Classifying Patterns of Understory Vegetation
Mixed-Oak Forests in Two Ecoregions of
Pennsylvania
•
•
Melanie J. Kaeser, Peter J. Gould, Marc E. McDill, Kim C. Steiner, and James C. Finley
Understory plots in mixed-oak stands in the Ridge and Valley and Allegheny Plateau ecoregions of Pennsylvania were classified to summarize the range of
understory conditions common to the region and to evaluate relationships between understory vegetation, overstory conditions, and advance tree regeneration.
Factor analysis and duster analysis were used to identify understory classes. Vegetation cover data including 15 variables (plant species, bare ground, and rocky
ground cover) and 4 composite variables (plant species groups) were observed on 1,208 204-m2 plots. The eight vegetation classes resulting from the analyses
were associated with different overstory compositions and advance regeneration densities. The understory classes identified in this study may provide insight
on the distribution and diversity of problematic and nonproblematic vegetation types.
Keywords: vegetation classification, duster analysis, mixed-oak forests, oak regeneration
orest understory vegetation has long been recognized as rele­
vant to forest management decisions because of irs potential
to reflect site quality and indicate forest productivity (Ca­
jander 1926, Coile 1938, Daubenmire 1976). The influence of
understory vegetation on desirable tree regeneration is a major bio­
logical and economic concern in the eastern United Stares, and
elsewhere, because it often interferes with the establishment of tree
regeneration and suppresses growth of desirable tree seedlings
(Horsley 1986, Horsley 1991, Moser er al. 1996, George and Bazzaz
1999). Interfering understory vegetation in eastern forests could be
described as a secondary consequence of intense browsing by white­
tailed deer ( Odocolieus virginianus Boddaert; Trumbull et a!. [1989],
Horsley er al. [2003]). White-tailed deer inhibit the growth and
survival of desirable tree seedlings by browsing, which promotes the
expansion of less palatable competing vegetation such as ferns,
grasses, and trees such as striped maple (Marquis and Brenneman
1981, Horsley et al. 2003). These conditions result in decreased
plant species diversity, which affects community composition and
forest stand structure (Kittredge and Ashton 1990).
Other factors also have contributed to the development of prob­
lematic understory vegetation. Wildfires, which likely limited the
expansion of some understory species in past centuries, have been
nearly eliminated from eastern forests since the early 20th century
(Abrams 1992, Brose et al. 2001, Shumway er al. 2001). Selective
harvesting in more recent decades has also stimulated understory
development (Fredericksen et al. 1999). These and other factors
contribute to widespread regeneration problems in the eastern
U nired States, particularly in oak (Quercus spp.) stands (Lorimer
1993). The particular role of understory vegetation in rhe oak re­
generarion problem is difficult to ascertain, in part because under,
F
story compositions and structure are poorly documented at a re­
gional level.
Although the negative impacts of understory vegetation on tree
regeneration have been widely reported, forest understory vegeta­
tion also has positive management values. Dibble er al. (1999) found
some understory species to be useful indicators of favorable condi­
tions for red spruce (Picea rubens Sarg.) regeneration in Maine. On
harsh sites, understory vegetation can facilitate tree regeneration
establishment by protecting seedlings from frost and browse damage
(Buckley et al. 1998). Kirschbaum and Anacker (2005) have pro­
posed using preferentially browsed understory vegetation as an in­
dicator of deer impact in northeastern forests for management pur­
poses. Additionally, understory communities are more responsive
than overstories to recent disturbances and continuous stressors
(Halpern 1989, Laughlin et a!. 2004). Understanding these rela­
tionships can provide forest managers with more derailed knowl­
edge of forest ecosystems.
Rather than viewing understory vegetation simply as a hindrance
to desirable tree regeneration, there is a potential to incorporate both
their positive and negative values into management decisions. Un­
derstanding these values is particularly important in low-intensity
silviculrural systems (sensu Nyland 2002, p. 35) that rely on natural
regeneration to restock harvest stands. Such systems are commonly
applied in the mixed hardwood forests of the eastern United
States. Understories composed of often-complex mixtures of
nontree vegetation and tree regeneration are the source of repro­
duction in these systems. Understory composition at the time of
harvest determines, in large part, the later composition of regen­
erated stands (Gould et al. 2005). In preparation for regeneration
harvests, forest managers must cultivate understories that have
Received January 11, 2007; accepted May 16, 2007.
Melanie Kaeser (mkaeser@jonesctr.org), Peter Gould (pgould@fi.fodus), Marc McDill {meml4@psu.edu), Kim Steiner {steiner@psu.edu}, and james Finley (jj4@psu.edu}, The
Pennsylvania State University, University Park, PA 16802. Funding for this project was provided in part by the Pennsylvania Department ofComervation and Natural Resources
(DCNR) Bureau ofForestry.
Copyright © 2008 by the Sociery of American Foresters.
38
NORTH. J. APPL. FOR. 25(1) 2008
Table 1. Variables included in the analyses and used as input to
the factor analysis by stratum.
