Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Tracking Natal Dispersal in a Coastal Population of a Migratory Songbird Using Feather Stable Isotope (δ2H, δ34S) Tracers

Tracking Natal Dispersal in a Coastal Population of a Migratory Songbird Using Feather Stable Isotope (δ2H, δ34S) Tracers

  • Samuel Haché, 
  • Keith A. Hobson, 
  • Erin M. Bayne, 
  • Steven L. Van Wilgenburg, 
  • Marc-André Villard
PLOS
x

Abstract

Adult birds tend to show high fidelity to their breeding territory or disperse over relatively short distances. Gene flow among avian populations is thus expected to occur primarily through natal dispersal. Although natal dispersal is a critical demographic process reflecting the area over which population dynamics take place, low recapture rates of birds breeding for the first time have limited our ability to reliably estimate dispersal rates and distances. Stable isotope approaches can elucidate origins of unmarked birds and so we generated year- and age-specific δ2H and δ34S feather isoscapes (ca. 180 000 km2) of coastal-breeding Ovenbirds (Seiurus aurocapilla) and used bivariate probability density functions to assign the likely natal areas of 35 males recruited as first-year breeders into a population located in northwestern New Brunswick, Canada. Most individuals (80–94% depending on the magnitude of an age correction factor used; i.e. 28–33 out of 35) were classified as residents (i.e. fledged within our study area) and estimated minimum dispersal distances of immigrants were between 40 and 240 km. Even when considering maximum dispersal distances, the likely origin of most first-year breeders was<200 km from our study area. Our method identified recruitment into our population from large geographic areas with relatively few samples whereas previous mark-recapture based methods have required orders of magnitude more individuals to describe dispersal at such geographic scales. Natal dispersal movements revealed here suggest the spatial scale over which many population processes are taking place and we suggest that conservation plans aiming to maintain populations of Ovenbirds and ecologically-similar species should consider management units within 100 or at most 200 km of target breeding populations.

Introduction

Dispersal is a key component of animal demography because it controls rates of immigration and emigration among populations [1][3]. Most songbirds are characterized by low natal philopatry (but see [4]) and the rare data available suggest that first-year breeders may disperse over tens [5][7] or even hundreds of kilometres [8][11] prior to establishing their first breeding territory. Conversely, experienced breeders show high site fidelity [12][14]. Hence, gene flow and connectivity among avian populations should mainly reflect natal dispersal [15], which is commonly estimated as the straight-line distance moved by an individual from its natal area to its first breeding site [12]. Empirical estimates of natal dispersal distances are required to determine the spatial scale over which breeding populations of widely-distributed species interact and, thus, to define relevant conservation units. Data from mark recapture studies and indirect estimates based upon spatial correlations in abundance both suggest that movements likely occur over relatively short distances and that long-distance movements are rare (e.g. [5][7]). Unfortunately, such studies are generally inefficient for measuring long-distance dispersal ([16], but see [7]). Recently, intrinsic markers such as stable isotopes, trace elements, genetic markers, species assemblages of parasites, and diseases have been suggested as potential means by which dispersal movements can be tracked without marking individuals [16]. In particular, stable-hydrogen isotope ratios in feathers (δ2Hf) have proven to be useful for detecting bird movements at continental scales [17].

Using δ2Hf, Hobson et al. [9] obtained the first estimates of dispersal in Ovenbird (Seiurus aurocapilla) and American Redstart (Setophaga ruticilla) in western Canada based on the number of first-year birds that were “isotopic outliers” relative to expected baseline isotopic values for six study sites. Studds et al. [10] used this marker to generate a species-specific isoscape for Bicknell’s Thrush (Catharus bicknelli) and assignment tests to determine the area of likely origin of first-year breeders (i.e. second year individuals; hereafter SY). That study provided minimum natal dispersal distances and evidence for spatio-temporal discrepancy in demographic connectivity. Although single isotope approaches provide important insights into natal dispersal, it is generally recognized that multiple isotopes or combinations of techniques improves the spatial resolution for assigning individuals to their natal area (e.g. [18][21]). Stable isotope measurements of sulfur (δ34S) may be particularly useful in distinguishing individuals growing feathers in coastal regions vs. inland since marine-derived sulfates are generally more enriched in 34S than terrestrial sources and volatilized sulfates from ocean spray can be deposited inland over considerable distances directly or through precipitation [22][24].

We investigated the use of two stable isotope tracers to estimate natal dispersal distances in Ovenbird. Given that our study area was located within 160 km of the Gulf of St Lawrence, we jointly used δ2Hf and δ34Sf to obtain a finer-scale assignment than using δ2Hf alone. We anticipated significant longitudinal structure in δ34Sf based on distance from the coast and prevailing weather. We first created species-, age-, and year-specific feather isoscapes [25] for our study region and then assigned first-year breeders to their natal origin using these isoscapes and likelihood-based assignment tests based on bivariate probability density functions. Our specific objectives were to estimate the proportion of resident (hatched in the study area) vs. immigrant SY breeders recruited into our local breeding population and to determine the minimum distance moved by immigrants. Based on the typically low natal philopatry reported in songbirds, we predicted that most SY breeders would be immigrants. We had no a priori expectations regarding the spatial extent of natal dispersal movements by Ovenbirds as they have never been quantified. Most information available on distance of natal dispersal has been provided by mark-recapture studies. While mark-recapture studies are biased against detection of longer-distance dispersers, the isotope approach is biased against short-distance dispersers [9] since these will often be classified as residents. Nevertheless, an important objective of this study was to quantify the extent and frequency of long-distance natal dispersal to identify the area over which population dynamics of Ovenbirds take place.

Methods

Study Area

Since 2006, an individually-marked subpopulation of male Ovenbird has been monitored in northwestern New Brunswick, Canada (47°23′ N67°40′W; Black Brook district; Figure 1) to quantify the effects of an experimental harvest treatment on population dynamics [14], [26]. Black Brook is a 2000 km2 managed forest district owned by J.D. Irving Ltd. The landscape is a mosaic of spruce plantations and deciduous, mixedwood, and coniferous stands at the interface of the northern hardwoods and Acadian forest regions [27].

thumbnail
Figure 1. δ2Hf (A) and δ34Sf (B) Ovenbird isoscapes based on samples from ASY males collected in 2010 and calibrated 2011 and 2012 data.

