j RaptorRes. 38(1):26-34 ¸ 2004 The Raptor Research Foundation, Inc.

MOVEMENTS AND SURVIVAL OF FLEDGLING COOPER'S HAWKS IN AN URBAN ENVIRONMENT

R. WILLIAMMANNAN, 1 WENDY A. ESTES,AND WILLIAM J. MATTER Schoolof RenewableNatural Resources,University of ,7•tcson, AZ 85721 U.S.A.

ABSTP,ACT.--Cooper's Hawks (Accipitercooperii) nest in urban and suburban areasacross , but little is known about movements,habitat use, or survivalof fledglingsin thesesettings. We followed 40 radio-tagged,fledgling Cooper'sHawks hatched in Tucson,Arizona in 1999 or 2000, for up to 6 mo to estimatesurvival, and describepatterns of movementand the environmentsthey use while dispersing. The typical pattern of movementfor hawkswe tracked through early winter consistedof sedentary behavior in the natal m'ea,followed by relativelylong movementsbeginning 11-13 wk after hatching, and finally sedentarybehavior again when they settledinto a fall/winter home range. Distancesbetween relocationsof individual hawkswere, on average,greater for females (/= 6.8 km, range = 0.02-51.7 km, SD = 9.8) than males (œ= 3.8 km, range = 0.05-20.8 km, SD = 5.4; t-test,P = 0.02). Home range size for nine hawksduring their first fall/winter averaged771 ha (SD = 403). Distancefrom center of home range to natal site averagednem'ly twice as far for females (œ= 10.9 km, range = 4.2-19.5 km, SD = 6.4) as males (œ= 6.0 km, range = 2.2-13.3, SD = 5.0), but the difference was not significant (t-test,P = 0.23). Survivalof radio-taggedhawks was 67% through 180 d. Hawks used a vm'ietyof environments prior to settling for the winter, but were found most frequently (35% of locations) in riparian areas.We found no discernablepattern of habitat selectionfor land usecategories inside winter home ranges.We speculatethat the abundanceof food may facilitatesurvival of post-fiedging,dispersing hawks in Tucson.

KEYWORDS: Cooper'sHawks; Accipter cooperii; dispersal;habitat selection; home range size;, urban environ- ments; Tucson, Arizona.

MOVIMIENTOS Y SUPERVIVENCIA DE POLLUELOS DEL GAVILAN DE COOPER EN UN AMBIEN- TE URBANO

RESUMEN.--E1gavil•tn de Cooper (Accipitercooperii) anida en fireasurbanas y suburbanasa lo largo de Norteam6rica,y se conoce poco acercade susmovimientos, uso de h•tbitat,o sobrela supervivenciade losjuveniles en estaslocalidades. Seguimos 40 gavilanesde Cooper juveniles equipadoscon radios de telemetria empollados en Tucson, Arizona en 1999 o 2000, pot cerca de 6 mesespara estimar su supervivencia,y describir los patrones de movimiento y los ambientesque ellos usan durante su dis- persi6n. E1 patr6n tipico de movimiento pm'a los gavilanesque seguimosa principios del invierno consisti6de un comportamiento sedentarioen el firea natal, seguidopot movimientosrelativamente lm'gosque comienzanen la semana 11-13 luego de romper el cascm'on,y finalmente un comporta- miento sedentm'iode nuevo cuandose establecendentro de su rango de acci6nde la temporadaotofio/ invierno. Las distanciasentre las reubicacionesindividuales de los gavilanesfueron, en promedio, mils grandespara las hembras (/= 6814 m, rango = 16-51 673 m, SD = 9752) que para los machos (i = 3776 m, rango = 46-20 759 m, SD = 5356) (test de T, P = 0.02.) E1 tamafiodel rango de acci6npara nuevegavilanes durante su primer otofio/invierno promedi6 771 ha (SD = 403.) La distanciadel centro del rango de acci6n al sitio natal fue en promedio cerca de dos vecesmils lejana para las hembras (œ = 10.9 km, rango = 4.2-19.5 km, SD = 0.387) que para los machos(/ = 6.0 km, rango = 2.2-13.3, SD = 5.0), pero la diferencia no fue significativa(test de T, P = 0.23). La supervivenciade los gavilanes monitoreadosrite del 67% pm'a 180 dias. Los gavilanesusm'on una variedad de ambientesantes de establecersepara el invierno, sin embargose encontraronmas frecuentemente (35% de lasubicaciones) en zonasriparias. No encontramosningfin patr6n discernible en la selecci6ndel h•tbitatpara las cate- gorias del uso del suelo dentro de los rangosde acci6n invernales.Especulamos que la abundanciade alimento puede facilitar la supervivenciade los gavilanesjuveniles que se dispersanen Tucson. [Traducci6n de G6sarM•trquez]

