INTRODUCTION

The North American Great Lakes (hereafter “Great Lakes”) collectively form one of the most important freshwater ecosystems in the world. Like many coastal regions, intensive population growth and urbanization have converted much of the Great Lakes coastal environment into human-dominated landscapes (Danz et al. 2007, Neumann et al. 2015, Culp et al. 2017). Precolonial coastal forests have been replaced extensively by agricultural lands and urban developments. Conversion of coastal habitats to urban and agricultural lands has altered the distribution and abundance of forest-dependent bird species, which has often resulted in regional declines in biodiversity (Freemark and Collins 1999, Rayner et al. 2015, Matuoka et al. 2020). Breeding birds of the Great Lakes eastern and boreal forests have experienced significant population declines since the 1970s—not only rare or threatened species, but also widespread and common species (Rosenberg et al. 2019). These declines may have a disproportionate influence on food web dynamics and ecosystem function (Rosenberg et al. 2019).

The Great Lakes basin consists of a mosaic of aquatic, wetland, and terrestrial habitats. The upland coastal zone of the Great Lakes is dominated by forests, which have experienced widespread reductions in size, distribution, quality, and composition in recent decades (Wolter et al. 2006, Schulte et al. 2007, Hollenhorst et al. 2011, Hupperts et al. 2019). These forests play a critical role in supporting regional breeding bird diversity; however, an estimated 22–45% decline in regional breeding populations of landbirds has been associated with habitat loss caused by human land use (Soulliere et al. 2020). For example, significant decreases in coniferous forest types and forest-dependent bird species have been documented within the Great Lakes region (Danz et al. 2007, Miller et al. 2007, Schulte et al. 2007, Ford and Flaspohler 2010, Hollenhorst et al. 2011, Gnass Giese et al. 2015, Niemi et al. 2016, Guiry et al. 2020). Similar human-caused declines have been demonstrated for Great Lakes wetland bird communities (e.g., Brazner et al. 2007, Howe et al. 2007, Tozer 2016, Grand et al. 2020). However, less is known about how changes in human land use influence ecosystem function in this coastal zone, particularly for birds that rely on forested habitats.

To characterize and understand the effect of habitat loss on bird diversity, it is important to identify how human land use influences the taxonomic and functional attributes of avian assemblages (Sekercioglu 2006, Whelan et al. 2008). Traditionally, community diversity has been measured using taxonomic-based metrics only; however, this approach ignores the diversity of ecosystem functions represented by species within the community. The inclusion of functional characteristics (i.e., traits) can provide a more comprehensive characterization of differences in diversity (O’Connell et al. 2000, Bishop and Myers 2005, Mouchet et al. 2010). For example, loss of function in an ecosystem occurs when a group of species that rely on the same resource or feeding mode disappears entirely. Grouping species based on shared resource use allows for broader generalizations about conservation needs than are possible by considering species individually (Lesak et al. 2011, Weisberg et al. 2014). Identifying changes in the range of functional traits present in an assemblage can help frame the potential impacts of human land use on ecological communities and the services they provide, including insect population control, pollination, and seed dispersal (Luck et al. 2012). For birds, common measures of functional diversity often include traits associated with food type and foraging behavior, body mass, and migratory strategy (Matuoka et al. 2020). Understanding the effects of forest loss and fragmentation on avian functional and taxonomic diversity is necessary to effectively inform and prioritize conservation actions (Bełcik et al. 2020).

We used breeding bird survey data collected as part of the Great Lakes Environmental Indicators Projects (GLEI-1: 2002–2003 and GLEI-2: 2012) (Niemi et al. 2007, 2009, Johnson et al. 2015) to explore the effects of human land use on functional and taxonomic diversity of upland birds in the Great Lakes coastal zone. The primary objectives of the GLEI-1 and GLEI-2 projects were to quantify a watershed-scale stress gradient based on human land use/inputs (e.g., agriculture, population, point source pollution) and to develop a robust sampling design to assess multi-taxa biotic responses across this defined stress gradient throughout the U.S. Great Lakes basin (Danz et al. 2005, Johnson et al. 2015, Host et al. 2019).

Bird surveys were conducted across a gradient of human land uses in nearshore upland zones, defined as the terrestrial region from the shoreline to 1 km inland (Danz et al. 2005); we call this region the “upland” coastal zone. The primary objectives of our analyses were to (1) document how different measures of taxonomic and functional diversity change in response to increased human land use, based on a quantitative stress gradient, (2) determine the degree to which changes in diversity measures were associated with spatial confounding factors, and (3) identify associations between species abundance, species traits, and land cover types throughout the upland coastal zone of the Great Lakes. We hypothesized that both taxonomic and functional diversity would decrease with increasing levels of human land use (Matuoka et al. 2020) and that functional groups associated with forest resources would be particularly vulnerable to human impacts in the upland coastal zone.

METHODS

Study area

We analyzed data collected at 199 survey locations (sites) spread across the upland coastal zone of the five Great Lakes in the United States and Canada (GLEI-1: n = 156 sites in the United States; GLEI-2: n = 20 sites in the United States and n = 23 sites in Canada) (Fig. 1). The distribution of sites by Great Lake was as follows: Superior (n = 47), Michigan (n = 58), Huron (n = 34), Erie (n = 37), and Ontario (n = 23). The study area was located primarily within the Laurentian Mixed Forest Province and Eastern Broadleaf Forest Province ecoregions (Bailey et al. 1994, Crins et al. 2009). These ecoregions separate the Great Lakes basin into two broad classes of roughly equal size: North, which is composed of Laurentian Mixed Forest and the Boreal Shield, and South, representing the Eastern Broadleaf Forest Province and the Mixedwood Plains. The northern portion of the Great Lakes basin lies either entirely within the Boreal Shield or within the Boreal Hardwood Transition zone dominated by coniferous and broadleaf deciduous forests (Bailey et al. 1994, Crins et al. 2009). The southern portion of the Great Lakes basin consists primarily of prairie, broadleaf deciduous forest (beech–maple forest), and oak savanna, which has largely been converted to agricultural lands (Bailey et al. 1994). Of the 199 sites, 113 were located in the northern portion and 86 were in the southern portion of the Great Lakes basin (Fig. 1). We included both lake (i.e., each of the five Great Lakes) and ecoregion (North or South) in our study to determine the relative importance of these potential spatial confounding factors in explaining variability in functional and taxonomic diversity measures in the upland coastal zone of the Great Lakes.

Site selection and quantification of the human land use gradient

The GLEI-1 project incorporated both local-scale (e.g., soils, point source pollution) and large-scale (e.g., land cover, atmospheric deposition) environmental variables (n = 207 variables) to quantify a watershed-level stress gradient across the Great Lakes basin (Danz et al. 2005, 2007). This provided the basis for determining sampling locations for the bird data collected in 2002–2003 based on an environmentally stratified random design (Danz et al. 2005). Watershed delineation followed the methods of Hollenhorst et al. (2007) and Host et al. (2011). Esri’s ArcHydro model based on 20-m digital elevation models (Canada) and 10-m digital elevation models (U.S.) were used to delineate watersheds across the Great Lakes basin (n = 5971 watersheds).

