Northeastern Naturalist Vol. 21, No. 1
M.J. Glennon
2014
NENHC-51
2014 NORTHEASTERN NATURALIST 21(1):NENHC-51–NENHC-71
Dynamics of Boreal Birds at the Edge of Their Range in the
Adirondack Park, NY
Michale J. Glennon*
Abstract - The Adirondack Park in northern New York is located at the southern range
extent for several bird species that inhabit lowland boreal forest habitats, which in the
Adirondacks are naturally fragmented and intermixed with eastern temperate forest types.
I examined occupancy dynamics of 8 bird species in lowland boreal forest wetlands, evaluating
the influence of variables associated with climate change and habitat fragmentation,
including wetland size and connectivity, on colonization and extinction dynamics for the
period 2007–2011. Occurrence data from point-count surveys conformed to predictions
of metapopulation theory with respect to extinction, with most species more likely to experience
local extinction from smaller, more isolated wetlands. Responses to latitude and
elevation were variable. Proximity of human infrastructure was the most consistent driver
of short-term dynamics across species, with two-thirds more likely to colonize low-impact
sites and become locally extinct from more-impacted sites. Evidence for metapopulation
structure suggests that improved connectivity among wetlands and reduction of human impact
near wetlands should be conservation goals for these species in the park.
Introduction
The Adirondack Park in northern New York State represents the southern range
extent for several species of boreal forest birds within eastern North America.
These populations are subject to the stresses imposed by conditions at the periphery
of a species’ range, and they are also geographically isolated from conspecific
populations found further north. These birds are thought to be vulnerable to climate
change because they prefer northern boreal habitat types expected to be sensitive to
warming temperatures (Moore 2002, Niemi et al. 1998, Pastor et al. 1998). Habitats
in the Adirondacks are naturally fragmented and less continuous than the Canadian
boreal to the North, with patches of boreal wetland habitat surrounded by temperate
forest habitat types (Jenkins 2010). Additionally, habitats within the Adirondack
landscape are further fragmented by small amounts of agriculture and developed
land uses.
Little is known about the population status of these boreal specialists in this region
of the US. The North American Breeding Bird Survey (BBS; Sauer et al. 2012)
provides information on long-term trends in abundance of avian species, but trend
information is often rated as deficient in credibility for these peatland-associated
species. The roadside nature of BBS routes, combined with the rarity of species
and habitats such as these, often means that marginal populations of birds near the
edges of their ranges are not well sampled (Sauer et al. 1995). Mountain Birdwatch
*Wildlife Conservation Society, Adirondack Program, 132 Bloomingdale Avenue, Saranac
Lake, NY 12983; mglennon@wcs.org.
Manuscript Editor: Jeremy Kirchman
Northeastern Naturalist
NENHC-52
M.J. Glennon
2014 Vol. 21, No. 1
(Scarl 2012) is a land-bird monitoring program for montane species in the northeastern
US, but the targets of the program and the species for which published
trend information is available do not overlap with the low-elevation boreal species
described here. New York State Breeding Bird Atlas data (Andrle and Carroll 1988,
McGowan and Corwin 2008) provide the best information on the distribution of
low-elevation boreal species in New York, but are limited as a source of information
on trends. Comparisons between atlas surveys conducted 20 years apart reveal
declines in occupancy across the state for some boreal bird species, and increases
for others (McGowan and Corwin 2008). Because of these limitations, scientists
at the Wildlife Conservation Society, Adirondack Program (hereafter WCS), have
been monitoring a suite of 12 species in lowland boreal habitats of the Adirondacks,
including an intensive period of data collection between 2007 and 2011. In this
paper, I use occupancy data from this survey to explore short-term trends in and
potential drivers of boreal bird dynamics at their southern range extent in this part
of the northeast US.
Given the fragmentation of habitats for boreal forest birds in the northeast, one
might expect that these species follow the predictions of metapopulation theory,
which generally apply to any species inhabiting a patchy habitat or any spatially
structured population (Akçakaya et al. 2006, Hanski 1998). Sjögren-Gulve and
Hanski (2000) suggest that metapopulation models are best applied to systems of
relatively small habitat patches that are highly fragmented and cover maximally
some 20% of the landscape. Lowland boreal habitats in the Adirondacks cover approximately
11% of the landscape and are scattered throughout the 19,700-km2 park
(Jenkins 2010). According to metapopulation theory, long-term population dynamics
should be driven by size and connectivity of habitat patches. Therefore, boreal
birds should be more likely to colonize large, well-connected habitat patches and
to experience local extinction in smaller, more isolated patches (Hames et al. 2001,
Hanski 1998, Pulliam 1988).
In addition to habitat patch size and connectivity, habitat degredation from climate
change and anthropogenic development may also affect population dynamics
of boreal birds in this region. Because these species are on the edge of their range
and at the southern fringe of a northern habitat, climate change may be influencing
long-term population trends in the Adirondacks. As such, they may be moving
up in either latitude or elevation or both (Parmesan 2006). Several authors have
noted actual or predicted changes in the ranges of boreal bird species as a result
of climate change (Brommer 2004, La Sorte and Thompson 2007, Parmesan 2006,
Thomas and Lennon 1999, Zuckerberg et al. 2009). Zuckerberg et al. (2009) found
that southern range boundaries of New York birds shifted northward 11.4 km in
the time between the 1985 and 2005 atlases of breeding birds in New York State
(Andrle and Carroll 1988, McGowan and Corwin 2008). Though occurring over a
longer time interval than the dataset described here, such observed shifts in patterns
of bird occupancy suggest that range boundaries may be changing rapidly for
some species. Thomas and Lennon (1999) found that the northern range margins
of British birds shifted northwards by 18.9 km in a 20-year period, and similarly,
Northeastern Naturalist Vol. 21, No. 1
M.J. Glennon
2014
NENHC-53
Brommer (2004) identified poleward shifts of 19 km in the range margins of Finish
birds in only a 12-year period of time. A meta-analysis of range-boundary changes
for more than 1700 species in the Northern Hemisphere estimated that northern and
upper-elevational boundaries had shifted, on average, 6.1 km per decade northward
or 6.1 m per decade upward (Parmesan and Yohe 2003). The rapidity of these shifts
suggests that the influence of elevation or latitude on colonization and extinction
probabilities of boreal birds over the 5-year time scale of the current study may be
detectable. Other potential impacts of climate change, and therefore mechanisms
through which bird abundance or occupancy patterns may shift, include peat loss,
bog contraction, and tree invasion resulting from increased summer drought created
by higher temperatures and lower rainfall (Moore 2002). Information on the current
occurrence of such changes in the peatlands of the Adirondacks is not available.
Multiseason occupancy data, however, when paired with covariate information, allow
for an opportunity to test whether dynamic processes of avian species in these
habitats are related to latitudinal or elevational patterns.
In addition to the potential impacts of habitat fragmentation and climate change,
there also exists the possibility that, because these are very specialized species
in the Adirondacks associated with a limited habitat type, they may be particularly
sensitive to human alteration and general habitat degradation or subject to
competition with more generalist species that often thrive in human-dominated
environments (DeVictor et al. 2007, Glennon and Porter 2005). The Adirondack
Park is unique in its mix of public and private land uses in a state park setting, with
approximately 50% of its landscape in state ownership or conservation easement,
and the other 50% providing for a variety of uses by its 130,000 year-round residents
and multitudes of visitors. Dynamics of lowland boreal birds here may also
be influenced by proximity and activities of humans in relation to their habitats. As
such, they may be more likely to colonize sites with low human impact and abandon
sites with high human impact.
Using the 5-year intensive survey dataset from WCS, I examined the mechanistic
processes underlying patterns of occupancy change for 8 of 12 monitored species.
Specifically, I tested underlying assumptions of metapopulation dynamics, and then
examined the effects of wetland size, connectivity, latitude, elevation, and development
on short-term dynamics using a multi-year occupancy modeling approach. I
used a model-selection framework and multi-model inference to draw conclusions
about the status and trends of these 8 species in the Adirondacks and to identify those
drivers that may have the strongest influence on observed patterns.
Field-Site Description
The Adirondack Park is located in the northern part of New York State and encompasses
an area of 19,700 km2. Elevations range from 30 to 1600 m, and the dominant
vegetation is a mixture of boreal and northern hardwood forest types (Glennon
and Porter 2005). The predominant habitat type in the park is Northern Hardwood
and Conifer Forest, followed by Boreal Upland Forest and Northern Swamp
(Anderson et al. 2013). The boreal habitats that are the subject of this study consist
Northeastern Naturalist
NENHC-54
M.J. Glennon
2014 Vol. 21, No. 1
of bogs, fens, wooded wetlands, and open river corridors in the Adirondack Park.
