Research Paper

INTRODUCTION

A successful implementation of recovery measures for species at risk depends on the precise definition of critical habitat. This becomes particularly crucial for species with widespread distributions, given the significant variability in biophysical attributes across geographical scales and regions within the species’ occupancy area (Crosby et al. 2019). Understanding where these variations occur is an important first step, enabling the identification of management units, the development of regional models, and the capacity to propose appropriate recovery planning scenarios (Leston et al. 2024).

The Canada Warbler (Cardellina canadensis; hereafter CAWA) is a good example of a bird exhibiting regional variability in habitat selection, which has limited the implementation of its recovery strategy since its designation as threatened in Canada in 2008 (COSEWIC 2008). In 2003, the long-term decline of the species began to slow down, and the population has increased steadily since 2012. In 2020, to reflect that improvement, CAWA was downgraded by the COSEWIC to “Special Concern,” a status that focuses on managing threats and habitat effectively to prevent future decline (COSEWIC 2020); however, the species remains listed as Threatened under the Species at Risk Act (2024). Regardless of its status, identifying the biophysical attributes used by the species during the breeding period is still pertinent to support appropriate management and prevent future declines.

The breeding range of CAWA covers most of the southern part of the boreal forest of Canada and extends through the northeastern United States in the Appalachian regions (Reitsma et al. 2020). This species nests on or near the ground and forages in vegetation with dense foliage within 3–5 m of the ground (Sabo and Holmes 1983, Goodnow and Reitsma 2011, Becker et al. 2012). Indeed, abundance, occupancy, and breeding success have all been positively linked with a dense and complex understory vegetation (Chace et al. 2009, Goodnow and Reitsma 2011, Krikun et al. 2018). At coarse scale, CAWA inhabits a wide range of deciduous and mixed forests, but has shown a range-wide association with deciduous and mixed forest stands with tall trees and closed canopy (see Reitsma et al. 2020). Fine scale variables appear to be more crucial in assessing the overall quality of the breeding habitat, such as soil wetness, the number of perch trees, edge effects, forest gaps, and topography, depending on the region (see Reitsma et al. 2020).

Numerous studies have been conducted over the last decade in western Canada (Hunt et al. 2017, Krikun et al. 2018), in the Appalachian forests of the United States (Hallworth et al. 2008a, 2008b, Chace et al. 2009, Becker et al. 2012), and in the Maritimes region (Westwood et al. 2019, 2020). However, there has been limited research in the central part of its breeding range, specifically in Ontario and Québec (Cupiche-Herrera et al. 2024). Across its eastern range, CAWA is frequently associated in forested wetlands, a pattern that has also been documented in more southern regions (see Reitsma et al. 2020). In Québec, however, field reports from citizen science programs suggest instead that the highest densities of CAWA are found in well-drained forests of the boreal and hemiboreal zones, where forested wetlands are rare (Toussaint 2019, eBird 2022). To better understand these regional differences, we examined CAWA breeding habitat selection across multiple forest landscapes that provide a representative overview of the habitat structures typical of the eastern Canadian provinces.

METHODS

Study areas

Our study was conducted between 2016 and 2018 in three distinct forested regions of Québec, each reflecting a different ecological context: the mixed boreal upland forest, the agroforestry landscape of the St. Lawrence lowlands, and the riverine deciduous forest of southern Québec (Fig. 1).

Mixed boreal forest

We surveyed the mixed boreal upland forest (hereafter mixed boreal forest) study area during the summers of 2016 and 2017 (see Appendix 1a). The mostly forested study area is located within the balsam fir—white birch bioclimatic domain where forest stands are dominated by balsam fir (Abies balsamea), trembling aspen (Populus tremuloides), and white birch (Betula papyrifera). Understory vegetation is characterized by sapling stands of mountain maple (Acer spicatum), white birch, beaked hazelnut (Corylys cornuta), and balsam fir (Saucier et al. 2009).

