ArticlePDF Available

Patterns of vegetation composition and diversity in pine-dominated ecosystems of the Outer Coastal Plain of North Carolina: Implications for ecosystem restoration

Authors:
Patterns of vegetation composition and diversity in pine-dominated
ecosystems of the Outer Coastal Plain of North Carolina: Implications for
ecosystem restoration
Stephen Mitchell
a,
, Kyle Palmquist
b,c
, Susan Cohen
d
, Norman L. Christensen
a
a
Nicholas School of the Environment, Duke University, Durham, NC 27708, United States
b
Curriculum for the Environment and Ecology, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, United States
c
Department of Botany, University of Wyoming, Laramie, WY 82071, United States
d
Naval Facilities Engineering Command, Marine Corps Base Camp Lejeune, Camp Lejeune, NC 28542, United States
article info
Article history:
Received 6 March 2015
Received in revised form 30 July 2015
Accepted 31 July 2015
Available online xxxx
Keywords:
Longleaf pine
Loblolly pine
Restoration ecology
Pond pine
Prescribed fire
Fire suppression
abstract
Terrestrial ecosystems of the Atlantic coastal plain have experienced considerable change over the past
two centuries, largely due to agricultural activities and fire suppression and exclusion. Many areas that
were once dominated by open longleaf pine (Pinus palustris) woodlands now support closed canopy
stands of loblolly pine (Pinus taeda) with a dense midstory of broadleaved shrubs and trees. In recent
years, efforts to restore the herbaceous plant communities typically found in fire-maintained longleaf
pine woodlands have focused on the use of midstory thinning to produce savanna-like conditions and
to facilitate the restoration of historical fire regimes through prescribed burning. Previous efforts to
restore longleaf pine stands have focused on the potential of fire-suppressed longleaf pine woodlands,
which have been met with some success. However, it is unclear what the potential is for loblolly pine
stands to act as a ’surrogate’ environment for the restoration of the often species-rich herbaceous layer
of longleaf pine woodlands. To assess the effectiveness of longleaf pine restoration treatments in existing
loblolly pine stands, we analyzed the drivers of plant community composition in loblolly pine stands with
mechanical midstory removal treatments, untreated loblolly pine stands, longleaf pine stands, and pond
pine dominated high-pocosin systems. We sampled 75 plots, from which more than 200 individual plant
taxa were identified, with species richness (number of species per 0.1 ha) ranging from 9 in pond pine
pocosins to 118 in longleaf pine woodlands. Plant species richness and composition varied in response
to soil properties, with the first NMS ordination axis correlated with soil properties related to soil mois-
ture and organic matter content (SOM), and the second NMS ordination axis correlated to the concentra-
tion of certain soil nutrients (P, Ca), the variability of which may be due, in part, to historic fertilizer
applications. While stand types were largely distinct from each other in their vegetation composition,
there was nevertheless some compositional overlap among some longleaf and loblolly pine stands.
Areas of compositional overlap appear to have somewhat similar soil properties, whereby the soils found
in overlapping loblolly pine stands were closer to those found in longleaf pine stands (i.e. low SOM con-
tent). Thus, an assessment of the soil properties of loblolly pine stands may allow for an identification of
candidate sites for which longleaf pine restoration treatments may be most effective.
Ó2015 Published by Elsevier B.V.
1. Introduction
Three pine species, longleaf (Pinus palustris), loblolly (P. taeda)
and pond (P. serotina), are the dominant overstory trees in a variety
of wetland and upland ecosystems over much of the lower coastal
plain (also called the Atlantic coastal flatlands (Wear and Greis,
2002), of North and South Carolina, USA (Wells, 1932;
Christensen, 2000). Longleaf pine is thought to have been the most
important canopy tree across much of this landscape prior to
European settlement, extending from xeric sandy sites to much
moister conditions on shallow organic soil. Pond pine was most
important on sites with perennially wet organic soils (Woodwell,
1958; Christensen, 1988). Loblolly pine may have been important
in coastal maritime forests, along riparian corridors, and in ‘‘flat-
woods’’ on poorly drained soils (Ashe, 1915; Wahlenberg, 1960;
http://dx.doi.org/10.1016/j.foreco.2015.07.035
0378-1127/Ó2015 Published by Elsevier B.V.
Corresponding author.
E-mail addresses: stephen.mitchell@duke.edu,srm12@duke.edu (S. Mitchell).
Forest Ecology and Management xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Forest Ecology and Management
journal homepage: www.elsevier.com/locate/foreco
Please cite this article in press as: Mitchell, S., et al. Patterns of vegetation composition and diversity in pine-dominated ecosystems of the Outer Coastal
Plain of North Carolina: Implications for ecosystem restoration. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.foreco.2015.07.035
Christensen, 2000), but the pre-settlement abundance and distri-
bution of this species on this landscape is not well understood.
Land use and a variety of human-caused disturbances over the last
3 centuries have greatly altered the distribution and relative abun-
dance of these species (Landers et al., 1995; Owen, 2002; Wear and
Greis, 2002; Frost, 2006; Noss et al., 2015). Owing to a combination
of timber harvest, naval stores, and clearing for agriculture, lon-
gleaf pine forests have been reduced by more than 98 percent com-
pared to presettlement conditions (Landers et al., 1995; Noss et al.,
1995; Schultz, 1999). Now, loblolly pine is the most abundant pine
on this landscape, except on deep organic soils where pond pine
dominates, and xeric sands where longleaf pine still dominates.
Roughly half of these loblolly stands were naturally seeded follow-
ing land abandonment and the other half planted as plantations
(Schultz, 1999; Wear and Greis, 2002).
Several studies have explored the changes in plant species rich-
ness (number of species) and community composition (identities
and relative abundances of the community of species) among lon-
gleaf and pond pine dominated forests in relation to soil character-
istics. Soil organic matter, bulk density and base cation availability
are highly correlated with variations in species richness and com-
position in longleaf and pond pine stands (Walker and Peet, 1984;
Christensen, 1988; Drewa et al., 2002; Kirkman et al., 2004; Peet,
2006; Carr et al., 2009; Peet et al., 2014). Species richness is highest
in frequently burned longleaf stands on moist and cation-rich soils;
such stands are among the most species rich temperate ecosystems
in North America (Walker and Peet, 1984). However, longleaf pine
is now most commonly found on sandy, well-drained soils with
lower levels of species richness, perhaps because such systems
were historically less attractive for agriculture or intensive
silviculture.
Much change in the composition and diversity of these pine
ecosystems has been caused by alteration of their fire regimes
(Garren, 1943; Christensen, 1981; Frost et al., 1986). Frequent fire
(every one to five years) is necessary for the maintenance of plant
species richness, plant species composition, and vegetation struc-
ture in longleaf pine woodlands (Walker and Peet, 1984; Frost
et al., 1986; Kirkman et al., 2004). With fire exclusion or longer fire
return intervals, species richness declines, woody components
increase, and the understory becomes dense and closed
(Heyward, 1939; Lewis and Harshbarger, 1976; Brockway and
Lewis, 1997; Glitzenstein et al., 2003). The specific location of the
transition from longleaf to pond pine dominated forests along soil
moisture gradients is a function of fire frequency; longer fire return
intervals move that transition to wetter conditions (Kologiski,
1977; Christensen, 1981; Frost et al., 1986; Frost, 2006).
Because of the high biodiversity, rarity, and endemism of the
longleaf pine ecosystem and its reduced areal extent, there has
been much interest in the restoration of longleaf pine stands
throughout the Southeastern US. In existing longleaf stands where
short-term (610 years) fire suppression has resulted in only mod-
est change in understory structure and composition, reestablish-
ment of frequent, low-severity fire regimes has been, by itself, an
effective restoration tool (Walker and Silletti, 2006). By contrast,
the legacy of long-term (>10 years) fire suppression presents addi-
tional challenges to the restoration of vegetation associated with
fire-maintained longleaf pine woodlands. First, excessive litter
and soil organic matter (SOM) accumulation associated with fire
suppression decreases reproductive success of established species
and prevents colonization events by new species (Walker and
Silletti, 2006), as many plant species within this ecosystem, includ-
ing longleaf pine, require bare, mineral soil for germination and
seedling establishment. The accumulation of SOM is also a barrier
to the establishment of herbaceous species characteristic of lon-
gleaf woodlands from the soil seed bank (Cohen et al., 2004).
