ABUNDANCE AND DISTRIBUTION OF THE AND AMERICAN IN THE SAN FRANCISCO BAY AREA, CALIFORNIA

JOHN P. KELLY and KATHERINE L. ETIENNE, CypressGrove ResearchCenter, AudubonCanyon Ranch, Marshall, California 94940 JENNIFER E. ROTH, Point Reyes Observatory,4990 ShorelineHighway, StinsonBeach, California94970

ABSTRACT:During the breedingseason of 1999, we surveyedfrom roadsin the San FranciscoBay areato determinethe regionalabundance and distributionof the Common Raven ( corax) and American Crow (C. brachyrhynchos).Ravens concentratedalong the outercoast and occurredin relativelylow numbersin some interiorareas, whereas the numberof crowsincreased significantly from the outer coastto interiorand bayshore locations. Both occurred in significantlygreater densitiesalong urban and suburban survey routes than along rural routes, but dramatic exceptionswere evident in someareas. Data from BreedingBird Surveys, Christmas Bird Counts,and breedingbird atlases yielded similar distribution patterns. Breeding BirdSurveys and ChristmasBird Counts revealed strong regional increases in both species,but annual trends varied substantially at localscales. This variation, as well as significantabundance variation among survey routes, suggested considerable local differencesin eitherhabitat suitability or capacityfor furtherpopulation growth. Our resultssuggested that ravenand crow populationsmay increasein both ruraland developedareas undergoing rapid urbanization and that localconditions, rather than whetherthe habitatis ruralor urban,may influenceregional patterns.

Abundancesof the Common Raven (Corvus corax) and American Crow (C. brachyrhynchos)have increasedsubstantially over much of (Marzluffet al. 1994, Boarman and Heinrich 1999, Sauer et al. 2001). The Common Raven occursthroughout California but is scarcein much of the CentralValley, the centralcoast from the Santa LuciaMoun- tainssouth to northwesternVentura County, and irrigatedportions of the Coachella,Imperial and lowerColorado River valleys in the southeastcorner of the state.The AmericanCrow is widespreadbut absentfrom somedrier westernparts of the SanJoaquin Valley, interior foothills, and the southeast- ern deserts(Small 1994). Throughouttheir range, ravensappear to be invadingagricultural areas to a greaterextent than urbanareas (but are common in many urban areas), whereascrows appear to be invading urbanizedareas more rapidlythan agriculturalareas (Marzluff et al. 1994, Marzluffand Restani1999). Both speciesare residentthrough most of the San FranciscoBay area, a highlyurbanized region with large sectionsthat remainundeveloped or usedfor agriculture.Both specieshave increased in numberrecenfiy in the southernportion of the region(Coston 1998). The overallstatus and distributionof these corvidswithin the San FranciscoBay area, however,have not been addressed. To determine the abundance and distribution of ravens and in the San FranciscoBay area,we surveyedthe regionduring the springof 1999 (Figure1). Becausefield observations were limited to repeatedroad surveys alongparticular routes within a singleseason, we comparedthe resultswith existingBreeding Bird Survey (BBS; Sauer et al. 2001), AudubonChrisb•nas

202 WesternBirds 33:202-217, 2002 COMMON RAVEN AND AMERICAN CROW IN THE SAN FRANCISCO BAY AREA

11

/• Mt. Diablo

San 12 Pacific Ocean

San Jose

0 km North 50 km

Figure1. Routesused in road surveysof Common Ravensand AmericanCrows in the San FranciscoBay area. Labels,route-identification numbers; thin lines,county boundaries(see Figures 5 and 6 for countynames).

Bird Count (CBC; Butcher 1990), and breedingbird atlas (BBA; Robbins 1990) datato providea more thoroughperspective on regionaldistribution and to evaluatelocal and regionaltrends in abundanceof thesespecies.

STUDY AREA

The studyarea coveredall nine countiesadjacent to San FranciscoBay: Marin, Sonoma, Napa, Solano, Contra Costa, Alameda, Santa Clara, San Mateo, andSan Francisco(Figure 1). The studyarea did not extendinto the deltaeast of the confluenceof the Sacramentoand San Joaquinrivers. We did not conductroad surveysalong the outer coastof Sonoma County or alongthe EastBay shoreline,although these areas are includedin analyses of BBS, CBC, and BBA data. Analysisof BBS data includedsome routes that extendedslightly beyond the bay area counties,into coastalSanta Cruz Countyto the south,coastal Mendocino County to the north, San Joaquin County to the east, and Yolo County to the north (Figure2).

203 COMMON RAVEN AND AMERICAN CROW IN THE SAN FRANCISCO BAY AREA

187

Napa

TB

Point

Pacific Ocean

0krn North 50krn 319 9

Figure2. BreedingBird Surveyroutes (bold lines and associatednumber codes) and ChristmasBird Count circles (letter codes) in the San FranciscoBay area. Thin lines, countyboundaries (see Figures 5 and 6 for countynames). See Tables2, 3, and 4 for details on each route or circle.

Above the coastalterraces and alluvialshorelines, the area is characterized by rollinghills and mountainsof the Coast Range. Moist coastredwood (Sequoiasempervirens), Douglas fir (Pseudotsugamenziesii), and mixed broad-leavedevergreen forests dominate the outercoastal drainages, giving way to wider expansesof grassland,chaparral, and woodlandto the east.Nonnative eucalyptus (predominantly blue gum, Eucalyptus globulus) occursthroughout the regionin patchesmostly smaller than one hectareor as narrow windbreaks. Urbanand suburban habitats are concentratedin centralSonoma County near Santa Rosa, in southern Napa County near Napa, and south of Sonoma,Napa andSolano counties along bay shorelines. The shorelineand terracesalong the outercoast and northernSan Pabloand Suisunbays are generallyrural and open. Rural areas are primarilyused for agriculture,with range cattle productiondominating most areas, dairy ranchinglocally dominantin the PointReyes area, and farmsand vineyardscharacterizing much of the northernportion of the region.

