Ornithol Sci 18: 215 – 219 (2019)

SHORT COMMUNICATION First tracking of post-breeding migration of the Ruddy King sher coromanda by GPS data logger

Shingo UEMURA1, Ayumu HAMACHI2, Kunihiro NAKACHI3 and Masaoki TAKAGI4,#

1 Graduate School of Sciences, Hokkaido University, Kita-ku, Sapporo-shi, Hokkaido 060–0810, 2 Prec Institute Inc., Chiyoda-ku, Tokyo 102–0083, Japan 3 Wild Society of Miyako, Higashi-nakasone 711–3, Hirara, Miyakojima-shi, Okinawa 906–0007, Japan 4 Faculty of Science, Hokkaido University, Kita-ku, Sapporo-shi, Hokkaido 060–0810, Japan

ORNITHOLOGICAL Abstract The migratory strategies of are poorly understood. In the first study of its kind, by means of Global Positioning System data loggers, we were SCIENCE able to describe the migration and wintering sites of the Ruddy Halcyon © The Ornithological Society coromanda population breeding on Miyako Island, Japan. were found to reach of Japan 2019 altitudes of almost 4,000 m during their migration to winter on Polillo and Tablas islands in The .

Key words GPS data logger, Halcyon coromanda, Migration, Philippines, Ruddy Kingfisher

The majority of kingfisherAlcedinidae species have understanding of the evolution of migration in king- pan-tropical distributions, and most (104 out of 114 fishers. species) are resident (Woodall 2017; Gill & Donsker In this paper, we describe the first use of Global 2018). The migratory strategies of the few migratory Positioning System (GPS) data loggers to track the species of Alcedinidae are poorly understood; how- autumn migration and wintering locations of Ruddy ever, a phylogeny of kingfishers by Andersen et al. Kingfisher from a breeding population on Miyako (2017) indicates that their migratory strategies may Island, Japan. This kingfisher is particularly -suit have evolved independently. able for migration research using GPS data loggers Ruddy Kingfisher Halcyon coromanda occurs in because of their body mass and high recapture rate in South-East to Nepal, south , northeast the study area (Uemura unpublished data). China, Korea and Japan. Ten subspecies of the Ruddy Kingfisher Halcyon coromanda have been recog- nised, three of which are migratory (Woodall 2017; MATERIALS AND METHODS Gill & Donsker 2018). In Japan, two subspecies (H. c. major and H. c. bangsi) occur and are exclusively Study area and species summer visitors (The Ornithological Society of Japan The study was conducted in Ohno Sanrin For- 2012). H. c. bangsi breeds in the Tokara Islands, est, on Miyako Island in the Ryukyu Archipelago, Amami Islands, and Ryukyu Islands. Although they Okinawa, Japan (24°48′30.7″N, 125°18′57.1″E). were thought to winter in the Philippines and Talaud This secondary evergreen forest is flat and domi- Islands (Woodall 2017), none of the birds ringed in nated by Pinus luchuensis and Bischofia javanica. Japan have been recorded or recaptured there (Bio- Ruddy Kingfishers arrive in the study area during diversity Center of Japan 2019), and their migration early April, excavate their nest cavities in tree snags, routes and wintering sites have remained unknown. and commence egg laying in May and June (Uemura Revealing the migratory movement of a migratory unpublished data). They leave Miyako Island by late subspecies of the Ruddy Kingfisher may help our September (Wild Bird Society of Miyako (2014)). The recapture rate is 32% in our study area (Uemura (Received 18 October 2018; Accepted 18 February 2019) unpublished data). # Corresponding author, E-mail: [email protected]