A refinement of the GLEI-1 stressor gradient included only five variables quantified at the watershed-level (percent agriculture, percent urban land use, road density, population density, and point source density). Population density was derived from census data (people per square kilometer, area-weighted average from census blocks, summarized by watershed) (Host et al. 2011, 2019). Road density was summarized as kilometer of road length per square kilometer of land area within the watershed, weighted by road class (Host et al. 2011). Point source stressors included sewage, pathogens, polycyclic aromatic hydrocarbons, solvents, nutrients, salts, and pharmaceuticals, with density calculated as the sum of weighted point source scores by watershed (Host et al. 2011). Each of these five variables was scaled from 0 to 1 based on basin-wide data. An additional variable—the sum of the scaled variables within each watershed (SumRel)— was used as an index of cumulative stress. Low SumRel values were associated with areas of low anthropogenic stress, while high values were associated with high levels of stress (Host et al. 2011). The SumRel cumulative stress index was used as the site selection gradient for GLEI-2 sampling locations. Sites were selected to represent the full extent of the gradient from low to high human land use and from low to high forest cover (Fig. A1.1).

Development of a human land use gradient that would represent environmental conditions in both the United States and Canada was limited by differences in data availability and compatibility. Given agriculture and development are the two dominant human land uses across the basin, we quantified two indices of land use within each watershed: (1) percent agricultural land use (%Ag) based on the National Land Cover Database (NLCD), Canadian Provincial Land Cover (PLC), and National Land and Water Information Service (NLWIS) from Land Information Ontario, and (2) percent development land use (%Dev) based on NLCD, PLC, and NLWIS as well as population density and road density. High-energy coastline polygons (United States) and lines (Canada) were buffered by 500 km inland and in either direction along the shoreline, and values for both indices were determined by area-weighting values from overlapping watersheds based on the area of each contributing watershed (Johnson et al. 2015). All data used to develop the human land use gradient were from 2000–2001 data sources. Both indices were scaled from 0 to 1 based on basin-wide data. A composite index of human land use was calculated using Euclidean distances from the 0,0 origin to x,y coordinate of scaled Ag and Dev indices (Host et al. 2019).

We buffered each bird point count location (n = 14–15) within each site by 1 km and merged overlapping polygons (Fig. 1). Within each merged polygon, the proportion of agricultural land use (Ag) (i.e., cropland), proportion of development land use (Dev), and composite human land use values were summarized using the Zonal Statistics and National Water-Quality Assessment tools in ArcMap. We used the composite values to represent human land use in this study, which ranged from 0.02 to 0.88, with low values representing sites with little human land use and high values representing sites with extensive human land use (Fig. A1.2). Due to variation in the distance of each point count location to open water, there was variation in the area of each buffered site, with mean area ± SD = 1370 ± 360 m.

Land cover data

We used the 2001 and 2011 U.S. Geological Survey NLCD for upland sites located in the United States (30 × 30 m resolution) to align with time periods in which bird surveys were conducted (2002–2003; 2012) (Homer et al. 2007, 2015), and the 2000 Ontario Land Cover (OLC) data, which were available for upland sites located in Canada (25 × 25 m resolution) (OLC 2002). Using ArcGIS version 10.2.2 (ESRI 2011), we created a 15-km buffer around the Great Lakes, clipped the NLCD and OLC shapefiles to the buffer, and reclassified data into nine common land cover classes (Table A1.1). We then buffered each bird point count location within each site by 1 km and merged overlapping polygons (Fig. 1). Within each buffered site, we used the National Water-Quality Assessment Area-Characterization tool in ArcMap to calculate total land area and determine the proportion of land area covered by each of the nine land cover classes. Areas of open water were removed from the analyses because we were interested only in effects of human land use occurring within the upland coastal zone. We also removed “developed” and “cropland” from the analyses because those two land cover classes were used to develop the composite land use gradient. The dominant land cover classes, based on the proportion of the five remaining land cover classes and the composite human land use gradient, were “agriculture and development” and “forest” (Fig. A1.3).

Bird surveys

Survey locations

Site selection for the 2002–2003 surveys was determined based on an environmentally stratified sampling design, described in “Site selection and quantification of the human land use gradient” from the GLEI-1 project. Site selection for 2012 (GLEI-2) followed a similar process but used the refined GLEI-1 stress gradient (Host et al. 2011) and included sites located in Canada. Once sampling locations were initially selected, accessibility was evaluated using aerial photographs and maps in GIS (Danz et al. 2005). If a sampling location was deemed inaccessible, another randomly chosen location within the coastal segment was evaluated. This process was repeated until the desired number of sites was deemed appropriate for inclusion across the human land use gradient. Sites may have also been removed from sampling based on conditions encountered during site visits (e.g., inaccessibility due to property ownership or safety concerns).

We surveyed 199 sites during two sample periods (2002–2003 and 2012). Each sample (= “site”) consisted of 14–15 unlimited-distance, 10-minute, fixed-location point counts, which resulted in a total of 2982 bird point count surveys. Surveyors were provided with site maps which included a polygon or line delineating the extent of the high-energy shoreline at each site (Johnson et al. 2015). Point count locations were then randomly chosen by the surveyor, based on accessibility, to fall within the intended survey area, while maintaining a minimum distance of 500 m between point count locations, and were located along secondary or smaller roads, when possible (Miller et al. 2007). Sites were sampled once during the breeding season from late May to early July. All birds seen or heard during a point count survey were recorded on a data sheet binned by distance, with a 100-m radius circle centered on the observer. Observations were recorded as occurring either within 0–100 m from the observer or >100 m. Observation type (i.e., behavior) was also recorded for each individual, and included singing, calling, drumming, visual observation, aerial foraging, or flyover (Miller et al. 2007). Birds were sampled between 0.5 hours before sunrise to 4 hours after sunrise and only during favorable weather conditions (e.g., no rain and low wind speed). For this study, we restricted the data set by excluding birds that were observed outside the 100-m survey radius, flyovers (i.e., individuals not using the habitat within the survey location), migrant species that do not breed in the Great Lakes region, and species commonly associated with open water or shoreline habitat (Table A1.2). We chose to restrict data in this way to reduce variability caused by differences in species detectability, to avoid making assumptions about habitat use for birds detected only as flyovers (Niemi et al. 2016), and to remove water- and shoreline-dependent species (e.g., Caspian Tern [Hydroprogne caspia], American Bittern [Botaurus lentiginosus]), which are not ecologically relevant to upland coastal habitats.