204 COMMON RAVEN AND AMERICAN CROW IN THE SAN FRANCISCO BAY AREA

METHODS

From 26 Marchthrough 21 June 1999 we countedCommon Ravens and AmericanCrows along 18 surveyroutes averaging 49 + 14.0 (standard deviation)km in length, establishedalong roads throughoutthe region (Figure1). We recruited40 experiencedbirders, capable of distinguishing crowsand ravens in thefield, as volunteer observers to conductmid-morning roadsurveys along each route twice monthly on standardizeddates. Because each observerconsistently surveyed the sameroute, ratherthan rotating among routes,observer variability could have influencedthe resultsto an unknownextent. Surveyteams consisting of one driverand one observer traveledat speedsof 56 to 72 km/h. Eachsurvey began from a standard startingpoint. For each raven or crow observation,observers recorded the distancealong the surveyroute, speciesname, group size (numberof individualswithin 100 m of each other),and perpendiculardistance (< 100 m, 100-200 m, or >200 m) from the road. We distributedsurvey routes systematically to sampletwo land-usetypes (>75% ruralvs. >75% urban/suburban)and threesubregions (outer coast, bay shore, and inland).We apportionedthe routesto approximatethe extent of each land-usetype and subregionwithin the studyarea and maximizedthe distancesbetween routes to ensureindependence of counts (Figure1). To keep surveyspeed down, we avoidedfreeways and major highways.In urbanizedareas, we selectedroutes that allowedobservers to maintainnormal survey speeds, although temporaw slowing was necessary on someoccasions. Given these constraints, suitable survey routes for each land-usetype and subregionwere sometimes determined by availableroads. Open/ruralhabitat in the interiorof the easternportion of the studyarea andsalt ponds in the southernpart of the baywere undersampled relative to their extents. We usedthe numberof birdsobserved per kilometerof surveyroute to index corvid densities. We then examined differences between rural and urban/suburbanhabitats, and amongcoastal, interior, and bayshoreloca- tionsusing analysis of variance.Before analysis, we (natural)log-transformed the data so that densitiesdid not differ significantlyfrom normality(P < 0.05). Post hoc testsfor differencesamong groupsincluded Bonferroni adjustmentsfor experimentwiseerror. Because routes were selected system- aticallyon the basisof land use, subregion,and availabilityof roads,we modeledvariation among routes nested within land use and location types as fixedeffects. However, because routes covered a substantialportion of the studyarea and were selected to achievea representativesample of regional populations,we alsotested (wherever indicated in Results)the effectsof land useand subregionagainst "random" variation among routes as a proxyfor other,unknown influences on generalregional patterns (Bennington and Thayne 1994). Indicationsof such influencesshould be interpretedcau- tiouslybecause selection of surveyroutes was systematic rather than random. We alsoexamined BBS andCBC data,to determineregional trends in the numbersof ravensand crowsand to compareabundances with patterns suggestedby ourroad-survey data. The BBSis also a roadsidesurvey, based on 50 three-minutepoint countsconducted along thousandsof 39.4-km

205 COMMON RAVEN AND AMERICAN CROW IN THE SAN FRANCISCO BAY AREA routes over much of North America (Sauer et al. 2001). Because BBS surveysdate back as far as 1966, they providea perspectiveon changesin regionalbreeding-season abundance. We examinedlong-term trends in winter abundancewith CBC data, which are based on total counts con- ductedannually within standard circles of 24.1-km diameter(458 km2; Butcher1990). We analyzedBBS and CBC trendsby linearroute regres- sion,with regional trends expressed in percentchange per yearweighted by mean abundanceand numberof yearscovered at eachBBS routeor CBC location(Butcher et al. 1990, Geisslerand Sauer 1990). We estimatedthe mean and variance of regional trends from 400 random bootstrap subsamplesdrawn with replacementfrom the originalsample of trends (Geisslerand Sauer 1990, Thomas and Martin 1996). CBC countswere standardizedto reflect expectedabundances and trends for the average effort (175 party-hours)in the studyarea over all countcircles and years, usingthe proceduredescribed by Butcherand McCulloch(1990; survey efforthad no significanteffect on ravennumbers, P > 0.90). We modified calculationof the exponentrelating survey effort to numberof birdscounted byincluding terms in themodel (dummy variables) to accountfor unbalanced effects of CBC locations on the overall relation between effort and abun- dance.We did not adjustfurther for differencesin effort, which include annual differencesin distancecovered per party-hourand distributionof effort amonghabitat types of varyingsuitability (e.g., ranchland vs. open water), althoughwe consideredhabitat differencesamong count circles when interpretingthe results.We emphasizethat trendsand abundances based on CBC data should be evaluated with considerable caution and confirmedby additionalstudy because of theseand otherpotential sources of error,including problems with identification,counting, weather, access, and habitatcoverage (Butcher 1990). We furtherevaluated regional breeding distributions by compilingbreed- ing bird atlas (BBA) data within the studyarea. BBAs use standardized criteriato determinethe breedingstatus (possible, probable, confirmed) of eachbird species within 5-km 2 blocks (Robbins 1990) and are organized and developedseparately by county(Marin County: Shuford 1993; Sonoma County:Burridge 1995; NapaCounty: R. Leongand B. Grummer,unpubl.; Contra CostaCounty: S. Glover,unpubl.; Alameda County: B. Richmond andH. Cogswell,unpubl.; Santa Clara County: Santa Clara Atlas Comm., unpubl.;San Mateo County:R. Johnson,unpubl.; San FranciscoCounty: M. Eaton,unpubl.). Where atlas blocks spanned BBA (county)boundaries, we usedthe statuscode from the BBA that indicatedthe greatestlikelihood of breedingfor that location.