Second, longleaf pine seedlings, as well as many of the associated
herb-layer species native to longleaf savannas, regenerate best in
an open, park-like setting (Brockway and Outcalt, 1998) and an
encroachment of the forest mid-story can significantly limit
recruitment via light limitation (Hiers et al., 2007). Thus, in stands
where decades of long-term (>10 years) fire suppression have
allowed the ingrowth of dense hardwood understories and midsto-
ries, restoration has relied on both mechanical thinning of the mid-
story and the restoration of appropriate fire regimes (Provencher
et al., 2001; Varner et al., 2005; Walker and Silletti, 2006;
Brockway et al., 2009; Outcalt and Brockway, 2010; Steen et al.,
2013).
There is great interest in restoring longleaf pine to places now
dominated by loblolly pine (Brockway et al., 2009; Schwilk et al.,
2009; Knapp et al., 2011). Most of these efforts involve thinning
of understory and midstory woody plants, similar to methods
implemented in fire-suppressed longleaf stands. Here, the
short-term goal is to restore an open stand structure similar to lon-
gleaf pine savannas. Over the long-term, such treatments are com-
bined with frequent applications of prescribed fire and may
include the planting of longleaf pine and some of its associate spe-
cies with the goal of eventually restoring the longleaf ecosystem
(see Walker and Silletti (2006) for a detailed discussion of restora-
tion targets and protocols). However, there are significant chal-
lenges to restoration. First, loblolly now occurs across a wide
variety of soil/site conditions, not all of which once supported lon-
gleaf. Thus, restoration will likely succeed on only a subset of sites
currently supporting loblolly. Second, most areas that now support
loblolly (e.g. plantations) have been altered by a long history,
including over a century of agriculture and/or intensive forest
management. Seed banks of longleaf pine endemics in these sys-
tems have been exhausted and few ecological legacies of longleaf
pine ecosystems remain. Furthermore, areas now dominated by
loblolly have not been burned for a long time, resulting in the inva-
sion of woody species in the understory and accumulations of litter
and soil organic matter, conditions that do not facilitate the estab-
lishment of longleaf pine and associated herb-layer species.
Reduction of duff via prescribed fire may be effective in reducing
soil organic matter but must be done with caution, as prescribed
burns set in stands with an exceptionally dry duff layer are more
likely to result in crown scorch and overstory tree mortality
(Varner et al., 2007, 2009).
An understanding of the environmental factors that control the
distribution of plant species in pine forests of the Atlantic coastal
plain is essential for implementing successful efforts to restore
their structure, composition, and disturbance regimes. Here, using
data from 75 permanent sample plots located in areas representing
a wide range of soil/site conditions, we evaluated the role of envi-
ronmental gradients and stand structure on plant species richness
and composition among loblolly, pond, and longleaf pine stands
located in the Atlantic coastal flatlands of North Carolina. We asked
the following questions:
(1) What environmental factors best explain the differences in
the herb-layer composition of pond pine pocosins, loblolly
pine forests, and longleaf pine woodlands?
(2) To what extent do the herb-layer composition and trends in
species richness in loblolly pine forests differ from that of
longleaf pine woodlands and pond pine pocosins?
(3) Which species are most indicative of the site conditions
characterized by fire-maintained pine woodlands, and
which species are most indicative of a departure from those
site conditions?
(4) What are the environmental conditions that might best
facilitate restoration of the herb-layer of longleaf pine
woodlands in stands that are not dominated by longleaf
pine?
2S. Mitchell et al. / Forest Ecology and Management xxx (2015) xxx–xxx
Please cite this article in press as: Mitchell, S., et al. Patterns of vegetation composition and diversity in pine-dominated ecosystems of the Outer Coastal
Plain of North Carolina: Implications for ecosystem restoration. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.foreco.2015.07.035
Of particular interest is whether there is any compositional
overlap between longleaf pine woodlands and loblolly pine and/or
pond pine forests. Areas of compositional overlap between loblolly
and longleaf pine stands may aid in the identification of candidate
restoration sites and the environmental properties for which
efforts to restore the vegetation native to longleaf pine woodlands
may be most promising. Conversely, areas of compositional over-
lap between loblolly and pond pine stands may be indicative of
stands in which efforts to restore longleaf pine vegetation are unli-
kely to succeed.
2. Methods
2.1. Study area
We conducted our research at Marine Corps Base Camp Lejeune
(MCBCL) within the Atlantic Coastal Flatlands Section of the Outer
Coastal Plains Mixed Forest Province. The climate is classified as
warm-humid temperate, with average annual precipitation of
1420 mm and mean annual temperature of 13 °C(MCBCL, 2006).
Elevation in our study area ranges from 10 to 30 m above mean
sea level (AMSL). Seventy-five study sites were located in forest
stands that were selected to represent a range of soil characteris-
tics (as documented in soil series classification by Barnhill
(1992)), which are determined largely by the characteristics of par-
ent material (e.g., unconsolidated sediments and limestone) and
soil moisture conditions. Longleaf pine woodlands generally occur
on shallow organic and mineral soils; the depth of the water table
in these ecosystems varies from a few centimeters to more than
1 m, depending on topography, creating a gradient between xeric
uplands and poorly drained, lowland stands. The study area also
contains high-pocosin forest-wetlands (Christensen, 2000;
Harding and Walters, 2002) on poorly drained organic soils.
Some plant species that are abundant in pocosins, such as Ilex gla-
bra and Persea palustris, also occur in loblolly and longleaf forest
stands on moister soils and where fires have been suppressed for
long periods. Loblolly pine forests at MCBCL are largely
fire-suppressed and, on average, have higher SOM than longleaf
pine stands and often contain a dense hardwood midstory with
stems of sweet gum (Liquidambar styraciflua) and red maple (Acer
rubrum). Historically, MCBCL was used primarily for timber pro-
duction and, to a lesser extent, agriculture, but no agricultural
usage has occurred since its designation as a military installation
in 1941 (MCBCL, 2006).
3. Vegetation sampling
Between June and September of 2009 and 2010, permanent
20-m 50-m (0.1-ha) vegetation monitoring plots were estab-
lished in each of 71 loblolly and longleaf stands, while 4 pond
pine/pocosin stands were sampled in 2011. Among these, 46 were
dominated by longleaf pine, 25 were dominated by loblolly pine
and 4 were dominated by pond pine. Based on criteria from
Schafale (2012), longleaf pine stands were divided into two groups.
23 plots were assigned to Dry Pine/Oak Savannas (DPOS), where
scrub oaks (e.g., Quercus laevis,Q. geminata and Q. hemisphaerica)
were abundant in the understory. The remaining 23 plots were
assigned to a Wet Pine Savannas (WPS) group, in which the under-
story was dominated by graminoids and herbs. Of the loblolly
plots, 16 had been subjected to hardwood mid-story removal with
a Hydro-Ax
Ò
mulching device (Prentice Forestry, http://www.
prenticeforestry.com) in the year prior to sampling. This thinning
removed all stems <20 cm dbh and created open-canopy condi-
tions similar to those of regularly burned longleaf stands
(Provencher et al., 2001; Brockway et al., 2009). The other 9
loblolly pine plots had a dense, untreated hardwood mid-story.
Thinned loblolly, unthinned loblolly and high pocosin stands were
designated as (LT), (LU) and (HPO), respectively. The 25 loblolly
pine treatment plots were randomly located within composition-
ally homogeneous stands >5 ha that exhibited evidence of fire
exclusion. The remaining 50 monitoring points were previously
located by MCBCL staff to represent gradients in soil properties,
vegetation composition, and disturbance history, while avoiding
any edge effects with neighboring stands. Thus, each 0.1 ha sample
plot is treated as a sample universe in its own right and is not nec-
essarily intended to be statistically representative of the entire
stand within which it is contained (Peet et al., 1998).
Each plot was permanently marked with heavy steel conduit,
and subsequently sampled for woody and herbaceous plant com-
position and cover using the Carolina Vegetation Survey (CVS) pro-
tocol (Peet et al., 1998). Individual plots were located to avoid
obvious vegetation transitions and represent relatively uniform
soil and topographic conditions. Within each plot, all living and
dead stems (greater than 1-cm dbh) were tallied by species and
dbh. For the herbaceous layer, the plot was further subdivided into
ten 10-m 10-m subplots. The cover of each herb species was
assessed in the intensive plots as prescribed by Peet et al. (1998).