Functional traits

We categorized species based on functional traits and habitat specialization. Previously published studies (Luck et al. 2013, Coetzee and Chown 2016, Jacoboski and Hartz 2020, Matuoka et al. 2020) helped us identify four classes of noncorrelated functional traits that were determined most likely to be influenced by human land use modification (Luck et al. 2013). Functional traits included body size, diet and foraging behaviors, nesting location, migratory strategy, and breeding habitat (Table 1). The morphometric trait (body mass [g]) is associated with metabolic rate, foraging rate, and life span (Luck et al. 2013). Because the primary influence of an organism on ecosystem function is associated with what and how they eat (Luck et al. 2013), we included measures of diet (i.e., primary diet and diet plasticity) and foraging behavior (i.e., method, substrate, and substrate plasticity) for each species. Nesting location was included as a functional trait because changes in forest structure and size could negatively impact birds that nest on the ground or in trees (Luck et al. 2012). Migratory strategy (i.e., long distance, short distance, resident, nomadic, and irregular) was included because the energetic cost associated with migration could have impacts on individual fitness if the necessary resources were lacking or larger distances of travel were required to obtain them (Alves et al. 2013). We also included habitat specialization (i.e., primary breeding habitat and breeding habitat plasticity) in the analyses because the ability of a species to use a variety of habitat types will increase their capacity to persist in a given landscape when changes in habitat availability occur either naturally or due to human modifications (Luck et al. 2013).

Functional traits associated with each species in our study were sourced from Wilman et al. (2014) and Billerman et al. (2020) (Table 1). We calculated measures of habitat plasticity using methods similar to those outlined in Luck et al. (2013), which used the frequency of species occurrence across different habitat types as a measure of plasticity. To estimate habitat plasticity in our study, we quantified site-level land cover data based on NLCD classifications. We first determined the dominant land cover class at each site using the merged NLCD and OLC data previously described. We considered any single land cover class that comprised ≥ 0.60 to be dominant cover. If no single land cover class was at or greater than 0.60, we included the next most dominant land cover class or classes until the threshold value was reached. For example, a site where “forest” = 0.40 and the next most abundant cover type was “developed” = 0.32, we added the two values to create a new “land cover class” (i.e., ForDev = 0.72). This process was repeated for all sites and resulted in 42 unique land cover classes. Any unique land cover classes created during this process were used solely to derive plasticity measures because they were only relative scores used to describe the degree of habitat specialization for each species detected in our study region. We summed the frequency of occurrence for each species in each land cover class and used that value (ranging from 1 to 42) as our measure of habitat plasticity. Therefore, low values represented species with high habitat specialization (e.g., Swainson’s Thrush [Catharus ustulatus] detected in 5 of 42 land cover classes), mid-range values (e.g., Alder Flycatcher [Empidonax alnorum] detected in 23 of 42 land cover classes) represented species with moderate habitat specialization, and high values represented species with low habitat specialization (e.g., Song Sparrow [Melospiza melodia] detected in 42 of the 42 land cover classes). The resulting plasticity measure was inherently coarse but adequate for providing a relative index of habitat specialization.

Statistical analyses

We considered each site (n = 199) a sampling unit for our analyses and calculated three multidimensional indices of functional diversity—Rao’s quadratic entropy (RaoQ), functional evenness (FEve), and functional dispersion (FDis) (Mason and Mouillot 2013)—along the human land use gradient using the R package “FD” (Laliberte and Legendre 2010). RaoQ measures the degree of functional dissimilarity among species in a community based on relative abundance (Botta-Dukát 2005). Functional evenness considers how regularly species abundances are distributed within the functional space (Villéger et al. 2010) and assumes that resource availability is even and is maximized by an even distribution of both species and abundance in trait space (Mason et al. 2005, Mouchet et al. 2010). Functional dispersion is the weighted mean distance in functional space of individual species relative to the weighted centroid of all species (Laliberte and Legendre 2010). Having both continuous and categorical traits, we used Gower distance to convert “species × trait matrix” to a distance matrix.

To examine the relationship between functional and taxonomic diversity indices and to identify how they vary relative to species richness (SR), we calculated Shannon’s species diversity index (H’) and SR using the R package “vegan” (Oksanen 2017) and compared those measures of diversity to that of RaoQ. We then used hierarchical partitioning analysis to determine the relative importance of spatial confounding factors (i.e., lake and ecoregion) in explaining variability in functional and taxonomic diversity indices relative to the human land use gradient. Although both ecoregions and lakes describe differences in large-scale environmental conditions, they still provide different information. For example, both can indicate latitudinal differences in species ranges, but the ecoregion tells us more about differences in land use, climate, and forest type across the basin (Danz et al. 2005). Data were analyzed and significance was assessed based on Z-scores produced using 100 randomizations in the R package “hier.part” (Mac Nally and Walsh 2004). Statistical significance was based on the upper 0.95 confidence limit (Z ≥ 1.65). The goodness-of-fit measure was R-squared using family = “Gaussian”, and it represents the maximum proportion of variation explained by the best combination of predictors based on hierarchical models. To facilitate ease of reading, we use the abbreviations FEve, FDis, SR, and H’ only when referring to results, with the exception of RaoQ.

Finally, we identified the strength and direction of the relationships between species abundance, functional traits, and land cover composition using RQL and fourth-corner analyses (Wang et al. 2020). We calculated “environment (R) × trait (Q) × species (L)” interactions using the function “traitglm”, with family = “negative binomial” and method = “glm1path” in the R package “mvabund” (Wang et al. 2020). Method “glm1path” applies a model selection algorithm (Least Absolute Shrinkage and Selection Operator [LASSO]), which reduces the number of predictor terms in the model to enhance prediction accuracy (Wang et al. 2020). The coefficients from the model output were used to describe variation in species traits relative to land cover composition in our study area. We used multivariate glm to identify species-specific associations with the composite human land use gradient and forest land cover class. To calculate parameter estimates for individual species, we used the function “manyglm” with family = negative binomial (link = “log”) distribution in R package “mvabund” to fit two univariate generalized linear models to multivariate abundance data (manyglm = species abundance ~ proportion forest) [model 1] and (manyglm = species abundance ~ proportion human land use gradient) [model 2]. We checked model assumptions using the plot function and assessed significance using analysis of variance (ANOVA) with the function “anova.manyglm” in package “mvabund” (Wang et al. 2020). All analyses were conducted in R version 3.6.1 (R Core Team 2019).

RESULTS

After imposing restrictions on distance of observation (≤ 100 m) and removing flyover observations, migrants, and species associated with open water and shoreline habitats, a total of 140 species and 50,203 individuals were included in the analyses (Table A1.2). For breeding bird assemblages in the upland coastal zone of the Great Lakes, functional evenness, species richness, and Shannon’s diversity index declined significantly along the human land use gradient (FEve: R = -0.49, p = 0.001; SR: R = -0.16, p = 0.03; H’: R = -0.41, p = 0.001), while RaoQ and functional dispersion remained relatively constant (Fig. 2).

Results from the hierarchical partitioning analysis indicated that spatial confounding factors (i.e., lake and ecoregion) accounted for variation in all functional and taxonomic diversity measures (Fig. 3). Although the percent independent effect (%I) was highest for a lake effect in all hierarchical partitioning models, the goodness-of-fit measures suggested the maximum proportion of variation explained by the best combination of predictors was low for SR (goodness-of-fit = 0.09), FDis (0.02), and RaoQ (0.03) but higher for FEve (0.28) and H’ (0.25). Z-scores indicated that a significant proportion of the variation in FEve and H’ was accounted for by lake (27% and 40%, respectively) and ecoregion (19% and 9%, respectively) (Fig. A1.4). The human land use gradient accounted for a significant proportion of variation in FEve (53%), H’ (51%), and SR (45%) (Fig. A1.4). There was a moderate positive association between H’ and RaoQ (R = 0.47; p = 0.001) (Fig. A1.5).

