W est Indian Whistling-ducks in Turks and Caicos I 17

W etland birds in I: a search for W est Indian Whistling-ducks D e n d r o c y g n a a r b o r ea

Geoff M. Hilton1,Tim Cleeves2,Tony Murray3, Baz Hughes4 and Ethlyn Gibbs W illiams5

Wildfowl & Wetlands Trust, Slimbridge, Gloucestershire GL2 7BT, U.K. Current address: Royal Society for the Protection of Birds.The Lodge, Sandy, Bedfordshire SG 19 2DL, U.K. Email: [email protected]

2 Royal Society for the Protection of Birds, N orthW est England Office,Westleigh Mews, Wakefield Road, Denby Dale, Huddersfield, West Yorkshire HD8 8QD, U.K. Email: [email protected]

3 Royal Society for the Protection of Birds.The Lodge, Sandy, Bedfordshire SG 19 2DL, U.K. Current address: National Parks and Wildlife Service, Lagduff More, Ballycroy, Mayo, Ireland. Email: [email protected]

4 Wildfowl & Wetlands Trust, Slimbridge, Gloucestershire GL2 7BT, U.K. Email: [email protected]

! National Trust of the Turks & Caicos, ,Turks & Caicos, British W est Indies.

Email: [email protected]

The West Indian Whistling-duck (Dendrocygna arborea) is globally threatened due to significant population declines during the Twentieth Century. Turks and Caicos is a potentially important range-state for the species, but its status there has never been firmly established. A survey of West Indian Whistling-ducks was conducted in the Turks and Caicos islands during February and March 1999 to make a preliminary assessment of the distribution and abundance of the species in the territory, and to test the viability of several different survey methods, including aerial surveys, tape- playback of the species' calls, and running transects through different habitat types. Only three-five West Indian Whistling-ducks were recorded at two sites on suggesting the species may be genuinely scarce on the islands, overlooked, or seasonally absent. The possibility that West Indian Whistling-ducks behave as somewhat nomadic opportunists in response to unpredictable changes in wetland conditions is raised.

Key Words: Black-billed Whistling-duck, Caribbean wetlands, survey methods, UK Overseas Territories;.

©Wildfowl & Wetlands Trust Wildfowl (2000) 51: I 17-126 I 18 W est Indian Whistling-ducks in Turks and Caicos

The W est Indian Whistling-duck large areas of undisturbed wetland habitat, Dendrocygna arborea was once widespread and few wetlands have been drained or in the Caribbean region. Its populations substantially altered. Human population have declined drastically during the last densities are very low outside the two century due mainly to habitat loss and more developed islands of Providenciales hunting, and it has been designated and Grand Turk, and two of the largest globally Vulnerable (Collar et al. 1994), islands - West Caicos and East Caicos - are although no detailed past or present uninhabited (Figure I). population estimates are available (Collar Despite this apparently high habitat et al. 1992, 1994; Rose & Scott 1997). A availability, there are very few published large population possibly persists in Cuba, records of W est Indian Whistling-ducks though there are no published population from the territory. A 1930 expedition estimates, while Cayman Islands, Bahamas, reported the species on Stubbs Cay, Fort Jamaica, Dominican Republic, Antigua & George Cay & Grand Turk (Walsh- Barbuda and Puerto Rico hold smaller McGehee et al. 1998). In 1987 there were numbers (Collar et al. 1994). records from North Caicos of 15 birds on The Turks and Caicos Islands have Bellfield Landing Pond and I I birds on generally been considered to be a part of Sawgrass Pond, in July and September the species’ range (e.g. Collar et al. 1992) respectively (Walsh-McGehee et al.. 1998). and the territory could potentially be a In 1989 a flock of ca. 20 birds, including key range-state because there are very immatures was seen at Big Pond, North

r 25km 72.00W

' 22.00N North Caicos

G rand Turk

F

Figure I . Map of the Turks and Caicos Islands W est Indian Whistling-ducks in Turks and Caicos I 19

