Research Paper

INTRODUCTION

Effective conservation and management of migratory bird populations requires an understanding of the components of full life-cycle dynamics (Marra et al. 2015a). Migration is a crucial component of the annual life cycle of many bird species and often results in higher mortality rates than during breeding and wintering periods (Sillett and Holmes 2002, Newton 2006, Paxton et al. 2017, Borowske et al. 2018). Although numerous studies have found that intrinsic and extrinsic processes that occur across the migratory landscape can affect individual survival and fitness (Ward et al. 2018), conservation strategies continue to focus on breeding and wintering priorities (Schuster et al. 2019, Lin et al. 2020). The entwined nature of migration with other life-cycle components makes it vital to understand the spatial and temporal patterns of migration. Only with that understanding can we answer questions about how to identify, prioritize, and protect those places used by migrants en route, and inform the creation of realistic and comprehensive conservation strategies (Moore et al. 1995).

Technological advances have allowed researchers to use VHF or UHF telemetry to track long-distance movements of small passerines (< 50 g) with automated radiotelemetry (Mills et al. 2011, Taylor et al. 2011) within the Motus Wildlife Tracking System network (www.motus.org). Motus comprises a collaborative, continental-scale network of automated telemetry receiver stations that can detect transmitters on many individuals simultaneously with moderate (± 1–15 km) positional accuracy (Taylor et al. 2017). Detections of tagged animals are not limited by personnel availability, site access, or time constraints, although there are still constraints due to distribution of receiver stations within the network, tag distance from receiving stations, tag output and burst interval, anthropomorphic (electromagnetic interference, buildings) and natural (mountains, weather) features of the landscape, and animal behavior (Taylor et al. 2017). Application of the automated tracking technology to threatened species can potentially identify vulnerabilities and critical needs that have previously been difficult to study, such as migratory routes, phenology, and the presence, location, and duration of stopovers. By filling key information gaps in our understanding of migration, we can better inform conservation strategies to protect a species across its full life cycle.

The Saltmarsh Sparrow (Ammospiza caudacuta) is an obligate species of tidal salt marshes (Greenlaw et al. 2020), a globally vulnerable habitat. With rapid population declines throughout its range (Correll et al. 2016) and extinction predicted to be possible between 2030 and 2060 (Field et al. 2017, 2018, Roberts et al. 2019), the species is identified as globally Endangered by the International Union for Conservation of Nature Red List (BirdLife International 2023) and is being considered for Endangered Species Act protection in the USA. Saltmarsh Sparrows are short distance migrants that occupy discontinuous patches of coastal marsh along the North American Atlantic coast. This extreme habitat specialization limits the extent of their movements during migration (Greenlaw et al. 2020). Breeding populations occur coastally from central Maine to the lower Chesapeake Bay in eastern Maryland and northeastern Virginia, with birds spending the non-breeding period from Cape Cod, Massachusetts, south to Florida (Greenlaw et al. 2020).

Knowledge of the Saltmarsh Sparrow’s breeding and wintering ecology has grown substantially in recent years; however, much is unknown about its migratory behavior (Greenlaw et al. 2020). Departing from their breeding grounds after completing their post-breeding molt in mid-September to late-October (Borowske et al. 2017, Greenlaw et al. 2020), most Saltmarsh Sparrows migrate south between September and December and start to move north again in March (Greenlaw et al. 2020). Banding data suggest that birds from across the breeding range mix throughout the wintering range with little evidence for population segregation (Borowske 2015), suggesting low migratory connectivity. Precise migration routes, timing, speed, and use of stopovers remain poorly known, however, despite evidence that adult mortality occurs primarily during migration (Borowske et al. 2018).

Considering these uncertainties and the need for action to protect Saltmarsh Sparrows, we applied automated radio telemetry to describe fundamental features of the species’ migratory ecology and behavior. Specifically, our objectives were to (1) document the timing, routes, and speed of their southbound and northbound migrations, (2) determine whether these differed among populations in different parts of the species range, and (3) identify areas and timing of migratory stopovers. Improving understanding of migration in these ways has been identified by management agencies as necessary to guide land protection and management decisions, including actions such as the siting of near-shore wind turbines.

METHODS

Field methods

Saltmarsh Sparrows were captured with mist nets during July-October of 2014–2016 at known breeding sites in Maine (York County), Massachusetts (Essex County), and Rhode Island (Newport County); and during March-April of 2016–2017 at locations in South Carolina (Kiawah Island, Charleston County) and Georgia (Glynn and Camden Counties; Fig. 1, Table 1; Appendix 1, Table A1.1). Most fall deployments occurred in September–October, but earlier deployments were used in our 2014 pilot year. Trapping locations were known breeding and nonbreeding sites where automated telemetry stations already existed or where we could easily establish new stations nearby.

Fall departing birds were aged as adults or hatch years based on molt and plumage patterns or skull pneumatization, and spring birds were aged to second year or after second year based on molt when possible (Pyle 1997; Appendix 1, Table A1.2). We collected standard morphometric measurements from all individuals, including tarsus length, wing chord, and body mass. Many (31%) captured individuals had been previously banded for breeding studies, during which females were sexed based on the presence of a brood patch and males based on a cloacal protuberance. Newly captured individuals were banded with a uniquely numbered United States Geological Survey aluminum leg band. We collected breast feathers or blood samples from birds that could not be sexed in the field (non-breeding adults that had not been previously banded and hatch year birds); blood samples (10–20 µl) were collected from the brachial vein and transferred to Nobuto blood filter strips (Sterlitech, Kent, Washington), then stored at room temperature. Molecular sexing of individuals was performed on DNA extracted from blood and feather samples by PCR amplification of the CHD1 gene using primers P2/P8, following methods developed by Fridolfsson and Ellegren (1999), as in Hill et al. (2013) and Benvenuti et al. (2018). Poor quality feather samples and missing samples resulted in numerous individuals with undetermined sex, causing small sample sizes of known sexes across seasons and capture locations and precluding investigations into migration differences between sexes (Appendix 1, Table A1.2).

