High fidelity to wintering, stop-over and breeding sites shown by a Long-billed Curlew Numenius americanus tracked with satellite telemetry on migratory flights across North America

Abstract

Capsule We studied the migratory movements and site fidelity of a male Long-billed Curlew using satellite telemetry in North America; the bird completed three migratory cycles and showed strong fidelity to stop-over, breeding, and wintering sites, not only on a geographical scale but also on a local scale across the years.

Studying and understanding the annual cycles of animal species, particularly migratory ones, is vital to establishing effective conservation measures for their habitats throughout their range. Determining the levels of connectivity among breeding, migratory, and wintering areas is a priority for fully understanding population dynamics (Heckscher et al. 2011). Avian migration studies have increased and improved dramatically due to the recent availability of small satellite tracking devices known as platform transmitter terminals or PTTs (Martell et al. 2001, Kenow et al. 2002, Warnock & Takekawa 2003).

The Long-billed Curlew Numenius americanus, the largest shorebird in North America, is a species of high conservation concern in both the US and Canadian shorebird conservation plans (Donaldson et al. 2000, Brown et al. 2001) and a species of special concern with both the US Fish and Wildlife Service and the Canadian Wildlife Service (U.S. Fish and Wildlife Service 2008, COSEWIC 2011). This concern is due to apparent population declines (Brown et al. 2001). The Long-billed Curlew breeds in prairies of the Great Plains and Great Basin, and in inter-montane valleys of the western US and southwestern Canada. It winters along the western Florida and central California coasts, significantly in California's Central Valley, and south through Mexico (American Ornithologists’ Union 1998, Dugger & Dugger 2002, Shuford et al. 2009, Sesser 2013).

Little information exists concerning the Long-billed Curlew's migration timing and routes, therefore identifying critical wintering and migration staging areas are priority conservation actions for the species (Fellows & Jones 2009). The only available data on Long-billed Curlew migratory patterns are from birds captured in three breeding sites of the USA (Nevada, Oregon, and Montana; Page et al. 2014). Here, we describe the migratory routes, migration timing, and fidelity to wintering, stop-over, and breeding sites of a male Long-billed Curlew captured at a wintering site in Mexico and tracked over two years with satellite telemetry.

We captured five wintering Long-billed Curlews (four males and one female) using leg-hold noose mats (Bub 1991, Mehl et al. 2003) in mid-October 2009. The sex of each bird was determined in the field by the length of its bill (Dugger & Dugger 2002) and was subsequently confirmed by molecular analysis of DNA. The birds were captured in a livestock pond (24°57′30″N, 100°43′24″W) in the community of La Hediondilla, Galeana, Nuevo Leon, Mexico. We selected the four birds (three males and one female) with the best apparent health (heavier and without ectoparasites) and fitted them with satellite transmitters to track their movements. Transmitters were attached with a single thread leg-loop harness (Sanzenbacher et al. 2000). The PTT and harness together weighed approximately 3.6% of the captured curlews’ average body mass (549 g), which is below the 5% recommended by Gaunt et al. (1997). All four curlews were tracked during the entire winter season in Mexico in order to obtain winter habitat use; we were able to successfully track migratory paths of only one male. This male was ringed and fitted with a 20-g solar-powered satellite transmitter (Microwave Telemetry, Columbia, MD, USA) with a duty cycle of 10 h on and 24 h off, and a life expectancy of 3 years.

We obtained the PTT's locations from the ARGOS satellite tracking system (ARGOS 2008) operated by Collecte Localisation Satellites (CLS America, Inc., Largo, MD). Latitude and longitude in decimal degrees, date, time, and location error were received from ARGOS within 24 h of satellite contact with the PTT. All auxiliary locations (codes 0, A, B, Z) were discarded because of their high error radius (>500 m), which makes them unsuitable for interpreting animal movements; only those locations with a 350-m accuracy or less (codes 3 and 2) were plotted, using Arc View 3.2 (ESRI, Redlands, CA).