215 S. UEMURA et al.

Capture, marking and tagging the battery of the tag. All fixes were made at 2130 We captured three pairs of Ruddy Kingfishers JST. As the location data were downloaded from the using mist-nets, in front of their nests from 19 to 23 GPS data loggers as geographic coordinates no fur- July 2016. Each captured individual was weighed to ther processing was necessary to determine locations. the nearest 1 g using a 300 g spring scale (40300, We eliminated fixes with dilution of precision values Pesola AG); their body masses ranged from 94 g lower than 10. This threshold is sufficient for a reli- to 119 g (mean±SD 103.7±7.7 g n=6). Individuals able point indication. Locations were mapped using were ringed with numbered aluminium and coloured Google Earth (7.1.2.2041, Google, US). The onset plastic rings (Japanese Ministry of the Environment) of migration was defined as the date after which all for identification in the field, and blood samples were subsequent locations occurred away from the breed- taken to confirm their sex (Fridolfsson & Ellegren ing area. A wintering site was defined as the last site 1999). GPS data loggers (PinPoint-10, Loteck Wire- in which a bird remained for more than a month. less Inc.) were attached to the back of each kingfisher using Rappole-Tipton style harnesses made of 2 mm RESULTS tubular Teflon ribbon (Rappole and Tipton 1991) (Fig. 1). The GPS loggers and the harness ware weighed Three of the six marked kingfishers were recap- to the nearest 0.1 g using a digital scale (KP-104, tured on Miyako Island in 2017: male A on 20 April, Tanita Corporation). The GPS loggers, including the male B on 10 June, and female C on 11 June. On harness, weighed approximately 2.4 g, amounting to the dates of recapture male A weighed 96 g (4 g 2.0% to 2.6% of each individual’s body mass. lighter than at initial capture), male B weighed 104 g (2 g lighter), and female C weighed 112 g (18 g Recapture heavier). Although each of the kingfishers showed In the breeding season (April to June) 2017, we some feather wear (from their harnesses) on their used mist-nets to recapture the tagged kingfishers in backs and knees, none exhibited plumage or skeletal the Ohno Sanrin Forest. We tried to recapture them disorders. We did not observe or recapture the other almost everyday as far as the weather permitted. marked kingfishers. When a focal kingfisher was recaptured, we retrieved When we tagged focal individuals in 2016, pair the GPS data logger and re-weighed the bird. A was rearing three nestlings, pair B was incubating two eggs, and pair C was rearing three nestlings. GPS locations and analysis After we tagged them, all nestlings raised by pair A The non-breeding season was divided into three and C fledged successfully, and one nestling from time periods, August to October, November to Febru- pair B fledged while one egg did not hatch (Table 1). ary, and March to April, and GPS data collection was The body mass of the 12 day-old nestlings ranged individually programmed to each interval. For the from 97.7 to 116.2 g (mean±SD 107.2±6.0 g n=7) first and third periods, the loggers were set to track while it was 57.6 to 123.1 g (mean±SD 105.3±11.4 their migration every three days while for the second g n=49) for all 12 day-old nestlings in 2016. Male A period it was set to record every nine days to save occupied the same territory in 2017 as in 2016, mated with female C in 2017 and successfully fledged four nestlings. In 2016, female C had mated with male B. We were able to download data from males A and B, but not from female C. Female C’s GPS data logger antenna had disappeared and further missing parts prevented any data from being downloaded (Vandentillaart personal communication). Male A remained in its breeding territory on Miyako Island until 24 September 2016 (Fig. 2A and 2B). Two locations were obtained during male A’s migration, the first, on 30 September, was off Okinawa Island, about 220 km northeast of Miyako Fig. 1. Rappole-Tipton style harness attached to a Ruddy Island at an altitude of 3,945 m, and the second, Kingfisher. on 6 October, was off Polillo Island at an altitude

216 The migration of the Ruddy Kingfisher

Table 1. Nest ID, capture date, recapture date, body weight at capture, body weight at recapture, number of eggs and nestlings at capture, number of fledglings of tagged individuals.

