THE SOUTHWESTERN NATURALIST 59(4): 542–547 DECEMBER 2014

SCAVENGING OF MIGRATORY CARCASSES IN THE SONORAN DESERT

* ANDREW M. ROGERS,MICHELLE R. GIBSON,TYLER POCKETTE,JESSICA L. ALEXANDER, AND JAMES F. DWYER

EDM International, Inc., 4001 Automation Way, Fort Collins, CO 805252 (AMR, MRG, TP, JLA, JFD) Present address of MRG: School of Biological and Biomedical Sciences, University of Durham, South Road, Durham DH1 3LE, U.K. *Correspondent: [email protected]

ABSTRACT—In this study we report avian and mammalian scavengers foraging on migratory bird carcasses in the Sonoran Desert. We used remote cameras to monitor carcasses we found along a power line right-of-way (n = 25). We documented four species scavenging 10 carcasses (kit fox, Vulpes macrotis,n= 4; coyote, Canis latrans,n= 3; , Corvus corax,n= 2, and greater , Geococcyx californianus,n= 1) and recorded coyote tracks at three additional carcasses. Neither remote cameras nor tracks indicated the scavenger species of the remaining carcasses. Our data suggest migrant might provide an important food source for resident scavengers, particularly in desert habitats where food can be scarce. Our study also supports prior assertions that failure to account for removal of carcasses by scavengers might cause errors in estimates of mortality.

RESUMEN—En este estudio reportamos observaciones de mam´ıferos y aves carroneros˜ aliment´andose de cad´averes de aves migratorias en el desierto de . Usamos c´amaras remotas para monitorear cad´averes (n = 25) que encontramos en la zona de restricci´on a lo largo de un cable de luz. Se registraron cuatro especies consumiendo carrona˜ de 10 cad´averes (zorra nortena,˜ Vulpes macrotis,n= 4; coyote, Canis latrans,n= 3; cuervo comun,´ Corvus corax,n= 2, y correcaminos norteno,˜ Geococcyx californianus,n= 1), y se registraron huellas de coyote en 3 cad´averes adicionales. Ni c´amaras remotas ni huellas indicaron la especie carronera˜ para el resto de los cad´averes. Nuestros datos sugieren que las aves migratorias son potencialmente una fuente importante de alimento para las especies de carroneros˜ residentes, particularmente en los h´abitats des´erticos donde la comida puede ser escasa. Nuestro estudio tambi´en apoya previas declaraciones de que no considerar la remoci´on de cad´averes por animales carroneros˜ puede ocasionar errores en las estimaciones de la mortalidad.