Isoscapes were based on ordinary Gaussian and ordinary spherical semivariogram models, respectively. Blue points are the 26 sampling locations and in black is the Black Brook district.

https://doi.org/10.1371/journal.pone.0094437.g001

Feather Samples, Benchmark Values, and Isoscapes

In 2010, most territorial males in the 250-ha area monitored (10×25 ha study plots; see [14] for details) that were not marked in previous breeding seasons (i.e. recruits) were captured (89%; 55/62) using mist-nets and playbacks of conspecific songs and marked with three color bands and a numbered aluminium band. Territorial individuals were also captured and marked in two additional 25-ha plots as part of a translocation experiment (see [28] for details; n = 12 and 16). Study plots were located in mature (81–160 years) and overmature (>161 years) deciduous stands that had not been altered in the past 30 years (n = 5) or managed through partial harvesting (n = 7; [14], [28]). Two 3rd rectrices were plucked from each captured individual for aging and isotope analyses. Males were aged as SY or after second-year (ASY) individuals following the criteria of Donovan and Stanley [29] adapted by Bayne and Hobson [30]. In 2010–2012, we also collected feather samples from 48 returning ASY males banded in the previous year (hereafter “returning ASY”; n = 23, 14, and 11 for 2010, 2011, and 2012, respectively). During the same period, an additional 225 territorial males were captured within ca. 5 km of 22 additional locations outside the study area (10.2±1.4, mean ± SD, individuals per location; 16, 3, and 3 sampling locations in 2010, 2011, and 2012, respectively). Birds were captured in young (30–50 years), immature (51–80 years), and mature (81–160 years) deciduous and deciduous-dominated mixedwood forests. The total sampling area covered ca. 180 000 km2 (340×530 km; mean distance between closest sampling locations was 64.6 km, range = 28.9–134.0 km; Figure 1). Only feather samples from ASY males captured at these sampling locations were used to generate the feather isoscapes because they were believed to provide local values of δ34Sf and δ2Hf based on the assumption that ASY individuals have high breeding site fidelity [14], [31], [32] and moulting occurs in the vicinity of the breeding territory [33] (but see [34]). In total, we measured δ2Hf for all SY (2010) and returning ASY males from Black Brook (2010–2012), and for all ASY males from the 22 additional sampling locations (2010–2012). Values of δ34Sf were also measured for all SY males (2010) from Black Brook, but only for returning ASY males from 2010 and 2012, owing to logistical constraints. For the same reason, only three individuals per additional sampling location were analyzed for δ34Sf (n = 66).

Stable-hydrogen isotope ratios from feathers collected in the same breeding area can vary among years [35], [36]. As our goal was to assign SY males captured in 2010 (feathers grown in 2009) to their natal area the previous year, we controlled for a potential yearly variation in isotope values by creating a year-specific δ2Hf isoscape reflecting expected values for feathers grown in 2009. Thus, we only used δ2Hf values from returning ASY males (Black Brook) captured in 2010 (feathers grown in 2009) as the benchmark values for the year-specific δ2Hf isoscape. We could not visit our entire study region in a single year to create a year-isoscape and so applied a correction factor for δ2Hf values from samples collected at locations surrounding Black Brook in 2011 (feathers grown in 2010; n = 3) and 2012 (feathers grown in 2011; n = 3) based on the yearly variation we saw in ASY males at Black Brook. Specifically, we compared mean δ2Hf values from returning ASY males (Black Brook) captured in 2010 vs. 2011 and 2010 vs. 2012. We then used these differences to adjust δ2Hf from ASY males captured at sampling locations outside Black Brook in 2011 and 2012. Only δ2Hf values from six sampling locations had to be calibrated as the ASY males from the remaining 16 locations were all captured in 2010. This approach assumes that year-to-year variation in δ2H from growing-season precipitation is similar across the total sampling area. Values of δ34Sf were not calibrated to account for annual variation as there was no a priori reason to expect annual variation in this isotope. However, we still tested for a year effect using a t-test.

Mean δ34Sf and δ2Hf from individuals captured at the 12 study sites in Black Brook were pooled in 4 locations (3 sites per location) covering the extent of the study area (Figure 1). Sites within each of these locations were 3–6 km apart and each location was separated by at least 7 km (see also [14], [28]). Study sites were pooled to provide a similar number of ASY males for each of the 26 sampling locations used to generate feather isoscapes. Mean values from all sampling locations were assigned a centroid and used to generate the isoscapes. Spatial autocorrelation among mean values from each sampling location was modelled for both isotopes using semivariance analyses and kriging interpolations. We investigated several semivariogram models (ordinary spherical, circular, exponential, Gaussian, or linear), and selected the model that minimized the root-mean-square error. Isoscapes were created in ArcGIS 10 using the Geostatistical Analyst extension (ESRI, Redlands, CA).

Research was approved by the Université de Moncton Animal Care and Use Committee (Permit Numbers: 10-06, 11-04, and 12-02), Canadian Wildlife Service (Permit Numbers: SC2710 and SC2751), and U.S. Fish and Wildlife Service (Permit Number: MB11009A-0).

Stable Isotope Analyses

Surface oils were removed from feathers using a 2∶1 chloroform:methanol solution. Samples were prepared according to [37] and analyses were conducted at the Stable Isotope Hydrology and Ecology Laboratory of Environment Canada (δ34Sf in 2010–2012 and δ2Hf in 2010 and 2012) and the Colorado Plateau Stable Isotope Laboratory (δ2Hf in 2011). High-temperature (1350°C) flash pyrolysis generating an H2 pulse for each sample (350±20 μg) was used to obtain δ2Hf measurements by continuous-flow isotope-ratio mass spectrometry (CF-IRMS). To account for exchangeable hydrogen in keratins, comparative equilibration was done using in-house keratin working standards (BWB = −108‰, CFS = −147.7‰, and CHS = −187‰; [38]). The two laboratories used the same protocol and standards for δ2Hf measurements. Values of δ34Sf were also measured by CF-IRMS (3500±100 μg of tissue per sample). All results are expressed as nonexchangeable deuterium (δ2Hf) and sulphur (δ34Sf) isotopic ratios in units of per mil (‰) and normalized to Vienna Standard Mean Ocean Water – Standard Light Antarctic Precipitation (VSMOW-SLAP) and Canyon Diablo Triolite, respectively. Based on within-run analyses of keratin standards, we assumed measurement error to be ca. ±2‰ for both δ2Hf and δ34Sf.

