A species’ range center is hypothesized to provide optimal habitat conditions (Brown 1984), while at range edges habitat may be less suitable (Lesica and Allendorf 1995), population dynamics more variable and populations less viable (Hoffmann and Blows 1994, Linder et al. 2000). In nature, peripheral populations may have lower reproductive success (Wright et al. 2007, Hollander et al. 2011, Golawski et al. 2016), greater reproductive success (Pidgeon et al. 2001, Hargrove and Rotenberry 2011), or reproductive success that does not differ from core populations (Barrientos et al. 2009). Differences in competitive ability, plasticity, adaptation, or exogenous factors related to local habitat patches, may be associated with demographic rates more strongly than proximity to the core range or large scale macroecological patterns in traits, e.g., latitudinal trends in clutch size, or number of breeding attempts (Ruskin et al. 2017). As populations exceed carrying capacity, birds may occupy habitat of lower suitability (Fretwell and Lucas 1969, Fretwell 1972, Hartman 1996, Wright et al. 2007) or similar quality habitat that was previously unoccupied (Hartman 1996). Often it is young or socially subordinate males that first occupy habitat (Van Horne 1983, Holmes et al. 1996, Weinberg and Roth 1998, Braillet et al. 2002).
The Kirtland’s Warbler (Setophaga kirtlandii) historically nested in young (~5–20 years old) dense jack pine (Pinus banksiana) forests in sandy glacial outwash areas in Lower Michigan, USA (Bocetti et al. 2020). Prior to Euro-American settlement, nesting habitat was maintained by stand-replacing wildfire (Donner et al. 2008) with approximately a 60-year return interval (Cleland et al. 2004). Natural regeneration of breeding habitat virtually ended after Euro-American settlement because of wildfire suppression and conversion to residential and agricultural uses, resulting in a severe decline in Kirtland’s Warbler population size over time (Mayfield 1972, 1983, Probst and Weinrich 1993, Bocetti 1994). The species was included on the endangered species list in 1966 (Federal Register 1967). The primary management strategy was establishment of jack pine plantations of high stem density, with small openings, mimicking naturally regenerated habitat (Kepler et al. 1996). Together with control of Brown-headed Cowbirds (Molothrus ater; Shake and Mattsson 1975, Kelly and DeCapita 1982, Cooper et al. 2019), this resulted in population recovery and the species was removed from the Federal List of Endangered and Threatened Wildlife (Federal Register 2019).
Kirtland’s Warblers expanded into Michigan’s Upper Peninsula in 1994 (Probst et al. 2003, Donner et al. 2008), and into Ontario, Canada and Wisconsin in 2007 (Trick et al. 2008, Donner et al. 2009, Richard 2014). Annually increasing population size coupled with a stable amount of breeding habitat in Lower Michigan likely fueled this expansion (Probst et al. 2003, Donner et al. 2008, 2009). Unlike Kirtland’s Warblers in Lower Michigan, which breed in young, dense jack pine habitat, Kirtland’s Warblers in Renfrew County, Ontario and Adams County, Wisconsin, breed in red pine (Pinus resinosa) plantations that include lesser components of naturally regenerated jack pine and oak (Quercus spp.; Trick et al. 2008, Anich et al. 2011, Richard 2014). Although Kirtland’s Warblers were previously observed breeding in red pine plantations occasionally (Mayfield 1960, Orr 1975, Probst 1986), it is unclear whether red pine plantations support viable populations of the species or are marginal habitat supporting low demographic rates. Data on nesting success and habitat use are needed to determine the long-term viability of populations using red pine plantations.
Habitat selection is a hierarchical process influencing fitness of individuals (Hutto 1985, Chalfoun and Martin 2007, Luepold et al. 2015) at increasingly fine resolution (Battin and Lawler 2006, Guttery et al. 2017), from the geographic range to the habitat patch, territory, and the nest site (Johnson 1980). Habitat use is an outcome of selection (Jones 2001). At the habitat patch scale, Kirtland’s Warbler habitat use may be a product of selection for tree species composition (Mayfield 1953, Walkinshaw 1983, Probst 1988), landscape context, tree density, and ground vegetation composition. Kirtland’s Warblers in Adams County, Wisconsin may have selected red pine-dominated plantations because of their structural similarity to jack pine, and low availability of suitably aged jack pine habitat. Proximity to nearby occupied patches and the presence of conspecifics likely influenced selection at the patch level (Mayfield 1953, 1960, Bocetti 1994, Donner et al. 2009, Anich and Ward 2017).
At the territory scale, habitat use is influenced by factors including habitat structure (Luepold et al. 2015), adult predation risk (Lima 2009), food availability (Chalfoun and Martin 2007), presence and breeding success of conspecifics (Danchin et al. 1998, Lima 2009, Anich and Ward 2017), territory quality (own or conspecifics; Danchin et al. 1998, Hoover 2003), and previous breeding success (Hoover 2003, Howlett and Stutchbury 2003, Lima 2009). Kirtland’s Warblers forage more often in jack pine trees than on the ground, in deciduous trees, snags, or woody debris (Fussman 1997). Males establish breeding territories (Mayfield 1960, Walkinshaw 1983), and males select areas with relatively high pine density, often in the vicinity of conspecifics (Anich and Ward 2017).
Nest site selection occurs at yet a finer scale. Females should select nest sites that increase fitness (Lima 2009), and nest site selection is impacted by many factors, including thermal conditions (Nelson and Martin 1999, Hoekman et al. 2002, Davis 2005, Warren and Anderson 2005, Fisher and Davis 2010), visibility or openness (Götmark et al. 1995), nest depredation risk (Davis 2005, Luepold et al. 2015), brood parasitism risk (Forsman and Martin 2009), and post-fledging habitat needs (Fisher and Davis 2011). Understanding factors that influence nest success is important for understanding avian population dynamics (Hoekman et al. 2002).
Recently established Kirtland’s Warbler populations provided an opportunity to determine how habitat use and reproductive success differ in the range core versus the range periphery. Further, because young jack pine habitat has been considered a requirement for breeding, understanding reproductive success and habitat use in red pine plantations may provide insights into the species’ niche flexibility that can guide management.
Our overarching goal was to characterize Kirtland’s Warbler breeding productivity and habitat use in red pine plantations at a set of nested scales. We quantified nest success and identified habitat features associated with territories and nest sites. We expected Kirtland’s Warbler nest success would increase with nest concealment and decrease as the breeding season advanced because of increased efficiency by predators (Grant et al. 2005, Grant and Shaffer 2012). We expected that males would select territories with dense low vegetation, and abundant blueberries because such vegetation provides potential nest sites and foraging opportunities, and blueberries may be important food late in the breeding season (Fussman 1997, Deloria 2000, Deloria-Sheffield et al. 2001). Finally, we expected red pine plantations used by Kirtland’s Warblers to be structurally similar to jack pine habitat they use in the range core.
Our study area was in Adams County, Wisconsin (Fig. 1). This landscape is a glacial outwash plain with well drained, sandy soil and a continental climate (Wisconsin Department of Natural Resources 2015). Vegetation includes jack pine, scrub oak, barrens, and jack and red pine plantations, forested and non-forested wetland, grassland, and agriculture (Wisconsin Department of Natural Resources 2015).