E-mail address:[email protected]. edu

26 MARCH 2004 POST-FLEDGLING COOPER'S HAWKS 27

Natal dispersalin is the movement of fledg- nests in their first fall and winter (Boal 1997), but lings away from their nests and often involves a most reported movements during natal dispersal search by dispersingindividuals for a place to re- are considerablyless than this distance (e.g., me- side and potentially breed. Distancestraveled by dian = 6.4 km for males, N= 10; 14.4 and 790 dispersingbirds and their survivalcan significantly km for two females; Rosenfield and Bielefbldt affect the genetic structure, demography, and via- 1992, Rosenfieldet al. 1996). Thus, the initial stag- bility of populations (e.g., Pulliam and Dan- es of natal dispersalby Cooper's Hawks hatched •n ielson 1991, Payne and Payne 1993, Globert et al. relatively large metropolitan areas could occur 2001). Natal dispersal usually is measured as the within these developed environments. straight-line distance between natal nestsand sites Knowledge of the patterns of movement and where birds first breed, if they survive (Howard habitat use of dispersingCooper's Hawks in urban 1960, Greenwoodand Harvey 1982). But dispersal settingscould be used to identify environments to generally takes place in landscapescomposed of protect or enhance as cities grow, assumingthat patches that vary in size and quality relative to the hawk populations are a desired feature of the ur- needs of dispersing individuals; thus, movements ban landscape. Furthermore, Boal and Mannan between natal and breeding sites are not likely to (1999) suggested that models of population be linear (Wiens 2001). Furthermore, selection of growth of urban Cooper's Hawks are needed to a place to reside by dispersingbirds may be con- understand whether some urban areas represent ducted in stages(e.g., search, settlement, residen- "source" or "sink" populations (Pulliam 1988). cy; Stamps2001), each associatedwith different be- Estimatesof survival of hawks during dispersalare haviors and patterns of movement. Identifying critical to such modeling efforts (e.g., Lande 1988, patterns of movement in all stagesof dispersalis Boyce1992, Beissingerand Westphal1998). We fol- critical to understandingwhy dispersingbirds set- lowed Cooper's Hawks hatched in an urban setting tle and breed where they do. Also, information for up to 6 mo after fiedging to: (1) estimate their about environments that facilitate movement and survival from late summer through early winter; survivalduring dispersal(e.g., Miller et al. 1997) and (2) determine how far they dispersefrom na- are important in the developmentof habitat man- tal sites,the kinds of environmentsthey use while agement plans for birds, especiallyif plans encom- dispersing,and the characteristicsof areasthey use passbroad spatial scales(e.g., Strong and Bancroft when they settle during their first winter. 1994, Miller et al. 1997) and fragmented land- STUDY AREA scapes(e.g., Temple 1989). Natal dispersalhas been studied in a variety of We marked and tracked fledgling Cooper's Hawks m bird speciesover the last decade, but information and near Tucson, Arizona (32 N ø, 111 Wø). The Tucson metropolitan area encompassesabout 70 000 ha with an about this process remains limited compared to estimated human population of 803 600 residents. Tuc- other aspectsof population demographyand other son includes developmentsranging from commercial dis- kinds of movements (Koenig et al. 2000, Walters tricts and high-density housing to suburban areas w•th 2000, Clobert et al. 2001). Among birds of prey, low-densityhousing. Parks, golf courses,and open space various stagesof natal dispersalhave been studied are scattered throughout residential areas. Tucson is lo- cated in the Sonoran Desert and supports remnants of primarily in speciesthat are the focus of manage- lower and upper Sonoran vegetation types and riparian ment and conservation efforts and in environ- corridors (Brown et al. 1979), but much of the natural ments that are relatively undeveloped or rural vegetation has been removed or replaced with nonnative (e.g., Wyllie 1985, Walls and Kenward 1995, Ganey plants. et al. 1998, Harmata et al. 1999, Restani and Mat- METHODS tox 2000). Cooper's Hawks (Accipitercooperii) nest in highly-developedenvironments (i.e., urban and We used bal-chatri traps (Bloom 1987) to capture fledgling Cooper's Hawks at nests monitored in a long- suburban areas) in several places across North term study (Boal and Mannan 1999, Mannan and Boal America (see Rosenfield and Bielefeldt 1993 for 2000). We captured fledglings when they were old review, Stewart et al. 1996, Boal and Mannan 1998, enough to hunt on their own and when their rectrices 1999), but little is known about the movements, were fully emerged (-->55 d old). We marked each cap- tured fledgling with a Department of Interior leg band habitat use, or survival of fledglings in these set- and a colored leg band with a unique alpha code, and tings. Existing information suggeststhat Cooper's attached a radiotransmitter (model RI-2C [5 g] in 1999 Hawks may move up to 100 km from their natal and model PD-2 [3.5 g] in 2000; Holohil SystemsLtd, 28 MANNAN ET AL. VOL. 38, NO. 1