Though the Adirondacks as a whole lie in the transition zone between the temperate
and boreal regions, there are extensive areas in the park that have summer
temperatures characteristic of the south edge of the true boreal and are characterized
by boreal community types that are maintained by boreal processes such as ice buildup
on river shores (Jenkins 2010). Boreal habitats are distributed in small patches
throughout the Adirondacks but are most concentrated in a band running from the
north-central part of the park to the southwestern edge. These habitats contain both
high- and low-elevation components; this paper deals solely with low-elevation
boreal communities and does not address the montane boreal. As recently characterized
by Anderson et al. (2013), these boreal communities fall primarily into Northern
Swamp and Northern Peatland macrogroups, with dominant habitat types within
those macrogroups including Northern Appalachian Acadian Conifer Hardwood
Acid Swamp, Boreal Laurentian-Acadian Acidic Basin Fen, and Boreal Laurentian
Bog. These are wet, acid, carbon-accumulating habitats with mean summer temperature
less than 18 °C (64° F) and predominantly coniferous vegetation. Dominant tree species
include Picea mariana (Mill.) Britton, Sterns, & Poggenburg (Black Spruce), Picea
glauca (Moench) Voss (White Spruce), and Larix laricina (Du Roi) K. Koch (Tamarack),
and dominant shrubs include Andromeda polifolia L. (Bog Rosemary), Ledum
groenlandicum Oeder (Labrador Tea), and Kalmia polifolia Wangenh. (Bog Laurel).
Study sites ranged in latitude from 43°40'8.141" to 44°41'39.559"N and in elevation
from 397 to 594 m (Table 1).
Table 1. Study sites at which a suite of 12 boreal bird species were monitored in the Adirondack
Park, NY, 2007–2011, including wetland area, location, and ownership at the time of the study. New
York State Forest Preserve ownership includes state land designations of Wilderness, Wild Forest,
and Primitive Area, recreational uses of which are described in the New York State Land Master Plan
(Adirondack Park Agency 1987). Wilderness is the most restrictive designation.
Name Area (km2) Latitude (N) Longitude (W) Ownership
Barnum Bog 0.63 44°27'13" 74°15'46" Private
Beaver Brook Bog 0.78 43°48'48" 74°42"45" NYS: Wilderness
Beaver Brook Fen East 0.69 43°49'49" 74°40'18" NYS: Wild Forest
Beaver Brook Fen West 0.69 43°49'43" 74°41'3" NYS: Wild Forest
Bigelow Road 5.47 44°25'18" 74°5'57" NYS: Wild Forest
Black Brook North 4.19 44°21'7" 74°31'45" Private
Black Pond Swamp 3.75 44°19'50" 74°17'27" NYS: Wild Forest
Bloomingdale Bog 5.45 44°22'44" 74°8"32" NYS: Wild Forest
Bloomingdale Bog II 3.34 44°24'23" 74°7'33" NYS: Wild Forest
Bloomingdale Bog III 5.47 44°25'13" 74°7'2" NYS: Wild Forest
Blue Mountain Road 0.10 44°29'5" 74°25'49" NYS: Wild Forest
Bog Stream 0.82 44°3'34" 74°33'1" Private
California Road 5.85 44°41'39" 74°12'36" NYS: Wild Forest
Chubb River 0.26 44°15'21" 74°1'11" NYS: Wilderness
Chubb River II 2.50 44°14'35" 74°1'36" NYS: Wilderness
Dexter Road Bog 0.83 44°37'17" 74°27'43" NYS: Wild Forest
Eastern Brown's Tract Inlet Fen 1.62 43°48'29" 74°39'54" NYS: Wild Forest
Ferd’s Bog 0.59 43°47'18" 74°44'58" NYS: Wilderness
Helldiver 0.55 43°40'8" 74°41'38" NYS: Wild Forest
Northeastern Naturalist Vol. 21, No. 1
M.J. Glennon
2014
NENHC-55
Methods
Bird data collection
Focal species. The bird data analyzed for this paper are the result of a long-term
monitoring program run by the Adirondack Program of the WCS for the purpose
of understanding the distribution, abundance, and population trends of a suite of
boreal birds in the park. All of the focal species are at or close to the southern extent
of their eastern North American range in the Adirondack Park and all are known
to occur in the Canadian boreal. A set of species was selected from those deemed
Table 1, continued.
Name Area (km2) Latitude (N) Longitude (W) Ownership
Hitchens Bog 1.04 44°6'34" 74°39'17" NYS: Primitive Area
Horseshoe Bog 1.55 44°7'57" 74°38'46" NYS: Primitive Area
Jones Pond Bog 0.20 44°27'8" 74°32'18" NYS: Wild Forest
Jones Pond Outlet 0.19 44°27'4" 74°12'41" NYS: Wild Forest
Jones Pond Road 0.20 44°27'7" 74°11'59" NYS: Wild Forest
Kildare Bear Creek 5.23 44°19'30" 74°33'38" Private
Kildare Bog 0.42 44°19'55" 74°32'33" Private
Last Gasp Fen 1.54 43°51'14" 74°36'57" NYS: Wild Forest
Little Cherry Patch Pond 0.07 44°18'0" 73°56'22" NYS: Wild Forest
Lost Ponds 0.04 43°41'9" 74°40'19" NYS: Wild Forest
Lower St. Regis 0.19 44°26'2" 74°14'16" Private: easement
Madawaska 1.50 44°30'27" 74°24'38" NYS: Primitive Area
Marion River Fen Central 2.43 43°50'2" 74°34'37" NYS: Wild Forest
Marion River Fen East 2.43 43°50'12" 74°33'8" Private
Marion River Fen West 2.43 43°49'56" 74°35'26" NYS: Wild Forest
Massawepie Mire 6.06 44°13'31" 74°40'40" Private: easement
Meachum Lake Swamp 1.86 44°32'41" 74°17'48" NYS: Wild Forest
Meno Bog 0.50 44°30'55" 74°28'0" NYS: Wild Forest
Moose Pond Road 0.37 44°22'25" 74°8'43" NYS: Wild Forest
Osgood River 3.84 44°28'1" 74°13'42" NYS: Wild Forest
Paul Smiths Bog Complex 0.38 44°25'20" 74°14'37" Private: easement
Quebec Brook 2.05 44°29'40" 74°20'34" NYS: Primitive Area
Red River 0.12 43°41'2" 74°44'52" Private: easement
Regis-Spitfire 0.63 44°25'20" 74°15'28" Private: easement
Rock Pond 1.31 44°21'56" 74°33'14" Private: easement
Round Lake Bog 1.08 44°3'24" 74°34'18" Private
Route 55 Bloomingdale 3.34 44°24'49" 74°7'46" NYS: Wild Forest
Sabattis Circle Road 0.32 44°4'13" 74°32'30" Private: easement
Sevey Bog 2.14 44°15'29" 74°41'49" Private
Silver Lake Bog 0.57 44°28'50" 73°53'18" Private: easement
Slush Pond 0.51 44°28'10" 74°18'37" NYS: Wild Forest
South Inlet Fen North 0.76 43°47'58" 74°37'3" NYS: Wilderness
South Inlet Fen South 0.76 43°47'17" 74°37'48" NYS: Wilderness
Spring Pond Bog 4.19 44°22'12" 74°30'10" Private: easement
Spring Pond Bog South 4.19 44°21'37" 74°30'57" Private: easement
Sumner Stillwater 0.06 43°41'19" 74°39'42" NYS: Wild Forest
Ton-Da-Lay 2.08 44°22'41" 74°28'30" Private: easement
Twin Brook Bog 0.81 44°34'58" 74°29'33" Private: easement
Western Brown's Tract Inlet Fen 1.62 43°48'12" 74°41'17" NYS: Wild Forest
Northeastern Naturalist
NENHC-56
M.J. Glennon
2014 Vol. 21, No. 1
to best represent the lowland boreal habitats of the Adirondacks and to be best
sampled with a point-count methodology. Those species are Picoides dorsalis Baird
(American Three-toed Woodpecker), Picoides arcticus Swainson (Black-backed
Woodpecker), Contopus cooperi Swainson (Olive-sided Flycatcher), Empidonax
flaviventris Baird and Girard (Yellow-bellied Flycatcher), Perisoreus canadensis
L. (Gray Jay), Poecile hudsonicus Forster (Boreal Chickadee), Oreothlypis peregrina
Wilson (Tennessee Warbler), Setophaga tigrina Gmelin (Cape May Warbler),
Setophaga castanea Wilson (Bay-breasted Warbler), Setophaga palmarum Gmelin
(Palm Warbler), Melospiza lincolnii Audubon (Lincoln’s Sparrow), and Euphagus
carolinus Müller (Rusty Blackbird).