Individuals ≥ 15.2 g were outfitted with coded VHF (166.380 MHz) transmitters (tags) with attachment tubes (0.45g; NTQB-3-2; Lotek Wireless Inc., Newmarket, Ontario, Canada) and a leg-loop harness (Rappole and Tipton 1991, Streby et al. 2015), such that tag weight did not exceed 3% of their total body weight (Barron et al. 2010). In 2014 and 2015, we constructed leg-loop harnesses with 1.0 mm fabric elastic cord (Mandala Trading Inc., Austin, Texas, USA). In 2016 and 2017, we constructed leg-loop harnesses from 0.5 mm elastic sewing thread (Gutermann, Gutach-Breisgau, Germany) based on Streby et al. (2015). Each tag emitted unique digitally coded VHF transmissions every 9.7–10.1 seconds, that allowed automated telemetry stations to detect multiple tags within the same frequency range broadcasting simultaneously. The manufacturer estimated a tag lifespan of about 64 days with the pulse interval used in this study; we observed a maximum tag detection period of 79 days.

From July 31-August 22 of 2014, we deployed tags only on known reproductive males (n = 18 of birds tagged in this initial year), avoiding females to minimize risk to the population if tags affected survival. Males were identified by the presence of a cloacal protuberance at their time of banding. This pilot component aimed to evaluate if Saltmarsh Sparrows could safely carry tags and avoid entanglement. Finding no major issues, we deployed 7 additional tags in September of 2014 and an additional 149 tags on adult and hatch year birds of either sex over the 3 subsequent years (2015–2017; Table 1; see Appendix 1, Table A1.2 for age and sex distributions of tagged birds). Tags deployed in 2014 had an 18 cm antenna constructed of an uncoated metal alloy. The manufacturer changed the antenna material to a plastic-coated metal alloy for 2015 deployments due to material availability, which resulted in several bird entanglements. For 2016 and 2017, we used a thinner nitinol antenna because the original uncoated metal alloy antenna was no longer manufactured. Given known problems with entanglement and tag removal and damage by birds that spend time moving through dense grasses (Hill and Elphick 2011), sparrows were manually tracked at their breeding capture sites and at the South Carolina wintering capture site for up to several weeks following release in 2014–2016 deployments, using a Lotek SMX 800 (Lotek Wireless Inc., Newmarket, Ontario, Canada) to check for tag loss, bird entanglement, or mortality. Manual tracking was not feasible at Georgia wintering locations due to the challenging logistics of accessing capture sites (e.g., boat access).

Receiver array and tag detections

The Motus network of stationary automated telemetry receiver stations was used to monitor movements of individuals that were captured, radio-tagged, and released at the study sites (Fig. 1). Generally, the spatial and temporal distribution of receivers in the Motus network increased throughout the study as the network expanded, although receiving stations occasionally were removed or became unavailable for various reasons. Most stations used three or more high-gain Yagi antennas positioned at various orientations (Taylor et al. 2017). Under optimal conditions, stations can detect a tag on a flying bird from 12–20 km away (Taylor et al. 2017, Morbey et al. 2018). When birds are on or near the ground, however, typical detection range is reduced to 0.5–2 km (Taylor et al. 2011; A. D. Smith, personal observation). Data collected from the automated telemetry system allowed for continuous passive monitoring of all deployed tags.

Data analysis

Processing automated telemetry data

We processed data using methods similar to prior Motus studies (e.g., Crysler et al. 2016, Duijns et al. 2017, 2019). We eliminated false detections during post-processing by considering several derived metrics of detection structure related to a tag’s frequency, burst interval, and other signal qualities, as well as the noise context of the receiving station and other valid detections of the tag. We assigned each tag deployment to one or more “home” receiving stations located within 10 km of the capture location. Tag deployments were associated with 0–3 home stations (median = 1). Twenty-six tag deployments on wintering sparrows in South Carolina and Georgia in 2017 did not occur within 10 km of an active station and were not assigned a home station; these birds could not contribute to analyses related to departure but could contribute to other analyses. Individuals that were never detected beyond the tagging location (with or without a home station) were excluded from all analyses.

Departure timing

Dates of departure from breeding or wintering sites were estimated for each individual using the last time stamp at the home station (cf. Mills et al. 2011, Taylor et al. 2011). Departure dates were only calculated for sparrows detected at a home receiving station and another station on the same night. For 40 individuals with known night of departure during fall deployments, we compared departure date among fall tag deployment locations using a linear mixed model with deployment year as a random effect. Following a likelihood ratio test evaluating an overall effect of tag deployment location on fall departure dates, we estimated pairwise comparisons between locations, adjusting degrees of freedom for multiple comparisons using Kenward-Roger approximations with the emmeans package (Lenth et al. 2023) in R (version 4.2.2; R Core Team 2022). We lacked sufficient departure data to perform an equivalent analysis for spring departures.

Travel routes and stopovers

We inferred individual travel routes from detection histories (i.e., detection locations, dates, and times) against a backdrop of receivers active during the year of study (Fig. 1). We visualized migration paths for each individual by assuming birds moved directly between the locations of consecutive detections. Typically, this approach produced simplified flight trajectories unlikely to represent actual fight paths (e.g., broken lines in Fig. 2), but useful for visualization of movements. In many cases, however, patterns and timing of detection, as well as derived estimates of ground speed, allowed us to presume direct or nearly direct flights between detecting receiving stations (e.g., solid lines in Fig. 2).