The cover of each herb species was assessed in the intensive plots
as prescribed by Peet et al. (1998), as described in the methods sec-
tion. For each of the four intensive modules, cover% was repre-
sented by cover class, ranging from trace coverage 1 = trace
(0.25%), 2 = 0–1% (0.5%), 3 = 1–2% (1.5%), 4 = 2–5% (3.5%), 5 = 5–
10% (7.5%), 6 = 10–25% (17.5%), 7 = 25–50% (37.5%), 8 = 50–75%
62.%), 9 = 75–95% (85%), and 10 = >95% (97.5%). Final coverage%
for the species matrix was calculated by the average% coverage
of the four intensive modules. For ‘‘trace’’ species not occurring
within the four intensive modules, coverage% was estimated for
the entire plot area. Residual subplots were surveyed to identify
all species not found in any of the four intensive subplots. Thus,
species richness estimates are based on the total number of species
identified in the entire plot. Species composition here refers to
identities and relative abundances of the community of species.
In some (11) cases, closely related species could not be reliably dis-
tinguished for estimates of percent coverage in the field, and these
were lumped into genus-level complexes. Liatris pilosa and Liatris
virgata, for example, were classified as simply Liatris
[pilosa +virgata].
4. Disturbance history
Since 1995, MCBCL has maintained a GIS database of all known
fires. Perimeters and dates of each fire were mapped with a GIS
data recorder and stored as a shape file in a geodatabase main-
tained by base personnel. We used these data to calculate the time
since most recent fire, as well as fire frequencies from 1995 until
the sampling date. Both categories of longleaf pine stands (DPOS
and WPS) had a fire frequency commensurate with the historical
mean fire return intervals for longleaf pine of 3 years (Table 1).
As expected, longleaf pine stands were also burned fairly recently,
2 years for prior to sampling for both DPOS and WPS (Table 1). By
contrast, thinned (LT) and unthinned loblolly pine (LU) stands had
fire frequencies of 9 and 8 years, respectively. Loblolly pine
stands had been burned, on average, 2 years prior to sampling for
both LU plots and LT plots. Pocosin plots were limited in number,
and all had a fire frequency of 4 years, with all fires occurring
1–2 years prior to sampling. These fire return intervals are con-
siderably shorter than those of historical pocosin fire regimes
(40–60 year return intervals, Christensen, 1981, 2000). Thus,
recent fire regimes in pocosin systems may be driven by an impe-
tus for fuels and fire hazard management.
S. Mitchell et al. / Forest Ecology and Management xxx (2015) xxx–xxx 3
Please cite this article in press as: Mitchell, S., et al. Patterns of vegetation composition and diversity in pine-dominated ecosystems of the Outer Coastal
Plain of North Carolina: Implications for ecosystem restoration. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.foreco.2015.07.035
5. Soil sampling
A uniform sample of the top 0–10 cm of mineral soil (soil
beneath layers of litter and duff) was collected using a 5-cm diam-
eter piston corer at each of four points located 10 m from the cen-
ter point of each 0.1 ha plot. Each soil sample was subsequently
analyzed by Brookside Laboratories (New Knoxville, OH). Soil pH
was measured using a glass electrode in a 1:1 slurry of soil and dis-
tilled water (McLean, 1982). Percent soil organic matter (SOM) was
determined by weight loss after ignition at 360 °C. Aluminum (Al),
boron (B), calcium (Ca), copper (Cu), iron (Fe), phosphorus (P),
potassium (K), magnesium (Mg), manganese (Mn), sodium (Na),
sulfur (S), and zinc (Zn) were extracted according to (Mehlich,
1984) and concentrations were determined using
plasma-emission spectroscopy. P concentrations in the Mehlich
extract were measured colorimetrically. Cation exchange capacity
(CEC) was measured by summation of all cations as mEq/100 g of
soil (Ross, 1995). Percent hydrogen (%H) was estimated based on
the fraction of total CEC not occupied by the metallic cations listed
above. Thus, %H is inversely correlated with percent exchangeable
bases.
6. Statistical analyses
We tested for significant differences in herb-layer plant compo-
sition among these five stand types using a Multi-Response
Permutation Procedure (MRPP, see Mielke and Berry (2007)).
MRPP is a nonparametric procedure for testing the hypothesis of
no difference between two or more groups of entities. As such,
MRPP was used to examine the magnitude of the differences in
community composition between these five stand types. MRPP cal-
culates the statistic A, a descriptor of within-group homogeneity
compared to the random expectation. This is known as
chance-corrected within-group agreement. Values for Aare com-
monly below 0.1, even when the observed delta differs signifi-
cantly from the expected, while an A> 0.3 is considered rather
high (McCune and Grace, 2002). Of particular interest to our study
were differences between longleaf pine stands and loblolly pine
stands with mid-story removal treatments.
Non-metric multidimensional scaling (NMS) ordination
(Kruskal, 1964) was used to analyze trends in species composition
among plots. NMS analyses were carried out using the S
ørenson
dissimilarity metric for 1000 iterations and a stability criterion of
0.00001. Only species occurring in 3 or more of our 0.1 ha study
sites were included in the ordinations. To account for covariation
among soil properties, we conducted a Principle Components
Analysis (PCA) of soil pH, soil bulk density, and soil nutrient con-
centrations (Peet et al., 2014). To ensure that the soil data con-
formed to a multivariate normal distribution prior to PCA, we
conducted a Mardia test (Mardia, 1970) for multivariate normality
using the MVN statistical package (Korkmaz et al., 2014) in R soft-
ware. Bulk density and pH did not require any data transformation
to meet assumptions for multivariate normality, while soil nutrient
data needed log transformations for P, Ca, K, Na, and B. Other soil
nutrients were not amenable the assumptions of multivariate
normality following log transformation and were thus excluded
from the PCA. We regressed NMS axes against the PCA axes to
ascertain the relationship between soil properties and the compo-
sition of herb-layer vegetation. Additionally, to separate the
respective contributions of soil properties and disturbance history
(i.e. fire frequency, and time since most recent fire), we utilized the
variance partitioning function from the vegan statistical package in
R software (Dixon and Palmer, 2003) to ascertain the unique con-
tributions of soil properties (PCA Axes 1 and 2) vs. those of distur-
bance history in determining plant composition as represented by
NMS analysis.
Finally, upon confirming statistically significant differences in
our five stand types through the MRPP analysis, we performed an
Indicator Species Analysis (ISA, Dufrêne and Legendre (1997)).
ISA generates an indicator value for species ranging from 0 to
100. A maximum value of 100 is achieved when all individuals of
a taxon occur at all sites of a single pre-defined group and no other
sites. The significance of the indicator value for each species was
generated with a Monte Carlo randomization procedure with
9999 permutations. Of particular interest is whether more species
showed an apparent preference for treated over untreated loblolly
pine stands and how indicative species vary between loblolly
stands and fire-maintained longleaf pine stands, as well as how soil
properties are associated with the fidelity of certain species. Thus,
ISA was conducted for two types of grouping variables. First, we
used our five stand types [Dry Pine/Oak Savanna (DPOS), Wet
Pine Savanna (WPS), Loblolly Unthinned (LU), Loblolly Thinned
(LT), and High Pocosin (HPO)] as categorical variables.
Additionally, we conducted ISA on soil PCA scores for PCA axes 1
and 2. Scores were grouped into positive (+) and negative (-) PCA
score values for both axes. All analyses (MRPP, NMS, and ISA) were
performed using PC-ORD software version 6.247 Beta (McCune and
Mefford, 2002).
7. Results and discussion
7.1. Soil variability and plant composition
MRPP pairwise comparisons indicated that each stand type
(DPOS, WPS, LU, LT, and HP) was significantly different from all
other groups. Within group agreement values (A) for most compar-
isons were >0.1, indicative of significant differences between most
groups (Table 2). Lower values of A for DPOS vs HP, LU vs LT, and LT
vs HPO are probably due to the smaller sample sizes of the LU and
HPO groups (Mielke and Berry, 2007). Differences among our com-
munity groups/stand types can be seen in the 2-D plot of our NMS
ordination (stress = 16.41, instability < 0.00001 after 67 iterations),
even though variation within community groups and overlap
among groups is generally high (Fig. 1). However, much of this
variation can be explained by soil properties. Means and ranges
for soil variables in each community group are displayed in
Table 3, with soil variable eigenvectors for PCA axes 1 and 2 shown
in Table 4. Soil PCA axis 1 explained 40.8% of the variation in the
soil data matrix, while soil PCA Axis 2 explained an additional
24.2%. PCA axis 1 was highly correlated with bulk density (BD),
K, and Na. While there was considerable overlap in group ranges
Table 1
Mean values of fire frequency and time since most recent fire for each community group. Ranges are shown in parentheses. Values of 0.5 are given for all fires occurring during the
year of sampling, but prior to the actual sampling.