Stratum 0
Rocks
Stratum I
Hayscenred fern
Bracken fern
Huckleberry
Blueberry
Red maple
Witch-hazel
Striped maple
Mountain-laurel
Bare
Oaks
Figure 1. Study area consisting of 45 mixed-oak stands in Pennsylvania
by ecoregion.
the capacity to reproduce well-stocked stands dominated by de­
sirable species. Given understory management's imporrance to
successful silviculture in the eastern United States, there is a need
to better understand understory vegetation and relate it to other
parts of the forest, namely, overstory trees and tree regeneration.
The specific objectives of this study were (i) to summarize un­
derstory vegetation patterns in central Pennsylvania mixed-oak for­
ests and (ii) to characterize any significant relationships among these
vegetation patterns and overstory composition, tree regeneration
size, and densiry. Numerous studies have focused on the influence of
problematic understory vegetation on tree regeneration for manage­
ment purposes. This study differs from others in that we take a
broad, multispecies approach and evaluate both positive and nega­
tive values of understory vegetation. The identification of under­
story vegetation classes can help managers understand links among
understory vegetation, overstory composition, and regeneration.
Methods
Study Area
The study area included 1,208 plots in 45 mixed-oak stands
distributed in the Allegheny Plateau and Ridge and Valley ecore­
gions of central Pennsylvania (Figure 1; Bailey [1994]). All stands
were located on Pennsylvania State Forest lands managed by the
Pennsylvania Department of Conservation and Natural Resources
(DCNR) Bureau of Forestry. The total area of the 45 stands used in
our study was 837 ha with stands ranging in size from 6 to 34 ha.
The mixed-oak forest rype is dominant in the Ridge and Valley
ecoregion of central Pennsylvania and in the southern portion of the
Allegheny Plateau ecoregion in western and central Pennsylvania
(Ward 1983). Oaks were the dominant overstory species on our
stands, comprising 56% of the overstory basal area on average, rang­
ing from 32 to 76%. Elevation ranged from 262 m above mean sea
level (msl) in the Ridge and Valley ecoregion to 683 m above msl on
the Allegheny Plateau. Site index ranged from 48 to 79 on these
stands.
Data Collection
Data were collected between 1996 and 2002 as part of a long­
term study focused on regeneration on Pennsylvania State Forest
lands.[1] Depending on stand area, 15-30 204-m2 circular perma­
nent plots were systematically established in a square grid to cover
each stand. Species and dbh were recorded for overstory trees at the
plot level. Each plot contained four permanent 4-m2 circular sub­
plots established in each of the four cardinal directions. Tree regen-
Conifers
Stratum 2
Striped maple
Red maple
Blackgum
Mountain-laurel
Witch-hazel
Oaks
Conifers
Rocky ground cover
D. punctilobula (Mich:x.) Moore
P. aquilinum (L.) Kuhn
G. baccata (Wang.) K. Koch
Vaccinium spp.
A. rubrum L.
H. virginiana L.
A. pemylvanicum L.
K latifolia L.
Absence of vascular plants
Q. rubra L., Q. velutina Lam., Q. coccinea Muenchh.,
Q. montana L., Q. alba L.
P. strobus L., T canadensis (L.) Carr.
A. pensylvanicum L.
A. rubrum L.
N sylvatica Marsh.
K latifolia L.
H. virginiana L.
Q. rubra L., Q. velutina Lam., Q. coccinea Muenchh.,
Q. montana Willd. Q. alba L.
P. strobus L., T canadensis (L.) Carr.
eration stem counts were recorded by species at the 4-m2 subplot
level. Ocular estimation was used to record the percentage cover of
understory vegetation in two strata: stratum 1 (0 -152 em) and
stratum 2 (152-610 em; Table 1; Steiner et al. [2002]). Percentage
cover of bare ground was estimated in stratum 1 only. Understory
cover by oak species (Quercus rubra L., Quercus alba L., Quercus
montana Willd., Quercus velutina Lam., and Quercus coccinea
Muenchh.) was combined into an "oak" class and cover by conifer­
ous species (Pinus strobus L. and Tsuga canadensis [L.] Carr.) was
combined into a "conifer" class. Regeneration size classes included
tree seedlings less than 30 em (small seedlings) and 30-152 em
(large seedlings) in height and were recorded as the number of tree
seedlings by species. We combined tree regeneration stem counts
and percentage vegetation, bare ground, and rocky ground cover in
each class over the four subplots to provide an estimate for the
204-m2 plot level for subsequent analyses.
Statistical Analyses
Principal factor analysis and hierarchical cluster analysis were
used to identify vegetation classes (Hair et al. 1992, McGarigal et al.
2000). Principal factor analysis was used to reduce the number of
original variables into a smaller set of compound dimensions (Hair
er al. 1992). The original data matrix consisted of 1,208 plots X 19
variables (plant species or species groups, bare ground, and rocky
ground cover). Several understory herbaceous species were dropped
from the analysis due to their low occurrence in the data set. Seven
factors were retained from the principal factor analysis and were
grouped using cluster analysis with the factor scores. Ward's mini­
mum-variance linkage was used for the fusion method (Ward
1963). Local peaks in the pseudo F-statistic and a peak or larger
t2-statistic between dendrogram nodes were evaluated to determine
the number of vegetation classes (McGarigal et al. 2000).