Caicos, with two at Bellfield Landing Pond, Methods and two near Conch Bar, Middle Caicos (P. Bradley pers. comm.). In 1997 a pair with Fieldwork was conducted between 24 downy young was seen on a pond on February and 4 April 1999. This coincided Middle Caicos on 2 December (Walsh- with the dry season in the territory, which McGehee et al. 1998). This is the only in this year produced an unusually intense known breeding record for the Territory. drought such that many lagoons were In October 1998 two birds were seen on partially or completely dry.The work was Montpelier Pond, Middle Caicos conducted on five islands: Providenciales, (Pienkowski & Cross unpublished 1999). W est Caicos, North Caicos, Middle Caicos There is a need to determine the and East Caicos. South Caicos, Grand Turk distribution and abundance ofW est Indian and the smaller cays were not visited (see Whistling-ducks in Turks and Caicos, in Figure I). Five different techniques were order to give a clearer picture of the global used to survey whistling-ducks: look-see status of the species, as well as directing counts at lagoons at dawn and dusk, aerial conservation actions and protected area survey, transect counts through salt flats designation within the territory. Such work and mangroves, dusk counts on flight lines is particularly relevant at present because between mangroves and lagoons, and plans have been published for a major interviews with local people.Tape playback cruise-liner development on the pristine of the species’ distinctive call was used in wetland island of East Caicos (Anon. order to elicit a call response from unseen 1998). This paper reports on a search for birds. W est Indian Whistling-ducks in the Turks and Caicos conducted in 1999.There were Study Sites two aims: firstly to determine the distribution and abundance of the species North, Middle and East Caicos have rather in the territory, and secondly to evaluate a similar habitat characteristics. The number of different survey methods. northern (Atlantic) coasts are rocky or Habitat use by W est Indian Whistling- sandy beaches with substantial wave ducks appears to vary between different action. A band of xerophytic scrub on low parts of the species’ range; birds generally limestone ridges runs inland from the feed at night, and roost during the day (del Atlantic coast towards the southern Hoyo et al. 1992). In the Bahamas the main (Caribbean) side of the islands. Within the feeding areas were fresh or saline lagoons scrub zone are a number of brackish or within scrub (Staus 1998a, 1998b). Birds saline lagoons (known locally as ponds). also fed on intertidal flats, on fruits in the Most of these are landlocked, though some crowns of Royal Palms Roystonea sp. and in have narrow sea inlets or subterranean agricultural crops (particularly corn). connections to the sea. Mangroves Roosting sites have been found in (Rhizophora mangle, Laguncularia racemosa mangroves, around lagoons, or in dry scrub and Avicennia germinans) fringe the lagoons (del Hoyo et al. 1992; Staus 1998a). Turks and xerophytic scrub Mangroves in the and Caicos holds all of these habitat types. territory are typically poorly developed, seldom exceeding 4 m in height. The lagoons are shallow - water depths very rarely exceeded I m during this visit - with 120 W est Indian Whistling-ducks in Turks and Caicos

Table I. Numbers of lagoons and survey coverage during W est Indian Whistling-duck surveys in Turks and Caicos, February-April 1999.

Island Number of Lagoons Number visited at Number visited present' dawn/dusk2 but dry

West Caicos 14 3 0

South Caicos 2 0 0

Middle Caicos 26 8 1

North Caicos 13 9 4

East Caicos 29 1 1 1

Providenciales 16 10 1

Grand Turk 7 0 0

Larger Cays 13 0 0

TO TAL 120 41 7

1 Estimated from 1:25,000 maps of the territory 2 Includes only sites holding water, where a wetland bird census was carried out

