FORKTAIL 30 (2014): 79-83 Status and roosting characteristics of Collared Crow Corvus torquatus at the Mai Po Nature Reserve, Hong Kong DAVID J. STANTON, BENA R. SMITH & KATHERINE K. S. LEUNG For unknown reasons, the number of Collared Crow Corvus torquatus communally roosting at Mai Po Nature Reserve has steadily increased between 2004 and 2013, and are higher in summer than the preceding winter. The Deep Bay population appears to be isolated and there is no evidence of recruitment locally or the population being supplemented by northern migrants in winter. Collared Crows form large pre¬ roost gatherings in several locations close to the final roost. When they depart to roost after sunset, they do so as a single species flock, showing a preference for dense stands of mangrove trees, probably Kandelia obovata, in the intertidal mangal (mangrove zone) as their final roost location. The majority of birds arrive from habitats to the south of the reserve, suggesting that the local population in Deep Bay is not homogeneous throughout the contiguous network of commercial fishponds. Fishponds are being ecologically degraded or lost over much of the Collared Crow's range, a result of development and agricultural intensification. Given the paucity of information about this species in most of its range, and that it is suspected to be in global decline, the Mai Po Nature Reserve and its environs are clearly an important site for this Near Threatened species. INTRODUCTION Study area The study area is the Mai Po Nature Reserve (hereafter Mai Po), a Collared Crow Corvus torquatus is a near-endemic to China, being 377 ha wetland located in the Mai Po Inner Deep Bay Ramsar Site resident in the south and east, reaching southern Hebei in the north¬ (22.485°N 114.035°E) (Figure 1). east, Gansu in the west and Yunnan and Hainan in the south-west The interior of Mai Po is a mix of impounded brackish ponds (Cheng 1987, BirdLife International 2014). Its range extends to (including traditionally managed shrimp ponds—locally called^?/ north Vietnam, and it is a vagrant to Taiwan (Brazil 2009, BirdLife wai—and shallow water ponds with islands) and rain-fed ponds/ International 2014). It is an uncommon and localised resident in marshes. The major vegetation types are mangal, composed the Hong Kong Special Administrative Region (Hong Kong), predominantly of Kandelia obovata, and reedbed dominated by frequently encountered in the Deep Bay area (Carey etal. 2001). common reedgrass Pbragmites australis. Open water covers 50.3% The global population of the species is in decline (BirdLife of the area (WWF-HK 2013). International 2014), with a range contraction being reported by Earth embankments (or bunds) between ponds andgei wai are birdwatchers and field researchers across China (DJS pers. obs.). typically lined with individual or small clusters of trees composed Loss of food supply owing to agricultural intensification and an of chinaberry Melia azedarach, elephant’s ear Macaranga tanarius, associated overuse of pesticide are considered the major factors, Chinese tallow tree Sapium sebiferum and several standing snags although human persecution in some areas also occurs (dos Anjos (dead trees), although recently disturbed areas are either bare et al. 2009, BirdLife International 2014). In recognition ofreported ground or short grass, interspersed with snags. population declines, a Near Threatened global status was assigned The western part of Mai Po is a strip of mature intertidal mangal to Collared Crow in 2008 (BirdLife International 2014). In 500 1,000 m wide (Figure 3). Dominant species include K. obovata, contrast to the global trend there appears to be a secure population spiny bears breeches Acanthus ilicifolius,Aegiceras comiculatum and in Hong Kong (Kilburn 2009), with an estimated 200 birds in 2007 black mangrove Avicennia marina (WWF-HK 2013). A 3-m high (BirdLife International 2014). immigration fence, constructed in the early 1960s between the impounded ponds and the intertidal mangal, controls human access Status in Hong Kong to the intertidal area. Deep Bay, an eastern branch of the Pearl River estuary located at the north-west of Hong Kong (22.485°N 114.035°E), is the major area for Collared Crows (Carey et al. 2001, WWF-HK 2009a,b). Figure 1. Map of Deep Bay showing the location of the study area. During the day, this population disperses as individuals or small foraging groups throughout the area, with highest concentrations seen feeding on bunds within commercially