Site selection. An initial list of potential field sites was compiled by consulting
a variety of data sources including Adirondack Park Agency wetlands inventory
data, New York State Breeding Bird Atlas data (Andrle and Carroll 1988, McGowan
and Corwin 2008), postings to the Northern New York Breeding Bird Listserv, and
local expert opinion. The final list of study sites was then determined by selecting
from within the potential list to include a number of the major well-known boreal
wetlands of the Adirondack Park and a random sample of smaller, less-known
locations. Because the lowland boreal habitats of the Adirondacks are relatively
disjunct and many are located in remote and roadless areas, our design precluded
a completely random selection of study sites. The best possible effort was made to
include a mix of known boreal wetlands in which some of these species had been
documented in the past and numerous sites that had never been surveyed.
Avian monitoring. WCS conducted unlimited-distance point counts to assess presence/
absence of our target species along transects of 5 points spaced at least 250 m
apart within boreal wetland habitats (Ralph et al. 1995). In a small number of particularly
large wetlands or wetland complexes, multiple transects were placed in order to
adequately represent the bird community present, but spaced with a minimal distance
separation of 300 m to maintain independence. All points were surveyed for 10-minutes
between the hours of 5:00 and 9:00 am. Survey start and end dates for each year
varied with weather conditions and song activity. All sampling occurred during the
primary breeding season on survey dates ranging from the last week of May to the
third week of July, with the majority of sites sampled in June. At each sample point,
birds were recorded by species, time period of detection (i.e., 0–3 minutes, 3–5
minutes, 5–10 minutes), and activity (i.e., singing, calling, individual seen). Point
counts were conducted by trained observers, the majority of whom conducted counts
at the same locations for 3 or more of the project years. During counts, we recorded
the date, start and end time for each survey, ambient temperature, and sky and wind
conditions. We measured sky conditions on a scale from 0 to 6 ranging from clear
or a few clouds to rain, and wind on a Beaufort scale from 0 to 5 ranging from calm
to small trees swaying. Surveys were halted in the event of wind or sky conditions
that affected behavior or precluded detection of birds (e.g., 5 or 6 on the sky scale and
4 or 5 on the wind scale). Surveys in boreal river corridors were conducted by boat.
We employed spatial replication of sample points rather than temporal replication,
to reduce costs and allow for the calculation of detection probabilities (MacKenzie
et al. 2006). The sites themselves, and not the five points within each site, serve as
Northeastern Naturalist Vol. 21, No. 1
M.J. Glennon
2014
NENHC-57
the experimental units for the purposes of analysis. Estimation of parameters from
spatial, rather than temporal replication is also employed by the BBS (Hines et al.
2010). In most cases, the sites were large and uniform enough to accommodate a linear
transect of 5 points, but in some cases points were placed in a nonlinear fashion,
maintaining a minimal distance separation of 250 m. We have sampled more than 80
locations over the course of the project; 58 of those were sampled consistently for the
period between 2007 and 2011 and are the subject of the current analysis (Table 1).
Analysis
GIS datasets. I used 3 primary GIS datasets to calculate variables of hypothesized
importance to boreal bird dynamics: wetland cover-type maps, a regional
human-footprint layer (Woolmer et al. 2008), and a digital elevation model. Area
and connectivity of study wetlands were calculated from wetland cover-type data
provided by the Adirondack Park Agency (APA) and made available as a layer on
their Shared Adirondack Park Geographic Information CD-ROM ver 1.0. These
maps were produced by APA staff and exist for all watersheds in the Adirondacks.
Polygon vector files are provided in which all park wetlands are classified by system,
which describes the complex of wetlands and deepwater habitats that share
the influence of similar hydrologic, geomorphologic, chemical, or biological factors
(e.g., palustrine, lower perennial riverine), and by class, which describes the
general appearance of the habitat in terms of either the dominant life form of the
vegetation or the physoiography and composition of the substrate (e.g., forested/
evergreen, broad-leaved deciduous scrub/shrub; Cowardin et al. 1979). I consulted
with local wetland ecologists to determine which of the class types within the
cover-type wetlands corresponded to the boreal habitat types in which our sampling
was conducted and used those cover-type polygons to calculate the area of each
wetland sampled. In order to assess wetland connectivity, I used the cluster and
outlier analysis tool in ArcMap 10, which identifies spatial clusters of features with
high or low value based on an analysis field. I ran the tool on the wetland polygons
described above, with wetland area as the analysis field. The tool calculates a Moran’s
I value, which is positive when the feature is part of a cluster and negative
when the feature represents an outlier. I used a fixed-distance band of 5 km in order
to evaluate the statistical properties of the wetland area data at a specific spatial
scale. I chose this distance because it is representative of landscape scale for birds
and comparable to other studies that have investigated the impacts of habitat size
and isolation on various avian species (Brown and Dinsmore 1986, Hames et al.
2001, Smith and Chow-Fraser 2010). Calculated Moran’s I values for each study
wetland were used as an index of habitat connectivity.
I used a regional human-footprint dataset to characterize the relative human
influence at each of the study wetlands. The human footprint is a representation of
the magnitude of human transformation of the landscape and was originated at a
global scale by Sanderson et al. (2002). Woolmer et al. (2008) used an adaptation of
Sanderson’s methodology to map the human footprint at the Northern Appalachian
ecoregional scale. GIS Layers from Woolmer et al.’s (2008) analysis were made
available through the Two Countries One Forest Northern Appalachian online atlas
(www.2c1forest.org). Scores range from 0 to 100 and represent the relative impact
Northeastern Naturalist
NENHC-58
M.J. Glennon
2014 Vol. 21, No. 1
associated with human settlement, access, land-use change, and electrical power infrastructure.
This dataset provides a relative measurement of human transformation
of the natural landscape across the park. I used the average human-footprint score
across each of the 5 points along each study transect to characterize human impact
at each study wetland.
I obtained information on elevation of study wetlands from a digital elevation
model also provided by the Adirondack Park Agency, and calculated a latitude for
each transect by mapping their UTM coordinates and using ArcMap 10 to determine
the latitude of the centroid for each transect. Together, these datasets resulted in 5
variables used in occupancy models to characterize study wetlands: area (wetarea),
connectivity (connect), latitude (utmy), elevation (elev), and human footprint (HF).
Occupancy modeling. To investigate dynamics of boreal birds in the Adirondacks,
I used the multi-season model implemented in program Presence (Hines
2006) to calculate detection (p), occupancy (ψ), colonization (γ), and extinction (ε)
probabilities for 2007–2011 for each of the species for which adequate data were
obtained (detections at 15% or more of study locations; George and Zack 2008).
Those included Black-backed Woodpecker, Olive-sided Flycatcher, Yellow-bellied
Flycatcher, Gray Jay, Boreal Chickadee, Palm Warbler, Lincoln’s Sparrow, and
Rusty Blackbird. Data for the other 4 target species (Three-toed Woodpecker, Tennessee
Warbler, Cape May Warbler, and Bay-breasted Warbler) constituted ≤12
detections of each species in the entire 5-year dataset and therefore could not be
used in an occupancy-modeling framework.
Occupancy probability is defined as the probability of a site being occupied
within a given season, while detection probability denotes the probability of a species
being detected, given its presence. Colonization probability is defined as the
probability that an unoccupied site in season t is occupied by the species in season
t + 1, and extinction probability denotes the probability that a site occupied in season
t is unoccupied by the species in season t + 1 (MacKenzie et al. 2006). Program
Presence uses the logit link and a maximum-likelihood approach to linearize the
relationships among covariates and probabilities of detection, occupancy, colonization,
and extinction (Glennon and Kretser 2013, Hines 2006). I used Akaike’s information
criterion (AIC) to select among competing models calculated in Presence
(Akaike 1974, Burnam and Anderson 2002).
I first modeled detection for each species while holding other parameters constant
and tested 6 variables for their influence on detection probability including
wind, sky (relative cloud cover), date, time, temperature, and observer. Here, I
used the default parameterization of the multi-season model, which estimates initial
occupancy, colonization, and extinction probabilities directly, and is the most
numerically stable (MacKenzie et al. 2006). Upon determining the best single
predictor of detection probability, I ran a set of 9 models (Table 2) incorporating
best-detection variables to first establish the appropriateness of the assumption of
Markovian changes in occupancy and metapopulation structure for these birds, as
well as to determine the evidence for stable or dynamic populations. This initial
phase of the analysis addressed whether for each species (1) occupancy is static or
dynamic, (2) changes in occupancy are random or Markovian, (3) populations are
Northeastern Naturalist Vol. 21, No. 1
M.J. Glennon
2014
NENHC-59
at equilibrium, and (4) probability of occupancy and local extinction are dependent
on patch size, as predicted by metapopulation theory.