Our study area was ~418 km west of the core Michigan Kirtland’s Warbler breeding area and consisted of red pine plantations in which jack pine was a naturally regenerating component. Kirtland’s Warbler breeding attempts were documented in Adams County beginning in 2007 by nest monitors employed through Wisconsin Department of Natural Resources (WDNR; Trick et al. 2008, 2009, Grveles 2009, Anich et al. 2011). During our study the plantations were under a conservation easement with the WDNR (Michigan Department of Natural Resources et al. 2015). Little management had occurred in the plantations. Kirtland’s Warblers occupied 10 red pine plantations between 2015 and 2017 that were planted between 2005 and 2009 and were 2.7 ha to 46.9 ha (mean = 24.8 ha), with a combined area of 248.1 ha. Eight of the 10 plantations were clustered together, separated by roads (~6–27 m wide) or 75–480 m of non-suitably aged red pine plantations (Fig. 1). Two additional occupied plantations were ~8.4 and ~15.4 km northeast of the cluster of eight. The occupied plantations were embedded within a matrix of red and jack pine plantations of various ages, with minor components of natural forest, wetland, agriculture, pastures, and private residences. Brown-headed Cowbirds were trapped near occupied plantations from approximately 10 April to 10 July each year and humanely euthanized, by U.S. Department of Agriculture Animal and Plant Health Inspection Service personnel.
Most males were individually identifiable by colored plastic leg bands (Refsnider et al. 2009, Trick et al. 2009, Anich et al. 2011). Male age was determined at the time of banding (Pyle 1997, Probst et al. 2007), and individual return histories were known from yearly resightings (Fig. A1.1).
Beginning in mid-May (2015–2017), we observed singing males, recorded the locations of singing perches using handheld GPS units (95% typical use GPS accuracy < 10 m), and noted behaviors indicating they were paired. We recorded an average of 2.4 singing locations per individual approximately every four days, starting when males were first observed and ending upon completion of nesting, for an average of 17 singing locations per male (min = 3, max = 37, SD = 7.75). We considered singing locations to represent males’ territories. We classified males as paired if they were associating with a female, singing with food, singing muffled songs, or if we observed copulation.
Nest monitoring was conducted under guidelines of WDNR, U.S. Fish and Wildlife Service (USFWS), and the Kirtland’s Warbler Recovery Team. Because of the species’ endangered status at the time, the guidelines required that nests be observed from a distance (Olah 2019). Once a male was classified as paired, we attempted to locate the nest. After locating the nest, we put flagging tape ~10 m away. We observed nests from the flag, approximately every two days, observing only long enough to determine whether nests were active or until 30 minutes elapsed. In the rare event that adults seemed agitated by our presence, we left and observed from farther away on the next visit. If we found nests accidentally, we noted nest contents at that time. If nests were found by observing male behaviors only and not approached, we inferred nest stage based on adult behaviors. We inferred that a nest contained eggs if we infrequently saw the male approach the nest with food and rarely observed the female off the nest (Bocetti et al. 2020). We inferred eggs had hatched if both adults were frequently visiting the nest with food (Bocetti et al. 2020). We inferred hatch date based on the shift from infrequent to frequent adult activity at the nest and estimated an expected fledge date (hatch day + average nestling stage length [9.4 days]; Bocetti et al. 2020). If after two 30-minute observations on sequential days, we did not observe activity near the nest during the nestling stage we assumed the nest had failed and approached to look for evidence of the cause of failure. We approached nests after they became inactive to look for unhatched eggs or evidence of depredation. To estimate nest success and contents for nests that were not approached while active, we used information such as adults feeding fledglings or the presence of unhatched eggs. We considered nests successful if at least one Kirtland’s Warbler fledged. Nest failure was assigned to one of four potential causes: depredation (clear evidence of a predator at a nest), abandonment (cold eggs, dead nestlings, no observed adult activity in two days), parasitism (nests contained only cowbird eggs or nestlings), or unknown. If we approached a nest and observed it had been parasitized, we removed the Brown-headed Cowbird nestlings, per guidelines.
We defined productivity as the number of Kirtland’s Warbler young that fledged from nests. Because of restrictions placed on nest approaches, we did not know the number of nestlings in a subset of nests, for which we determined the possible minimum and maximum number fledged. For example, if we did not approach a nest prior to chicks fledging but we observed three fledglings with the adults and found one unhatched egg in the nest we concluded that a minimum of three young fledged (three fledglings observed) and a maximum of four young fledged (usual clutch size is five, but one egg remained unhatched thus five could not have fledged). Thus, we estimated high and low averages of young fledged per year, assuming that the true number fell within this range.
We collected vegetation data in 2015–2017, 3–14 days after a nest attempt ended, at nests and at an equal number of points located in a random direction and distance of between 1–55 m (mean = 26.2 m, min = 3.3 m, max = 55 m) from the nest (hereafter, random points). All nests were within the associated male’s territory boundaries (as defined by a minimum convex polygon; see territory vegetation measurements section, below), and all but two random points fell within the territory boundary. We characterized live trees and shrubs within 10 m of sampling points using point-centered-quarter (PCQ) methodology (Cottam and Curtis 1956, Warde and Petranka 1981). In each quadrant we recorded the height and distance to the nearest shrub, the nearest tree 1–3 m tall, and the nearest tree > 3 m tall. For trees, we measured the height from the ground of the lowest live branch (Buech 1980). Using our PCQ method, nests and random points within 20 m of each other could include trees or shrubs that were sampled twice, but post-hoc we found that only 3 of 210 trees and 2 of 105 shrubs were sampled twice. Our protocol reflected our expectation that trees of the two height classes may provide different resources, such that shorter trees may have lower live branches offering greater nest concealment (per Mayfield 1960) and taller trees may offer more attractive song perches.
We estimated percent ground cover (bare ground, moss, lichen, litter, woody debris, grass, sedge (Carex spp.), blueberry shrubs, shrubs, trees) in 1x1 m plots centered on the nest or random point in three categories (0%, < 50%, > 50%) to maximize detection of differences. We measured vegetation-height density (Robel et al. 1970) at two heights (1 m, 0.25 m) that were one and four meters from each point, in each cardinal direction. For each combination of height and distance, we averaged the four measures, resulting in four estimates of vegetation-height density. We measured dead herbaceous vegetation at three spots centered on the nest or for random points, at the angles of an equilateral triangle with sides of 10 cm (outer diameter of Kirtland’s Warbler nests) and averaged these values. We estimated nest concealment from above by placing a cardboard disc 5.8 cm in diameter (interior nest diameter) in the nest. We estimated the percent of the disc obscured by vegetation from eye level (~1.5 m) from four equidistant locations, each 0.5 m from the nest, and averaged the estimates. We did not estimate concealment from above at random points.
We recorded singing locations of males on 56 territories (min = 3, max = 37, mean = 17 locations per territory, SD = 7.75) between 2015 and 2017. We used asymptote tests (Laver and Kelly 2008) to determine whether the number of singing locations was sufficient for calculating territory boundaries. Specifically, for each individual we first randomly selected n singing locations and from them created 90% minimum convex polygons, repeating 100 times for each 1 unit increase in number of singing locations from n = 4 to n = maximum number of singing locations, in the “adehabitatHR” package in R (Calenge 2006). Then we created 95% confidence intervals around the mean of the 100 bootstrapped estimates of territory size at each number of singing locations (Fig. A2.1). If the 90% minimum convex polygon based on all singing locations was within the 95% confidence interval, we considered the territory well defined. We had too few singing locations to do this for five territories. Using this assessment, 46 were well defined (13 from 2015, 16 from 2016, 17 from 2017).
To assess habitat within and outside of territories we collected vegetation data in 2016 and 2017. Of the 38 territories delineated in 2016 and 2017, six were not well defined using the above criteria so we excluded those data. From the remaining 33, we randomly selected three of each male’s singing locations at which to compare vegetation with vegetation collected at random points located outside of all territory 90% minimum convex polygons; no random points fell within any territory boundaries as defined. This occurred after nesting ended to avoid disturbing nests. Vegetation sampling began in early July and ended in early August. We averaged data for each male so that the unit of analysis was the territory (hereafter, territory points), and averaged data collected at random points outside of territories. We characterized live trees and shrubs at sampling points using the PCQ methodology. We estimated tree density and height by species across territory points and across random points.
