soul• (1986), "decayin one factor (suchas population size)can exacerbatenot only itselfbut also the behaviorof other factors(such as inbreeding and fragmentation)." Our resultsindicate that the populationof CaliforniaGnatcatchers on the Palos Verdes Peninsulais characterizedby a number of factorsthat place it at a high risk of extinction. Only about 30 breedingpairs of gnatcatchers now remain on the peninsula, and the surroundingmegalopolis of Los Angeles makes immigrationof birds from other portions of the species' range in southernCalifornia impossible or extremelyunlikely. Furthermore, gnatcatchersremaining on the peninsulaare distributedover a very frag- mented landscape,with approximately50% of the total breedingpopulation concentratedin three relativelysmall areas that could easilybe disruptedby developmentor fire. Soul• (1986) suggestedthat extinction due to demo- graphic stochasticityis a real threat for populationsof 20 or fewer breeding females.On the Palos Verdes Peninsulafactors not yet fully understoodbut most likely related to weather variations have caused one-year declines exceeding 50%.

346 CALIFORNIA GNATCATCHERS ON THE PALOS VERDES PENINSULA

$o 2o SagebrushScrub LemonadeberryScrub

zo 40I• • •_ ,,,[] CAGN•_,ABSENT , o Cactus Scrub Sage-Lemonadeberry Scrub 15 [] CAGNAlLSENT z•l$ []ß cRc.•CAGNABS•'rPRESENT ß CAGNPRESENT lO u.

0 ;• 4 e e 10 12 14 16 18 20 22 24 ;'6 PATCH SIZE (ha) PATCH SIZE (ha)

Figure4. CaliforniaGnatcatcher distribution by sizeof habitatpatch on the Palos VerdesPeninsula, 1993-1997.

Conservation of California Gnatcatchers in the Palos Verdes area will requirecareful protection of nearlyall areasof currentlyoccupied coastal sage scrub, especiallyin the Agua Amarga Canyon, Edward'sCanyon (McCarrell'sCanyon), and KlondikeCanyon areas. Additional efforts should also focuson increasingthe overallamount of sagebrushscrub through restoration,especially on lands adjacentto currentlyoccupied habitat. However,given the PalosVerdes gnatcatcher population's current highly threatenedstatus, plans to mitigate loss of presentlyoccupied habitat throughlong-term habitat restoration projects may not be realistic. While conclusionsregarding the gnatcatcher'slong-term status on the peninsulamust await completion of furtherstudies and planning,at thistime we see no evidencesuggesting that the Palos Verdes populationis not severelythreatened. Under presentknowledge, it wouldcertainly be fool- hardyto baseplanning decisions elsewhere in coastalsouthern California on the mere fact that CaliforniaGnatcatchers now tenuouslypersist on the peninsula.

SUMMARY

The population of the California Gnatcatcher on the Palos Verdes Peninsulaof LosAngeles County provides a uniqueopportunity for studying the species'biology in a fragmentedand highlyisolated landscape. Annual surveyslocated 51, 56, 26, 39, and 38 breedingpairs from 1993 to 1997. Substantialareas of apparentlysuitable habitat were not usedby breeding gnatcatchers.Approximately 50% of each year's breedingpairs were concentratedin approximately33% of the peninsula's642 ha of coastal sagescrub. Of discretehabitat patches occupied by breedingbirds, only 21% were occupiedduring all 5 years.Patch size effectively predicted both the

347 CALIFORNIA GNATCATCHERS ON THE PALOS VERDES PENINSULA

6O

50-

40-

30-

ZO

10 W -I-

1 2 3 4 5

Figure5. Meanpatch size and number of yearsof occupancyby CaliforniaGnatcatch- ers on the Palos Verdes Peninsula, 1993-1997. Error bars, 2 standarderrors. likelihoodof gnatcatcheroccupancy of patchesof sagebrushscrub and the year-to-yearconsistency of occupancyof scrubpatches (all subassociations combined).Our resultssuggest that the PalosVerdes population of Califor- nia Gnatcatchersis at a high risk of extinction becauseof its small size, concentrateddistribution, and isolation from other populationsin Los Angelesand Orange counties.

