—— / Raptor Res. 26(3);146-151 © 1992 The Raptor Research Foundation, Inc. RAPTOR PREDATION ON ROCK PTARMIGAN THE CENTRAL IN {Lagopus mutus) CANADIAN ARCTIC Richard G. Cotter^ and David A. Boag^ Department ofZoology, University ofAlberta, Edmonton, AB, Canada T6G 2E9 Christopher C. Shank Department ofRenewable Resources, Government of the Northwest Territories, Yellowknife, NWT, Canada XIA 2L9 We Abstract. studied the nature and timing of raptor predation on a population of Rock Ptarmigan {Lagopus mutus) in the central Arctic of the Northwest Territories from 1987 to 1989. The density of Rock Ptarmigan at Windy Lake declined during the 3 years of the study from 5.8 to 3.5 to 2.8 birds per 100 ha. The number of nesting pairs of raptors within 10 km of Windy Lake also declined during this period from 9 to 10 to 5. Nesting raptors included Gyrfalcons {Falco rusticolus), Peregrine Falcons {Falco peregrinus) and Golden Eagles (Aquila chrysaetos). Common Ravens {Corvus corax) were also present. During the breeding seasons of 1988 and 1989, 15% of the adult breeding population of Rock Ptarmigan were depredated. There was no sex or age bias among Rock Ptarmigan killed, but a temporal bias was recorded over the course of the breeding season. Predation was the proximate cause of death for 917o of Rock Ptarmigan found dead over the three summers. Falcons accounted for 957o of all predation, the majority likely by Gyrfalcons. Depredacion por raptoras que victiman a las de la especie Lagopus mutus en el Artico central canadiense Extracto. Hemos estudiado la naturaleza y epoca de la depredacion causada por raptoras en una poblacion de Lagopus mutus en el Artico central de los Territorios del Noroeste canadiense, desde 1987 a 1989. La densidad de estas raptoras en Windy Lake decline durante los tres aiios de estudio desde 5.8 (1987), a 3.5 (1988), y a 2.8 (1989) aves por 100 hect. El numero de parejas raptoras nidificantes dentro de 10 km de Windy Lake tambien decline durante este periodo desde 9 (1987), a 10 (1988), a 5 (1989). Las raptoras nidificantes incluyeron especies tales como Falco rusticolus, F. peregrinus, y Aquila chrysaetos. Los Corvus corax tambien estuvieron presentes. Durante las estaciones de reproduccion de 1988 y 1989, el 157o de la poblacion reproductora de las L. mutus fue depredada. No hubo parcialidad por sexo 6 edad entre las L. mutus que fueron muertas, pero si se registro una parcialidad temporal durante el curso de la estacion reproductora. La depredacion fue la causa mas probable para la muerte de 917o de L. mutus, a las que se las encontro muertas en el curso de tres veranos. Los halcones contribuyeron con el 957o de toda la depredacion, la que, al parecer, en su mayoria fue causada por F. rusticolus. [Traduccion de Eudoxio Paredes-Ruiz] In most predator-prey studies, including those of mammals and raptors. Although the respective stud- raptors and grouse, emphasis has been placed on ies provided much information on the diet of Gyr- either the predator or the prey. For example, in falcons (Poole and Boag 1988) and levels and pat- northern Canada, Poole (1987) studied the ecology terns ofpredation on ptarmigan (Hannon and Gruys and food habits of Gyrfalcons {Falco rusticolus), a 1987), neither study examined the numerical rela- We major predator of ptarmigan {Lagopus spp.), where- tionship between predators and prey. report on as Hannon and Gruys (1987) studied the impact of the nature and timing of predation on Rock Ptar- predation on Willow Ptarmigan {L. lagopus) by migan in the same area where the diet and feeding behavior of Gyrfalcons (Poole and Boag 1988) and the interrelationships within the raptor guild were reported (Poole and Bromley 1988). Our specific ’ Present address: Richard C. Cotter, 8506d 169 St., Ed- monton, AB, T5R 2W9. objectives were to determine numerically which ^ Present address: David A. Boag, 6746 Amwell Drive, predators were involved in ptarmigan predation, their R.R. Brentwood Bay, BG, VOS lAO. relative