We categorized percent vegetation cover in four 1 m² plots located 3 m from a sampling point in each cardinal direction, using similar cover categories and the same percent categories used for nest site vegetation sampling. We used a majority rule to consolidate information from the 12 plots: if more than half had the same percent cover class for a ground cover category then we assigned the majority percent cover value, and if the plots were evenly split between percent cover values, we assigned the middle cover value (< 50%). Vegetation-height density was also measured as is described for nest site vegetation sampling.
We defined a patch as an area of contiguous pine plantation of uniform age. We calculated tree densities and frequencies within each occupied patch by combining tree data collected at Kirtland’s Warbler locations and random points within the same patch. We calculated tree density in patches using PCQ methodology. We defined trees as those ≥ 2.5 cm in diameter at 50 cm height, and shrubs as woody plants ≥ 50 cm tall with multiple stems at or above 10 cm above ground. We estimated relative frequency, relative density, and average tree height, overall and by tree species.
We calculated daily survival rates using logistic exposure (Shaffer 2004). We modeled daily survival for all years combined, and separately for 2016 and 2017 but not for 2015 because only one nest failed that year. All analyses were conducted in R version 3.6.2 (R Core Team 2019). We created generalized linear (binomial) models with the following link function:
The link function included exposure, where t = the number of days between successive observations. We excluded known re-nesting attempts (n = 4) from survival models because we could not account for male identity in models because our small dataset necessitated limiting model size. We used unstandardized variables in models so that we could make interpretable predictions in the units of the variables (Greenland et al. 1986, Luskin 1991, Grace and Bollen 2005, Menard 2011). We created a global generalized linear model using the R package MASS (Venables and Ripley 2002) and ranked all possible subsets of variable combinations by AICc using the package MuMIn (Bartoń 2018). We excluded collinear variables (Spearman’s rank order correlation ≥ 0.7), here and in all subsequent analyses. In the global logistic exposure model, we included these predictors: concealment from above, vegetation-height density 1.25, and season day (the number of days after the first nest was found each season). We considered models with AICc differences ≤ 2 (∆AICc) from the minimum AICc model to be in the confidence set (Burnham and Anderson 2004). To compare models, we calculated Akaike weights (wi) and evidence ratios (Burnham and Anderson 2004). Variable importance was calculated over the confidence set and the entire model set (Burnham and Anderson 2004). As a measure of model fit, we calculated generalized R² values (Nakagawa and Schielzeth 2013) with the MuMIn package (Bartoń 2018), and area under the receiver operator curve (AUC) using the modEva package (Barbosa et al. 2016). Within individual models in the confidence set we considered variables to be significant if P ≤ 0.1, following Arnold (2010). We assumed that variables with higher variable importance values and that were significant in individual models would be most useful in distinguishing between nest success and failure. To generate period survival estimates from the null model, we raised the estimated daily survival rate to the power of the period length (14 days for incubation through hatching, 24 days for incubation through fledging), such that the predictor function was:
We calculated percent of nests parasitized annually. We used a binomial logistic regression model to assess whether proximity to cowbird traps influenced the likelihood of nest parasitism.
We compared vegetation at 38 nests and 35 random locations, using generalized linear models with logit-link functions,
following the modeling procedure described for nest survival, above, combining data from 2016 and 2017. The response variable had a binary outcome of “nest” or “random location.” The global model included the following predictor variables, none of which were strongly correlated: vegetation-height density 1.25, vegetation-height density 4.1, depth of dead vegetation on the ground, blueberry cover, pine branch cover, and distance to nearest trees. We chose global model variables based on our expectation that nests would be better concealed from predators, brood parasites, or the elements where vegetation cover and density were greater, and trees were closer together than random points.
As described for our nest survival models, we considered models ≤ 2 AICc from the minimum AICc model to be in the confidence set (Burnham and Anderson 2004), and used Akaike weights, evidence ratios, variable importance values, R² values, and AUC to compare models. We examined the effect of individual predictor variables on the probability of a point being a nest or not, while holding other model variables constant (Shaffer and Thompson 2007).
We compared vegetation at 32 territory points and 26 random points (outside of territories) using generalized linear models with logit-link functions (Eqn. 3) using the modeling procedure described above. The response variable had a binary outcome of “territory” or “non-territory.” The global model included the following variables, none of which were strongly correlated: vegetation-height density 1.1, vegetation-height density 1.25, vegetation-height density 4.1, vegetation-height density 4.25, blueberry shrub cover, pine branch cover, bare ground cover, shrub distance, and tree distance. Our choice of variables was based on the expectation that territories would include dense, low vegetative cover, and abundant blueberries, a potentially important food source. Although we initially expected shrub cover to be an important factor differentiating territories from non-territories, we excluded it because all but one point had 0% shrub cover.
As described above, we considered models ≤ 2 AICc from the minimum AICc model to be in the confidence set and we assessed territory models in the same way as nest survival and nest site models. We examined the effect of individual predictor variables on the probability of a point being a territory point or not, while holding other model variables constant (Shaffer and Thompson 2007).
We assessed whether males used specific tree species in two ways using chi-square tests. First, we compared the number of territory and random points in which the nearest tree was red pine, jack pine, or oak. Second, we compared the relative frequency of tree species present on all territory points with the relative frequency of tree species at all random points.
We found the first nest on 2 June in 2015 and 2016, and 5 June in 2017. The last fledging date was 24 July in 2015, 12 July in 2016, and 27 July in 2017. In 2015, 51 young fledged from 12 nests; in 2016, 22–23 young fledged from 7 nests; and in 2017, 39–48 young fledged from 11 nests (Fig. 2). Six nests were never approached while active: 4 in 2015, and 2 in 2016. We considered the 4 nests in 2015 to be successful based on adult behaviors indicating the presence of fledglings. Both nests that were not approached in 2016 became inactive during the nestling stage, and subsequent inspection revealed they had failed.
In 2015, only one of the 13 nests failed (Fig. 3) and we did not model nest survival for that year. In 2016, 7 of the 16 nests were successful; failure was attributed to depredation and abandonment. Of the 2016 nests, four were parasitized by Brown-headed Cowbirds. The cowbird chicks (n = 4, one per nest) were removed. The relationship between cowbird removal and ultimate nest failure cause was not clear. In 2017, failures (4 nests) were attributed to depredation and brood parasitism (Fig. 3).
In nest survival models, concealment from above had the highest variable importance in 2016 and in the combined year model (VI = 0.28 and 0.48 respectively; Table 1), however the nature of this relationship varied and was not significant at α = 0.1 in any model. In the 2017 model set, season day had the greatest variable importance value (VI = 0.52; Table 1), exhibiting a slight negative relationship with survival that was not significant at α = 0.1. We predicted daily survival rates and period survival rate estimates (through hatching and through fledging) for 2016, 2017, and the combined-year data. We used the null models because they were within the confidence set in each case (Table 2). The estimated daily survival rate was 0.95 (95% CI = 0.90, 0.98) in 2016, 0.93 (95% CI = 0.84, 0.98) in 2017, and 0.97 (95% CI = 0.94, 0.98) across all years. Our period survival estimates through hatching were 0.48 (95% CI = 0.21, 0.71) in 2016, 0.39 (95% CI = 0.09, 0.70) in 2017, and 0.62 (95% CI = 0.43, 0.76) across all three years. Our period survival estimates through fledging were 0.28 (95% CI = 0.07, 0.55) in 2016, 0.20 (95% CI = 0.02, 0.55) in 2017, and 0.43 (95% CI = 0.23, 0.62) for all three years.