Carp, Ontario, Canada) to a central rectrix (Samuel and dences/ha]; high-density residential areas [-->7.4 resi- Fuller 1994). Estimated life of transmitters was 6 mo dences/ha]; open spacewith low human use [e.g., cem- (model PD-2) or 9 mo (model RI-2C). No more than two eteries, neighborhood parks, and natural open space], fledglings (usually a male and female) from any nest open spacewith high human use [e.g., golf courses,dis- were radio-tagged,and nestsfrom which fledglingswere trict and regional parks, and schools],and other [e.g, marked were all >5 km from the edge of the metropol- roadways,and commercial,industrial, and agricultural lean area. eas]). If hawks establishedhome ranges during fall and We relocated radio-taggedhawks, while they remained winter, we used the convex polygon method to generate in their nest areas, at least once per week by "homing" area-observation curves (Odum and Kuenzler 1955) for (White and CarrotE 1990) with Telonics TR-2 receivers each home range to assesswhether our sample of loca- and RA-14 flexible, two-element, yagi antennas (Telonics- tions adequately described home range size for the pe- Electronics Consultants, Mesa, AZ). After hawks left their riod of interest.We then estimatedhome range sizeswith nest areas, we attempted to relocate them by scanning the kernal method (90% isopleth; Worton 1989). We for their radio signalsfrom 40 elevated positions (e.g., chose the 90% isopleth becauseit seemedto minimize hillsides, tops of buildings) once or twice per week. Ele- fragmentation of home ranges,while excludingareas not vated positions were scattered throughout most of the usedby hawks.We applied a smoothingtechnique to the Tucson metropolitan area so that any radio-taggedhawk boundaries of home rangesbased on least squarescross present likely would be detected. If a signalwas detected validation (h). We first calculated h for the set of locations from an elevated position, we estimated its general loca- for each hawk. We then estimated size of home ranges, tion based on signal strength, and then attempted to lo- and applied the averageh to all home ranges (Kenward cate the hawk by "homing" in an automobile and on 2001). We used only locationsof hawksidentified by sight foot. In areaswithout elevatedpositions, we drove along or triangulation, and those estimated to be within 30 m parallel roads throughout the area and scannedfor radio of the observer,to calculatehome ranges.We used pro- s•gnals.If a hawk setfled into a relativelysmall area where gram RANGES V (Kenward and Hoddler 1996) to gen- it could be found with consistency(i.e., establisheda erate area-observation curves, and Arcview Version 3.2 home range), we attempted to relocate it up to five times (Environmental SystemsResearch Institute 1996) with per week at various times of the day by homing. We al- the extension " Movements" (Hooge and Eichen- lowed at least 12 hr between relocations for any given laub 1997) to measure distances between locations and hawk to minimize the risk of dependency among loca- estimate size of home ranges. tions. We described habitat use by dispersing hawks before Once or twice per month, we flew in a single engine they settled into winter home ranges by calculating the aircraft in parallel transectsover the Tucson metropoli- percent of locations (all hawkscombined) in each land tan area and scannedfor radio signals.We also scanned use category. We could not evaluate habitat selection for radio-taggedhawks while the aircraft wasflown along (i.e., comparing use versusavailability) during this period major watercoursesup to 100 km from Tucson.If a signal becauseour sample of locationsper hawk wassmall and from a radio-taggedhawk was detected from the aircraft, we did not have a databaseidentifying land use on the we recorded the general area of the detection, and re- entire area used by dispersinghawks. turned to the area in an automobile to search for the We evaluated habitat selection inside winter home hawk. ranges by first overlayingoutlines of the home rangeson Cooper's Hawks in Tucson are acclimated to the pres- the digital database of land use categoriesin Tucson ence of humans, and single observers can approach (Shaw et al. 1996). We calculated the coverage of each hawkswithout eliciting flight (Mannan and Boal 2000). category in each home range, compared the coverages Therefore, we attempted to see the hawks at each relo- to land use patterns shownon aerial photographstaken cation. Sometimes hawks were out of sight (e.g., in a in 2001 to assessaccuracy, and made corrections in two fenced backyard), but were closeenough that their radio home ranges.We used compositionalanalysis (Aebischer signalcould be detectedwith the receiverconnected only et al. 1993) to compare the proportion of coverage of to the antenna cable (i.e., with the antenna detached). each category to the proportion of hawk locations in In such cases, we estimated the location of the hawk to each category among all home ranges.We used the Kap- be within 30 m of the strongestsignal. If a hawk wasout- lan-Meier (1958) method to estimate survival of radio- of-sight,in an area where we did not have permissionto tagged hawks. We conducted statisticalanalyses in the access, and not within 30 m of the observer, we estimated JMP IN 3 Windows Version statisticalpackage (Sall et al its position by triangulation. We marked all locationson 2001) or Resource Selection AnalysisSoftware for Win- an atlas of city streets. dows (Copyright 1999, Fred Leban). Analyses.We plotted all locationsof hawkson a digital representationof the streetsand land use categoriesin RESUI•TS Tucson (Shaw et al. 1996). We first measured distances between all consecutivelocations of each hawk, including We radio-tagged 21 hawks in 1999 (10 females locations estimatedfrom aircraft and elevated positions and 11 males), and 19 hawks in 2000 (8 females (i e., in instanceswhere we could not locate hawksby and 11 males). We relocated the 40 hawks a total homing). We assigned all locations, except those esti- mated from aircraft and elevatedpositions, to one of five of 527 times (58.3% by sight, 22.0% with the an- categories based on kind of development and level of tenna detached from the receiver in combination human use (i.e., low-densityresidential areas [<7.4 resi- with triangulation,7.4% by triangulation alone, 1VL•ac• 2004 POST-FLEDGLING COOPER'S HAmkS 29