Mixed Pine/Oak Savanna Wet Pine Savanna Loblolly unthinned Loblolly thinned High pocosin
Abbreviation DPOS WPS LU LT HPO
Number of sites N=23 N=23 N=9 N=16 N=4
MFRI (1995–2010) 3.12 (1.36–14) 3.53 (1.56–20) 8.30 (1.36–20) 9.43 (3–20) 4 (4)
Time since most recent fire 2.43 (0.5–8.5) 2.15 (0.5–8.5) 1.89 (0.5–8.5) 1.81 (0.5–8.5) 2.25 (1.5–2.5)
4S. Mitchell et al. / Forest Ecology and Management xxx (2015) xxx–xxx
Please cite this article in press as: Mitchell, S., et al. Patterns of vegetation composition and diversity in pine-dominated ecosystems of the Outer Coastal
Plain of North Carolina: Implications for ecosystem restoration. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.foreco.2015.07.035
(Table 3), the high correlation of bulk density with PCA axis 1 was
very clearly related to the gradient of soil moisture conditions from
sandy, well drained DPOS sites to moister, more poorly drained
WPS sites, to very wet HPO sites. Thus, soil properties related to
soil moisture conditions were strongly correlated with vegetation
composition, evidenced by a very strong relationship between soil
PCA axis 1 and NMS axis 1 (R
2
= 0.57, p< 0.01, Fig. 2). In general,
Dry Pine-Oak Savanna (DPOS) stands had the highest NMS axis 1
scores, followed by Wet Pine Savannas (WPS), while High
Pocosin (HPO) stands had the lowest NMS axis 1 scores; mean axis
1 scores for Loblolly-Unthinned (LU) and Loblolly-Thinned (LT) fell
between those for WPS and HPO (Table 5).
PCA axis 2 was highly correlated with PO
4
-P, pH, and B, and was
moderately correlated with Ca (Table 4). We found a moderately
strong relationship between PCA axis 2 and NMS axis 2
(R
2
= 0.30, p< 0.01, Fig. 2). One potential explanation for the rela-
tionship between PCA axis 2 and NMS axis 2 is the variation in soil
properties that often results from historic applications of phos-
phate fertilizer and lime on former agricultural lands, many of
which later transition to loblolly pine stands. The effects of such
additions can persist for years. Richter et al. (2006) examined the
long-term soil chemistry of a loblolly pine forest with a history
of fertilization and found evidence for a resupplying of the most
labile fractions of soil P from release of organic and inorganic P
associated with Fe and Al oxides and Ca compounds. Our findings
are somewhat consistent with a scenario in which the previous fer-
tilization and liming of agricultural land may have facilitated a
slow release of P associated with Ca. Stands with high values for
the 2nd PCA Axis were strongly dominated by the LU and LT
groups, which had mean PO
4
-P values of 7.01 and 6.38
l
g/g com-
pared to 3.80, 2.88, and 3.40
l
g/g for DPOS, WPS, and HPO. LU
Table 2
Results of pairwise comparisons from MRPP analysis. Values for Aare commonly
below 0.1, even when the observed delta differs significantly from the expected
(McCune and Grace, 2002). An A> 0.3 is fairly high.
Pairwise comparisons T-Value Ap<
DPOS vs WPS 11.823 0.156 0.00001
DPOS vs LU 19.184 0.153 0.00001
DPOS vs LT 13.097 0.127 0.00001
DPOS vs HPO 4.031 0.022 0.004
WPS vs LU 8.258 0.127 0.00001
WPS vs LT 3.294 0.109 0.02
WPS vs HPO 10.201 0.118 0.00001
LU vs LT 2.620 0.029 0.02
LU vs HPO 15.314 0.103 0.00001
LT vs HPO 10.847 0.091 0.00001
Fig. 1. NMS ordination results showing longleaf pine stands tightly clustered on the
right side of ordination space. Loblolly pine stands, with and without midstory
removal, exhibited a much higher range of compositional variation in ordination
space. Our 2-D NMS ordination of plant community composition had a final stress
of 16.41 and an instability <0.00001 after 67 iterations.
Table 3
Mean values of soil variables for each community group. Ranges are shown in parentheses. Statistically significant differences (P< 0.05) among groups (as determined by ANOVA
and Duncan’s Multiple Range Tests) are indicated by letter superscripts.
Dry Pine/Oak Savanna Wet Pine Savanna Loblolly unthinned Loblolly thinned High pocosin
Abbreviation DPOS WPS LU LT HPO
Number of Sites N=23 N=23 N=9 N=16 N=4
BD (g/cm
3
) 1.16 (0.63–1.40)
a
1.16 (0.81–1.40)
a
0.63 (0.18–1.09)
b
0.63 (0.31–0.98)
b
0.70 (0.30–1.10)
ab
pH 4.39 (3.8–4.9) 4.14 (3.8–4.6) 4.12 (3.63–4.55) 4.05 (3.83–4.68) 4.50 (3.9–5.1)
CEC (mEq/100 g) 4.24 (2.0–6.1)
a
4.09 (2.3–9.5)
a
6.86 (3.26–10.42)
b
8.11 (5.69–11.48)
b
4.37 (2.4–6.3)
a
SOM (%) 3.52 (1.45–12.7)
a
4.17 (0.97–10.6)
a
30.40 (2.85–85.45)
b
25.04 (5.92–68.34)
b
18.66 (2.7–55.8)
b
PO
4
-P (
l
g/g) 3.8 (1.0–8.0)
a
2.88 (1.0–7.0)
a
7.01 (3.0–13.0)
b
6.38 (2.5–11.5)
b
3.40 (2.8–3.8)
a
Ca (
l
g/g) 186.86 (117.7–286.8)
a
143.54 (54.9–276.3)
a
229.21 (54.9–595.0)
b
288.64 (170.5–518.3)
b
157.00 (103.3–190.5)
a
Mg (
l
g/g) 34.27 (16.3–65.5) 30.45 (15.0–65.8)
a
59.78 (24.0–114.0) 61.63 (42.3–94.0) 63.95 (38.8–127.0)
K(
l
g/g) 21.82 (11.5–54.8) 20.06 (9.0–37.8) 41.97 (18.25–87.75) 37.27 (23.5–63.0) 21.50 (15.3–31.5)
Na (
l
g/g) 25.39 (15.5–38.3) 25.40 (17.8–46.3) 37.22 (20.25–67.75) 31.66 (21.3–52.5) 32.55 (26.5–37.0)
B(
l
g/g) 0.17 (0.1–0.37) 0.17 (0.1–0.49) 0.26 (0.15–0.38) 0.24 (0.15–0.35) 0.14 (0.1–0.30)
Fe (
l
g/g) 53.75 (27.0–188.3) 119.04 (21.8–382.5) 107.83 (38.0–204.3) 114.41 (29.3–232.8) 46.40 (22.3–82.3)
Mn (
l
g/g) 4.88 (0.8–11.3) 2.06 (0.8–6.8) 1.76 (09–4.0) 2.35 (1.1–4.5) 1.82 (0.8–3.5)
Cu (
l
g/g) 0.48 (0.16–0.91) 0.77 (0.27–1.72) 0.72 (0.41–1.22) 0.67 (0.35–1.93) 0.37 (0.28–0.56)
Zn (
l
g/g) 1.03(0.58–1.93) 1.26 (0.46–4.49) 1.29 (0.66–2.90) 2.16 (0.62–12.12) 0.69 (0.51–0.92)
Al (
l
g/g) 151.07 (69.5–401.5)
a
298.50 (52.5–798.0)
a
568.27 (121.5–1211.5)
b
483.77 (113.75–1003.5)
a
185.25 (110.0–274.5)
b
Table 4
Soil variable eigenvectors for PCA axes 1 and 2, each scaled to its standard deviation.
Scaled in this manner, eigenvectors are equivalent to the correlation coefficients
between stand PCA scores and the values for each soil variable for each sample area.
PCA axes 1 and 2 account for 40.8% and 24.2% of the total multivariate variance.
Shaded values are significant at the P< 0.001 level.
S. Mitchell et al. / Forest Ecology and Management xxx (2015) xxx–xxx 5
Please cite this article in press as: Mitchell, S., et al. Patterns of vegetation composition and diversity in pine-dominated ecosystems of the Outer Coastal
Plain of North Carolina: Implications for ecosystem restoration. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.foreco.2015.07.035
and LT groups had higher Ca as well, at 229.21 and 288.64
l
g/g,
compared to 186.86, 143.54, and 157.00
l
g/g for DPOS, WPS, and
HPO (Table 3). However, we cannot confirm a legacy effect of fer-
tilization without long-term soil records.