managed fish ponds. Nearby, on the northern shoreline adjacent to the city of Shenzhen (22.518°N 114.002°E; Figure 1), the species is rarely recorded and occurs generally singly (J. Martinez verbally). Away from Deep Bay, the other main area for the species is Shuen Wan near Tolo Harbour, north-east Hong Kong (22.452°N 114.201°E), with a maximum count of 71 pre-roosting birds in April 2011 (HKBWS 2013). Outside these areas, observations of Collared Crow typically involve no more than five birds. On Po Toi Island (22.167°N 114.252°E), a regularly watched site in southern Hong Kong, there is only a single record (G. Welch verbally). The species disappeared from a cluster of former haunts on Lantau Island following construction of Hong Kong International Airport (22.301°N 113.917°E) (Carey 2009). 80 DAVID J. STANTON, BENA SMITH AND KATHERINE K.S. LEUNG Forktail 30 (2014) Pre-roost and roost behaviour by Hong Kong Observatory) and continued until birds had settled Communal roosting by corvids is well documented (Cramp & at the final roost site in Mai Po or light conditions inhibited Simmons 1977, Francis 1998, Everding&Jones 2004, dos Anjos e/ counting. Following sunset, twilight provides between 20- al. 2009, Seng 2009) and although it has been reported for Collared 50 minutes for observations, and in most cases crows would depart Crow (Carey etal. 2001, Wong & Young 2009, WWF-HK 2009a,b, to roost within this time period. 2011, 2013), we expand further here on a locally well-known A single observation point within an 800 m radius of the final Collared Crow roost. roost was used during each survey; the position of this point was The benefits of communal roosting are well described and dependent on the location of the pre-roost site and a certain amount include a decrease in the chance of predator approach going of flexibility was necessary throughout the counts. Observations unobserved, some physical protection against adverse weather, the were made with binoculars and naked eye. All birds seen arriving facilitation of the meeting and pairing of unrelated individuals and at the pre-roost location were counted and field notes kept of arrival probably maximising the chances of finding rewarding food sites direction (simply as the lour cardinal points of the compass), local the next day (Moore & Switzer 1998, Dali 2002, dos Anjos et al. movements, habitat use, any noteworthy behaviour and association 2009). These ‘information centres’ have previously been attributed with other corvid species. A maximum count was derived from each to both roost sites and breeding colonies (Zahavi 1971, Ward & survey sample; this was usually when birds finally departed to roost. Zahavi 1973, Sonerud et al. 2001, Wright et al. 2003, dos Anjos et However, when a final roost count could not be made, the al. 2009). It has been suggested that young corvids tend to follow maximum count irom the survey was taken based on experience of older individuals and recruitment to the communal roost is pre-roost behaviour and confidence that the maximum count had common (dos Anjos et al. 2009). already been attained. Counts were conducted seasonally between Prior to roosting, many corvids often congregate at locations December 2004 and August 2013; four every winter (between away from the final roost site, forming what is known as a pre-roost December and February) and four every summer (between July and (Moore & Switzer 1998, Hansen et al. 2000, Everding & Jones August), each count being approximately two weeks apart. 2004). These pre-roost gatherings have rarely been studied and their The final roost was determined to be the location where the function is poorly understood;it has been suggested that pre-roost birds that had departed irom the pre-roost locations settled as a gatherings of corvids are not simply a consequence of many larger flock and ceased movements completely for the overnight individuals approaching the same roost area but, like communal roost. Pre-roost locations were deemed to be those areas where birds roosting, have a particular function (Moore & Switzer 1998, congregated in the 60-minute period prior to the final roost. Birds Hansen et al. 2000, Winiecki 2000). may be mobile at these pre-roost locations, but generally form a loose congregation whilst feeding or interacting with other Study objectives individuals. The core part of this study was to look at the long-term population trend of Collared Crows roosting at Mai Po and to form a preliminary assessment of their roosting