New discoveries in landbird migration using geolocators, and a flight plan for the future

BIRD MIGRATION IS a spectacular natural phenomenon that has generated wonder and interest for centuries. Feats of migration inspire amazement—individual birds that weigh less than 200 g may log more than 80,000 km annually (Egevang et al. 2010), travel more than 600 km day-1 (Stutchbury et al. 2009, Åkesson et al. 2012), and cross huge geographic barriers such as oceans (Bairlein et al. 2012) and inhospitable deserts (Tøttrup et al. 2012b). Despite the vast geography covered during migration, many birds return to the same territories year after year. Although incredible progress has been made in our understanding of bird migration (Newton 2008), many gaps remain in our knowledge of the migration of small birds. The development of miniaturized tracking technology has produced a wave of research into the migratory behavior of small birds (Fig. 1). The inaugural application of miniaturized geolocators (or “geologgers”) on small songbirds in 2007 (Stutchbury et al. 2009) initiated a rapid increase in the number of studies of small landbird migration; there are currently more than 100 permits in North America alone for attaching geolocators to small birds. This technology has been so enthusiastically applied because it provides information critical to conservation and management of declining songbird populations (Faaborg et al. 2010a), as well as the opportunity to test long-standing hypotheses related to endogenous control mechanisms, navigation, and energetics (Robinson et al. 2010). Although more accurate devices may someday be available for tracking small birds, geolocators are currently the only option for migrants that weigh <50 g (Bridge et al. 2011). The main goal of many geolocator studies to date has been the description of little-known migratory routes and wintering sites (e.g., Beason et al. 2012, Stach et al. 2012). As this technique becomes more widely applied (both geographically within species and taxonomically across a broad spectrum of small landbirds), researchers can begin to test hypotheses about migration, non-breeding ecology, and behavior to inform conservation measures. Many migratory species are declining; thus, a comprehensive understanding of the annual cycle is timely and important for management of species at risk. The purpose of our review is to summarize, for the first time, patterns emerging from geolocator studies. We review new data on (1) migratory connectivity, (2) migratory routes and stopovers, (3) intratropical migration of wintering birds, and (4) migration schedules. We then explore questions that can be answered with emerging geolocator studies, and provide a “flight plan” for future work as direct-tracking technology becomes increasingly smaller and more broadly applied. Geolocators are archival light-recording devices that are mounted on the lower back (for most small birds) following the Rappole and Tipton (1991) leg-loop harness design. The tags record light levels in relation to an internal timer. This allows the determination of sunrise and sunset times and, thus, day length and solar midday and midnight upon retrieval of the tag from the bird after it has completed its migration, usually the next year. Latitude and longitude are estimated from these light data, typically with a computer program such as LOCATOR (British Antarctic Survey) or tripEstimation in R (Sumner et al. 2009). Locations are determined using the “threshold method,” which uses calibration to determine the average sun elevation angle (the angle of the sun on the horizon) when a sunrise or sunset transition is defined, or using a “template-matching” method (Sumner et al. 2009, Lisovski et al. 2012). For more details on analysis methods, see Bridge et al. (2011), Lisovski et al. (2012), and McKinnon et al. (2013). Since their deployment on the first migratory landbirds in 2007, geolocators have been used to track individual birds in the Palearctic-Tropical, Nearctic-Neotropic, and Austral migratory systems. Colors that frame the photographs match the colors that indicate migratory routes. One individual's spring migration is shown for each subspecies (two subspecies are shown for Purple Martin and Swainson's Thrush, and three for Northern Wheatear), except for Fork-tailed Flycatcher (yellow) and Thrush Nightingale (bright green), whose fall migrations are shown. Maps are modified from references in Table 1. Photo credits: Red-eyed Vireo, Wood Thrush, and Veery: Lang Elliot; Swainson's Thrush: Darren Irwin; Red-backed Shrike: Per Eckberg; Purple Martin and Fork-tailed Flycatcher: Harold Stiver; Yellow-billed Cuckoo: Karthryn Mann; Snow Bunting: Sebastien Descamps; Northern Black Swift: Steven Daly; Common Swift: Steve James; Eurasian Hoopoe, Northern Wheatear, and Thrush Nightingale: Mikkel W. Kristensen. Geolocator accuracy varies, depending on bird behavior, geographic location, habitat, and