Across the five Great Lakes, the largest proportion of land cover classes represented in our study included “forest” and “agriculture and development”; i.e., the composite human land use index (Fig. A1.3). Therefore, we present results on the species traits and habitat associations for these two correlated land cover classes (R = -0.75, p = 0.00) (Fig. A1.6). Species traits and species composition that showed positive associations with forest often resulted in negative associations with increasing human land use (Fig. 4). For example, 89% of species associations that were significantly positive for one land cover type were significantly negative for the other (Table A1.3). Thus, we describe only correlations between species traits and species composition as they related to the proportion of forest land cover, while noting any correlations that were not complimentary. We acknowledge that due to the partial correlation of lake and ecoregion and the strong correlation between human land use and ecoregion, discerning the influence of each was unachievable with a basin-wide analysis. Additionally, it is likely that the land cover class “grassland and pasture” was dominated by non-native grasslands used for agricultural purposes, particularly in Ontario, where this land cover class made up > 20% of the land cover classified within the surveyed areas. Therefore, it is possible that the actual effect of land use may be underrepresented in this study due to these confounding factors.

Results from the fourth-corner analysis predicted that basin-wide, functional traits associated with species that nest on the ground, eat primarily invertebrates, and forage on the bark or leaves of trees were positively associated with the proportion of forest (Fig. 4). In contrast, species that nest on platforms and in holes, are classified as omnivores, and are dependent on ground or water for foraging substrates were negatively associated with the proportion of forest (Fig. 4). Species defined as being long-distance or irregular migrants were positively associated with the proportion of forest, while those defined as resident species showed a negative association with it (Fig. 4). Functional traits that were significantly associated with increasing human land use only included increases in species whose primary diet was plants and seeds, and a slight increase in species with higher body mass, but declines in tree-nesting species and in species that require water as a foraging substrate (Fig. 4).

When comparing species-specific associations with the proportion of forest and proportion of human land use within our study area, using multivariate glms, we found the number of species with positive (+) and negative (-) associations were similar (with increasing proportion forest: 36 [-] and 38 [+]; with increasing proportion human land use: 40 [-] and 33 [+]) (Table A1.3). Regarding functional traits, most species that were positively associated with the proportion of forest land cover had diets that consisted primarily of invertebrates (74%) and were primarily species that forage on foliage insects or bark (63%) (Table A1.2). Species that were negatively associated with the proportion of forest were those commonly associated with open woodlands and grasslands (e.g., Cedar Waxwing [Bombycilla cedrorum], Grasshopper Sparrow [Ammodramus savannarum]) and those commonly associated with human-dominated landscapes (i.e., urban or agricultural lands), four of which were non-native species: House Sparrow (Passer domesticus), House Finch (Carpodacus mexicanus), Rock Pigeon (Columba livia), and European Starling (Sturnus vulgaris) (Table A1.3). Of the 36 species that declined with increasing proportion of forest, 61% were ground foragers, 69% were short-distance migratory or resident species, and ~50% had diets composed primarily of plants and seeds or were categorized as omnivores and ~50% had primarily invertebrate-based diets (Tables A1.1 and A1.3).

DISCUSSION

The negative effects of deforestation on breeding bird communities have been documented within the Great Lakes region and in many forested ecosystems worldwide (Robinson et al. 1995, Howe et al. 2007, Miller et al. 2007, Gibson et al. 2011, Gnass Giese et al. 2015, Matuoka et al. 2020, Rurangwa et al. 2021). Our study is the first to compare functional diversity of upland breeding bird communities with traditional measures of community diversity across a human land use gradient in the Great Lakes region. This novel approach is valuable because inclusion of functional traits helps clarify the ecosystem services provided by species and the dependencies of those species on measurable features of the environment (Laureto et al. 2015). Relationships between the human land use gradient and diversity measures were variable, which highlights the importance of using multiple metrics to assess changes in community composition and diversity.

Habitat loss is often accompanied by increases in generalist species and high species turnover rates (De Coster et al. 2015). Our study showed declines in functional evenness and species diversity in response to human land use, which suggests that niche space may be underutilized, which could increase vulnerability to invasive species (Mason et al. 2005, Jonason et al. 2017). This is consistent with other studies that have found functional evenness to increase with native forest cover (Barbaro et al. 2014, Coetzee and Chown 2016). Moreover, when functional dispersion remains constant but functional evenness declines across a defined gradient, as in our study, changes are most likely occurring in species with lower abundances (Barbaro et al. 2014, Coetzee and Chown 2016).

Replacement of one species with another species that exhibits the same functional traits does not influence functional trait diversity; therefore, important negative effects of habitat loss might be masked or underestimated when only measures of functional diversity are considered (De Coster et al. 2015). This has been documented in other forest bird communities where habitat loss resulted in species loss and changes in species composition but little or no loss of functional diversity (Devictor et al. 2010, De Coster et al. 2015, Coetzee and Chown 2016, Matuoka et al. 2020). The high degree of redundancy measured by functional diversity indices at present does not appear effective in capturing the intricate life history characteristics embodied in all species of a community. This lack of sensitivity, when changes in bird species communities are observed, highlights the challenges of using functional diversity indices alone to describe ecosystem function and stability. Because birds are more restricted in their range of ecological strategies than expected, relative to their high species richness, it is suggested that taxonomic and functional diversity could be generated by different processes (Cooke et al. 2019). Understanding drivers of functional diversity and the ramifications of losing species will be critical for addressing ecosystem-level conservation (Bishop and Myers 2005).

Fragmentation of forests has been shown to change structural composition and microhabitat conditions (e.g., increased habitat homogeneity) and is often accompanied by declines in forest specialists during the breeding season, including reductions in abundance and foraging rates of foliage gleaning birds (Niemi et al. 2004, Zuckerberg and Porter 2010, Luck et al. 2013, Murray et al. 2014, Weisberg et al. 2014, Ehlers Smith et al. 2018, Endenburg et al. 2019, Bełcik et al. 2020). Our study indicated that many of the species that declined with increased human land use were insectivorous, long-distance migratory species which rely on forests for nesting and foraging. Population declines in species with these life history traits have been observed throughout human-altered forested ecosystems globally, which suggests that forest-dependent species are increasingly vulnerable to habitat modification and reductions in habitat availability (Schulze et al. 2019). Several species categorized as having irregular migration strategies—i.e., dispersing from their normal range often to exploit opportunistic food sources (e.g., Evening Grosbeak [Coccothraustes vespertinus] and Purple Finch [Haemorhous purpureus])—were shown to decline with increased human land use. Species that readily used urbanized and agricultural lands for nesting (e.g., human-created structures such as platforms, nest boxes, or eaves) and species that predominantly eat plants and seeds, by contrast, increased.