little or no emergent or submerged were completely dry), representing vegetation. Substrates are mostly silt or approximately 33% of the total number of clay, occasionally of bare limestone. Moving lagoons in the Turks and Caicos (according towards the southern side of the islands, to 1:25,000 scale maps of the territory) the land gets lower, until the scrub is (Table I) .The aim was to locate as many replaced by extensive salt flats, which run sites as possible where the species was to the Caribbean shore.The salt flats have present, and thus to derive a minimum isolated scrub outcrops and large areas of estimate of the size of the population in shallow unvegetated water. Fringing the the territory, as well as determining its Caribbean shore are further poorly distribution. Therefore preferentially developed patches of mangroves, in selected site were apparently suitable for general no more than two metres in Whistling-ducks (e.g. larger sites with good height. W est Caicos and Providenciales habitat, sites with historical or anecdotal contain the same habitat types, but lack records of the species). O f the named extensive salt flats. ponds with historical records of the species, Bellfield Landing Pond and Dawn/dusk surveys at lagoons Sawgrass Pond (North Caicos) were Forty one lagoons were surveyed at dawn visited but were dry; Big Pond (North or dusk (a further seven were visited but Caicos) and Montpelier Pond (Middle Caicos) were surveyed. W est Indian Whistling-ducks in Turks and Caicos 121

On dawn surveys observers arrived at accessibility. No dedicated transects in sites ca. 30 minutes before sunrise, at xerophytic scrub were conducted. It would which time it was fully dark. On dusk have been necessary to cut paths in order surveys observers remained at the site to create transect routes. Visibility on until 30-60 minutes after sunset, by which either side of a transect route would be time they would have been unable to see very restricted. Given the enormous or identify any arriving Whistling-duck extent of the scrub habitat, Whistling- unless it had called. One-three observers ducks would have had to be were present at each site, and distributed extraordinarily numerous for there to be themselves around the lagoon such that any chance of encountering them. coverage of the open water and fringing vegetation was maximised. A 60-second Dusk counts on flight lines

tape of W est Indian Whistling-duck calls Dusk counts were conducted during four was played at maximum volume through evenings on East Caicos giving a total two 10-amp speakers, three-five times per linear coverage of 4.8km of flight line. visit. Observers were stationed along putative Aerial survey flight lines between mangrove-fringed coastal areas and lagoons, 30 minutes An aerial survey of East Caicos was before sunset. Observers stood 400m conducted between 0925-1 I 10 hours on 2 apart and watched silently for over-flying March. A top-winged single-engine four­ birds until 30 minutes after sunset. seater ‘plane was flown at 90-100 knots Interviews with local people and a height of 70 - 100 m.Two flights over all lagoons in the scrub zone were carried Eight people were interviewed about West out, making a third pass at 50 m if birds Indian Whistling-ducks. Interviewees were remained unidentified. A grid was then asked whether they were aware of the flown over the entire area of saltflats and species; where they had seen it and in what mangroves, passing within ca. 400m. numbers; whether they had seen the (horizontally) of all points. species recently or in the more distant Mangrove, saltflat and scrub past; they were asked to describe the tran sects birds’ habitat and behaviour.

The Caribbean-coast mangrove habitat Results was covered by one 8km walked transect, two motorboat transects totalling 4km and one canoe transect of 3km, all Dawn/dusk lagoon surveys conducted during daylight hours. Observers paused and played a 60-second During the lagoon surveys, W est Indian tape of W est Indian Whistling-duck once Whistling-ducks were found at two sites, every 250 m.Thus ca. 5 % of the Caribbean- both on East Caicos. One site, Jacksonville coast mangrove habitat of the islands was Pond (7I°34’W, 2I°45’N), held two - four visited. Twenty kilometres of transects birds during a dawn visit. One bird called were walked through the saltflat habitat. spontaneously and responded immediately Locations of transects were dictated by to tape playback of a whistling-duck call. 122 W est Indian Whistling-ducks in Turks and Caicos