Malesß Females ]

2O

15

• 10

o•,•

5

o 9 11 13 15 17 19 21 23 25 27

Weeks After Hatching Figure 1. Mean distancebetween consecutive locations in 2-wkperiods for radio-taggedCooper's Hawks after fledg- lng from nestsin Tucson,Arizona, 1999-2000 (N = 4-17 locationsfor males,and 4-13 for females,depending on period; mean number of daysbetween locations from 11-22 wk after hatching wassimilar for males [• = 6.7, SD = 5 2] and females [• = 7.6, SD = 7.4; t-test,P = 0.38]).

7.4% from elevatedpositions, and 4.9% from air- We estimated size of home range during fall/ craft). Six hawks (two females and four males) lost winter for nine of 10 hawks (Table 1). Seven of the the tail feather on which the radio was attached, nine home ranges had become relatively stable in and five (three females and two males) died before area by the end of our sampling period (-<5.1% meaningful information on their movementscould increase in area over at least the last week of sam- be collected. Of the remaining 29 radio-tagged pling), but home ranges of two females were still hawks, 19 (eight females and 11 males) either were increasing (8.8 and 13.2%), when the hawk died located sporadically (i.e., N--< 9) throughout the or the radio failed (Table 1). We did not estimate fall, or early in the fall but never again, and 10 (five home range size for one hawk becausethe sample females and five males) remained within 20 km of of locations was small (N = 11) and was obtained the Tucson metropolitan area at least into late Oc- during one month (December). Home range size tober. for the nine hawks averaged771 ha (SD = 403), Movements. The typical pattern of movements and did not differ (t-test, P = 0.80) between males for hawkswe were able to track through earlywin- (• = 804 ha, SD = 456) and females (• = 731, SD ter consistedof sedentary behavior (i.e., short in- = 387; Table 1). Distance from center of home ter-location distances) in the natal area, followed range to natal site was nearly twice as far for fe- by relatively long movementsbeginning 11-13 wk males (• = 10.9 km, range = 4.2-19.5 km, SD = after hatching, and finally by sedentarybehavior 6.4) as males (2 = 6.0 km, range = 2.2-13.3 km, again when hawks establisheda fall/winter home SD = 5.0), but the difference was not significant range (Fig. 1). Inter-location distanceswere, on av- (t-test;P = 0.23). Eight of nine home ranges en- erage, greater for females (• = 6.8 km, range = compassedone or two traditional nest sitesof Coo- 0.02-51.7 km, SD = 9.8) than males (• = 3.8 km, per's Hawks, but none of the dispersinghawks in- range = 0.05-20.8 km, SD = 5.4; t-test,P = 0.02; corporated their natal nest sitesin their fall/winter Fig. 1). One female, for example, made two move- ranges. ments of >59 km in a period of <20 d. Siblings Habitat. We found hawksin a variety of environ- did not movetogether after they left the natal area, ments after they left their natal areas and before and neither males nor females moved in discern- they settled into fall/winter home ranges; environ- able patterns prior to establishing fall/winter ments used included riparian areas (35%; a sub- home ranges. We generally found hawks of both category of open space), high-density residential sexesin scatteredlocations throughout Tucson be- areas (25%), low-densityresidential areas (22.5%), fore they settled (Fig. 2). and parks and golf courses(17.5%) (N = 40 lo- 30 MANNAN ET AL. VOL. 38, NO. I

3

North

5 km

Figure2. Sequentialmovements of two hawks(female = solidline; male = dashedline) betweenleaving their natal area and settlinginto a fall/winter home range in Tucson,Arizona, 1999-2000. Multiple locationsin the natal area are representedby a single location. MARCH 2004 POST-FLEDGLING COOPER'S HAWKS 31

Table 1. Size of home rangesduring fall/winter, and distancefrom center of thesehome rangesto natal nestsfor post-fledglingCooper's Hawks in Tucson,Arizona, 1999-2000.