Agroforest

We surveyed the agroforestry landscape of the St. Lawrence lowlands study area (hereafter agroforest), located within a 50 km radius around the Abenakis First Nation of Odanak (see Appendix 1b), in spring 2018. This study area comprises 54% agricultural lands, 25% forests, 13% water bodies, and 8% urban areas. The region is also part of two bioclimatic domains, namely the sugar maple (Acer saccharum)—basswood (Tilia americana) domain and the sugar maple—bitternut hickory (Carya cordiformis) domain, which are characterized by deciduous forest stands mainly composed of sugar maple and red maple (Acer rubrum) with a significant presence of basswood, bitternut hickory, ashes (Fraxinus spp.), and oaks (Quercus spp.; Saucier et al. 2009). Understory vegetation is characterized by saplings and seedlings of the same species as well as shrub vegetation of various species.

Riverine forest

We conducted fieldwork in the riverine deciduous forest of the southern Québec study area (hereafter riverine forest) in spring 2018 (see Appendix 1c). This study area lies within the sugar maple—yellow birch (Betula alleghaniensis) and the sugar maple—basswood bioclimatic domains (Saucier et al. 2009). It features 65 lakes, numerous marshes and swamps, and is traversed by the Kinonge River valley.

Site selection and bird survey

Mixed boreal forest

We conducted CAWA surveys along all road networks (~14 km) by at least two walking observers. Bird surveys took place between 11 and 26 June 2016, and between 15 and 28 June 2017. While walking, observers broadcasted CAWA intraspecific calls (lasting between 5 and 10 minutes) at every 100 m road segment of CAWA potential habitat (hereafter referred to as stations), and the occurrence of the species was recorded when ≥ 1 individual was heard or observed. These surveys were designed to capture individuals in order to study their local movements and migration departure using Motus technology (see Bégin-Marchand et al. 2022).

Agroforest

In order to maximize potential detection of CAWA, we preselected stations based on prior mentions in public databases (e.g., Toussaint 2019, eBird 2022) and from 1:20,000 digital forest maps targeting stands with maples or pines (Pinus spp.) as dominant/co-dominant species, as well as wet areas (Reitsma et al. 2020). We added additional stations afterward based on their accessibility. We retained a total of 61 stations, with centers spaced at least 2 km apart, for this study.

We carried out a single point count located 50 m away from a forest edge at every station. We made a 4-hour recording (starting 30 minutes before dawn) with unlimited radius using an automatic recording unit (Song Meter SM4, Wildlife Acoustics; hereafter ARU) at each point count. We relocated ARUs to a new point count as soon as we obtained a quality recording (i.e., on days without rain or wind). We conducted each point count between 1 and 23 June 2018. All recordings were processed entirely by a single observer using spectrograms computed with Raven Pro software (Bioacoustics Research Program 2014) to detect the occurrence of CAWA. The observer could replay recordings as many times as needed and refer to online sound libraries.

Riverine forest

We used a 10-minute point count with a 50 m limited radius to identify CAWA occurrence at stations located 50 m from the shore of a wetland or lake (> 0.5 ha). A total of 93 stations, spaced at least 250 m apart, were visited between sunrise and 9:00 am when conditions were favorable (no rain and only light wind) by an experienced observer on two occasions (with at least 7 days between visits) between 4 and 29 June 2018. CAWA intraspecific calls were broadcasted for 150 seconds if the species was not detected during the first 5 minutes to maximize detection.

Vegetation survey

Mixed boreal forest

Within the road network surveyed, we retained all stations with CAWA occurrence and randomly selected an equivalent number of stations without CAWA to survey the vegetation (a total of 16 stations in 2016 and 22 stations in 2017, including both presence and absence). At each station, we established a linear transect across the road, and we sampled two vegetation plots on both sides of the transect: one with a 1.5 m radius (7 m²) located 3 m perpendicular to the road within the sapling stand that delimits the edge of the road (hereafter referred to as the forest edge), and one with an 8 m radius (200 m²) located 10 m within the forest stands (Fig. 2a). In the sapling stand (7 m² plot), we counted all saplings (DBH < 9 cm) by species and assigned them to a diameter class. In the forest stand (200 m² plot), we recorded the diameter at breast height of all tree species (DBH ≥ 9 cm) using a caliper (± 0.05 cm), and we also surveyed a 7 m² plot centered within the 200 m² plot with the same information collected as in the sapling stands. Finally, we estimated from the road the width and height of the foliage at the forest edge (± 0.25 m).