Assignment Test

Previous studies have reported differences in δ2Hf between juvenile (i.e. nestlings, hatch-year, or SY birds) and ASY individuals growing feathers at the same site ranging between 6–32.6‰ [10], [36], [39]. Our goal was to assign SY males assumed to have grown their rectrices mostly during the first few weeks postfledging (i.e. rectrices only start emerging from feather sheaths a few days prior to fledging; S.H., MAV, and EMB, pers. obs.). Thus, to link SY results to an ASY feather isoscape, a correction factor to account for the documented effect of age was required. Based on a long-term study, Studds et al. [10] reported that δ2Hf was on average ∼6‰ lower in first-year breeders/fledglings than in adult Bicknell’s Thrushes at the northern edge of their study area. This portion of the Bicknell’s Thrush range (i.e. southern Quebec, New Brunswick, and Gaspe Peninsula) corresponded to the extent of our isoscapes and this comparison based on values from first-year breeders/fledglings is more appropriate than values from nestling Ovenbirds obtained by Haché et al. [36]. Indeed, it is unclear whether comparisons between δ2Hf from growing (nestlings) versus inert (first-year breeders/fledglings) tissue provide appropriate age correction factors. Also, the Bicknell’s thrush is expected to be of the same guild as the Ovenbird and hence should make a reasonable surrogate species. Hobson et al. [40] found no evidence for an age effect (hatching- and second-year vs. ASY) on δ2Hf. Although we believe that the most parsimonious approach was to use a range of age correction factors from no correction to adding 6‰ to all SY males [10], [40], we generated assignment tests based on five different age correction factors (+6‰, +3‰, no correction, –3‰, and –6‰) to examine how sensitive the results of assignment tests were to a range of correction factors. The correction factor was applied to δ2Hf of each SY male prior to assigning their origin to the ASY feather isoscapes.

We generated spatially-explicit assignment tests using bivariate normal probability density functions (BNPDF; [41]) to determine the likely origin of SY males. The likelihood that a given raster cell (j; 2 km×2 km cell size) within the feather isoscapes was a potential origin was estimated based on the following equation:Where f(xiyi|μj) is the likelihood that an individual (i) with given δ2Hf (xi) and δ34Sf (yi) values originated from a raster cell j, while μ is the mean, σ is the standard deviation, and ω is the correlation between δ2Hf (x) and δ34Sf (y) predicted for each raster cell (j) within the feather isoscapes. Predicted mean isotopic composition for each raster cell (j) was generated from the calibrated isoscapes (δ2Hf and δ34Sf). The parameter μj is therefore a vector of means for each raster cell (j) within the feather isoscapes presented as:

Each BNPDF was parameterized using both δ2Hf and δ34Sf isoscapes and the following variance-covariance matrix:where σ2δ2Hf and σ2δ34Sf are the expected variance for each isotope ratio and σδ2Hf, σδ34Sf is the covariance between the pair of isotope ratios. The variance-covariance matrix was validated to define the likelihood a cell within the extent of our isoscape corresponded to a potential origin. This was achieved by using a subset of 66 samples from ASY males (i.e. the three ASY males from each of the 22 sampling location outside Black Brook with values for both δ2Hf and δ34Sf; 36% of all ASY in the dataset). We assumed a spatial stationarity of variance-covariance across our study area [41]. Standard deviations representing variation within sampling location were 10.5‰ and 3.8‰ for δ2Hf and δ34Sf, respectively, and the correlation coefficient from the covariance was 0.29. For all SY males (2010), we generated a binary raster layer to classify each cell as a likely (1) or unlikely (0) origin based on a 2∶1 odds ratio (i.e. “1” was assigned to cells that were consistent with the upper 67% of the cumulative probabilities and “0” was assigned to all other cells). Odds ratios between 2∶1 and 4∶1 provided the best depiction of the geographic distribution of assigned versus true origins for samples from multiple individuals ([21] on advantages of using 2∶1 odds ratio). We estimated the likely origin of all SY males by summing the binary surfaces. We conducted similar analyses to assign the 23 returning ASY males captured in 2010 (Black Brook) to the isoscapes to determine the spatial resolution and classification accuracy (i.e. percent of returning ASY with an assigned area that overlapped their known origin) provided by this approach. To compare predictions from results of models using different odds ratio, we also generated assignment tests for all SY males using a 4∶1 odds ratio and two age correction factors (±6‰). Lastly, we used the predicted δ2Hf and δ34Sf from our isoscape for the four sampling locations within Black Brook (Figure 1) and generated fictional individuals for which their origin was considered “unknown”. We assigned their likely area of origin to the bivariate isoscape and calculated the minimum detectable dispersal distance as the perpendicular distance from a sampling location to the closest area assigned as unlikely origin. This value reflected the minimum distance from which an individual could have been assigned as immigrant. This is not a universal minimum detectable dispersal distance and only reflects the minimum detectable distance given our isoscapes and the location of our study site within those isoscapes. Assignment tests were generated using scripts employing the raster [42] package in R v 2.15.2. For each SY male assigned as an immigrant (i.e. assigned origin did not overlap the sampling location where it was captured; as opposed to a resident), we calculated the minimum linear distance between breeding location and natal area [10] using ArcGIS v 10.