PERIOD TRACKED PERCENT DISTANCE TO HAWKID a IN HOME RANGE N b INCREASEc NESTd (km) AREA(HA) M327 16 Sep 99-18 Feb 00 20 4.3 2.2 492 M457 12 Sep 00-2 Dec 00 32 3.6 9.1 537 M885 11 Aug 99-8 Mar 00 73 0.0 2.2 556 M854 16 Aug 99-26 Oct 99 13 5.l 3.1 854 M313 9 Sep 00-11 Dec 00 23 3.2 13.3 1580 F215 26 Aug 00-26 Nov 00 32 8.8 4.2 409 F977 22 July 99-15 Feb 00 63 0.0 10.8 593 F276 23 Sep 00-3 Dec 00 30 13.2 9.2 628 F259 14 Aug 00-13 Nov 00 16 0.0 19.5 1294 M = males; F = females. Number of relocationsidentified by sight or triangulation,and those estimatedto be within 30 m of the observer. Increasein area in the home range over at leastthe last week of sampling,representing at most the last five locations. Distancefrom center of the winter home range to navalnest.

cations for all hawks combined, excluding those land use categories inside home ranges (X2 = estimated from aircraft and elevated positions). 6.058, df = 4, P = 0.19; Table 2). Once hawkssettled, their home rangesalso varied Survival.Survival of radio-taggedhawks through in composition.Some encompassedmostly high- 180 d was 67%. Two males and four females died density residential areas, others encompassed while we were tracking them, but there was no dif- mostly low-density residential areas, and others ference in survival (Wilcoxon test, P = 0.66) be- were dominated by cemeteries, neighborhood tween males (75%) and females (64%). Of the six parksand natural open space,or golf coursesand hawks that died, two females were killed by colli- district and regional parks (Table 2). We found no sions with cars, one male was electrocuted, and discernablepattern of selection among hawksfor three died of unknown causes. Of the 10 hawks we

Table 2. Percentagesof locationswithin fiveland usecategories ain home rangesof fledglingCooper's Hawks during their first fall/winter in Tucson, Arizona, 1999-2000.

PERCENTOF LOCATIONS(PERCENT OF HOMERANGE) b

OPEN SP^CE OPEN SP^CE LOW-DENSITY HIGH-DENSITY WITH LOW WITH HIGH HAWK ID c RESIDENTIAL RESIDENTIAL HUMAN USE HUMAN USE OTHER

