2010 NORTHEASTERN NATURALIST 17(1):1–18
Moose Foraging in the Temperate Forests of
Southern New England
Edward K. Faison1,2,*, Glenn Motzkin1, David R. Foster1,
and John E. McDonald3
Abstract - Moose have recently re-colonized the temperate forests of southern New
England, raising questions about this herbivore’s effect on forest dynamics in the
region. We quantified Moose foraging selectivity and intensity on tree species in relation
to habitat features in central Massachusetts. Acer rubrum (Red Maple) and Tsuga
canadensis (Eastern Hemlock) were disproportionately browsed; Pinus strobus
(White Pine) was avoided. Foraging intensity correlated positively with elevation,
distance to development, and watershed type and negatively with time since forest
harvest, explaining 26% of the variation. Moose may interact with forest harvesting
to contribute to a decline in Red Maple and Eastern Hemlock and an increase
in White Pine in intensively browsed patches. Nonetheless, foraging impacts may
diminish over time, as increasing temperatures and sprawling development increasingly
restrict suitable Moose habitat.
Introduction
Large herbivores—often interacting with climate, fire, and human disturbance—
can profoundly shape vegetation communities (McNaughton 1988,
Zimov et al. 1995). Megafaunal effects are particularly strong in ecotones
where climate is conducive to more than one vegetation type and foraging
and trampling disturbances can shift vegetation toward one type or the other
(Scholes and Archer 1997, Zimov et al. 1995). In forested biomes, persistent
and rapid tree regeneration and low densities of browsing mammals
typically preclude large vegetation shifts by herbivores (e.g., from forest to
grassland; Forsyth 1985, Marks 1983); however, large mammals can still
have important effects on species composition and biodiversity, tree density,
successional pathways, and nutrient cycling in forests (Cote et al. 2004, Pastor
et al. 1988).
Since the late 1980s, Alces alces L. (Moose) have recolonized their prehistorical
range in southern New England (SNE) forests to approximately
the July 20 °C isotherm (Foster et al. 1998, 2002). Moose populations have
increased rapidly since their arrival and are currently estimated at 1000
resident animals in Massachusetts and Northern Connecticut (Massachusetts
Wildlife, Westborough, MA, unpubl. data; Connecticut Department
of Environmental Protection, Hartford, CT, unpubl. data). Recent range expansions
and recolonizations of Moose have also occurred in other parts of
1Harvard Forest, Harvard University, 324 North Main Street, Petersham, MA 01366.
2Current address - Highstead, PO Box 1097, Redding, CT. 3US Fish and Wildlife
Service, 300 Westgate Center Drive, Hadley, MA 01035-9589. *Corresponding author
- efaison@highstead.net.
2 Northeastern Naturalist Vol. 17, No. 1
North America, including the midwestern United States, the Rocky Mountain
States, and the coastal rain forests of British Columbia (Darimont et al.
2005, Karns 1997). Although the primary factors controlling many species’
southern range limits are competition and predation, temperature is the
primary factor determining the Moose’s southern limit (Brown et al. 1996,
Karns 1997).
Moose were relatively common in 17th-century Massachusetts, but uncommon
in Connecticut (Godin 1977, Trefethen 1953). Conditions at that
time contrasted strikingly with the present. Predators such as Canis lupus
L. (Gray Wolf), Puma concolor L. (Mountain Lion), and Native Americans
were relatively common; human settlements were primarily small and
dispersed; and temperatures were considerably colder under a Little Ice
Age climate (Cronon 1983, Foster et al. 2004, Godin 1977). Southern New
England disturbances were generally limited to localized surface fires, localized
forest clearing by Native Americans, and infrequent (100–150 year
intervals) canopy-replacing windstorms (Foster et al. 2004, Whitney 1994).
Forest cover in southern New England included greater amounts of Fagus
grandifolia Ehrhart (American Beech), Quercus spp. (oak)—especially Q.
alba L. (White Oak)—Castanea dentata (Marshall) Borkhausen (American
Chestnut), and lesser amounts of Pinus strobus L. (White Pine), Betula lenta
L. (Black Birch), and especially Acer rubrum L. (Red Maple) (Abrams 1998,
2006; Hall et al. 2002).
Fires have been actively suppressed in SNE since the early 20th century
(Whitney 1994). Since the early 1800s, average temperatures have warmed
by about 1.5 °C, a trend particularly pronounced in the winter months; and
human development has increased dramatically in the region, reducing the
suitable habitat for the boreal-adapted Moose (Foster et al. 2004, Keim and
Rock 2001). However, in the past several decades, widespread low-intensity
logging of sawtimber-sized White Pine and Quercus rubra L. (Red Oak) has
created patches of increased regeneration (Kittredge et al. 2003, McDonald
et al. 2006). Logging combined with the extirpation of a chief predator—
the Gray Wolf—and a ban on Moose hunting have potentially enhanced
the quality of the remaining habitat in this region, at least in the short term
(Telfer 1984, Thompson and Stewart 1997).
Due to the early extirpation of Moose from most of its temperate forest
range, there is almost no information about their habitat selection or foraging
behavior in the mixed hardwood-White Pine-Tsuga canadensis L. (Carriere)
(Eastern Hemlock) forests typical of the region. Our main objectives in this
study were: (1) to quantify the selectivity and intensity of Moose foraging
to anticipate possible impacts of this large herbivore on forest vegetation in
Central Massachusetts and (2) to identify habitat features influencing Moose
foraging activity at landscape and site scales. To address these goals, we
quantified species-level availability and use of woody plants in two large
forested areas and evaluated browse intensity in relationship to forest characteristics,
landscape features, and harvesting activity.
2010 E.K. Faison, G. Motzkin, D.R. Foster, and J.E. McDonald 3
Field Site Description
The Quabbin and Ware River Watershed forests in the Central Uplands
of Massachusetts were selected for investigation because these undeveloped
areas support the highest Moose densities in the region (an estimated 0.5
individuals/km² in Quabbin and 1.1 individuals/km² in Ware River; D. Clark,
Director of Natural Resources, Quabbin/Ware River Watershed, Belchertown,
MA, pers. comm.) and because they include extensive forest harvesting
and considerable variation in forest composition and habitat characteristics
that could influence browse patterns (Table 1, Fig. 1; McDonald et al. 2006).