The average number of young produced per nest was between 2.55 and 2.77 (all years combined; Fig. 2). In 2015, one nest was parasitized by Brown-headed Cowbirds, the cowbird nestling was removed, and Kirtland’s Warbler young fledged. In 2016, cowbirds parasitized six nests, the cowbird nestlings were removed, and Kirtland’s Warbler young fledged from two of the nests. In 2017, two nests contained only cowbird eggs or nestlings, which were removed. An additional nest in 2017 containing Kirtland’s Warbler and cowbird eggs was depredated.
The brood parasitism rate was 22.7% (2015 = 7.7%, 2016 = 37.5%, 2017 = 20%). Nests were on average within 1.28 km of a cowbird trap (min = 0.20 km, max = 7.19 km, SD = 1.6 km). Parasitized nests were on average 0.45 km further from cowbird traps than unparasitized nests, however distance to nearest trap was not significant in binomial regression (P = 0.38). Brood parasitism rates in Michigan ranged from 0% to 1.6% between 2015 and 2018, even with reduced cowbird management (Cooper et al. 2019).
Red pine was the nearest tree to 63% of nests (n = 27), Quercus spp. was closest to 21% of nests (n = 9), jack pine was closest to 9% of nests (n = 4), and other tree species were closest to 7% of nests (n = 3). Nests were in live grasses (31%, n = 13), blueberry (19%, n = 8), red pine duff (17%, n = 7), low live red pine branches (10%, n = 4), live sedge (10%, n = 4), dead sedge (7%, n = 3), other dead vegetation (5%, n = 2), and dead grass (2%, n = 1). Dead herbaceous vegetation depth around the perimeter of nests varied from 5 cm to 15.73 cm (5–10 cm deep at 40% of nests, 10–13 cm deep at 48% of nests, and 14–15.73 cm deep at 12% of nests).
Nest sites had greater dead herbaceous vegetation depth, vegetation-height density 1.25, and tree cover relative to random points within territories. In the confidence set of nest site models (Table 3), dead herbaceous vegetation depth, vegetation-height density 1.25, and tree branch cover had equal variable importance values (VI values = 0.48; Table 1). Across all nest site models, dead herbaceous vegetation depth had greatest variable importance value (VI value = 1.0), followed by vegetation-height density 1.25 (VI value = 0.96), and tree branch cover (VI value = 0.87; Table 1). As vegetation-height density 1.25 increased from 0 cm to 20 cm, the probability of nest placement increased from 0.03 to 1.0. As dead herbaceous vegetation depth increased from 0 cm to 15.73 cm the probability of nest placement increased from < 0.001 to 1.0. The predicted probability of nest placement increased from 0.23 when tree branch cover was 0%, to 0.97 when tree branch cover was < 50%, to 1.0 when tree branch cover was > 50%.
Of males observed in Adams County from 2008 to 2019, 58% (min = 25%, max = 80%) were after second-year males (hatched at least two years previous), 30% (min = 13%, max = 47%) were second-year males (hatched the previous year), and 12% (min = 0%, max = 38%) were of unknown age (Fig. A1.1).
Average density of trees within territories (1577 trees ha-1) was greater than outside of territories in the same patch (1094 trees ha-1; Fig. 4). Red pine occurred at greatest density both in territories (1292 trees ha-1, n = 331) and outside of territories (885 trees ha-1, n = 284), followed by jack pine (territories = 265 trees ha-1, n = 68; outside of territories = 203 trees ha-1, n = 65; Fig. 4). The relative frequencies of red pine and jack pine were similar in and outside of territories (territories: red pine = 66%, jack pine = 32%; outside of territories: red pine = 67%, jack pine = 32%; Fig. 4).
In predicting whether points were within or outside of territories, distance to the nearest tree (which characterizes tree density of the entire territory [AO, AP, personal observation]) discriminated most strongly (variable importance [VI] value = 0.92); the shorter the distance, the greater the probability a point was within a territory. Vegetation height density 1.25 (VI value = 0.80), and vegetation-height density 1.1 (VI = 0.70; Table 1) were also important discriminators (Table 4). The predicted probability of a point being within a territory increased from 0.38 to 1.0 when vegetation-height density 1.25 increased from 0 cm to 150 cm while other variables were held constant at their mean. The probability of a point being within a territory decreased from 0.83 to 0.04 when vegetation-height density 1.1 increased from 0 cm to 133 cm while other variables were held at their mean. The probability of a point being within a territory decreased from 0.89 to 0.05 when average distance to the nearest tree increased from 1.8 m to 5.4 m while other variables were held at their mean.
Between territory and random points, we found no difference in the proportion of points in which the nearest tree was red pine, jack pine, or oak (chi-square test, χ = 0.4, P = 0.8). Red pine was the closest tree at 77–80% of points, jack pine was the closest tree at 19–21% of points, and oak was the closest tree at 1–2% of points.
On average, occupied patches consisted of 70% red pine, 28% jack pine, and 2% oak. Average tree height in occupied patches was 4.1 m (SD = 1.1 m, n = 2552 trees) and was identical for red and jack pine, n = 2138 and 391 respectively), while the average height of oaks was 2.8 m (SD = 1.5, n = 13). Low live branch height of red pines was closer to the ground (mean = 8.7 cm, SD = 11.6, n = 2138) than jack pine (mean = 51.4 cm, SD = 46.0, n = 391). Mean tree density in occupied patches was 1937 trees ha-1 (SD = 912 trees ha-1, n = 8 patches).
In a population of Kirtland’s Warblers breeding at the range periphery in red pine-dominated plantations, a relatively little used habitat type, nesting success was similar to that in long-occupied jack pine habitat used by the core population. Both nest survival (WI = 0.43, all years combined; MI = 0.32, Bocetti et al. 2020) and average number of young fledged per nest (Wisconsin, 2.5–2.8, 2015–2017; Michigan, 2.8–3.6; Shake and Mattsson 1975, Kelly and DeCapita 1982, Bocetti 1994, Rockwell et al. 2012) were comparable in the peripheral and core populations. This suggests that nesting success is not reduced in red pine-dominated plantations.
Although Kirtland’s Warbler nests were closer to red pines than to other tree species, nests were placed directly in grass, blueberry, or red pine duff and rarely directly under low live tree branches. In Michigan, nests were typically placed in open areas of low vegetation within one meter of a pine thicket (Bocetti et al. 2020), with vegetation directly covering nests mainly composed of grasses, sedges, and blueberry (Mayfield 1960). Nest sites had deeper dead herbaceous vegetation, greater vegetation height-density, and greater cover of low tree branches within 0.5 m of the nest than non-nest points, suggesting that concealment of nests by vegetation, even if not touching the nest, is important. Perhaps this is because Kirtland’s Warblers use surrounding vegetation to conceal their nest approaches (Mayfield 1960). Tall dense vegetation may offer nest concealment from predators or shelter from the elements (Martin and Roper 1988, Nelson and Martin 1999, Davis 2005, Warren and Anderson 2005, Fisher and Davis 2010). Dead vegetation is also likely a primary source of nest cover early in the season before full green up (Mayfield 1960, Davis 2005, Warren and Anderson 2005, Fisher and Davis 2010). Many Kirtland’s Warbler nests are built before ground vegetation has fully emerged; thus females may select sites with high proportions of dead vegetation because of benefits (concealment, thermal etc.) those features may confer. We collected vegetation data after a nesting effort ended regardless of nest fate. This delayed measurement may have resulted in measuring conditions other than what the birds experienced at nest initiation, leading to spurious conclusions (McConnell et al. 2017), although we note that vegetative growth rates slowed considerably as the summer progressed, mitigating differences in time of use versus time of measurement.