characteristics. This serves RESULTS to build up the knowledge base to aid the conservation of this species elsewhere in its range and to generate information for A total of 71 counts between December 2004 and August 2013 management decisions at Mai Po. (one count in July 2005 was cancelled due to bad weather) recorded Pre-roost and roost counts were considered the most a maximum of 167 birds (18 July 2013) and a minimum of 31 birds appropriate way of gathering data to meet the study objectives. This (3 February 2005). During each survey, when roosting was paper does not attempt to explain the purpose of pre-roost or roost recorded, there was only one final communal Collared Crow roost gatherings. at Mai Po. Total numbers roosting in both winter and summer periods increased over the eight-year study period at average annual rates METHODS of+9.6 and +8.6 individuals respectively. Winter mean count was 60.1 ± 22.2 birds (range 30-118, n = 36) and summer mean Roost counts and observations count was 100.0 ± 30.2 birds (range 32-167, n = 35). Differences Counts and determination of both pre-roost and final roost between the winter and summer numbers were significant (one¬ locations commenced at least 60 minutes before sunset (as recorded way ANOVAr/./. = l,p<0.001) (Figure 2). Figure 2 Maximum counts of Collared Crows roosting at Mai Po in summer and winter between 2004 and 2013. Forktail 30 (2014) Collared Crow Corvus torquatus at the Mai Po Nature Reserve, Hong Kong 81 Pre-roost gathering and communal roosting Association with other corvids Individuals were often present at regular pre-roost locations before On nine occasions, Large-billed Crow Corvus macrorybnchos and the start of the count. In the 60-minute period prior to sunset, pre¬ House Crow C. splendens were recorded among pre-roosting roosting birds arrived singly or in small groups of 2-10 birds. Birds Collared Crows. These individuals left the pre-roost locations typically congregated at the southern and south-western parts of earlier than the Collared Crow flock and in a different direction; Mai Po (Figure 3) and were observed both loafing and foraging on none was seen to join the Collared Crow roost. Up to 74 Eurasian bunds or islands, or perched on snags or on branches of living trees Magpie Pica pica roosted in a separate area of the intertidal mangal devoid of, or with limited, foliage. All the pre-roost locations were (KKSL pers. obs.) and were regularly observed at the same pre¬ within 150-950 m ol the final roost site. Birds generally departed roost locations used by Collared Crows. to the main night roost location as a single group led by several individuals that flew above the congregation, cawing and circling the pre-roost locations for several minutes. The lead birds would DISCUSSION then drop into the final roost, followed by the main group, and did Overall numbers not re-emerge. Birds settled at the final roost on average 22.4 ± 9.3 minutes after sunset (range -6 to +45 minutes). Counts when At Mai Po, numbers of Collared Crows observed gathering and final roost time could not be determined are excluded. going to roost within the intertidal mangal have steadily increased The final roost was located inside the intertidal mangal (Plate over the eight-year study period, reaching high counts of 167 in 3), with exceptions on eight counts (11.3% of total) when birds summer (2013) and 118 individuals in winter (2012/13). remained at the pre-roost locations and were not seen going into Numbers were consistently higher in each summer period than roost within the survey period. Given the strong site fidelity of the the preceding winter period; this is assumed to be due to the roosting crows for the intertidal mangrove zone, it is assumed on recruitment of young birds following breeding success. Collared these occasions that the population entered the roost sometime after Crows commence breeding in early February with a clutch size of twilight. Whilst the final roost location within the wide belt of 2-6 eggs; young fledge at the end of March (dos Anjos etal. 2009). intertidal mangal could not be determined with precision, the same This corresponds to the winter/summer pattern seen in the roosting general area of dense mangal was utilised consistently (Figure 3). flock and perhaps the drop in winter numbers is a result of natural mortality. However, given the similarities in plumage of different- Direction of arrival to pre-most locations age birds and the varying light levels at survey times, it was