weather (Fudickar et al. 2012, Lisovski et al. 2012). Ground truthing at multiple sites prior to fall migration found that geolocators mapped 91% (83 of 91) of Purple Martins accurately to within 100 km latitude and longitude (Fraser et al. 2012; scientific names of species not provided in the text can be found in Table 1). Ground-truthing tests with stationary forest birds on their tropical wintering grounds have shown that geolocators can place birds within a few degrees of latitude (365 km) and <1° of longitude (66 km) of actual locations (McKinnon et al. 2013). These levels of accuracy are sufficient to elucidate unambiguous patterns of connectivity, migration timing, and so on for most species. Determination of finer-scale movements (e.g., within ∼100 km) and finer-scale mapping of birds while stationary at non-breeding sites is currently limited by technology. Geolocators, by relying on day length to determine latitude, are not able to determine the location of birds near the vernal and autumnal equinoxes (approximately 20 March and 22 September, respectively) when day length is the same everywhere. However, longitude during this time is still as accurate as at other times of the annual cycle (Fudickar et al. 2012). Aside from the accuracy and analysis issues detailed above (see also Fudickar et al. 2012, Lisovski et al. 2012, McKinnon et al. 2013), researchers have encountered several shortcomings of geolocator studies on small birds in terms of field work and study design. One of the first limitations for geolocator studies was harness and geolocator failure. This has become solvable for many species as geolocator models and harness designs continue to be refined on the basis of field and laboratory data (Bowlin et al. 2010). We encourage the publication of details of successful and failed geolocator attachment methods to improve future studies. Battery failure occurs even with the most reliable small geolocator models (e.g., 10–15% failure rate for British Antarctic Survey MK 10 and 16 on songbirds). Researchers must take into account not only return rates of the individual birds, but also potential rates of harness or geolocator failure when determining how many geolocators should be deployed. Another important issue to consider is the impact on the study species. Most geolocator studies assume little impact on survival or behavior of the birds tracked. Although true tests of effects of geolocators on migratory behavior are not possible, evidence suggests that return rates of birds with geolocators are not significantly lower than those without geolocators (E. S. Bridge et al. unpubl. data). For sensitive species, low returns may be solvable through changes in geolocator or harness design (e.g., Purple Martins had very low return rates in the first 2 years, but shortening the light stalk of the geolocator solved the problem; B. J. M. Stutchbury et al. unpubl. data). Pilot testing with dummy geolocators and various harness types is a low-cost way to establish whether negative effects occur before time, effort, and money are invested into real geolocator deployments. In some cases, the retrieval rate of birds wearing geolocators is low because of low site fidelity (i.e., deployments on juveniles or at stopover sites). Whether or not deployment on species or demographic groups with low return rates is valuable or ethical depends on the study species and questions. In sum, the collective experience gained around the world from tracking small birds with geolocators in the past 5 years means that most researchers should be able to confidently proceed with geolocator tracking of small birds because they can (1) identify appropriate questions, species, study sites, and methods needed to get migration data; (2) understand the inherent limitations of geolocators (battery failure, temporal and issues with latitude, and data from (3) that they effects on birds and studies in and (4) return and retrieval rates in to determine a of and for the questions One of the of tracking migratory birds and wintering is to determine migratory is from a as the of in wintering locations of individual birds from geographically populations et al. and The of populations in has for and 2007, and and, conservation of migratory birds et al. 2007, et al. 2011). patterns of migratory have been determined for some species by using analysis of at the site of interest and et al. et al. by mapping et al. or by using a of these et al. 2012). Geolocators can patterns (i.e., locations of birds within km) in most cases, be by or which can only birds to broad available in the or to the of groups et al. 2011). of data is the North et al. 2011). However, for many species, they are to about migratory (e.g., Northern Black et al. and Purple Martin et al. For geolocators to provide information on the of migratory connectivity, birds should be from multiple wintering sites to determine the of of populations at in the annual cycle (Fig. deployment of geolocators at such a broad has been see et al. 2012, et al. et al. 2013). However, geolocators are in for populations (or Table and can be used in with other data, such as analysis of et al. 2012) and et al. to provide a of species using the Wood Thrush and Purple in Wood from a in in a small of the wintering grounds in and (Stutchbury et al. 2009, 2011). from other sites and three wintering sites a of populations and to in the of their birds and in and and and birds (e.g., and North to in the and et al. unpubl. data). of migratory patterns using geolocator within subspecies of Purple Martin shown by to indicate multiple but across subspecies shown by (Fraser et al. and three populations by of Red-backed to the same wintering (Tøttrup et al. 2012b). This how multiple populations are needed to determine there was little in of the subspecies of Purple Martin et al. 2012). from across the during the non-breeding et al. 2013). In Purple birds from across the found in the same wintering in the (Fig. et al. 2012). from a on km and had with birds from km This of populations at wintering sites is a of migratory However, at the subspecies Purple Martins Purple Martins from the North subspecies that using geolocators had a wintering in (Fraser et al. 2012) that not at with the wintering of the subspecies (Fig. These the of and in patterns of et al. found connectivity, also at the of in Snow with birds wintering in through and birds wintering in the alone suggests a of within multiple populations from a broad wintering site in However, geolocator data from the a migratory at the species a broad of et al. 2012). et al. subspecies of Swainson's Thrush at a migratory and found that subspecies by km at their sites had migratory routes and wintering in Snow a of was within subspecies of Swainson's birds with geolocators from a in had a wintering in relation to birds from British et al. 2013). Most studies using geolocators have on of connectivity, such as the broad patterns However, as more data are of can be migratory in Purple et al. used to the wintering birds from the same populations and for and wintering latitude and and wintering et al. used returns and to of individual at and sites, testing whether the of birds in was or with the of birds in the very using geolocator data from Swainson's et al. 2013). of patterns and for more which to understanding of the of migratory behavior and, of patterns of in species that are small for geolocator The of individual migratory routes and stopover sites, from to is of the most and of geolocators to the study of migration in small birds. Geolocators have geographically and patterns of migration within species (Tøttrup et al. and (Bairlein et al. 2012), and spring (e.g., Red-eyed et al. and fall (e.g., Wood Stutchbury et al. 2011). Many species using geolocators have shown in migratory routes (i.e., migration; Table 1). For most Wood fall migration routes of spring migration routes et al. 2012). In the Red-backed from three sites in also a of migration, with in fall in relation to spring routes (Tøttrup et al. 2012b). These patterns may be related to patterns or in the of stopover et al. these hypotheses have not been in small species in migratory routes. and subspecies of Swainson's Thrush in migratory et al. 2012), as and subspecies of the Northern (Bairlein et al. 2012). be but within populations have been as Eurasian from the of their very migration of a migratory et al. 2010), and Northern from also in their migratory et al. in migratory routes is also in the migration of from a across the of in three the from to et al. 2011). is whether (or in migratory routes of in a is because of endogenous control (i.e., and or is a to et al. weather effects on migration by Northern and found that and migration in fall but not in of individual Wood that of 10 birds used a spring at the of from year to at a geographic longitude when was not et al. 2012). studies (i.e., tracking the same in more than using geolocators are because of the of small birds and, thus, the number of geolocators needed to a sufficient of design that is very important for questions about Geolocators have also that some species take very et al. Stutchbury et al. Åkesson et al. 2012, et al. Table these the time for migratory on the basis of models of stopover that sites stopover because birds can and migration (Newton 2008), but geolocators have to a new on stopover In the of these has into the of and researchers have made the that than needed to should be that are as important for conservation as and wintering et al. 2012, et al. 2012b). at sites with be they improve migratory or on at sites (Newton et al. 2011). Swainson's and Red-eyed are and on a that is by many migrants at tropical sites before and during spring migration and stationary during spring migration in of these species et al. 2012, et al. be by These sites may be important for for migration, as in but this to be