RESULTS

Differencesin corvidnumbers among survey routes, nested within land- usetype or subregion,were greaterthan expectedby the variabilitywithin routes(F -- 40.65; df -- 10, 110; P < 0.0001), indicatingthat variation amongsurvey routes accounted for significantregional variability in corvid densities(Table 1, Figure1). On average,89% + 2.3 (standarderror) of ravensand 92% + 1.8 of crowswere observedwithin 200 m of survey

206 COMMON RAVEN AND AMERICAN CROW IN THE SAN FRANCISCO BAY AREA

Table 1 Number of Common Ravens and American Crows Observed DuringRoad Surveysin the San FranciscoBay Area Duringthe Breeding Season of 1999

Birdsper kilometer Common Raven American Crow Route LandUse Subregion Length(km) N Mean (SE)a Mean (SE)

1 Rural Coast 19.4 12 0.584 (0.290) 0.014 (o.010) 2 Rural Coast 59.4 18 0.040 (0.005) 0.222 (0.030) 3 Rural Coast 33.0 7 0.241 (0.081) 0.322 (0.078) 4 Rural Coast 41.6 7 0.169 (0.024 o.o34 (0.014) 5 Rural Coast 50.8 7 0.128 (0.018 0.000 (0.000) 6 Rural Bay 41.0 7 0.038 (0.019 0.056 (0.025) 7 Rural Interior 50.7 2 0.089 (0.010 o.335 (0.177) 8 Rural Interior 79.8 6 0.029 (0.013 0.090 (0.023) 9 Rural Interior 65.2 6 0.070 (0.036 0.116 (0.039) 10 Rural Interior 40.9 4 0.006 (0.006 0.000 (0.000) 11 Rural Interior 51.0 8 0.007 (0.005 2.457 (0.567) 12 Rural Interior 41.8 3 0.016 (0.016) 0.318 (0.076) 13 Urban/Suburban Coast 39.6 7 0.375 (0.086) 0.007 (0.005) 14 Urban/Suburban Bay 46.2 7 0.012 (0.009) 0.444 (0.079) 15 Urban/Suburban Bay 50.4 6 0.205 (0.016) 0.251 (0.054) 16 Urban/Suburban Bay 41.8 6 0.134 (0.040) 0.017 (0.017) 17 Urban/Suburban Interior 63.1 7 0.238 (0.036) 0.253 (0.044) 18 Urban/Suburban Interior 65.4 5 0.003 (0.003) 0.281 (0.075) aSE, standarderror.

routes, with 69% + 4.4 of ravens and 70% + 7.1 of crows observedwithin 100 m. The proportionof crowsor ravensthat occurred in pairs,suggesting possiblebreeding status, did not differ significantly by land use, subregion, or route (P > 0.05).

American Crow In general,densities of AmericanCrows along rural and urban/suburban surveyroutes did not differsignificantly (F = 1.76; df = 1,110; P = 0.19). If route 11, an open agriculturalarea with unusuallyhigh numbersof crows (Table1) wasexcluded, however, crows were significantlymore abundant alongurban/suburban survey routes than alongthe remainingrural routes (F = 10.9; dr= 1, 103; P < 0.01; Figure3). When testedagainst random route effects (see Methods), the number of crows in rural and urban/ suburbanareas did not differsignificantly (F = 0.39; df = 1, 15; P = 0.54). Densitiesof crowsalong survey routes did vary significantly by subregion (F = 50.5; df = 2, 110; P < 0.0001), withnumbers increasing significantly fromthe outercoast to the bayshore to the interior(P < 0.01). If route11, an interior route with unusuallyhigh numbersof crows (Table 1) was excluded,densities in bayshoreand interior locationsdid not differ (P >

207 COMMON RAVEN AND AMERICAN CROW IN THE SAN FRANCISCO BAY AREA

0.2 ß CommonRaven [•] AmericanCrow

0.1

0.0 Rural Urban/Suburban Figure 3. Mean number of Common Ravensand AmericanCrows observedper kilometerof surveyroute by rural or urban/suburbanland usein the San Francisco Bay area. Valuesfor the AmericanCrow excluderoute 11 (see text). Error bars, standard errors.

0.99; Figure4). Despitesignificantly lower densitiesof crowsalong outer- coastroutes, differences by locationwere not significantlygreater than the estimatedrandom variationacross the region (randomeffects F -- 1.54; dr=2, 15; P -- 0.25). Land use and subregioninfluenced the numberof crowson each route independently(F = 0.13; df = 2, 112; P > 0.87).

