Saker Falcon diet: the implications of habitat change Mark Watson and Roger Clarke Bruce Pearson

ABSTRACT A study of the Saker Falcon Falco cherrug in northeast in 1997 revealed a diet dominated by susliks , but including some birds, mainly gamebirds (Galliformes) and crows (Corvidae).Traces of prey remains from the previous season included smaller of that vary greatly in abundance from year to year. Saker breeding success was lower in those areas with taller vegetation, which held fewer susliks.With the collapse of collective farming in Kazakhstan, nomadic herders no longer graze large areas of and alps, and the onset of vegetation succession appears to have reduced the abundance of susliks in some areas. Most -species declines are due to agricultural intensification, but here is an example of negative conservation implications of the abandonment of grazing across large areas of Kazakhstan. If widespread, these land-use changes, together with pressure from falconers, could threaten total numbers of Saker Falcons.

he breeding range of the Saker Falcon of southern and Kazakhstan, Falco cherrug encompasses parts of the plateaux of ,Tibet, and west and TAustria, the Czech Republic, Slovakia, central China, and extends to the edge of the , the former Yugoslavia, Romania, Great Plain of eastern China. Small outlying Bulgaria and , stretches across the populations breed to the south, in Turkey

136© British Birds 93: 136-143, March 2000 Watson & Clarke: Saker Falcon diet and Iran. Six were recognised by period, to monitor prey deliveries and to Baumgart (1991), but Pfeffer (1992) and corroborate the result of the pellet analysis. White (1995) drew attention to possible tax- Pellet contents and prey remains were onomic revisions. identified by comparing them with collec- The World population of Sakers has been tions. We identified hair by using estimated at 35,000-40,000 pairs (Baumgart microscopy to see the diagnostic patterns of 1991), and in Europe this falcon has been the medulla (Teerinck 1991), and bird classed as endangered (Tucker & Heath feathers from the structure of the downy 1994). Sakers eat mainly mammals, especially barbules at the base (Brom 1986). susliks (ground-) Spermophilus, Quantification of prey was on the basis: which are abundant on short vegetation in detection in a pellet or as remains = 1 indi- steppe grassland across much of the Western vidual. There are two weaknesses in this, Palearctic (Cramp & Simmons 1980; Nowak especially for pellets found at nests. First, a 1991; Tucker & Heath 1994). In Kazakhstan, single prey individual may occur in all the human depopulation has resulted in the pellets from a brood; and, secondly, remains abandonment of grazing in many areas. from more than one prey individual of the Succession has caused some of these areas to same species may be indistinguishable in a become unsuitable for susliks. Sakers are single pellet.These weaknesses are, however, already threatened by the illegal trade in birds common to the other studies that we used for falconry,so the impact of land-use change, for comparison. which reduces the abundance of its main Diet-diversity indices 1 were calculated for prey, could be significant. Three recent our own study, and for four others for com- studies, in Austria (Frey & Senn 1980), parison. In our analysis, prey were assigned Hungary (Bagyura et al. 1994) and Slovakia to one of three classes: mammals, birds and (Obuch & Chavko 1997), showed, however, reptiles. Beetles (Coleoptera) occurred in 12 that some Sakers could breed successfully by (9%) of the pellets, but were ignored as feeding mainly or exclusively on birds. inconsequential to diet. Suslik abundance was measured in the Study areas and methods nine study plots. A simple index was calcu- Fieldwork was carried out in northeast Kaza- lated, scoring 1 for each sighting, call and khstan from mid April to the end of June in active burrow on up to six 400-m transects. 1997. Nine study plots were grouped in two Three of the transects were walked in areas areas, one in the Alakol Basin and one in the where susliks were apparent, and three were foothills of the Altai Mountains.The Alakol selected at random. Resulting total scores for area was hot and arid, with a rocky substrate, each area were divided by the total distance while the Altai area was wetter and sub- of transects to give an index of suslik abun- alpine in character. The study plots were dance per 100 m of transect. remote from permanent human habitation For each study area, suslik habitat quality and had traditionally been subject to sea- was scored on a scale of 1-9, low scores sonal grazing by horses, sheep and cattle. reflecting long vegetation and poor habitat This transhumance had ended in both areas and high scores indicating short vegetation by 1995. and good habitat. Since there was insuffi- Breeding success was calculated as the cient time to do this systematically at ran- mean number of young fledged per pair that domised points, it was based on a subjective laid (productivity). During each nest visit, general assessment within a radius of 2 km pellets and prey remains were collected from of each nest site. the nest itself and the ground below, as well Lastly,Saker breeding success, suslik abun- as from nearby plucking posts and resting dance and habitat quality in the nine plots places. One nest was watched for a 24-hour were examined using regression analysis.