Studies of avian migration typically focus on when, with anthropogenic structures also have been widely where, why, and how birds migrate (Bowlin et al., 2010). reported (Longcore et al., 2013; Sporer et al., 2013), but Survival during migration is of interest, particularly there is little understanding of the interactions between because it is often lower than survival during sedentary ecological barriers and anthropogenic structures. The periods. However, because dead migrants are rarely ecological role dead migrants might play in the commu- found, details describing the timing or location of nities surrounding the anthropogenic structures has not mortality during migration are scarce (Sillet and Holmes, been investigated. 2002). Where avian survival during migration has been We hypothesized that the carcasses of migratory birds investigated, mortality has been attributed to exhaustion would be an important prey source to desert scavengers while traversing ecological barriers (Newton, 2008), and predicted that if migratory bird carcasses could be collision with anthropogenic structures (Longcore et al., found and monitored, desert scavengers would be shown 2013; Sporer et al., 2013), loss of stopover habitat consuming these carcasses. Direct observations of scaven- (Sutherland, 1996; Schwarzer et al., 2012), and predation gers might not provide an unbiased estimator of scavenger (Schmaljohann and Dierschke, 2005). activity because some scavenger species could be more Deserts are an important ecological barrier for avian difficult to detect. For example, some of the potential migrants moving between Europe and Africa (Newton, scavengers in our study were hunted and were cryptically 2008; Strandberg et al., 2010) and between North colored (e.g., coyote, Canis latrans) while others were America and South America (Patten et al., 2003). protected and were boldly colored (e.g., common raven, Migratory birds frequently die while traversing deserts Corvus corax). Remote cameras can provide useful infor- during migration, but almost no information exists mation when human observers might affect behaviors of describing the fate of their carcasses. Avian collisions interest (Dwyer and Doloughan, 2013). In this study we December 2014 Rogers et al.—Scavenging of migratory bird carcasses 543 used direct observations and remote cameras to provide Corporation, Cody, KS). Each camera was programmed to the first documentation of residents scavenging the capture three, eight-megapixel photographs at 5-s intervals each carcasses of migratory birds in the Sonoran Desert of time the camera was triggered. We initially used high-sensitivity southern . settings on the cameras but, during presurvey trials, found that at these settings the cameras exhausted their memories and power supplies recording wind-driven movements of vegetation. MATERIALS AND METHODS—Study Area—We conducted our Thus, we used medium-sensitively settings throughout the study study along a power line right-of-way between El Centro, to balance oversensitivity to vegetation movement with under- California (3284703100N, 11583304700W) and Ocotillo, California sensitivity to scavengers. (3284401900N, 11585903900W) on land owned by the U.S. Bureau of Land Management. We selected our survey area because large The cameras recorded color photographs illuminated via numbers of migrants breeding in North America and wintering ambient light during the day, and black and white photographs in Central and South America pass through this portion of the illuminated via infrared at night, allowing 24-h continuous Sonoran Desert (Patten et al., 2003), because collisions with observation of carcasses. At each carcass we monitored, we power lines are of management interest (Ponce et al., 2010; placed one remote camera under a nearby creosote bush. Barrientos et al., 2012; Sporer et al., 2013), and because desert Creosote bushes were common in the study area and provided passages can be particularly difficult for migrant passerines visual cover, reducing the likelihood that cameras would be (Newton, 2008; Strandberg et al., 2010). Our study area was noticed by potential scavengers or people. Creosote bushes entirely within the Sonoran Desert with elevations from 1–300 m could have obscured the infrared sensors of the camera. To be above sea level (El Centro is below sea level). Vegetation was sure each camera had a clear view of each monitored carcass, we sparse, dominated by creosote bush (Larrea tridentate) and laced any branches that would obscure the carcass behind ocotillo (Fouquieria splendens) and, to a lesser extent, Opuntia adjacent branches outside the view of the camera. We wrapped species of cholla and prickly pear , indigo bush each camera in burlap and the branches of creosote bushes to (Psorothamnus species), and occasional mesquite trees (Prosopis further break up the boxy shapes of the cameras. We then species). Rainfall averaged <13 cm per year (Western Regional revisited each monitored carcass every 24–48 h. The substrate in Climate Center, 2013) and summer temperatures regularly the study area was a mix of sand and gravel. If the carcass was reached 408C. absent, we recorded any tracks within 5 m, if present, and Data Collection—Between 15 March 2013 and 15 June 2013, retrieved the camera. Because the cameras were triggered based we walked transects daily through our study area in search of partially on detection of body heat, being in an already warm migratory bird carcasses. We defined our survey period to environment could have decreased the likelihood of cameras coincide with the peak of avian migration through our study triggering when scavengers were present. Recording tracks area (Patten et al., 2003). Each transect was 500 m long and 50 enabled us to evaluate scavenging in cases where the camera transects were surveyed 14–16 times. Transects were completed did not capture an image of the scavenger. by three observers walking parallel to one another along a power line right-of-way. Each observer surveyed a 25-m wide portion of RESULTS—While walking transects, we recorded 26 the total transect, with each observer slightly overlapping the instances of a common raven in flight dropping to the areas surveyed by the adjacent observer so that total transect ground with nothing in its feet or and then width was 65 m. Each observer zig-zagged back and forth within immediately flying up with a migratory bird carcass in their area of responsibility within each transect (as in Faanes, its beak. Scavenged birds identified to species were two 1987; Barrientos et al., 2012), and walked slowly at about 3–4 black-headed grosbeaks (Pheucticus melanocephalus), one km/hr (2 mph; as in Murphy et al., 2009). black-throated gray warbler (Setophaga nigrescens), one We began transects at local sunrise and continued until orange-crowned warbler (Vermivora celata), one white- approximately 6 h after sunrise. While walking transects, we winged dove (Zenaida asiatica), and two yellow warblers recorded observations of foraging behaviors of potential scavengers; for example, a common raven (Corvus corax)in (Setophaga petechia). Five warblers could not be identified flight with nothing in the feet or beak, dropping to the ground, to species (Family Parulidae) and 14 birds could be and then immediately flying up again in possession of a identified only as passerines (Order Passeriformes), based migratory bird carcass. When this occurred, we followed the on size, as the raven departed with the carcass in its beak. common raven to identify the scavenged animal to species level, These scavenging events occurred primarily during the if possible, or to family or order otherwise. We also recorded the morning (mean – SE = 0757 h – 20 min; min = 0543 h, locations of the migratory bird carcasses we encountered. To max = 1138 h) with two scavenging events observed in minimize the possibility that our presence might influence March, 14 in April, 10 in May, and none in June after scavenger activity, we did not collect, mark, move, or remove any young ravens had fledged and family groups of ravens migratory bird carcass we encountered. We recorded carcass moved away from the power line corridor. locations with a WAAS-enabled GPSmap 62s receiver (Garmin We used remote cameras to monitor the carcasses of 21 International, Olathe, KS) and used the global positioning system device to return to carcass locations. We used direct passerines and four nonpasserines (n = 25, Table 1). observations and remote cameras to document scavenging of Cameras recorded kit fox (Vulpes macrotis), coyote, migratory bird carcasses. common raven, and greater roadrunner (Geococcyx We also used remote cameras to document scavenging californianus) scavenging 10 of the carcasses (Fig. 1). events. We used three remote cameras, one HC500 (Reconyx, These scavenging events occurred primarily during the Inc., Holmen, WI) and two Bushnell Trophy Cams (Bushnell night (n = 6), with fewer recorded during the early 544 The Southwestern Naturalist vol. 59, no. 4