Differences in tree seedling regeneration densities, understory
vegetation, and overstory compositions were evaluated between veg­
etation classes using analysis of variance. Mean densities of regener­
ation by oaks, red maple (Acer rubrum L.; the most common regen­
eration species), and other species combined were compared using
NORTH. J. APPL FoR. 25(I) 2008
39
Table 2.
labels.
Names of the vegetation classes and corresponding
Vegetation class
Abbreviation
Sparse understory and midstory vegetation
Blueberry low shrub understory
Hayscenred fern---<lominated understory
Mountain-laurel tall shrub understory
Huckleberry low shrub understory
Witch-hazel-dominated subcanopy/shrub understory
Striped maple-dominated deciduous subcanopy
Coniferous subcanopy
sv
BB
HF
ML
HB
WH
SM
co
two height classes (less than 30 em and 30 -152 em). Total overstory
stocking was computed for each plot using Gingrich's equation
(Gingrich 1967). Modified importance values (Curtis and Mcin­
tosh 1950, 1951) were calculated for common overstory tree species
by averaging relative density and relative basal area and expressing it
as a percentage. Mean values for total overstory stocking and relative
importance values for common species were compared across clus­
ters. Duncan's multiple range tests were used to obtain pairwise
comparisons among percentage vegetation cover means, tree regen­
eration density means, relative importance value means, and over­
story stocking means across the eight vegetation classes. SAS version
8.2 (SAS Institute, Inc., 2001, p. 3) was used to perform all multi­
variate and univariate analyses.
Results
The seven factors retained in the factor analysis accounted for
93% of the variance among the original variables. Our cluster anal­
ysis resulted in eight vegetation classes that were named based on the
dominant vegetation in each class (Table 2). With eight classes, 59%
of the total variation occurred between classes and 41o/o occurred
within the classes (Figure 2).
Description of Understory Vegetation Classes
Sparse Understory and Midstory Vegetation (SV)
The SV understory class contained the greatest number of plots
(22%), of which the majority was located in the Ridge and Valley
ecoregion (Table 3). These plots were found in 41 stands and this
class was the dominant class in 10 stands in the Ridge and Valley.
On average, this vegetation class had a very low percentage of vege­
tative cover in stratum 1 (82% was "bare") and low vegetative cover
in stratum 2 (Table 4).
0.125
s
� 0.100
'
Blueberry Low Shrub Understory (BB)
The BB understory class accounted for 21 o/o of plots and was
similarly common on the Allegheny Plateau and in the Ridge and
Valley ecoregions (Table 3). Plots in this class were found in 41
stands and it was the dominant class in 7 stands in the Ridge and
Valley and 3 stands on the Allegheny Plateau. Average vegetative
cover in stratum 1 was composed primarily of blueberry (Vaccinium
spp.), with an average cover of 20.7o/o (Table 4). Minor components
of stratum 1 included red maple, huckleberry ( Gaylussacia baccata
[Wang.] K. Koch), and bracken fern (Pteridium aquilinum [L.]
Kuhn). The abundant blueberry and smaller amounts of huckle­
berry formed an ericaceous low shrub layer. Red maple cover was
8.9% in stratum 2 and was the major constituent of this stratum.
Hayscented Fern-Dominated Understory (HF)
The HF understory class was identified on 17% of the total
number of plots and was present on 25 stands. This class was dom­
inant on five stands on the Allegheny Plateau and five stands in the
Ridge and Valley ecoregions. Plots in this vegetation class were
characterized by a dense hayscented fern (Dennstaedtia punctilobula
[Michx.] Moore) layer, covering 46% of stratum 1 on average (Ta­
ble 4). The plots in this class had low overall cover in stratum 2.
Mountain-Laurel Tall Shrub Understory (ML)
The ML understory class comprised 15% of the plots and was
found on 27 stands (Table 3). Plots in this class were dominant on
two stands on the Allegheny Plateau and four stands in the Ridge
and Valley ecoregions. The plots in this vegetation class were dom­
inated largely by ericaceous shrubs, with mountain-laurel (Kalmia
latifolia L.) as the dominant species, with an average cover of 33.9%
in stratum 1 (Table 4). Mountain-laurel was found growing in very
dense thickets on some plots, with coverage as high as 100%. Blue­
berry and huckleberry were smaller components of this vegetation
class in stratum 1. Stratum 2 cover was patchy and consisted pri­
marily of red maple, blackgum (Nyssa sylvatica Marsh.), and
mountain-laurel.
Huckleberry Low Shrub Understory (HB)
The HB understory class included 11o/o of plots and was present
on 27 stands (Table 3). This class was the dominant class on five
stands, all in the Ridge and Valley ecoregion. Plots in this class were
characterized by a low, ericaceous shrub layer in the understory.
Huckleberry was the major shrub species in this vegetation class,
comprising over one-half of the average vegetative cover in stratum
1 (28.7%; Table 4). Smaller components of stratum 1 included
blueberry and mountain-laurel. Blackgum was well represented in
stratum 2, and red maple and oak species were smaller components
of this stratum.
t 0.07�
I
� 0.000 i---+--+-+----+---+--1:..,_-+---1s
.
r
0.025
Figure 2. Dendrogram of eight clusters resulting from Ward's minimum­
variance linkage fusion depicting the plots on the x-axis and dissimilarity
on the y-axis.