Agroforest

We centered a circular plot with an 11.28 m radius (400 m²) at the point count for all 61 stations surveyed to identify, count, and measure the diameter of all trees at breast height (DBH ≥ 9 cm) using a caliper (± 0.05 cm). We assigned saplings (DBH < 9 cm) to a diameter class within five circular micro-plots (4 m²) located in the center of the 400 m² plot and at the four cardinal points, 11.28 m away from the center (Fig. 2b). We measured the mean canopy height using a clinometer, and we estimated canopy closure with a sighting tube (Noon 1981). At 2 m intervals along the N-S and E-W diameters of the circular plot (20 points per plot), the sighting tube was pointed overhead, and we recorded the presence or absence of deciduous and coniferous cover.

Riverine forest

We sampled riverine vegetation at all 26 stations where CAWA occurred and at 26 randomly selected stations where the species was absent. At each station, we placed three 25 m² square plots, 20 m apart along a 50 m transect perpendicular to the shoreline (starting 5 m from the natural high-water line; Fig. 2c). At both ends of the transect, we measured the mean canopy height with a clinometer, we evaluated canopy closure using a densitometer, and we recorded tree basal area using a forestry prism (factor 2). We identified all saplings (DBH < 9 cm) and assigned them to a diameter class in each plot.

Statistical analysis

Vegetation structure

We adjusted the habitat variables collected during these three vegetation surveys to ensure comparable parameters across the three study areas (Table 1). For each study area, we constructed a comparable set of logistic regression models to predict the probability of occurrence of CAWA at stations according to a priori hypotheses based on stand-scale fixed effects and covariates (see Appendix 2 for model description). Because of the limited number of sampled stations in each study area, we limited the candidate models to a maximum of three parameters to respect the concept of parsimony (Burnham and Anderson 2002). All models were ranked using the Akaike Information Criterion corrected for small samples (Burnham and Anderson 2002) and we applied a model averaging procedure to obtain conditional parameter estimates of the models having the lowest AICc values (given a ΔAICc < 2). We ensured that all variables used in models were not correlated (-0.7 > r > 0.7) and did not present multicollinearity (VIF < 3).

We performed model evaluation and averaging using the R-packages lme4 (Bates et al. 2015) and MuMin (Bartoń 2016), respectively. We also used lsmeans (Lenth 2016) to assess a posteriori comparisons of least-squares means (α = 0.05).

Plant community

Following Krikun et al. (2018), we used a set of statistical methods for each study area to evaluate habitat composition at stations based on the density of woody plants (i.e., small and large saplings, trees). First, we used a permutational ANOVA (PERMANOVA; Anderson et al. 2008) to assess significant differences in the general understory vegetation composition between stations with the presence and absence of CAWA. A resemblance matrix of all stations was computed using the Bray-Curtis similarity metric on the square root of each plant species density. We then evaluated which plant species were indicators of CAWA presence or absence using the Indicator species analysis (Dufrêne and Legendre 1997), which relies on a permutation procedure to test significant associations (α = 0.05). Finally, we used a non-metric multidimensional scaling (NMDS) to plot differences in the plant community structure between stations with and without CAWA as well as significant indicator plant species. These analyses were all performed using the R-package vegan (Oksanen et al. 2022).

RESULTS

Vegetation structure

We observed CAWA in 19, 11, and 25 stations in the mixed boreal forest, the agroforest, and the riverine forest, respectively. For the mixed boreal forest, one model that included the dominant canopy type as well as the density of small and large saplings accounted for 93% of the Akaike weight on the probability of presence of CAWA while in the agroforest, the top model included only the dominant canopy type and the density of small saplings representing 61% of the Akaike weight. In the riverine forest, the Null model was the best-supported model, preventing any meaningful inference about habitat selection (Table 2).