impact, the relative vulnerability of different 1, 146 ^ September 1992 Raptor Predation on Rock Ptarmigan 147 age and sex classes of Rock Ptarmigan, and the tim- Table 1. Number of nesting pairs of raptors within 10 ing of predation relative to their annual cycle. We km of Windy Lake, Northwest Territories, in 1987-89. also examined the extent to which differences in these variables were related to seasonal and annual Species 1987 1988 1989 differences. Gyrfalcon 3 4 2 Study Area Peregrine Falcon 3 3 2 The study area, centered around Windy Lake, North- Golden Eagle 1 1 0 west Territories (68“05'N 106°40'W), lies near the center Raven^ 2 2 1 of the Kilgavik study area used by Poole (1987) in his Total 9 10 5 study of Gyrfalcons. The area consists of gently rolling tundra punctuated by small rock outcrops. The main geo- ® Considered as a “functional raptor” (White and Cade 1971) logical features are granitic intrusions, and diabase dykes and sills (Fraser 1964). Lines of cliffs and blocks of rock confirmed by the presence of either leg bands or radio at rising abovethe surrounding terrain provide nesting ledges a nearby Gyrfalcon nest site or plucking post. Birds with for Gyrfalcons, Peregrine Falcons {Falco peregrinus). transmitters that were killed by predators were found and Golden Eagles {Aquila chrysaetos) and Common Ravens their status documented. The fate of males without trans- (Corvus corax), the major avian predators of adult ptar- mitters could not be ascertained in all cases, and the re- migan (Cramp and Simmons 1980, Poole 1987). For a corded number of males preyed upon during the breeding description on the flora, fauna and climates of the region season is, therefore, a minimum. Predation rate was cal- see Poole and Bromley (1988). culated as the percent ofresident males or females, present at the onset ofthe breeding season, that were subsequently Methods lost to predators. We surveyed all known and potential raptor nest sites Survival rates of banded ptarmigan were calculated us- within the Kilgavik region studied by Poole and Bromley ing the Kaplan-Meier (K-M) product limit method, as (1988) each spring (May-early June) in 1987-89 to de- modified by Pollock et al. (1989) for staggered entry of termine occupancy status. Particular attention was given animals. Survival rates were calculated on a weekly basis. to the area within a 10 km radius of Windy Lake where, The K-M method calculates a survival function (S[f]) on two nearby study areas, most Rock Ptarmigan were which is “the probability of an arbitrary animal in a captured and marked. Rock Ptarmigan were captured at population surviving t units of time from the beginning of the onset ofthe breeding season (early June) using a noose the study” (Pollock et al. 1989). Ptarmigan, whose fate pole or ground nets (Hannon 1983). Sex and age were was known only up to a certain date, were eliminated from recorded for all individuals. Birds were classed as adult analysis fromthat dateon (White and Garrot 1990). When (>;1 yr old) or yearling (<1 yr old) based on the pigmen- dates of last observation and documented death spanned tation of the eighth and ninth primaries (Parker et al. several days, amedian date wasused. Statistical differences 1985). All birds were fitted with four color-coded bands; in the survival function between age and sex groups were one numbered aluminum and three numbered plastic (Na- examined out using the most conservative of the three tional Band and Tag Co., Newport, KY). “approximate chi-square” tests presented by Pollock et al. In 1988 and 1989 all female Rock Ptarmigan were fitted (1989). with 12-g “necklace” radio transmitters (Biotrack, Ware- Data sets of observed frequencies were compared using 2x2 ham, U.K.). Only half of the males were fitted with radio contingency analyseswith the G-test ofindependence transmitters; in 1988 with 18-g “backpack” radio trans- and Williams Correction (Sokal and Rohlf 1981). The mitters (Telemetry Systems, U.S.A.) and in 1989 with significance level was P < 0.05 for all tests. 