As species range shifts continue to occur during both breeding and non-breeding periods in response to a changing climate, continued monitoring efforts in the upland coastal zone of the Great Lakes is necessary (Culp et al. 2017, Grinde et al. 2017, Zurell et al. 2018, Bateman et al. 2020). Monitoring data will allow us to assess the impacts of future changes in human land use, identify how restoration actions influence bird community assemblages, and identify priority areas for conservation in the Great Lakes region. For example, identifying potential thresholds of functional diversity across the human land use gradient could help identify areas for conservation and restoration that could be used by landowners, stakeholders, and agencies for coordinated conservation actions (Zuckerberg and Porter 2010). Conservation and restoration efforts often focus on single-species that have been designated as species of conservation concern by state or federal agencies. However, because declines in common and widespread species can disproportionately influence ecosystem function (Rosenberg et al. 2019), we advocate for using a trait-based approach to aid in the development of indicator metrics. Trait-based indicator metrics for the Great Lakes coastal zone can be used to evaluate the condition of ecosystems, help identify sensitive functional groups, and set conservation priorities. Specifically, identifying which traits are most strongly tied to changes in forest resources and which species possess multiple traits with strong forest associations can provide an index of vulnerability for both species and functional groups in the region.

Additional studies of functional composition are necessary to identify habitat-scale features that promote higher taxonomic and functional diversity in the upland coastal zone. We sought to describe diversity patterns for birds across the Great Lakes basin at a scale that was ecologically meaningful and at which conservation actions are often most effective (Zhao et al. 2013, Mattsson et al. 2020). However, because birds respond to habitat characteristics at multiple spatial scales (e.g., microhabitat [m²], forest patch [< 100 ha], and landscape [> 100 ha]) (Niemi et al. 2004, Grinde et al. 2017), using a multiscale approach to assess differences in species–habitat relationships at the lake or ecoregion level is an important next step in developing landscape- and habitat-scale plans. For example, if the suitability of a forest patch for supporting forest-dependent birds is contingent upon the amount of forest surrounding the patch, that would suggest that management goals for sustaining forest-dependent guilds would best be met by landscape-scale management plans. Identifying how different traits associated with forest-dependent species respond at multiple spatial scales may provide additional insight into the importance of habitat connectivity and would aid in identifying the appropriate scale at which management should be implemented for different functional groups across the basin. Overall, enacting best forest management practices that have been shown to be beneficial for forest birds, such as maintaining or increasing heterogeneity in forest structure and composition, including retention of threshold densities of woody debris or cavity-bearing old trees, will help increase functional diversity for species with diverse traits (Schulze et al. 2019, Remeš et al. 2021). Forest management that includes permanent canopy cover and selective harvests has also been shown to be more effective in promoting forest bird abundance and biodiversity than unmanaged forests (Schulze et al. 2019). However, habitat-specific (e.g., aspen, pine) or bird species-specific forest management guidelines that can be used across ownership (e.g., private, federal, state, county) for the coastal zone would be beneficial and an important next step in aiding conservation efforts.

CONCLUSIONS

Human land use modifications, coupled with predicted impacts of climate change (e.g., changes in food availability, increased competition), pose major threats to landbird populations in the Great Lakes region (Rosenberg et al. 2016, Culp et al. 2017). Neotropical migrants are particularly vulnerable to these changing conditions as their southern ranges contract while no measurable shifts are being observed at the northern leading edge (Zurell et al. 2018, Rushing et al. 2020). Many long-distance migrants that are dependent on insects during the breeding season in northern climates, including foliage gleaning and aerial insectivores, are in population decline in North America, including in the Great Lakes region (Nebel et al. 2010, Rosenberg et al. 2019). The long-term consequences of these population changes and subsequent reductions in insectivory may have serious impacts on the stability and ecological health of forested ecosystems in the Great Lakes region. Additional research is needed to identify what is causing this negative relationship (e.g., pesticide use, lack of suitable stopover habitat). Understanding the ecological consequences associated with loss of avian taxonomic and functional diversity will be fundamental to mitigating these effects through targeted conservation actions.

Our study highlights the importance of using multiple metrics to assess changes in community composition and diversity of breeding birds relative to human land use in the upland coastal zone of the Great Lakes. We were able to document changes in diversity measures and functional traits using coarse-scale habitat classifications and land cover data, which suggests that landscape-level analyses were able to effectively describe important drivers of avian diversity in the region. By assessing the relationship between functional traits and land cover, we were able to describe which traits were most strongly associated with forest cover, and therefore, which traits and suite of species possessing those traits were most likely to be negatively impacted by forest loss or degradation. This information can then be used to predict future changes in diversity based on different land use and climate change scenarios and can inform landscape-level conservation actions aimed at mitigating effects of forest loss on avian biodiversity and ecosystem function.

LITERATURE CITED

Alves, J. A., T. G. Gunnarsson, D. B. Hayhow, G. F. Appleton, P. M. Potts, W. J. Sutherland, and J. A. Gill. 2013. Costs, benefits, and fitness consequences of different migratory strategies. Ecology 94:11-17. https://doi.org/10.1890/12-0737.1

Bailey, R. G., P. E. Avers, T. King, and W. H. McNab. 1994. Ecoregions and subregions of the United States (map). W. H. McNab and R. G. Bailey, editors. U.S. Forest Service.

Barbaro, L., B. Giffard, Y. Charbonnier, I. van Halder, and E. G. Brockerhoff. 2014. Bird functional diversity enhances insectivory at forest edges: a transcontinental experiment. Diversity and Distributions 20:149-159. https://doi.org/10.1111/ddi.12132

Bateman, B. L., C. Wilsey, L. Taylor, J. Wu, G. S. LeBaron, and G. Langham. 2020. North American birds require mitigation and adaptation to reduce vulnerability to climate change. Conservation Science and Practice 2:e242. https://doi.org/10.1111/csp2.242

Bełcik, M., M. Lenda, T. Amano, and P. Skórka. 2020. Different response of the taxonomic, phylogenetic and functional diversity of birds to forest fragmentation. Scientific Reports 10:20320. https://doi.org/10.1038/s41598-020-76917-2

Billerman, S. M., B. K. Keeney, P. G. Rodewald, and T. S. Schulenberg. 2020. Birds of the world. Cornell Laboratory of Ornithology, Ithaca, New York, USA.

Bishop, J. A., and W. L. Myers. 2005. Associations between avian functional guild response and regional landscape properties for conservation planning. Ecological Indicators 5:33-48. https://doi.org/10.1016/j.ecolind.2004.10.001

Botta-Dukát, Z. 2005. Rao’s quadratic entropy as a measure of functional diversity based on multiple traits. Journal of Vegetation Science 16:533-540. https://doi.org/10.1111/j.1654-1103.2005.tb02393.x

Brazner, J. C., N. P. Danz, A. S. Trebitz, G. J. Niemi, R. R. Regal, T. Hollenhorst, G. E. Host, E. D. Reavie, T. N. Brown, J. A. M. Hanowski, C. A. Johnston, L. B. Johnson, R. W. Howe, and J. J. H. Ciborowski. 2007. Responsiveness of Great Lakes wetland indicators to human disturbances at multiple spatial scales: a multi-assemblage assessment. Journal of Great Lakes Research 33:42-66. https://doi.org/10.3394/0380-1330(2007)33[42:ROGLWI]2.0.CO;2

Coetzee, B. W. T., and S. L. Chown. 2016. Land-use change promotes avian diversity at the expense of species with unique traits. Ecology and Evolution 6:7610-7622. https://doi.org/10.1002/ece3.2389

Cooke, R. S. C., F. Eigenbrod, and A. E. Bates. 2019. Projected losses of global mammal and bird ecological strategies. Nature Communications 10:2279. https://doi.org/10.1038/s41467-019-10284-z

Crins, W. J., P. A. Gray, P. W. Uhlig, and M. C. Wester. 2009. The ecosystems of Ontario. Part 1: ecozones and ecoregions. Government of Ontario, Peterborough, Ontario, Canada.