difficult Birds not already at pre-roost locations 30 minutes prior to sunset to identify the age-class structure within these pre-roost gatherings. arrived either from the south (49.0% of the total number of birds) Seasonal fluctuations in the number of individuals joining roosts or from areas to the north and east of the study area (40.4%). are usual amongst communally roosting corvids, increasing in early Relatively few birds (10.6%) arrived from the west, i.e. over the summer and falling during autumn (dos Anjos et al. 2009). intertidal mangal. The state of tides was not recorded, but given Preliminary findings from a study in progress confirm that this the low percentage of birds coming from the west tidal influence seasonal pattern is found in the Collared Crow roost at Mai Po. was believed to be minimal. The reason for this gradual population increase is not known. To the authors’ knowledge there have been no significant changes to the habitats of, or management strategies at, Mai Po, and no Figure 3. Pre-roost and final roost locations of Collared Crows at Mai Po. apparent land use or habitat changes in the wider area during the R = location of final roost; P = regular pre-roost locations; pond/ge/ wai study period. Given that lower numbers are seen in winter, there is numbers shown. no evidence to show that the winter population at Mai Po is N supplemented by a winter influx of northern birds withdrawing southward to milder coastal areas of southern China, as reported by dos Amjos etal. (2009). ha addition, our study in progress suggests it is unlikely that there is recruitment locally from populations outside Deep Bay. Further investigation is required to explain the shift of the population away from a state of equilibrium, as population density is determined by birth and death rates, and by immigration and emigration rates. Pre-roost gathering and communal roosting Pre-roost gatherings typically formed in open areas, usually on bare tree branches or snags but also on bare or sparsely vegetated islands and bunds; all appeared to be conspicuous from the air. The trigger for the final movement to roost is not known. Immediately prior to roosting, birds became vocal with constant movements and jostling by small groups between trees. Birds departed to roost well after sunset, settling at the roost on average 22.4 ± 9.3 minutes post-sunset. Variation in settling times may be attributed to seasonality or local weather conditions. In addition, other parameters that may influence timing of roosting include the presence of security floodlights along the immigration fence, which are switched on around dusk with no discernible pattern, along with the periodic drain-down of ponds or gei wai, allowing birds to forage later into the evening. All surveys at Mai Po recorded Collared Crows pre-roosting and roosting communally. Elsewhere in Hong Kong this species 500m only appears to roost communally in the Plover Cove area (Carey 82 DAVID J. STANTON, BENA SMITH AND KATHERINE K.S. LEUNG Forktail 30(2014) G N U E E L L HA NE N RI MARTI KATHE Plate 1. Collared Crow Corvus torquatus gather at pre-roost on a snag Plate 3. Intertidal mangal outside the immigration fence—the final (dead tree), Mai Po Nature Reserve, Hong Kong, September 2013. roost location of the Collared Crows, Mai Po Nature Reserve, Hong Kong, December 2005. from habitats in the south. It is not known why there should be such a preference for the habitats to the south of the roost site given the apparent similarities to other fish pond areas throughout the Deep Bay area. Further investigation into why specific areas are selected as foraging grounds is required. Association with other corvids Communal roosting involving different corvid species has been described previously (Francis 1998, Wright et al. 2000, dos Anjos etal. 2009) and Eurasian Magpies notably share pre-roost locations with the Collared Crows at Mai Po. However none was recorded E going to the same final roost with the Collared Crows, indicating L A H that roosting in a single monospecific flock is preferable. A single N TI Collared Crow was observed at a pre-roost gathering of Large-billed R A M Crows in southern Hong Kong, and presumably roosted as part of Plate 2, Collared Crow gather at pre-roost on a Melia azedarach, Mai the flock (DJS pers. obs.). Po Nature Reserve, Hong Kong, September 2013. Elsewhere in its range. Collared Crow is known to associate with Large-billed Crow (dos Anjos et al. 2009). Qu et al. (2005) etal. 2001, HKBWS 