To ascertain the relative contributions of soil properties and dis-
turbance history on the first NMS axis of plant composition, we
conducted a variance partitioning of soil PCA axis 1, fire frequency,
and time since most recent fire. Soil PCA Axis 1 was highly corre-
lated with NMS axis 1 (discussed above, and see Fig. 2) and our
variance partitioning found that soil PCA axis 1 uniquely explained
43% of the variation in NMS axis 1, while fire frequency uniquely
explained only 2% and time since most recent fire uniquely
explained 0%. Thus, fire frequency, while correlated with the 1st
NMS axis of plant composition (R
2
= 0.17, p< 0.01), did not
uniquely explain a significant proportion of the variation of it, sug-
gesting that soil properties have a much stronger relationship with
plant community composition than fire frequency. Similarly, for
NMS axis 2, we conducted a variance partitioning of soil PCA axis
2, fire frequency, and time since most recent fire. PCA axis 2
uniquely explained 30% of the variation in NMS axis 2, while both
fire frequency and time since most recent fire uniquely explained
0%. However, it would be highly premature to suggest that fire fre-
quency does not have a significant influence on plant community
composition. Instead, the relationship between fire frequency and
vegetation may be manifested, in part, through the relationship
of fire frequency on soil properties, whereby frequent fires reduce
the accumulation of soil organic matter. Fire frequency did have a
slightly positive correlation to PCA axis 1 (R
2
= 0.12, p< 0.01) of soil
properties. However, such a relationship may also reflect a propen-
sity by forest managers to conduct prescribed fires more frequently
in stands that are low in soil organic matter accumulation to begin
with.
7.2. Indicator species analysis by stand type
Indicator Species Analysis using five stand types as categorical
variables revealed 53 species with a statistically significant associa-
tion (p< 0.05) with a particular group (Table 6). Of these, 23 were
significant indicators for the DPOS group (e.g. Aristida stricta,
Vaccinium tenellum), 16 for the WPS group (e.g. Rhexia alifanus,
Pteridium aquilinum), 4 for the LT group (e.g. A. rubrum,L. styraciflua),
and 9 for the HPO group (e.g. Gordonia laisianthus,Lyonia lucida). No
species was a significant indicator for the LU group by itself. It should
be noted that 3 of the 4 LT indicators were seedlings of A. rubrum,
Pinus taeda, and L. styraciflua. All three are shade intolerant, and their
importance in these thinned stands undoubtedly reflects higher
irradiance at the forest floor. Thus, species that are indicative of
loblolly pine are all indicators of disturbance via midstory removal.
By contrast, several herbaceous layer species were associated
with the HPO group (e.g. Rhychospora plumosa,Desmodium spp.,
Rhexia mariana, and Carphephorus paniculatus).
7.3. Indicator species analysis by soil properties
The Indicator Species Analysis using positive (+) and negative
() PCA axis score groups from the 1st and 2nd PCA axes found
54 species that exhibited a statistically significant association
(p< 0.05) for at least 1 soil PCA axis group, with 19 species showing
a statistically significant association for both the 1st and 2nd soil
PCA axis groups. Our ISA found 28 species with that were
Fig. 2. NMS axes 1 and 2 plotted against soil PCA axes 1 and 2, respectively. Points are highlighted by their respective stand type. The soil variables indicative of the PCA axes
are displayed in Table 3.
Table 5
Mean NMS Axis 1 and Axis 2 scores for five community groups. Ranges are indicated in parentheses.
Dry Pine/Oak Savanna Wet Pine Savanna Loblolly unthinned Loblolly thinned High pocosin
Abbreviation DPOS WPS LU LT HPO
Number of Sites N=23 N=23 N=9 N=16 N=4
NMS axis 1 score 0.733 (0.633 to 1.235) 0.378 (0.719 to 1.049) 0.902 (1.532 to 0.0736) 0.8214 (1.39 to 0.190) 1.073 (1.455 to 0.628)
NMS axis 2 score 0.011 (0.500 to 0.665) 0.247 (0.334 to 0.968) 0.407 (1.534 to 0.642) 0.369 (1.214 to 0.400) 0.911 (0.711 to 1.101)
6S. Mitchell et al. / Forest Ecology and Management xxx (2015) xxx–xxx
Please cite this article in press as: Mitchell, S., et al. Patterns of vegetation composition and diversity in pine-dominated ecosystems of the Outer Coastal
Plain of North Carolina: Implications for ecosystem restoration. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.foreco.2015.07.035
significantly associated with the positive soil PCA axis 1 group,
which were primarily soils with high bulk density and low soil
organic matter generally associated with longleaf pine stands
(Table 7). Conversely, we found 10 species associated with the neg-
ative PCA Axis 1 group, which were primarily soils with low bulk
density and high soil organic matter, typical of loblolly pine stands.
Soil PCA axis 2 values were associated with a PO
4
-P gradient more
than anything else (Table 4), with positive values associated with
high PO
4
-P availability, and negative values associated with low
PO
4
-P availability. We found 17 species that were significantly
associated with the positive soil PCA axis 2 value group, and 18
species that were significantly associated with the negative soil
PCA axis 2 value group (Table 7). Thus, soil properties can exert
considerable influence on species associations in these ecosystems.
7.4. Herb-layer coverage and species richness
Mean values for species richness (#/0.1 ha, min–max) for both
DPOS (44.5, 24–72) and WPS (57.4, 25–118) groups were signifi-
cantly higher than the LU (22.5, 7–48), LT (26.1, 13–41), and HPO
(18.2, 9–29) groups based on ANOVA and multiple range compar-
isons (p< 0.01). Hedman et al. (2000) compared longleaf and
loblolly pine stands and likewise found higher levels of species
richness among (35, 25–95) longleaf pine. Although mean species
richness was higher in the WPS than the DPOS community group,
the difference was not statistically significant (p> 0.074).
Nevertheless, 7 WPS stands supported more than 75 specie-
s/0.1 ha, whereas no DPOS stands had more than 72 species.
There were no significant differences in species richness among
the LU, LT and HPO groups. Species richness was negatively corre-
lated with PCA axis 1 and positively correlated with NMS axis 1
(Fig. 3). However, species richness in stands with the very highest
scores was generally lower than would be predicted by a simple
linear regression line, and the highest species richness scores were
observed in stands with high-intermediate PCA and NMS axis 1
scores (i.e., WPS).
Mean herb-layer cover (%, min–max) was about two times
higher in the DPOS (89.2%, 30.9–170.6), WPS (90.0%, 27.5–148.7),
and HPO (116.8%, 99.1–133.6) groups compared to the LU (33.9%,
7.1–67.1) and LT (45.2%, 11.2–84.5) groups (p< 0.01). Others have
also found noted higher herb-layer coverage among longleaf pine
stands compared to loblolly pine stands, with longleaf pine stands
exhibiting 45% more coverage (Hedman et al., 2000). The herb
layer in the HPO stands was predominantly low woody shrubs,
whereas herbaceous species comprised the majority of DPOS and
WPS group herb layer cover. No significant difference in herb layer
cover between LU and LT groups was evident 12–18 months after
thinning treatments.
7.5. Restoration implications
Overall, our results suggest that a utilization of loblolly pine
stands as surrogate sites for the restoration of longleaf pine vege-
tation remains an uncertain endeavor that will be highly con-
strained by site characteristics and disturbance history. Soil
properties, particularly bulk density (BD), were highly correlated
with plant community composition, whereby high SOM and low
BD were not characteristic soil conditions of fire-maintained lon-
gleaf pine woodlands (Cohen et al., 2004; Walker and Silletti,
2006). While other studies have examined the relationship
between soil properties and herb-layer species composition, our
work is in contrast to many previous case studies of longleaf pine
restoration focused on the restoration of longleaf stands in which
SOM content is already low. Studies of restoration in
fire-suppressed longleaf pine stands in northern Florida, USA
(Eglin Air Force Base) are primarily concentrated on longleaf stands
on xeric soils with high bulk density and low SOM (<0.9%).
Kirkman et al. (2013) examined the long-term (15 year) response
of herb-layer vegetation in fire-suppressed longleaf pine to mid-
story treatments with frequent (every 3 year) post-treatment pre-
scribed fires and found that treated stands showed a positive
restoration trajectory of the herbaceous layer toward that of lon-
gleaf pine woodlands (Rodgers and Provencher, 1999; Kirkman
et al., 2013).
The restoration of longleaf pine woodlands has not been limited
to mature longleaf stands; conversion of abandoned agricultural
land to longleaf pine woodlands has also taken place in sites that
Table 6
Results from indicator species analysis with stand as the categorical variables. Species
are listed in order of decreasing indicator value (% of perfect indication). No species
was found to exhibit preference for Loblolly-unthinned.