Another made with geolocators is the of intratropical migrations of migratory birds within their 1). Although there was evidence that some species are in (Newton 2008), data from geolocators that multiple in may be 1). For of Purple Martins more than and some had to sites km (Fraser et al. 2012). movements studies because the of with the non-breeding as birds The of multiple wintering sites is important for conservation and management of species at risk. these sites and the and to sites remain the and of is in as by a study of the tracking that birds most of their time at sites of and non-breeding sites for birds in the can be to the more widely intratropical and migration of tropical species (Faaborg et al. to determine this is in some species or and not to the deployment of geolocators on small migration was estimated using or more on migration (e.g., et al. et al. One of this is that multiple populations of birds are and, temporal changes within populations are not (Newton this actual migration rate of be is a method used to migration birds are in location and in is that when birds or at each location is which of migration and such data are for many species and only a of Despite location with geolocator geolocators currently provide the most accurate method to the migration rate of individual birds from to and during spring and fall In this migration and or and is usually using migration by the number of on migratory knowledge of stopover and migratory routes and is more using geolocators for species that with light data, such as the Purple The first geolocator study of a songbird that migration in was than estimated using other methods (Stutchbury et al. 2009). Purple Martins from America to the at km and Wood from America at km for spring migration in songbirds km day-1 (Newton studies using geolocators have that many species typically travel km day-1 on spring migration (Fig. has been that birds at a rate (Newton data from Table migration in spring and fall with migration using a and found We also the that is related to with birds (Newton 2008), but birds not than smaller birds can also spring and fall migration rates of individual birds and species for the first time (Fig. a found that spring migration rate is significantly than fall by species Many can migration or but it is that birds at rates when they are of time to their (Newton et al. 2012b). However, the of in spring not migration because birds may also at by tropical wintering sites date from in many cases, the of date at sites, with in rate little to in et al. 2012, et al. et al. et al. more geolocator studies, be able to the of for a (e.g., Purple Martin and Wood or a (e.g., Red-eyed and Swainson's rate of spring also be important to establish or spring date and, thus, from on date may be related to and et al. and 2011). is possible, with to test how individual and at wintering sites, not only date but also spring migration rate and migration and rate also remain and in are of fall migration for some species (e.g., et al. We can also geolocators to the of of is whether birds can to at wintering sites or on migration, or whether these are endogenous control et al. 2011). This is important to in the of because can be when of migration not with et al. geolocators to track the same individual Wood in multiple years in of spring migration, with tropical non-breeding sites, on within of in years et al. 2012). of migration suggests that little individual in the migration of some species and the of endogenous Red-backed and Thrush found to their spring date at geolocator tracking that this was in to an and a stopover by the birds at a stopover site in the of (Tøttrup et al. This suggests that weather may changes in migration of timing, and of other migration be tracking of and but with geolocators deployments and a species with return Geolocators have the first detailed migrations of small Although many studies are on small they have such as very in fall and intratropical and very rapid migration rate and The number of direct-tracking studies for studies to test for the that the of these for individual species also more and to explore migration behavior (e.g., Stutchbury et al. et al. 2012) and conservation (Fraser et al. 2012). migration migration provides a that can be using data from individual of small migratory birds. the migration be to stopover transition from migration to and and migration, and and schedules. For migration can be used to and of or which birds should take to et al. applied migration to study the migratory and routes of Northern using In the of to stopover with and the birds also to and in a with However, to migration Northern not using the routes and sites, and in spring they not to such as and during fall and spring in other studies (see above and Table also not to migration migratory birds to be in many which a of the rate of migration and stopover is that geolocators can data only for and, thus, it is not