1 2 Using Levins’ (1968) formula: diet diversity = 1/∑pi , where pi is the proportion of the diet in prey class i.This measures both diversity and evenness of diet, in that it gives values in the range of 1 to n, where n is the number of prey classes in the diet. High values (approaching n) indicate a broader diet, more evenly divided among the prey classes.

British Birds 93: 136-143, March 2000 137 Watson & Clarke: Saker Falcon diet

Table 1. Prey items identified from pellets and remains at nests of Saker our ten pellet records of Falcons Falco cherrug in northeast Kazakhstan, April-June 1997. crows came from the same nest. The registration of Uneaten Saker feathers in one pellet prey Pellets remains simply reflected the fact MAMMALS that surviving chicks ate Red-cheeked Suslik Spermophilus erythrogenys 94 3 their siblings as broods Long-tailed Suslik Spermophilus undulatus 7 1 were reduced, presumably 1 Steppe lagurus 0 as a result of limited food Common Vole arvalis 0 1 Unidentified vole 2 – availability. Unidentified small mammal 1 0 The results of observa- Siberian Mole Rat Myospalax myospalax 0 1 tions of feeding behaviour Common Hamster Cricetus cricetus 0 1 at the nest watched for 24 Stoat Mustela erminea 0 2 Bobak Marmota bobac 0 8 hours are summarised in Artiodactyla sp. 1 0 table 2.The chicks were fed Unidentified mammals 3 6 five times in the period from 14.55 hours on 13th 24 Mammals subtotal 108 May to 15.30 hours on 14th (82%) (48%) May. Both parents were BIRDS hunting, but on two occa- Saker Falcon Falco cherrug 1 0 sions the male brought Black Grouse Tetrao tetrix 0 1 food and made a food-pass Chukar Partridge Alectoris chukar 0 2 Grey Partridge Perdix perdix robusta 0 7 to the female; as with other Common Quail Coturnix coturnix 0 1 raptors, it is the female that Unidentified gamebird 0 1 usually feeds the nestlings Rock Dove Columba livia 0 1 (Pfeffer 1990). It is possible 0 Common Cuckoo Cuculus canorus 1 that the male brought more European Bee-eater Merops apiaster 1 0 Calandra Lark Melanocorypha calandra 0 1 items than we recorded, as Sky Lark Alauda arvensis 0 1 food-passes could have Unidentified lark 1 – occurred out of sight. Pied Wagtail Motacilla alba 0 1 The nest observations Pied Wheatear Oenanthe pleschanka 0 2 Eurasian Jackdaw Corvus monedula 0 3 supported the findings of Carrion Crow Corvus corone 0 2 the pellet analysis, showing Unidentified crow 10 – reliance on susliks, but also Rosy Starling Sturnus roseus 0 1 that Sakers switch to other – Unidentified small passerines 4 species according to their Unidentified birds 3 2 availability. For example, in Birds subtotal 21 26 the canyons surrounding (16%) (52%) this nest, there were flocks REPTILES of up to 60 Eurasian Jack- Unidentified 2 (2%) 0 daws Corvus monedula TOTALS 131 50 containing many recently fledged juveniles that the falcons could catch easily. Diet Our sample included mammals of smaller Prey found in pellets and prey remains at body size than susliks, but species which nests in northeast Kazakhstan (table 1) raptors can find easily and catch in large showed that the diet was dominated by Red- quantities in irruption years (Sludskiy 1969). cheeked Susliks S. erythrogenys, but other- For example, we recorded the Steppe wise included a range of mammal and bird Lemming Lagurus lagurus and Common species. Second to susliks, the most frequent Vole Microtus arvalis in pellets, but only as prey appeared to be crows (Corvidae) and traces from unoccupied nests in one study gamebirds (Galliformes), although the latter area in the Altai; these were judged to be were evident only in remains, and seven of remains from the preceding season, but they

138British Birds 93: 136-143, March 2000 Watson & Clarke: Saker Falcon diet

Table 2. Frequency, duration and types of prey delivered to a brood of four female Saker chicks at about 20 days old, northeast Kazakhstan, May 1997.

Feed Food-pass from Day number Time (hrs) Item Fed by male

13th May 1 16.13-16.32 Suslik Female No 2 18.19-18.32 Suslik (juv.) Female Yes 14th May 1 08.34-08.51 Eurasian Jackdaw (juv.) Female No 2 10.12-10.27 Suslik (juv.) Female Yes 3 15.00-15.19 Suslik Female No

Table 3. Comparison of five studies of Saker Falcon Falco cherrug diet in the breeding season. For explanation of diversity index, see ‘Methods’.