Common Raven Densitiesof Common Ravens were significantlygreater along survey routes in urban/suburbanthan in rural areas (F -- 5.75; df -- 1, 107; P = 0.02; Figure3). In contrast,the highestnumbers of ravensoccurred in the pastoralzone of Point ReyesNational Seashore (route 1; Table 1, Figure1). Land-useeffects did not differsignificantly from estimatedrandom variation in the region(random route effects F = 0.70; df -- 1, 16; P -- 0.41). Ravendensities varied significantly by subregion(F -- 18.66; df = 2,107; P < 0.0001), with numbersincreasing significantly from interiorto bayshore to outer-coastsurvey locations (P < 0.05; Figure4). The test for random route effectsrevealed a significantcontrast between the outer-coastroutes and the bay shoreor interior(F = 5.12; df -- 1, 15; P -- 0.04). Therefore, higherconcentrations of ravenscharacterize the outercoast. Land use and subregioninfluenced the numberof ravenson eachroute independently (F = 0.003; df = 2, 108; P > 0.95).

Other SurveyData Numbersof ravensand crowsobserved along 15 BBS routesin the San FranciscoBay areareveal distribution patterns similar to thosesuggested by our roadsurvey, including dramatic abundance variation from route to route (Table 2). BBS data suggestsignificant overall annual increasesin both ravens(5.16% + 0.15; P < 0.001) and crows(3.21% + 0.13; P < 0.001).

208 COMMON RAVEN AND AMERICAN CROW IN THE SAN FRANCISCO BAY AREA

0.2 ß CommonRaven [•] AmericanCrow

(1) 13. 0.1

0.0 Outer Bay Interior Coast Shore

Figure4. Mean numberof CommonRavens and AmericanCrows observedper kilometerof surveyroute in outer coast, San FranciscoBay shore, and interior subregionsin the San FranciscoBay area. Valuesfor the AmericanCrow exclude route 11 (see text). Error bars, standarderrors.

BBS trendsvary greatlyacross the region,however, suggesting effects at smallerscales (Table 2). BBS data imply strongincreases in the numberof ravensalong the southernouter coast of the studyarea and in centralNapa County. They suggestsignificant increases in crowsnear Santa Rosa and alongpart of the outer Sonomacoast. Crows apparently increased along route 203 in the eastern interior of Contra Costa and Alameda counties from 1972 to 1991 butdeclined along nearby BBS route 303 from 1992 to 1999 (Table2, Figure2). Inspectionof scatterplotsfor all individualroutes, however,suggested that suchdifferences may dependon the numberof yearssurveyed, because linear slopes measured over periods of lessthan 8- 10 yearsmay contrastwith longer-termtrends. CBC data reveal significantoverall annual increasessince 1950 in numbersof both ravens(6.57% + 0.12; P < 0.001) and crows(3.31% + 0.05; P < 0.001), increasesthat acceleratedduring the 1980s and 1990s (ravens:7.23% + 0.16, P < 0.001; crows: 5.02% + 0.08, P < 0.001). Patternsof variationin abundanceby CBC circleare generallyconsistent with road-surveyand BBS resultsand suggestsubstantial local variation in numbersof both ravensand crows(Tables 3 and 4). In neitherBBS or CBC resultsdid we find significantcorrelations, positive or negative,between the two species'abundances or trends(P < 0.05), suggestingtheir distributions are independent.In contrast,our roadsurveys implya significantinverse relationship between raven and crowabundances (r -- -0.55, P < 0.05), suggestingtheir distributions are complementary.For example, raven and crow numbersvaried dramaticallyamong urbanized routes,with high numbersof ravensin San Franciscobut relativelyfew crows,and surprisinglylow numbersof ravensalong the urbanizedcorridor

209 COMMON RAVEN AND AMERICAN CROW IN THE SAN FRANCISCO BAY AREA

Talkie 2 BreedingBird SurveyTrends and Mean Number of Birdsper Survey Route for the Common Raven and American Crow in the San FranciscoBay Area

American Crow Common Raven

BBS Years Trend Mean Trend Mean birds routea Period surveyed(% per year) b perroute (%per year) b perroute

14 1968-2000 31 -3.27 0.55 2.33** 20.48 15 1969-2000 27 -11.12 2.52 10.00 1.81 16 1968-2000 18 6.80 6.39 19.92 1.17 71 1971-2000 24 1.82 20.75 6.45 7.83 83 1975-2000 25 1.34 17.18 0.36 7.65 172 1972-2000 19 -10.92 4.05 0.00 0.05 186 1972-2000 28 9.07** 52.50 0.32 1.86 187 1972-1988 14 12.30'* 24.78 0.00 0.21 189 1972-2000 20 3.04 25.30 9.42 0.45 193 1972-2000 25 19.89' 1.12 -5.40** 8.88 194 1972-1995 23 -1.40 0.07 18.10' 0.52 202 1972-1997 26 0.12 29.54 10.93'* 4.81 203 1972-1991 20 13.07'* 13.30 -6.83 0.15 303 1992-1999 7 -8.13'* 124.43 6.75 6.43 319 1992-1996 5 0.00 0.00 37.02* 18.20 aSeeFigure 2 for location. bLineartrend significant at *P < 0.05, **P < 0.01. of easternMarin Countywhere crowswere relativelyabundant (routes 13 and 14, Table 1). Compositebreeding bird atlases for the regionreflect distributions that are generallyconsistent with the road surveysand BBS (Figures5 and 6). We emphasizethat atlas data for the bay area representa wide range of samplingperiods dating back as far as the late 1970s andshould therefore be interpretedwith considerable caution and careful comparisons with other availabledata. BBAs indicateCommon Raven concentrationsalong the outercoast, particularly in the PointReyes area and along the San Francisco and San Mateo coast,in urbanizedareas along the southernSan Francisco Bay shoreline,in somehigher elevations of easternContra Costa, Alameda, and Santa Clara counties,and in parts of Napa county.Such "concentra- tions"are basedon a greaterfrequency of 5-km2 blockswith breeding ravens,rather than greaterraven abundances per se. However,broad areas with breedingconfirmed over many adjacentblocks are likelyto support more ravensthan areaswhere breeding is confirmedin only a few blocks. Breedingravens appear to be relativelyscarce in large areas of central Contra Costa, Alameda, and Santa Clara countiesand in relativelylow numbersin centralSonoma County. BBAs reveal that the AmericanCrow's breedingdistribution is relativelycontinuous in mostbayshore and interior areas(although habitat is not necessarilysaturated) but sparseor brokenin manyparts of the outercoastal drainage (Figure 6). In additionto resembling