40
NORTH. J. APPL. FOR. 25(1) 2008
Witch-Hazel-Dominated Subcanopy!Shrub Understory (WH)
Eight percent of plots were grouped into the WH understory
class, all of which were located in the Ridge and Valley ecoregion
(Table 3). Plots in this class were found on 24 stands but was the
dominant class on only 3 stands all located in the Ridge and Valley.
The WH class was characterized by shrubs and small trees, primarily
mountain-laurel, blueberry, and witch-hazel (Hamamelis virginiana
L.; Table 4). Witch-hazel was the principal species in stratum 2,
contributing on average 14.3% vegetative cover. Red maple and
blackgum were smaller components of stratum 2.
Table 3.
Frequency and percentage of plots by vegetation class and ecoregion.
Vegetation class
Ridge and Valley
Allegheny Plateau
Total
Table 4.
sv
BB
222
24%
34
12%
256
22%
179
20%
74
25%
253
21%
ML
HB
WH
SM
co
Total
125
14%
51
17%
176
15%
128
14%
6
2%
134
11%
101
II%
0
0%
101
8%
50
5%
3
1%
53
4%
22
2%
5
2%
27
2%
916
HF
89
10%
119
41%
208
17%
292
1208
Average percentage cover by species and vegetation class.
Class
Frequency
Stratum 0
Rocks
Stratum 1
Striped maple
Red maple
Witch-hazel
Mountain-laurel
Blueberry
Bracken fern
Huckleberry
Hayscenred fern
Oak spp.
Coniferous spp.
Bare
Stratum 2
Striped maple
Red maple
Witch-hazel
Blackgum
Mountain-laurel
Oak spp.
Coniferous spp.
Orher
sv
256
11.8b
BB
253
5.1'
0.2b
l.Ob
0.2'
2Y
5.5d
0.6b
1.6'·d
l.l'
1.4'
1.1b
82.3"
0.2b
8.2"
0.2'
0.8'
20.7"
3.0"
r.d
2.2b.
1.3'
3.9b
0.2'
51.5d
0.1 b
9.7"
0.5'
4.3b
0.1b
8.9"·b
0.2'
2.6b.c
0.3'
1.3'
2.2b
5.9"·b
0.1'
3.6"·b
0.2'
n. .r
3.8 b
HF
208
ML
176
2.6'
4.1'
0.2b
0.7b
0.1'
1.8'
3.1 d
l.Ob
0.3d
45.6"
1.0'
0.1'
42.9'.]
0.1b
0.7b
0.3'
33.9"
12.8b
l.3b
4.2b
1.8'
1.3'
0.1'
4t.3'f
0.2b
2.2b
0.3'
3.5'
10.7b.r
0.4b
28.7"
0.6'
6.6"
0.2'
44.5'
0.3b
7.3"·b
0.4'
5.5 b
0 .4b
7.8''·b
0.4'
10.2"
c
0.7b.
4.7"
2.1b .c
0.1'
2.6'
SM
101
134
11.8b
0.1b
10.0"
0.1'
0.7'
0.5'
1.0'
0.1'
2.7'
WH
HB
11.8b
0.5b
2.1b
4.7"
11.7b
53
24.5"
9.6"
1.8b
0.4'
1.8'
4.5d
o.ob
l. Td
5.3b
9.2'
0.1b
r.d
2.1b.
.r
2.9b
2.6b'
1.5b
2.1'
0.1'
66.7b
59.3'
J.Ob
7.2"-b
14.3"
4.9b
c
l.Ob.
.c
b
2.2
2.3b
n. .c
4.2 b
4.7"
0.3'
2.7'
16.0"
9.8"
0.5'
3.3b.c
0.1'
1.6'
0.0'
6.2"
co
P value
27
5.9'
<0.0001
O.lb
l.3b
1.3b
12.7b
r
9.9b.
0.4b
.r
3.9b
5.6b
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
1.2'
22.9"
38.?1
0.1b
5.8b
3.0b
2.2b.r
1.4b
4.3"
26.5"
3.3b.c
<0.0001
0.0447
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
Means with the same letter are nor significant at the 0.05 level.
Striped Maple-Dominated Deciduous Subcanopy (SM)
Four percent of plots were grouped into the SM understory class
and were found on 16 stands (Table 3). This class was dominant on
rwo stands in the Ridge and Valley ecoregion. Striped maple (Acer
pensylvanicum L.) was the dominant species characterizing this class,
accounting for 9.6% of the total cover in stratum 1 and 16% in
stratum 2. Blueberry and hayscenred fern were smaller components
of stratum 1. Rocky ground cover was significantly more abundant
in this class than in any other vegetation class, comprising 25% of
the total ground cover on average.
pared with all other classes, with stocking values of 82 and 85%,
respectively.
Modified importance values calculated for overstory tree species
are summarized in Table 5. Northern red oak and red maple had
significantly greater importance values in the HF class than in any
other class (P
0.0274 and P
0.0275, respectively). Conversely,
chestnut oak was not important in the HF class, but it was fairly
important in the other classes, especially the HB class. White pine
and hemlock were significantly more important in the overstory of
the CO class than in all other classes (P < 0.0001)
Coniferous Subcanopy (CO)
The CO vegetation class was characterized by coniferous tree
species in the understory and was the smallest class, occurring on 2%
of plots (Table 3). Plots in this class occurred on 13 stands, but this
class was not the dominant class on any stands. Total cover was high
in strata 1 and 2, with white pine and eastern hemlock covering 23%
of stratum 1 and 27% of stratum 2 on average (Table 4). Moun­
tain-laurel and blueberry were also well represented in stratum 1.