Results

Of the 10.2 male Ovenbirds captured on average (±1.4, SD; n = 225) at each of the 22 sampling locations outside Black Brook, 6.5 were ASY (±2.1; range of 3–10 individuals per location; n = 142; 63%). When added to the 23 ASY males captured in Black Brook in 2009 and subsequently recaptured in 2010 (5.8±2.2 ASY per sampling location), samples from a total of 165 and 89 ASY males were used to produce the δ2Hf and δ34Sf isoscapes, respectively (Figures 12). In Black Brook, we also recaptured 14 returning ASY males in 2011 and 11 in 2012. There was year-to-year variation in δ2Hf from returning ASY males (2010 = −67.0±5.3‰, 2011 = −88.1±4.1‰, and 2012 = −78.9±5.1‰; see also [36]). Thus, to correct for this year effect in δ2Hf, we added 21.1‰ to the δ2Hf of 20 ASY males captured at sampling locations (n = 3) outside Black Brook in 2011 and 11.9‰ to those of 25 ASY males captured at the three other sampling locations in 2012. Mean δ34Sf values differed significantly between years (2010 = 6.79±0.58‰ and 2012 = 6.30±0.29‰; t32 = −2.6, p = 0.014), but this difference was negligible when considering the 2‰ measurement error. In 2010, we captured 35 SY males breeding in Black Brook (δ2Hf = −68.0±10.2‰ and δ34Sf = 7.06±0.86‰; Figure 2).

thumbnail
Figure 2. Histograms of δ2Hf (A) and δ34Sf (B) from returning ASY (n = 23) and SY (n = 35) males breeding in the Black Brook district, New Brunswick, in 2010.

https://doi.org/10.1371/journal.pone.0094437.g002

The δ2Hf isoscape was generated using ordinary kriging based upon a Gaussian semivariogram model (root-square-mean error = 4.59). The model-estimated parameters were a major range of 120.7 km, nugget of 0.81, and sill of 31.8. Data for δ34Sf were kriged to create an isoscape based upon a spherical semivariogram model (root-square-mean error = 1.03). The semivariogram model parameters for the δ34Sf isoscape included a major range of 120.7 km, nugget of 0, and sill of 11.8. The δ2Hf isoscape showed generally more enriched values in southeastern portion of the isoscape (∼50‰) with gradual depletion along a southeastern-northwestern gradient, whereas the δ34Sf isoscape corresponded to the anticipated coastal effect with more enriched values mostly along the coast (10–20‰) creating an east-west gradient (Figure 1).

The assigned origin of 22 of the 23 ASY males of known origin overlapped their sampling location (classification accuracy of 95.6%; Figure 3A). Similarly, 94.3% of the SY males (33/35) were classified as residents under four of the five correction factors (no correction, −3‰, +3‰, and +6‰; 2∶1 odds ratio; Figures 3B, S1B, and S2). The two individuals classified as an immigrants had minimum dispersal distances ranging from 40–54 km and 221–237 km (Table S1). The assignments based on a −6‰ correction factor had 80% of the SY males (28/35; Figure S1A) classified as residents. Using this correction factor, the seven immigrants had minimum dispersal distances ranging from 23–210 km, but only one individual had a minimum dispersal distance>61 km (Table S1). The minimum detectable dispersal distance of our four sampling locations in Black Brook was 40±14 km (Figure S3). Using a 4∶1 odd ratios provided more conservative estimates with only one and two individuals classified as immigrants under the +6‰ and −6‰ age correction factors, respectively (Table S1; Figures S4).

thumbnail
Figure 3. Geographic distribution of the assigned origin of 23 returning ASY (A) and 35 SY (B) male Ovenbirds known to have bred in the Black Brook district (in black), New Brunswick.

Assignments of each male to the δ2Hf and δ34S isoscapes were based on using bivariate normal probability density functions and no δ2Hf age correction factor (2∶1 odds ratio). Maps represent the sum of all binary raster layers of each individual assignment. Blue points are the 26 sampling locations and in black is the Black Brook district.

https://doi.org/10.1371/journal.pone.0094437.g003

The maximum distance of dispersal (i.e. maximum distance of likely origin) of SY males could not be estimated because in all cases, it reached the edge of the isoscape, suggesting that they could have originated from an area beyond our defined feather isoscape. However, the population-level assigned origin (i.e. sum of the binary surfaces for each SY male) indicated that likely origins of SY males only reached edges of the isoscapes in small areas in the southeast and northeast, suggesting that the majority of SY males tended to have a maximum natal dispersal distance<200 km from Black Brook (Figures 3B, S1B, and S2). Under the −6‰ age correction factors, a larger proportion of SY males might have originated beyond the northwestern and southwestern edges of the isoscape (Figure S1A). This effect was even greater when we used a 4∶1 odds ratio with a −6‰ age correction factor (Figure S4A). Interestingly, this edge effect was not as important for assignment tests based on a 4∶1 odds ratio and +6‰ age correction factor (Figure S4B). The assigned area of returning ASY males (Figure 3A) tended to be similar to that of SY males based on four of the six age correction factors (no correction, –3‰, +3‰, and +6‰), suggesting that the maximum natal dispersal distance is a conservative estimate as it corresponds to the maximum spatial resolution provided by this approach.

Discussion

By combining age- and year-specific δ34Sf and δ2Hf Ovenbird isoscapes and bivariate assignment tests along with the most parsimonious age correction factors (i.e. no correction, +3‰, and +6‰), we showed that most (94%) SY males recruited in our coastal-breeding population likely originated locally (i.e. hatched in our 2000 km2 study area; residents). Given that stable isotope analyses have been shown to be better-suited to identify long-distance dispersal than mark-recapture studies, this finding was unexpected based on the low natal philopatry generally reported in songbirds [12]. However, our range of dispersal distances (40–200 km) did overlap with those provided by mark recapture studies and indirect estimates based upon time-lagged, pairwise correlations in abundance (10–100 km; e.g. [5][7]). When summing the assigned areas of all SY males, our estimates of maximum distances of natal dispersal still suggested that movements generally occurred within 200 km (Figure 3B and S2). This information on natal dispersal movements offers insight into the area over which many population processes are taking place and, ultimately, how populations are structured. Hence, our approach can be used to identify the spatial scale at which conservation plans for the Ovenbird and ecologically-similar songbirds should be implemented (see also [6], [15]).