Jacksonville Pond is saline. It has an Caicos (up to the low hundreds), although unusually extensive, dense and tall several stated that they had not seen mangrove fringe (up to 5m high) and has Whistling-ducks for some time. All reports two small mangrove-covered islands. Again were from lagoons or of nocturnal feeding unusually, there are submerged in maize crops - though these habitats macrophytes. The water is deeper than receive more human visits than other most lagoons (many of which were entirely habitats on the territory. It was rather less than 0.5m deep), though apparently difficult to evaluate the reports. Although less than I m deep in most parts. people are aware of the Whistling-duck as At a second (unnamed) site on East a distinct type, in practice they rarely Caicos (7 1 »29’ W , 21=43’ N), a single bird differentiate it from other duck species. was heard calling briefly 30 minutes after sunset. It did not respond to tape playback, Discussion but there was heavy rain and strong winds. The bird was apparently on a large area of dried-out lagoon that had been re-flooded Occurrence ofW est Indian Whistling-ducks to a depth of ca. 10cm by heavy rain during the previous hour. W est Indian Whistling-ducks were Aerial survey apparently extremely rare on Turks and Caicos during February and March 1999. N o Whistling-ducks were seen during the This may mean that there is no significant aerial survey. It was apparent that only population of the species in the territory; birds that were on open water well clear however there are two alternative of fringing vegetation were likely to be suggestions: Firstly, that the birds were seen and identified. Ground-truthing of present in much larger numbers, but that aerial survey counts indicated that only they were not found. Secondly that they Greater Flamingos Phoenicopterus ruber, were scarce at the time of the survey, but heron and egret species, and American occur in greater numbers at other times of Coot Fulica americana could be detected the year. and counted from the air. It is unlikely that large numbers of Whistling-ducks were overlooked. Mangrove and saltflat transects, Approximately one third of all the lagoons dusk counts on flight lines in the territory were visited at either dawn No W est Indian Whistling-ducks were or dusk. Coverage was actually better than found during transects in mangrove or this figure implies, because: ( I) larger sites saltflat habitat, or during dusk counts on with historical records were visited flight lines. preferentially; (2) many of the lagoons that were not visited were completely dry; (3) Interviews with local people Grand Turk and South Caicos were not visited as these islands are so heavily Middle-aged or elderly respondents were developed that Whistling-ducks are highly all aware of Whistling-ducks, particularly unlikely to persist. the habit of feeding in maize crops at night. If Whistling-ducks arrive at lagoons for Several reported seeing substantial feeding well after dusk, and leave well numbers, particularly on Middle and East before dawn, as occurs at some sites (P. W est Indian Whistling-ducks in Turks and Caicos 123