We found that occupied red pine-dominated plantations in Adams County had similar proportions of pine (~90% pine; in MI up to 20% broadleaf; Smith 1979, Probst 1988) but lower densities of trees (1860 trees ha-1) than jack pine plantations used by Kirtland’s Warblers in Lower Michigan (2000–3345 jack pine ha-1; Bocetti 1994, Houseman and Anderson 2002). Additionally, live red pine branches were nearer to the ground than live jack pine branches in our study site, which could explain why Kirtland’s Warblers would occupy lower density red pine plantations.
Within occupied patches, territories had slightly greater tree density than random locations (~480 trees ha-1 greater). Relative frequencies of red and jack pine were not different between territories and random locations. If males were using jack pine in greater proportion than was present in occupied patches, territories should have a higher proportion of jack pine than locations outside of territories. Point level measures of distance to trees (our proxy of point-level tree density) indicates that trees within territories are closer together than trees outside of territories. On territories, vegetation-height density measured at 0.25 m height was greater than in areas outside of territories but within the same patch, suggesting that males establish territories in areas with greater low vegetation cover, possibly because it offers better foraging opportunities, more potential nest sites, or lower detection by predators. However, we also found that territories had lower vegetation-height density measured at 1 m height than outside of territories. It is less clear why this negative association exists with density of tall vegetation, but we speculate that dense tall grass may be more difficult for Kirtland’s Warblers to fly through. Although we expected blueberry cover to be an important habitat feature, we did not find that to be the case. Although blueberry cover may be important for foraging opportunities or nest concealment, it is possible that at scales larger than the nest site Kirtland’s Warblers may accept a wide range of ground cover compositions (Probst 1988, Probst and DonnerWright 2003). It is also possible that blueberry is unimportant on territories because fruiting does not occur until territories begin to break down as the nesting period ends.
The population of Kirtland’s Warblers in Adams County is likely still growing. All suitable territories in occupied plantations may not be filled, nor all suitable habitat patches occupied. The contrast in habitat attributes between territory and non-territory points may be subtle if non-territory points are also suitable but unoccupied (Fretwell and Lucas 1969, Fretwell 1972, Hartman 1996). Conspecific attraction also influences habitat occupancy in Kirtland’s Warblers (Mayfield 1960, Bocetti 1994, Anich and Ward 2017), and likely enhances suitability of occupied patches through (positive) Allee effects (Allee et al. 1949). If red pine-dominated habitat in Adams County was of marginal quality, we would expect most new males observed each year to be second year birds or birds that were subordinate. We did not find evidence of this. Most newly observed males were after-second-year birds, suggesting that red pine plantations are perceived to be of similar quality to jack pine plantations in the core range. However, we cannot rule out the possibility that newly arrived after-second-year birds in Adams County were subordinate to territory holders in the core habitat, as metrics associated with dominance were not measured.
Demographic rates are often hypothesized to be greatest at a species’ geographic range center, coinciding with peak species abundance, and lowest at the range periphery. This pattern is found in multiple species (Wright et al. 2007, Hollander et al. 2011, Golawski et al. 2016). However, demographic rates are not always greater at the core of a species’ range. Local habitat quality may more strongly influence fecundity across geographic ranges than large scale environmental gradients (Pidgeon et al. 2001, Hargrove and Rotenberry 2011, Ruskin et al. 2017). Or there may be no differences in reproductive success between populations across a geographic range (Barrientos et al. 2009). Kirtland’s Warblers align with this latter group, in that nest survival rates and number of young fledged per nest in peripheral habitat and the core of the breeding range are similar.
In response to habitat creation and Brown-headed Cowbird control, the Kirtland’s Warbler population has recovered and is no longer listed as federally endangered (Federal Register 2019). A goal stated in the Kirtland’s Warbler Conservation Plan is to provide suitable breeding habitat in areas peripheral to the core range for 10% (100 pairs) of the Kirtland’s Warbler population (Michigan Department of Natural Resources et al. 2015). Kirtland’s Warblers will always remain conservation-reliant because wildfire suppression prevents natural habitat creation, and cowbird parasitism depresses their productivity. The challenge in the post-delisting period is to continue creating suitable breeding habitat and managing Brown-headed Cowbirds. Although Kirtland’s Warblers in the core range appear to be resilient to relaxed levels of cowbird control (Cooper et al. 2019), the small population in Adams County suffers from high levels of brood parasitism even with active cowbird management and a declining Brown-headed Cowbird population in Wisconsin (-2.67% over the period 2007–2017, North American Breeding Bird Survey, Patuxent Wildlife Research Center-Bird Population Studies, https://www.mbr-pwrc.usgs.gov/). The cowbird population trend is comparable to that in Michigan over this same 10-year period (-3.06%, ibid). Low parasitism rates in Michigan are attributed to low cowbird population size, associated with reforestation (Cooper et al. 2019); however, in Adams County WI, agriculture is prevalent, which could account for the differences in nest parasitism rates in the two states.
Managers will likely need to minimize costs associated with creating Kirtland’s Warbler habitat, and managers in Wisconsin will likely need to continue managing cowbirds. One option for increasing timber value is inter-planting red and jack pine trees in plantations, a strategy included in the most recent Kirtland’s Warbler Conservation Plan (Michigan Department of Natural Resources et al. 2015).
Our findings, together with others from red pine-dominated plantations (Anich et al. 2011, Richard 2014), suggest that suitable Kirtland’s Warbler breeding habitat can consist of various combinations of red pine and jack pine in a managed plantation. When the Kirtland’s Warbler population was lowest, management focused on habitat known to sustain the population: young jack pines intermixed with grassy openings. Now that the species is no longer threatened with extinction, managers have more flexibility to plant a range of tree species composition proportions that Kirtland’s Warblers will use as breeding habitat, which may facilitate lower costs of managing habitat for this conservation-reliant species.
All authors conceptualized the project, designed the research, and participated in writing and editing. AO collected and formally analyzed the data, and wrote the original draft. AP and CR participated in analysis of the data, reviewing and editing, and funding acquisition for the project. KG, DL, and SW assisted with project logistics.
This work was supported by the National Science Foundation Graduate Research Fellowship (Grants. DGE-1256259 and DGE-1747503), and the National Institute of Food and Agriculture, USDA (grant number WIS01590). Funding was provided to A.O. by the American Ornithologists Union, the Association of Field Ornithologists, and the Wisconsin Society for Ornithology. No funder had input into the content of the manuscript. We thank the Natural Resources Foundation of Wisconsin for assistance. Thank you, J. Cummings, K. Reinstma, L. White, N. Anich, L. Serna, Z. Osberg, and A. Jocham for help with data collection. Thank you, N. Livingston, for being a long-time friend of Kirtland’s Warblers and providing housing. We thank C. Bocetti, D. Donner, R. White, for comments on previous manuscript versions. Use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government. Findings and conclusions in this article are those of the authors and do not necessarily represent the views of USFWS. All activities were approved under the following permits or protocols: USFWS Endangered Species Recovery Permit TE99059B-2, WDNR Endangered and Threatened Species Permit #1052, USGS Federal Bird Banding permit #23398, and University of Wisconsin - Madison Animal Care and Use Protocol A005409.
Allee, W. C., A. E. Emerson, O. Park, T. Park, and K. P. Schmidt. 1949. Principles of animal ecology. Saunders, Philadelphia, Pennsylvania, USA.