Both watersheds are ≥10 km from a large urban area (Table 1).
Average temperatures within 10–15 km of the study area are -3.5 °C
in winter and 19.6 °C in summer (Harvard Forest, unpubl. weather station
data 2001–2005), values that exceed documented heat-stress thresholds
for Moose in the boreal forest by 1.5 °C in winter and 5.6 °C in summer
(Schwartz and Renecker 1997). Ursus americanus Pallas (Black Bear),
a chief predator of Moose calves, are present in both watershed forests
(Ballard and Van Ballenberghe 1997, Foster et al. 2002). Odocoileus virginianus
Zimmerman (White-tailed Deer), a potential competitor and disease
transmitter to Moose, occur at low densities (2–7 individuals/km²) in both
watersheds (Boer 1997, McDonald et al. 2007).
Quabbin Watershed Forest
The 22,000-ha Quabbin Watershed Forest (42°17'N, 72°21'W; Fig. 1) is
the largest tract of conservation land in southern New England (Kittredge et
al. 2003). Soils are primarily till-derived and acidic, and elevations range from
116 to 382 m, with a mean of 232 m. Oak-pine forests predominate with lesser
amounts of Eastern Hemlock, northern hardwoods, and conifer plantations
(Kyker-Snowman et al. 2007). Oaks comprise 31% of the forest’s basal area,
White Pine 28%, Red Maple 13%, Eastern Hemlock 9%, Black Birch 5%,
Fraxinus americana L. (White Ash) 4%, and Prunus serotina Ehrhart (Black
Cherry) <1% (Kyker-Snowman et al. 2007). The Quabbin forest is closed to
public vehicles and development, but has an extensive network of unpaved
roads. Moose colonized the watershed in about 1993 (B. Spencer, former chief
forester, Quabbin Watershed Forest, Belchertown, MA, pers. comm.).
Table 1. Important Moose habitat characteristics of two watershed forests in central Massachusetts
(Drawbridge et al. 2003, Kyker-Snowman et al. 2007)
Watershed characteristics Quabbin Ware River
Size (ha) 22,000 9500
Average elevation (m) 232 267
% conifer cover 22 23
% conifer swamp <1 4
% of land harvested since 1984 30 22
Relative size of patch clearcuts Smaller Larger
% open wetlands 3 6
Closest (km) large urban area (>150,000 people) 14 10
4 Northeastern Naturalist Vol. 17, No. 1
Figure 1. Central Massachusetts study area with 120 study sites (shown as white
circles with black dots). Inset is the location of the two study areas, the Quabbin (left)
and Ware River (right) Watershed forests, in southern New England. Data from US
EPA 2001 National Land Cover Data, Massachusetts Department of Conservations
and Recreation, and MassGIS.
2010 E.K. Faison, G. Motzkin, D.R. Foster, and J.E. McDonald 5
Ware River Watershed Forest
The 9500-ha Ware River Watershed Forest (42°25'N, 72°01'W; Fig. 1)
lies 10 km east of the Quabbin area. The glacial till-covered uplands are
interspersed with extensive valley outwash deposits and support oak-pine
forests with a large proportion of forested and open wetlands (Drawbridge
et al. 2003). White Pine accounts for 37% of the basal area, oaks 30%, Red
Maple 16%, Eastern Hemlock 7%, and Black Cherry 4% (Drawbridge et al.
2003). Elevations range from 190 to 365 m with a mean of 267 m. The Ware
River Watershed Forest is undeveloped but has a network of trails and unpaved
roads, many of which are open to public vehicular access (Drawbridge
et al. 2003). Moose colonized the Ware River about 1993 (H. Eck, chief
forester, Quabbin/Ware River Watershed, Belchertown, MA, pers. comm.).
Methods
Plot selection
We established plots following a stratified-random design, stratifying
by forest types and recent harvesting history. Three upland forest
types (hemlock, oak-pine, and northern hardwoods) and two wetland
forest types (swamp hardwoods and Picea spp. [spruce] - Abies balsamea
(L.) Miller (Balsam Fir) were identified on forest-type maps for each
watershed (Degraaf and Yamasaki 2001; Massachusetts Department of
Conservation and Recreation, Belchertown, MA, unpubl. data). Upland
areas harvested since 1984 were identified based on maps filed for all commercial
harvests in Massachusetts (McDonald et al. 2006). Eight harvested
and 8 unharvested sites were randomly selected in each forest type. In
addition, 8 oak-pine hilltops were sampled to broaden the range of topographical
positions. In cases where the vegetation and harvesting history in
the field did not match the mapped cover type, the site was sampled and the
designations corrected. To ensure independence of plots, sample locations
were separated by a minimum distance of 700 m, a distance greater than
the average daily movements of Moose in winter (Phillips et al. 1973). We
distributed 120 sample plots among the 2 watersheds in approximate proportion
to their forest area (Fig. 1).
Vegetation surveys
Each study site consisted of two 100-m² circular plots for sampling trees,
tall shrubs, and browsing activity. Species and DBH were recorded for all
trees >2.5 cm DBH, and all tall shrubs greater than 1.8 m high were recorded
by species. In a nested 10-m² circular subplot within each 100-m² plot, all
tree stems <2.5 cm DBH were recorded by species to determine seedling
density and composition (Degraaf and Yamasaki 2001, Higgins et al. 1996).
The centers of the two 100-m² plots were 30 m apart, and data from the 2
subplots were summed.
6 Northeastern Naturalist Vol. 17, No. 1
Moose foraging surveys
At each site, Moose foraging was assessed on trees and shrubs. To distinguish
Moose from White-tailed Deer foraging, we only recorded browse
above 1.8 m, the approximate height limit for year-round deer foraging
(Curtis and Sullivan 2001). Moose forage at all heights between 0–3 m,
but feeding trials have shown that, on >80% of deciduous tree species, they
consume most browse between 1.5 and 2.5 m (Bergstrom and Danell 1986).
We assumed that recording browse above 1.8 m would exclude most deer
browse but capture the predominant foraging activity of Moose; however,
we recognized that in recently harvested areas this sampling system would
not capture Moose browsing on low stump sprouts.