TABLE 1—Remote cameras documented scavenging of migrant bird carcasses (n = 25) in the Sonoran Desert.

Scavenged species Scavenger speciesa

Common name Scientific name Common name Scientific name

Ash-throated flycatcher Myiarchus cinerascens N/A N/A Black-throated sparrow Amphispiza bilineata —— Brant goose Branta bernicla Coyote Canis latrans Lazuli bunting2 P. amoena Coyote tracks2 C. latrans Lazuli bunting2 P. amoena Coyote tracks C. latrans Lincoln’s sparrow Melospiza lincolnii —— MacGillivray’s warbler Oporornis tolmiei —— Mourning dove Zenaida macroura Coyote C. latrans Mourning dove Z. macroura N/A N/A Nashville warbler Vermivora ruficapilla —— Orange-crowned warbler Vermivora celata —— Red-winged blackbird Agelaius phoeniceus —— Townsend’s warbler Setophaga townsendi Common raven Corvus corax Townsend’s warbler S. townsendi Common raven C. corax Unidentified Empidonax flycatcher Empidonax species Kit fox Vulpes macrotis Western tanager Piranga ludoviciana N/A N/A White-winged dove Zenaida asiatica Kit fox V. macrotis Willow flycatcher Empidonax traillii N/A N/A Willow flycatcher E. traillii N/A N/A Wilson’s warbler2 Wilsonia pusilla Coyote tracks C. latrans Wilson’s warbler W. pusilla —— Wilson’s warbler W. pusilla Coyote C. latrans Wilson’s warbler W. pusilla Greater roadrunner Geococcyx californianus Wilson’s warbler W. pusilla Kit fox V. macrotis Yellow warbler Setophaga petechia Kit fox V. macrotis a N/A indicates the carcass was not scavenged during the monitoring period; — indicates the carcass was scavenged but the scavenger species was not identified. b Three carcasses were within the frame of the same camera setup. morning before 0630 (n = 2) or later in the day (n = 2). DISCUSSION—Most of the carcasses we found occurred Coyote tracks indicated the scavenger species at three during the peak of spring migration in our study area carcasses where cameras failed, but we could not identify (Patten et al., 2003). Little is known about the fate of the the time of day these carcasses were scavenged. Seven carcasses of migratory birds that die during migration. carcasses were scavenged, but neither cameras nor tracks Our data suggest migratory birds might provide an indicated the scavenger species. Combining these 20 important food source for resident scavengers in desert events, one scavenging event occurred in March, seven in habitats where food resources can be rare. Common April, 10 in May, and two in June. Five carcasses were not ravens are regularly reported as facultative scavengers scavenged during the monitoring period. (Boarman and Heinrich, 1999; Matley et al., 2012), and Combining both types of scavenging events (direct our observations of common ravens flying directly from observations and remote cameras), carcasses scavenged scavenging sites to nest sites (TP, pers. obs.) suggests the by birds were consistently taken during the day (26/26 common ravens in our study area used scavenged documented via direct observation; 3/3 documented via carcasses to provision nestlings. Prior studies of the diets remote cameras; 100%) and carcasses scavenged by of kit foxes indicated that a low proportion (6.9%) of mammals were taken primarily at night (6/7 documented fecal scats collected in the Chihuahuan Desert contained via remote cameras; 86%; this excludes three carcasses avian remains (Moehrenschlager et al., 2007). In nonde- where the mammalian scavenger was identified by tracks, sert habitat near Bakersfield, California, White et al. seven carcasses where the scavenger was not identified at (1985) found slightly higher values, with 8.6% of kit fox all, and five carcasses which were not scavenged). scats containing avian remains. Prior studies of the diets Combining all carcasses, we recorded three events in 10 of coyotes in the Sonoran Desert also indicated that a low days of monitoring in March, 22 events in 22 days of proportion (2.6%) of fecal scats collected in fall monitoring in April, 23 events in 23 days of monitoring in contained avian remains (Hern´andez et al., 1994), May, and three events in 10 days of monitoring in June. though scats collected year-round had higher proportions Migratory bird carcasses were more likely to be present in of birds (7.4%; McKinney and Smith, 2007). Birds were April and May (v2 = 9.11, df = 3, P = 0.028). also relatively rare in the diets of coyotes in the December 2014 Rogers et al.—Scavenging of migratory bird carcasses 545