Anich, N. M., J. A. Trick, K. M. Grveles, and J. L. Goyette. 2011. Characteristics of a red pine plantation occupied by Kirtland’s Warblers in Wisconsin. Wilson Journal of Ornithology 123:199-205. https://doi.org/10.1676/10-057.1
Anich, N. M., and M. P. Ward. 2017. Using audio playback to expand the geographic breeding range of an endangered species. Diversity and Distributions 23:1499-1508. https://doi.org/10.1111/ddi.12635
Arnold, T. W. 2010. Uninformative parameters and model selection using Akaike’s Information Criterion. Journal of Wildlife Management 74:1175-1178. https://doi.org/10.2193/2009-367
Barbosa, A. M., J. A. Brown, A. Jiménez-Valverde, and R. Real. 2016. modEva: Model evaluation and analysis. [online] URL: http://hdl.handle.net/10174/20946
Barrientos, R., A. Barbosa, F. Valera, and E. Moreno. 2009. Breeding parameters of the Trumpeter Finch at the periphery of its range: a case study with mainland expanding and island populations. Journal of Arid Environments 73:1177-1180. https://doi.org/10.1016/j.jaridenv.2009.06.001
Bartoń, K. 2018. MuMIn: multi-model inference. [online] URL: https://cran.r-project.org/web/packages/MuMIn/index.html
Battin, J., and J. J. Lawler. 2006. Cross-scale correlations and the design and analysis of avian habitat selection studies. Condor 108:59-70. https://doi.org/10.1093/condor/108.1.59
Bocetti, C. I. 1994. Density, demography, and mating success of Kirtland’s Warblers in managed and natural habitats. Dissertation. Ohio State University, Columbus, Ohio, USA. [online] URL: http://rave.ohiolink.edu/etdc/view?acc_num=osu1487856076416827
Bocetti, C. I., D. M. Donner, and H. F. Mayfield. 2020. Kirtland’s Warbler (Setophaga kirtlandii), version 1.0. In A. F. Poole, editor. Birds of the World. Cornell Lab of Ornithology, Ithaca, New York, USA. https://doi.org/10.2173/bow.kirwar.01
Braillet, C., A. Charmantier, F. Archaux, A. Dos Santos, P. Perret, and M. M. Lambrechts. 2002. Two Blue Tit Parus caeruleus populations from Corsica differ in social dominance. Journal of Avian Biology 33:446-450. https://doi.org/10.1034/j.1600-048X.2002.02956.x
Brown, J. H. 1984. On the relationship between abundance and distribution of species. American Naturalist 124:255-279. https://doi.org/10.1086/284267
Buech, R. R. 1980. Vegetation of a Kirtland’s Warbler breeding area and 10 nest sites. Jack-Pine Warbler 58:58-72.
Burnham, K. P., and D. R. Anderson. 2004. Multimodel inference: understanding AIC and BIC in model selection. Sociological Methods and Research 33:261-304. https://doi.org/10.1177/0049124104268644
Calenge, C. 2006. The package “adehabitat” for the R software: a tool for the analysis of space and habitat use by animals. Ecological Modelling 197:516-519. https://doi.org/10.1016/j.ecolmodel.2006.03.017
Chalfoun, A. D., and T. E. Martin. 2007. Assessments of habitat preferences and quality depend on spatial scale and metrics of fitness. Journal of Applied Ecology 44:983-992. https://doi.org/10.1111/j.1365-2664.2007.01352.x
Cleland, D. T., T. R. Crow, S. C. Saunders, D. I. Dickmann, A. L. MacLean, J. K. Jordan, R. L. Watson, A. M. Sloan, and K. D. Brosofske. 2004. Characterizing historical and modern fire regimes in Michigan (USA): a landscape ecosystem approach. Landscape Ecology 19:311-325. https://doi.org/10.1023/B:LAND.0000030437.29258.3c
Cooper, N. W., C. S. Rushing, and P. P. Marra. 2019. Reducing the conservation reliance of the endangered Kirtland’s Warbler through adaptive management. Journal of Wildlife Management 83:1297-1305. https://doi.org/10.1002/jwmg.21714
Cottam, G., and J. T. Curtis. 1956. The use of distance measures in phytosociological sampling. Ecology 37:451-460. https://doi.org/10.2307/1930167
Danchin, E., T. Boulinier, and M. Massot. 1998. Conspecific reproductive success and breeding habitat selection: implications for the study of coloniality. Ecology 79:2415-2428. https://doi.org/10.1890/0012-9658(1998)079[2415:CRSABH]2.0.CO;2
Davis, S. K. 2005. Nest-site selection patterns and the influence of vegetation on nest survival of mixed-grass prairie passerines. Condor 107:605-616. https://doi.org/10.1093/condor/107.3.605
Deloria, C. M. 2000. Kirtland’s Warbler (Dendroica kirtlandii) diet and its relationship to warbler age, sex, and jack pine stand characteristics. Thesis. Michigan State University, East Lansing, Michigan, USA. https://doi.org/doi:10.25335/M5TT4G389
Deloria-Sheffield, C. M., K. F. Millenbah, C. I. Bocetti, P. W. Sykes Jr, and C. B. Kepler. 2001. Kirtland’s Warbler diet as determined through fecal analysis. Wilson Bulletin 113:384-387. https://doi.org/10.1676/0043-5643(2001)113[0384:KSWDAD]2.0.CO;2
Donner, D. M., J. R. Probst, and C. A. Ribic. 2008. Influence of habitat amount, arrangement, and use on population trend estimates of male Kirtland’s Warblers. Landscape Ecology 23:467-480. https://doi.org/10.1007/s10980-008-9208-9
Donner, D. M., C. A. Ribic, and J. R. Probst. 2009. Male Kirtland’s Warblers’ patch-level response to landscape structure during periods of varying population size and habitat amounts. Forest Ecology and Management 258:1093-1101. https://doi.org/10.1016/j.foreco.2009.05.029
Federal Register. 1967. Notices; Office of the Secretary; Native Fish and Wildlife; Endangered Species, 32 Fed. Reg. 3961, Vol. 48, No. 6. 32, pg. 4001 (Saturday March 11, 1967). [online] URL: https://www.loc.gov/item/fr032048/
Federal Register. 2019. Endangered and Threatened Wildlife and Plants; Removing the Kirtland’s Warbler From the Federal List of Endangered and Threatened Wildlife. Vol. 84, No. 196. 50 CFR 17. [online] URL: https://www.federalregister.gov/documents/2019/10/09/2019-22096/endangered-and-threatened-wildlife-and-plants-removing-the-kirtlands-warbler-from-the-federal-list
Fisher, R. J., and S. K. Davis. 2010. From Wiens to Robel: a review of grassland-bird habitat selection. Journal of Wildlife Management 74:265-273. https://doi.org/10.2193/2009-020
Fisher, R. J., and S. K. Davis. 2011. Post-fledging dispersal, habitat use, and survival of Sprague’s Pipits: are planted grasslands a good substitute for native? Biological Conservation 144:263-271. https://doi.org/10.1016/j.biocon.2010.08.024
Forsman, J. T., and T. E. Martin. 2009. Habitat selection for parasite-free space by hosts of parasitic cowbirds. Oikos 118:464-470. https://doi.org/10.1111/j.1600-0706.2008.17000.x
Fretwell, S. D. 1972. Populations in a seasonal environment. Princeton University Press, Princeton, New Jersey, USA.
Fretwell, S. D., and H. L. Lucas. 1969. On territorial behavior and other factors influencing habitat distribution in birds. I. Theoretical development. Acta Biotheoretica 19:16-36. https://doi.org/10.1007/BF01601953
Fussman, J. L. 1997. Foraging ecology of Kirtland’s Warblers in managed and natural breeding habitat. Thesis. Ohio State University, Columbus, Ohio, USA. [online]: http://rave.ohiolink.edu/etdc/view?acc_num=osu1260989687
Golawski, A., E. Mróz, and Z. Kasprzykowski. 2016. Breeding performance of the White-winged Tern (Chlidonias leucopterus) in two habitat types at the edge of their distribution range. Ornis Fennica 93:225-234.