Regeneration
Relationships with Overstory Stocking
Toral overstory stocking percent was similar for all vegetation
classes except the SV class. The plots in the SV class had an average
of 102% stocking, which was significantly greater than the other
vegetation classes (P < 0.0001). Although not significantly differ­
em, the HF and ML classes had the lowest overstory stocking com-
=
=
Statistically significant differences were detected among average
oak, red maple, and "other" species regeneration[2] densities across
the eight vegetation classes (Table 6). The HF and ML classes had
relatively high densities of small oak seedlings (less than 30 em),
averaging 20,5 10 and 19,175 stems/ha, respectively. However, the
rwo classes had the lowest densities of large oak seedlings (30 -152
em), averaging 1 16 and 7 17 stems/ha, respectively. The HB and BB
understory classes had, by a large margin, the greatest large oak
seedling densities compared with the other vegetation classes. The
BB understory class had the highest density of both small and large
red maple seedlings, with 70,217 and 10,660 stems/ha, respectively.
In contrast, the CO vegetation class had the lowest number of small
and large red maple seedlings (26,702 and 388 stems/ha, respec­
tively). The SM and WH understory classes had the greatest number
NORTH.]. A.PPL. FOR. 25(1) 2008
41
5.
Table
Modified importance values for canopy tree species across vegetation classes.
Vegetation class
Species
sv
BB
HF
Red oak
Chestnut oak
Black oak
White oak
Scarlet oak
Red maple
Blackgum·
White pine
Hemlock
Other spp.
Sum of Oak spp.
22"·b
19a,b
5"
c
8b.
19n,b,c
17"·b
8"
IIa.b
2"
22b
5"· b
3b
1b
13b
4"
21b
4b
2b
ob
14b
24"
13b
6"
b,c
5
2"
31"
4b
Ib
ob
i4b
56
59
50
ML
HB
WH
SM
c
17"'b,
22"·b
8"
,c
JO"·b
c
13b,
29"
c
I I b.
25"·b
c
16"·b.
20"· b
r
6"
!2"·b
3"
20b
b
4
3b
lb
15b
57
6"
2'
I"
25"·b
3b
Ib
ob
b.c
7
2"
22b
9"
b
I
ob
c
IOb.
58
3"
22b
8"·b
Ib
lb
c
8 b,
60
26"
45
co
Pvalue
10'
20"·b
2"
17"
4"
17b
b
4
15"
5"
6'
53
0.0274
0.0548
0.6567
0.0679
0.6818
0.0275
0.0324
<0.0001
<0.0001
0.0014
Means in the same row followed by the same lerrer are nor significantly differenr at rhe 0.05 level.
Table 6.
Average oak, red maple, and other seedling densities (stems/hal by height (em) and vegetation class.
Vegetation class
Oaks
Red maple
Other
Total
Height (em)
sv
BB
<30
30-152
<30
30-152
<30
30-152
<30
30-152
14,807"-b
726'·"
63,642"·b
722b
14,797"-b
b
2,745
70,217"
10,660"
5,995'·"
959'
91,009
14,364
7,875'
c
1,055b,
86,324
2,503
HF
20,510"
116"
53,795"b
b
420
3,776"
534'
78,080
1,070
ML
19,175"
717'· "
c
43,770b.
558b
3,348"
707'
66,293
1,982
HB
20,206"
4,502"
c
43,819b,
1,960b
d
5,090'·
778'
69,115
7,240
WH
SM
co
13,519".1'
1,507'
bc
48,363"·
1,347b
11,609b
1,040b,c
9,536b
d
I,329'·
6!,870"·b
1,270b
8,238b
798'·d
73,492
3,894
15,990"
4,097"
87,396
6,697
26,702'
388b
3,798"
l,784b
P value
0.0015
<0.0001
0.0002
<0.0001
<0.0001
<0.0001
38,739
2,970
Means followed by the same lener are nor significanrly differenr at rhe 0.05 level.
of seedlings in the "other" tree regeneration category, averagtng
15,990 and 1 1,609 stems/ha, respectively.
Discussion
Understory vegetation is an important bur relatively little studied
component of eastern mixed-oak forests. The understory classes
outlined in this study appear to be associated with the composition
and density of overstory trees and advance tree regeneration. The
interactions among understory vegetation, overstory composition,
and tree regeneration likely are closely related to regeneration suc­
cess and future forest composition (George and Bazzaz 1999).