In the mixed boreal forest, the most influential variable on the occurrence of the CAWA was the density of large saplings (11.71 95CI [5.82; 20.71]), and forest stands dominated by deciduous cover had a higher probability of species occurrence than coniferous ones (3.75 95CI [0.88; 7.98]; Fig. 3a). Small saplings also had a weak positive effect on the occurrence of CAWA in this region (0.80 95CI [0.22; 1.72]; Fig. 3b). In the agroforest, CAWA had a higher probability of occurrence in mixed forest stands than in coniferous ones (2.29 95CI [0.58; 4.19]), and the density of small saplings had a slightly positive effect (0.61 95CI [0.06; 1.26]; Fig. 4). The average density of small and large saplings was the lowest in the agroforest and the highest in the riverine forest (Fig. 5). The difference between the three study areas was much more pronounced for small saplings than for large saplings.

Plant community

In the mixed boreal forest and the agroforest, the overall forest plant community was significantly different between stations with the occurrence of CAWA (F_1,36 = 4.38, p < 0.01 and F_1,59 = 1.62, p = 0.05, respectively), but we did not find a difference in the riverine forest (F_1,51 = 0.95, p = 0.49) (Fig. 6). Mountain maple saplings (small and large), small saplings of beaked hazelnut, and large saplings of pin cherry (Prunus pensylvanica) were positively associated with the occurrence of CAWA in the mixed boreal forest while small saplings of trembling aspen, white and black spruces and large saplings of balsam fir were negatively associated with the species’ occurrence. In agroforest, small saplings of balsam fir, rhodora, viburnum, and blueberry, and large saplings of red maple were positively associated with the occurrence of CAWA. In the riverine forest, only small saplings of white birch had a positive association with CAWA occurrence while American beech (Fagus grandifolia) trees had a negative association (see Appendix 3).

DISCUSSION

Our study assessed the fine-scale vegetation structure associated with the occurrence of CAWA during the breeding period in three different landscapes in Québec: the mixed boreal upland forest, the agroforestry landscape of the St. Lawrence lowlands, and the riverine deciduous forest of southern Québec. Cover type associated with the occurrence of CAWA varied across our study areas: a deciduous-dominated canopy was favored in mixed landscape of the upland boreal forest, while a coniferous component was favored in the deciduous agroforestry landscape of the St. Lawrence lowlands. The plant community associated with the presence of CAWA also differed between our three study areas: it was positively associated with saplings of mountain maple, beaked hazelnut, and pin cherry in the mixed boreal forest, whereas in the agroforest, it was positively associated with saplings of balsam fir, red maple, rhodora, and blueberry. The occurrence of CAWA was only positively associated with saplings of white birch in the riverine forest. However, we highlighted a consistent positive association between understory density (both small and large saplings) and CAWA occurrence, despite differences in magnitude of this association across study areas.

The relationship between nesting CAWA and a dense forest understory has been documented extensively across the species’ range (Flockhart et al. 2016, Reitsma et al. 2020, Dimmig et al. 2022). Because CAWA nests close to the ground in an open cup-nest, the nests’ concealment provided by a high density of stems appears to be a crucial factor ensuring nesting success (Goodnow and Reitsma 2011, Hunt et al. 2017), likely offering protection against predation. Flockhart et al. (2016) also demonstrated that CAWA density was higher, and their home range smaller, as the percentage of shrub cover increased. This supports the idea that CAWA concentrate in high-quality habitats with sufficient protection, where the quantity of food may not be limiting (Venier et al. 2012). In the present study, the average density of small saplings varied between study areas, likely due to differences in the plant community composition of this stratum; however, the difference between stations with and without CAWA was evident in the mixed boreal forest and the agroforest. On the other hand, the similarity in shrub layer density between stations in the riverine forest is likely related to the fact that they were all in the riparian zone, limiting the ability to distinguish between good and less favorable habitats. Riparian zones and proximity to watercourses are generally recognized favorable habitats for CAWA because of the high density of the shrub layer (Ball et al. 2016, Dimmig et al. 2022, Cupiche-Herrera et al. 2024), suggesting that habitat availability may not be a limiting factor, and that competition could instead drive local spatial dynamics (Flockhart et al. 2016).