12-g “necklace” transmitters. In 1987 work on ptarmigan ceased after banding was Results completed in mid-June. In 1988 and 1989 the fate ofadult The number of raptors nesting within 10 km of males and females was followed through to the end of each breeding season by regularly surveying (every 3-4 d) their Windy Lake declined from 1987-89 (Table 1). Fal- respective territories. If an individual ptarmigan was not cons comprised 71% of all nesting raptors with an immediately visible, it was located by radio telemetry, or, almost equal number of Gyrfalcons and Peregrine inthecaseofmales notradio-marked, by intensive searches Falcons (Table 1). There was also a decline in ptar- of its territory. migan densities per 100 ha between 1987 and 1989; When a ptarmigan was killed, we attempted to identify the predator from the location and nature ofthe ptarmigan from 5.8 in 1987 to 3.5 in 1988 and to 2.8 birds in remains. All remains were examined closely when the 1989. identity of the predator was not obvious at the kill site; a Between June 1987 and July 1989, 96% (123/ description of the location, date, and condition of the re- 128) of resident Rock Ptarmigan were marked and mains was recorded for comparison with descriptions in released, the overall sex ratio favoring males (1:0.83; the literature for potential predator species involved (Ei- narsen 1956, Jenkins et al. 1964, Nielsen 1986). In some Table 2). The survival function (S[^]) of these birds May cases of presumed Gyrfalcon predation, identification was over the breeding season (29 to 31 July), for 148 Cotter et al. VoL. 26, No. 3 • MALES 0.6 S' —Cc33D “Cc13D3 “CC33D “CC33D “33 3 —33 —33 O(M) m OJ C33 <CMo CO 1^ CM Figure 1. The Kaplan-Meier survival function (modified for staggered entry of animals) of male and female Rock Ptarmigan over the breeding seasons of 1988 and 1989 at Windy Lake, Northwest Territories. 1988 and 1989 combined, was 0.82. There were no son, all years combined, was 14% (6% in 1987, 15% differences between the sexes (Fig. 1; females 0.87, in 1988, and 14% in 1989). Predation rate did not males 0.78, ^ 0.08, P > 0.95), or between age differ among years (G = 0.61, df = 2, T = 0.74). classes (adults 0.81, yearlings 0.84, = 0.04, P = Based on data from 1988 and 1989, more Rock = 0.84). Eighteen percent of birds (23/128) resident Ptarmigan were killed in June {N 14) than in for one or more breeding seasons were found dead. July {N = 2) (G = 10.35, df = 1, P < 0.01). There Mortality over the breeding season must be consid- were no differences between the proportion of males ered minimal because the fate of some birds was not and females killed in either June (G = 0.50, P = known at the end of the season. Assuming that all 0.48) or July (G = 2.46, P = 0.12). A comparison ofthese birds died, either during the breeding season of all known raptor predation events from the breed- or before returning the following spring, annual ing seasons of 1988 and 1989 revealed no difference mortality would rise to 57%, somewhat higher among (G = 0.01, df = 2, P = 0.92) between males (15%, N = N = females (61%) than males (50%). 61) and females (147o, 49). There was Predation was the proximate cause of death in 21 also no difference in the percentage of adults (14%, N = N = of 23 ptarmigan found dead over the three summers. 63) and yearlings (147o, 42) killed during A = predator was identified for all 21 depredated Rock the breeding seasons of 1988 and 1989 (G 0.00, Ptarmigan found on the study area. All but one of P > 0.95), and this was true for both males (G = the birds were killed by falcons (Table 3). The other, 0.24, P = 0.62) and females (G = 0.10, P = 0.75). a female, was taken by either Red {Vulpes vulpes) or Arctic fox (Alopex lagopus). During the study, falcons killed an equal number of males (10) and Discussion females (10). Seventeen of the 21 kills occurred dur- The density of Rock Ptarmigan at the onset of ing the breeding season, 