Estimates of minimum natal dispersal distances suggest that most SY males can be considered as residents and, in the rare instances where individuals were classified as immigrants (n = 2 based on the most parsimonious age correction factors), immigrants would have originated within 240 km of our study area. The Black Brook district covers 2000 km2 (ca. 80 km× 25 km), suggesting that a high proportion of males recruited into the district’s Ovenbird population likely hatched “locally”. Based on δ2Hf outliers, Hobson et al. [9] considered 9.5% (±8.1%; n = 6) of SY male Ovenbirds breeding in western Canada as long-distance dispersers which is consistent with the percentage of immigrants (5.7%) in our study. However, they could not separate short-distance dispersers from residents. A band-recovery study from Europe [8] and a North American study using one-year time-lagged correlations in Breeding Bird Survey data [6] both estimated mean natal dispersal distances<100 km for several landbird species. Alternatively, Studds et al. [10] showed that minimal natal dispersal movements in Bicknell’s Thrush reached 700 km and they considered that 59% of first-year breeders were immigrants. Unlike the Ovenbird, Bicknell’s Thrush occurs at low abundance, population sizes are declining [43], and they have a small, highly-patchy breeding range [44]. However, the study sites surveyed by Studds et al. [10] might be smaller than ours (Black Brook, ca. 2000 km2) and spatial scale must be considered when making such comparisons. Nonetheless, Paradis et al. [8] showed that interspecific variation in natal dispersal distance was negatively correlated with population size and the spatial extent of the breeding range, consistent with the observed differences between Ovenbird and Bicknell’s Thrush populations. In contrast, Prothonotary Warbler (Protonotaria citrea) and Tree Swallow (Tachycineta bicolor) also have different population sizes and breeding ranges [45], [46], yet show similar in natal dispersal distances. Median natal dispersal distance in Tree Swallow is 2.3 km (1.3% of individuals dispersed 50–210 km; [5]), as opposed to 1.4 km (maximum of 17 km) in Prothonotary Warbler [7]. However, those results are from nestbox studies and inferences from such studies to populations breeding in natural nest sites may be misleading [47], [48]. Those results were also from mark recapture studies and although those authors obtained dispersal distance distributions, their approaches cannot be used to estimate emigration rates unless we account for sampling effort. Using intrinsic markers such as stable isotopes results in every capture being a recapture [16] and provides unbiased estimates of immigration rates given the spatial resolution provided by this approach (40 km in this study).

The assigned origin of all SY males reached the edge of our isoscape, limiting our ability to generate accurate estimates of maximum natal dispersal distances. This was especially the case when using a −6‰ age correction factor and assignment tests based on a 4∶1 odds ratio. However, when using the most parsimonious age correction factors, the likely origin only reached smaller areas at the edge of the isoscape (ca. 190 km from Black Brook; Figure 3B and S2), suggesting that we likely captured most of the area that would have been assigned had we examined isoscapes of larger spatial extent. This is based on the assumption that similar isotopic composition is unlikely to occur over large areas beyond the extent of our isoscape as a result of the latitudinal and longitudinal gradients reported for δ34Sf and δ2Hf, respectively. Similar coastal conditions at the same latitude also do not occur elsewhere in the Ovenbird’s breeding range [49]. Nevertheless, we recognize that dual isotopic mapping over the entire range of the species would be required to consider all possible isotopic origins of birds arriving in our study area. Also, the assigned likely natal origin of the SY males monitored in this study (Figure 3B) covered an area similar to the assigned origin of the returning ASY males from Black Brook (known origin; Figure 3A) suggesting that our estimates of maximum distance of natal dispersal are constrained by the spatial resolution (i.e. uncertainty) provided by this approach and might even be shorter. Given the spatial resolution provided, it is also possible that some individuals classified as residents might have immigrated over short distances and were within the minimum range of detectability using the isotope approach. This support results from previous studies suggesting that stable isotope analyses would have limited ability to quantify short-distance dispersal (e.g. [9]). Lastly, the classification accuracy of ASY of known origin was 95.6% and 94% (33/35) of SY males were considered resident. Hence, it is unclear whether the two immigrants are indeed immigrants or this pattern was a result of the classification accuracy from our bivariate approach.

We minimized sources of error in our δ2Hf isoscape by controlling for interspecific, age, and year effects which are all known to influence δ2Hf [36], [39], [40]. Elevation is another factor influencing δ2Hf [50], [51]. Most of the area covered by our isoscapes varied from 0–600 m in elevation [52], but we are confident that our sampling design integrated the majority of this variation within our δ2Hf isoscape. Other potential factors explaining spatial variation in the δ2Hf isoscape are fine-scale variation in temperature, moisture level, and amount of precipitation [53] which might be especially important near coastlines (see also[34]).

Unlike δ2Hf, we had no a priori reason to expect inter-annual variation in δ34Sf. Similarly, δ34Sf is unlikely to be influenced by the age of an individual because little diet-tissue isotopic discrimination occurs in these isotopes as a result of metabolic activity [54], [55]. Overall, the spatial variation observed in the δ34Sf isoscape corresponds to the anticipated coastal effect (deposition of volatilized marine sulfates) previously observed [22][24], [56], [57]. However, in some instances, the overall gradient in δ34Sf across our isoscape as a function of distance to the ocean seems to have been altered by local processes that we do not fully understand (reviewed by [22], [54], [58]).

We calibrated δ2Hf of SY males based on age-related variation in δ2Hf provided by Hobson et al. [40] and a 10-year study on Bicknell’s Thrush [10]. The assumption that this range in the correction factor (0–6‰) is consistent among songbird species and breeding seasons needs to be validated as it could have important implications when assigning origin of SY males to ASY feather isoscapes. Only subtle differences were detected between these results and the assigned origin of SY males based on −3‰ and −6‰ correction factors (Table S1). It has often been suggested that 2∶1 odds ratio would offer the most reasonable likelihood of correct versus incorrect assignments of individuals to geographic origin (reviewed by [21]), but this remains a relatively subjective threshold and we believe that is important to compare these results from those more conservative odd ratios (e.g. 4∶1 and 9∶1).

Low recapture rates often prevent the use of extrinsic markers to evaluate natal dispersal. The main advantage of the isotope approach is that it allows tracking natal dispersal of a completely random sample of a population of first-year breeders, whereas mark-recapture studies require an initial marked population that can be biased. Multivariate assignment tests have been used to assign the geographic origin of numerous taxa (e.g. birds [59], [60], invertebrates [61], [62], and mammals [63]), but most studies documented migratory patterns and fewer explored age-specific dispersal movements. Our results are consistent with those from other studies suggesting that although philopatry is low in songbirds, natal dispersal typically occurs over relatively short distances. Findings from this study could be used to test predictions about effects of habitat fragmentation on immigration rates and dispersal movements (e.g. [64]). However, future studies should consider: 1) using larger isoscapes, i.e. whole breeding range, to provide estimates of both minimum and maximum dispersal distances; 2) investigating natal dispersal in females; and 3) including additional chemical markers (e.g. δ13C, δ15N, 87Sr/86Sr, and trace elements; [65][67]) to determine the maximum spatial resolution available to investigate biologically relevant spatial scales for management planning to conserve bird populations. Nonetheless, we recommend that to conserve Ovenbird populations and those from ecologically-similar songbird species, the relevant demographic unit, i.e. the area of origin of potential recruits, corresponds to a radius of 100 or at most 200 kilometers (see also [68], [69]).