Bradley verbally), they would have been widespread at some time. Interviews with missed in the current survey. There is a local people suggested that Whistling- suggestion that at sites where hunting ducks were a relatively common sight pressure is high, birds are more wary, and between approximately 60 and 20 years delay flight until full darkness, leaving again before present, and that they probably well before dawn. For example at Long used agricultural crops and lagoons. Island in the Bahamas, birds tend to fly to Nowadays they are clearly not sufficiently feeding sites at around dawn and dusk abundant to be seen by local people in the (Staus 1998a, 1998b), whereas in Jamaica course of their normal activities. However, flights tend to occur in full darkness agriculture and hunting have almost ceased (Williams & Rees 1997). Hunting pressure on the Territory, and local people very on W est Indian Whistling-ducks in the rarely visit the extensive wetlands. Thus Turks and Caicos is low to non-existent the lack of recent records may simply (pers, obs.), such that the birds are unlikely represent a reduction in the likelihood of to be particularly wary; furthermore East people encountering Whistling-ducks. and W est Caicos are uninhabited. The second suggestion - that birds were Only a small proportion of the potential seasonally absent from the islands - is roosting habitat in coastal mangroves and difficult to assess. During the study visit scrub was covered. Hence most birds in there was an unusually intense drought on these areas would have been missed. the islands. Most of the lagoons were at However, most of the better-developed least partially dry and lacking in mangrove vegetation occurred round macrophytes, phytoplankton and lagoons rather than on the coast, and invertebrates, while the saltflats were hence one might predict that lagoon- mainly dry and unproductive. A rapid flush feeding birds would roost at the lagoons of growth following rain might perhaps too. The possibility that large numbers of make both habitats much more suitable. birds used the scrub for roosting, as they Is it possible that W est Indian Whistling- do in the Bahamas (Staus 1998a), cannot ducks are somewhat nomadic, and move be dismissed, given the many thousands of between island sites in response to local hectares of this impenetrable habitat on variations in rainfall? The species is the islands. The potential feeding areas on generally thought to be relatively the saltflats did not appear suitable. They sedentary (del Hoyo et al. 1992), and held an extraordinarily low density and vagrancy, which might be an indicator of a diversity of birds during the study period, tendency to nomadism, has very rarely and there was apparently little primary been reported (Collar et al. 1992). productivity (no visible macrophytes or However, an ability to move rapidly and phytoplankton), and almost no visible exploit locally favourable conditions would invertebrate life. appear to be adaptive in the Caribbean Although breeding of W est Indian region where the climate is unpredictable. Whistling-ducks on the islands has been The bird’s breeding season appears to be proved relatively recently (Walsh- rather variable, indicating a flexible McGehee et al. 1998), there is no evidence response to wetland conditions (Madge & to suggest that there has ever been a large Burn 1987).Thus the possibility that Turks population. The wide scatter of previous and Caicos holds a significant population records hints at the bird having been of W est Indian Whistling-ducks at other 124 W est Indian Whistling-ducks in Turks and Caicos seasons or in other years cannot be viable census technique, since they are too excluded. Indeed, a confirmed breeding labour intensive. In Turks & Caicos the record was received from Conch Bar, mangrove area is much less substantial, and Middle Caicos, following the onset of significant proportions of the habitat could heavy rains in June I 999. Potential source be covered either in canoes or on foot populations are in Cuba, which lies with relatively low effort in a search for ca.300km from the Caicos Islands and roosting birds. Similarly the use of flight probably holds the largest W est Indian line counts at dusk or dawn allows good Whistling-duck population, and in the coverage to be achieved with relatively low southern Bahamas, 100-300km from the effort. However, there is no certainty that Caicos Islands, where there are also the birds do fly between mangroves and thought to be substantial populations lagoons. (Collar et al. 1992). W e have shown that W est Indian Whistling-ducks respond to tape playback Evaluation of survey methods of their calls, and this could be a valuable tool in future survey work. W e have not, Because so few W est Indian Whistling- however, quantified the probability of ducks were recorded, it is rather difficult to evaluate the different survey methods. getting a response to playback from a bird Hence this section is based largely on that is present, nor investigated seasonal subjective impressions. variation in response. Dawn/dusk lagoon surveys appear to be In this study, interviews with local people proved to be of little use in a promising means of locating birds. Available evidence suggests that these tend elucidating the current distribution of the to be the most favoured feeding sites species. However, this may not always be the case. In Turks & Caicos very few people (Staus 1998a), and our findings support this idea. Lagoons of the sort visited in work on the land or