Götmark, F., D. Blomqvist, O. C. Johansson, and J. Bergkvist. 1995. Nest site selection: a trade-off between concealment and view of the surroundings? Journal of Avian Biology 26:305-312. https://doi.org/10.2307/3677045
Grace, J. B., and K. A. Bollen. 2005. Interpreting the results from multiple regression and structural equation models. Bulletin of the Ecological Society of America 86:283-295. https://doi.org/10.1890/0012-9623(2005)86[283:ITRFMR]2.0.CO;2
Grant, T. A., and T. L. Shaffer. 2012. Time-specific patterns of nest survival for ducks and passerines breeding in North Dakota. Auk 129:319-328. https://doi.org/10.1525/auk.2012.11064
Grant, T. A., T. L. Shaffer, E. M. Madden, and P. J. Pietz. 2005. Time-specific variation in passerine nest survival: new insights into old questions. Auk 122:661-672. https://doi.org/10.1093/auk/122.2.661
Greenland, S., J. J. Schesselman, and M. H. Criqui. 1986. The fallacy of employing standardized regression coefficients and correlations as measures of effect. American Journal of Epidemiology 123:203-208. https://doi.org/10.1093/oxfordjournals.aje.a114229
Grveles, K. M. 2009. The first annual census of the Kirtland’s Warbler (Dendroica kirtlandii) in Wisconsin. Passenger Pigeon 71:123-130.
Guttery, M. R., C. A. Ribic, D. W. Sample, A. Paulios, C. Trosen, J. Dadisman, D. Schneider, and J. A. Horton. 2017. Scale-specific habitat relationships influence patch occupancy: defining neighborhoods to optimize the effectiveness of landscape-scale grassland bird conservation. Landscape Ecology 32:515-529. https://doi.org/10.1007/s10980-016-0462-y
Hargrove, L., and J. T. Rotenberry. 2011. Breeding success at the range margin of a desert species: implications for a climate-induced elevational shift. Oikos 120:1568-1576. https://doi.org/10.1111/j.1600-0706.2011.19284.x
Hartman, G. 1996. Habitat selection by European beaver (Castor fiber) colonizing a boreal landscape. Journal of Zoology 240:317-325. https://doi.org/10.1111/j.1469-7998.1996.tb05288.x
Hoekman, S. T., I. J. Ball, and T. E. Fondell. 2002. Grassland birds orient nests relative to nearby vegetation. Wilson Bulletin 114:450-456. https://doi.org/10.1676/0043-5643(2002)114[0450:GBONRT]2.0.CO;2
Hoffmann, A. A., and M. W. Blows. 1994. Species borders: ecological and evolutionary perspectives. Trends in Ecology & Evolution 9:223-227. https://doi.org/10.1016/0169-5347(94)90248-8
Hollander, F. A., H. van Dyck, G. San Martin, and N. Titeux. 2011. Maladaptive habitat selection of a migratory passerine bird in a human-modified landscape. PLoS ONE 6(9):e25703. https://doi.org/10.1371/journal.pone.0025703
Holmes, R. T., P. P. Marra, and T. W. Sherry. 1996. Habitat-specific demography of breeding Black-Throated Blue Warblers (Dendroica caerulescens): implications for population dynamics. Journal of Animal Ecology 65:183-195. https://doi.org/10.2307/5721
Hoover, J. P. 2003. Experiments and observations of Prothonotary Warblers indicate a lack of adaptive responses to brood parasitism. Animal Behaviour 65:935-944. https://doi.org/10.1006/anbe.2003.2154
Houseman, G. R., and R. C. Anderson. 2002. Effects of jack pine plantation management on barrens flora and potential Kirtland’s Warbler nest habitat. Restoration Ecology 10:27-36. https://doi.org/10.1046/j.1526-100X.2002.10103.x
Howlett, J. S., and B. J. M. Stutchbury. 2003. Determinants of between-season site, territory, and mate fidelity in Hooded Warblers (Wilsonia citrina). Auk 120(2):457–465.
Hutto, R. L. 1985. Habitat selection by nonbreeding, migratory land birds. Pages 455-476 in M. L. Cody, editor. Habitat selection in birds. Academic, New York, New York, USA.
Johnson, D. H. 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology 61:65-71. https://doi.org/10.2307/1937156
Jones, J. 2001. Habitat selection studies in avian ecology: a critical review. Auk 118:557-562. https://doi.org/10.1642/0004-8038(2001)118[0557:HSSIAE]2.0.CO;2
Kelly, S. T., and M. E. DeCapita. 1982. Cowbird control and its effect on Kirtland’s Warbler reproductive success. Wilson Bulletin 94:363-365. [online] URL: https://www.jstor.org/stable/4161651
Kepler, C. B., G. W. Irvine, M. E. DeCapita, and J. Weinrich. 1996. The conservation management of Kirtland’s Warbler Dendroica kirtlandii. Bird Conservation International 6:11-22. http://doi.org/10.1017/S0959270900001271
Laver, P. N., and M. J. Kelly. 2008. A critical review of home range studies. Journal of Wildlife Management 72:290-298. https://doi.org/10.2193/2005-589
Lesica, P., and F. W. Allendorf. 1995. When are peripheral populations valuable for conservation? Conservation Biology 9:753-760. https://doi.org/10.1046/j.1523-1739.1995.09040753.x
Lima, S. L. 2009. Predators and the breeding bird: behavioral and reproductive flexibility under the risk of predation. Biological Reviews 84:485-513. https://doi.org/10.1111/j.1469-185X.2009.00085.x
Linder, E. T., M.-A. Villard, B. A. Maurer, and E. V. Schmidt. 2000. Geographic range structure in North American landbirds: variation with migratory strategy, trophic level, and breeding habitat. Ecography 23:678-686. https://doi.org/10.1111/j.1600-0587.2000.tb00311.x
Luepold, S. H. B., T. P. Hodgman, S. A. McNulty, J. Cohen, and C. R. Foss. 2015. Habitat selection, nest survival, and nest predators of Rusty Blackbirds in northern New England, USA. Condor 117:609-623. https://doi.org/10.1650/CONDOR-14-215.1
Luskin, R. C. 1991. Abusus Non Tollit Usum: standardized coefficients, correlations, and R2s. American Journal of Political Science 35:1032-1046. https://doi.org/10.2307/2111505
Martin, T. E., and J. J. Roper. 1988. Nest predation and nest-site selection of a western population of the Hermit Thrush. Condor 90:51-57. https://doi.org/10.2307/1368432
Mayfield, H. 1953. A census of the Kirtland’s Warbler. Auk 70:17-20. https://doi.org/10.2307/4081053
Mayfield, H. 1960. The Kirtland’s Warbler. Cranbrook Institute of Science, Bloomfield Hills, Michigan, USA.