Supporting Information

Figure S1.

Geographic distribution of the assigned origin of 35 SY male Ovenbirds known to have bred in the Black Brook district, New Brunswick. Assignments of SY males to the δ2Hf and δ34S isoscapes were based on using bivariate normal probability density functions and −6‰ (A) and −3‰ (B) δ2Hf age correction factors (2∶1 odds ratio). Maps represent the sum of all binary raster layers of each individual assignment. Blue points are the 26 sampling locations and in black is the Black Brook district.

https://doi.org/10.1371/journal.pone.0094437.s001

(TIF)

Figure S2.

Geographic distribution of the assigned origin of 35 SY male Ovenbirds known to have bred in the Black Brook district, New Brunswick. Assignments of SY males to the δ2Hf and δ34S isoscapes were based on using bivariate normal probability density functions and +3‰ (A) and +6‰ (B) δ2Hf age correction factors (2∶1 odds ratio). Maps represent the sum of all binary raster layers of each individual assignment. Blue points are the 26 sampling locations and in black is the Black Brook district.

https://doi.org/10.1371/journal.pone.0094437.s002

(TIF)

Figure S3.

Minimum detectable dispersal distance (in grey) for individuals captured at one of the four sampling locations (in blue) within the Black Brook district (in black). Predicted δ2Hf and δ34Sf from our isoscapes were used for the four sampling locations within the Black Brook district to generate fictional individuals for which their origin was considered “unknown”. The likely area of origin was assigned to the bivariate isoscape and we calculated the minimum detectable dispersal distance as the perpendicular distance from a sampling location to the closest area assigned as unlikely origin (0).

https://doi.org/10.1371/journal.pone.0094437.s003

(TIF)

Figure S4.

Geographic distribution of the assigned origin of 35 SY male Ovenbirds known to have bred in the Black Brook district, New Brunswick. Assignments of SY males to the δ2Hf and δ34S isoscapes were based on using bivariate normal probability density functions and –6‰ (A) and +6‰ (B) δ2Hf correction factors (4∶1 odds ratio). Maps represent the sum of all binary raster layers of each individual assignment. Blue points are the 26 sampling locations and in black is the Black Brook district.

https://doi.org/10.1371/journal.pone.0094437.s004

(TIF)

Table S1.

SY male Ovenbirds (n = 35) classified as resident (1) or immigrant (0) depending on whether or not the assigned area of origin overlapped the study area (Black Brook). Assignment tests were based on 5 different age correction factors. In parentheses is the minimum distance of dispersal for each immigrant. Unless indicated otherwise (4∶1), assignment tests were based on a 2∶1 odds ratio. ID represents individual birds.

https://doi.org/10.1371/journal.pone.0094437.s005

(DOCX)

Acknowledgments

We are grateful to P. Bertrand for his help with data collection. Stable isotope analyses were assisted by C. Gryba and G. Koehler. We also thank the staff of J.D. Irving Ltd. for their advice and logistical support.

Author Contributions

Conceived and designed the experiments: SH EMB MAV KAH. Analyzed the data: SH SLVW. Wrote the paper: SH KAH MAV EMB SLVW. Stable isotope analyses: KAH.