Using support for metapopulation dynamics and Markovian changes in occupancy
from the initial model set, I next modeled the impacts of 5 covariates (wetland
area, connectivity, latitude, elevation, and human footprint) on extinction and colonization
dynamics for each species over the 5-year period between 2007 and 2011,
incorporating the best predictor for detection. This model set also included 2 equilibrium
models for comparison against dynamic models: one in which equilibrium
was defined in terms of constant occupancy probability (Stationary Markov) and
one in which it was defined in terms of constant colonization and extinction probabilities
(MacKenzie et al. 2006). I did not place any covariates on occupancy itself,
assuming it to be reflective of past dynamics (Sjögren-Gulve and Hanski 2000) and
adequately captured by covariates placed on colonization and extinction. Finer-scale
habitat quality is also likely to impact occupancy dynamics of these species, but past
analyses of these data have demonstrated much greater support for large-scale factors
in controlling occupancy (M.J. Glennon, unpubl. data). The purpose of the second
part of the analysis was specifically to assess the impact of landscape-scale drivers on
short-term dynamics of these birds in the Adirondacks.
In this second phase of the analysis, I tested a set of models whereby I asked if
(1) occupancy and/or rates of colonization and extinction are constant, (2) colonization
dynamics depend on wetland area, connectivity, latitude, elevation, and human
infrastructure, and (3) extinction dynamics depend on wetland area, connectivity,
latitude, elevation, and human infrastructure. I did not have plausible biological
explanations for modeling every possible combination of covariates and chose, for
simplicity, to hold one rate constant and vary the other within the model set, resulting
in a set of 12 models for each species (Table 3). It is possible, of course, that
both colonization and extinction rates vary at the same time and, as such, I draw inferences
from the betas and model-averaged estimates of γ and ε for all models. My
predictions were that (1) wetland area and connectivity would positively influence
colonization and negatively influence extinction—bigger, more-connected wetlands
are expected to be of higher quality for birds than smaller, more-isolated sites,
(2) latitude and elevation would positively influence colonization and negatively
Table 2. Nine models of occupancy (ψ), colonization (γ), and extinction (ε) probability used to examine
metapopulation structure and equilibrium assumptions for 8 bird species in boreal wetlands of the
Adirondack Park, NY, 2007–2011. Covariates are explained in Methods.
Model Predicted dynamics reflective of
ψ (.) Static occupancy
ψ (2007), γ(year), ε = 1- γ Random changes, population not at equilibrium
ψ (2007), γ(.), ε = 1- γ Random changes, population at equilibrium
ψ (.), γ(.), ε(.) Markovian changes, population at equilibrium
ψ (.), γ(year), ε(year) Markovian changes, population not at equilibrium
ψ (.), γ(.) Stationary Markov process
ψ (.), ε(wetarea) Constant occupancy, area-driven extinction
ψ (.), γ(.), ε(wetarea) Markovian changes, area-driven extinction
ψ (wetarea), γ(.), ε(wetarea) Markovian changes, area-driven occupancy and extinction
Northeastern Naturalist
NENHC-60
M.J. Glennon
2014 Vol. 21, No. 1
influence extinction—given their location at the southern range extent, it is expected
that these birds may move northward and up in elevation over time given climate
change, and (3) human infrastructure would negatively influence colonization, and
positively influence extinction—these specialist birds will generally avoid human
habitats and/or be outcompeted by more abundant, generalist species near humaninfluenced
areas. I used model-averaged estimates of colonization and extinction to
calculate occupancy rates for each of the years between 2007 and 2011 in order to
examine trends over time. The default model parameterization calculates colonization
and extinction probabilities, as well as occupancy for year 1(ψt); occupancy for
each subsequent season is calculated as:
ψt+1 = ψt(1 - εt) + (1 - ψt)γt ,
where ψ represents occupancy probability, and γ and ε represent colonization and
extinction probabilities, respectively (MacKenzie et al. 2006).
Results
A total of 1305 detections were made for all species over the 5-year time frame,
with the majority of detections occurring for Yellow-bellied Flycatcher (30%), Lincoln’s
Sparrow (23%), and Yellow Palm Warbler (20%), and far fewer detections
made of Black-backed Woodpecker (8%), Gray Jay (8%), Olive-sided Flycatcher
(6%), Boreal Chickadee (3%), and Rusty Blackbird (2%). Wetland area ranged from
0.04–6 km2 (mean = 1.77 km2), elevation ranged from 397–594 m (mean = 512 m),
and latitude ranged from 43°40'8"N–44°41'40"N. Most wetlands (90%) were associated
with positive values for Moran’s I, indicating that they were within clusters.
Z scores calculated from Moran’s I values for these wetlands indicated that the
majority of them (83%) were large wetlands within clusters of other large wetlands
(P < 0.05). Human-footprint values for individual wetlands ranged from 3.8–47.5
(mean = 20), and these values generally did not vary greatly among points within
wetlands (mean coefficient of variation = 23%).
Table 3. Twelve models used to predict probability of occupancy (ψ), colonization (γ), and extinction
(ε) for 8 bird species in boreal wetlands of the Adirondack Park, NY, 2007–2011. Covariates are
explained in Methods.
Model Predicted dynamics dependent on
ψ (.), γ(.), ε(.) Markovian changes in yearly occupancy of wetlands.
ψ (.), γ(.) Stationary Markov process.
ψ (.), γ(wetarea), ε(.) Area-driven colonization rates.
ψ (.), γ(connect), ε(.) Connectivity-driven colonization rates.
ψ (.), γ(utmy), ε(.) Latitude-driven colonization rates.
ψ (.), γ(elev), ε(.) Elevation-driven colonization rates.
ψ (.), γ(HF), ε(.) Human impact-driven colonization rates.
ψ (.), γ(.), ε(wetarea) Area-driven extinction rates.
ψ (.), γ(.), ε(connect) Connectivity-driven extinction rates.
ψ (.), γ(.), ε(utmy) Latitude-driven extinction rates.
ψ (.), γ(.), ε(elev) Elevation-driven extinction rates.
ψ (.), γ(.), ε(HF) Human impact-driven extinction rates.
Northeastern Naturalist Vol. 21, No. 1
M.J. Glennon
2014
NENHC-61
I found that no single variable best predicted detection probability for all boreal
species. Time of day and observer were the best predictors of detection probability
for 2 species each, while detection probabilities for the 4 remaining species were
best predicted by wind, sky, date, and temperature, respectively (Table 4). The results
of top models in the first model set indicated that changes in occupancy were
driven by Markovian, rather than random changes. Though constant-occupancy
models were supported for some species, the majority of models supported for 6 of
8 species were dynamic models, indicating that occupancy was not constant over
the 5-year period. Most species also demonstrated some support for area-driven
extinction, as predicted by metapopulation theory (Table 4). These results provided
justification for the structure of the second phase of the analysis, which explicitly
examined drivers of colonization and extinction, and which assumed Markovian
changes in occupancy.
Results of the second model set indicated that most species were controlled more
strongly by extinction rather than colonization dynamics (Table 5). Among drivers
of extinction and colonization dynamics, the strongest predictors by total model
weight across all species were the effect of elevation on colonization and latitude
Table 4. Summary of model selection results from analysis of underlying dynamics for 8 bird species
monitored in boreal wetlands in the Adirondack Park, NY, 2007–2011. Covariates are explained
in Methods; only the results of top models are shown (ΔAIC ≤ 2.0). L = likelihood; # = number of
parameters. Species: B-b W = Black-backed Woodpecker, O-s F = Olive-sided Flycatcher, Y-b F =
Yellow-bellied Flycatcher, G J = Gray Jay, B C = Boreal Chickadee, P W = Palm Warbler, L S = Lincoln’s
Sparrow.