NE Kazakhstan SE Kazakhstan Slovakia 1978- Hungary 1986-91 Austria (Frey & 1997 (this study) 1984-87 (Pfeffer 1995 (Obuch & (Bagyura et al. Senn 1980) 1990) Chavko 1997) 1994)

REMAINS PELLETS REMAINS PELLETS REMAINS Upland Lowland REMAINS Sample size 131 50 834 2,317 1,236 704 197 Percentages: MAMMALS 82 48 55 16 23 32 19 Susliks 77 8 22 6 21 30 11 Gerbils 0 0 29 0 0 0 0 BIRDS 16 52 45 84 77 68 81 Pigeons 0 2 ? 44 66 47 28 REPTILES 2 0 0 0 0 0 0

Diversity 1.41 2.00 1.98 1.37 1.55 1.78 1.50 index

Table 4. Productivity, habitat quality and suslik abundance by study area, northeast Kazakhstan, April-June 1997. For method of calculations of productivity, habitat score and suslik abundance, see ‘Methods’.

Number of Suslik abundance Study area breeding attempts Mean productivity Habitat score index ALAKOL DEPRESSION 3 1.00 7 2.84 A 2 1.00 5 1.56 B 1 0.00 4 1.30 C 1 4.00 7 0.42 D 3 2.00 8 4.58 E 1 1.00 7 2.26 F ALTAI FOOTHILLS G 1 2.00 5 1.25 H 0 0.00 1 0.29 J 4 3.00 9 3.00 show that Sakers switch to these species in the study area J (S. Starikov verbally). years when their numbers are high. Although We compared our results with those of no or voles were trapped or seen other studies (table 3). On the basis of in any of the study areas in 1997, in June pellets, our study revealed low diet diversity 1996, a point count of 27 Steppe Lemmings and the greatest dependence on susliks. had been made 400 m from a Saker nest in A similar study by Pfeffer (1990) recorded

British Birds 93: 136-143, March 2000 139 Watson & Clarke: Saker Falcon diet significant numbers 3.5 of gerbils Meriones y = 0.4794x + 0.2266 3 in Saker pellets R2 = 0.3816 from southeast 2.5 Kazakhstan; this 2 species is similar in 1.5 size to susliks. 1 Results based on 0.5 prey remains showed Saker productivity 0 the highest propor- tions of birds, and 0 1 2 3 4 5 those based on suslik abundance pellets showed the lowest.This was par- Fig. 1. Saker Falcon Falco cherrug breeding success and food: the relationship ticularly striking in between mean productivity per area and mean suslik abundance per area (P<0.05). our sample from northeast Kazakh- stan. 5 y = 0.3744x – 0.6493 Most of the 4 pellets we collected R2 = 0.4622 were from nest- 3 lings. Of 101 pellets 2 containing suslik 1 remains, 76% inc- luded claws, where- 0 as only 10% con- Saker productivity -1 0 2 4 6 8 10 tained the teeth or jaws, and not one suslik habitat quality contained a skull. Teeth were unlikely Fig. 2. Saker Falcon Falco cherrug breeding success and suslik habitat quality: the relationship between mean productivity per area and suslik habitat-quality to have been totally score (1 = long vegetation, low habitat quality; 9 = short vegetation, high digested, as verte- habitat quality)(P<0.05). brae and the ends of limb bones were commonly found in pellets. area where suslik abundance was measured. We concluded that adults generally removed This illustrates the difficulty of assessing the heads of susliks, and fed the rest to their hunting habitat over a large area in rugged young. terrain. In area H, where suslik abundance was Breeding success, suslik lowest and vegetation succession most abundance and habitat quality advanced after grazing had been abandoned, We compared Saker breeding success, suslik three pairs of falcons held territories but did abundance and suslik habitat quality across not breed. In 1996, three nests there had pro- study plots (table 4). Saker breeding success duced 14 young (I. S.Vorbiev verbally). was positively related to areas with high Seasonal variation in prey availability also suslik abundance (fig. 1), and in turn to areas affected diet composition, as susliks aestivate containing good habitat for susliks in terms and remain below ground in response to of short vegetation (fig. 2). Area D was high air temperatures and the dying-back of excluded from the analysis in fig. 1, as the green plants in midsummer in arid steppe suslik-abundance index for this area was very (Nowak 1991). In the Alakol area, suslik avail- low, at 0.42, yet the nest there had the ability was much reduced by the end of June highest productivity of four female young (daily suslik counts were 20% of those made and the earliest fledging date. The adults in April). By contrast, this was not a factor in were seen returning to the nest with susliks the Altai, where conditions were cooler and caught beyond a cliff-line and outside the wetter owing to higher altitude.