FIG. 1—Migrant passerine carcasses scavenged in the Sonoran Desert. a) Kit fox scavenging Wilson’s warbler. b) Coyote scavenging mourning dove. c) Common raven scavenging Townsend’s warbler. d) Greater roadrunner scavenging Wilson’s warbler (see Table 1 for scientific names).

Chihuahuan Desert of New , where birds occurred It might be that injured scavengers in our study area in only 1.6% of scats collected in spring and 1.5% of scats depended on the carcasses of migrating birds. If so, then collected year-round (Hern´andez et al., 2002). Our study cascading effects from the mortality of avian migrants suggests that, though the proportion of birds in the scats within an ecological barrier could have influenced the of mammalian scavengers might be consistently low in ecological community of our study area. Future research general, those studies might not well-represent individual investigating coyote scats in our study area, particularly scavengers occupying areas where migrant passerine with emphasis on seasonal difference in scat composition, carcasses occur in disproportionately high numbers. would help resolve the potential differences between our We observed kit fox dens, a pack of coyotes including findings and those of previous researchers and would juveniles, and four common raven nests in our study area clarify the ecological role of the carcasses of migratory (TP, unpubl. data). Each of these species was breeding birds in migration corridors. during our study; coyotes March–May (Webb et al., 2004), Though remote cameras enabled us to document some kit foxes March–September (Zoellick et al., 1989), and scavenging events we would not have seen otherwise, the common ravens May–July (Smith et al., 1981). Thus, we cameras failed to detect all scavengers. In these cases we speculate that scavenged migratory bird carcasses were speculate carcasses were scavenged in one of two ways. likely provided to offspring in each of these species. We First, based on our observations of common ravens in also observed an emaciated kit fox, a coyote with only flight dropping to the ground and then immediately three legs, and a common raven with a broken mandible. flying up again with a scavenged migratory bird carcass, we 546 The Southwestern Naturalist vol. 59, no. 4 suggest that in some instances scavenging events might DEVAULT,T.L.,AND A. R. KROCHMAL, 2002. Scavenging by snakes: have occurred more quickly than our cameras were an examination of the literature. Herpetologica 58:429–436 capable of capturing photographs. Second, during tran- DWYER,J.F.,AND K. DOLOUGHAN. 2013. Testing systems of avian sects we also regularly encountered desert iguanas perch deterrents on electric power distribution poles. Human–Wildlife Interactions 7:39–54. (Dipsosaurus dorsalis) which occasionally scavenge carrion DWYER,J.F.,AND R. W. MANNAN. 2007. Preventing raptor (Norris, 1953). Because our cameras operated by detect- electrocutions in an urban environment. Journal of Raptor ing differences in temperature, and because iguanas are Research 41:259–267. ectothermic species, we suggest that if reptiles like desert FAANES, C. A. 1987. Bird behavior and mortality in relation to iguanas scavenged carcasses (DeVault and Krochmal, power lines in prairie habitats. Fish and Wildlife Technical 2002) our cameras might not have detected them. We Report 7. United States Department of the Interior, do not know if both, either, or neither of these hypotheses Publications Unit, Fish and Wildlife Service, Washington, are correct. Nevertheless, because scavenged carcasses D.C. were consumed diurnally and nocturnally, and human HERNANDEZ´ , L., M. DELIBES, AND F. HIRALDO. 1994. Role of reptiles and arthropods in the diet of coyotes in extreme desert areas surveys for carcasses were largely conducted diurnally, our of northern Mexico. Journal of Arid Environments 26:165– study supports prior assertions that failure to account for 170. removal of carcasses by scavengers might bias studies HERNANDEZ´ , L., R. R. PARMENTER,J.W.DEWITT,D.C.LIGHTFOOT, AND toward lower estimates of mortality (Dwyer and Mannan, J. W. LAUNDRE´. 