Eggs and Weight (g) Weight (g) Fledglings ID Nest ID 2016 Capture Recapture nestlings /Capture /Recapture 2016 /Capture male A A 2016/7/19 2017/4/20 100 96 3 nestlings 3 female A A 2016/7/19 103 3 nestlings 3 male B B 2016/7/17 2017/6/10 106 104 2 eggs 1 female B B 2016/6/15 119 2 eggs 1 male C C 2016/7/21 100 3 nestlings 3 female C C 2016/7/26 2017/6/11 94 112 3 nestlings 3

Fig. 2. (A) Autumn migration routes, wintering areas and altitudes of two Ruddy Kingfishers. The solid line represents male A and the dashed line male B. White squares (□) indicate wintering site or stop over site of male A, black squares (■) indicate wintering area of male B. (B) Location points in Ohno Sanrin Forest, Miyako Island, Japan, between 1 August and 24 September 2016. Blue points represent male A and yellow points represent male B. The large circles represent their nests. (C) Locations of male B on Polillo Island, The Philippines, between 3 October and 17 November 2016. of 2,353 m. Male A continued its migration until known to be 44 m above sea level. The altitude data at least 9 October, but the battery ran out after the obtained during migration were based on more than logger failed to record its locations on 12 and 15 four satellites, thus it was thought to be credible. The October. Male A may have landed on Tablas Island distance between Miyako Island and Tablas Island is on 9 October because it’s GPS data logger recorded 1,429 km, whereas that between Polillo Island and an altitude of 98.2 m at a point where the elevation is Tablas Island is 300 km. The total migration distance

217 S. UEMURA et al. including the point off Okinawa Island and off Polillo and initially moved northeast before turning south. Island was 1,815 km. Seven days later, male A arrived off Polillo Island. Male B was last recorded in its breeding terri- The two data points obtained were both over the tory on Miyako Island on 24 September 2016; by sea and indicated that male A migrated at high alti- 3 October it had reached The Philippines. No data tude. Above 2,000 m, stable winds blow in what is were obtained during migration because the GPS data called the free atmosphere; there, the wind is scarcely logger failed to record locations on 27 and 30 Sep- affected by either geographical features or convection tember. Male B spent the period from 3 October to at currents (Roland 1988). Flight above this altitude least 17 December around the Dangla River, Polillo may be especially efficient for long distance migra- Island (Fig. 2C), after which the data logger battery tion (Kerlinger & Moore 1989). The initial flight of failed. The GPS data logger recorded altitudes of 65 male A to the northwestward may indicate that he m to 117 m at points (mean±SD 90±15 m n=10) became caught in a typhoon wind. The 17th Typhoon where the elevation is known to be about 40 m above in 2016 approached Miyako Island during 26 and sea level. The distance between Miyako Island and 27 September. Typhoon wind and clouds reach the Polillo Island is 1,153 km. Although the GPS data altitude at which male A was flying. Male A landed loggers had been programmed to fix data until the on Tablas Island on 9 October, but the lack of subse- birds returned to their breeding grounds in spring quent data prevented us from judging whether it win- 2017, their batteries ran out after just a few months. tered there or continued its migration further south. Due to the limited data we obtained, we could not reveal their migration and wintering sites fully. DISCUSSION However, the wintering sites inferred for males A The weights of males A and B, and female C did and B were consistent with the wintering range of not change much between tagging in 2016 and recap- H. c. bangsi described by Woodall (2017) as The ture in 2017. After we deployed tags on them in 2016, Philippines and Talaud Islands (Woodall 2017). all pairs reared chicks of normal body mass. Further- Although males A and B bred in neighbouring ter- more, male A and female C successfully fledged four ritories on Miyako Island, Japan, their wintering sites nestlings (the norm for this population (Hamachi et in The Philippines were far apart. The relationship al. unpublished data)) after their GPS data loggers among breeding and overwintering locations may be were removed. It appears that the GPS data loggers weak in Ruddy Kingfishers. did not constitute a deleterious burden for the birds. In this study, only two locations during migration The GPS data showed that both male (A and B) between Miyako and Tablas islands were obtained Ruddy Kingfishers remained within their breeding from male A. In the future, improvement of log- territories from their capture in July until late Sep- gers will make it possible to obtain more data dur- tember 2016 when they began their migration. The ing migration between breeding and wintering sites possible height reached by birds in the area was by setting data loggers to record locations more fre- approximately 40–60 m. The altitude data showed quently (e.g. twice a day). As the data logger batter- an error range of −40 to +60 m. This confirms that ies lasted less than six months it should be possible to the GPS data provided by these tags were reliable save power by commencing data recording from late enough even when the birds were in lowland forest. September onwards when Ruddy Kingfishers begin After its migration, male B remained around the migrating. Dangla River on Polillo Island for more than 45 days, indicating that it wintered in the area, which is domi- ACKNOWLEDGMENTS nated by mangrove forest (Hampson et al. 2003). Male B’s GPS data logger recorded its altitude as We thank: the members of the Laboratory of 65.3 to 117.2 m. The area is known to be about 40 Sociology, Osaka City University, and the Laboratory m above sea level, and given that tropical rain for- of Biodiversity, Hokkaido University, for field and est trees can reach heights of more than 70 m and laboratory assistance; N. Nakamura and T. Yamazaki mangrove trees may reach up to 40 m, the altitude of the Yamashina Institute for Ornithology, Abiko, data recorded appear to correctly assess the bird’s for helpful advice concerning the usage of tracking position in forest. devices; and The Japanese Environmental Agency Male A commenced its migration by 30 September for permits (Permit numbers 11–10, 11–13, 11–14,