We defined stopover locations as detections at a single station, other than a home station, spanning > 8 hours with no intervening detections elsewhere, or spanning > 10 hours among multiple stations located within 30 km of each other (cf. Crysler et al. 2016), based on a maximum detection radius of ~15 km for a given Motus station. We considered as ambiguous any stopover events that occurred when a bird was not subsequently detected at other locations. We excluded cases in which a bird was known to have stopped for more than 10 days and was within its potential wintering range, on the basis that these birds might have reached their destination making it impossible to assess whether they were true stopovers. Although such time frames are necessarily arbitrary and might underestimate average stopover lengths, we based ours on the maximum documented stopover time for a closely related species to minimize that risk (Crysler et al. 2016). To visualize areas used for stopovers, we buffered by a 15-km radius the stations associated with stopover for each individual, then overlaid and summed these buffers across individuals to generate a heat map of stopovers for each tagging location. For fall migration, data were sufficient to evaluate changes in stopover duration over the course of the migration with a lognormal linear mixed model. We included tag deployment year and individual (four individuals contributed two stopovers) as random effects, although the random effect variance could not be estimated for individuals and was removed from the model. Including only the first detected stopover from the four individuals contributing multiple stopovers did not substantively change model estimates, so we retained the full stopover duration data.

Speed of travel and wind assistance

We estimated migration speeds for Saltmarsh Sparrows during migratory flights from detection times at receiving stations separated by at least 150 km (to exclude local movements and reduce bias from uncertainty in an individual’s precise location during a detection; Duijns et al. 2019), and we restricted estimates of migration speed to flights of 12 hours or less. These constraints produced 46 flight trajectories from 40 individuals for analysis of travel speed and wind assistance.

We calculated total trajectory length as the shortest orthodromic distance connecting all detecting receiving stations passed during the flight between the beginning and ending receiving stations, and trajectory displacement length as the shortest orthodromic distance between the beginning and ending receiving stations. We calculated net ground speed for a flight as the trajectory displacement length divided by the time between the last detection at the beginning receiving station and the first detection at the ending receiving station. We note that these are not exact ground speeds due to variability in detection ranges and distances among stations. We permitted an individual to contribute multiple flights (i.e., 40 birds contributed 46 flights), provided there was no receiver station overlap in the trajectories. Excluding multiple flights for an individual did not alter the conclusions so we reported from the full flight trajectory data set. For each migratory flight, we also estimated tailwind support at departure using surface wind conditions (i.e., 1000 hPa pressure level) at the measurement time closest to departure. We assumed the “preferred” direction of movement (as required by the calculation of tailwind support) was the bearing that would lead the bird to the ending receiving station along a great-circle route. We evaluated the correlation between net ground speed for a migratory flight and the estimated initial tailwind support at departure. We estimated trajectory lengths using the trajr package (McLean and Volponi 2018) and acquired wind component data for each migratory flight from the NCEP/NCAR Reanalysis project (Kalnay et al. 1996) via the RNCEP package (Kemp et al. 2012).

All linear and generalized linear mixed model analyses used the lme4 package (Bates et al. 2015; version 1.1.23). Prior to inference, we assessed model residual and other diagnostic plots to confirm an adequate model fit (e.g., normality and homogeneity of variance in residuals) with the DHARMa package (Hartig and Lohse 2022). Sample sizes for the various analyses are summarized in Appendix 1, Table A1.3.

Recaptures

Due to the high breeding site fidelity of Saltmarsh Sparrows, we expected that surviving individuals would return to the sites where they were tagged. We used targeted mist-netting to attempt to capture previously tagged birds during the breeding seasons of 2015–2017 to assess tag retention and potential tag effects on individuals.

RESULTS

We fitted 174 Saltmarsh Sparrows with tags, 136 tagged on the breeding grounds and 38 on the wintering grounds (Table 1; Appendix 1, Tables A1.1-A1.2). After post-processing, we retained more than 237,000 detections from 122 individuals (70%); 52 individuals were not detected on any stations nor met the data validation standards. We removed an additional 34 individuals from the analysis because they lacked detections beyond their tag deployment location and did not contribute departure or movement data: 17 tags deployed in July or August as part of our 2014 pilot (batteries failed before migration), 7 in each of fall 2015 and 2016, 2 from spring 2016, and 1 from spring of 2017. Among southbound deployments, in which all 136 individuals were associated with home stations, individuals detected only by their home station tended to be those with transmitters deployed early in the fall and whose last detection dates were earlier than those detected during migration, suggesting that battery failure prior to departure may explain the lack of detections at other sites in many of these birds. The data used for migratory analysis represented 68 Saltmarsh Sparrows from the breeding grounds and 20 from the wintering grounds (Table 1). On average, birds used in the analysis were detected by 5.7 (range 1–16) receiving stations during the life of the tag.

Departure from tagging location

We could determine the exact night of departure for 44 individuals (40 fall, 4 spring). Fall migration commenced between early October and mid-November (Fig. 3). Departure dates differed among the three tag deployment locations (likelihood ratio test: χ²₂ = 18.4, P < 0.001), with Maine departures averaging 10–15 days earlier than Massachusetts and Rhode Island departures (t_36.5 = -4.17, P < 0.001; t_36.6 = -3.33, P = 0.008, respectively; Fig. 3); Massachusetts and Rhode Island departure dates did not differ (t_35.5 = 1.17, P = 0.48). We detected apparent exploratory fall movements by two adults and one juvenile, which were detected at nearby receivers before returning to their home site. Northward movements started as early as mid-April from Georgia locations and continued into mid-May from both Georgia and South Carolina.

Travel routes

During southbound fall migrations, only 10 of 46 Maine and Massachusetts Saltmarsh Sparrows were detected on Cape Cod, Nantucket, or Martha’s Vineyard despite excellent coverage from operational receiver stations, suggesting that most birds took an overland route to the southern New England coast. From there, Maine and Massachusetts birds followed similar southbound routes to Rhode Island birds (Fig. 2A). Detection histories, including absences from receiving stations known to be active and detecting other animals, supported with ground speeds estimated from detection data, suggested that many southbound Saltmarsh Sparrows made overwater movements between southern New England and the mid-Atlantic (Fig. 2A; Appendix 1, Fig. A1.1). Detection histories also suggested that north and southbound sparrows may have migrated inland over North Carolina, bypassing much of the Outer Banks (Fig. 2B), although detection data in these cases was insufficient to corroborate with ground speed estimates (Appendix 1, Fig. A1.1). Similarly, some northbound Georgia birds were not detected at active coastal stations in South Carolina that detected other individuals, suggesting that they may have followed a more direct inland route to the mid-Atlantic (Fig. 2B).