The dense understory vegetation, with relatively little midstory
cover, that characterizes the HF and ML vegetation classes appears
problematic for the development of advance regeneration. Under­
story vegetation does not appear to inhibit seedling establishment in
these classes, but it appears to restrict subsequent growth. Small
advance oak regeneration densities in the HF and ML vegetation
classes were among the highest of the eight classes, but densities of
large seedlings were the lowest. The expansion of fern and suppres­
sion of advance oak regeneration has been partially attributed to
large deer populations (Steiner and Joyce 1999). Competition from
fern and excessive deer browse likely has a cumulative impact on the
height growth and survival of oak regeneration. Additionally, Mess­
ier et al. ( 1989) found that an abundance of hayscented fern in forest
understories suppresses desirable tree seedlings by decreasing light
quantity and quality for overtopped seedlings beneath the herba­
ceous layer. Light levels beneath mountain-laurel canopies are also
very low and have been reported as low as approximately 2% of full
sunlight, which is well below optimal levels for seedling growth
(Chapman 1950). Patterns are also evident for red maple regenera­
tion and regeneration of other species. Fern-dominated understories
appear to be an indicator of and perhaps a causal factor in the
42
NORTH. J. APPL. FOR. 25(1) 2008
succession from oak dominance to red maple dominance. In this
study, the HF vegetation class was associated with overstory north­
ern red oak and red maple and occurred more frequently in the
Allegheny Plateau ecoregion versus the Ridge and the Valley ecore­
gions. Similar stands in this ecoregion were found to regenerate to
red maple-dominated stands after silviculrural clearcutring (Gould
et al. 2005). This shift in composition follows a regional pattern of
succession from oaks to more shade-tolerant species (Lorimer 1993,
Abrams 1998).
In contrast to the other understory classes, plots in the SV class
have failed to develop understory and midstory vegetation. Two
factors that may be responsible for the lack of understory vegetation
include intense browsing by white-tailed deer and low light levels
caused by high overstory stocking. Marquis and Brenneman ( 1981)
noted that very high deer populations are capable of virtually elim­
inating understory vegetation, creating a distinct "browse line."
Deer populations in Pennsylvania have been near historic high levels
for several decades and have had a profound impact on forest vege­
tation (Marquis and Brenneman 1981, Horsley er al. 2003). Addi­
tionally, the SV class has a greater amount of overstory stocking
compared with any other vegetation class. Thus, unlike the HF and
ML classes, oak seedlings in the SV class a.re not suffering from the
shading associated with dense understory vegetation, but instead
may be affected by greater shading from the overstory. It is possible
that SV sites might have been dominated by dense understory veg­
etation, as did HF and ML sires, if overstory shade levels had nor
prevented the growth of fern and mountain-laurel. In this case,
reducing overstory stocking in the SV class could quite possibly
result in invasion by fern, mountain-laurel, or other aggressive un­
desirable vegetation, resulting in even greater management chal­
lenges. The influence of external factors (i.e., deer) and understory
light levels should be explored further to explain the deficiency of
advance regeneration and understory vegetation in this vegetation
class.
Ericaceous low shrub communities (HB and BB vegetation
classes) differ markedly in understory compositions and tree regen­
eration compositions from the other classes. Large oak seedlings are
more abundant in these classes than in any other vegetation class.
Advance oak regeneration in these ericaceous low shrub understory
classes could be favored by the short height attained by these species
compared with other understory vegetation rypes. Ericaceous low
shrub species (i.e., blueberry and huckleberry) generally do not grow
as tall and dense as hayscented fern and mountain-laurel and hence
do not form a closed understory canopy that creates extremely low
light levels in the understory. Additionally, ericaceous communities
have an affiniry for infertile sites with well-drained acidic soils (Rog­
ers 1974). The development of advance oak regeneration is favored
on poor, dry sites because there is less interference from the herba­
ceous and woody vegetation that thrive on high-qualiry sites
(Gottschalk 1983). The combination of abundant oak in the over­
story, lower site qualiry, and fairly high levels of advance regenera­
tion in the BB and HB vegetation classes suggest that these condi­
tions contribute to a relatively stable oak ecosystem.
The SM, WH, and CO classes are structurally similar, with rel­
atively high levels of midstory (152- 610 em above the forest floor)
vegetative cover. The shading associated with midstory forest vege­
tation appears less problematic to regeneration than the low shade of
the HF and ML classes (Oliver and Larson 1990). Moderate densi­
ties of large and small advance oak regeneration,and regeneration of
other tree species were found in these classes. Midstory composi­
tions vary between classes, with the SM and WH classes containing
shade-tolerant trees char reach a relatively small stature at maturiry.
The CO class appears ecologically unique in mixed-oak stands be­
cause of midstory conifer cover and an association with overstory
eastern white pine and, to a lesser degree, eastern hemlock. Often,
conifers are poorly represented but valued for structural diversiry in
mixed-oak forests, and usually are not viewed as a management
obstacle. The presence of striped maple in the SM class is a manage­
ment concern in some areas because of irs abiliry to respond aggres­
sively to resource availabiliry after disturbance (Hibbs et a!. 1980).
When striped maple is present in the understory before cutting, it
frequently becomes the dominant vegetation rype after cutting, re­
sulting in the suppression or exclusion of desirable tree species
(Horsley and Bjorkbom 1983). It is essential to consider the trajec­
tory of regeneration in chis class, as in other classes, after disturbance
to ensure that striped maple does not aggressively expand and dom­
inate the overstory.