210 COMMON RAVEN AND AMERICAN CROW IN THE SAN FRANCISCO BAY AREA

ß ...... ß ß .* .0.*© ß OO Napa ß e.O. ß ß ß OOe..e 989-1993 0. o..o, ß oil

1986-1992Sonoma,• ß ß ßß ß ßß•1ß * ßß:' ß

ß Solano (no data)

Marin ,. * ' 1976-1982 ß I I I I ßßß ß .ß .ß ß ß ContraCosta ß ß =* , ..< ß ß ß

San Francisco 1991-1993 ß ß ßß ßI . Alameda ß ß ß ß ß 1993-1997 Pacific ß ' ß ß Ocean ß ß ß ß ß ß ß ßß • • SantaClara San1988-1997Mateo• ß ßß:i• '1993 0•=North •0•= ß :: ':/ Figure5. Breedingdistribution of theCommon Raven in the SanFrancisco Bay area, basedon breedingbird at]as data. Filled circles indicate likelihood of breedingin each 5-kinblock: large circle, confirmed; medium circle, probable; small circle, possible; no circle,no ravensobserved. Field data were collected in the yearsindicated under each COUntyname. patternsemerging from the road surveyand BBS (Tables1 and 2), BBA patternsare generallyconsistent with winter distributions indicated by CBC results.CBCs, however, suggest lower concentrations of ravensalong the eastside of southernSan FranciscoBay andhigher concentrations of crows in westernSonoma County (Tables 3 and 4).

DISCUSSION

Increasesin the numbersof ravensand crowsin the San FranciscoBay areaare consistentwith increases throughout the westernUnited States, but in the bay area raven distributiondiffers from patternsevident at larger scaleswith regard to urbanization(Marzluff et al. 1994).BBS data imply that Common Raven densitiesin the western United States are negatively

211 COMMON RAVEN AND AMERICAN CROW IN THE SAN FRANCISCO BAY AREA

ß ß ß ß ß ß. ß ß . .., Napa ß 0000 ß ß ß ß ß • 1989-1993 ß .0. ß ß O. ß ß 0-

Sonoma ß -. eOOOe. ß 1986-1992 ß eelell. ß * ß ß ß ß Solano ß ß (nodata)

Madn ,• ß ß ß ' ß 1976-1982 lee .eee ContraCosta ß .Joel ß oil, oeec .... eee• 1998-2000

San Francisco

1991-1993 Alameda 1993-1997 Pacific Ocean

San Mateo 1988-1997

North

ß Figure6. Breedingdistribution of theAmerican Crow in the SanFrancisco Bay area, basedon breedingbird arias data. Filled circles indicate likelihood of breedingin each 5-kinblock: large circle, confirmed; medium circle, probable; small circle, possible; no circle,no ravensobserved. Field data were collected in theyears indicated under each county name. associatedwith humandensities (although less strongly in coastalstates), with ravensapparently invading agricultural areas to a greaterextent than urbanareas (Marzluff et al. 1994). In contrast,we foundgreater raven densitiesalong urbanized survey routes, on average,than along rural survey routes,a pattern similarto that in the MojaveDesert (Knight et al. 1993). However,both our highestand lowestraven densities were alongrural routes,the highestin the agriculturalzone of PointReyes National Seashore (dominatedby dairyranches), and the lowestin the ruralinterior of Alameda and Napa counties. Common Ravensoccurred in significantlygreater densities along the outercoast than in the interiorof the regionor nearthe bay (althoughthe east bay shorelinewas not includedin our road survey).In suchcoastal habitats,ravens feed on livestockcarcasses, grain in cattlefeedlots, and a

212 COMMON RAVEN AND AMERICAN CROW IN THE SAN FRANCISCO BAY AREA

Table 3 ChristmasBird Count Trendsand Mean Number of Birds per Count Circle for the American Crow in the San FranciscoBay Area

1950-1998 1980-1998 Trend(% Meanbirds Trend(% Meanbirds CountCircle a Years peryear) b percircle Years peryear) b percircle

Angwin(AG) 39 3.66'** 516.10 18 -0.69 681.75 AfioNuevo (AN) 26 3.01 2.50 18 -3.34 3.36 Arroyo CheapThrills (CT) 10 2.96 1543.27 8 1.65 1604.26 Benicia(BE) 37 17.77'** 50.86 19 15.24'** 96.46 ContraCosta (CC) 43 9.06*** 46.90 19 8.73** 69.46 CrystalSprings Reservoir(CS) 41 6.33*** 16.36 18 23.92*** 32.99 Hayward-Fremont(HF) 29 1.45 72.99 17 10.90'** 62.85 MountHamilton (MH) 21 1.24 101.75 19 4.25 97.27 Oakland(OA) 45 4.01'** 43.96 18 5.31' 56.73 PaloAlto (PA) 39 14.81'** 22.86 19 13.24'** 43.39 PointReyes (PR) 27 1.56' 605.44 19 -0.15 653.41 PutahCreek (PC) 26 10.52'** 316.29 19 16.28'** 382.64 SanFrancisco (SF) 29 10.51'** 36.97 15 23.48'** 70.28 SanJose (SJ) 47 0.18 180.47 19 9.61'** 148.48 SantaRosa (SR) 36 0.49 1547.83 19 4.70'** 1581.53 Southern MarinCounty (SM) 23 7.56*** 494.32 19 6.39'** 556.13 TomalesBay (TB) 13 2.63 365.71 0 Western Sonoma County(WS) 31 4.03*** 241.11 19 2.19 284.10