ACKNOWLEDGMENTS

Financialsupport for variousaspects of thiswork wasprovided by the U.S. Navy, Southern California Edison, Ed Almanza and Associates,the National Fish and WildlifeFoundation, the PalosVerdes Peninsula Land Conservancy,the trusteesof Manomet Center for ConservationSciences, and an anonymousdonor. Jack Cameron,Barbara Courtois, Jim and LyndaLuttrell, Amy Miller,Nancy Nicolai, Tim Overbey, and Mike Walther all contributedimportant field assistance.George Cox and John Rotenberrygave usefulcomments that improvedan early draft of the manuscript.The U.S. Navy providedaccess to the DefenseFuel Support Point. The EnvironmentalSystems Research Institute, Inc. (ESRI)provided GIS capabilitiesin the form of PC ArcInfosoftware; Stacie Grove of Manomet patientlyexplained how to make it work. Barbaraand Sarah Atwood encouragedthe seniorauthor during his travels to southern California. This researchwas conductedunder U.S. Fish and Wildlife Service endangered speciesrecovery permit PRT-800922, U.S. Fish and Wildlife Service master bird- bandingpermit 09996, Californiascientific-collecting permits and a California Departmentof Fishand Game memorandum of understandingdated 25 August1992 (as amended).

348 CALIFORNIA GNATCATCHERS ON THE PALOS VERDES PENINSULA

0 I 2 Kilometers

Figure 6. Principal locations of California Gnatcatcherson the Palos Verdes Penin- sula, 1993-1997. Black areas, habitat that supported >5% of each year's total population;gray areas, other regionsof coastalsage scrub. Area 1, KlondikeCanyon and vicinity; area 2, Agua Amarga Canyon; area 3, Edward's (McCarrell's)Canyon and vicinity. Fine black lines, city boundaries.

LITERATURE CITED

Atwood, J. L. 1993. CaliforniaGnatcatchers and coastalsage scrub: The biologicalbasis for endangeredspecies listing, in Interfacebetween Ecology and Land Development in California(J. E. Keeley, ed.), pp. 149-169. S. Calif. Acad. Sci., Los Angeles. Atwood, J. L., and Noss, R. E 1994. Gnatcatchersand development:A "train wreck" avoided? Illahee 10:123-132. Calif. Dept. Fish & Game. 1993. Southern California coastal sage scrub Natural Communities ConservationPlan: Scientific review panel conservationguidelines and documentation(order from Calif. Dept. Fish & Game, 1416 9th Str.eet, Sacramento 95814). DeSimone, S., and Burk, J. H. 1992. Local variation in fioristics and distributional factors in Californian coastal sage scrub. Madrofio 39:170-188. Fink, A. 1966. Time and the Terraced Land. Howell-North Books, Berkeley. Gales, D. M. 1974. Handbook of Wildflowers, Weeds, Wildlife, and Weather of the South Bay and Palos Verdes Peninsula.FoldaRoll Co., Palos Verdes Peninsula. Gilpin, M. E., and SoulS, M. E. 1986. Minimum viable populations:Processes of species extinction, in ConservationBiology: The Science of Scarcity and Diver- sity (M. E. SoulS, ed.), pp. 19-34. Sinauer Assoc., Sunderland,MA.

349 CALIFORNIA GNATCATCHERS ON THE PALOS VERDES PENINSULA

Hanski, I. 1991. Single-speciesmetapopulation dynamics: Concepts, models and observations,in MetapopulationDynamics: Empirical and Theoretical Investiga- tions (M. Gilpin and I. Hanski, eds.), pp. 17-38. AcademicPress, London. Jones and Stokes Assoc., Inc. 1993. Methods used to survey the vegetation of Orange County parks and open space areas and the Irvine Company property. Prepared for County of Orange, Environmental Management Agency, Santa Ana (availablefrom Wilson Ornithol. So½., Mus. Zool., Univ. Mich., Ann Arbor, MI 48109-1079). Kirkpatrick, J. B., and Hutchinson, C. F. 1977. The community composition of California coastal sage scrub. Vegetatio 35:21-33. Munz, R A. 1970. A CaliforniaFlora. Univ. of Calif. Press, Berkeley. Thorne, R. E 1976. The vascularplant communitiesof California,in Plant Commu- nities of Southern California(J. Latting, ed.), pp. 1-31. Calif. Native Plant So½. Spec. Publ. 2. Rea, A.M., and Weaver,K. L. 1990. The taxonomy,distribution, and statusof Cactus Wrens. W. Birds 21:81-126. Reid, T. S., and Murphy, D. D. 1995. Providing a regional context for local conservationaction. BioS½ienceSuppl. 1995 (Science and BiodiversityPolicy): 84-90. Shaffer, M. L. 1981. Minimum populationsizes for speciesconservation. BioS½ience 31:131-134. SoulS, M. E. 1986. ConservationBiology: The Science of Scarcity and Diversity. Sinauer Assoc., Sunderland, MA. Westman, W. E. 1981. Factorsinfluencing the distributionof speciesof Californian coastal sage scrub. Ecology 62:439-455.

Accepted 6 July 1998

350