Species ISA value p-value
Dry Pine/Oak Savanna (DPOS)
Quercus laevis 66 0.002
Quercus incana 60.4 0.0026
Aristida stricta 59 0.0002
Gaylussacia dumosa 59 0.0068
Pinus palustris 47.4 0.0068
Vaccinium tenellum 46.9 0.0112
Toxicodendron pubescens 46.7 0.0056
Liatris [pilosa + virgata] 46.6 0.0144
Cnidoscolus stimulosus 42.9 0.013
Andropogon spp. 42.7 0.0264
Pityopsis graminifolia 41.4 0.0304
Sericocarpus tortifolius 40.1 0.0112
Carphephorus bellidifolius 39.6 0.0214
Diospyros virginiana 38.5 0.0298
Rhus copallina 38.5 0.0338
Ionactis linariifolius 35.4 0.0292
Tragia urens 31.7 0.0466
Morella pumila 31.5 0.0436
Quercus hemisphaerica 29.7 0.0372
Euphorbia ipecacuanhae 29 0.0404
Scleria nitida 27.1 0.0412
Chrysopsis gossypina 24.5 0.0344
Cirsium vulgare 17.4 0.0332
Wet Pine Savanna (WPS)
Rhexia alifanus 56.8 0.0096
Pteridium aquilinum 56.2 0.0172
Xyris caroliniana 48 0.0108
Lachnocaulon anceps 45.9 0.0092
Lyonia mariana 45.6 0.0132
Carphephorus odoratissimus 43.7 0.0138
Pinus serotina 40.5 0.0386
Lobelia nuttallii 36 0.0242
Pterocaulon pycnostachyum 33.4 0.0494
Gymnopogon brevifolius 30.4 0.0326
Hypericum crux-andreae 26.4 0.0416
Woodwardia virginica 26 0.0348
Desmodium tenuifolium 25.5 0.0464
Eupatorium pilosum 25.3 0.043
Carphephorus tomentosus 23 0.0482
Ludwigia virgata 21.7 0.0374
High pocosin (HPO)
Rhynchospora plumosa 94.2 0.0002
Gordonia lasianthus 91 0.0002
Desmodium spp. 90.8 0.0002
Zenobia pulverulenta 75 0.0006
Ilex coriacea 70.1 0.003
Lyonia lucida 66.1 0.0006
Magnolia virginiana 62 0.0062
Rhexia mariana 61 0.0006
Carphephorus paniculatus 21.2 0.0398
Successional loblolly-thinned (LT)
Acer rubrum 72.4 0.0006
Pinus taeda 60.1 0.0014
Smilax laurifolia 41.9 0.0376
Liquidambar styraciflua 41.8 0.0344
S. Mitchell et al. / Forest Ecology and Management xxx (2015) xxx–xxx 7
Please cite this article in press as: Mitchell, S., et al. Patterns of vegetation composition and diversity in pine-dominated ecosystems of the Outer Coastal
Plain of North Carolina: Implications for ecosystem restoration. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.foreco.2015.07.035
likewise have low (1%) SOM (Markewitz et al., 2002).
Furthermore, the vegetation associated with fire-maintained lon-
gleaf pine stands is not limited to soils with high bulk density
and low (1%) SOM content. Other sites occurring in a more mesic
environment, such as longleaf pine-depressional wetland land-
scapes, have higher (6%) SOM in the upper portion of the soil pro-
file (Craft and Chiang, 2002). Higher SOM content can also be found
in mixed loblolly-longleaf pine forests, with as much as 8–11%
SOM in the upper (0–10 cm) portion of the soil profile (Binkley
et al., 1992). Median SOM content among longleaf pine stands
was rather low at 4%, although it was considerably higher in
some stands. Sites with 10–18% SOM still contained many of the
species that are highly correlative of fire-maintained longleaf pine
woodlands, including (A. stricta), creeping blueberry (Vaccinium
crassifolium), and dwarf huckleberry (Gaylussacia dumosa,
Table 6). Thus, while it appears that efforts to restore the
herb-layer of longleaf pine woodlands are more likely to succeed
on sites in which there is minimal SOM accumulation (<15%), a
successful restoration of vegetation in our study sites remains
uncertain and many require many years of monitoring before the
effectiveness of restoration treatments can be fully ascertained.
Re-introduction of fire to the landscape is essential to ongoing
restoration efforts, and the effects of a re-introduction of fire can
take many years to become apparent. Frequent burning (every 2–
3 years) can discourage the post-treatment sprouting and growth
of trees and shrubs (Brockway et al., 2009; Outcalt and
Brockway, 2010) while reducing the risks of high severity fire prop-
agation via mid-story ladder fuels (Agee and Skinner, 2005;
Table 7
Indicator species analysis results for species with a statistically significant association with at least one soil PCA Axis group. Positive (+) PCA axis 1 values were associated with
soils with high bulk density and low soil organic matter, while negative () soil PCA axis 1 value sare associated low bulk density and high soil organic matter. Positive (+) soil PCA
axis 2 values were primarily associated with higher PO
4
-P availability, while negative (i) values were associated with lower PO
4
-P availability.
Species PCA Axis 1 Group ISA Value Mean StDev pPCA Axis 2 Group ISA Value Mean StDev p
Acer rubrum 74.6 27.4 5.83 0.0001 36.5 27.2 5.73 0.0797
Andropogon spp. + 65 43.4 5.24 0.0013 + 37.4 43.4 5.13 0.9212
Aristida stricta + 70.7 34.9 4.56 0.0001 + 45.5 34.9 4.47 0.0277
Arundinaria tecta + 18.9 26.6 6.69 0.9054 55.3 26.4 6.58 0.0001
Aster spp. + 3.1 5.9 2.44 0.9096 12.1 6 2.46 0.0363
Cnidoscolus stimulosus + 36.3 20.3 4.61 0.0054 + 24 20.1 4.42 0.1852
Carphephorus bellidifolius + 37.2 20.2 4.59 0.0046 + 35 20.1 4.46 0.0075
Carphephorus odoratissimus + 63.5 32.9 5.08 0.0002 + 47.6 32.8 5 0.013
Carphephorus tomentosus + 17.4 9.8 3.38 0.0376 9.8 9.7 3.43 0.4609
Carex spp. 8.5 5.6 2.24 0.2965 12.1 5.5 2.3 0.0332
Dichanthelium mattamuskeetense + 3.5 8.3 3.07 0.9709 15 8.3 3.08 0.0316
Dichanthelium portoricense + 14.5 10.9 3.59 0.1764 + 23.8 10.8 3.48 0.003
Dichanthelium spp. 26.3 24.3 5.07 0.2941 56.6 24.2 4.94 0.0001
Diospyros virginiana + 47.4 28.4 5.23 0.0036 + 34.9 28.2 5.05 0.1148
Eupatorium capillifolium 16.1 12.9 4.14 0.212 21.6 12.9 4.06 0.0441
Euphorbia ipecacuanhae + 13.1 10.8 3.54 0.2333 + 21 10.6 3.41 0.0122
Gaylussacia dumosa + 73.8 35.7 6.01 0.0001 + 52.1 35.6 5.93 0.0134
Gaylussacia frondosa + 62.7 48.2 5.24 0.013 49.5 48 5.19 0.3264
Gentiana autumnalis + 23.9 12.2 3.98 0.0115 + 16.6 12.2 3.9 0.1712
Hypericum tenuifolium + 41.4 21 4.72 0.0025 15.7 20.9 4.64 0.9441
Ilex opaca 32.6 24.3 5.16 0.081 39.2 24.1 5.06 0.0131
Iris verna + 52.7 31 4.66 0.0009 + 54.2 31 4.66 0.0007
Kalmia carolina + 29.3 15.5 4.54 0.0102 21.3 15.4 4.36 0.1081
Liatris [pilosa +virgata] + 51.8 27.3 5.26 0.001 + 41.6 27.2 5.12 0.0128
Liquidambar styraciflua 45.1 30.1 5.99 0.0212 52.6 29.8 5.82 0.0013
Lyonia lucida 54.5 32.2 5.63 0.0021 + 37.3 32.1 5.55 0.17
Lyonia mariana + 36.4 21.2 4.88 0.0088 + 17 21.2 4.8 0.7969
Morella pumila + 21.1 14.3 4.25 0.0862 + 27.9 14.2 4.13 0.0083
Nyssa sylvatica 41.9 27.2 5.68 0.0159 49.8 27.1 5.56 0.0007
Osmunda cinnamomea 83.5 41.4 8.63 0.0001 67.5 41.1 8.71 0.0014
Osmunda regalis 15.8 9 3.22 0.0539 17.3 8.9 3.23 0.0234
Panicum virgatum 21.4 10.2 3.53 0.0081 19.3 10.1 3.42 0.0213
Persea palustris 77.9 51 6.48 0.0002 62.1 50.7 6.32 0.0586
Pinus serotina + 38.5 27.2 5.67 0.0456 + 39.1 27 5.56 0.0381
Pinus palustris + 70 36.2 5 0.0001 + 51.6 36.1 4.83 0.0072
Pinus taeda 59.5 33.9 5.35 0.0004 38.4 33.7 5.17 0.173
Pityopsis graminifolia + 58 30.2 5.5 0.0005 + 28.3 30.1 5.4 0.548
Pterocaulon pycnostachyum + 34.5 16.9 4.54 0.0015 + 26.8 16.9 4.44 0.0343
Quercus incana + 27.4 16.2 4.56 0.0292 + 21.2 16 4.38 0.1341
Quercus laevis + 43.1 22.2 5.61 0.0013 + 33.9 22.1 5.51 0.0317
Quercus marilandica + 21 12.6 3.82 0.0455 + 21.2 12.5 3.79 0.0366
Rhynchospora plumosa + 16.3 14.1 4.37 0.2887 + 28 14.1 4.28 0.0053
Rubus spp. + 5.8 6 2.42 0.7036 12.1 6 2.42 0.0332
Sassafras albidum + 41.1 26.9 6 0.0207 20 26.8 5.96 0.883