In each 100-m² plot, all woody stems with live twigs between 1.8 and
3.0 m and rooted in the plot were recorded by species, and presence/absence
of browsing was noted. Only stems that were unequivocally browsed
(i.e., torn and ragged) were recorded as “yes”, and no distinction was
made between recent and old browsing; thus, our browsing surveys likely
reflected a browsing history that extended back months and even years
(McInnes et al. 1992). Moose also strip bark from the trunks of saplings
and pole-sized trees and walk on or pull down small trees to browse the
nutritious leading shoots, breaking stems (Schwartz and Renecker 1997).
We noted saplings broken along the major stem (with browsed leading
shoots) and recorded bark-stripped stems >1.8 m high. Moose predominantly
feed on leaves of deciduous woody plants and aquatic vegetation in
summer and on twigs and bark in winter and early spring (Renecker and
Schwartz 1997). Therefore, our data likely captured predominantly late
fall to early spring foraging.
Foraging-intensity index
To quantify the level of impact by Moose at each site, we developed
a foraging-intensity index that combined three variables: browsing, bark
stripping, and stem breakage. Moose are capable of impacting woody
plants up to 6 cm DBH by pulling down and breaking the stem and by
browsing terminal shoots (Renecker and Schwartz 1997). Above 6 cm
DBH, trees are too large for Moose to pull down. Although Moose regularly
browse lower branches of larger trees, this likely has minimal impact
on the tree, and we therefore restricted our analysis of browse to stems ≤6
cm DBH. Moose can, however, impact larger trees by bark stripping and
are therefore able to affect forest communities from the seedling layer up
into the canopy (Miquelle and Van Ballenberghe 1989). With this in mind,
at each site, we summed the proportion of stems ≤6 cm DBH browsed, the
proportion of stems ≤6 cm DBH broken, and the proportion of trees >2.5
cm DBH bark-stripped. We then divided this sum by 3 in order to express
the index relative to 100. We limited index calculations to sites with ≥4
available stems per category to avoid spuriously high or low proportions
that might result from very small samples.
2010 E.K. Faison, G. Motzkin, D.R. Foster, and J.E. McDonald 7
Explanatory variables
Telfer (1984) outlined major habitat features influencing Moose distribution,
which include climate, landform, vegetation, and animal communities
(including humans). Based on these habitat features, we selected 6 predictor
variables (five continuous and one categorical) for analysis in relation to
Moose foraging intensity: elevation, solar radiation (hillshade), time since
forest harvest, distance to development and major roads, “watershed” (i.e.,
Ware River or Quabbin), and Eastern Hemlock sapling density. Elevation is
inversely related to winter temperature in central Massachusetts, and Moose
begin to show winter heat stress at temperatures similar to January averages
for central Massachusetts (Foster et al. 1998, Kanda et al. 2005, Schwartz
and Renecker 1997). Hillshade is derived from an analysis performed with
ArcGIS 9.2 to provide a proxy for relative radiation load using winter altitudes
and azimuths for the town of Barre, MA (US Naval Observatory 2008).
Time since harvest had a maximum value of 22 years because forest-cutting
plans were not filed prior to the 1983 Massachusetts Forest Cutting Practices
Act (McDonald et al. 2006). Distance to development/major roads was also
calculated with ArcGIS, and “developed” land included residential, industrial,
commercial, transportation, waste disposal, and recreation sites, but was
primarily residential development in our study area. “Major” roads did not
include unpaved forest roads. A maximum distance of 500 m was used for
this variable, as wildlife is generally influenced by development up to this
distance at a site scale (Duerksen et al. 1996). The categorical variable “watershed”
was related to broader landscape features relevant to Moose habitat,
including percentage of open wetlands, percentage of conifer swamps, and
relative size of patch clear cuts (Table 1). Lastly, Eastern Hemlock sapling
density reflected the number of Eastern Hemlock stems (>1.8 m high and ≤6
cm DBH) at each site.
Statistical analysis
Multiple linear regression was used to analyze foraging intensity (dependent
variable) in relation to six explanatory variables. We examined
correlation coefficients for the 6 variables, and none of them were significantly
correlated (R < 0.18, P > 0.08; Table 2), which increased the
Table 2. Correlation coefficients (R) of predictor variables entered into multiple regression and
stepAIC model.
Distance to Eastern
Harvest development Hemlock
Variable age /roads Watershed Hillshade Elevation saplings
Harvest age 0.015 0.127 0.009 -0.081 0.038
Distance to development/roads 0.069 0.012 -0.015 0.001
Landscape 0.081 0.178 0.145
Hillshade 0.077 0.041
Elevation 0.166
Eastern Hemlock saplings1
1Density of Eastern Hemlock saplings (≥1.8 m high, ≤6 cm DBH) at each site.
8 Northeastern Naturalist Vol. 17, No. 1
reliability of the final model (Gotelli and Ellison 2004, Graham 2003). To
further test the reliability of the coefficients obtained from the multiple
regression analysis, we performed a “stepAIC” (both directions) with
Akaike’s information criterion (AIC) to select the best model from the
predictor variables. We used chi-square tests to examine browse selectivity
among individual tree species. Despite using multiple comparisons, we
chose not to use Bonferroni adjustments, which are deemed excessively
conservative by Gotelli and Ellison (2004). We did, however, use Yates
continuity corrections for each test (Gotelli and Ellison 2004). Tree species
are presented as “selected” or “avoided” based on whether there was a
significant difference between their availability and use. We used analysis
of variance (ANOVA) to compare foraging intensities among the 5 forest
types. All data were arc-sine or log transformed to achieve normality, and
all analyses were performed with S-Plus 8.0 software (Insightful Corporation,
Seattle, WA).
Results
At sites where Moose sign was observed, the average percentage of stems
browsed was 36% at Ware River and 18% at Quabbin. Overall, only 3% of
our sites were intensively browsed (>75% of stems browsed). Bark stripping
occurred on 1% of trees >2.5 cm DBH and <30 cm DBH, and 91% of the
incidences occurred at Ware River sites. Red Maple accounted for 98% of
bark-stripped stems, and 2% of our sites were intensively stripped (>50%
of stems stripped). Stem breakage by Moose occurred on 1% of all small
saplings (2.5–6.0 cm DBH). Among major tree species (≤6.0 cm DBH) in
harvested sites, White Pine was avoided (P < 0.001), and Red Maple (P <
0.01) and total hardwoods (P < 0.001) were selected for browsing (Table 3).