2002. Coyote diets in the Chihuahuan Desert, 2007; Ponce et al., 2010), particularly when a diverse suite more evidence for optimal foraging. Journal of Arid of predators are active throughout the day and night. Environments 51:613–624. Scavenging events tended to happen relatively soon after LONGCORE, T., C. RICH,P.MINEAU,B.MACDONALD,D.G.BERT,L.M. sunrise as ravens flew along the length of the power line SULLIVAN,E.MUTRIE,S.A.GAUTHREAUX,JR., M. L. AVERY,R.L. right-of-way. Because most scavenging events occurred CRAWFORD,A.M.MANVILLE II, E. R. TRAVIS, AND C. DRAKE. 2013. Avian mortality at communication towers in the United States nocturnally or relatively soon after sunrise, our data also and Canada: which species, how many, and where? Biological suggest that survey time period strongly influences Conservation 158:410–419. detection probability. The later in the day that surveys MATLEY, J. K., R. E. CRAWFORD, AND T. A. DICK. 2012. Observation of occur, the greater the likelihood that carcasses are common raven (Corvus corax) scavenging Arctic cod (Bor- scavenged before surveys begin. eogadus saida) from seabirds in the Canadian High Arctic. Our study was relatively limited in scope, including Polar Biology 7:1119–1122. only a single spring migration in an area where spring MCKINNEY,T.,AND T. W. S MITH. 2007. Diets of sympatric bobcats and fall avian migrations occur annually. Future research and coyotes during years of varying rainfall in central comparing multiple years of surveys, and comparing . Western North American Naturalist 67:8–15. surveys during spring and fall migration, would likely MOEHRENSCHLAGER, A., R. LIST, AND D. W. MACDONALD. 2007. Escaping intraguild predation: Mexican kit foxes survive reveal additional species of migratory birds consumed by while coyotes and golden eagles kill Canadian kit foxes. scavengers as well as potential differences in the species Journal of Mammalogy 88:1029–1039. composition of scavengers and migratory bird carcasses by MURPHY, R. K., S. M. MCPHERRON,G.D.WRIGHT, AND K. L. season. SERBOUSEK. 2009. Effectiveness of avian collision averters in preventing migratory bird mortality from powerline strikes in We thank R. Abe and E. Gilbreath for assistance with data the central Platte River, Nebraska. Final Report to the U.S. collection, A.M. Dwyer, D. Eccleston, and R. E. Harness for Fish and Wildlife Service, Grand Island, Nebraska, USA. comments on an early draft of this work, and D. P. Ordonez˜ for NEWTON, I. 2008. The migration ecology of birds. Academic translating our abstract into Spanish. Press, London, U.K. NORRIS, K. S. 1953. The ecology of the desert iguana Dipsosaurus LITERATURE CITED dorsalis. Ecology 34:256–287. PATTEN, M. A., G. MCCASKIE, AND P. U NITT. 2003. Birds of the BARRIENTOS, R., C. PONCE,C.PALACIN´ ,C.A.MARTIN´ ,B.MARTIN´ , AND Salton Sea: status, biogeography, and ecology. University of J. C. ALONSO. 2012. Wire marking results in a small but California Press, Berkeley, California. significant reduction in avian mortality at power lines: a BACI PONCE, C., J. C. ALONSO,G.ARGANDONA˜ ,A.G.FERNANDEZ´ , AND M. designed study. PLoS ONE 7:e32569. CARRASCO. 2010. Carcass removal by scavengers and search BOARMAN,W.I.,AND B. HEINRICH. 1999. Common raven (Corvus accuracy affect bird mortality estimates at power lines. corax). The birds of North America online (A. Poole, editor). Animal Conservation 13:603–612. Ithaca: Cornell Laboratory of . Available at: SCHMALJOHANN,H.,AND V. DIERSCHKE. 2005. Optimal bird http://bna.birds.cornell.edu/bna/species/476. Accessed 11 migration and predation risk: a field experiment with July 2013. northern wheatears. Journal of Animal Ecology 74:131–138. BOWLIN, M. S., I.-A. BISSON,J.SHAMOUN-BARANES,J.D.REICHARD,N. SCHWARZER, A. C., J. A. COLLAZO,L.J.NILES,J.M.BRUSH,N.J. SAPIR,P.P.MARRA,T.H.KUNZ,D.S.WILCOVE,A.HEDENSTROM¯ , DOUGLASS, AND H. F. PERCIVAL. 2012. Annual survival of red C. G. GUGLIELMO,S.A˚ KESSON,M.RAMENOFSKY, AND M. WIKELSK. knots (Calidris canutus rufa) wintering in Florida. Auk 2010. Grand challenges in migration biology. Integrative and 129:725–733. Comparative Biology 50:261–279. SILLET, T. S., AND R. T. HOLMES. 2002. Variation in survivorship of a December 2014 Rogers et al.—Scavenging of migratory bird carcasses 547