Culp, L. A., E. B. Cohen, A. L. Scarpignato, W. E. Thogmartin, and P. P. Marra. 2017. Full annual cycle climate change vulnerability assessment for migratory birds. Ecosphere 8:e01565. https://doi.org/10.1002/ecs2.1565

Danz, N. P., G. J. Niemi, R. R. Regal, T. Hollenhorst, L. B. Johnson, J. A. M. Hanowski, R. P. Axler, J. J. H. Ciborowski, T. Hrabik, V. J. Brady, J. R. Kelly, J. A. Morrice, J. C. Brazner, R. W. Howe, C. A. Johnston, and G. E. Host. 2007. Integrated measures of anthropogenic stress in the U.S. Great Lakes basin. Environmental Management 39:631-647. https://doi.org/10.1007/s00267-005-0293-0

Danz, N. P., R. R. Regal, G. J. Niemi, V. J. Brady, T. Hollenhorst, L. B. Johnson, G. E. Host, J. M. Hanowski, C. A. Johnston, T. Brown, J. Kingston, and J. R. Kelly. 2005. Environmentally stratified sampling design for the development of Great Lakes environmental indicators. Environmental Monitoring and Assessment 102:41-65. https://doi.org/10.1007/s10661-005-1594-8

De Coster, G., C. Banks-Leite, and J. P. Metzger. 2015. Atlantic forest bird communities provide different but not fewer functions after habitat loss. Proceedings of the Royal Society B: Biological Sciences 282:1-8. https://doi.org/10.1098/rspb.2014.2844

Devictor, V., D. Mouillot, C. Meynard, F. Jiguet, W. Thuiller, and N. Mouquet. 2010. Spatial mismatch and congruence between taxonomic, phylogenetic and functional diversity: the need for integrative conservation strategies in a changing world. Ecology Letters 13:1030-1040. https://doi.org/10.1111/j.1461-0248.2010.01493.x

Ehlers Smith, D. A., X. Si, Y. C. Ehlers Smith, and C. T. Downs. 2018. Seasonal variation in avian diversity and tolerance by migratory forest specialists of the patch-isolation gradient across a fragmented forest system. Biodiversity and Conservation 27:3707-3727. https://doi.org/10.1007/s10531-018-1622-y

Endenburg, S., G. W. Mitchell, P. Kirby, L. Fahrig, J. Pasher, and S. Wilson. 2019. The homogenizing influence of agriculture on forest bird communities at landscape scales. Landscape Ecology 34:2385-2399. https://doi.org/10.1007/s10980-019-00895-8

Environmental Systems Research Institute (ESRI). 2011. ArcGIS desktop. Environmental Systems Research Institute, Redlands, California, USA.

Ford, M. T., and D. J. Flaspohler. 2010. Scale-dependent response by breeding songbirds to residential development along Lake Superior. Wilson Journal of Ornithology 122:296-306. https://doi.org/10.1676/09-137.1

Freemark, K., and B. Collins. 1999. Landscape ecology of birds breeding in temperate forest fragments. NCASI Technical Bulletin :179.

Gibson, L., T. M. Lee, L. P. Koh, B. W. Brook, T. A. Gardner, J. Barlow, C. A. Peres, C. J. A. Bradshaw, W. F. Laurance, T. E. Lovejoy, and N. S. Sodhi. 2011. Primary forests are irreplaceable for sustaining tropical biodiversity. Nature 478:378-381. https://doi.org/10.1038/nature10425

Gnass Giese, E. E., R. W. Howe, A. T. Wolf, N. A. Miller, and N. G. Walton. 2015. Sensitivity of breeding birds to the “human footprint” in western Great Lakes forest landscapes. Ecosphere 6:1-22. https://doi.org/10.1890/ES14-00414.1

Grand, J., S. P. Saunders, N. L. Michel, L. Elliott, S. Beilke, A. Bracey, T. M. Gehring, E. E. Gnass Giese, R. W. Howe, B. Kasberg, N. Miller, G. J. Niemi, C. J. Norment, D. C. Tozer, J. Wu, and C. Wilsey. 2020. Prioritizing coastal wetlands for marsh bird conservation in the U.S. Great Lakes. Biological Conservation 249:108708. https://doi.org/10.1016/j.biocon.2020.108708

Grinde, A. R., G. J. Niemi, B. R. Sturtevant, H. Panci, W. Thogmartin, and P. Wolter. 2017. Importance of scale, land cover, and weather on the abundance of bird species in a managed forest. Forest Ecology and Management 405:295-308. https://doi.org/10.1016/j.foreco.2017.09.057

Guiry, E. J., M. Buckley, T. J. Orchard, A. L. Hawkins, S. Needs-Howarth, E. Holm, and P. Szpak. 2020. Deforestation caused abrupt shift in Great Lakes nitrogen cycle. Limnology and Oceanography 65:1921-1935. https://doi.org/10.1002/lno.11428

Hollenhorst, T. P., T. N. Brown, L. B. Johnson, J. J. H. Ciborowski, and G. E. Host. 2007. Methods for generating multi-scale watershed delineations for indicator development in Great Lake coastal ecosystems. Journal of Great Lakes Research 33:13-26. https://doi.org/10.3394/0380-1330(2007)33[13:MFGMWD]2.0.CO;2

Hollenhorst, T. P., L. B. Johnson, and J. Ciborowski. 2011. Monitoring land cover change in the Lake Superior basin. Aquatic Ecosystem Health & Management 14:433-442. https://doi.org/10.1080/14634988.2011.628242

Homer, C. G., J. Dewitz, J. Fry, M. Coan, N. Hossain, C. Larson, N. Herold, A. J. McKerrow, J. N. VanDriel, and J. Wickham. 2007. Completion of the 2001 National Land Cover Database for the conterminous United States. Photogrammetric Engineering and Remote Sensing 73:337-341.

Homer, C., J. Dewitz, L. Yang, S. Jin, P. Danielson, G. Xian, J. Coulston, N. Herold, J. Wickham, and K. Megown. 2015. Completion of the 2011 National Land Cover Database for the conterminous United States - representing a decade of land cover change information. Photogrammetric Engineering and Remote Sensing 81:345-354.