The observed differences in the specific composition of plant species within the shrub layer across the study areas align with the marked variations in their overall habitats. The relationship between CAWA and certain plant species offers valuable insights into CAWA’s general habitat preferences within different landscapes, particularly because several indicator species are associated with poorly drained areas (Saucier et al. 2009). In our study, mountain maple is commonly associated with seepage drainage sites in the mixed boreal forest, where wet soil conditions are prevalent near the surface, while rhodora and blueberries are linked to poorly drained areas in the lowland agroforest (Saucier et al. 2009). These patterns extend to a broader scale, as CAWA is frequently found in wetter, marshier habitats (Reitsma et al. 2020). Regional differences in the shrub layer composition suggest that the structure provided by dense shrub cover, rather than specific plant species, may be the key factor driving CAWA habitat selection (Hallworth et al. 2008a).

The differences observed in canopy composition at CAWA sites between the mixed boreal forest and the agroforest study areas suggest that, on a broader scale, this species is associated with landscapes featuring a mixed component (Reitsma et al. 2020, Cupiche-Herrera et al. 2024). In the mixed boreal forest, as well as in the western Canadian boreal forest (Krikun et al. 2018), CAWA is associated with stands featuring an increased proportion of deciduous species within a predominantly coniferous landscape. Conversely, in agroforest, which reflects conditions found in south-eastern forests of its range (Hallworth et al. 2008b), CAWA is associated with stands that have an increased proportion of coniferous species within a predominantly deciduous landscape.

Limitations and methodological considerations

This study did not fully account for imperfect detectability of CAWA during point counts, which may have led to an underestimation of occupancy rates. Although we conducted surveys during the peak breeding period in Québec (Rousseau and Drolet 2017) and under favorable conditions, differences in survey methodologies across study areas may have influenced detection probability in distinct ways. In the mixed boreal forest, surveys relied on playback calls at regular intervals, a method known to increase detectability by eliciting territorial responses from males (Marion et al. 1981, Rae et al. 2015). In contrast, surveys in the agroforestry landscape used passive acoustic monitoring with automated recording units (ARUs), which allow for unbiased repeated sampling but may under-detect individuals that vocalize infrequently or in areas with high ambient noise (Venier et al. 2012, Campos-Cerqueira and Aide 2016, Darras et al. 2019). The riverine forest surveys combined fixed-radius point counts (50 m) with playback when initial detection failed, balancing the advantage of eliciting responses with the consistency of fixed-radius sampling (Hutto et al. 1986). However, riparian areas tend to have naturally high background noise due to water movement and wind, potentially affecting the detection of CAWA vocalizations (Pacifici et al. 2008).

To ensure comparability across study areas, we assumed similar detection probabilities despite differences in methods and habitats. Although standardized protocols and experienced observers helped reduce bias, some variation likely remains. Playback surveys may overestimate occupancy compared to passive ARUs, especially in noisy environments like riparian forests. These differences highlight the need for cautious interpretation across landscapes. Future studies should consider hierarchical occupancy models or repeated surveys to better account for imperfect detection.

Despite these limitations, the consistency of habitat relationships observed across study areas suggests that the key habitat parameters identified remain robust. Although differences in survey methodologies may have influenced detection rates, they also reinforce the generalizability of our findings by demonstrating the structural habitat associations of CAWA across a range of landscapes.

CONCLUSION

The present study highlights variations in the breeding habitat selection of CAWA across three regions with contrasting landscapes in terms of the specific composition of the understory and canopy cover. Our study reinforces the notion that CAWA can inhabit a wide range of different habitats across its range (Crosby et al. 2019, Reitsma et al. 2020, Leston et al. 2024). While highlighting an overall preference for mixed forest compositions, our study emphasizes the scale-dependent nature of habitat selection where the species selected forest stands with a higher proportion of deciduous trees within predominantly coniferous landscapes, and vice versa. However, within forest stands, our results underscore the predominance effect of forest structure (i.e., dense shrub density) over the specific vegetation composition, as previously reported for its role in nest concealment (Rangen et al. 1999, Reitsma et al. 2008, Krikun et al. 2018) and reproductive performance (Hallworth et al. 2008a).