Species Model AIC ΔAIC AIC wt L # -2LogLike
B-b W ψ(wetarea), γ(.), ε(wetarea), p(wind) 713.23 0.00 0.4754 1.0000 7 699.23
ψ(.), γ(.), ε(wetarea), p(wind) 713.27 0.04 0.4660 0.9802 6 701.27
O-s W ψ(.), γ(.), p(time) 571.9 0.00 0.3448 1.0000 4 563.90
ψ(.), γ(.), ε(.), p(time) 572.18 0.28 0.2998 0.8694 5 562.18
ψ(.), γ(.), ε(wetarea), p(time) 573.39 1.49 0.1637 0.4747 6 561.39
Y-b F ψ(.), γ(.), p(date) 1461.88 0.00 0.4878 1.0000 4 1453.88
ψ(.), γ(.), ε(.), p(date) 1463.68 1.80 0.1983 0.4066 5 1453.68
G J ψ(.), γ(.), ε(wetarea), p(obs) 687.78 0.00 0.298 1.0000 6 675.78
ψ(wetarea), γ(.), ε(wetarea), p(obs) 688.21 0.43 0.2404 0.8065 7 674.21
ψ(.), γ(.), p(obs) 688.41 0.63 0.2175 0.7298 4 680.41
ψ(.), γ(.), ε(.), p(obs) 689.49 1.71 0.1267 0.4253 5 679.49
B C ψ(wetarea), γ(.), ε(wetarea), p(obs) 364.37 0.00 0.8125 1.0000 7 350.37
P W ψ(wetarea), γ(.), ε(wetarea), p(temp) 998.42 0.00 0.2811 1.0000 7 984.42
ψ(.), γ(.), ε(wetarea), p(temp) 998.7 0.28 0.2444 0.8694 6 986.70
ψ(.), γ(.), ε(.), p(temp) 999.39 0.97 0.1731 0.6157 5 989.39
ψ(.), ε(wetarea), p(temp) 999.46 1.04 0.1671 0.5945 5 989.46
ψ(.), γ(.), p(temp) 999.97 1.55 0.1295 0.4607 4 991.97
L S ψ(.), γ(year), ε(year), p(time) 1240.98 0.00 0.7285 1.0000 11 1218.98
R B ψ(.), γ(.), p(sky) 235.16 0.00 0.3972 1.0000 4 227.16
ψ(.), ε(wetarea), p(sky) 236.16 1.00 0.2409 0.6065 5 226.16
ψ(.), γ(.), ε(.), p(sky) 236.85 1.69 0.1706 0.4296 5 226.85
Northeastern Naturalist
NENHC-62
M.J. Glennon
2014 Vol. 21, No. 1
on extinction processes. In both cases, however, agreement with predictions was
mixed (Table 6). In general, no single covariate had strong effects on colonization
or extinction dynamics across species, and there was high variability among species
in their response to individual covariates. With respect to colonization, the strongest
responses by species were as follows: Black-backed Woodpecker to wetland area;
the 2 flycatcher species to human footprint; Gray Jay, Rusty Blackbird, and Palm
Warbler to elevation; and Boreal Chickadee and Lincoln’s Sparrow to latitude. In
terms of extinction probability, the different species were most strongly influenced
as follows: Black-backed Woodpecker and Rusty Blackbird by connectivity; Olivesided
Flycatcher, Boreal Chickadee, and Lincoln’s Sparrow by latitude; Gray Jay
by elevation; Yellow-bellied Flycatcher by human footprint; and Palm Warbler by
wetland area (Table 6).
In addition to examining model weights, it is also of value to examine signs
of the betas to determine the degree of consistency with which species responded
to covariates and the degree of agreement with predictions. Human footprint and
wetland area were the most consistent predictors of colonization probability across
species, and size and connectivity of wetlands—as well as human footprint—were
the most consistent predictors of extinction probability (Table 6).
Trends calculated from modeled colonization and extinction probabilities indicated
that 4 of the 8 species modeled are demonstrating a pattern of declining
occupancy in boreal wetlands in the Adirondacks, although the relative rate of
decline is variable among them (Table 7). Rusty Blackbird and Yellow-bellied
Flycatcher occupancy remained stable over the 5-year period, while only Lincoln’s
Sparrow and Palm Warbler demonstrated a pattern of increasing occupancy. In no
Table 5. Summary of model selection results from analysis of drivers of dynamics for 8 bird species
monitored in boreal wetlands in the Adirondack Park, NY, 2007–2011. Covariates are explained
in Methods; only the results of top models are shown (ΔAIC ≤ 2.0). L = likelihood; # = number of
parameters. Species: B-b W = Black-backed Woodpecker, O-s F = Olive-sided Flycatcher, Y-b F =
Yellow-bellied Flycatcher, G J = Gray Jay, B C = Boreal Chickadee, P W = Palm Warbler, L S = Lincoln’s
Sparrow.
Species Model AIC ΔAIC AIC wt L # -2LogLike
B-b W ψ(.), γ(.), ε(connect), p(wind) 711.25 0.00 0.5411 1.0000 6 699.25
O-s F ψ(.), γ(.), ε(utmy), p(time) 569.46 0.00 0.3420 1.0000 6 557.46
Y-b F ψ(.), γ(.), ε(HF), p(date) 1454.08 0.00 0.6826 1.0000 6 1442.08
ψ(.), γ(HF), ε(.), p(date) 1455.96 1.88 0.2666 0.3906 6 1443.96
G J ψ(.), γ(elev), ε(.), p(obs) 682.37 0.00 0.5231 1.0000 6 670.37
B C ψ(.), γ(.), ε(utmy), p(.) 362.14 0.00 0.7605 1.0000 5 352.14
P W ψ(.), γ(elev), ε(.), p(temp) 997.63 0.00 0.2513 1.0000 6 985.63
ψ(.), γ(.), ε(wetarea), p(temp) 998.70 1.07 0.1472 0.5857 6 987.60
ψ(.), γ(.), ε(.), p(temp) 999.39 1.76 0.1042 0.4148 5 989.39
L S ψ(.), γ(.), ε(utmy), p(time) 1274.74 0.00 0.5422 1.0000 6 1228.74
R B ψ(.), γ(elev), ε(.), p(sky) 232.86 0.00 0.2714 1.0000 6 220.86
ψ(.), γ(wetarea), ε(.), p(sky) 233.02 0.16 0.2505 0.9231 6 221.02
ψ(.), γ(utmy), ε(.), p(sky) 233.63 0.77 0.1847 0.6805 6 221.63
Northeastern Naturalist Vol. 21, No. 1
M.J. Glennon
2014
NENHC-63
Table 7. Model-averaged parameter estimates of occupancy (ψ), colonization (γ), and extinction (ε) for 8 bird species in boreal wetlands in the Adirondack
Park, NY, 2007–2011.
Black-backed Olive-sided Yellow-bellied Boreal Lincoln’s Rusty
Parameter Woodpecker Flycatcher Flycatcher Gray Jay Chickadee Palm Warbler Sparrow Blackbird
ψ2007 0.71 ± 0.12 0.55 ± 0.12 0.77 ± 0.08 0.71 ± 0.13 0.36 ± 0.11 0.39 ± 0.07 0.63 ± 0.09 0.22 ± 0.11
ψ2008 0.64 ± 0.12 0.49 ± 0.12 0.75 ± 0.08 0.68 ± 0.13 0.32 ± 0.11 0.44 ± 0.07 0.65 ± 0.09 0.22 ± 0.11
ψ2009 0.58 ± 0.12 0.45 ± 0.12 0.75 ± 0.08 0.67 ± 0.13 0.29 ± 0.11 0.48 ± 0.07 0.67 ± 0.09 0.23 ± 0.11
ψ2010 0.54 ± 0.12 0.43 ± 0.12 0.75 ± 0.08 0.67 ± 0.13 0.27 ± 0.11 0.51 ± 0.07 0.68 ± 0.09 0.23 ± 0.11
ψ2011 0.50 ± 0.12 0.42 ± 0.12 0.75 ± 0.08 0.66 ± 0.13 0.26 ± 0.11 0.53 ± 0.07 0.69 ± 0.09 0.23 ± 0.11
γ 0.08 ± 0.12 0.14 ± 0.09 0.44 ± 0.14 0.37 ± 0.29 0.09 ± 0.08 0.13 ± 0.05 0.17 ± 0.10 0.10 ± 0.11
ε 0.13 ± 0.10 0.22 ± 0.14 0.15 ± 0.07 0.19 ± 0.10 0.28 ± 0.28 0.07 ± 0.04 0.07 ± 0.04 0.34 ± 0.25
Table 6. Model weights for single-covariate models used to predict colonization and extinction rates of 8 boreal bird species in the Adirondack Park, NY,
2007–2011. Prediction indicates the direction of the expected relationship between predictor variables and colonization and extinction rates; asterisks denote
that the effect of the variable on colonization or extinction was as predicted (i.e., the sign of the betas matched predictions). Covariates are explained
in Methods.