The differences in departure timing observed among fall migration tagging locations were also reflected in geographic (latitudinal) and temporal patterns of southbound detection (Fig. 4). For example, we observed Maine individuals leaving earliest and often reaching more southerly locations ahead of Massachusetts and Rhode Island individuals, providing some evidence that Maine birds bypass the Massachusetts and Rhode Island birds, at least temporarily, during migration (Fig. 4D–F). Geographic and temporal patterns of detection were similar among Massachusetts and Rhode Island birds, although some Massachusetts birds were among the last to make southbound movements (Fig. 4D–F). During spring, Georgia individuals often departed earlier and proceeded north ahead of birds tagged and departing from South Carolina locations (Fig. 4B–C).

Number of stopovers, stopover locations, and duration

We identified the general location of 21 stopover events involving 17 Saltmarsh Sparrows during fall migration (Fig. 5A–C) and 4 stopover events from 4 sparrows during spring migration (Fig. 5D). Thus, only 25% of individuals with detections away from the tagging location provided information on stopover locations during the fall, and only 9% of individuals in the spring. We excluded from analysis 3 sparrows with stopovers longer than 10 days, with no subsequent detections in the network, or within their potential wintering range, primarily because we could not be sure that these individuals had functional transmitters or were still alive. These were an individual tagged in Maine that stopped along the western Connecticut coast from 14 Nov–1 Dec 2016, another individual tagged in Maine that stopped along the central New Jersey coast from 8 Nov–1 Dec 2016, and an individual tagged in Rhode Island that used the Rhode Island coast from 4–16 Nov 2015. We also excluded three individuals that appeared to show exploratory movements as noted above: one tagged in northern Massachusetts and detected for less than a day at a station along the southern Massachusetts coast in early December, and two Rhode Island individuals detected at stations near the tagging location but subsequently detected back at their home station. In fall, individuals from all studied breeding populations made stopovers on the coast of Connecticut and Rhode Island, particularly between Hammonassett Beach State Park, Connecticut, and Trustom Pond National Wildlife Refuge (NWR), Rhode Island (Fig. 5A–C). Maine breeding birds also made multiple stops after a short movement to Parker River NWR in northeastern Massachusetts (Fig. 5A). We detected stopovers of multiple Massachusetts breeders in southeastern New Jersey near Edwin B. Forsythe NWR and the North Brigantine Natural Area (Fig. 5B). Among northbound birds, we detected multiple stopovers along the Delmarva Peninsula from Prime Hook NWR, Delaware, to Assateague Island, Virginia (Fig. 5D).

Most fall stopovers lasted < 2 days (median 37 h, range 15–235 h) and all 4 recognized spring stopovers lasted < 1 day; Fig. 6). Stopovers initiated later in fall migration were shorter than those initiated earlier (likelihood ratio test: χ²₁ = 5.9, P = 0.01), declining by roughly 32% (95% CI: 10–47%) per week throughout the fall (Fig. 7).

Speed of travel and wind assistance

We quantified migration speeds for 39 fall migratory flights from 33 individuals and 7 spring migratory flights from 7 individuals. The median trajectory length was 262 km (range: 156610– km) and the median straight-line distance between beginning and ending receiver stations was 244 km (range: 152–424 km). All 46 migratory flights occurred at night, with the initial detection occurring after local sunset at the initial receiver location and the final detection occurring before local sunrise at the final receiver location. Mean initial flight detections were 2.3 h after local sunset (SD: ± 1.5 h; range: 1.0–7.1 h), and mean flight duration was 5.4 h (SD: ± 2.1 h; range: 1.8–9.8 h).

Median net ground speed was estimated to be 15.3 m/s (55 km/h) for fall flights and 17.3 m/s (62 km/h) for spring flights (Fig. 8A). Net ground speeds correlated significantly with tailwind support in the fall (n = 39, Pearson’s r = 0.34, p = 0.035; Fig. 8B). In the spring, a similar pattern was apparent, but the small sample size made it difficult to draw clear conclusions (n = 7, Pearson’s r = 0.40, p = 0.38; Fig. 8B). However, 93% of all measured migratory flights were associated with tailwinds at departure.

Recaptures, entanglements and dropped tags

We recaptured 7 individuals in the breeding season following fall deployment (2015 deployments: 2; 2016 deployments: 5). Beginning in 2016, we switched to a thinner diameter elastic thread to construct harnesses, expecting it to reduce chafing and irritation and increase the probability of tag loss after battery failure (~70 days). The thinner elastic thread did not appear to increase the likelihood of the harness coming off or reduce wear. All recaptured birds retained their nanotags with harness for at least eight months and exhibited moderate chafing and feather loss on the femur and back. All tags had been destroyed, presumably by the birds, most commonly via broken antennas or removed batteries. Six (86%) of these birds had been detected by receiving stations. In 2017, we recaptured a female Saltmarsh Sparrow twice after tag deployment, with five weeks between recaptures. The initial recapture revealed chafing and feather loss on the thighs, but upon the final capture feathers had regrown and appeared in a condition similar to other untagged individuals. Two individuals were recaptured one year post deployment on the wintering grounds. In both instances, the birds had lost the tag and harness, and did not show any signs of tag wear or irritation, suggesting an expected lifetime of 8–12 months for 0.5 mm elastic thread in this salt marsh environment.

We discovered no entanglement problems during the 2014 pilot year (uncoated metal alloy antenna) or in 2016–2017 using a thinner, uncoated nitinol antenna. In 2015, however, manual tracking led to the discovery of a bird caught in the vegetation after its release, causing us to expand our manual search efforts. Tag antenna design changed from an uncoated metal antenna in 2014 to a metal alloy antenna of similar diameter but with flexible plastic coating in 2015. Of the 52 birds fitted with tags with plastic-coated antennas, 4 became entangled when the tag antenna twisted around marsh vegetation forming a knot or kink. One individual was found alive and was released after removing its harness and tag. One tag was found entangled in the vegetation with no sign of the bird, which we suspect successfully freed itself from the harness. Two entanglements resulted in mortalities.