In unharvested sites, Eastern Hemlock (P < 0.05) was selected, and White
Pine was avoided (P < 0.01; Table 4). Red Maple was selected for bark stripping
in both harvested and unharvested forests (Tables 3 and 4). Hemlock
saplings were broken in greater proportion to their availability in unharvested
areas and overall (P < 0.01; Tables 4 and 5), while Betula populifolia
Marsh (Gray Birch) saplings were disproportionately broken in harvested
sites (P < 0.001; Tables 3–5).
Foraging intensity
One hundred sites met the inclusion criteria of the foraging-intensity
index, and index values ranged from 0–0.40 across these sites. Five of the
100 sites were excluded from the subsequent multiple regression and stepAIC
analysis because of missing values. The multiple regression model
included three significant variables (P <0.05) that explained 26% of the
variation in foraging intensity: elevation (positive; P = 0.004), watershed
(Ware River; P = 0.02), and time since harvest (negative; P = 0.03); distance
to development was marginally significant (positive; P = 0.05). StepAIC
analysis produced a final model with the same four explanatory variables as
2010 E.K. Faison, G. Motzkin, D.R. Foster, and J.E. McDonald 9
predictors of Moose foraging intensity. Lastly, foraging intensities did not
differ across forest types (P = 0.84).
Discussion
Selective foraging and long-term forest dynamics
Moose selected Eastern Hemlock, Red Maple, and total hardwood
browse and avoided White Pine. Moose also stripped Red Maple saplings
and poles and broke Eastern Hemlock and Gray Birch saplings in greater
proportion to their availability (Tables 3–5). White Pine has increased in
the region since the time of European settlement (Hall et al. 2002), in part
as a result of selective browsing by domestic cattle in late 19th- and early
20th-century abandoned fields (Bromley 1935, Fisher 1918). Selective
browsing by Moose may continue to promote the increase of White Pine in
this landscape.
Red Maple has increased in the region as a result of 20th-century fire suppression,
selective harvesting of oaks, and perhaps increased precipitation
(Abrams 1998, Hall et al. 2002, Keim and Rock 2001). Selection of Red
Maple twigs and bark by Moose and reductions of Red Maple seedlings in
Table 3. Moose selection and avoidance of trees in harvested sites of both watershed forests.
Chi-square values indicate the degree of difference between used and available stems.
Unharvested sites n % used % available χ2 Choice4
Browsing1
Red Maple 120 30 17 16.7** S
Black Birch 176 25 25 0.01
Carpinus caroliniana Walt. 10 1 1 0.15
(American Hornbeam)
American Chestnut 12 3 2 1.50
White Ash 19 0.4 3 4.30
White Pine 212 9 31 43.1*** A
Black Cherry 58 14 8 5.40
Red Oak 28 4 4 0.12
Eastern Hemlock 14 3 2 0.31
Other trees 43 10 6 3.10
Hardwoods 465 89 67 39.1*** S
Conifers 227 11 33 39.1*** A
Bark stripping²
Red Maple 118 100 21 18.0*** S
Other trees 440 0 79 18.0*** A
Stem breakage³
Gray Birch 5 33 2 24.9*** S
Red Oak 16 33 6 7.50
Other trees 264 33 91 26.3*** A
¹Trees ≤6.0 cm DBH.
²Trees 2.5–29.0cm DBH.
³Trees 2.5–6.0cm DBH.
4Choice: S = selected, A = avoided.
*P < 0.05, **P < 0.01, ***P < 0.001.
10 Northeastern Naturalist Vol. 17, No. 1
the region by deer (Healy 1997) suggest that Moose browsing could limit
Red Maple in some areas. Winter browsing, captured by our sampling, is
generally less harmful to hardwoods like Red Maple than is summer browsing
(Canham et al. 1994), and bark-stripped trees often recover unless
severely girdled (>66% of circumference; Gill 1992). Although we did not
measure the percent circumference of bark wounds, we observed that most
Red Maples were damaged at a level below 66%. Nonetheless, an increased
incidence of disease and stem breakage can occur at the wound site of barkstripped
stems (Gill 1992, Miquelle and Van Ballenberghe 1989).
Like Balsam Fir in the boreal forests of eastern North America,
Eastern Hemlock appears to be an important winter browse species for
Moose in the temperate forests of SNE; Eastern Hemlock is also selected
by deer in this region and has declined in the past decade due to Adelges
tsugae Annand (Hemlock Woolly Adelgid) and associated salvage logging
(Kittredge and Ashton 1995, Orwig et al. 2002). Moose browsing
may contribute to this on-going decline and could have a bigger impact
on Eastern Hemlock than on Red Maple because of Eastern Hemlock’s
slower growth rate than hardwoods (Kelty 1986), the generally poorer
Table 4. Moose selection and avoidance of trees in unharvested sites of both watershed forests.
Chi-square values indicate the degree of difference between used and available stems.
Unharvested sites n % used % available χ2 Choice4
Browsing1
Balsam Fir 13 6 3 3.50
Red Maple 93 24 19 1.00
Betula alleghaniensis Britt. (Yellow Birch) 17 8 3 3.20
Black Birch 63 8 13 2.10
American Hornbeam 36 10 7 0.58
White Pine 135 9 28 15.30** A
White Oak 11 5 2 2.80
Red Oak 17 4 3 0.15
Quercus velutina Lam. (Black Oak) 16 3 3 0.001
Eastern Hemlock 35 17 7 9.80* S
Other trees 38 6 8 0.20
Hardwoods 295 68 60 1.80
Conifers 193 32 40 1.80
Bark stripping2
Red Maple 142 98 24 98.00*** S
American Chestnut 4 2 1 1.40
Other trees 441 0 75 107.70*** A
Stem breakage3
Red Maple 64 17 23 0.13
Red Oak 8 17 3 3.70
Eastern Hemlock 30 67 11 17.50** S
Other trees 177 0 66 10.00* A
1Trees ≤6.0 cm DBH.