References

  1. 1. Colbert J, Danchin E, Dhondt A, Nichols JD (2001) Dispersal. New York: Oxford University Press. 452 p.
  2. 2. Bullock JM, Kenward RE, Hails R (2002) Dispersal Ecology. Oxford: Blackwell Science. 458 p.
  3. 3. Hanski I, Gaggiotti OE (2004) Ecology, genetics and evolution of metapopulations. California: Elsevier. 696 p.
  4. 4. Dale S, Lunde A, Steifetten Ø (2005) Longer breeding dispersal than natal dispersal in the ortolan bunting. Behavioral Ecology 16: 20–24.
  5. 5. Winkler DW, Wrege PH, Allen PE, Kast TL, Senesac P, et al. (2005) The natal dispersal of tree swallows in a continuous mainland environment. Journal of Animal Ecology 74: 1080–1090.
  6. 6. Tittler R, Villard MA, Fahrig L (2009) How far do songbirds disperse? Ecography 32: 1051–1061.
  7. 7. McKim-Louder MI, Hoover JP, Benson TJ, Schelsky WM (2013) Juvenile survival in a neotropical migratory songbird is lower than expected. PLoS ONE 8: e56059.
  8. 8. Paradis E, Baillie SR, Sutherland WJ, Gregory RD (1998) Patterns of natal and breeding dispersal in birds. Journal of Animal Ecology 67: 518–536.
  9. 9. Hobson KA, Wassenaar LI, Bayne E (2004) Using isotopic variance to detect long-distance dispersal and philopatry in birds: an example with ovenbirds and American redstarts. Condor 106: 732–743.
  10. 10. Studds CE, Mcfarland KP, Aubry Y, Rimmer CC, Hobson KA, et al. (2012) Stable-hydrogen isotope measures of natal dispersal reflect observed population declines in a threatened migratory songbird. Diversity and Distributions 18: 919–930.
  11. 11. Sutherland GD, Harestad AS, Price K, Lertzman KP (2000) Scaling of natal dispersal distances in terrestrial birds and mammals. Conservation Ecology 4: 16.
  12. 12. Greenwood PJ, Harvey PH (1982) The natal and breeding dispersal of birds. Annual Reviews of Ecology and Systematics 13: 1–21.
  13. 13. Holmes RT, Sherry TW (1992) Site fidelity of migratory warblers in temperate breeding and neotropical wintering areas: implications for population dynamics, habitat selection, and conservation. In Hagan JM, Johnston DW, editors. Ecology and Conservation of Neotropical Migrant Landbirds. Washington: Smithsonian Press. 563–575.
  14. 14. Haché S, Villard MA (2010) Age-specific response of a migratory bird to an experimental alteration of its habitat. Journal of Animal Ecology 79: 897–905.
  15. 15. Tittler R, Fahrig L, Villard MA (2006) Evidence of large-scale source-sink dynamics and long-distance dispersal among wood thrush populations. Ecology 87: 3029–3036.
  16. 16. Hobson KA, Norris DR (2008) Animal migration: a context for using new techniques and approaches. In: Hobson KA, Wassenaar LI, editors. Tracking animal migration with stable isotopes. New York: Academic Press. 1–19.
  17. 17. Hobson KA, Wassenaar LI (2008) Tracking animal migration with stable isotope. New York: Academic Press. 144 p.
  18. 18. Donovan T, Buzas J, Jones P, Gibbs HL (2006) Tracking dispersal in birds: assessing the potential of element markers. Auk 123: 500–511.
  19. 19. Szép T, Hobson KA, Vallner J, Piper SE, Kovács B, et al. (2009) Comparison of trace element and stable isotope approaches to the study of migratory connectivity: an example using two hirundine species breeding in Europe and wintering in Africa. Journal of Ornithology 150: 621–636.
  20. 20. Coulton DW, Clark RG, Hebert CE (2010) Determining natal origins of birds using stable isotopes (δ34S, δD, δ15N, δ13C): model validation and spatial resolution for mid-continent Mallards. Waterbirds 33: 10–21.
  21. 21. Procházka P, Van Wilgenburg SL, Neto JM, Yosef R, Hobson KA (2013) Using stable hydrogen isotopes (δ2H) and ring recoveries to trace natal origins in a Eurasian passerine with a migratory divide. Journal of Avian Biology 44: 001–010.
  22. 22. Jamieson RE, Wadleigh MA (2000) Tracing sources of precipitation sulfate in eastern Canada using stable isotopes and trace metals. Journal of Geophysical Research 105: 549–556.
  23. 23. Kelly JF, Ruegg KC, Smith TB (2005) Combining isotopic and genetic markers to identify breeding origins of migrant birds. Ecological Applications 15: 1487–1494.
  24. 24. Zazzo A, Monahan FJ, Moloney AP, Green S, Schmidt O (2011) Sulphur isotopes in animal hair track distance to sea. Rapid Communication in Mass Spectrometry 25: 2371–2378.
  25. 25. West JB, Bowen GJ, Dawson TE, Tu KP (2010) Isoscapes: understanding movement pattern and process on earth through isotope mapping. New York: Springer-Verlag. 501 p.
  26. 26. Haché S, Villard MA, Bayne EM (2013) Experimental evidence for an ideal free distribution in a breeding population of a territorial songbird. Ecology 94: 861–869.
  27. 27. Mosseler A, Lynds JA, Major JE (2003) Old-growth forests of the Acadian forest region. Environmental Reviews 11: S47–S77.
  28. 28. Villard MA, Haché S (2012) Conifer plantations consistently act as barriers to movement in a deciduous forest songbird: a translocation experiment. Biological Conservation 155: 33–37.
  29. 29. Donovan TM, Stanley CM (1995) A new method of determining Ovenbird age on the basis of rectrix shape. Journal of Filed Ornithology 66: 247–252.
  30. 30. Bayne EM, Hobson KA (2001) Effects of habitat fragmentation on pairing success of ovenbirds: Importance of male age and floater behavior. Auk 118: 380–388.
  31. 31. Bayne EM, Hobson KA (2002) Apparent survival of male ovenbirds in fragmented and forested landscapes. Ecology 83: 1307–1316.
  32. 32. Porneluzi PA (2003) Prior breeding success affects return rates of territorial male ovenbirds. Condor 105: 73–79.
  33. 33. Pyle P (1997) Identification guide to North American birds. Part 1. California: Slate Creek Press. 732 p.
  34. 34. Hallworth MT, Studds CE, Sillett TS, Marra PP (2013) Do archival light-level geolocators and stable hydrogen isotopes provide comparable estimates on breeding-ground origin? Auk 130: 273–282.
  35. 35. Langin KM, Reudink MW, Marra PP, Norris DR, Kyser TK, et al. (2007) Hydrogen isotopic variation in migratory bird tissues of known origin: implications for geographic assignment. Oecologia 152: 449–457.
  36. 36. Haché S, Hobson KA, Villard MA, Bayne EM (2012) Assigning birds to geographic origin using feather hydrogen isotope ratios (δ2H): importance of year, age, and habitat. Canadian Journal of Zoology 90: 722–728.
  37. 37. Wassenaar LI (2008) An introduction to light stable isotopes for use in terrestrial animal migration studies. In Hobson KA, Wassenaar LI, editors. Tracking animal migration with stable isotopes. New York: Academic Press. 21–44.
  38. 38. Wassenaar LI, Hobson KA (2003) Comparative equilibration and online technique for determination of non-exchangeable hydrogen of keratins for use in animal migration studies. Isotopes in Environmental and Health Studies 39: 211–217.