IdentifYing understory vegetation classes is an important step
toward better integrating understory conditions into forest manage­
ment decisions in the central Appalachians. Our plot-level classifi­
cation serves as a tool to sharpen our focus on the oak regeneration
problem by providing insight on the distribution and diversiry of
problematic and nonproblematic vegetation rypes. This broad mul­
tiple-species approach can help managers to consider those condi­
tions that contribute to oak regeneration failures and to recognize
conditions that favor oak regeneration. Forest managers can use this
information to enhance their understanding of relationships be­
tween the overstory, understory, and regeneration, while research
addressing the oak regeneration problem continues. Understanding
the trajectories of the understory classes after stand disturbance is
imperative in developing silvicultural recommendations aimed at
regenerating mixed-oak stands. By considering these trajectories in
future studies, conditions that are unfavorable for oak regeneration
can be identified and oak regeneration failures may be prevented.
Endnotes
[I] Penn State's Oak Regeneration Project is a long-term study funded by the
Pennsylvania DCNR Bureau of Forestry.
[2] Species in the "other" category include A. pensylvanicum, Amelanchier spp., N
sylvatica, Sassaftas albidum, Prunus serotina, P. strobus, Fraxinus americana,
Carya glabra, Robinia pseudoacacia, Prunus pensylvanica, Betula spp., Betula
lenta, Crataegus spp., Acer saccharum, Carya cordiformis, Carya tomentosa, Cor­
nus florida, Fagus grandifolia, Liriodendron tulipiftra, Ostrya virginiana, Picia
abies, Pinus rigida, Tilia americana, T canadensis, Castanea dentata, Populus
grandidentata, Carpinus caroliniana, Magnolia acuminata, and Populus tremu­
loides.
Literature Cited
ABRAMS, M.D.
1992.
Fire and the development of oak forests. BioScience
42:346-353.
M.D. 1998. The red maple paradox. BioScience 48:355-363.
United States. Map (1:7,500,000).
USDI Geological Survey, Reston, VA.
BROSE, P., T. SCHULER, D. VAN LEAR, AND J. BERST. 2001. Bringing fire back: The
changing fire regimes of the Appalachian mixed-oak forests. J For.
ABRAMS,
BAILEY, R.G. 1994. Ecoregions and subregions ofthe
99(11):30-35.
D.S., T.L. SHAJUK, AND J.G. ISEBRANDS. 1998. Regeneration of northern
red oak: Positive and negative effects of competitor removal. Ecology 79:65-78.
CAJANDER, A.K. 1926. The theory of forest types. Acta For. Penn. 29(3):1-108.
CHAPMAN, G.L. 1950. The influence of mountain-laurel undergrowth on
environmental conditions and oak reproduction. Ph.D. dissertation, Yale Univ.,
New Haven, CT. 157 p.
COILE, T.S. 1938. Forest classification: classification of forest types with special
reference to ground vegetation.] For. 36:1062-1066.
CURTIS, J.T., AND R.P. MciNTOSH. 1950. The inter-relations of certain analytic and
synthetic phytosociological characters. Ecology 31:434-455.
CURTIS, ].T., AND R.P. MC INTOSH. 1951. An upland forest continuum in the
prairie-forest border region of Wisconsin. Ecology 32:476-496.
DAUBENMIRE, R. 1976. The use of vegetation in assessing the productivity of forest
lands. Bot. Rev. 42:115-143.
DIBBLE, A.C., ].C. BRJSSETrE, AND M.L. HUNTER]R. 1999. Putting community data
to work: Some understory plants indicate red spruce regeneration habitat. For.
Ecol. Manag. 114:275-291.
FREDERJCKSEN, T.S., B.D. ROSS, W. HOFFMAN, M.L. MORRJSON, AND J. BEYEA.
1999. Short-term understory plant community responses to timber-harvesting
intensity on non-industrial private forestlands in Pennsylvania. For. Ecol. Manag.
BUCKLEY,
116:129-139.
L.O., AND F.A. BAZZAZ. 1999. The fern understory as an ecological filter:
Growth and survival of canopy-tree seedlings. Ecology 80(3):833-845.
GINGRJCH, S.F. 1967. Measuring and evaluating stocking and stand density in
upland hardwood forests in the Central States. For. Sci. 13:38-53.
GOTTSCHALK, K.W. 1983. Management strategies for successful regeneration: oak­
hickory. P. 190-213 in Proc. of Regenerating hardwood stands, Finley, ]., R.S.
Cochran, and J.R. Grace (eds.). The Pennsylvania State Univ., University Park,
PA.
GOULD, P.J., K.C. STEINER, ].C. FINLEY, AND M.E. MCDILL. 2005. Development
pathways following the harvest of oak-dominated stands. For. Sci. 51(1):76-90.
HAlR, J.F. JR., R.E. ANDERSON, R.L. TATHAM, AND W.C. BLAC K. 1992. Multivariate
data analysis with readings. Macmillan, New York. 544 p.
HALPERN, C.B. 1989. Early successional patterns of forest species: Interactions of life
history traits and disturbance. Ecology 70:704-720.
HIBBS, D.E., B.F. WILSON, AND B.C. FISCHER. 1980. Habitat requirements and
growth of striped maple (Acer pensylvanicum L.). Ecology 6! (3):490-496.
HORSLEY, S.B. 1986. Evaluation of hayscented fern interference with black cherry.
Am. j Bot. 73:663-669.