2Trees 2.5–29.0cm DBH.
3Trees 2.5–6.0cm DBH.
4Choice: S = selected, A = avoided.
*P < 0.05, **P < 0.01, ***P < 0.001.
2010 E.K. Faison, G. Motzkin, D.R. Foster, and J.E. McDonald 11
ability of conifers to exhibit compensatory growth to browsing (Persson
et al. 2005), and Eastern Hemlock’s apparent susceptibility to stem breakage
(Miquelle and Van Ballenberghe 1989).
In the boreal regions of eastern North America, Moose browsing shows
some parallel patterns to SNE. One dominant conifer (Balsam Fir) and a
few deciduous species—Betula papyrifera Marshall (Paper Birch), Populus
tremuloides Michx. (Quaking Aspen), and Populus balsamifera L. (Balsam
Poplar)—are selected by Moose. A second major conifer (spruce) is avoided
(Pastor et al. 1988, Thompson et al. 1992). Selective and intensive foraging
in the boreal forest has shifted forest composition from early successional
hardwoods to late successional spruce in some areas, speeding up forest succession
and reducing available soil nitrogen (by increasing the proportion
of low-nutrient spruce litter that is returned to the soil; Pastor et al. 1988,
Thompson et al. 1992).
The successional implications of our foraging data are difficult to generalize.
Selection by Moose of Red Maple and early successional Gray Birch
in regenerating stands dominated by hardwoods may speed up succession
by allowing oaks to dominate the stand more quickly than they otherwise
would (Oliver and Larson 1996). However, selection of Eastern Hemlock
and Red Maple and avoidance of White Pine on very moist sites could
slow down succession, as Eastern Hemlock and Red Maple outgrow other
species at an earlier age in moist locations and persist as late-successional
Table 5. Moose selection and avoidance of trees in all sites of both watershed forests. Chisquare
values indicate the degree of difference between used and available stems.
All sites n % used % available χ2 Choice4
Browsing1
Red Maple 213 28 18 16.10** S
Black Birch 239 20 20 0.01
White Pine 347 9 29 57.20*** A
Black Cherry 71 10 6 5.40
Oak spp. 75 7 6 0.02
Eastern Hemlock 49 7 4 4.20
Hardwoods 760 83 64 38.00*** S
Conifers 420 17 36 38.00*** A
Bark stripping2
Red Maple 260 98 23 122.90*** S
Other trees 885 2 77 122.90*** A
Stem breakage3
Red Maple 121 9 21 0.99
Oak spp. 32 27 6 8.98
Eastern Hemlock 36 36 6 15.30** S
Other trees 383 27 69 7.60
1Trees ≤6.0 cm DBH.
2Trees 2.5–29.0cm DBH.
3Trees 2.5–6.0cm DBH.
4Choice: S = selected, A = avoided.
*P < 0.05, **P < 0.01, ***P < 0.001.
12 Northeastern Naturalist Vol. 17, No. 1
dominants (Oliver and Larson 1996, Spurr 1956). White Pine does occur as
a less frequent, late-successional species in Central Massachusetts on dry
soils (Spurr 1956). In contrast to the decline of soil fertility associated with
an increase in spruce in the boreal model, an increase in White Pine in SNE
would not necessarily result in a decline in soil fertility; soils beneath White
Pine stands can have higher nitrogen mineralization rates than those beneath
adjacent hardwood stands (Binkley and Valentine 1991).
Oaks, a valuable timber species in SNE and also producing the most important
food source (acorns) for eastern forest animals (McShea and Healy
2002, University of Connecticut Cooperative Extension Service 1994–
2008), were not selectively browsed, bark stripped, or broken by Moose.
Oak seedlings, however, declined sharply in the Quabbin Reservation from
deer browsing during the 20th century (Healy 1997). Relative abundance
of overstory oaks has declined across the region and most of the eastern
United States since pre-settlement times, perhaps due to fire suppression and
increased precipitation (Abrams 2006, Keim and Rock 2001). Oak recruitment
into the sapling layer is currently very low in many areas in the eastern
United States, but in southern New England, oak sapling densities are higher
today than they were two decades ago (Abrams 2006, Moser et al. 2006).
Foraging intensity and effect on forests
When projecting impacts by moose to temperate forests of southern
New England, it is instructive to examine how different foraging intensities
by Moose and other herbivores have affected plant communities in other
regions. In the boreal forests of Isle Royale, MI, tree biomass production
was related to past browsing intensity. Vegetation at a site with heavy past
browsing (76% of woody stems browsed) was substantially altered—e.g.,
unpalatable spruce were the only trees growing above browse height, while
Balsam Fir and Paper Birch were suppressed (McInnes et al. 1992, Pastor et
al. 1988). In contrast, the tree species mix at a site with moderate past browsing
(51% of stems browsed) was not changed.
Shifts in community composition have been noted when >50% of the
stems of a particular tree species were bark stripped by herbivores such as
White-tailed Deer and Cervus elaphus L. (Elk) (Miquelle and Van Ballenberghe
1989). Only 3% of our sites matched the >75% browsing intensity
associated with strong impacts to Isle Royale forests, and only 2% of our
sites had >50% Red Maple stems stripped. Thus, our data imply that Moose
(at current population densities) may have serious impacts on forests at only
scattered localities in SNE.
Interactions with forest harvesting
Foraging intensity was unrelated to forest type, but was correlated with
time since forest harvesting. Moose are attracted to early seral stages of
forests (5–15 years) because of high stem densities, large shoots produced
by stump sprouting, and high quality forage resulting from increased soil
nitrogen mineralization stimulated by fire and forest harvesting (Pastor et
2010 E.K. Faison, G. Motzkin, D.R. Foster, and J.E. McDonald 13
al. 1988, Rea and Gillingham 2001, Renecker and Schwartz 1997). Had we
sampled browsing <1.8 m in height (and been able to distinguish Moose
from deer browsing), we would have captured intensive browsing of low
stump sprouts in recent harvests and perhaps strengthened the relationship
between foraging intensity and time since harvest. This sampling limitation
may account for some of the 74% unexplained variation in our multiple regression
model.