1 occurred in autumn, and the breeding season at Windy Lake declined over 3 occurred between mid-summer and winter (in the course ofthis 3-year study. In the larger Kilgavik 1987). The predation rate during the breeding sea- region, the pattern was similar, with the decline September 1992 Raptor Predation on Rock Ptarmigan 149 commencing in 1986 after 3 consecutive years of Table 2. Number of resident Rock Ptarmigan at the population increase (Poole and Boag 1988, Cotter onset (approximately 1 June) of the breeding season at 1991). The number of nesting raptors in the vicinity Windy Lake, Northwest Territories. A difference in num- of Windy Lake declined from 1987-89 as well. The ber of males and females indicates unpaired individuals decline in numbers of breeding raptors, which co- Enclosed in parentheses are percentage yearlings. incided with that of the Rock Ptarmigan, suggests Total that these raptors assess food resources early in the season before beginning the reproductive cycle. Size of Density® Study Site Fe- (birds/ However, Shank and Poole (1992) noted that in the Year (ha) Males MALES 100 ha) larger Kilgavik region the number of productive Gyrfalcon pairs declined only slightly during those 1987 310 9 9 5.8 years, from 11 in 1987 to 9 in 1988, and with no (33) (67) change from 1988 to 1989. This discrepancy between 1988 1720 33 28 3.5 Windy Lake and the larger Kilgavik region makes (42) (46) it difficult to understand whether this population of 1989 1695 28 21 2.8 Gyrfalcons is responding to the ehanges in numbers (7) (70) of breeding ptarmigan. Court (1986) suggests that ^8% (10/128) of resident Rock Ptarmigan were outside of area Peregrine Falcons often will hold territory early in used for calculating density. the season but abandon it shortly thereafter and apparently forego reproduction in that year. Poole season, falcons have equal success in taking male and Bromley (1988) observed among Gyrfalcons at and female ptarmigan. Kilgavik that the percentage of territorial pairs to There was a temporal bias in levels of predation initiate laying ranged between 64 and 91%; however, on Rock Ptarmigan over the course of the breeding regardless of the annual variation there were no season, with most predation having oceurred in June differences between years in either mean brood size or productivity. The fact that those birds which did and very little in July. There are few data from other studies on relative levels of predation during nest and produce young fledged normal numbers of the breeding season; nevertheless, Nielsen (1986) chicks suggests that the precision with which food and Poole (1987) both reported seasonal variations resources for breeding is measured is high. in the apparent vulnerability of ptarmigan to Gyr- It has been suggested that different mortality rates falcon predation. Nielsen (1986) observed two sea- observed in male and female Rock Ptarmigan (Wee- sonal peaks in vulnerability, the first during the ter- den and Theberge 1972) are a consequence of dif- ritorial stage in the early part ofthe breeding season. ferences in conspicuousness of the sexes during the incubation phase of the breeding season (Bergerud and Mossop 1984). Prior to hatching, which in the Table 3. Number of known depredations on male and Kilgavik region occurs in early July (Cotter 1991), female Rock Ptarmigan at Windy Lake, Northwest Ter- females have been cryptic (both in plumage and ritories, in 1987-89. behavior) for a period of 3-4 wk, whereas males have been cryptic for only a week or so (MacDonald Predator 1970). In this study there was no sex bias in ptar- migan mortality over the course of the breeding sea- Pere- sons of 1988 and 1989, and this was true on a month- Gyr- grine Falcon Fox ly basis (June and July) as well. Poole (1987) found Year Sex falcon Falcon SPP.® SPP.