2013). Observations in the wider environs of described up to 20 Collared Crows roosting with other corvids over Hong Kong reveal this species is present in lower densities (singles six consecutive winters in Henan province, China. These were in or pairs) and it does not appear to roost communally as a single mixed flocks with Daurian Jackdaw C. danuricus (1,000+), Eurasian flock. Jackdaw C. monedula (500+), Rook C. frugilegus (4,000+) and The intertidal mangal in Deep Bay is well documented as a Large-billed Crow (5,000+); numbers in parentheses indicate the favoured roost site for Collared Crows (WWF-HK 2009a,b, Wong maximum count over the study period, illustrating the low & Young 2009, WWF-HK 2011,2013). This dense habitat type is proportion of Collared Crows in these roosting flocks. in decline in southern China (UNEP 2008) and globally (UNEP 2010, AFCD 2011). The fact that Collared Crows choose to roost within intertidal mangal as opposed to stands of terrestrial mature CONCLUSION trees which are present nearby, adds potential importance to the conservation of mangal for this species. Investigation of other roost Given the paucity of information on the Collared Crow in most of sites is required. its range (dos Anjos etal. 2009) and that it is believed to be in global The strip of mature intertidal mangal currently used by Collared decline (Kilburn 2009, BirdLife International 2014), Mai Po and Crows is not considered to be under any immediate threat. its environs are clearly an important site for this Near Threatened However, siltation (leading to colonisation by terrestrial climbing species. The increase in the local population at Mai Po is plants), annual defoliation of A. marina by the larvae of the moth encouraging, and it would appear the existing management and the Nephopterix syntaractis, and competition between native mangrove current site condition are favourable for roosting Collared Crows. trees and the recently introduced exotic mangrove species Further studies are required to establish the reasons behind the Sonneratia caseolaris and S. apetala (AFCD 2011) could have continued population growth at Mai Po. Furthermore, it is hoped implications for the Deep Bay Collared Crow population. that wider studies can be employed to reveal the factors affecting Monitoring of these threats and their impact should be considered. the population status and dynamics elsewhere in the region. Direction of arrival to pre-roost locations Birds tended to arrive at pre-roost gatherings from the south, ACKNOWLEDGEMENTS suggesting that the local population in Deep Bay is not homogeneous throughout the wetlands. There are approximately We thank Tommy K.C. Cheung and Roger H. Lee for their help in collecting 1,150 ha of fish pond habitat in Hong Kong (AFCD 2013). A larger data and WWF-HK for usage of its'monitoring data and the Agriculture, area of suitable fish pond habitat is available to the north-east of Fisheries and Conservation Department of Hong Kong Special Administrative MaiPo (Figure 1), but a larger proportion of the population arrives Region for its conservation work in the Ramsar site. Paul Leader provided Forktail 30(2014) Collared Crow Corvus torquatus at the Mai Po Nature Reserve, Hong Kong 83 exhaustive comments on earlier drafts which greatly improved this manuscript United Nations Environment Programme (UNEP) (2008) Nationalreportson and for which we are extremely grateful. We thank Martin Hale for allowing mangroves in the South China Sea. UNEP/GEF/SCSTechnical Publication us to use his images. No. 14. United Nations Environment Programme (UNEP) (2010) World's most comprehensive assessment of mangroves reveals drastic loss to global REFERENCES economy and livelihoods, despite positive acceleration in restoration effort.'World Mangrove Atlas'highlights the importance of and threats AFCD (2011) Mai Po Inner Deep Bay Ramsar Site management plan. Hong Kong to mangroves. Press release downloaded from http:// Special Administrative Region: Agriculture, Fisheries and Conservation www.rona.unep.org/documents/news/20100714_Mangroves.pdf on Department. 