migratory songbird throughout its annual cycle. Journal of WEBB, W. C., W. I. BOARMAN,J.T.ROTENBERRY. 2004. Common Animal Ecology 71:296–308. raven juvenile survival in a human-augmented landscape. SMITH, G. J., J. R. CARY, AND O. J. RONGSTAD, 1981. Sampling Condor 106:517–528. strategies for radio-tracking coyotes. Wildlife Society Bulletin WESTERN REGIONAL CLIMATE CENTER. 2013. Southern California 9:88–93. climate summaries. Desert Research Institute, Reno, , SPORER, M. K., J. F. DWYER,B.D.GERBER,R.E.HARNESS, AND A. K. USA. Available at: http://www.wrcc.dri.edu/climate-maps/. PANDEY. 2013. Marking power lines to reduce avian collision Accessed 11 July 2013. near the Audubon National Wildlife Refuge, North Dakota. WHITE, P. J., K. RALLS, AND C. A. VANDERBUILT WHITE. 1985. Overlap Wildlife Society Bulletin: doi:10.1002/wsb.329. in habitat and food use between coyotes and San Joaquin kit STRANDBERG, R., R. H. G. KLAASSEN,M.HAKE, AND T. A LERSTAM. foxes. Southwestern Naturalist 40:342–349. 2010. How hazardous is the Sahara Desert crossing for ZOELLICK, B. W., N. S. SMITH, AND R. S. HENRY. 1989. Habitat use migratory birds? Indications from satellite tracking of and movements of desert kit foxes in western Arizona. raptors. Biology Letters 6:297–300. Journal of Wildlife Management 53:955–961. SUTHERLAND, W. J. 1996. Predicting the consequences of habitat loss for migratory populations. Proceedings of the Royal Submitted 21 October 2013. Society of London. Series B: Biological Sciences 263:1325– Acceptance recommended by Associate Editor, M. Clay Green, 28 April 1327. 2014.