With 25–30% of the forest land in central Massachusetts harvested
between 1984–2003 and White Pine growing with Red Maple and other
hardwoods in many regenerating cut blocks, the current rate of logging
could largely determine spatial pattern and extent of Moose impacts to forest
composition (McDonald et al. 2006). In the boreal zone, landscape-scale
fires and large clearcuts (Delong and Tanner 1996, Hunter 1993) produce
extensive areas of abundant regeneration that can reduce the impact of even
very high Moose densities (Pastor et al. 1988, Peterson 1995). The small
scale and intensity of Massachusetts harvests (average 15 ha; Kittredge et
al. 2003) suggests that regeneration in SNE forests may be more susceptible
to Moose impacts than in boreal forests.
Controls over landscape-level patterns
Foraging intensity was related to elevation and specific watershed features
and suggests that in addition to recent forest harvests, thermoregulation
and year-round availability of forage away from human settlements may be
important controls over Moose foraging. Several aspects of the Ware River
area appear to be favorable for Moose. A higher fraction of the watershed
consists of coniferous forested wetlands than the Quabbin, providing Moose
with cool locations during the summer and abundant tall-shrub browse (Pastor
et al. 1988; Table 1). The Ware River also has twice the percentage of
open wetlands than the Quabbin; these wetlands supply abundant aquatic
plants (Table 1). Despite the Quabbin’s greater proportion of harvested
forests, the Ware River has historically been managed with larger patch
clearcuts, creating larger areas of concentrated browse that allow Moose to
forage with less energy expenditure (Telfer 1984).
The relationship between foraging intensity and elevation suggests that
topography and thermoregulation could limit the ability of Moose to impact
forests at lower elevations in the region and ultimately to spread south
into the lowlands of southern New England. Microclimatic studies indicate
that temperatures in the region decrease approximately 0.7–1.3 °C over an
equivalent elevation rise of our study area (Kanda et al. 2005, Ross 1958),
a significant amount given that Moose in the boreal forest begin to exhibit
heat stress in winter at temperatures (-5 °C) that are below winter averages
in central Massachusetts (-3.5 °C; Harvard Forest, unpubl. weather station
data 2001–2005; Schwartz and Renecker 1997). With some climate models
predicting a rise in regional temperatures of 3.1–5.3 °C over the next
century and the loss of forestland in Massachusetts and Connecticut accelerating
over the past 30 years (Foster et al. 2005, Hurtt and Hale 2001),
14 Northeastern Naturalist Vol. 17, No. 1
much (if not all) of the region could be rendered unsuitable for Moose by
the end of the century. Suitable Moose habitat may become restricted to
remote, elevated areas with abundant swamps and open wetlands, reducing
the extent of their impacts to forest dynamics in this region.
Moose are unlikely to be limited by predation in SNE, as only one of
their three chief predators, Black Bear, currently inhabits the region. Potential
recolonization habitat for the Gray Wolf in the northeastern United
States is limited to northern New England and New York (Mladenoff and
Sickley 1998).
Acknowledgments
This paper is a contribution from the Harvard Forest Long-term Ecological
Research Program. Funding for this project was provided by the National Science
Foundation, the Anna B. Bliss Fund and Living Diorama Scholarship Fund, and the
Harvard University Program for Research and Training in Ecology from the Andrew
W. Mellon Foundation. Special thanks to Aaron Ellison for statistical assistance.
Thanks also to Todd Fuller, Bruce Spencer, Herm Eck, Derek Beard, Thom Kyker-
Snowman, Dan Clark, and Bill Woytek for helpful correspondence and to Josh Rapp
for field assistance.
Literature Cited
Abrams, M.D. 1998. The Red Maple paradox. Bioscience 48:355–364.
Abrams, M.D. 2006. Ecological and ecophysiological attributes and responses to fire
in eastern oak forests. Pp. 74–89, In M.B. Dickinson (Ed.). Fire in Eastern Oak
Forests: Delivering Science to Land Managers. General Technical Report NRSP-
1. US Department of Agriculture, Forest Service, Northern Research Station.
Newtown Square, PA. 303 pp.
Ballard, W.B., and V. Van Ballenberghe. 1997. Predator/prey relationships. Pp. 247–
273, In A.W Franzmann and C.C. Schwartz (Eds.). Ecology and Management of
the North American Moose. Smithsonian Institution, Washington, DC. 733 pp.
Bergstrom, R., and K. Danell. 1986. Moose winter feeding in relation to morphology
and chemistry of six tree species. Alces 22:91–112.
Binkley, D., and D. Valentine. 1991. Fifty-year biogeochemical effects of Green Ash,
White Pine, and Norway Spruce in a replicated experiment. Forest Ecology and
Management 40:13–25.
Boer, A.H. 1997. Interspecific relationships Pp. 337–349, In A.W Franzmann and
C.C. Schwartz (Eds.). Ecology and Management of the North American Moose.
Smithsonian Institution, Washington, DC. 733 pp.
Bromley, S.W. 1935. The original forest types of southern New England. Ecological
Monographs 5:61–89.
Brown, J.H., G.C. Stevens, and D.M. Kaufman. 1996. The geographic range: Size,
shape, boundaries, and internal structure. Annual Review of Ecology and Systematics
27:597–623
Canham, C.D., J.B. McAninch, and D.M. Wood. 1994. Effects of the frequency, timing,
and intensity of simulated browsing on growth and survival of tree seedlings.
Canadian Journal of Forest Research 24:817–825.
Cote, S.D., T.P. Rooney, J.P. Tremblay, C. Dussault, and D.M. Waller. 2004. Ecological
impacts of deer overabundance. Annual Review of Ecology, Evolution, and
Systematics 35:113–147.
2010 E.K. Faison, G. Motzkin, D.R. Foster, and J.E. McDonald 15
Cronon, W. 1983. Changes in the Land: Indians, Colonists, and the Ecology of New
England. Hill and Wang, New York, NY. 241 pp.
Curtis, P.D., and K.L. Sullivan. 2001. Wildlife damage management fact sheet:
White-tailed Deer. Cornell Cooperative Extension, Ithaca, NY.
Darimont, C.T., P.C. Paquet, T.E. Reimchen, and V. Crichton. 2005. Range expansion
by Moose into coastal temperate rainforests of British Columbia, Canada.