DISCUSSION

Our use of automated telemetry has enabled the most detailed analysis of Saltmarsh Sparrow migratory movements to date. We revealed new insights about the species’ migratory routes and strategies, timing and speed of migratory movements and flights, and areas of stopover for southbound and northbound individuals. Patterns of detection indicated that longer Saltmarsh Sparrow flights regularly occurred over water or inland away from coastal areas, suggesting that Saltmarsh Sparrows use straight-line movements, resulting in a shorter traveling distance. However, detections clearly supported the notion that stopovers require coastal habitats for this obligate user of coastal salt marshes. Ensuring adequate salt marsh habitat occurs at regular intervals along the Atlantic Coast will be essential for Saltmarsh Sparrow persistence and recovery. The projected magnitude of marsh losses (Ganju et al. 2017, Watson et al. 2017, FitzGerald and Hughes 2019, Klingbeil et al. 2021), particularly when viewed through the lens of historical salt marsh loss (Gedan et al. 2009), suggests this may become a serious issue in the next three decades.

Southbound Saltmarsh Sparrows typically did not complete large proportions of their migratory route in single flights and instead exhibited short-hop behavior to complete their migration (Fig. 4 D–F; O’Reilly and Wingfield 1995, Wright et al. 2018). In addition, most fall stopovers were short (43% < 1 day; 57% < 2 days), suggesting sparrows often use stopovers only to rest and feed during daylight hours, followed by continuation of migration. Stopover duration decreased on average as fall migration progressed and sparrows neared their wintering grounds. These findings are consistent with those for the closely related Ipswich Savannah Sparrow (Passerculus sandwichensis princeps; Crysler et al. 2016, Bliss 2020), in which individuals followed an energy-minimizing strategy (Alerstam and Hedenström 1998) and exhibited frequent, short migratory flights and stopovers. Given our definition of stopovers, it is possible that birds also made rare long stops, but it was impossible to distinguish these from birds reaching wintering sites or experiencing transmitter failure in our study.

All detected southbound stopovers longer than four days occurred in southern New England (coastal Massachusetts to coastal Connecticut), and geographic patterns of stopover (Fig. 5) reinforce the conclusion that the Connecticut and Rhode Island coast is an important area of stopover for southbound individuals, particularly among Maine breeders. The central New Jersey coast may be of similar importance to Massachusetts breeders. A relatively sparse Motus receiver network south of Virginia and waning battery life of transmitters left us unable to identify fall stopover locations south of the mid-Atlantic. Motus network sparsity likely hindered our ability to detect spring stopovers as well. However, the short length of the spring stopovers that we could identify may also explain the low number that were detected. We potentially saw few spring stopovers because Saltmarsh Sparrows rarely make them, consistent with longstanding ideas about faster migration in spring. The ability to identify key stopover locations is important to the whole life-cycle conservation of a species (McKinnon et al. 2017, Van Loon et al. 2017, Smetzer and King 2018), given that the success of a migratory journey, as well as an individual’s survival and breeding output following migration, are highly dependent on the availability and quality of stopover sites (Moore et al. 1995, Calvert et al. 2009, McGowan et al. 2011).

A concern with all telemetry studies is that tags alter an animal’s behavior or survival. Given problems with antenna entanglement in prior studies of Saltmarsh Sparrow and other grassland passerines (Hill and Elphick 2011, van Vliet and Stutchbury 2018), we tried to minimize this possibility. Several birds suffered antenna entanglements when using tags with a plastic-coated metal alloy antenna rather than an uncoated metal alloy antenna. As a result, we worked with the manufacturer to modify antenna design to avoid the plastic coating and saw no further adverse effects. Our experiences caution that care must be taken when using tags on birds that move regularly among dense grasses and similar vegetation. Our recapture work demonstrated that tags often do not fall off, even when using materials designed or expected to do so, and that minor feather damage and abrasions can occur. In short, seemingly insignificant differences in harness deployment (materials and fit) and antenna design can produce very different outcomes for individuals. Telemetry has proven to be essential to understanding the details of migration and has greatly enhanced conservation of many species, but our results reinforce the need for continuous monitoring of tag effects, especially with species of high conservation concern (Hill and Elphick 2011).

Despite the insights gained from this work, there is much to be learned, particularly in a more refined understanding of migratory connectivity between breeding and wintering areas. This is specifically true regarding movement through, and stopover and overwintering within, the southern half of the Saltmarsh Sparrow’s range. Rectifying this situation will require increased coverage and density in the Motus receiving station network. Receiving network growth along the south Atlantic Coast and improved transmitter battery longevity since the conclusion of this study should greatly improve resolution for subsequent work. Further, targeted network development in eastern North Carolina and coastal Florida will greatly improve inferences of migratory routes for Saltmarsh Sparrows. Creative and timely partnerships will be necessary to maintain and expand this existing network for the benefit of all.

Although not designed to examine post-breeding exploratory behavior (Mettke-Hofmann and Gwinner 2003, Mitchell et al. 2010), our study provides additional hints that it happens on the breeding grounds (cf. Greenlaw et al. 2020), yet we have little understanding of how common it is and whether it affects subsequent settlement at breeding sites. Given the degree of breeding site fidelity for Saltmarsh Sparrows, information about how the species locates breeding grounds may become more important as the amount of suitable breeding habitat declines and the locations of breeding habitat changes. Future telemetry studies will be crucial for understanding the consequences of habitat loss and climate change by revealing further insight into the full life-cycle dynamics of Saltmarsh Sparrows.