Black-backed Olive-sided Yellow-bellied Boreal Lincoln’s Rusty
Covariates Prediction Woodpecker Flycatcher Flycatcher Gray Jay Chickadee Palm Warbler Sparrow Blackbird Total wt
Colonization
wetarea + 0.15 0.03* 0.00* 0.02* 0.02* 0.04 0.02 0.25* 0.55
connect + 0.01 0.05 0.00 0.16* 0.00* 0.04 0.02 0.02* 0.29
utmy + 0.03 0.04* 0.01 0.07* 0.03* 0.09* 0.11 0.18 0.55
elev + 0.03* 0.04 0.01* 0.52 0.01 0.25 0.01* 0.27* 1.14
HF - 0.00* 0.08 0.27* 0.01* 0.02* 0.05 0.02* 0.03* 0.47
Extinction
wetarea - 0.20 0.05* 0.00* 0.04* 0.01* 0.15* 0.01* 0.01* 0.46
connect - 0.54 0.03 0.00* 0.02* 0.09* 0.07* 0.01* 0.06* 0.84
utmy - 0.01 0.34* 0.00 0.02* 0.76* 0.04* 0.54* 0.02 1.72
elev - 0.01* 0.12 0.00 0.06 0.00 0.05 0.14 0.01 0.41
HF + 0.00* 0.03 0.68* 0.04* 0.02* 0.04 0.01* 0.02* 0.84
Northeastern Naturalist
NENHC-64
M.J. Glennon
2014 Vol. 21, No. 1
case was the stationary occupancy model included in the top model set (ΔAIC less than
2.0) for any species (Table 5).
Discussion
The geography of the Adirondack boreal region, the location of these species at
their southern range extent within eastern North America, and the patterns identified
here suggest that boreal bird populations in the Adirondack Park are dynamic.
This analysis also demonstrates that occupancy by several species appears to be
in decline. Understanding the processes that drive their dynamics can enhance the
ability of land managers to influence their long-term conservation (MacKenzie et
al. 2006). For species occurring in the patchy habitats of the Adirondack boreal
region, metapopulation biology and the lessons that arise from it are also valuable
to conservation planning. I predicted that wetland size and connectivity would
positively influence colonization and negatively influence extinction of birds in
boreal wetlands in the Adirondack Park. Patterns in boreal bird dynamics matched
these expectations of metapopulation theory strongly for extinction but not for
colonization. Seven of the 8 species modeled were more likely to disappear from
smaller, isolated wetlands, but only 5 of 8 were more likely to colonize larger
wetlands, and only 3 of 8 were more likely to colonize more connected wetlands.
As suggested by Hames et al. (2001) and others, the effect of isolation must be
measured in comparison to the dispersal abilities of the organism under study. It
is possible that these effects were inconsistent across species and stronger with
respect to extinction dynamics because birds are highly vagile and thus less sensitive
to isolation of their habitats than other vertebrate taxa. Opdam (1991) and
others have suggested, however, that the effects of isolation on birds should not
be dismissed a priori (Hames et al. 2001, Villard et al. 1995). Hames et al. (2001)
also state that resident birds are expected to be more affected by isolation of habitat
patches than are long-distance migrants. I did find a greater influence of wetland
connectivity on resident birds than on migrants (Table 6), but the direction of the
influence was as predicted for only 2 of the 3 resident species. In general, I found
stronger evidence for the influence of wetland size and connectivity on extinction
rather than colonization dynamics, and uneven support among species, with some
(e.g., Black-backed Woodpecker, Rusty Blackbird) much more highly influenced
by size and connectivity of boreal wetlands than others and not necessarily in the
ways predicted. Most investigations of boreal birds elsewhere have not employed
a metapopulation framework because such an approach may be less appropriate in
more continuous boreal habitats in northern regions. Support for birds exhibiting
metapopulation dynamics has been found for other species, however, including the
work of Hames et al. (2001) in investigations of Piranga olivacea Gmelin (Scarlet
Tanager) breeding in fragmented habitats in eastern North America, and Smith and
Chow-Fraser (2010) who examined the influence of isolation on marsh bird communities
in southern Ontario.
I also predicted that latitude and elevation would positively influence colonization
and negatively influence extinction for these bird species. Because they are
Northeastern Naturalist Vol. 21, No. 1
M.J. Glennon
2014
NENHC-65
largely at their southern range extent in the Adirondacks, I expected that these birds
may move northward and up in elevation over time given observed and predicted
changes in distribution and abundance of other northern bird species across the
globe in response to climate change (Virkkala and Rajasärkkä 2011, Virkkala et al.
2008, Waite and Strickland 2006, Zuckerberg et al. 2009). Both latitude and elevation
were strong predictors of colonization and extinction dynamics, but only for
a few species, and with inconsistent results. Some species did have higher probability
of extinction at southern sites and at sites of lower elevation, while others
demonstrated opposite patterns. The inconsistency of responses among species to
these 2 predictor variables suggests that other factors may be playing a larger role
in controlling these species’ dynamics than does climate change over this small
window of time, and that the responses of individual species to climate change are
not likely to be uniform or highly predictable. Though these species are at their
southern range extent and expected to be sensitive to climate change, the short
duration of the dataset in comparison to climate-driven processes may preclude detection
of changes driven solely by warming. Zuckerberg et al. (2011) pointed out
the importance of urbanization and behavioral adaptation in modifying the impact
of climate change on birds. Human land use—in particular, residential and commercial
development—along with other stressors are likely interacting with any
impacts of climate change on these birds in the Adirondacks.
I predicted that proximity of human infrastructure would impact these highly
specialized species by negatively influencing colonization and positively influencing
extinction rates. Although these boreal wetlands are, in general, not in close
proximity to roads and development in the Adirondacks, our previous research in
the region, as well as that of others, has demonstrated that the impacts of development
can permeate significant distances within undeveloped areas (Bock et al.
1999, Glennon and Kretser 2013, Odell and Knight 2001, Smith and Chow-Fraser
2010) and that development tends to benefit more generalist species at the expense
of habitat specialists in this landscape and elsewhere (DeVictor et al. 2007, Glennon
and Kretser 2013, Glennon and Porter 2005, Hansen et al. 2005, Maestas et
al. 2001, Smith and Chow-Fraser 2010). When considering both colonization and
extinction, human footprint was the most consistent predictor of boreal bird dynamics
across species, with 6 of 8 modeled species more likely to colonize areas of low
human infrastructure and more likely to experience local extinction from highly
impacted areas. It is likely that human infrastructure, in the form of roads and
houses, is close enough to boreal wetlands in some places that the negative consequences
that can accompany human development are apparent. In the case of these
highly specialized bird species, it is possible that the influx of more cosmopolitan
species, those better able to exploit a variety of habitats and food sources, may
result in competition through which these boreal specialists ultimately lose out.
Such processes have been documented for marsh birds in areas proximal to urbanization
(Smith and Chow-Fraser 2010) and for the songbird community in upland
forested habitats of the Adirondacks (Glennon and Porter 2005). They have also
been hypothesized as a potential driver of the alarming decline in Rusty Blackbird
Northeastern Naturalist
NENHC-66
M.J. Glennon
2014 Vol. 21, No. 1
abundance (Greenberg and Matsuoka 2010). It is possible that competitive interactions
around food or other resources between several of these species and more
widespread family members (e.g., Gray Jay and Cyanocitta cristata L. [Blue Jay],
Boreal Chickadee and Poecile atricapillus L. [Black-capped Chickadee]) may
impact their success in wetlands more proximal to human-dominated areas; such
species are commonly detected in the study wetlands described here (M.J. Glennon,
unpubl. data). Niemi et al. (1998) identified a need for a better understanding of
interactions with predators and competitors and effects to population variability as
a critical knowledge gap for boreal bird species.
I suspect that isolated wetland populations of boreal birds are functioning as
metapopulations in the Adirondack Park, but that their dynamics with respect to
response to wetland size and connectivity are tempered by other factors such as
warming temperatures and adjacent human infrastructure. Modeled occupancy
rates for 4 of the 8 study species demonstrate declining patterns; 2 appear stable,
and only Lincoln’s Sparrow and Palm Warbler appear to be increasing in the Adirondack
landscape. Colonization and extinction rates among these species are
highly variable, with 4 species demonstrating higher colonization rates than extinction
rates, and the other 4 species demonstrating the opposite pattern. Several of the
species for which I observed declining occupancy in the Adirondacks do not appear
to show declining abundance across North America from the large-scale monitoring
effort of the North American Breeding Bird Survey, including Black-backed Woodpecker,
Boreal Chickadee, and Gray Jay, though Boreal Chickadee does appear to
be declining in the Atlantic Northern Forest. These data may be deficient for this
species and Black-backed Woodpecker (BBS; Sauer et al. 2012). Contrastingly,
Lincoln’s Sparrow, which increased in occupancy between 2007 and 2011 in the
Adirondacks, shows no significant survey-wide trend (Sauer et al. 2012). Rates of
colonization and extinction among the 8 modeled species are modest for the most
part, but there are notable exceptions.