  39. 39. Gow EA, Stutchbury BJM, Done T, Kyser TK (2012) An examination of stable isotope (δD) variation in adult and juvenile feathers from a migratory songbird. Canadian Journal of Zoology 90: 585–594.
  40. 40. Hobson KA, Van Wilgenburg SL, Wassenaar LI, Larson K (2012) Linking hydrogen (δH) isotopes in feathers and precipitation: sources of variance and consequences for assignment to isoscapes. PLoS ONE 7: e35137.
  41. 41. Royle A, Rubenstein DR (2004) The role of species abundance in determining breeding origins of migratory birds with stable isotopes. Ecological Applications 14: 1780–1788.
  42. 42. Hijmans RJ, Van Etten J (2012) raster: Geographic analysis and modeling with raster data. R package version 2.0–12. Available: http://CRAN.R-project.org/package=raster. Accessed 12 September 2013.
  43. 43. Rich TD, Beardmore CJ, Berlanga H, Blancher PJ, Bradstreet MSW, et al.. (2004) Partners in Flight North American Landbird Conservation Plan. Cornell Lab of Ornithology. Available: http://www.partnersinflight.org/cont_plan/default.htm. Accessed 12 September 2013.
  44. 44. Rimmer CC, Mcfarland KP, Ellison WG, Goetz JE (2001) Bicknell’s Thrush (Catharus bicknelli). In Poole A, editor. The Birds of North America Online. Cornell Lab of Ornithology. Available: http://bna.birds.cornell.edu/bna/species/592. Accessed 12 September 2013. https://doi.org/ http://bna.birds.cornell.edu/bna/species/592
  45. 45. Petit LJ (1999) Prothonotary Warbler (Protonotaria citrea). In Poole A, editor. The Birds of North America Online. Cornell Lab of Ornithology. Available: http://bna.birds.cornell.edu/bna/species/408. Accessed 12 September 2013. https://doi.org/ http://bna.birds.cornell.edu/bna/species/408
  46. 46. Winkler DW, Hallinger KK, Ardia DR, Robertson RJ, Stutchbury BJ, et al.. (2011) Tree Swallow (Tachycineta bicolor). In Poole A, editor. The Birds of North America Online. Cornell Lab of Ornithology. Available: http://bna.birds.cornell.edu/bna/species/011. Accessed 12 September 2013. https://doi.org/ http://bna.birds.cornell.edu/bna/species/011
  47. 47. Lambrechts MM, Adriaensen F, Ardia DR, Artemyev AV, Atienzar F, et al. (2010) The design of artificial nestboxes for the study of secondary hole-nesting birds: a review of methodological inconsistencies and potential biases. Acta Ornithologica 45: 1–26.
  48. 48. Wesolowski T (2011) Reports from nestbox studies: a review of inadequacies. Acta Ornithologica 46: 13–17.
  49. 49. Porneluzi P, Van Horn MA, Donovan TM (2011) Ovenbird (Seiurus aurocapilla). In Poole A, editor. The Birds of North America Online. Cornell Lab of Ornithology. Available: http://bna.birds.cornell.edu/bna/species/088. Accessed 12 September 2013.
  50. 50. Poage MA, Chamberlain P (2001) Empirical relationships between elevation and the stable isotope composition of precipitation and surface water: consideration for studies of paleoelevation change. American Journal of Science 301: 1–15.
  51. 51. Hobson KA, Wassenaar LI, Mila B, Lovette I, Dingle C, et al. (2003) Stable isotopes as indicators of altitudinal distributions and movements in an Ecuadorean hummingbird community. Oecologia 136: 302–308.
  52. 52. NASA Land Processes Distributed Active Archive Center (2012) GDEx: ASTER Global DEM – SRTM 90m. US Geological Survey, USA. Available: http://gdex.cr.usgs.gov/gdex/. Accessed 2013 Sep 12.
  53. 53. Dansgaard W (1964) Stable isotopes in precipitation. Tellus 16: 436–468.
  54. 54. Krouse HR, Grinenko LN, Grinenko VA, Newman L, Forrest J, et al.. (1991) Case studies and potential applications. In Krouse HR, Grinenko VA, editors. Stable isotopes: natural and anthropogenic sulphur in the environment. Scope 43. New York: John Wiley & Sons. 308–425.
  55. 55. Hobson KA (2008) Applying isotopic methods to tracking animal movements. In: Hobson KA, Wassenaar LI, editors. Tracking animal migration with stable isotopes. New York: Academic Press. 45–78.
  56. 56. Lott CA, Meehan TD, Heath JA (2003) Estimating the latitudinal origins of migratory birds using hydrogen and sulfur isotopes in feathers: influence of marine prey base. Oecologia 134: 505–510.
  57. 57. Valenzuela LO, Chesson LA, O’Grady SP, Cerling TE, Ehleringer JR (2011) Spatial distributions of carbon, nitrogen, and sulfur isotope ratios in human hair across the central United States. Rapid Communication in Mass Spectrometry 25: 861–868.
  58. 58. Chambers LA, Trudinger PA (1979) Microbiological fractionation of stable sulfur isotopes: a review and critique. Geomicrobiology Journal 1: 249–293.
  59. 59. Wunder MB, Kester CL, Knopf FL, Rye RO (2005) A test of geographic assignment using isotope tracers in feathers of known origin. Oecologia 144: 607–617.
  60. 60. Evans Ogden LJ, Hobson KA, Lank DB, Bittman S (2006) Stable isotope analysis reveals that agricultural habitat provides an important dietary component for nonbreeding Dunlin. Avian Conservation and Ecology 1: 3.
  61. 61. Miller NG, Wassenaar LI, Hobson KA, Norris DR (2010) Monarch butterflies cross the Appalachians from the west to recolonize the east coast of North American. Biology Letters 7: 43–46.
  62. 62. Flockhart DT, Wassenaar LI, Martin TG, Hobson KA, Wunder MB, et al. (2013) Tracking multi-generational colonization of the breeding grounds by monarch butterflies in eastern North America. Proceedings of the Royal Society B 280: 20131087.
  63. 63. Henaux V, Powell LA, Hobson KA, Nielsen CK, LaRue MA (2011) Tracking large carnivore dispersal using isotopic clues in claws: an application to cougars across the Great Plains. Methods in Ecology and Evolution 2: 489–499.
  64. 64. Garrard GE, McCarthy MA, Vesk PA, Radford JQ, Bennett AF (2012) A predictive model of avian natal dispersal distance provides prior information for investigating response to landscape change. Journal of Animal Ecology 81: 14–23.
  65. 65. Kosciuch KL, Rivers JW, Sandercock BK (2008) Stable isotopes identify the natal origins of a generalist brood parasite, the brown-headed cowbird Molothrus ater. Journal of Avian Biology 39: 364–367.
  66. 66. Poesel A, Nelson DA, Gibbs HL, Olesik JW (2008) Use of trace element analysis of feathers as a tool to track fine-scale dispersal in birds. Behavioral Ecology and Sociobiology 63: 153–158.
  67. 67. Sellick MJ, Kyser TK, Wunder MB, Chipley D, Norris DR (2009) Geographic variation of strontium and hydrogen isotopes in avian tissue: implications for tracking migration and dispersal. PLoS ONE 4: e4735.
  68. 68. Ward MP (2005) The role of immigration in the decline of an isolated migratory bird population. Conservation Biology 19: 1528–1536.
  69. 69. Schaub M, Jokober H, Stauber W (2013) Strong contribution of immigration to local population regulation: evidence from a migratory passerine. Ecology 94: 1828–1838.