Diversity and Distributions 11:235–239.
DeGraaf, R.M., and M. Yamasaki. 2001. New England Wildlife: Habitat, Natural
History, and Distribution. University Press of New England, Hanover, NH.
482 pp.
Delong, S.C., and D. Tanner. 1996. Managing the pattern of forest harvest: Lessons
from wildfire. Biodiversity and Conservation 5:1191–1205.
Drawbridge, S., H. Eck, B. Spencer, D. Clark, T. Kyker-Snowman, P. Church, M.
Fluet, M. Hopkinson, D. Small, J. French, and J. Zimmerman. 2003. Ware River
Watershed Land Management Plan 2003–2012. Massachusetts Department of
Conservation and Recreation Division of Water Supply Protection, Office of
Watershed Management. Boston, MA. 205 pp.
Duerksen, C. J., N.T. Hobbs, D.L. Elliot, E. Johnson, and J.R. Miller. 1996. Managing
Development for People and Wildlife: A Habitat Protection Handbook for
Local Governments. Clarion Associates, Denver, CO. 122 pp.
Fisher, R.T. 1918. Second-growth White Pine as related to the former uses of the
land. Journal of Forestry 16:253–254.
Forsyth, A. 1985. Mammals of the American North. Camden House Publishing,
Camden East, ON, Canada. 351 pp.
Foster, D.R., G. Motzkin, and B. Slater. 1998. Land-use history as long-term broadscale
disturbance: Regional forest dynamics in Central New England. Ecosystems
1:96–119.
Foster, D.R., G. Motzkin, D. Bernardos, and J. Cardoza. 2002. Wildlife dynamics in
a changing landscape. Journal of Biogeography 29:1337–1357.
Foster, D., G. Motzkin, J. O’Keefe, E. Boose, D. Orwig, J. Fuller, and B. Hall. 2004.
The environmental and human history of New England. Pp. 43–100, In. D.R.
Foster and J.D. Aber (Eds.). Forests in Time. Yale University Press, New Haven,
CT. 477 pp.
Foster, D., D. Kittredge, B. Donahue, G. Motzkin, D. Orwig, A. Ellison, B. Hall,
B. Colburn, and A. D’Amato. 2005. Wildlands and woodlands: A Vision for the
forests of Massachusetts. Harvard Forest, Petersham, MA. 24 pp.
Gill, R.M.A. 1992. A review of damage by mammals in north temperate forests: 3.
Impact on trees and forests. Forestry 65:363–388.
Godin, A.J. 1977. Wild Mammals of New England. Johns Hopkins University Press,
Baltimore, MD. 304 pp.
Gotelli, N.J., and A.M. Ellison. 2004. A Primer of Ecological Statistics. Sinauer
Associates, Sunderland, MA. 510 pp.
Graham, M.H. 2003. Confronting multicollinearity in ecological multiple regression.
Ecology 84:2809–2815.
Hall, B., G. Motzkin, D.R. Foster, M. Syfert, and J. Burk. 2002. Three hundred years
of forest and land-use change in Massachusetts, USA. Journal of Biogeography
29:1319–1335.
Healy, W.M. 1997. Influence of deer on the structure and composition of oak forests
in central Massachusetts. Pp. 249–266, In W.J. McShea, H.B. Underwood, and
J.H. Rappole (Eds.). The Science of Overabundance. Smithsonian Institution,
Washington, DC. 402 pp.
16 Northeastern Naturalist Vol. 17, No. 1
Higgins, K.F., J.L. Oldemeyer, K.J. Jenkins, G.K. Clambey, and R.F. Harlow. 1996.
Vegetation sampling and measurement. Pp. 567–591, In T.A. Bookhout (Ed.).
Research and Management Techniques for Wildlife and Habitat. Allen Press,
Lawrence, KS. 740 pp.
Hunter, M.L. 1993. Natural fire regimes as spatial models for managing boreal forests.
Biological Conservation 65:115–120.
Hurtt, G., and S. Hale. 2001. Future climates of the New England Region. Pp.
26–31, In The New England Regional assessment Group (Eds.). Preparing For a
Changing Climate: The New England Regional Assessment Overview. US Global
Change Research Program, University of New Hampshire, Durham, NH. 96 pp.
Kanda, L.L., T.K. Fuller, P.R. Sievert, and K.D. Friedland. 2005. Variation in winter
microclimate and its potential influence on Virginia Opossum (Didelphis virginiana)
survival in Amherst, Massachusetts. Urban Ecosystems 8:215–225.
Karns, P.D. 1997. Population, distribution, density, and trends. Pp. 125–139, In A.W
Franzmann and C.C. Schwartz (Eds.). Ecology and Management of the North
American Moose. Smithsonian Institution, Washington, DC. 733 pp.
Keim, B., and B. Rock. 2001. The New England region’s changing climate. Pp.
8–17, In The New England Regional assessment Group (Eds.). Preparing For a
Changing Climate: The New England Regional Assessment Overview. US Global
Change Research Program, University of New Hampshire, Durham, NH. 96 pp.
Kelty, M.J. 1986. Development patterns in two hemlock-hardwood stands in southern
New England. Canadian Journal of Forest Research 16:885–891.
Kittredge, D.B., and P.M.S. Ashton. 1995. Impact of deer browsing on regeneration
in mixed stands in southern New England. Northern Journal of Applied Forestry
12:115–120.
Kittredge, D.B., A.O. Finley, and D.R. Foster, 2003. Timber harvesting as ongoing
disturbance in a landscape of diverse ownership. Forest Ecology and Management
180:425–442.
Kyker-Snowman, T., D. Clark, H. Eck, J. French, D. Morin, R. Stone, M. Fluet, S.
Campbell, D. Small, P. Lyons, and J. Zimmerman. 2007. Quabbin Reservoir Watershed
system: Land management plan. 2007–2017. Massachusetts Department
of Conservation and Recreation Division of Water Supply Protection, Office of
Watershed Management. Boston, MA. 341 pp.
Marks, P. 1983. On the origin of the field plants of the Northeastern United States.
American Naturalist 122:210–228.
McDonald, J.E., D.E. Clark, and W.A. Woytek. 2007. Reduction and maintenance of
a White-tailed Deer herd in central Massachusetts. Journal of Wildlife Management
71:1585–1593.