Three species may warrant particular attention in the Adirondacks. Boreal
Chickadee, Olive-sided Flycatcher, and Rusty Blackbird showed high extinction
probabilities in Adirondack wetlands, and Boreal Chickadee and Olive-sided Flycatcher
appear to demonstrate declining trends in New York as well as on some
larger scales. All 3 species declined in occupancy across New York State in the
period between 1985 and 2005 (McGowan and Corwin 2008). Rusty Blackbird did
not show declines in the short-term period described here, but its occupancy rate
is so low that large changes may be difficult to detect. All three of these birds were
more likely to experience local extinction from smaller, more isolated wetlands,
and both Boreal Chickadee and Rusty Blackbird appeared sensitive to human infrastructure
in close proximity to wetland sites. Rusty Blackbird is known to be
declining at alarming rates and is the subject of much current effort to elucidate
causes and mechanisms for its decline (Greenberg and Matsuoka 2010). Olive-sided
Flycatcher is also experiencing declines and is listed as a species of concern by
several management entities (Altman and Sallabanks 2012). Boreal Chickadee has
not generated widespread concern, and its status is unknown in the high-elevation
Northeastern Naturalist Vol. 21, No. 1
M.J. Glennon
2014
NENHC-67
boreal habitats it also uses in the Adirondacks, but its high probability of extinction,
coupled with low colonization rates and declining occupancy, may indicate an
ongoing decline in this landscape.
Several of the predictor variables I tested for their influence on boreal bird dynamics
are, unfortunately, confounded to a large degree by the geography of the
Adirondack boreal region. The majority of the large and highly connected wetland
complexes in the Adirondacks are located in the northwest part of the park, and
boreal habitats at more southern latitudes in the southwest region are generally
smaller and more isolated. As such, it is difficult to separate the effects of wetland
size and connectivity from the effects of latitude. The interrelatedness of these factors
presents both opportunities and challenges for conservation, however. Larger,
more connected wetlands are more likely to retain birds, but most wetlands in the
Adirondacks (and everywhere) are small (Semlitsch and Bodie 1998). Smaller wetlands
are also more likely to be nearby and impacted by human infrastructure and
smaller, more isolated wetlands in the Adirondacks are generally located at higher
elevations. Because these birds probably function in metapopulations across this
landscape, all boreal wetlands probably have some importance to them, and smaller,
isolated wetlands may serve as sink habitats that are of lesser quality but important
to long-term maintenance of these species on the landscape as a whole. This conclusion
presents a challenge to conservation in the form of protecting both large
and small wetlands across the landscape. Searching for a lower area threshold for
occupancy by boreal birds is probably not a relevant exercise (Opdam 1991), and
conservationists and land managers in the Adirondacks must strive to make the case
for the maintenance of a complex of boreal wetlands across the park rather than assuming
that we can safely ignore the small ones. Seimlitsch and Bodie (1998) point
out the importance of small, isolated wetlands as critical for maintaining regional
biodiversity because they harbor very large numbers of species. This is the case in
the Adirondacks. The total number of species detected at least once in point counts
at boreal wetlands is 136—roughly 70% of the total known breeding bird species
richness for the park—in comparison to an average species richness of 70 on counts
conducted in upland forest sites (M.J. Glennon, unpubl. data). The majority of small
boreal wetlands may be of lower quality to birds, but they are likely to contain a
large fraction of the total individuals in a population (Pulliam 1988) simply because
there are many more of them than larger ones. These potential sink habitats can be
critical in maintaining long-term viability of populations (Davis and Howe 1992,
Doak 1995, Howe et al. 1991, Pulliam 1988, Pulliam and Danielson 1991).
Although maintaining a complex of large and small wetlands across the park is
challenging, there are also opportunities for conservation. Several of the variables
demonstrated here to be important in controlling bird dynamics in the Adirondacks
are unlikely to change in the near future. Area of the large peatlands is likely to
remain relatively consistent in the near future because a large proportion of these
“charismatic megawetlands” are protected in the park. Their elevation and latitude
will not change. What will change, and what land managers have the greatest potential
to influence, is human footprint and connectivity. Decreasing the likelihood
Northeastern Naturalist
NENHC-68
M.J. Glennon
2014 Vol. 21, No. 1
of human impacts nearby to boreal wetlands, and maintaining the smaller, isolated
fringe habitats that probably provide important stepping stones for boreal birds in
this landscape will probably best serve their long-term maintenance. Minimizing human
infrastructure in and near these areas will have the added benefit of reducing the
likelihood of invasion by synanthropic species with which these birds may compete.
Climate change may render the long-term persistence of these species in the park
uncertain. On shorter time scales, however, additional research to understand the importance
of specific human impacts, coupled with efforts to buffer these habitats and
to maintain their functional connectivity through protection and management will
benefit this iconic community of birds found nowhere else in the state.
Acknowledgments
I am indebted to numerous field technicians who have collected bird data for this
project including M. McCormack, B. McAllister, G. Lee, A. Belford, F. Latif, B. Keelan,
K. Jablonski, E. Obercian, N. Laviola, and V. Stein. A. Ross, J. Racette, B. Swift, J.
Jenkins, D. Spada, and R. Curran provided helpful information on boreal birds and habitats.
I am also very grateful to landowners and managers who permitted access to their
lands for the purpose of this work. This project has benefitted tremendously from funding
provided by the New York State Department of Environmental Conservation State
Wildlife Grants Program, the Northern New York Audubon Joseph and Joan Cullman
Foundation Grants Program, and L. Master.
Literature Cited
Adirondack Park Agency. 1987. State of New York Adirondack Park state land master plan.
Adirondack Park Agency, Ray Brook, NY.
Akaike, H. 1974. A new look at the statistical model identification. IEEE Transactions on
Automatic Control, AC-19:716–723.
Akçakaya, H.R., G. Mills, and C.P. Doncaster. 2006. The role of metapopulations in Conservation.
Pp. 64–84, In D. Macdonald and K. Service (Eds.). Key Topics in Conservation
Biology. Wiley-Blackwell, Oxford, UK. 328 pp.
Altman, B., and R. Sallabanks. 2012. Olive-sided Flycatcher (Contopus cooperi). No. 502,
In A. Poole (Ed.).The Birds of North America Online. Cornell Lab of Ornithology,
Ithaca, NY. Available online at http://bna.birds.cornell.edu/bna/species/502. Accessed
June 2013.
Anderson, M.G., M. Clark, C.E. Ferree, A. Jospe, A. Olivero Sheldon, and K.J. Weaver.
2013. Northeast habitat guides: A companion to the terrestrial and aquatic habitat maps.
The Nature Conservancy, Eastern Conservation Science, Eastern Regional Office. Boston,
MA. Available online at http://nature.ly/HabitatGuide. Accessed November 2013.
Andrle, R.F., and J.R. Carroll (Eds.). 1988. The Atlas of Breeding Birds in New York State.
Cornell University Press, Ithaca, NY. 551pp.
Bock, C.E., J.H. Bock, and B.C. Bennett. 1999. Songbird abundance in grasslands at a
suburban interface on the Colorado High Plains. Studies in Avian Biology 19:131–136.
Brommer, J.E. 2004. The range margins of northern birds shift polewards. Annales Zoologici
Fennici 41:391–397.
Brown, M., and J.J. Dinsmore. 1986. Implications of marsh size and isolation for marsh bird
management. Journal of Wildlife Management 50:392–397.
Northeastern Naturalist Vol. 21, No. 1
M.J. Glennon
2014
NENHC-69
Burnham, K.P., and D.R. Anderson. 2002. Model Selection and Multimodel Inference:
A Practical Information-theoretical Approach, Second Edition. Springer-Verlag, New
York, NY. 488 pp.
Cowardin, L.M., V. Carter, F.G. Golet, and E.T. LaRoe. 1979. Classification of wetlands and
deepwater habitats of the United States. US Department of the Interior Fish and Wildlife
Service, Office of Biological Services, Washington, DC. 131pp.
Davis, G.J., and R.W. Howe. 1992. Juvenile dispersal, limited breeding sites, and the dynamics
of metapopulations. Theoretical Population Biology 41:184–207.