*^ similar proportions of each sex among prey remains 1987 Male 1 0 0 0 taken from Gyrfalcon nest sites at Kilgavik during Female 1 0 2 0 May-June and July-August of 1984-87. The lack 1988 Male 1 1 2 0 of any sex bias in ptarmigan predation during the Female 1 0 4 1 breeding season at Windy Lake and Kilgavik, par- 1989 Male 3 0 2 0 ticularly prior to hatching, indicates that in spite of Female 0 0 2 0 differences in conspicuousness (and, therefore, per- “ Killed by either a Gyrfalcon or a Peregrine Falcon. haps vulnerability) of the sexes during the breeding Killed by either a Red or an Arctic fox. 150 Cotter et al. VoL. 26, No. 3 and the second after the territorial system has col- could not be classified to species responsible. Even lapsed and chicks are capable of sustained flight, so, given that seven of eight kills, for which the commencing in late summer and lasting into early species of falcon was known, were by Gyrfalcons winter. The period in between coincides with a pe- suggests that most of the falcon kills were, in fact, riod in which ptarmigan are much more secretive by Gyrfalcon. In the central Canadian Arctic, there- and cryptic (MacDonald 1970). Poole (1987) re- fore, Gyrfalcons would seem to be the most impor- ported a similar seasonal variation in Rock Ptar- tant predator ofRock Ptarmigan during the breeding migan use by Gyrfalcons at Kilgavik. He also ob- season (see also Bergerud and Mossop 1984, Nielsen served an inverse correlation between the percentage 1986, Poole 1987, and Gardarsson 1988). of ptarmigan in the diet of Gyrfalcons and the avail- ability of juvenile Arctic Ground Squirrels (Sper- Acknowledgments mophilus parryi). At Windy Lake, juvenile ground We acknowledge the inspiration received from Fred squirrels emerged in early July (Poole 1987), and and Fran Hamerstrom whose work with grouse and rap- their emergence coincided with a sudden drop in the tors laid the foundations for many a study undertaken proportion of ptarmigan among the collected prey by their students and colleagues. We thank our many field remains. Thus predation pressure on ptarmigan may assistants, in particular Jay VanderGaast, and the Re- newable Resources staff in Cambridge Bay who helped reflect not only the relative vulnerability of other make this study possible. We wish to acknowledge the potential prey, but also their own behavior and cryp- invaluable assistance ofJimmy Ayalik, John Nanuak, and ticity. Adult male and female Rock Ptarmigan, some Doug Stern ofCambridge Bay and Umingmaktok fortheir ofwhich had previous breeding experience at Windy field assistance and sharing of knowledge of the region. Jim Hare and Lome LeClair assisted with statistical anal- Lake, incurred losses similar to those of yearlings ysis, and Rogier Gruys helped in identifying the species breeding for the first time. This suggests that any ofpredator ofthe ptarmigan kills that were collected. This potential increased vulnerability of young birds is study would not have been possible without the logistical overcome by the time they return to the breeding and financial support from the Department of Renewable Resources, Government of the Northwest Territories. grounds for the first time. Further support was provided by a Natural Sciences and The importance ofRock Ptarmigan as prey within Engineering Research Council of Canada operating grant the arctic raptor community is evident, particularly to DAB and a Post-graduate Scholarship to RCC, a Polar to Gyrfalcons the range of which overlaps to a large Continental Shelf Project grant to CCS, and a Northern extent the holarctic distribution of ptarmigan and Science Training Grant (Canadian Circumpolar Institute) RCC. to the diet of which has a high percentage of ptarmigan (Bergerud and Mossop 1984, Poole and Boag 1988). Literature Cited At Kilgavik, adult Rock Ptarmigan are known to be preyed upon by Gyrfalcons, Peregrine Falcons and Bergerud, A.T. and D.H. Mossop. 