M327 0.0 (6.1) 60.0 (51.4) 15.0 (21.5) 10.0 (11.7) 15.0 (9.3) M457 78.6 (58.5) 3.6 (12.5) 7.1 (16.7) 0.0 (1.2) 10.7 (11.1) M885 0.0 (6.7) 78.6 (52.8) 12.9 (17.3) 0.0 (1.9) 8.5 (21.3) M854 46.1 (37.4) 7.6 (13.4) 38.5 (33.5) 0.0 (4.8) 7.8 (10.9) M313 16.7 (7.4) 45.8 (34.3) 29.2 (37.6) 0.0 (1.2) 8.3 (19.5) F215 36.7 (12.9) 13.3 (28.2) 3.3 (6.6) 46.7 (37.7) 0.0 (14.6) F977 81.4 (72.1) 0.0 (3.3) 10.2 (18.4) 8.4 (5.1) 0.0 (1.1) F276 0.0 (0.0) 50.0 (54.3) 19.2 (17.8) 26.9 (2.4) 3.9 (25.5) F259 25.0 (24.7) 31.2 (24.9) 31.2 (39.7) 6.3 (6.4) 6.3 (4.3) a Low-densityresidential (<7.4 residences/ha);high-density residential (>7.4 residences/ha);open spacewith low human use = cemeteries,neighborhood parks and natural open space;open spacewith high human use = golf courses,district and regional parks, and schools,and other = roadways,and commercial industrial,and agriculturalareas. b X2 = 6.058, df = 4, P: 0.19; compositionalanalysis. "M = males; F = females. 32 •_ANNAN ET AL. VOL. 38, NO. 1 were able to track through early winter, six were ranges of Cooper's Hawks in Tucson during their known to be alive 2 yr after they were radio-tagged. first fall/winter were, on average, about 11 times larger than the home ranges of breeding males DISCUSSION (Mannan and Boal 2000), and typicallyoverlapped The wide-ranging movements we detected one or two traditional breeding ranges.We do not among fledgling Cooper's Hawks between 11 and know whether the home ranges establishedin the 22 wk after hatching are similar to those observed first fall/winter of life persist into the following in related speciesduring the early stagesof natal breeding season and beyond, but subsequentob- dispersal(e.g., EurasianSparrowhawks [Accipiter ni- servationssuggest that the Cooper's Hawks we ra- sus] in southern Scotland [Newton 1986:261]), dio-taggedremained in or closeto them for several and presumablywere explorations in search of a years. For example, four hawks (three males and place to settle for the winter. Environmental and one female) nested, attempted to nest, or died 2 social cues that triggered Cooper's Hawks to settle or 3 yr after they were radio-tagged at sites that for the fall/winter are unknown, but rich sources were encompassedby their first winter home rang- of food (e.g., concentrationsof birds at bird feed- es. Also, one male and one female hawk, radio- ers), and low levelsof intraspecific and interspecif- tagged in 2000, nested 2 yr later in sitesthat were ic competition are likely candidates.A variety of 1.0 and 4.6 km, respectively,from the edges of environmentswere used by hawksduring the "ex- their first winter home ranges. ploration" period but riparian systemswere used Home rangesof the hawkswe trackedwere dom- by hawks (35%) more than they generallyoccur inated by a variety of land use categories,and no on the landscape(e.g., 6% in the Tucson metro- consistentpattern of selection by hawks was evi- politan area, Shawet al. 1996), suggestingthat ri- dent among categories within home ranges. We parian corridors may be attractive to dispersing speculate that Cooper's Hawks during their first hawks as sources of food and cover, or facilitate fall/winter in Tucson can persistin a variety of ur- their movements in some manner in the urban en- ban environments because rich sources of food vironment. (e.g., concentrations of birds) are common and Females disperse farther than males in many widespread (Germaine et al. 1998). speciesof birds (Greenwoodand Harvey 1982). Survival of birds of prey during dispersalis low Rosenfield et al. (1996) reported that the median for many (e.g., Belthoff and Ritchison distancemoved by 10 male Cooper'sHawks during 1989, Rohner and Hunter 1996, Ganey et al. 1998, natal dispersalwas 6.4 km, whereas two females but see Harmata et al. 1999) probably because moved 14.4 and 79.0 km. We found that females fledglingsare relativelyinexperienced in acquiring moved greater distances between locations than food and avoiding predators and other agents of males, especiallyfrom 11-22 weeksafter hatching. mortality. We found survival to be relatively high The wide-rangingmovements of femalesmay cause among Cooper's Hawks in Tucson for 6 mo after them to settle farther from their natal nests than fledging. Abundance of food can influence survival males,but our small sampleof winter home ranges of dispersing birds (e.g., Rohner and Hunter and the restricted area in which we were able to 1996), and we propose that an abundanceof prey follow hawks precluded definitive assessmentof in Tucson may reduce mortality and offset agents this relationship. The movements of hawks we re- of mortality common in developedenvironments port herein also may be biasedin that some of the (e.g., collisionswith vehiclesand windows,electro- hawkswe could not locate may have moved farther cution). Our speculationthat Tucson providesam- than those we were able to track; thus, our data ple food for Cooper's Hawks is supported by evi- may underestimate the distancesmoved by both dence that total densityand biomassof birds often sexes.It is also possiblethat some of the hawkswe is higher in urban than nonurban areas (e.g., in located only sporadicallythroughout the fall never Tucson,Emlen 1974;elsewhere, Beissinger and Os- establisheda home range. borne 1982, Blair 1996, Marzluff et al. 1998). How- Settling and establishing home ranges, even if ever,our estimatesof survivalof fledgling Cooper's temporary, after some period of exploration ap- Hawks in Tucson may be biasedin at least two ways. pears to be common during the early stagesof na- First, we could have overestimated survival because tal dispersalof predatory birds (e.g., Beske 1982, some of the hawks we could not find may have Ferrer 1993, Walls and Kenward 1998). Home died. And second, attachment of radio-tagscould IV[ARCH 2004 POST-FLEDGI.ING COOPER'S HAWKS 33 have negativelyaffected the hawkswe studied (e.g., asisin Cooper'sHawks from Arizona.J. Wildl.Dis. 34' R. Reynolds,G. White, S. Joy, and R. Mannan un- 590-593. publ. data) and led to an underestimateof survival. BoYcE,M.S. 1992. Population viability analysis.Annu. Rev Boal (1997) modeled the dynamicsof the pop- Ecol. Syst.23:481-506. BROWN,D.E., C.H. LOWE,AND C.P. PASE. 1979. A digitized ulation of Cooper's Hawks in Tucson, and con- classification system for the biotic communities of cluded that it was declining at about 8%/yr, pri- North America with community (series) and associa- marily due to a disease(trichomoniasis) that killed tion examplesfor the southwest.J. Ariz.-Nev.Acad. Scz about 40% of the nestlingsand fledglings each 14 (Suppl. 1): 1-16. year (Boal et al. 1998, Boal and Mannan 1999). CLOBERT,J., E. DANCHIN,A.A. DHONDT,AND J.D. NICH- The estimate of survival of juvenile Cooper's OLS.[EDS.]. 2001. Dispersal.Oxford Univ. Press,Inc., Hawks in Tucson reported herein is higher than New York, NY U.S.A. what Boal (1997) used in his model; thus, notions EMI,EN,J.T. 1974. An urban bird community in Tucson, about whether Tucson is a "source" or "sink" pop- Arizona: derivation, structure,regulation. Condor76' 184-197. ulation for Cooper'sHawks could changewhen re- ENVIRONMENTAL SYSTEMS RESEARCH INSTITUTE. 1996. vised estimatesof post-fiedgingsurvival are com- Arcview GIS. Environmental SystemsResearch Insti- bined with updated information on adult survival, tute, Inc., Redlands, CA U.S.A. productivity, and mortality from trichomoniasis FERRER,M. 1993. Juvenile dispersalbehaviour and natal and other agents. philopatry of a long-livedraptor, the SpanishImperial