218 The migration of the Ruddy Kingfisher

11–19, and 11–21). This study was partly supported J F Ornithol 62: 335–337 by the Research Support Project of the Japan Bird Gill F & Donsker D (eds) (2018) IOC World Bird List Research Association Grant Number 2015–003. We (v8.2). would also like to thank Mark Brazil, Scientific Edit- Hampson K, Bennett D, Alviola P, Clements T, Galley ing Services, for assistance in the preparation of the C & Hilario MV (2003) Wildlife and conservation final draft of the manuscript. in the Polillo Islands. Polillo: Polillo Project Final Report. The Ornithological Society of Japan (2012) Check-list REFERENCES of Japanese Birds. 7th revised edition. The Ornitho- Andersen MJ, Smith BT, Mccullough JM et al. (2017) logical Society of Japan, Sanda. A phylogeny of kingfishers reveals an Indomalayan Rappole J & Tipton A (1991) New harness design for origin and elevated rates of diversification on oceanic attachment of radio transmitters to small passerines. islands. J F Ornithol 62: 335–337. Biodiversity Center of Japan (2019) Web-GIS Atlas Roland BS (1988) An introduction to Boundary of Birds (Bird Banding Survey, Data of recov- Layer Meteorology. Kluwer Academic Publishers, ery records). Available at http://www.biodic.go.jp// Dordrecht. birdRinging_en/index.html (accessed on 23 January Woodall PF (2017) Kingfishers (Alcedinidae). In: del 2019). Hoyo J, Elliott A, Sargatal J, Christie DA & de Juana Fridolfsson A-K & Ellegren H (1999) A Simple and E (eds) Handbook of the Birds of the World Alive. Universal Method for Molecular Sexing of Non-Rat- Lynx Edicions, Barcelona. ite Birds. J Avian Biol 30: 116–121. Wild Bird Society of Miyako (2014) Miyako yachou no Kerlinger P & Moore FR (1989) Atmospheric Structure kai 40 shuunen kinen shi (The 40th-anniversary issue and Avian Migration. Curr Ornithol 109–142. of the wild bird society of Miyako). Shimoji Print, Rappole J & Tipton A (1991) New harness design for Naha (in Japanese). attachment of radio transmitters to small passerines.

219