øSeeFigure 2 for location. •Lineartrend significant at *P < 0.05, **P < 0.01, ***P< 0.001. varietyof otherfoods found in grasslands,at publicpicnic areas, and along dunesand beaches(Roth et al. 1999). Ravensalso exploit garbage dumps (Marzluffet al. 1998), which are often locatedin rural areas. However, our roadsurvey indicated that high ravennumbers in coastalagricultural areas arematched by similarlyhigh numbers in the mosturbanized habitats around San FranciscoBay. Althoughour road survey did not includeurban areas in the eastbay, BBA data showthat breedingravens and crowsare establishedthere, and CBC data suggestsignificant increases in both species.Over the past20 years, ravens have established themselves and increased their numbers near salt pondsalong the southerneast bay shoreline(H. Cogswellpers. comm.). The numberof AmericanCrows in Californiaincreased dramatically with the spreadof agriculturein the early 1900s (Emlen 1940). BBS data, however, show that crow abundancesin the western United States are negativelyassociated with the extentof farmlandand positivelycorrelated with humanpopulation density (Marzluff et al. 1994). Consistentwith this pattern, our resultsrevealed that crowsconcentrate in urbanizedhabitats

213 COMMON RAVEN AND AMERICAN CROW IN THE SAN FRANCISCO BAY AREA

'ral•!e 4 ChristmasBird Count Trendsand Mean Numberof Birdsper Count Circlefor the CommonRaven in the San FranciscoBay Area

1950-1998 1980-1998 Trend(% Meanbirds Trend(% Meanbirds CountCircle a Years peryear) b percircle Years peryear) • percircle

Angwin(AG) 40 9.18'** 23.53 19 5.53* 35.84 Afio Nuevo(AN) 26 19.39'** 30.62 18 14.57'** 43.50 Arroyo CheapThrills (CT) 10 -1.00 60.40 8 0.84 57.75 Benicia(BE) 37 8.02*** 4.65 19 9.72** 8.32 ContraCosta (CC) 43 5.70*** 2.14 19 7.47** 4.21 CrystalSprings Reservoir(CS) 42 12.86'** 27.95 18 25.16'** 62.22 Hayward-Fremont(HF) 30 3.93* 2.07 18 12.87'* 2.89 MountHamilton (MH) 21 9.13'** 24.57 19 9.21'** 26.32 Oakland(OA) 46 8.99*** 8.59 18 17.86'** 21.17 PaloAlto (PA) 39 15.66'** 23.95 19 12.09'** 47.95 PointReyes (PR) 28 0.87 347.57 19 0.58 365.32 PutahCreek (PC) 27 11.76'** 9.89 19 18.13'** 13.37 SanFrancisco (SF) 29 7.21'** 40.69 15 18.32'** 73.07 San Jose(SJ) 47 5.69*** 7.09 19 25.07*** 14.11 SantaRosa (SR) 35 5.18'** 22.06 19 6.73*** 28.00 Southern MarinCounty (SM) 23 8.29*** 125.87 19 8.98*** 140.63 TomalesBay (TB) 13 22.39*** 170.31 0 Western Sonoma County(WS) 28 2.17'* 208.75 16 -0.75 230.38 aSeeFigure 2 for location. bLineartrend significant at *P < 0.05, **P < 0.01, ***P< 0.001. overmost of the San FranciscoBay areaand occur in relativelylow numbers in many rural areas. Yet there is a dramaticexception to this pattern in farmlandnortheast of SuisunBay (Route11), wherewe foundthe highest regionaldensities of crows;the reasonsfor thesehigh densities are unknown. Breedingravens and crowsgenerally move lessthan 3-7 km per day (Engeland Young 1992, Linz et al. 1992, Sullivanand Dinsmore 1992, Caccamiseet al. 1997, Roth et al. 1999). Therefore,the 50-km routesused in our roadsurvey were probablyof a lengthsuitable to capturedifferences in localabundance near the appropriatescale of landscapeuse. Although we did not measuredetection probabilities associated with distancefrom the road, ravensand crowsare relativelyconspicuous birds; their dramatically higherdensities within 100 rn of surveyroads suggest habits of foragingon highway-generatedcarrion (Knight and Kawashima1993). The effectsof landscapeproperties on populationgrowth carmotbe determinedfrom our count data because differences in birddensity might not be relatedto productivityor survivorship.Studies based on individuallymarked birdsare neededto determineif behavioraladjustments to urbanizationor