ACKNOWLEDGMENTS Eagle Aquila adalberti.Ibis 135:132-138. GANEY,J.L., W.M. BLOC•:,J.K. DWYER,B.E. STROHMEYER, We thank C. Boal, D. Ripper, R. Rosenfield, and two ANDJ.S. JENNESS.1998. Dispersalmovements and sur- anonymousreferees for reviewing drafts of this manu- vival rates of juvenile Mexican Spotted in north- script.The Arizona Game and Fish Department Heritage ern Arizona. Wilson Bull. 110:206-217. Fund supported the project (U98006). GERMAINE, S.S., S.S. ROSENSTOCK,R.E. SCHWEINSBURG, LITERATURE CITED AND W.S. RiCHARDSON.1998. Relationships among AEmSCHER,N.J., P.A. ROBERTSON,AND R.E. K•NW•mD. breeding birds, habitat, and residentialdevelopments 1993. Compositional analysisof habitat use from ani- in greater Tucson,Arizona. Ecol.Appl. 8:680-691. mal radio-trackingdata. Ecology 74:]313-]325. GREENWOOD,PJ. AND P.H. HARVEY.1982. The natal and BEISSINGER,S.R. AND D.R. OSBO•E. 1982. Effects of ur- breeding dispersalof birds. Annu. Rev. Ecol. Syst.13 banization on avian community organization. Condor 1-21. 84:75-83. HARMATA,A.R., G.J. MONTOPOLI,B. OAKLEAF,P.J. HAR- --AND M.I. WESTP•IaL. 1998. On the use of demo- MATA, AND M. RESTANI. 1999. Movements and survival graphic models of population viabilityin endangered of Bald Eagles in the Greater Yellowstone Ecosystem speciesmanagement. J. Wildl. Manage.62:821-841. J. Wildl.Manage. 63:781-793. BELTHOFF,J.R. AND G. RiTCHISON.1989. Natal dispersal HOOGE, P.N. AND B. EICHENLAUB. 1997. Animal move- of Eastern Screech-Owls. Condor 91:254-265. ments extension to Arcview. ver. l.1. Alaska Science BEsva•,A.E. 1982. Local and migratory movementsof ra- Center--Biological SciencesOffice, U.S. Geological dio-taggedjuvenile harriers. RaptorRes. 16:39-53. Survey,Anchorage, AK U.S.A. BI>aR, R.B. 1996. Land use and avian speciesdiversity HOWARD,W.E. 1960. Innate and environmental dispersal along an urban gradient. Ecol.Monog• 6:506-519. of individual vertebrates. Am. Midl. Nat. 63:152-161 BLOOM,P.H. 1987. Capturingand handling raptors.Pag- KAPLAN,E.L. AND P. MEIER. 1958. Nonparametric esti- es 99-123 in B.A. Giron Pendleton, B.A. Millsap, K.W. mation from incomplete observations.J. Am. Stat. As- Cline, and D.M. Bird [EDS.], Raptor management soc. 53:457-481. techniquesmanual. Natl. Wildl. Fed., ,DC KENWARD,R.E. 2001. A manual for wildlife radio tagging U.S.A. Academic Press,San Diego, CA U.S.A. BOAL,C.W. 1997.An urban environmentas an ecological --AND K.H. HODDLER.1996. RangesV: an analys•s trap for Cooper's Hawks. Ph.D. dissertation,Univer- systemfor biological location data. Institute of Ter- sity of Arizona, Tucson, AZ U.S.A. restrial Ecology,Dorset, U.K. --AND R.W. M_^•NAN. 1998. Nest-site selection by KOENIG,W.D., P.N. HOOGIE, M.T STANBACK,AND J. HAY- Cooper's Hawks in an urban environment. J. Wildl. DOCK. 2000. Natal dispersal in the cooperatively Manage.62:864-871. breeding Acorn Woodpecker.Condor 102:492-502. --AND --. 1999. Comparativebreeding ecology LANDE,R. 1988. Demographic models of the northern of Cooper's Hawks in urban and exurban areas of Spotted , occidentaliscaurina. Oecologia75' southeasternArizona. J. Wildl. Manage.63:77-84 601-607. --, --, AND K.S. HUDELSON. 1998. Trichomoni- MANNAN,R.W. AND C.W. BOAL.2000. Home range char- 34 MANNAN ET AL. VOL. 38, NO. 1