Mayfield, H. F. 1972. Third decennial census of Kirtland’s Warbler. Auk 89:263-268. https://doi.org/10.2307/4084205
McConnell, M. D., A. P. Monroe, L. W. Burger, and J. A. Martin. 2017. Timing of nest vegetation measurement may obscure adaptive significance of nest-site characteristics: a simulation study. Ecology and Evolution 7:1259-1270. https://doi.org/10.1002/ece3.2767
Menard, S. 2011. Standards for standardized logistic regression coefficients. Social Forces 89:1409-1428. https://doi.org/10.1093/sf/89.4.1409
Michigan Department of Natural Resources, U.S. Fish and Wildlife Service, and U.S. Forest Service. 2015. Kirtland’s Warbler breeding range conservation plan. Michigan Department of Natural Resources, Lansing, Michigan, USA. [online] URL: https://www.michigan.gov/documents/dnr/Kirtlands_Warbler_CP_457727_7.pdf
Nakagawa, S., and H. Schielzeth. 2013. A general and simple method for obtaining R2 from generalized linear mixed-effects models. Methods in Ecology and Evolution 4:133-142. https://doi.org/10.1111/j.2041-210x.2012.00261.x
Nelson, K. J., and K. Martin. 1999. Thermal aspects of nest-site location for Vesper Sparrow and Horned Larks in British Columbia. Studies in Avian Biology 19:137-143.
Olah, A. M. 2019. Kirtland’s Warbler habitat selection, nest and fledging survival and activity patterns in red pine dominated breeding habitat. Thesis. University of Wisconsin, Madison, Wisconsin, USA.
Orr, C. D. 1975. 1974 breeding success of the Kirtland’s Warbler. Jack-Pine Warbler 53:58-66.
Pidgeon, A. M., N. E. Mathews, R. Benoit, and E. V. Nordheim. 2001. Response of avian communities to historic habitat change in the northern Chihuahuan Desert. Conservation Biology 15:1772-1789. https://doi.org/10.1046/j.1523-1739.2001.00073.x
Probst, J. R. 1986. A review of factors limiting the Kirtland’s Warbler on its breeding grounds. American Midland Naturalist 116:87-100. https://doi.org/10.2307/2425940
Probst, J. R. 1988. Kirtland’s Warbler breeding biology and habitat management. Pages 28-35 in T. W. Hoekstra and J. Capp, editors. Integrating forest management for wildlife and fish. General Technical Report NC-122, U.S. Forest Service, North Central Forest Experiment Station, St. Paul, Minnesota, USA.
Probst, J. R., and D. DonnerWright. 2003. Fire and shade effects on ground cover structure in Kirtland’s Warbler habitat. American Midland Naturalist 149:320-334. https://doi.org/10.1674/0003-0031(2003)149[0320:FASEOG]2.0.CO;2
Probst, J. R., D. M. Donner, C. I. Bocetti, and S. Sjogren. 2003. Population increase in Kirtland’s Warbler and summer range expansion to Wisconsin and Michigan’s Upper Peninsula, USA. Oryx 37:365-373. https://doi.org/10.1017/S0030605303000632
Probst, J. R., D. M. Donner, and M. A. Bozek. 2007. Continuous, age-related plumage variation in male Kirtland’s Warblers. Journal of Field Ornithology 78:100-108. https://doi.org/10.1111/j.1557-9263.2006.00091.x
Probst, J. R., and J. Weinrich. 1993. Relating Kirtland’s Warbler population to changing landscape composition and structure. Landscape Ecology 8:257-271. https://doi.org/10.1007/BF00125132
Pyle, P., 1997. Identification guide to North American Birds Part 1: Columbidae to Ploceidae. Slate Creek, Point Reyes Station, California, USA.
R Core Team 2019. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.
Refsnider, R. L., J. A. Trick, and J. L. Goyette. 2009. 2008 capture and banding of Kirtland’s Warblers (Dendroica kirtlandii) in Wisconsin. Passenger Pigeon 71:115-122.
Richard, T. 2014. Characteristics of the breeding habitat of endangered Kirtland’s Warblers on a Canadian military installation. Journal of Field Ornithology 85:301-309. https://doi.org/10.1111/jofo.12069
Robel, R. J., J. N. Briggs, A. D. Dayton, and L. C. Hulbert. 1970. Relationships between visual obstruction measurements and weight of grassland vegetation. Journal of Range Management 23:295-297. https://doi.org/10.2307/3896225
Rockwell, S. M., C. I. Bocetti, P. P. Marra. 2012. Carry-over effects of winter climate on spring arrival date and reproductive success in an endangered migratory bird, Kirtland’s Warbler (Setophaga kirtlandii). Auk 129:744-752. https://doi.org/10.1525/auk.2012.12003
Ruskin, K. J., M. A. Etterson, T. P. Hodgman, A. C. Borowske, J. B. Cohen, C. S. Elphick, C. R. Field, R. A. Kern, E. King, A. R. Kocek, A. I. Kovach, K. M. O’Brien, N. Pau, W. G. Shriver, J. Walsh, and B. J. Olsen. 2017. Seasonal fecundity is not related to geographic position across a species’ global range despite a central peak in abundance. Oecologia 183:291-301. https://doi.org/10.1007/s00442-016-3745-8
Shaffer, T. L. 2004. A unified approach to analyzing nest success. Auk 121:526-540. https://doi.org/10.1093/auk/121.2.526
Shaffer, T. L., and F. R. I. Thompson. 2007. Making meaningful estimates of nest survival with model-based methods. Studies in Avian Biology 34:84-95.
Shake, W. F., and J. P. Mattsson. 1975. Three years of cowbird control: an effort to save the Kirtland’s Warbler. Jack-Pine Warbler 53:48-53.
Smith, E. L. 1979. Analysis of Kirtland’s Warbler breeding habitat in Ogemaw and Roscommon Counties, Michigan. Thesis. Michigan State University, Lansing, Michigan, USA.
Trick, J. A., K. M. Grveles, D. DiTommaso, and J. Robaidek. 2008. The first Wisconsin nesting record of Kirtland’s Warbler (Dendroica kirtlandii). Passenger Pigeon 70:93-102.
Trick, J. A., K. M. Grveles, and J. L. Goyette. 2009. The 2008 nesting season: first documented successful nesting of Kirtland’s Warbler (Dendroica kirtlandii) in Wisconsin. Passenger Pigeon 71:101-114.
Van Horne, B. 1983. Density as a misleading indicator of habitat quality. Journal of Wildlife Management 47:893-901. https://doi.org/10.2307/3808148
Venables, W. N., and B. D. Ripley. 2002. Modern applied statistics with S. Fourth edition. Springer, New York, New York, USA. https://doi.org/ 10.1007/978-0-387-21706-2 https://doi.org/10.1007/978-0-387-21706-2
Walkinshaw, L. H. 1983. Kirtland’s Warbler: the natural history of an endangered species. Cranbrook Institute of Science, Bloomfield Hills, Michigan, USA.
Warde, W., and J. W. Petranka. 1981. A correction factor table for missing point-center quarter data. Ecology 62:491-494. https://doi.org/10.2307/1936723
Warren, K. A., and J. T. Anderson. 2005. Grassland songbird nest-site selection and response to mowing in West Virginia. Wildlife Society Bulletin 33:285-292. https://doi.org/10.2193/0091-7648(2005)33[285:GSNSAR]2.0.CO;2
Weinberg, H. J., and R. R. Roth. 1998. Forest and habitat quality for nesting Wood Thrushes. Auk 115:879-889. https://doi.org/10.2307/4089507
Wisconsin Department of Natural Resources. 2015. Central Sand Plains ecological landscape. Chapter 10 in The ecological landscapes of Wisconsin: an assessment of ecological resources and a guide to planning sustainable management. PUB-SS1131L 2015, Wisconsin Department of Natural Resources, Madison, Wisconsin, USA. [online] URL: https://dnr.wi.gov/topic/Landscapes/documents/1805Ch10.pdf
Wright, L. J., R. A. Hoblyn, W. J. Sutherland, and P. M. Dolman. 2007. Reproductive success of Woodlarks Lullula arborea in traditional and recently colonized habitats. Bird Study 54:315-323. https://doi.org/10.1080/00063650709461491