Our study of Saltmarsh Sparrows highlights the value of large-scale telemetry studies across a species’ migratory pathway and the value of the Motus network for implementing such studies. We have demonstrated that such work is possible, even for small songbirds, although challenges such as the constraints imposed by the available receiver network and transmitter battery life remain even as the network expands and technology improves. Despite these challenges, collecting high-resolution movement information throughout the annual cycle will be crucial, especially for endangered and declining species. For instance, although our study has improved our knowledge of stopover biology in Saltmarsh Sparrows, we need better data to more fully identify sites of particular importance during migration. More refined understanding of flight paths and stopover sites would also allow better integration of bird conservation into human infrastructure development, e.g., the expansion of wind farms or other industrial projects that may influence survival during migration, and prioritization of sites for large-scale restoration. Given the challenges birds face during migration, providing high quality habitats for stopovers and maintaining the integrity of migratory pathways are likely to remain key to conservation planning. For the Saltmarsh Sparrow and other small migratory birds, automated telemetry offers a valuable and accessible approach to identifying those pathways and stopovers.

LITERATURE CITED

Alerstam, T., and A. Hedenstróm. 1998. The development of bird migration theory. Journal of Avian Biology 29:343–369. https://doi.org/10.2307/3677155

Barron, D. G., J. D. Brawn, and P. J. Weatherhead. 2010. Meta-analysis of transmitter effects on avian behaviour and ecology. Methods in Ecology and Evolution 1:180–187. https://doi.org/10.1111/j.2041-210X.2010.00013.x

Bates, D., M. Mächler, B. Bolker, and S. Walker. 2015. Fitting linear mixed-effects models using lme4. Journal of Statistical Software 67:1–48. https://doi.org/10.18637/jss.v067.i01

Benvenuti, B., J. Walsh, K. M. O’Brien, M. J. Ducey, and A. I. Kovach. 2018. Annual variation in the offspring sex ratio of Saltmarsh Sparrows supports Fisher’s hypothesis. Auk 135:342–358. https://doi.org/10.1642/AUK-17-148.1

BirdLife International. 2023. Species factsheet: Ammospiza caudacuta. Birdlife International, Cambridge, UK. http://datazone.birdlife.org/species/factsheet/22721129

Bliss, S. 2020. Using automated radio-telemetry to assess spring migratory ecology and survival of Ipswich Sparrows (Passerculus sandwichensis princeps). Thesis. Dalhousie University, Halifax, Nova Scotia, Canada. https://DalSpace.library.dal.ca//handle/10222/78527

Borowske, A. C. 2015. Effects of life history strategies on annual events and processes in the lives of tidal Marsh Sparrows. Dissertation. University of Connecticut, Mansfield, Connecticut, USA. https://digitalcommons.lib.uconn.edu/dissertations/877/

Borowske, A., C. R. Field, K. J. Ruskin, and C. S. Elphick. 2018. Consequences of breeding system for body condition and survival throughout the annual cycle of tidal Marsh Sparrows. Journal of Avian Biology 49:jav-01529. https://doi.org/10.1111/jav.01529

Borowske, A., C. Gjerdrum, and C. Elphick. 2017. Timing of migration and prebasic molt in tidal Marsh Sparrows with different breeding strategies: comparisons among sexes and species. Auk 134:51–64. https://doi.org/10.1642/AUK-16-116.1

Calvert, A. M., P. D. Taylor, and S. Walde. 2009. Cross-scale environmental influences on migratory stopover behaviour. Global Change Biology 15:744–759. https://doi.org/10.1111/j.1365-2486.2008.01733.x

Correll, M. D., W. A. Wiest, B. J. Olsen, W. G. Shriver, C. S. Elphick, and T. P. Hodgman. 2016. Habitat specialization explains avian persistence in tidal marshes. Ecosphere 7:e01506. https://doi.org/10.1002/ecs2.1506

Crysler, Z. J., R. A. Ronconi, and P. D. Taylor. 2016. Differential fall migratory routes of adult and juvenile Ipswich Sparrows (Passerculus sandwichensis princeps). Movement Ecology 4:3. https://doi.org/10.1186/s40462-016-0067-8

Duijns, S., A. M. Anderson, Y. Aubry, A. Dey, S. A. Flemming, C. M. Francis, C. Friis, C. Gratto-Trevor, D. J. Hamilton, R. Holberton, S. Koch, A. E. McKellar, D. Mizrahi, C. A. Morrissey, S. G. Neima, D. Newstead, L. Niles, E. Nol, J. Paquet, J. Rausch, L. Tudor, Y. Turcotte, and P. A. Smith. 2019. Long-distance migratory shorebirds travel faster towards their breeding grounds, but fly faster post-breeding. Scientific Reports 9:9420. https://doi.org/10.1038/s41598-019-45862-0

Duijns, S., L. J. Niles, A. Dey, Y. Aubry, C. Friis, S. Koch, A. M. Anderson, and P. A. Smith. 2017. Body condition explains migratory performance of a long-distance migrant. Proceedings of the Royal Society B: Biological Sciences 284:20171374. https://doi.org/10.1098/rspb.2017.1374

Field, C. R., T. S. Bayard, C. Gjerdrum, J. M. Hill, S. Meiman, and C. S. Elphick. 2017. High-resolution tide projections reveal extinction threshold in response to sea-level rise. Global Change Biology 23:2058–2070. https://doi.org/10.1111/gcb.13519

Field, C. R., K. J. Ruskin, B. Benvenuti, A. C. Borowske, J. B. Cohen, L. Garey, T. P. Hodgman, R. A. Longenecker, E. King, A. R. Kocek, A. I. Kovach, K. M. O’Brien, B. J. Olsen, N. Pau, S. G. Roberts, E. Shelly, W. G. Shriver, J. Walsh, and C. S. Elphick. 2018. Quantifying the importance of geographic replication and representativeness when estimating demographic rates, using a coastal species as a case study. Ecography 41:971–981. https://doi.org/10.1111/ecog.02424

FitzGerald, D. M., and Z. Hughes. 2019. Marsh processes and their response to climate change and sea-level rise. Annual Review of Earth and Planetary Sciences 47:481–517. https://doi.org/10.1146/annurev-earth-082517-010255