DeVictor, V., R. Julliard, D. Couvet, A. Lee, and F. Jiguet. 2007. Functional homogenization
effect of urbanization on bird communities. Conservation Biology 21:741–751.
Doak, D.F. 1995. Source-sink models and the problem of habitat degradation: General models
and applications to the Yellowstone Grizzly. Conservation Biology 9(6):1370–1379.
George, T.L., and S. Zack. 2008. Bird occupancy and richness in Ponderosa Pine forests
with contrasting forest structure and fire history. Canadian Journal of Forest Research
38:936–942.
Glennon, M.J., and W.F. Porter. 2005. Effects of land-use management on biotic integrity:
An investigation of bird communities. Biological Conservation 126:499–511.
Glennon, M.J., and H.E. Kretser. 2013. Size of the ecological-effect zone associated with
exurban development in the Adirondack Park, NY. Landscape and Urban Planning
112(2013):10–17.
Greenberg, R., and S.M. Matsuoka. 2010. Rusty Blackbird: Mystery of a species in decline.
The Condor 114(4):770–777.
Hames, R.S., K.V. Rosenberg, J.D. Lowe, and A.A. Dhondt. 2001. Site reoccupation in
fragmented landscapes: Testing predictions of metapopulation theory. Journal of Animal
Ecology 70:182–190.
Hansen, A., R. Knight, J. Marzluff, S. Powell, K. Brown, P. Gude, and K. Jones. 2005.
Effects of exurban development on biodiversity: Patterns, mechanisms, and research
needs. Ecological Applications 15:1893–1905.
Hanski, I. 1998. Metapopulation dynamics. Nature 396:41–49.
Hines, J.E. 2006. Presence2: Software to estimate patch occupancy and related parameters.
USGS–PWRC, Laurel, MD.
Hines, J.E., J.D. Nichols, J.A. Royle, D.I. MacKenzie, A.M. Gopalaswamy, N. Samba
Kumar, and K.U. Karanth. 2010. Tigers on trails: Occupancy modeling for cluster sampling.
Ecological Applications 20:1456–1466.
Howe, R.W., G.J. Davis, and V. Mosca. 1991. The demographic significance of sink populations.
Biological Conservation 57:239–255.
Jenkins, J.J. 2010. Climate Change in the Adirondacks: The Path to Sustainability. Comstock
Publishing Associates, Cornell University Press, Ithaca, NY, and London, UK. 183 pp.
La Sorte, F.A., and F.R. Thompson. 2007. Poleward shifts in winter ranges of North American
birds. Ecology 88:1803–1812.
MacKenzie, D.I., J.D. Nichols, J.A. Royle, K.H. Pollock, L.L. Bailey, and J.E. Hines. 2006.
Occupancy Estimation and Modeling: Inferring Patterns and Dynamics of Species Occurrence.
Elsevier, Burlington, MA. 324 pp.
Maestas, J.D., R.L. Knight, and W.G. Gilgert. 2001. Biodiversity and land-use change in the
American Mountain West. The Geographical Review 91:509–524
McGowan, K., and K. Corwin. 2008. The Second Atlas of Breeding Birds in New York
State. Cornell University Press, Ithaca, NY. 688 pp.
Moore, P.D. 2002. The future of cool temperate bogs. Environmental Conservation
29(1):3–20.
Northeastern Naturalist
NENHC-70
M.J. Glennon
2014 Vol. 21, No. 1
Niemi, G., J. Hanowski, P. Helle, R. Howe, M. Mönkkönen, L. Venier, and D. Welsh. 1998.
Ecological sustainability of birds in boreal forests. Conservation Ecology 2(2):17. Available
online at http://www.consecol.org/vol2/iss2/art17. Accessed 26 September 2005.
Odell, E.A., and R.L. Knight. 2001. Songbird and medium-sized mammal communities associated
with exurban development in Pitkin County, Colorado. Conservation Biology
15:1143–1150.
Opdam, P. 1991. Metapopulation theory and habitat fragmentation: A review of holarctic
breeding bird studies. Landscape Ecology 5(2):93–106.
Parmesan, C. 2006. Ecological and evolutionary responses to recent climate change. Annual
Review of Ecology Evolution and Systematics 37:637–669.
Parmesan, C., and G. Yohe. 2003. A globally coherent fingerprint of climate change impacts
across natural systems. Nature 421:37–42.
Pastor, J., S. Light, and L. Sovell. 1998. Sustainability and resilience in boreal regions:
Sources and consequences of variability. Conservation Ecology 2(2):16. Available online
at http://www.consecol.org/vol2/iss2/art16. Accessed 26 September 2005.
Pulliam, H.R. 1988. Sources, sinks, and population regulation. American Naturalist
132(5):652–661.
Pulliam, H.R., and B.J. Danielson. 1991. Sources, sinks, and habitat selection: A landscape
perspective on population dynamics. American Naturalist 137:S50–S66.
Ralph, C.J., S. Droege, and J.R. Sauer. 1995. Managing and monitoring birds using point
counts: Standards and applications, Pp. 161–169, In C.J. Ralph, J.R. Sauer, and S.
Droege (Eds.). Monitoring bird populations by point counts. Technical Report PSWGTR-
149. USDA Forest Service, Berkeley, CA. 187 pp.
Sanderson, E.W., M. Jaiteh, M.A. Levy, K.H. Redford, A.V. Wannebo, and G. Woolmer.
2002. The human footprint and the last of the wild. Bioscience 52:891–904.
Sauer, J.R., G.W. Pendleton, and S. Orsillo. 1995. Mapping of bird distributions from pointcount
surveys. Pp. 151–160, In C.J. Ralph, J.R. Sauer, and S. Droege (Eds.). Monitoring
bird populations by point counts. Technical Report PSW-GTR-149. USDA Forest
Service, Berkeley, CA. 187 pp.
Sauer, J.R., J.E. Hines, J.E. Fallon, K.L. Pardieck, D.J. Ziolkowski, Jr., and W.A. Link.
2012. The North American breeding bird survey: Results and analysis 1966–2011.
Version 12.13.2011, USGS Patuxent Wildlife Research Center, Laurel, MD. Available
online at: http://www.mbr-pwrc.usgs.gov/bbs/bbs.html. Accessed 27 June 2013.
Scarl, J.C. 2012. Mountain Birdwatch 2011: Annual report to the United States Fish and
Wildlife Service. Vermont Center for Ecostudies, Norwich, VT. 20 pp.
Semlitsch, R.D., and J.R. Bodie. 1998. Are small, isolated wetlands expendable? Conservation
Biology 12(5):1129–1133.
Sjögren-Gulve, P., and I. Hanski. 2000. Metapopulation viability analysis using occupancy
models. Ecological Bulletins 48:53–71.
Smith, L.A., and P. Chow-Fraser. 2010. Impacts of adjacent land use and isolation on marsh
bird communities. Environmental Management 45:1040–1051.
Thomas, C.D., and J.J. Lennon. 1999. Birds extend their ranges northwards. Nature
399:213–213.
Villard, M.A., G. Merriam, and B.A. Maurer. 1995. Dynamics in subdivided populations of
neotropical migratory birds in a fragmented temperate forest. Ecology 76:27–40.
Virkkala, R., and A. Rajasärkkä. 2011. Climate change affects populations of northern birds
in boreal protected areas. Biology Letters 7:395–398.
Virkkala, R., R.K. Heikkinen, N. Leikola, and M. Luoto. 2008. Projected large-scale range
reductions of northern boreal land bird species due to climate change. Biological Conservation
141:1343–1353.
Northeastern Naturalist Vol. 21, No. 1
M.J. Glennon
2014
NENHC-71
Waite, T.A., and D. Strickland. 2006. Climate change and the demographic demise of a
hoarding bird living on the edge. Proceedings of the Royal Society B 273:2809–2813.
Woolmer, G., S.C. Trombulak, J.C. Ray, P.J. Doran, M.G. Anderson, R.F. Baldwin, A. Morgan,
and E.W. Sanderson. 2008. Rescaling the human footprint: A tool for conservation
planning at an ecoregional scale. Landscape and Urban Planning 87:42–53.
Zuckerberg, B., A.M. Woods, and W.F. Porter. 2009. Poleward shifts in breeding bird distributions
in New York State. Global Change Biology 15(8):1866–1883.
Zuckerberg, B., D.N. Bonter, W.M. Hochachka, W.D. Koenig, A.T. DeGaetano, and J.L.
Dickenson. 2011. Climatic constraints on wintering bird distributions are modified by
urbanization and weather. Journal of Animal Ecology 80:403–413.