214 COMMON RAVEN AND AMERICAN CROW IN THE SAN FRANCISCO BAY AREA

anthropogenicfood sourceslead to changesin crow or ravenpopulation growthor distribution(Knight et al. 1987, Marzluffet al. 1998). Of particular concernis whether increasesin numbersof ravensand crowsmight signal increasesin nestpredation of othernative species, including waterfowl (Stiehl and Trautwein1991), the MarbledMurrelet (Brachyrarnphus rnarrnoratus; Singeret al. 1991, Nelsonand Hammer 1995), SnowyPlover (Charadrius alexandrinus; U.S. and Wildlife Service 1993), Least Tern (Sterna antillarurn; Fancher 1992, Avery 1995), Common Murre (Uria aalge; Thayeret al. 1999, Rothet al. 1999), heronsand egrets (Ardeidae; Parsons 1995, Kellyand Fischer 2000), andforest-nesting (Andren 1992, Bulerand Hamilton2000). However,corvid numbers may correlatepoorly with overallnest- rates, especially if changesin corvidabundance influencerates of nest predationby other species(Clark et al. 1995). Additionalstudies are neededto testfor the independenteffects of increasing corviddensities on the populationdynamics of otherspecies. With the exceptionof a greaterconcentration of ravensalong the outer coast,variation in abundanceof crowsand ravensamong coastal, bayshore, and interiorsubregions did not seemto exceedrandom differences in the region.Similarly, variation in corvidnumbers and trends suggests substantial variation with local conditions.Also, the inverse relationship between abundancesof ravensand crowsdid not correspondto differencesin rural vs. urbanizedhabitat and was not significantin other availabledata. Thus, factorsaffecting differences in speciesdistribution and habitatuse may be complex,with local-scalehabitat variation influencing the generalregional pattern. Further studyis neededto determinethe extent to which local conditionsreflect differences in habitatsuitability for foragingor breeding,or alternatively,differences in habitatsaturation and associatedcapacity for furtherpopulation growth.

ACKNOWLEDGMENTS

We thank the followingobservers for conductingroad surveys:Sarah Allen, MargaretBarson, Charles Benedict, Noelle Bon, SusanBuckner, Lynn Campbell,Ted and Zoe Chandik,Rigdon Currie, Nancy DeStefanis, Katie Etienne,Binny Fischer, LeslieFlint, Graeme Fraser, Harry andKate Fuller, Tom Gardali,Keith and Kaire Gish, Pat and Phil Gordon, Jim Hench, Conrad Jones, Gail Kabat, Barbara Lapp, Flora Maclise,Jo Maillard,Dan and Joan Murphy, RobertNorthrop, Linda Reichel,Rudi Richardson,Jared Roth, EllenSabine, Don and MarilynSanders, Danielle Saunders, RobinSmith, Sue Spoffard, and Paula Terry. We thankKatie Fehring, Elizabeth Lynn, Mark McCaustland,and Sue Guers for organizingand compilingthe data. Lynn Campbelland Graeme Fraser assisted in the developmentof the fieldprotocol. Doug Oman providedadvice on the statisticalanalysis. Howard L. Cogswell,Mark K. Sogge,and an anonymousreviewer provided helpful comments on the manuscript. This projectreceived financial support from the Marin CommunityFoundation and the Marin and Sequoiachapters of the NationalAudubon Society.

LITERATURE CITED

Andren,H. 1992. Corviddensity and nestpredation in relationto forestfragmenta- tion: A landscapeperspective. Ecology 73:794-804.

215 COMMON RAVEN AND AMERICAN CROW IN THE SAN FRANCISCO BAY AREA

Avery,M. L., Pavelka,M. A., Bergman,D. L., Decker,D. G., Knitfie,C. E., and Linz, G. M.. I995. Aversiveconditioning to reduceraven predationon California LeastTern . ColonialWaterbirds 18:131-138. Bennington,C. C., andThayne, W. V. 1994. Useand misuse of mixedmodel analysis of variancein ecologicalstudies. Ecology 75:717-722. Boarman, W. I., and Heinrich, B. 1999. Common Raven (Corvus corax), in The Birds of North America (A. Poole and E Gill eds.), no. 476. Birds N. Am., Philadelphia. Buler,J. J., and Hamilton, R. B. 2000. Predationof naturaland artificialnests in a southernpine forest.Auk 117:739-747. Burridge,B. 1995. SonomaCounty Breeding Bird Atlas.Madrone Audubon Soc., P.O. Box 1911, Santa Rosa, CA 95402. Butcher, G. S. 1990. Audubon ChristmasBird Counts, in Survey Designsand StatisticalMethods for the Estimationof Avian PopulationTrends (J. R. Sauer and S. Droege,eds.), pp. 5-13. U.S. Fishand WildlifeServ. Biol. Rep. 90. Butcher,G. S., and McCullochC. E. 1990. Influenceof observereffort on the number of individualbirds recorded on ChristmasBird Counts,in SurveyDesigns and StatisticalMethods for the Estimationof Avian PopulationTrends (J. R. Sauer and S. Droege,eds.), pp. 120-129. U.S. Fishand Wildlife Serv. Biol. Rep. 90. Butcher,G. S., Fuller,M. R., McAllister,L. S., and Geissler,P. H. 1990. An evaluation of the ChristmasBird Count for monitoringof populationtrends of selected species.Wildlife Soc. Bull. 18:129-134. Caccamise,D. F., Reed,L. M., Romanowski,J., and Stouffer,P. C. 1997. Roosting behaviorand groupterritoriality in AmericanCrows. Auk 114:628-637. Clark, R. G., Meger,D. E., andIgnatiuk, J. B. 1995. RemovingAmerican Crows and ducknesting success. Can. J. Zool. 73:518-522. Coston, C. 1998. ChangingBay Area bird populations:The corvids.The Stilt, newsletterof the San FranciscoBay Bird Observatory17:4-5. Emlen, J. T. 1940. The midwinter distributionof the crow in California. Condor 42:287-294. Engel,P. R., andYoung, L. S. 1992. Movementsand habitat use by CommonRavens from roostsites in southwesternIdaho. J. WildlifeMgmt. 56:596-602. Fancher,J. M. 1992. Populationstatus and trends of the CaliforniaLeast Tern. Trans. W. Sect. Wildlife Soc. 28:59-66. Geissler,P. H., and Sauer,J. R. 1990. Topicsin route-regressionanalysis, in Survey Designsand Statistical Methods for the Estimationof AvianPopulation Trends (J. R. Sauerand S. Droege,eds.), pp. 54-57. U.S. Fishand WildlifeServ. Biol. Rep. 90. Kelly,J.P., and Fischer,B. 2000. 1999 heronand egretmonitoring results at West Marin Island.ACR Tech. Rep. 90-3-10 (orderfrom AudubonCanyon Ranch, 4900 ShorelineHwy., StinsonBeach, CA 94970). Knight,R. L., Grout,D. J., andTemple, S. A. 1987. Nest-defensebehavior of the American Crow in urban and rural areas. Condor 89:175-177. Knight,R. L., andKawashima, J. Y. 1993. Responsesof ravenand Red-tailed Hawk populationsto linearright-of-ways. J. WildlifeMgmt. 57:266-271. Knight,R. L., Knight,H. A. L., andCamp, R. J. 1993. Ravenpopulations and land- usepatterns in the MojaveDesert, California. Wildlife Soc. Bull. 21:469-471. Linz, G. M., Knittle, C. E., and Johnson,R. E. 1992. Home range of breeding Common Ravens in coastal southern California. Southw. Nat. 37:199-202.