Sericocarpus tortifolius + 48.1 25 4.67 0.0003 + 31.1 24.9 4.57 0.1094
Smilax laurifolia 70.8 31 5.6 0.0001 47.7 31 5.58 0.0101
Toxicodendron radicans 13.5 7.5 2.94 0.0248 14.8 7.5 2.94 0.0315
Vaccinium crassifolium + 59.4 35.2 6.37 0.0019 + 55.9 35.2 6.31 0.004
Vaccinium formosum + 22.8 33.3 6.67 0.9814 46.5 33.2 6.5 0.0315
Vaccinium fuscatum + 30.8 28 5.27 0.266 41.9 27.8 5.17 0.0141
Vaccinium stamineum + 22.6 13.6 4.16 0.0404 + 28.5 13.5 4.02 0.0038
Vaccinium tenellum + 68.7 41.5 5.32 0.0002 + 47.5 41.3 5.28 0.1324
Vitus rotundifolia 26.9 20.7 5.16 0.1301 33.3 20.8 5.01 0.0145
Xyris caroliniana + 35.2 20.6 4.87 0.0118 + 16.6 20.5 4.81 0.7543
8S. Mitchell et al. / Forest Ecology and Management xxx (2015) xxx–xxx
Please cite this article in press as: Mitchell, S., et al. Patterns of vegetation composition and diversity in pine-dominated ecosystems of the Outer Coastal
Plain of North Carolina: Implications for ecosystem restoration. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.foreco.2015.07.035
Brockway et al., 2009). Frequent burning can also facilitate the
recolonization of the vegetation found in longleaf pine woodlands.
Grasses such as A. stricta and Andropogon spp. are sensitive to duff
accumulation and can increase in coverage following the combus-
tion of duff materials (Hiers et al., 2007), but their post-fire recov-
ery is not immediate. Outcalt and Brockway (2010) examined a
xeric longleaf pine forest in Alabama with a history of fire suppres-
sion to ascertain the response of herb-layer vegetation to burning
and thinning treatments. While grass species cover shortly after
prescribed fire was minimal, after 3 years they observed an
increase in grass species on the burned, thinned, and bur-
ned + thinned plots, with the largest gains in the burned + thinned
plots. After 6 years following the initial restoration treatments, the
thinning without burning treatment had 14% grass coverage, while
all of the treatments that included burning had approximately 30%
more grass coverage than when the study sites were first treated.
Thus, some of the more tangible effects of restoration treatments
may require several years to become apparent.
7.6. Summary
On the lower coastal plain of North Carolina, loblolly pine grows
on a wide variety of soil moisture conditions. Attempts to restore
understory composition and species richness typical of longleaf
pine savannas should be focused on the subset of loblolly pine
stands on the mesic to xeric portion of that soil moisture gradient.
Historical legacies of past agriculture such as higher concentrations
of calcium and phosphate are evident in some loblolly soils and
their potential effect on restoration should also be evaluated
(Brudvig et al., 2013). As might be expected from the apparent
habitat preferences exhibited by many plant species, all five of
our stand types (DPOS, WPS, LT, LU, and HPO) exhibited statisti-
cally significant differences from each other as indicated by the
MRPP analysis. However, there was some apparent overlap
between longleaf and loblolly pine stands with similar levels of
SOM. An overlap in herb-layer composition of longleaf and loblolly
pine stands may suggest some potential for restoring longleaf pine
vegetation in loblolly pine stands.
While the long-term success of implementing restoration treat-
ments in loblolly pine stands remains uncertain, our analysis
suggests that the application of mid-story thinning restoration
treatments to loblolly stands is more likely to shift the composition
of herbaceous plant communities toward that of longleaf pine
woodlands when performed on sites with low (i.e. <15%) amounts
of SOM. Forests that have only recently accumulated marginal
amounts of SOM as a direct result of fire suppression may be
potential candidates for the restoration of the plant communities
native to longleaf pine woodlands (Kirkman et al., 2013), as many
such forests may still have viable seed banks. Stands that are in
close proximity to high-quality longleaf pine stands may also war-
rant additional consideration for having the additional benefit of
seed availability. With midstory thinning and the reintroduction
of fire, herb composition and diversity in longleaf stands with a
history of fire suppression show definite restoration progress in
5–10 years (Outcalt and Brockway, 2010; Kirkman et al., 2013).
Given the legacy of land use that has removed much of the preset-
tlement flora and its seedbank, and the buildup of duff and soil
organic matter associated with fire suppression, restoration in
loblolly pine stands will likely take much longer.
Acknowledgements
This research was conducted under the Department of Defense
Coastal/Estuarine Research Program (DCERP, Project RC-1413)
funded by a grant from the Strategic Environmental Research and
Development Program (SERDP) to Norman L. Christensen. We also
thank Ezekiel Overbaugh, Courtney Spears, and Rachel Roberts for
their considerable fieldwork efforts. Finally, we thank land man-
agers at Camp Lejeune Marine Corps Base for access to field sites
and their assistance with research logistics.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.foreco.2015.07.
035.
References
Agee, J., Skinner, C., 2005. Basic principles of forest fuel reduction treatments. For.
Ecol. Manage. 211, 83–96.
Fig. 3. Herb-layer species richness plotted against soil PCA Axis 1 and NMS Axis 1. High pocosin (HPO) study sites had the lowest levels of species richness, while wet pine
savanna (WPS) sites had the highest species richness. Some longleaf pine stands had exceptionally high levels of species richness (>60 species) that weaken the strength of the
linear correlation between species richness and PCA Axis 1 and NMS Axis 1.
S. Mitchell et al. / Forest Ecology and Management xxx (2015) xxx–xxx 9
Please cite this article in press as: Mitchell, S., et al. Patterns of vegetation composition and diversity in pine-dominated ecosystems of the Outer Coastal
Plain of North Carolina: Implications for ecosystem restoration. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.foreco.2015.07.035
Ashe, W.W., 1915. Loblolly, Or North Carolina Pine. Edwards & Broughton Printing
Company, state printers and binders.
Barnhill, W., 1992. Soil survey of Onslow County, NC. U.S. Department of
Agriculture, Soil Conservation Service, Washington.
Binkley, D., Richter, D., David, M.B., Caldwell, B., 1992. Soil chemistry in a
loblolly/longleaf pine forest with interval burning. Ecol. Appl., 157–164.
Brockway, D.G., Lewis, C.E., 1997. Long-term effects of dormant-season prescribed
fire on plant community diversity, structure and productivity in a longleaf pine
wiregrass ecosystem. For. Ecol. Manage. 96, 167–183.
Brockway, D.G., Outcalt, K.W., 1998. Gap-phase regeneration in longleaf pine
wiregrass ecosystems. For. Ecol. Manage. 106, 125–139.
Brockway, D.G., Outcalt, K.W., Estes, B.L., Rummer, R.B., 2009. Vegetation response
to midstorey mulching and prescribed burning for wildfire hazard reduction
and longleaf pine (Pinus palustris Mill.) ecosystem restoration. Forestry 82,
299–314.
Brudvig, L.A., Grman, E., Habeck, C.W., Orrock, J.L., Ledvina, J.A., 2013. Strong legacy
of agricultural land use on soils and understory plant communities in longleaf
pine woodlands. For. Ecol. Manage. 310, 944–955.