140British Birds 93: 136-143, March 2000 Watson & Clarke: Saker Falcon diet tefan Danko tefan ˘ S 88. Saker Falcons Falco cherrug, Hungary, May 1990.

DISCUSSION disproportionately shared among young. Nevertheless, pellet analysis may still give Bias in dietary studies some distortion of respective frequencies of Studies of prey remains rather than pellets prey types, because a less frequent item (in are biased towards larger birds. Some caution our study, birds) is more noticeable against a is therefore required in interpreting the background of a frequent prey item where results. Prey remains from larger birds parts of many individuals are harder to distin- include many feathers and bones, particu- guish (susliks). This, however, could mean larly from wings and legs, which are easier to only that birds were even less important find and identify than, say, fragments of than is shown by the pellet data. mammal hair. Our samples revealed few susliks among prey remains and many in The implications of pellets (table 1), reflecting the fact that habitat change mammals are more entirely consumed than are birds. Combining data from pellets and Our data show variation in Saker breeding remains is often recommended for assessing success related to suslik abundance and raptor diet, in order to counter the biases of habitat quality within a single year. Since it is each method (Simmons et al. 1991). We, known that all the study plots were exten- however, regard pellets as generally the more sively grazed up to 1995, the data provide reliable indicator of relative proportions of only circumstantial evidence of habitat prey taken by Sakers, where the main prey change due to differential succession types are fairly even in size and therefore not between areas with different climates and

British Birds 93: 136-143, March 2000 141 Watson & Clarke: Saker Falcon diet

substrates. If the effects of these land-use were represented in our sample only as changes are widespread and alternative prey traces in pellets that had survived from the are not available, Saker numbers could be previous year. Brandt’s Vole Microtus brandti much reduced in the future. To be certain has been recorded frequently in the Saker’s that a trend is occurring, data are clearly diet in Mongolia (Baumgart 1991), where required from several years. Such a process higher Saker productivity (Ellis et al. 1995) is, however, already documented for Bulgaria, could be explained by switching to this where cessation of grazing in the early 1990s species in irruption years. In Tibet, a similar caused declines in susliks and reduced Saker phenomenon has been observed whereby numbers to fewer than ten pairs (Bob Scott Sakers were observed to concentrate at high in litt.). Up to 1990, the Bulgarian population density to feed on Black-mouthed Pikas was 20-40 pairs (Tucker & Heath 1994). Ochotona melanostoma (P. J. Leader ver- Declines in suslik abundance may also bally); although small in body size, this affect other raptors. A single sample of five mammal was superabundant and easy for the Long-legged Buzzard Buteo rufinus pellets falcons to catch. Fledglings of ground-nesting from area J contained only remains of Red- birds such as larks (Alaudidae) also provide cheeked Susliks. Steppe Eagles Aquila high concentrations of prey at specific times nipalensis and Eastern Imperial Eagles A. of year; in late June, Sakers have been heliaca in Central Asia are also dependent on observed catching young larks that cannot susliks and, since these birds are larger, they fly simply by chasing them on the ground may require a much higher biomass of prey (A. Levin verbally). than do Sakers in order to breed successfully. As a generalist predator, the Saker Falcon In Bulgaria, Golden Eagles A. chrysaetos and may be well adapted to taking other prey in Long-legged Buzzards are believed to have response to reduced abundance and avail- declined as a result of habitat change ability of susliks. For example, in a study affecting the abundance of their main prey from Hungary,where loss of meadows to cul- (Bob Scott in litt.). tivation and other development has resulted Lemmings, voles and other species in widespread loss of habitat for European showing cyclical population fluctuations Susliks Spermophilus citellus, one pair of H. D. Brandl 89. European Suslik Spermophilus citellus.

142British Birds 93: 136-143, March 2000 Watson & Clarke: Saker Falcon diet

Saker Falcons fed its chicks mainly on racing nerable to human persecution. pigeons Columba livia (Bagyura et al. 1994). Usually, it is agricultural intensification Pigeons are also important for Sakers in the that causes declines in grassland-species Czech Republic and Slovakia, where susliks assemblages (Tucker & Heath 1994). This formed only 6% of prey remains collected small-scale study, however, provided circum- from the late 1970s to the mid 1990s, stantial evidence that abandoning agriculture whereas 44% were of domestic pigeons by stopping large-scale grazing in Kazakhstan (Danko et al. 1994; Obuch & Chavko 1997). may further affect a species that is already at In Austria, pigeons constituted 28% of prey risk from over-harvesting for falconry items identified by Frey & Senn (1980). (Watson 1997). Monitoring of Saker Falcon Domestic pigeons are slightly larger than numbers in relation to habitat change on a susliks in body size, with a published mean wider scale and over longer periods is essen- weight of 425 g (Ratcliffe 1993). Feeding on tial to determining the real impact of this racing pigeons could make Sakers more vul- phenomenon.