HORSLEY, S.B. 1991. Using roundup and oust to control interfering understories in
Allegheny hardwood stands. P. 281-290 in Proc. of the 8th Central hardwood
forest conf, Mar. 4-6, 1991, McCormick, L.H., and K.W. Gotrschalk (eds.).
NEFES Gen. Tech. Rep. NE-148. The Pennsylvania State Univ., University
Park, PA.
HORSLEY, S.B., AND].C. BJORKBOM. 1983. Herbicide treatment of striped maple and
beech in Allegheny hardwood stands. For. Sci. 29(1):103-112.
HORSLEY, S.B., S.L. STOUT, AND D.S. DECALESTA . 2003. White-tailed deer impact
on the vegetation dynamics of a northern hardwood forest. Ecol. Applic.
GEORGE,
13(1):98-118.
C.D., AND B.L. ANACKER. 2005. The utility of Trillium and
Maianthemum as phytoindicators of deer impact in northwestern Pennsylvania,
USA. For. Ecol. Manag. 217:54-66.
KIRSCHBAUM,
NORTH.
J. APPL.
FOR. 25(1) 2008
4J
KnTREDGE, D.B. JR., AND P.M.S. AsHTON. 1990. Natural regeneration patterns in
even-aged mixed stands in southern New England. North ]. Appl. For.
7 :163-168.
LAUGHLIN, D.C., J.D. BAKKER, M.T. STODDARD, M.L. DANIELS, J.D. SPRINGER,
C.N. GILDAR, A.M. GREEN, AND W.W. COVINGTON. 2004. Towards reference
conditions: Wildfire effects on flora in an old-growth ponderosa pine forest. For.
Ecol. Manag. 199:137-152.
LORIMER, C.G. 1993. Causes of the oak regeneration problem. P. 14-39 in Proc.
Symp. oak regeneration: Serious problems, practical recommendations, Sept. 8-10,
1992, Knoxville, TN. US For. Serv. Gen. Tech. Rep. SE-84. US For. Serv.,
North Central For. Exp. Srn., Sr. Paul, MN.
MARQUIS, D.A., AND R. BRENNEMAN. 1981. The impact ofdeer onforest vegetation in
Pennsylvania. US For. Serv. Gen. Tech. Rep. NE 65. 7 p.
McGARIGAL, K., S. CUSHMAN, AND S. STAFFORD. 2000. Multivariate statistics for
wildlife and ecology research. Springer, New York. 281 p.
MESSIER, C.T., T.W. HONER, AND J.P. I<JMMINS. 1989. Photosynthetic photon flux
density, red:far-red ratio, and minimum requirement for survival of Gaultheria
shallon in western red cedar-western hemlock stands in coastal British Columbia.
Can.]. For. Res. 19:1470-1477.
MOSER, K.W., M.J. DUCEY, AND P.M.S. AsHTON. 1996. Effects of fire intensity on
competitive dynamics between red and black oaks and mountain-laurel. North
j.Appl. For. 13(3):119-123.
NYLAND, R.D. 2002. Silviculture: Concepts and applications, 2nd Ed. McGraw-Hill,
New York. 682 p.
44
NORTH. J. APPL. FOR. 25(1) 2008
OLIVER, C.D., AND B.C. LARSON. 1990. Forest stand dynamics. McGraw-Hill, New
York. 467 p.
ROGERS, R. 1974. Blueberries [forest ranges}. US For. Ser. Gen. Tech. Rep. NE-9.
15 p.
SAS INSTITUTE, INC. 1990. SASISTAT User's guide, ver. 6, 4th Ed. SAS Institute,
Inc., Cary, NC.
SAS INSTITUTE, INC. 200l. The SAS systemfor windows, ver. 8.2. SAS Institute, Inc.,
Cary, NC.
SHUMWAY, D.L., M.D. ABRAM S, AND C.M. RUFFNER. 2001. A 400-year history of
fire and oak recruitment in an old-growth oak forest in western Maryland, U.S.A.
Can.]. For. Res. 31:1437-1443.
STEINER, K.C., AND B.J. JOYCE. 1999. Survival and growth of a Quercus rubra
regeneration cohort during the five years following masting. P. 255-257 in Proc.
ofthe 9th Central hardwoodforest conf, Stringer, W., and D.L. Loftis (eds.). US
For. Serv. Gen. Tech. Rep. SRS-24.
STEINER, K.C., J. FINLEY, AND M.E. McDILL. 2002. Field manual: Bureau ofForestry
Regeneration Project, ver. 6.4. The Pennsylvania State Univ., School of Forest
Resources, University Park, PA. 40 p.
TRUMBULL, V.L., E.j. ZIELINSKI, AND E.C. AHARRAH. 1989. The impact of deer
browsing on the Allegheny forest type. North}. Appl. For. 6:162-165.
WARD, J.H. 1963. Hierarchical groupings to optimize an objective function.]. Am.
Stat. Assoc. 58:236-244.
WARD, W.W. 1983. Ecological and historical perspective: Pennsylvania oak types. P.
1-8 in Proc. ofRegenerating hardwood stands, Finley, J., R.S. Cochran, and J.R.
Grace (eds.). The Pennsylvania State Univ., University Park, PA.