Our study suggests that management for CAWA should not rely on a habitat per se (combination of cover type and age class) but rather on structural characteristics that can be found within different habitat types. Thus, a suitable habitat could comprise, for example, of a young beaked hazelnut understory in a mixed boreal forest (this study and Krikun et al. 2018), or a dense rhodora understory in a mixed deciduous forest (this study and Dimmig et al. 2022). It is also crucial to consider the differences in scales between management actions and the forest structure needed to support CAWA, while keeping in mind that favorable structural characteristics, such as those created by naturel gap dynamics or timber harvest, can be transient over time (Lambert and Faccio 2005). Therefore, in environments where ephemeral habitats are scarce, it would be relevant for forest planners to focus on other types of structurally more stable habitats (e.g., rhodora; Dimmig et al. 2022), or consider landscape-scale partial or total cutting rotations (Hallworth et al. 2008a).

Although this study has addressed key habitat components in Québec, certain aspects of CAWA ecology, such as conspecific and interspecific interactions (Flockhart et al. 2016), and finer-scale habitat elements like the number of perch trees, soil wetness, and edge effects (Reitsma et al. 2020), offer opportunities for future localized research. Investigating these factors will not only deepen our understanding of the species’ ecology but also provide insights to refine conservation strategies more effectively.

LITERATURE CITED

Anderson, M. J., R. N. Gorley, and K. R. Clarke. 2008. PERMANOVA+ for PRIMER: Guide to software and statistical methods. PRIMER-E, Plymouth, UK.

Ball, J. R., P. Sólymos, F. K. A. Schmiegelow, S. Haché, J. Schieck, and E. Bayne. 2016. Regional habitat needs of a nationally listed species, Canada Warbler (Cardellina canadensis), in Alberta, Canada. Avian Conservation and Ecology 11(2):10. https://doi.org/10.5751/ACE-00916-110210

Bartoń, K. 2016. MuMIn: Multi-model inference. R package version 1.15.6. https://doi.org/10.32614/CRAN.package.MuMIn

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

Becker, D. A., P. B. Wood, and P. D. Keyser. 2012. Canada Warbler use of harvested stands following timber management in the southern portion of their range. Forest Ecology and Management 276:1-9. https://doi.org/10.1016/j.foreco.2012.03.018

Bégin-Marchand, C., F. Gagnon, B. Drolet, J. Ibarzabal, and J. A. Tremblay. 2022. Insights on post-breeding movements from a northeastern population of Canada Warblers (Cardellina canadensis). Avian Conservation and Ecology 17(1):4. https://doi.org/10.5751/ACE-02013-170104

Brandt, J. P. 2009. The extent of the North American boreal zone. Environmental Reviews 17:101-161. https://doi.org/10.1139/A09-004

Burnham, K. P., and D. R. Anderson. 2002. Model selection and multimodel inference: a practical information-theoretic approach. Springer, New York, New York, USA.

Campos-Cerqueira, M., and T. M. Aide. 2016. Improving distribution data of threatened species by combining acoustic monitoring and occupancy modelling. Methods in Ecology and Evolution 7(11):1340-1348. https://doi.org/10.1111/2041-210X.12599

Chace, J. F., S. D. Faccio, and A. Chacko. 2009. Canada Warbler habitat use of northern hardwoods in Vermont. Northeastern Naturalist 16(4):491-500. https://doi.org/10.1656/045.016.n401

Committee on the Status of Endangered Wildlife in Canada (COSEWIC). 2008. COSEWIC assessment and status report on the Canada Warbler (Wilsonia canadensis) in Canada. COSEWIC, Ottawa, Ontario, Canada.

Committee on the Status of Endangered Wildlife in Canada (COSEWIC). 2020. COSEWIC assessment and status report on the Canada Warbler (Cardellina canadensis) in Canada. COSEWIC, Ottawa, Ontario, Canada.

Crosby, A. D., E. M. Bayne, S. G. Cumming, F. K. A. Schmiegelow, F. V. Dénes, and J. A. Tremblay. 2019. Differential habitat selection in boreal songbirds influences estimates of population size and distribution. Diversity and Distributions 25(12):1941-1953. https://doi.org/10.1111/ddi.12991

Cupiche-Herrera, V. J., A. R. Westwood, and B. E. McLaren. 2024. Field-validated species distribution model of Canada Warbler (Cardellina canadensis) in Northwestern Ontario. Avian Conservation and Ecology 19(2):10. https://doi.org/10.5751/ACE-02696-190210