Fridolfsson, A.-K., and H. Ellegren. 1999. A simple and universal method for molecular sexing of non-ratite birds. Journal of Avian Biology 30:116–121. https://doi.org/10.2307/3677252

Ganju, N. K., Z. Defne, M. L. Kirwan, S. Fagherazzi, A. D’Alpaos, and L. Carniello. 2017. Spatially integrative metrics reveal hidden vulnerability of microtidal salt marshes. Nature Communications 8:14156. https://doi.org/10.1038/ncomms14156

Gedan, K. B., B. R. Silliman, and M. D. Bertness. 2009. Centuries of human-driven change in salt marsh ecosystems. Annual Review of Marine Science 1:117–141. https://doi.org/10.1146/annurev.marine.010908.163930

Greenlaw, J. S., C. S. Elphick, W. Post, and J. D. Rising 2020. Saltmarsh Sparrow (Ammodramus caudacutus), version 1.0. In Birds of The World P. G. Rodewald, Editor). Cornell Lab of Ornithology, Ithaca, NY, USA. https://doi.org/10.2173/bow.sstspa.01

Hartig, F., and L. Lohse. 2022. DHARMa: residual diagnostics for hierarchical multi-level / mixed regression models. R Foundation for Statistical Computing, Vienna, Austria. https://doi.org/10.32614/CRAN.package.DHARMa

Hill, J. M., and C. S. Elphick. 2011. Are grassland passerines especially susceptible to negative transmitter impacts? Wildlife Society Bulletin 35:362a–367. https://doi.org/10.1002/wsb.84

Hill, J. M., J. Walsh, A. I. Kovach, and C. S. Elphick. 2013. Male-skewed sex ratio in Saltmarsh Sparrow nestlings. Condor 115:411–420. https://doi.org/10.1525/cond.2013.120116

Kalnay, E., M. Kanamitsu, R. Kistler, W. Collins, D. Deaven, L. Gandin, M. Iredell, S. Saha, G. White, J. Woollen, Y. Zhu, M. Chelliah, W. Ebisuzaki, W. Higgins, J. Janowiak, K. C. Mo, C. Ropelewski, J. Wang, A. Leetmaa, R. Reynolds, R. Jenne, and D. Joseph. 1996. The NCEP/NCAR 40-year reanalysis project. Bulletin of the American Meteorological Society 77:437–472.

Kemp, M. U., E. Emiel van Loon, J. Shamoun-Baranes, and W. Bouten. 2012. RNCEP: global weather and climate data at your fingertips. Methods in Ecology and Evolution 3:65–70. https://doi.org/10.1111/j.2041-210X.2011.00138.x

Klingbeil, B. T., J. B. Cohen, M. D. Correll, C. R. Field, T. P. Hodgman, A. I. Kovach, E. E. Lentz, B. J. Olsen, W. G. Shriver, W. A. Wiest, and C. S. Elphick. 2021. High uncertainty over the future of tidal marsh birds under current sea-level rise projections. Biodiversity and Conservation 30:431–443. https://doi.org/10.1007/s10531-020-02098-z

Lenth, R. V., B. Banfai, B. Bolker, P. Buerkner, I. Giné-Vázquez, M. Herve, M. Jung, J. Love, F. Miguez, J. Piaskowski, H. Riebl, and H. Singmann. 2023. emmeans: estimated marginal means, aka least-squares means. R Foundation for Computational Statistics, Vienna, Austria. https://doi.org/10.32614/CRAN.package.emmeans

Lin, H.-Y., R. Schuster, S. Wilson, S. J. Cooke, A. D. Rodewald, and J. R. Bennett. 2020. Integrating season-specific needs of migratory and resident birds in conservation planning. Biological Conservation 252:108826. https://doi.org/10.1016/j.biocon.2020.108826

Marra, P. P., E. B. Cohen, S. R. Loss, J. E. Rutter, and C. M. Tonra. 2015a. A call for full annual cycle research in animal ecology. Biology Letters 11:20150552. https://doi.org/10.1098/rsbl.2015.0552

McGowan, C. P., J. E. Hines, J. D. Nichols, J. E. Lyons, D. R. Smith, K. S. Kalasz, L. J. Niles, A. D. Dey, N. A. Clark, P. W. Atkinson, C. D. T. Minton, and W. Kendall. 2011. Demographic consequences of migratory stopover: linking red knot survival to horseshoe crab spawning abundance. Ecosphere 2:art69. https://doi.org/10.1890/ES11-00106.1

McKinnon, E. A., C. Artuso, and O. P. Love. 2017. The mystery of the missing warbler. Ecology 98:1970–1972. https://doi.org/10.1002/ecy.1844

McLean, D. J., and M. A. Skowron Volponi. 2018. trajr: an R package for characterisation of animal trajectories. Ethology 124:440–448. https://doi.org/10.1111/eth.12739

Mettke-Hofmann, C., and E. Gwinner. 2003. Long-term memory for a life on the move. Proceedings of the National Academy of Sciences 100:5863–5866. https://doi.org/10.1073/pnas.1037505100

Mills, A. M., B. G. Thurber, S. A. Mackenzie, and P. D. Taylor. 2011. Passerines use nocturnal flights for landscape-scale movements during migration stopover. Condor 113:597–607. https://doi.org/10.1525/cond.2011.100186

Mitchell, G. W., P. D. Taylor, and I. G. Warkentin. 2010. Assessing the function of broad-scale movements made by juvenile songbirds prior to migration. Condor 112:644–654. https://doi.org/10.1525/cond.2010.090136

Moore, F., S. Gauthreaux, Jr., P. Kerlinger, and T. R. Simons. 1995. Habitat requirements during migration: important link in the conservation of Neotropical landbird migrants. Pages 121–144 in T. Martin and D. Finch, editors. Ecology and management of neotropical migratory birds: a synthesis and review of critical issues. Oxford University Press, New York, New York, USA. https://doi.org/10.1093/oso/9780195084405.003.0005