to understand in migration spring migration rate and fall migration rate for birds using in of average indicate km day-1 for spring and km day-1 for without are from that not or data with the of the and Yellow-billed which are as and Eurasian Hoopoe, Yellow-billed Purple Wood Thrush, Northern Black Common Snow Swainson's Thrush Swainson's Thrush Red-backed Fork-tailed 5 Thrush 2 Northern Northern Northern and patterns in and migratory be to migration and for some demographic with for our understanding of and for conservation of and allows for tests of that patterns such as in spring migration and as well as patterns such as in fall of birds on their first fall migration (i.e., from the is currently et al. because return rates are in most species, but tracking of first spring migration is more in the wintering sites of migratory birds allows the deployment of geolocators on birds before their first spring migration and it to field tests of hypotheses related to and of migratory birds. As Wood on their first spring migration from the and to sites than et al. the details of migration stopover behavior, or and migratory have to be using of and intratropical that multiple sites and have been in many species hypotheses related to the and of these can be et al. provided evidence that intratropical movements of to changes in within their non-breeding et al. potential and intratropical of and that may to this in Stach et al. found a and intratropical of Thrush which suggests that birds are tracking or as they the has been that may wintering movements of an the Fork-tailed Flycatcher et al. and of the migrants and et al. These hypotheses are by weather patterns and bird movements using provided by (see Bridge et al. et al. 2011). The of this was by a of Northern movements with and et al. 2012b). studies should consider the of which data and analysis as well as tracking data available to the for or research can studies to test hypotheses and geographic in migration Åkesson et al. a in the spring migration of Common using they that this on the of in a small of before the birds the of in not in Common et al. 2013). stopover may be important sites for et al. 2009, et al. et al. that spring in Snow may be birds to in groups to by As with multiple sites, understanding the and of for migratory birds inform conservation and management for migratory species and studies at these sites to of and risk. Geolocators as a conservation are for more species, it be important to the effects of these and patterns in (or other on and 2010). tracking and mapping can also to determine and in the non-breeding to test hypotheses that to non-breeding (e.g., et al. 2012). et al. that birds with be more to than those with patterns because they are to little their to an to a also of species to and on the wintering grounds because populations to with rates of have few from other wintering sites to populations from et al. For the first time, geolocators populations to be mapped so that (e.g., and can be at the and wintering grounds of the same to understand and to take conservation at and wintering sites of the same The migrations of small landbirds are to be the most of the annual in species, data estimated that of annual during migration and occurs during migration, and is a critical conservation Geolocators provide information on birds that not migration, but birds that provide data on and when and which are migration For many Wood the in spring through a very of longitude near (Stutchbury et al. 2009, et al. 2012), which suggests that this site is important for conservation of this declining species. 10 Red-eyed by et al. also made in this small of and from three widely populations also a stopover in this et al. 2013). We can also migration routes of widely populations and the and of (e.g., and For a for using with an in birds have been et al. 2013). We can also test whether migration is with as is (Faaborg et al. and 2010). new of bird migration research has been with the of miniaturized The geolocators are small in to be on so the of migration data is each year. Despite small geolocators have about migratory connectivity, migration non-breeding sites, and migration For a of geolocator returns for the Common more information about migration and non-breeding sites for this species than 100 years of bird et al. 2012). Researchers should be that many of the hypotheses using geolocators also be to studies using other direct-tracking technology. Aside from new hypotheses and questions, data from geolocators have provided a for conservation of declining migratory birds. conservation of migratory landbirds was limited by the of the migration (Faaborg et al. have an for determining of species (e.g., Beason et al. 2012, et al. 2012) and for testing hypotheses related to during the non-breeding (Fraser et al. 2012). We in the North on this and M. for about this and is available at

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.