Darras, K., P. Batáry, B. J. Furnas, I. Grass, Y. A. Mulyani, and T. Tscharntke. 2019. Autonomous sound recording outperforms human observation for sampling birds: a systematic map and user guide. Ecological Applications 29(6):e01954. https://doi.org/10.1002/eap.1954

Dimmig, G., C. Rota, P. Wood, and C. Lituma. 2022. Understory structure and heterospecifics influence the occupancy of a ground-nesting species of conservation concern, the Canada Warbler. Avian Conservation and Ecology 17(1):20. https://doi.org/10.5751/ACE-02079-170120

Dufrêne, M., and P. Legendre. 1997. Species assemblages and indicator species: the need for a flexible asymmetrical approach. Ecological Monographs 67:345-366. https://doi.org/10.2307/2963459

eBird. 2022. eBird basic dataset. Version: EBD_relNovember-2022. Cornell Lab of Ornithology, Ithaca, New York, USA.

Flockhart, D. T. T., G. W. Mitchell, R. G. Krikun, and E. M. Bayne. 2016. Factors driving territory size and breeding success in a threatened migratory songbird, the Canada Warbler. Avian Conservation and Ecology 11(2):4. https://doi.org/10.5751/ACE-00876-110204

Goodnow, M. L. and L. R. Reitsma. 2011. Nest-site selection in the Canada Warbler (Wilsonia canadensis) in central New Hampshire. Canadian Journal of Zoology 89(12):1172-1177. https://doi.org/10.1139/z11-094

Hallworth, M. T., P. M. Benham, J. D. Lambert, and L. Reitsma. 2008a. Canada Warbler (Wilsonia canadensis) breeding ecology in young forest stands compared to a red maple (Acer rubrum) swamp. Forest Ecology and Management 255(3-4):1353-1358. https://doi.org/10.1016/j.foreco.2007.10.045

Hallworth, M. T., A. M. Y. Ueland, E. Anderson, J. D. Lambert, and L. R. Reitsma. 2008b. Habitat selection and site fidelity of Canada Warblers (Wilsonia canadensis) in central New Hampshire. Auk 125(4):880-888. https://doi.org/10.1525/auk.2008.07115

Hunt, A. R., E. M. Bayne, and S. Haché. 2017. Forestry and conspecifics influence Canada Warbler (Cardellina canadensis) habitat use and reproductive activity in boreal Alberta, Canada. Condor 119(4):832-847. https://doi.org/10.1650/CONDOR-17-35.1

Hutto, R. L. , S. M. Pletschet, and P. Hendricks. 1986. A fixed-radius point count method for nonbreeding and breeding season use. Auk 103(3):593-602. https://doi.org/10.1093/auk/103.3.593

Krikun, R. G., J. L. McCune, E. M. Bayne, and D. T. T. Flockhart. 2018. Breeding habitat characteristics of Canada Warblers in central Alberta. Forestry Chronicle 94(3):230-239.

Lambert, J. D., and S. D. Faccio. 2005. Canada Warbler: population status, habitat use, and stewardship guidelines for Northeastern forests. Vermont Institute of Natural Science Technical Report 05-4, Woodstock, Vermont, USA.

Lenth, R. V. 2016. Least-squares means: The R package lsmeans. Journal of Statistical Software 69(1):1-33. https://doi.org/10.18637/jss.v069.i01

Leston, L., F. V. Dénes, T. Docherty, J. A. Tremblay, Y. Boulanger, S. L. Van Wilgenburg, D. Stralberg, P. Solymos, S. Haché, K. St-Laurent, R. Weeber, B. Drolet, A. R. Westwood, D. D. Hope, J. R. Ball, S. J. Song, S. G. Cumming, E. M. Bayne, and F. K. A. Schmiegelow. 2024. A framework to support the identification of critical habitat for wide-ranging species at risk under climate change. Biodiversity and Conservation 33(1):603-628. https://doi.org/10.1007/s10531-023-02761-1

Marion, W. R., T. E. O’Meara, and D. S. Maehr. 1981. Use of playback recordings in sampling elusive or secretive birds. Studies in Avian Biology 6:81-85.