Acknowledgments Kalkvoralpen. Egretta 23: 31-38. Fieldwork was funded by The Environmental Research Levins, R. 1968. Evolution in Changing and Wildlife Development Agency, Abu Dhabi. We Environments. Princeton. thank Dr Nick Fox for making arrangements in Nowak, R. M. 1991. Walker’s Mammals of the World. Kazakhstan and for support in the UK. 5th edn. Baltimore. Obuch, J., & Chavko, J. 1997. The diet of the Saker References Falcon Falco cherrug in SW Slovakia. Buteo 9: 77-84. Bagyura, J., Haraszthy, L., & Szitta, T. 1994. Feeding Pfeffer, R. 1990. Saker Falcon. In Kovshar, A. F. (ed.), biology of the Saker Falcon Falco cherrug in Rare of Desert Regions. Kazak Academy of Hungary. In Meyburg, B.-U., & Chancellor, R. D. Sciences, Institute of Zoology. (eds.), Raptor Conservation Today. Mountfield. —— 1992. The Saker Falcon Falco cherrug in the Baumgart, W. 1991. Der Sakerfalke. Wittenberg Asiatic part of the former USSR. In Abstracts,IV Lutherstadt. World Conference on Birds of Prey,WWGBP. Brom,T.G. 1986. Microscopic identification of feathers Ratcliffe, D. 1993. The Peregrine Falcon. 2nd edn. and feather fragments of Palearctic birds. Bijdragen London. tot de Dierkunde 56: 181-204. Siegel, S., & Castellan, N. J. 1988. Nonparametric Brown, L., & Amadon, D. 1968. Eagles, Hawks and Statistics for the Behavioural Sciences.New York. Falcons of the World. Feltham. Simmons, R. E.,Avery, D. M., & Avery, G. 1991. Biases in Cade,T.J. 1982. The Falcons of the World. London. diets determined from pellets and remains: Chavko, J. 1995. Nesting of Saker Falco cherrug in correction factors for a mammal and bird-eating Slovakia in 1993 and 1994. Buteo 7: 175-181. raptor. J. Raptor Res. 25: 63-67. Cramp, S., & Simmons, K. E. L. (eds.) 1980. The Birds of Sludskiy, A. A. (ed.) 1969. Mammals of Kazakhstan. the Western Palearctic. vol. 2. Oxford. vol. 1. Science Press, Alma Ata. Danko, S˘., Divis,T., Dvorska, J., Dvorsky, M., Chavko, J., Teerinck, B. J. 1991. Hair of Western European Karaska, D., Kloubec, B., Kurka, P., Matusik, H., Peske, Mammals: Atlas and Identification Key. L., Schropfer, L., & Vacik, R. 1994. The state of Cambridge. knowledge of breeding numbers of birds of prey Tucker, G. M., & Heath, M. F. (eds.) 1994. Birds in Falconiformes and owls Strigiformes in the Czech Europe: their conservation status. BirdLife and Slovak Republics as of 1990 and their International, Cambridge. population trends in 1970-1990. Buteo 6: 1-89. Watson, M. 1997. Saker Falcon Ecology and Ellis, D. H., Ellis, M. H., & Tsengeg,T. 1995. Productivity Conservation in North East Kazakhstan. of Saker Falcons in Mongolia. In Proceedings of the Unpublished MSc thesis. University of Kent, Specialist Workshop, Abu Dhabi, United Arab Canterbury. Emirates, 14-16 November 1995, Middle East Falcon White, C. 1995. Round Table Discussion on the Research Group. Systematics and of the Saker Falcon. In Frey, H., & Senn, H. 1980. Zur Ernährung der Proceedings of the Specialist Workshop, Abu Dhabi, Wurgfalken Falco cherrug und Wanderfalken Falco United Arab Emirates,14-16 November 1995,Middle peregrinus in den niederösterreichischen East Falcon Research Group.

Mark Watson, The Durrell Institute of Conservation and Ecology, University of Kent, Canter- bury, Kent CT2 7NX (Address for correspondence: c/o 19 Priory Drive, Reigate, Surrey RH2 8AF)

Dr Roger Clarke, New Hythe House, Reach, Cambridge CB5 0JQ

British Birds 93: 136-143, March 2000 143