Host, G. E., T. N. Brown, T. P. Hollenhorst, L. B. Johnson, and J. J. H. Ciborowski. 2011. High-resolution assessment and visualization of environmental stressors in the Lake Superior basin. Aquatic Ecosystem Health & Management 14:376-385. https://doi.org/10.1080/14634988.2011.625340

Host, G. E., K. E. Kovalenko, T. N. Brown, J. J. H. Ciborowski, and L. B. Johnson. 2019. Risk-based classification and interactive map of watersheds contributing anthropogenic stress to Laurentian Great Lakes coastal ecosystems. Journal of Great Lakes Research 45:609-618. https://doi.org/10.1016/j.jglr.2019.03.008

Howe, R. W., R. R. Regal, J. A. Hanowski, G. J. Niemi, N. P. Danz, and C. R. Smith. 2007. An index of ecological condition based on bird assemblages in Great Lakes coastal wetlands. Journal of Great Lakes Research 33:93-105. https://doi.org/10.3394/0380-1330(2007)33[93:AIOECB]2.0.CO;2

Hupperts, S. F., Y. L. Dickinson, C. R. Webster, and C. C. Kern. 2019. Promoting structural and species diversity in Great Lakes northern hardwoods: a conceptual model and its application. Forestry 92:16-25. https://doi.org/10.1093/forestry/cpy026

Jacoboski, L. I., and S. M. Hartz. 2020. Using functional diversity and taxonomic diversity to assess effects of afforestation of grassland on bird communities. Perspectives in Ecology and Conservation 18:103-108. https://doi.org/10.1016/j.pecon.2020.04.001

Johnson, L. B., K. E. Kovalenko, G. E. Host, V. J. Brady, A. M. Bracey, and J. J. H. Ciborowski. 2015. Great Lakes environmental indicators testing and refinement: final report to US EPA GLNPO Project Identifier: EPAGLNPO-2010-NS-5-1071-795.

Jonason, D., J. Ekroos, E. Öckinger, J. Helenius, M. Kuussaari, J. Tiainen, H. G. Smith, and R. Lindborg. 2017. Weak functional response to agricultural landscape homogenisation among plants, butterflies and birds. Ecography 40:1221-1230. https://doi.org/10.1111/ecog.02268

Laliberte, E., and P. Legendre. 2010. A distance-based framework for measuring functional diversity from multiple traits. Ecology 91:299-305. https://doi.org/10.1890/08-2244.1

Laureto, L. M. O., M. V. Cianciaruso, and D. S. M. Samia. 2015. Functional diversity: an overview of its history and applicability. Natureza & Conservação 13:112-116. https://doi.org/10.1016/j.ncon.2015.11.001

Lesak, A. A., V. C. Radeloff, T. J. Hawbaker, A. M. Pidgeon, T. Gobakken, and K. Contrucci. 2011. Modeling forest songbird species richness using LiDAR-derived measures of forest structure. Remote Sensing of Environment 115:2823-2835. https://doi.org/10.1016/j.rse.2011.01.025

Luck, G. W., A. Carter, and L. Smallbone. 2013. Changes in bird functional diversity across multiple land uses: interpretations of functional redundancy depend on functional group identity. PLoS ONE 8. https://doi.org/10.1371/journal.pone.0063671

Luck, G. W., S. Lavorel, S. Mcintyre, and K. Lumb. 2012. Improving the application of vertebrate trait-based frameworks to the study of ecosystem services. Journal of Animal Ecology 81:1065-1076. https://doi.org/10.1111/j.1365-2656.2012.01974.x

Mac Nally, R., and C. J. Walsh. 2004. Hierarchical partitioning public-domain software. Biodiversity & Conservation 13:659-660. https://doi.org/10.1023/B:BIOC.0000009515.11717.0b

Mason, N. W. H., and D. Mouillot. 2013. Functional diversity measures. Encyclopedia of biodiversity. Second edition. Elsevier Ltd. https://doi.org/10.1016/B978-0-12-384719-5.00356-7

Mason, N. W. H., D. Mouillot, W. G. Lee, and J. B. Wilson. 2005. Functional richness, functional evenness and functional divergence: the primary components of functional diversity. Oikos 111:112-118. https://doi.org/10.1111/j.0030-1299.2005.13886.x

Mattsson, B. J., J. H. Devries, J. A. Dubovsky, D. Semmens, W. E. Thogmartin, J. J. Derbridge, and L. Lopez-Hoffman. 2020. Linking landscape-scale conservation to regional and continental outcomes for a migratory species. Scientific Reports 10:4968. https://doi.org/10.1038/s41598-020-61058-3

Matuoka, M. A., M. Benchimol, J. M. de Almeida-Rocha, and J. C. Morante-Filho. 2020. Effects of anthropogenic disturbances on bird functional diversity: a global meta-analysis. Ecological Indicators 116:106471. https://doi.org/10.1016/j.ecolind.2020.106471

Miller, C., G. J. Niemi, J. M. Hanowski, and R. R. Regal. 2007. Breeding bird communities across an upland disturbance gradient in the western Lake Superior region. Journal of Great Lakes Research 33:305-318. https://doi.org/10.3394/0380-1330(2007)33[305:BBCAAU]2.0.CO;2

Mouchet, M. A., S. Villéger, N. W. H. Mason, and D. Mouillot. 2010. Functional diversity measures: an overview of their redundancy and their ability to discriminate community assembly rules. Functional Ecology 24:867-876. https://doi.org/10.1111/j.1365-2435.2010.01695.x

Murray, F., A. Douglas, and M. Solan. 2014. Species that share traits do not necessarily form distinct and universally applicable functional effect groups. Marine Ecology Progress Series 516:23-34. https://doi.org/10.3354/meps11020

Nebel, S., A. Mills, J. D. McCracken, and P. D. Taylor. 2010. Declines of aerial insectivores in North America follow a geographic gradient. Avian Conservation and Ecology 5(2):1. https://doi.org/10.5751/ACE-00391-050201

Neumann, B., A. T. Vafeidis, J. Zimmermann, and R. J. Nicholls. 2015. Future coastal population growth and exposure to sea-level rise and coastal flooding – a global assessment. PLoS ONE 10:131375. https://doi.org/10.1371/journal.pone.0118571

Niemi, G. J., V. J. Brady, T. N. Brown, J. J. H. Ciborowski, N. P. Danz, D. M. Ghioca, J. A. M. Hanowski, T. P. Hollenhorst, R. W. Howe, L. B. Johnson, C. A. Johnston, and E. D. Reavie. 2009. Development of ecological indicators for the U.S. Great Lakes coastal region - a summary of applications in Lake Huron. Aquatic Ecosystem Health & Management 12:77-89. https://doi.org/10.1080/14634980802711943

Niemi, G. J., J. A. M. Hanowski, N. Danz, R. Howe, M. Jones, J. Lind, and D. M. Mladenoff. 2004. Hierarchical scales in landscape responses by forest birds. ASTM Special Technical Publication:56-68. https://doi.org/10.1520/STP11942S

Niemi, G. J., R. W. Howe, B. R. Sturtevant, L. R. Parker, A. R. Grinde, N. P. Danz, M. D. Nelson, E. J. Zlonis, N. G. Walton, E. E. G. Giese, and S. M. Lietz. 2016. Analysis of long-term forest bird monitoring data from National Forests of the western Great Lakes Region. Gen. Tech. Rep. NRS-159. U.S. Forest Service, Newtown Square, Pennsylvania, USA. https://doi.org/10.2737/NRS-GTR-159

Niemi, G. J., J. R. Kelly, and N. P. Danz. 2007. Environmental indicators for the coastal region of the North American Great Lakes: introduction and prospectus. Journal of Great Lakes Research 33:1-12. https://doi.org/10.3394/0380-1330(2007)33[1:EIFTCR]2.0.CO;2

O’Connell, T. J., L. E. Jackson, and R. P. Brooks. 2000. Bird guilds as indicators of ecological condition in the central Appalachians. Ecological Applications 10:1706-1721. https://doi.org/10.1890/1051-0761(2000)010[1706:BGAIOE]2.0.CO;2

Oksanen, J. 2017. Vegan: ecological diversity. R Package Version 2.4-4:11.