Noon, B. R. 1981. Techniques for sampling avian habitats. Pages 42-52 in D. E. Capen, editor. The use of multivariate statistics in studies of wildlife habitat. General Technical Report RM-GTR-87. U.S. Forest Service, Rocky Mountain Forest and Range Experiment Station, Fort Collins, Colorado, USA.

Oksanen, J., G. L. Simpson, F. G. Blanchet, R. Kindt, P. Legendre, P. R. Minchin, R. B. O’Hara, P. Solymos, M. H. H. Stevens, E. Szoecs, et al. 2022. vegan: Community ecology package. R package version 2.6-2. https://doi.org/10.32614/CRAN.package.vegan

Pacifici, K., T. R. Simons and K. H. Pollock. 2008. Effects of vegetation and background noise on the detection process in auditory avian point-count surveys. Auk 125(3):600-607. https://doi.org/10.1525/auk.2008.07078

Rae, L. F., D. M. Whitaker, and I. G. Warkentin. 2015. Variable effect of playback of chickadee mobbing calls on detection probability of boreal forest birds. Journal of Field Ornithology 86(1):51-64. https://doi.org/10.1111/jofo.12088

Rangen, S. A., R. G. Clark, and K. A. Hobson. 1999. Influence of nest-site vegetation and predator community on the success of artificial songbird nests. Canadian Journal of Zoology 77(11):1676-1681. https://doi.org/10.1139/z99-141

Reitsma, L. R., M. T. Hallworth, and P. M. Benham. 2008. Does age influence territory size, habitat selection, and reproductive success of male Canada Warblers in central New Hampshire? Wilson Journal of Ornithology 120(3):446-454. https://doi.org/10.1676/07-109.1

Reitsma, L. R., M. T. Hallworth, M. McMahon, and C. J. Conway. 2020. Canada Warbler (Cardellina canadensis), version 2.0. In P. G. Rodewald and B. K. Keeney, editors. Birds of the world. Cornell Lab of Ornithology, Ithaca, New York, USA. https://doi.org/10.2173/bow.canwar.02

Rousseau, F., and B. Drolet. 2017. The nesting phenology of birds in Canada. Canadian Wildlife Service Technical Report Series No. 533, Environment and Climate Change Canada, Québec, Canada.

Sabo, S. R., and R. T. Holmes. 1983. Foraging niches and the structure of forest bird communities in contrasting montane habitats. Condor 85(2):121-138. https://doi.org/10.2307/1367245

Saucier, J. P., P. Grondin, A. Robitaille, J. Gosselin, C. Morneau, J. H. Richard, J. Brisson, L. Sirois, A. Leduc, H. Morin, E. Thiffault, S. Gauthier, C. Lavoie, and S. Payette. 2009. Écologie forestière. Pages 133-279 in Manuel de Foresterie. l’Ordre des ingénieurs forestiers du Québec et Éditions Multi-Mondes, Montréal, Québec, Canada.

Toussaint, D. 2019. Canada Warbler. Pages 576-577 in M. Robert, M.-H. Hachey, D. Lepage and A. R. Couturier, editors. Second atlas of the breeding birds of Southern Québec. Regroupement QuébecOiseaux, Canadian Wildlife Service and Bird Studies Canada, Montréal, Québec, Canada.

Venier, L. A., S. B. Holmes, J. L. Pearce, and R. E. Fournier. 2012. Misleading correlations: the case of the Canada Warbler and spruce budworm. Journal of Wildlife Management 76(2):294-298. https://doi.org/10.1002/jwmg.257

Westwood, A. R., J. D. Lambert, L. R. Reitsma, and D. Stralberg. 2020. Prioritizing areas for land conservation and forest management planning for the threatened Canada Warbler (Cardellina canadensis) in the Atlantic Northern Forest of Canada. Diversity 12(2):61. https://doi.org/10.3390/d12020061

Westwood, A. R., C. A. Staicer, P. Solymos, S. Haché, T. Fontaine, E. M. Bayne, and D. Mazerolle. 2019. Estimating the conservation value of protected areas in Maritime Canada for two species at risk: the Olive-sided Flycatcher (Contopus cooperi) and Canada Warbler (Cardellina canadensis). Avian Conservation and Ecology 14(1):16. https://doi.org/10.5751/ACE-01359-140116