216 COMMON RAVEN AND AMERICAN CROW IN THE SAN FRANCISCO BAY AREA

Marzluff,J. M., Boon, R. B., and Cox, G. W. 1994. Historicalchanges in populations and perceptionsof nativepest and bird speciesin the West. StudiesAvian Biol. 15:202-220. Marzluff, J. M., Gehlbach,F. R., and Manuwal, D. A. 1998. Urban environments: Influenceson avifaunaand challengesfor the avian conservationist,In Avian Conservation:Research and Management(J. M. Marzluff and R. Sallabanks, eds.),pp. 283-299. IslandPress, Washington, D.C. Marzluff,J. M., and Restani,M. 1999. The effectsof forestfragmentation on avian nestpredation, in ForestFragmentation: Wildlife and ManagementImplications (J. A. Rochelle,L. A. Lehmann, and J. Wisniewski,eds.), pp. 155-169. Brill, Leiden, Netherlands. Nelson, S. K., and Hamer, T. 1995. Nest successand the effectsof predationon MarbledMurrelets, in The ecologyand conservationof the MarbledMurrelet in NorthAmerica: An interagencyscientific evaluation (C. J. Ralph,G. L. Hunt, M. G. Rafael,and J. F. Piatt,eds.), pp. 89-97. U.S. Dept. Agric.Forest Serv. Gen. Tech. Rep. PSW-GTR-152. Parsons,K. C. 1995. Heron nestingat Pea PatchIsland, Upper DelawareBay, USA: Abundanceand reproductivesuccess. Colonial Waterbirds 18:69-78. Robbins.C. S. 1990. Use of breedingbird atlasesto monitorpopulation change, in SurveyDesigns and StatisticalMethods for the Estimationof Avian Population Trends(J. R. Sauerand S. Droege,eds.), pp. 18-22. U.S. Fishand Wildlife Serv. Biol. Rep. 90. Roth, J. E., Kelly, J.P., Sydeman,W. J., Parker,M. W., and Allen, S. G. 1999. Ecosystem-levelmanagement of CommonRavens on the Point ReyesNational Seashore.Report to Point ReyesNational Seashore, Point ReyesStation, CA 94956. Sauer,J. R., Hines,J. E., and Fallon,J. 2001. The North AmericanBreeding Bird Survey, Resultsand Analysis1966-2000. Version 2001.2. U.S. Geol. Surv. PatuxentWildlife Res. Center, Laurel, MD. Shuford,D. W. 1993. The MarinCounty Breeding Bird Atlas. Bushtit Books, Bolinas, CA. Singer,W. W., Naslund,N. L., Singer,S. A., and Ralph, C. J. 1991. Discoveryand observations of two tree nests of the Marbled Murrelet. Condor 93:330-339. Small,A. 1994. CaliforniaBirds: Their Statusand Distribution.Ibis Publ., Vista, CA. Stiehl, R. B., and Trautwein,S. N. 1991. Variationsin dietsof nestingCommon Ravens. Wilson Bull. 103:83-92. Sullivan,B. D., and Dinsmore,J. J. 1992. Home range and foraginghabitat of American crows, Corvus brachyrhynchos,in a waterfowl breedingarea in Manitoba. Can. Field-Nat. 106:181-184. Thayer, J. A., Sydeman,J., Fairman, N. P., and Allen, S. G. 1999. Attendanceand effectsof disturbanceon coastalCommon Murre coloniesat Point Reyes, California. Waterbirds 22:130-139. Thomas,L., and Martin, K. 1996. The importanceof analysismethod for breeding bird surveypopulation trend estimates.Cons. Biol. 10:479-490. U.S. Fishand WildlifeService. 1993. Finalrule: Western Snowy Plover. Fed. Reg. 58:12865-12874.

Accepted 12 July 2002

217