Ontario Land Cover (OLC). 2002. Ontario land cover data. Ontario Ministry of Natural Resources, Toronto, Ontario, Canada.

R Core Team. 2019. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.

Rayner, L., K. Ikin, M. J. Evans, P. Gibbons, D. B. Lindenmayer, and A. D. Manning. 2015. Avifauna and urban encroachment in time and space. Diversity and Distributions 21:428-440. https://doi.org/10.1111/ddi.12293

Remeš, V., E. Remešová, N. R. Friedman, B. Matysioková, and L. Rubáčová. 2021. Functional diversity of avian communities increases with canopy height: from individual behavior to continental-scale patterns. Ecology and Evolution 11:11839-11851. https://doi.org/10.1002/ece3.7952

Robinson, S. K., F. R. T. Ill, T. M. Donovan, D. R. Whitehead, and J. Faaborg. 1995. Regional forest fragmentation and the nesting success of migratory birds. Science 267:1987-1990. https://doi.org/10.1126/science.267.5206.1987

Rosenberg, A. K. V, A. M. Dokter, P. J. Blancher, J. R. Sauer, A. C. Smith, P. A. Smith, J. C. Stanton, A. Panjabi, L. Helft, and P. P. Marra. 2019. Decline of the North American avifauna. Science 366:120-124. https://doi.org/10.1126/science.aaw1313

Rosenberg, K. V., J. A. Kennedy, R. Dettmers, R. P. Ford, D. Reynolds, J. D. Alexander, C. J. Beardmore, P. J. Blancher, R. E. Bogart, G. S. Butcher, A. F. Camfield, A. Couturier, D. W. Demarest, W. E. Easton, J. J. Giocomo, R. H. Keller, A. E. Mini, A. O. Panjabi, and T. Will. 2016. Partners in Flight landbird conservation plan. Partners in Flight Science Committee. Revision f:1-119.

Rurangwa, M. L., J. Aguirre-Gutiérrez, T. J. Matthews, P. Niyigaba, J. P. Wayman, J. A. Tobias, and R. J. Whittaker. 2021. Effects of land-use change on avian taxonomic, functional and phylogenetic diversity in a tropical montane rainforest. Diversity and Distributions 27:1732-1746. https://doi.org/10.1111/ddi.13364

Rushing, C. S., J. Andrew Royle, D. J. Ziolkowski, and K. L. Pardieck. 2020. Migratory behavior and winter geography drive differential range shifts of eastern birds in response to recent climate change. Proceedings of the National Academy of Sciences of the United States of America 117:12897-12903. https://doi.org/10.1073/pnas.2000299117

Schulte, L. A., D. J. Mladenoff, T. R. Crow, L. C. Merrick, and D. T. Cleland. 2007. Homogenization of northern U.S. Great Lakes forests due to land use. Landscape Ecology 22:1089-1103. https://doi.org/10.1007/s10980-007-9095-5

Schulze, E. D., D. Craven, A. M. Durso, J. Reif, M. Guderle, F. Kroiher, P. Hennig, A. Weiserbs, P. Schall, C. Ammer, and N. Eisenhauer. 2019. Positive association between forest management, environmental change, and forest bird abundance. Forest Ecosystems 6:3. https://doi.org/10.1186/s40663-019-0160-8

Sekercioglu, C. H. 2006. Increasing awareness of avian ecological function. Trends in Ecology and Evolution 21:464-471. https://doi.org/10.1016/j.tree.2006.05.007

Soulliere, G. J., M. A. Al-Saffar, K. R. VanBeek, C. M. Tonra, M. D. Nelson, D. N. Ewert, T. Will, W. E. Thogmartin, K. E. O’Brien, S. W. Kendrick, A. M. Gillet, J. R. Herkert, E. E. Gnass Giese, M. P. Ward, and S. Graff. 2020. Upper Mississippi/Great Lakes Joint Venture landbird habitat conservation strategy - 2020 revision.

Tozer, D. C. 2016. Marsh bird occupancy dynamics, trends, and conservation in the southern Great Lakes basin: 1996 to 2013. Journal of Great Lakes Research 42:136-145. https://doi.org/10.1016/j.jglr.2015.10.015

Villéger, S., J. R. Miranda, D. F. Hernández, and D. Mouillot. 2010. Contrasting changes in taxonomic vs. functional diversity of tropical fish communities after habitat degradation. Ecological Applications 20:1512-1522. https://doi.org/10.1890/09-1310.1

Wang, Y., U. Naumann, D. Eddelbuettel, J. Wilshire, and D. Warton. 2020. mvabund: statistical methods for analysing multivariate abundance data.

Weisberg, P. J., T. E. Dilts, M. E. Becker, J. S. Young, D. C. Wong-Kone, W. E. Newton, and E. M. Ammon. 2014. Guild-specific responses of avian species richness to LiDAR-derived habitat heterogeneity. Acta Oecologica 59:72-83. https://doi.org/10.1016/j.actao.2014.06.002

Whelan, C. J., D. G. Wenny, and R. J. Marquis. 2008. Ecosystem services provided by birds. Annals of the New York Academy of Sciences 1134:25-60. https://doi.org/10.1196/annals.1439.003

Wilman, H., J. Belmaker, J. Simpson, C. de la Rosa, M. M. Rivadeneira, and W. Jetz. 2014. Elton Traits 1.0: species-level foraging attributes of the world’s birds and mammals. Ecology 95:2027. https://doi.org/10.1890/13-1917.1

Wolter, P. T., C. A. Johnston, and G. J. Niemi. 2006. Land use land cover change in the U.S. Great Lakes basin 1992 to 2001. Journal of Great Lakes Research 32:607-628. https://doi.org/10.3394/0380-1330(2006)32[607:LULCCI]2.0.CO;2

Zhao, Q., E. T. Azeria, M. L. Le Blanc, J. Lemaître, and D. Fortin. 2013. Landscape-scale disturbances modified bird community. PLoS ONE 8. https://doi.org/10.1371/journal.pone.0081358

Zuckerberg, B., and W. F. Porter. 2010. Thresholds in the long-term responses of breeding birds to forest cover and fragmentation. Biological Conservation 143:952-962. https://doi.org/10.1016/j.biocon.2010.01.004

Zurell, D., C. H. Graham, L. Gallien, W. Thuiller, and N. E. Zimmermann. 2018. Long-distance migratory birds threatened by multiple independent risks from global change. Nature Climate Change 8:992-996. https://doi.org/10.1038/s41558-018-0312-9