McDonald, R.I., G. Motzkin, M.S. Bank, D.B. Kittredge, J. Burk, and D.R. Foster.
2006. Forest harvesting and land-use conversion over two decades in Massachusetts.
Forest Ecology and Management 227:31–41.
McInnes, P.F., R.J. Naiman, J. Pastor, and Y. Cohen. 1992. Effects of Moose browsing
on vegetation and litter of the boreal forest, Isle Royale, Michigan, USA.
Ecology 73:2059–2075.
McNaughton, S.J. 1988. Large mammals and process dynamics in African ecosystems.
Bioscience 38:794–800.
McShea, W.J., and W.M. Healy. 2002. Oaks and acorns as a foundation for ecosystem
management. Pp. 1–9, In W.J. McShea and W.M. Healy (Eds.). Oak Forest
Ecosystems: Ecology and Management for Wildlife. Johns Hopkins University
Press. Baltimore, MD. 432 pp.
2010 E.K. Faison, G. Motzkin, D.R. Foster, and J.E. McDonald 17
Miquelle, D.G., and V. Van Ballenberghe. 1989. Impact of bark stripping by Moose
on aspen-spruce communities. Journal of Wildlife Management 53:577–586.
Mladenoff, D.J., and T.A. Sickley. 1998. Assessing potential Gray Wolf restoration
in the Northeastern United States: A spatial prediction of favorable habitat and
potential population levels. Journal of Wildlife Management 62:1–10.
Moser, W.K., M. Hansen, W. McWilliams, and R. Sheffield. 2006. Oak composition
and structure in the Eastern United States. Pp. 49–61, In M.B. Dickinson (Ed.).
Fire in Eastern Oak Forests: Delivering Science to Land Managers. General
Technical Report NRS-P-1. Department of Agriculture, Forest Service, Northern
Research Station. Newtown Square, PA. 303 pp.
Oliver, C.D., and B.C. Larson. 1996. Forest Stand Dynamics. John Wiley and Sons,
New York, NY. 520 pp.
Orwig, D.A., D.R. Foster, and D.L. Mausel. 2002. Landscape patterns of hemlock
decline in New England due to the introduced Hemlock Woolly Adelgid. Journal
of Biogeography 29:1475–1487.
Pastor, J., R.J. Naiman, B. Dewey, and P. McInnes. 1988. Moose, microbes, and the
boreal forest. Bioscience 38:770–777.
Persson, I.L., K. Danell, and R. Bergstrom. 2005. Different Moose densities and
accompanied change in tree morphology and browse production. Ecological Applications
15:1296–1305.
Peterson, R.O. 1995. The Wolves of Isle Royale: A Broken Balance. Willow Creek
Press, Minocqua, WI. 189 pp.
Phillips, R.L., W.E. Berg, and D.B. Siniff. 1973. Moose movement patterns and range
use in northeastern Minnesota. Journal of Wildlife Management 37:266–278.
Rea, R.V., and M.P. Gillingham. 2001. The impact of the timing of brush management
on the nutritional value of woody browse for Moose, Alces alces. Journal
of Applied Ecology 38:710–719.
Renecker L.A., and C.C. Schwartz. 1997. Food habits and feeding behavior. Pp. 403–
439, In A.W. Franzmann and C.C. Schwartz (Eds.). Ecology and Management of
the North American Moose. Smithsonian Institution, Washington, DC. 733 pp.
Ross, P. 1958. Microclimatic and vegetational studies in a cold-wet deciduous forest.
Harvard Black Rock Forest Paper No. 24. Cornwall-on-the-Hudson, NY. 89 pp.
Scholes, R.J., and S.R. Archer 1997. Tree-grass interactions in savannas. Annual
Review of Ecology and Systematics 28:517–544.
Schwartz, C.C., and L.A. Renecker. 1997. Nutrition and energetics. Pp. 441–478,
In A.W. Franzmann and C.C. Schwartz (Eds.). Ecology and Management of the
North American Moose. Smithsonian Institution, Washington, DC. 733 pp.
Spurr, S.H. 1956. Forest associations in the Harvard Forest. Ecological Monographs
26:245–262.
Telfer, E.S. 1984. Circumpolar distribution and habitat requirements of Moose.
Pp.145–181, In R. Olson, R. Hastings, and F. Geddes (Eds.). Northern Ecology
and Resource Management. University of Alberta, Edmonton, AB, Canada.
456 pp.
Thompson, I.D., and R.W. Stewart. 1997. Management of Moose habitat. Pp. 377–
401, In A.W Franzmann and C.C. Schwartz (Eds.). Ecology and Management of
the North American Moose. Smithsonian Institution, Washington, DC. 733 pp.
Thompson, I.D., W.J. Curran, J.A. Hancock, and C.E. Butler. 1992. Influence of
Moose browsing on successional forest growth on Black Spruce sites in Newfoundland.
Forest Ecology and Management 47:29–37.
18 Northeastern Naturalist Vol. 17, No. 1
Trefethen, J.B. 1953. The Massachusetts land and its wildlife: A history of the
resident and migratory game birds and mammals of Massachusetts. M.A. Thesis.
University of Massachusetts, Amherst, MA.
University of Connecticut Cooperative Extension Service. 1994–2008. Southern
New England stumpage price survey results. Compiled by University of Massachusetts,
University of Connecticut, and the state Forestry Agencies of Connecticut,
Massachusetts, and Rhode Island. Available online at http://www.canr.
uconn.edu/ces/forest/pricesht.htm. Accessed 10 December 2008.
United States Naval Observatory. 2008. Naval Oceanography Portal. Available online
at http://aa.usno.navy.mil/data/docs/RS_OneDay.php. Accessed December
5, 2008.
Whitney, G.G. 1994. From Coastal Wilderness to Fruited Plain. Cambridge University,
Cambridge, UK. 488 pp.
Zimov, S.A., V.I. Chuprynin, A.P. Oreshko, F.S. Chapin, J.F. Reynolds, and M.C.
Chapin. 1995. Steppe-tundra transition: A herbivore driven biome shift at the end
of the Quaternary. American Naturalist 146:765–793.