Morbey, Y. E., C. G. Guglielmo, P. D. Taylor, I. Maggini, J. Deakin, S. A. Mackenzie, J. M. Brown, and L. Zhao. 2018. Evaluation of sex differences in the stopover behavior and postdeparture movements of wood-warblers. Behavioral Ecology 29:117–127. https://doi.org/10.1093/beheco/arx123

Newton, I. 2006. Can conditions experienced during migration limit the population levels of birds? Journal of Ornithology 147:146–166. https://doi.org/10.1007/s10336-006-0058-4

O’Reilly, K. M., and J. C. Wingfield. 1995. Spring and autumn migration in Arctic shorebirds: same distance, different strategies. American Zoologist 35:222–233. https://doi.org/10.1093/icb/35.3.222

Paxton, E. H., S. L. Durst, M. K. Sogge, T. J. Koronkiewicz, and K. L. Paxton. 2017. Survivorship across the annual cycle of a migratory passerine, the Willow Flycatcher. Journal of Avian Biology 48:1126–1131. https://doi.org/10.1111/jav.01371

Pyle, P. 1997. Identification guide to North American birds: a compendium of information on identifying, ageing, and sexing “near-passerines” and passerines in the hand. Slate Creek Press, Bolinas, California, USA.

R Core Team. 2022. R: a language and environment for statistical computing. R Foundation for Computational Statistics, Vienna, Austria. https://www.R-project.org/

Rappole, J. H., and A. R. Tipton. 1991. New harness design for attachment of radio transmitters to small passerines. Journal of Field Ornithology 62:335–337.

Roberts, S. G., R. A. Longenecker, M. A. Etterson, C. S. Elphick, B. J. Olsen, and W. G. Shriver. 2019. Preventing local extinctions of tidal marsh endemic Seaside Sparrows and Saltmarsh Sparrows in eastern North America. Condor 121:duy024. https://doi.org/10.1093/condor/duy024

Schuster, R., S. Wilson, A. D. Rodewald, P. Arcese, D. Fink, T. Auer, and J. R. Bennett. 2019. Optimizing the conservation of migratory species over their full annual cycle. Nature Communications 10:1754. https://doi.org/10.1038/s41467-019-09723-8

Sillett, T. S., and R. T. Holmes. 2002. Variation in survivorship of a migratory songbird throughout its annual cycle. Journal of Animal Ecology 71:296–308. https://doi.org/10.1046/j.1365-2656.2002.00599.x

Smetzer, J. R., and D. I. King. 2018. Prolonged stopover and consequences of migratory strategy on local-scale movements within a regional songbird staging area. Auk 135:547–560. https://doi.org/10.1642/AUK-18-4.1

Smith, A., B. Benvenuti, C. S. Elphick, N. Ernst, A. Kovach, K. O’Brien, and N. Pau. 2025. Data and analysis: migratory movements of Saltmarsh Sparrows revealed by automated telemetry. figshare. Dataset. https://doi.org/10.6084/m9.figshare.23582559.v1

Streby, H. M., T. L. McAllister, S. M. Peterson, G. R. Kramer, J. A. Lehman, and D. E. Andersen. 2015. Minimizing marker mass and handling time when attaching radio-transmitters and geolocators to small songbirds. Condor 117:249–255. https://doi.org/10.1650/CONDOR-14-182.1

Taylor, P. D., T. L. Crewe, S. A. Mackenzie, D. Lepage, Y. Aubry, Z. Crysler, G. Finney, C. M. Francis, C. G. Guglielmo, D. J. Hamilton, R. L. Holberton, P. H. Loring, G. W. Mitchell, D. R. Norris, J. Paquet, R. A. Ronconi, J. R. Smetzer, P. A. Smith, L. J. Welch, and B. K. Woodworth. 2017. The Motus Wildlife Tracking System: a collaborative research network to enhance the understanding of wildlife movement. Avian Conservation and Ecology 12(1):8. https://doi.org/10.5751/ACE-00953-120108

Taylor, P. D., S. A. Mackenzie, B. G. Thurber, A. M. Calvert, A. M. Mills, L. P. McGuire, and C. G. Guglielmo. 2011. Landscape movements of migratory birds and bats reveal an expanded scale of stopover. PLOS ONE 6:e27054. https://doi.org/10.1371/journal.pone.0027054

Van Loon, A., J. D. Ray, A. Savage, J. Mejeur, L. Moscar, M. Pearson, M. Pearman, G. T. Hvenegaard, N. Mickle, K. Applegate, and K. C. Fraser. 2017. Migratory stopover timing is predicted by breeding latitude, not habitat quality, in a long-distance migratory songbird. Journal of Ornithology 158:745–752. https://doi.org/10.1007/s10336-017-1435-x

van Vliet, H. E. J., and B. J. M. Stutchbury. 2018. Radiotagged fledgling Savannah Sparrows Passerculus sandwichensis at risk of entanglement in vegetation. Ibis 160:919–922. https://doi.org/10.1111/ibi.12615

Ward, M. P., T. J. Benson, J. Deppe, T. J. Zenzal, R. H. Diehl, A. Celis-Murillo, R. Bolus, and F. R. Moore. 2018. Estimating apparent survival of songbirds crossing the Gulf of Mexico during autumn migration. Proceedings of the Royal Society B: Biological Sciences 285:20181747. https://doi.org/10.1098/rspb.2018.1747

Watson, E. B., C. Wigand, E. W. Davey, H. M. Andrews, J. Bishop, and K. B. Raposa. 2017. Wetland loss patterns and inundation-productivity relationships prognosticate widespread salt marsh loss for southern New England. Estuaries and Coasts 40:662–681. https://doi.org/10.1007/s12237-016-0069-1

Wright, J. R., L. L. Powell, and C. M. Tonra. 2018. Automated telemetry reveals staging behavior in a declining migratory passerine. Auk 135:461–476. https://doi.org/10.1642/AUK-17-219.1