05/05/2014. AFCD (2013) Marine fish culture, pond fish culture and oyster culture. Ward, P. & Zahavi, A. (1973) The importance of certain assemblages of birds Downloaded from http://www.afcd.gov.hk/english/fisheries/fish_aqu/ as'information centres'for food-finding. Ibis 1 15: 517-534. fish_aqu_mpo/fish_aqu_mpo.html on 13/06/2013. Winiecki, A. (2000) The wintering strategy of Rooks Corvus frugilegus dos Anjos, L., Debus, S. J. S., Madge, S. C. & Marzluff, J. M. (2009) Family Linnaeus 1758, in Poznan, West Poland. Acta Zool. Cracoviensia 43:135- Corvidae (crows). Pp.494-640 in J. del Hoyo, A. Elliott & D. A. Christie, eds. 164. Handbook of the birds of the world, 14. Barcelona: Lynx Edicions. Wong, C. & Young, L. (2009) Coastal wetlands. Pp.25-68 in L. C. Wong, V. M. BirdLife International (2014) Species factsheet: Corvus torquatus. Downloaded Y Lam & G. W. J. Ades, Ecology of the birds of Hong Kong. Hong Kong from http://www.birdlife.org on 29/04/2014. Special Administrative Region: Kadoorie Farm and Botanic Gardens. Brazil, M. (2009) Birds of east Asia. London: Christopher Helm. Wright, J., Stone, R. E. & Brown, N. (2003) Communal roosts as structured Carey, G. J. (2009) Rocky and sandy coasts. Pp.19-24 in L. C. Wong, V. M. Y Lam information centres in the Raven, Corvus corax. J. Anim. Ecol. 72:1003- &G.W. J. Ades, eds. Ecology of the birds of Hong Kong. Hong Kong Special 1014. Administrative Region: Kadoorie Farm and Botanic Gardens. WWF-HK (2006) An extension to the existing boardwalk and new floating Carey, G. J., Chalmers, M. L., Diskin, D. A., Kennerley, P. R., Leven, M. R., mudflat bird-watching hide at Mai Po Nature Reserve for education Lewthwaite, R. W., Melville, D. S., Turnbull, M. & Young, L. (2001) The and conservation purposes. Project Profile to Hong Kong Special avifauna of Hong Kong. Hong Kong: Hong Kong Bird Watching Society. Administrative Region Government. Hong Kong: World Wide Fund for Cheng, T.-H. (1987) A synopsis of the avifauna of China. Beijing: Science Press. Nature Hong Kong. Cramp, S. & Simmons, K. E. L. (1977) The birds of the Western PalearcticA. WWF-HK (2009a) Deep Bay is an increasingly important wetland for Oxford: Oxford University Press. threatened bird species Collared Crow and Spoon-billed Sandpiper. Dali, S. R. X. (2002) Can information sharing explain recruitment to food from Press release downloaded from http://www.wwf.org.hk/en/news/ communal roosts? Behavioural Ecology 13:42-51. press_release/2009_press_release.cfm?1235 on 11/12/2013. Everding, S. E & Jones, D. N. (2004) Communal roosting in a suburban WWF-HK (2009b) Deep Bay - a global stronghold for Collared Crow? Article population of Torresian Crows (Corvus orru). Landscape and Urban downloaded from http://www.wwf.org.hk/en/whatwedo/ Planning 74: 21-33. conservation/wetlands/managermaipo/maipoupdates/?2121/on 11/ Francis, I. S. (1998) Observations at a winter Rook and Jackdaw roost in 12/2013. Aberdeenshire. Scottish Birds 19:129-133. WWF-H K (2011) Research and monitoring work five-year report, Mai Po Nature Hansen, H., Smedhaug, C.A. & Sonerud, G. A. (2000) Pre-roosting behaviour Reserve: September2005 to August 2010. Hong Kong: World Wide Fund of Hooded Crows (Corvus corone cornix). Canadian J. Zool. 78:1813-1821. for Nature Hong Kong. HKBWS (2013) Hong Kong Bird Report 2011: systematic list. Hong Kong Bird WWF-HK (2013) Mai Po Nature Reserve habitat management, monitoring and Report 2011:1-186. research plan:2013 2018. Hong Kong: World Wide Fund for Nature Hong Kilburn, M. (2009) Conservation of birds in Hong Kong. Pp.167-189 in L. C. Kong. Wong, V. M. Y Lam & G. W. J. Ades, eds. Ecology of the birds of Hong Kong. Zahavi, A. (1971)The function of pre-roost gatherings and communal roosts. Hong Kong Special Administrative Region: Kadoorie Farm and Botanic Ibis 113: 106-109. Gardens. Moore, J. E. & Switzer, P. V. (1998) Preroosting aggregations in the American David J. STANTON, AEC Ltd, 127 Commercial Centre, Palm Crow Corvus brachyrhyncos. Canadian J. Zool. 76: 508-512. Springs, New Territories, Hong Kong. Email: [email protected] Qu J. Z., Zou Q. & Li X. D. (2005) Roost and environment of mixed colony of Corvidae in winter in urban areas of Zhengzhou. J. Xinyang Normal Bena R. SMITH, WWF-Hong Kong, Mai Po Marshes Wildlife University (Natural Science Edition) 18(3): 304-306. (In Chinese with English Education Centre and Nature Reserve, Tam Kon Chau Road, New abstract.) Territories, Hong Kong. Email: [email protected] Seng, L. K. (2009) The avifauna of Singapore. Singapore: Nature Society (Singapore). Katherine K. S. LEUNG, WWF-Hong Kong, Mai Po Marshes Sonerud, G. A., Smedshaug, C. A. & Brathen, 0. (2001) Ignorant Hooded Crows Wildlife Education Centre and Nature Reserve, Tam Kon Chau follow knowledgeable roost-mates to food: support for the information Road, New Territories, Hong Kong. Email: [email protected] centre hypothesis. Proc. R. Soc. Lond. B 268: 827-831.