The PTT had an operational lifetime of 876 days (14 October 2009 to 7 March 2012). During this period the curlew completed two annual cycles and showed high fidelity to wintering, breeding, and stop-over sites (Table 1, Fig. 1).

Figure 1. Routes used by a male Long-billed Curlew tracked with satellite telemetry across North America during 2010 and 2011 spring and autumn migrations.

Table 1. Migratory parameters of an adult male Long-billed Curlew tracked by satellite telemetry during 2009–2012.

Parameter	Spring migration		Autumn migration
	2010	2011	2010	2011
Departure	24 March	18 March	23 June	27 June
Arrival Final Destination	16 April	21 April	20 October	4 November
Days spent in migration	24	35	120	131
DST^a/DSSSM^b	4/20	5/30	4/116	4/127
Linear distance (km)	2975	2845	3486	3137
Average distance between consecutive roosting places (km)	595	569	697	784
Parameter	Breeding Season		Wintering Season
	2010	2011	2009–2010	2010–2011	2011–2012
Days spent on breeding grounds (Canada)	68	68	-	-	-
Days spent on wintering grounds (Mexico)	-	-	161	263	247

^aDays spent travelling.

^bDays spent in stop-over sites during migration.

We defined spring migration as the period between when the curlew started moving consistently north from its wintering site until it reached breeding territory and remained in the same area (within 15 km) for at least 7 days; after this period we considered spring migration finished. The time between when the curlew departed breeding sites travelling south, and arrived at traditional wintering grounds, and remained within 15 km of the same area for at least 7 days, was considered autumn migration. Stop-over sites during migration included all places where birds did not move through the area in a consistent direction (south or north) for at least three days (Johnson et al. 2010).

The curlew travelled an average Euclidean linear distance of 2910 km during spring migrations; a slightly greater distance was recorded on autumn migrations, an average of 3311 km. The maximum distance covered in 1 trip was 1836 km during the 2011 autumn migration, the trip took 36–48 h to be completed.

The duration of the autumn migration was longer than the spring migration (average migration days in autumn = 126 days; spring = 30 days). The curlew also travelled longer distances during autumn migrations than during spring migrations, with an average distance between consecutive roosting places of 740 km in autumn versus 582 km recorded for the spring migration.

Spring migration stop-over sites

We identified one main stop-over site, which was used for 22 and 28 consecutive days in 2010 and 2011, respectively. It is a 1272 km² crop-land area located on the borders of Parmer, Deaf Smith, Castro (Texas) and Curry (New Mexico) counties (34°30′N, 102°35′W). Other sites used to a lesser extent include Big Horn and the northeast portion of Wyoming, and Petroleum and Garfield in Montana (1–3 days in 2010); Adobe Creek Reservoir, Fort Collins, and Cheyenne in Colorado (2–3 days in 2011).

Autumn migration stop-over sites

Two main stop-over sites were recorded during both years in the USA and Mexico, one for each country. The agricultural area located within the Reeves (31°23′N, 103°43′W), Culberson (31°18′N, 104°28′W), and Balmorhea counties (30°52′N, 103°44′W) in Texas was used by the curlew for 3 days and 1 day in 2010 and 2011, respectively. Other stop-over sites in the USA included Thomas, Kansas (2 days in 2010) and Curry, Texas (7 days in 2011). Within Mexico the curlew fed and rested during 114 and 122 consecutive days in 2010 and 2011, respectively, in the agricultural area located on the Durango and Coahuila states limit (26°10′N, 103°40′W and 25°32′N, 103°12′W, respectively).

Breeding grounds (canada)

The bird arrived at the same breeding site on 16 April 2010 and on 21 April 2011. The site is located halfway between the cities of Lethbridge and Medicine Hat in Alberta province (49°58′N, 111°46′W). Within the area, there is irrigated agricultural land and native pasture, with an oil field.