visit wetlands. Turks & Caicos are very simple to survey, Because we were unable to locate birds ourselves, it was difficult to corroborate because of their relatively small size and anecdotal reports. In areas where local lack of emergent vegetation (Hilton et al. 2000). Saltflats are more problematic: their people are more likely to come into enormous extent means that a large effort contact with the species, interviews should is needed to cover a significant proportion be seen as a potentially valuable tool for of the total area at dawn or dusk. locating key sites. Interviews were used Aerial surveys are probably not suitable with apparent success to locate surviving for detection of W est Indian Whistling- populations of W est Indian Whistling- ducks due to the difficulty of identifying ducks in Jamaica (Williams & Rees 1997). birds that aren’t on open water. Aerial The similarity of the crepuscular and surveys must also be done in good light, similar sized White-cheeked Pintail to when Whistling-ducks will usually be Whistling-ducks at a distance and in very roosting and very difficult to detect. low light levels was striking, despite the In areas where there are large areas of very obvious differences between the xerophytic scrub - such as the Bahamas species. Pintails tend to fly into lagoons in and Turks & Caicos - which serve as small parties at dusk, and on landing potential roosting habitat (Staus 1998a), it assume an alert posture with neck outstretched, making them appear very is unlikely that scrub transects will be a W est Indian Whistling-ducks in Turks and Caicos 125 long-necked. Reports of large, crepuscular and Martha and George Volz for help and ducks should therefore be treated with advice during the study. Mike Pienkowski caution when made by inexperienced and Sara Cross of the U K Overseas observers, such as the interviewees Territories Conservation Forum, and Jim questioned in this study. Stevenson and Debbie Pain (RSPB) all In conclusion, the initial search for the provided valuable assistance in the UK. location of birds should mainly be focused Patricia Bradley, Lisa Sorenson and Ann on dawn and dusk lagoon visits. In addition Sutton gave useful advice on survey it would be valuable to use dawn/dusk or techniques. RSPB and W W T funded this night-time visits to maize crops, and also study. to try night visits to lagoons, with use of call playback, in order to examine the References possibility that the birds do not move on to feeding areas until well after full darkness. Anon. (1998). Proposed destruction of natural The extensive wetlands on West, North, wetlands by East Caicos port and holiday development. Conservation News 16. UK Middle and East Caicos remain largely Overseas Territories Conservation Forum, unaltered and undisturbed by humans. December 1998. Hunting is at a very low ebb, and was much Collar, N.J., Crosby, M.J., & Stattersfield, A.J. more widespread in the past. The only (1994). Birds to Watch 2. The world list of notable change to W est Indian Whistling- threatened birds. Birdlife International, duck habitat has been the decline of Cambridge, UK. agriculture in the territory, which may have Collar, N.J., Gonzaga, L.P., Krabbe, N „ Madroño deprived the species of a food source.The Nieto, A., Naranjo, L.G., Parker,T.A., & Wege, habitat on the islands is very similar to that D.C. (1992). Threatened Birds of the Americas. described by Staus (1998a, 1998b) in the The ICBPIIUCN Red Data Book. ICBP, Bahamas, where there is a relatively large Cambridge, UK. population. It is therefore surprising that del Hoyo, J., Elliot, A. & Sargatal, J. (1992). so few birds were recorded in the present Handbook of the Birds of the World. Volume I: survey. It remains possible that Turks and Ostrich to Ducks. Lynx Edicions, Barcelona. Caicos is an important range state, and Hilton, G.M., Gibbs Williams, E., Cleeves, T. & further surveys at different times of the Murray,T. (2000). Wetland birds in Turks and year are needed to clarify the situation. Caicos Islands II: wetland bird communities. W ildfow l 5 1: 127-1 38 Acknowledgements Madge, S. & Burn, H. (1987). Wildfowl. An identification guide to the Ducks, Geese and Swans of the World. Christopher Helm, W e are very grateful for the help and London. hospitality of Natasha Cox, Elaine Pienkowski, M.W. & Cross, S.J. (1999). R ep o rt on Reynolds and Kathy Green at the Turks & Forum work with Turks & Caicos National Trust, Caicos National Trust. Also to Valdez October/November 1998. Unpublished report Thomas and Ernest Williams of the to the UK Overseas Territories Department of Environmental & Coastal Conservation Forum. Resources, and to Marsha Pardee- Rose, P.M. & Scott, D.A. (1997). W aterfowl Woodring, Kathleen Wood, Telford Outen Population Estimates, second edition. 126 W est Indian Whistling-ducks in Turks and Caicos

Wetlands International Publication 44. Wetlands International, Wageningen, Netherlands.

Staus, N.L. ( 1998a). Habitat use and home range of W est Indian Whistling-ducks. Jo u rn a l o f Wildlife Management, 62: 171 -178. Staus, N.L. (1998b). Behaviour and natural history of the W est Indian Whistling-duck Dendrocygna arborea on Long Island, Bahamas. W ildfow l 49: 194-206. Walsh-McGehee, M., Lee, D.S., & Wunderle.J.M. (1998). A report of aquatic birds encountered in December from the Caicos Islands. Bahamas Journal of Science 6: 28-33. Williams, I & Rees, D. Status (1997).The status and distribution of the West Indian Whistling- duck in W estern Jamaica. RSPB Sabbatical Report. RSPB, Sandy.