acteristicsof male Cooper's Hawks in an urban envi- lemetry. Pages 370-418 in T.A. Bookhout [ED.], Re- ronment. Wilson Bull. 112:21-27. search and management techniques for wildlife and M2mzEuvv,J.M., F.R. GEHLB^CH,•tND D.A. MANUWAL. habitats, Fifth Ed. The Wildlife Society,Bethesda, MD 1998. Urban environments: influences on avifauna U.S.A. and challengesfor the conservationist.Pages 283-292 SHAW,W.W., L.K. HARPdS,M. LMNGSTON,J.P. CHARPEN- inJ.M. Marzluff and R. Sellabanks,[EDS.], Avian con- T•ER,AN•) C. WISSLER.1996. Pima County habitat in- servation: research and management, Island Press, ventory, Phase II. Arizona Game and Fish Depart- Washington, DC U.S.A. ment, Contract No. 650028-001, Phoenix, AZ U.S.A. M•LLER,G.S., R.J. SMALL, ^ND E.C. MESLOW.1997. Habitat STAMPS,J.A. 2001. Habitat selection by dispersers:inte- selectionby Spotted Owls during natal dispersalin grating proximate and ultimate approaches. Pages westernOregon. J. Wildl.Manage. 61:140-150. 230-242 in J. Clobert, E. Danchin, A.A. Dhondt, and NEWTON,I. 1986. The sparrowhawk.T. & A.D. Poyser J.D. Nichols [EDS.], Dispersal. Oxford Univ. Press, Ltd., Town Head House, Gaiton, Waterhouses, Staf- Inc., New York, NY U.S.A. fordshire, U.K. STEWART,A.C., R.W. CAMPBELL,AND S. DICKIN. 1996. Use ODUM, E.P. AND EJ. KUENZLER.1955. Measurementsof of dawn vocalizationsfor detecting breeding Cooper's territory and home-rangesizes in birds. Auk 72:128- Hawks in an urban environment. Wildl. Soc. Bull. 24. 137. 291-293. P^YNE,R.B. ANDL.L. PAYNE.1993. Breeding dispersalin STRONG,A.M. AND G.T. BANCROFT.1994. Postfledgingdis- Indigo Buntings:,circumstances and consequencesfor persal of White-crowned Pigeons: implications for breeding successand population structure. Condor95: conservation of deciduous seasonal forests in the Flor- 1-25. ida Keys.Conserv. Biol. 8:770-779. PULLIAM,H.R. 1988. Sources,sinks, and population reg- TEMPLE,S.t. 1989. The role of dispersalin the mainte- ulation. Am. Nat. 132:652-661. nance of bird populationsin a fragmented landscape. --AND BJ. DANIELSON.1991. Sources, sinks, and Acta Congo.Int. Ornithol.20:2298-2305. WALks, S.S. AND R.E. K•NWARD. 1995. Movements of ra- habitat selection:a landscapeperspective on popula- tion dynamics.Am. Nat. 137:50-66. dio-taggedCommon BuzzardsButeo buteo in their first RESTANI,M. ANDW.G. MATTOX.2000. Natal dispersalof year. Ibis 137:177-182. --AND --'. 1998. Movements of radio-tagged PeregrineFalcons in Greenland.Auk 117:500-504. BuzzardsButeo buteo in early life. Ibis 140:561-568. ROHNER,C. ANDD.B. HUNTER.1996. First-yearsurvival of WALTERS,J.F. 2000. Dispersalbehavior: an ornithological Great Horned Owls during a peak and decline of the frontier. Condor 102:479-481. snowshoehare cycle. Can.J. Zool.74:1092-1097. WHITE, G.C. ANDR.A. GARROTT.1990. Analysisof wildlife ROSENFIELD,R.N. ANDJ.BIELEFELDT. 1992. Natal dispersal radio-trackingdata. Academic Press,Inc., New York, and inbreeding in the Cooper's Hawk. WilsonBull. NY U.S.A. 104:182-184. WIENS,J.A. 2001. The landscapecontext of dispersal.Pag- --AND --. 1993. Cooper's Hawks (Accipitercoop- es 96-109 inJ. Clobert, E. Danchin, A.A. Dhondt, and er/i) In A. Poole and F. Gill [EDS.], The birds of North J.D. Nichols lEDs.], Dispersal. Oxford Univ. Press, A•nerica, No. 75. The Academy of Natural Sciences, Inc., New York, NY U.S.A. Philadelphia, PA and The American Ornithologists' WORTON,B.J. 1989. Kernel methods for estimatingthe Union, Washington, DC U.S.A. utilization distribution in home range studies.Ecology , --, AND S.M. Vos. 1996. Skewed sex ratios in 70:164-168. Cooper'sHawk offspring.Auk 113:957-960. WYLLIE,I. 1985. Post-fledgingperiod and dispersal of SALL,J., A. LEHMAN,AND L. CREIGHTON.2001. JMP Start young sparrowhawksAccipiter nisus. Bird Study32:196- Statistics,2nd Ed. SASInstitute. Duxbury Press,Pacif- 198. ic Grove, GA U.S.A. SAMUEL, M.D. AND M.R. FULLER. 1994. Wildlife radiote- Received27 August2002; accepted7 September2003