Wintering grounds (mexico)

We identified three main zones used during all three seasons (2009–2012): General Cepeda in Coahuila (25°22′N, 101°28′W), and San Rafael and Llano La Soledad in Nuevo Leon (25°01′N, 100°39′W and 24°50′N, 100°41′W, respectively). These areas contain crop-lands, shrubs, and grasslands.

We documented migratory movements of one bird, and so it is difficult to draw general conclusions concerning the Long-billed Curlew's migration behaviour. However, because of the successful tracking of this bird, our results show strong fidelity to stop-over sites, breeding, and wintering grounds at a geographical and local scales, as the curlew used the same foraging and roosting sites across all years of our study.

Long-billed Curlews use the Central Plains/Playa Lakes Region of the American Central Flyway along with 37 other shorebird species. According to Fellows et al. (2001), the hydrology of most wetlands within the region has been altered by wetland drainage, agriculture practices, and urbanization, the latter particularly affects most shorebirds as they depend on the invertebrates found in the wetlands; insects and spiders make up a large part of the Long-billed Curlew's diet (Erlich et al. 1998). Our results suggest that the species also relies on upland habitats opportunistically as our tracked curlew widely used the agricultural fields of Texas–New Mexico (the USA) and Durango–Coahuila (Mexico) during spring and autumn migrations.

In our study, spring migration (to the breeding grounds) occurred in a very short time span whereas autumn migration occurred over a much longer period; during both trips the curlew travelled similar distances but its strategies differed behaviourally. Apparently our curlew follows the type of migration strategy described by O'Reilly & Wingfield (1995) as intermediate distance bout, in which the distribution of ideal foraging habitat may not be regular, thus longer flights (100–2000 km) may be interspersed with relatively short distance flights.

Our results show that long-distance migration segments were more frequent during autumn migration than during spring migration in practically the same period of time (four days of flight to complete each migration), implying greater physical wear to the bird. We infer this is the reason why, surprisingly, the curlew spent 77% of the total time of tracking in wintering grounds (Mexico). An interesting fact that supports our theory is the time spent in the area of Durango and Coahuila (118 consecutive days on average) before reaching its final destination. Because the months involved (June, July, August, and September) are traditionally considered migratory windows, rather than wintering months, and because the curlew is not in a hurry to arrive and breed, it is very likely that this area is the last refuelling point before reaching final wintering sites.

Long-billed Curlews arrive in Texas by mid-March during the spring migration, and those reproducing in the northern portion of the species range (British Columbia and Alberta, Canada) arrive by mid-April. During the autumn migration, Long-billed Curlews depart from Canada by the end of August (Dugger & Dugger 2002, Fellows & Jones 2009). The curlew we studied followed the phenological pattern described during the spring migration but left the breeding grounds almost two months earlier than expected. This male was an adult when captured, so it is plausible he departed early, as some adults do, while juveniles are still developing prior to their own migration later in the season (O'Reilly & Wingfield 1995, Fellows & Jones 2009).

Partly because of the widespread concern for shorebird conservation, much attention has been paid to understanding their migration patterns (Marks & Redmond 1994). Our ‘Canadian' curlew showed practically the same migration and timing of curlews nesting in Montana (Page et al. 2014), which all stopped along their routes, with the longest stops occurring in western Texas and eastern New Mexico (the USA) and boundaries of Chihuahua and Coahuila (Mexico) during southbound migration, and in western Texas and eastern New Mexico while northbound. These results suggest that Canadian and US curlews might migrate together to Mexican wintering grounds and back north, which highlights the importance of establish effective conservation measures at a tri-national level.

ACKNOWLEDGEMENTS

We specially thank the World Wildlife Fund for donating the satellite PTTs, D. Jorgensen for assistance in the attachment of the PTTs, L. Tibbits and H. Rodríguez-Vela for assistance with PTT location data and mapping, Geoffrey Holroyd for PTT location in Canada, Clay Green for comments and suggestions to this manuscript.

FUNDING

Financial support was received from U.S. Fish and Wildlife Service (Neotropical Migratory Bird Conservation Act), agreement number MX-N1441 and World Wildlife Fund Mexico, agreement number 0K60.