1984. The pair Golden Eagles, and ptarmigan remains have been bond in ptarmigan. Can. J. Zool. 62:2129-2141. found in raven nests (Poole and Bromley 1988). Cotter, R.C. 1991. Population attributes and repro- Although ravens may occasionally kill adult ptar- ductive biology of Rock Ptarmigan (Lagopus mutus) in the central Canadian Arctic. M.Sc. thesis. University migan (Nielsen 1986), they also scavenge ptarmigan of Alberta, Edmonton, AB, Canada. from Gyrfalcon caches (Poole and Bromley 1988). Court, G.S. 1986. Some aspects of the reproductive It is not surprising that Golden Eagles were not biology oftundra Peregrine Falcons. M.Sc. thesis. Uni- important predators as they rely on other prey for versity of Alberta, Edmonton, AB, Canada. the bulk of their diet (Cramp and Simmons 1980, Cramp, S. and K.E.L. Simmons. 1980. Handbook of Poole and Bromley 1988). In the Kilgavik region, the birds ofEurope, the Middle East and North Africa. Gyrfalcons and Peregrine Falcons comprised over The birds of the Western Palearctic. Vol. 2. Oxford half the population of nesting raptors within 10 km University Press, U.K. of the Windy Lake study sites. It is, therefore, not Einarsen, A.S. 1956. Determination of some predator surprising that these two falcons accounted for all species by field signs. Oregon State Monographs 10:1- 34. Rock Ptarmigan killed by avian predators during Fraser, J.A. 1964. Geological notes on northeastern the breeding season. It was not always possible to District of MacKenzie, Northwest Territories. Paper determine which falcon had made the kill from the 63-40. Map 45-1963, Geol. Surv. Can., Ottawa, ON, remains found since both species handle prey in sim- Canada. ilar fashion (Jenkins et al. 1964, Cramp and Sim- Gardarsson, A. 1988. Cyclic population changes and mons 1980). Thus more than half the falcon kills some related events in Rock Ptarmigan in Iceland. September 1992 Raptor Predation on Rock Ptarmigan 151 Pages 300-320 in A.T. Bergerud and M.W. Gratson and foraging characteristics ofGyrfalcons inthecentral [Eds.], Adaptive strategies and ptopulation ecology of Canadian Arctic. M.Sc. thesis. University of Alberta, Northern Grouse. Vol. 1. University of Minnesota Edmonton, AB, Canada. Press, Minneapolis, MN. AND D.A. Boag. 1988. Ecology of Gyrfalcons, Hannon, S.J. 1983. Spacing and breeding density of Falco rusticolus, in the central Canadian Arctic: diet Willow Ptarmigan in resfxmse to an experimental al- and feeding behavior. Can. Zool. 66:334-344. J. teration of sex ratio. /. Anim. Ecol. 52:807-820. AND R.G. Bromley. 1988. Interrelationships AND R.C. Gruys. 1987. Patterns of predation within a raptor guild in the central Canadian Arctic. in a Willow Ptarmigan population in Northern Can- Can. J. Zool. 66:2275-2282. ada. Proc. 4th Int. Grouse Symp., pages 44-50. Shank, C.C. and K.G. Poole. 1992. Status of Gyrfal- Jenkins, D., A. Watson and G.R. Miller. 1964. Pre- cons in the Northwest Territories, Canada. Proc. 4th dation and Red Grouse populations. J. Appl. Ecol. 1: World Conference on Birds of Prey and Owls, Berlin, 183-195. Germany. MacDonald, S.D. 1970. The breeding behavior of the SOKAL, R.R. AND F.J. Rohlf. 1981. Biometry. W.H. Rock Ptarmigan. Living Bird 9:195-238. Freeman and Co., New York. Nielsen, O.K. 1986. Population ecology of the Gyrfal- Weeden, R.B. and J.B. Theberge. 1972. The dynam- con in Iceland with comparative notes on the Merlin ics of a fluctuating population of Rock Ptarmigan in and the Raven. Ph.D. thesis. Cornell University, Ith- Alaska. Proc. 15th Int. Ornithol. Congr., pages 90- aca, NY. 106. Parker, H., H. Ottensen and E. Knudsen. 1985. Age White, C.M. and T.J. Cade. 1971. Cliff-nesting rap- determination in Svalbard Ptarmigan (Lagopus mutus tors and ravens along the Colville River in arctic Alas- hyperboreus). Polar Res. 3:125-126. ka. Living Bird 10:107-150. Pollock, K.H., S.R. Winterstein, C.M. Bunck and White, G.C. AND R.A. Garrot. 1990. Analysisofwild- P.D. Curtis. 1989. Survival analysis in telemetry life radio-tracking data. Academic Press, Inc., Boston, studies: the staggered entry design. J. Wildl. Manage. MA. 53:7-15. May Poole, K.G. 1987. Aspects of the ecology, food habits Received 2 April 1992; accepted 31 1992