Carr, S.C., Robertson, K.M., Platt, W.J., Peet, R.K., 2009. A model of geographical,
environmental and regional variation in vegetation composition of pyrogenic
grasslands of Florida. J. Biogeogr. 36, 1600–1612.
Christensen, N., 1981. Fire regimes in southeastern ecosystems. Fire Regimes
Ecosyst. Properties 112, 136.
Christensen, N.L., 1988. Vegetation of the southeastern coastal plain. In: Barbour,
M.G., Billings, W.D. (Eds.), North American Terrestrial Vegetation. Cambridge
University Press, Cambridge, pp. 317–363.
Christensen, N.L., 2000. Vegetation of the Coastal Plain of the southeastern United
States. In: Barbour, M., Billings, W.D. (Eds.), Vegetation of North America,
second ed. Cambridge University Press, Cambridge, United Kingdom,
pp. 397–448.
Cohen, S., Braham, R., Sanchez, F., 2004. Seed bank viability in disturbed longleaf
pine sites. Restor. Ecol. 12, 503–515.
Craft, C.B., Chiang, C., 2002. Forms and amounts of soil nitrogen and phosphorus
across a longleaf pine–depressional wetland landscape. Soil Sci. Soc. Am. J. 66,
1713–1721.
Dixon, P., Palmer, M., 2003. VEGAN, a package of R functions for community
ecology. J. Veg. Sci. 14, 927–930.
Drewa, P.B., Platt, W.J., Moser, E.B., 2002. Community structure along elevation
gradients in headwater regions of longleaf pine savannas. Plant Ecol. 160,
61–78.
Dufrêne, M., Legendre, P., 1997. Species assemblages and indicator species: the need
for a flexible asymmetrical approach. Ecol. Monogr. 67, 345–366.
Frost, C., 2006. History and future of the longleaf pine ecosystem. In: The Longleaf
Pine Ecosystem. Springer, pp. 9–48.
Frost, C.C., Walker, J., Peet, R.K., 1986. Fire-dependent savannas and prairies of the
Southeast: original extent, preservation status and management problems.
Wilderness and Natural Areas in the Eastern United States: A Management
Challenge. Stephen F. Austin State University, School of Forestry, Center for
Applied Studies, Nacogdoches, TX, pp. 348–357.
Garren, K.H., 1943. Effects of fire on vegetation of the southeastern United States.
Bot. Rev. 9, 617–654.
Glitzenstein, J.S., Streng, D.R., Wade, D.D., 2003. Fire Frequency Effects on Longleaf
Pine (Pinus palustris P. Miller) Vegetation in South Carolina and Northeast
Florida, USA.
Harding, S.R., Walters, J.R., 2002. Processes regulating the population dynamics of
red-cockaded woodpecker cavities. J. Wildl. Manage., Wiley on behalf of the
Wildlife Society
Hedman, C., Grace, S., King, S., 2000. Vegetation composition and structure of
southern coastal plain pine forests: an ecological comparison. For. Ecol. Manage.
134, 233–247.
Heyward, F., 1939. The relation of fire to stand composition of longleaf pine forests.
Ecology 20, 287–304.
Hiers, J.K., O’Brien, J.J., Will, R.E., Mitchell, R.J., 2007. Forest floor depth mediates
understory vigor in xeric Pinus palustris ecosystems. Ecol. Appl. 17, 806–814.
Kirkman, L.K., Barnett, A., Williams, B.W., Hiers, J.K., Pokswinski, S.M., Mitchell, R.J.,
2013. A dynamic reference model: a framework for assessing biodiversity
restoration goals in a fire-dependent ecosystem. Ecol. Appl. 23, 1574–1587.
Kirkman, L.K., Goebel, P.C., Palik, B.J., West, L.T., 2004. Predicting plant species
diversity in a longleaf pine landscape. Ecoscience 11, 80–93.
Knapp, B.O., Wang, G.G., Hu, H., Walker, J.L., Tennant, C., 2011. Restoring longleaf
pine (Pinus palustris Mill.) in loblolly pine (Pinus taeda L.) stands: effects of
restoration treatments on natural loblolly pine regeneration. For. Ecol. Manage.
262, 1157–1167.
Kologiski, R.L., 1977. The phytosociology of the Green Swamp, North Carolina
[Pinus, grass and forest systems]. Technical Bulletin-North Carolina Agricultural
Experiment Station (USA).
Korkmaz, S., Goksuluk, D., Zararsiz, G., 2014. MVN: An R Package for Assessing
Multivariate Normality. A Peer-reviewed, Open-access Publication of the R
Foundation for Statistical Computing, p. 151.
Kruskal, J., 1964. Nonmetric multidimensional scaling: a numerical method.
Psychometrika 29, 115–129.
Landers, J.L., Van Lear, D.H., Boyer, W.D., 1995. The longleaf pine forests of the
Southeast: requiem or renaissance? J. Forest. 93, 38–44.
Lewis, C.E., Harshbarger, T.J., 1976. Shrub and herbaceous vegetation after 20 years
of prescribed burning in the South Carolina Coastal Plain. J. Range Manage., 13–
18
Mardia, K.V., 1970. Measures of multivariate skewness and kurtosis with
applications. Biometrika 57, 519–530.
Markewitz, D., Sartori, F., Craft, C., 2002. Soil change and carbon storage in longleaf
pine stands planted on marginal agricultural lands. Ecol. Appl. 12, 1276–1285.
MCBCL, 2006. Integrated Natural Resource Management Plan (INRMP). US Marine
Corps, Camp Lejeune, NC.
McCune, B., Grace, J.B., 2002. Analysis of Ecological Communities. MjM Software
Design, Gleneden Beach.
McCune, B., Mefford, M.J., 2002. PC-ORD. Multivariate Analysis of Ecological Data.
Version 6.247 beta. MjM Software, Gleneden Beach, Oregon, U.S.A..
McLean, E., 1982. Soil pH and lime requirement. Methods of Soil Analysis. Part 2.
Chemical and Microbiological Properties, pp. 199–224.
Mehlich, A., 1984. Mehlich 3 soil test extractant: a modification of Mehlich 2
extractant. Commun. Soil Sci. Plant Anal. 15, 1409–1416.
Mielke, P.W., Berry, K.J., 2007. Permutation Methods: A Distance Function
Approach. Springer.
Noss, R.F., LaRoe, E.T., Scott, J.M., 1995. Endangered Ecosystems of the United States:
A Preliminary Assessment of Loss and Degradation. US Department of the
Interior, National Biological Service, Washington, DC, USA.
Noss, R.F., Platt, W.J., Sorrie, B.A., Weakley, A.S., Means, D.B., Costanza, J., Peet, R.K.,
2015. How global biodiversity hotspots may go unrecognized: lessons from the
North American Coastal Plain. Divers. Distrib. 21, 236–244.
Outcalt, K.W., Brockway, D.G., 2010. Structure and composition changes following
restoration treatments of longleaf pine forests on the Gulf Coastal Plain of
Alabama. For. Ecol. Manage. 259, 1615–1623.
Owen, W., 2002. The history of native plant communities in the South. Southern
forest resource assessment. In: Wear, D.N., Greis, J.G. (Eds.), USDA For. Serv.
Gen. Tech. Rep. GTR-SRS-53, pp. 47–61.
Peet, R.K., 2006. Ecological classification of longleaf pine woodlands. In: The
Longleaf Pine Ecosystem. Springer, pp. 51–93.
Peet, R.K., Palmquist, K.A., Tessel, S.M., 2014. Herbaceous layer species richness of
southeastern forests and woodlands: patterns and causes. The Herbaceous
Layer in Forests of Eastern North America, pp. 255–276.
Peet, R.K., Wentworth, T.R., White, P.S., 1998. A flexible, multipurpose method for
recording vegetation composition and structure. Castanea 63.
Provencher, L., Herring, B.J., Gordon, D.R., Rodgers, H.L., Galley, K.E.M., Tanner, G.W.,
Hardesty, J.L., Brennan, L.A., 2001. Effects of hardwood reduction techniques on
longleaf pine sandhill vegetation in northwest Florida. Restor. Ecol. 9, 13–27.
Richter, D.D., Allen, H.L., Li, J., Markewitz, D., Raikes, J., 2006. Bioavailability of
slowly cycling soil phosphorus: major restructuring of soil P fractions over four
decades in an aggrading forest. Oecologia 150, 259–271.
Rodgers, H.L., Provencher, L., 1999. Analysis of longleaf pine sandhill vegetation in
northwest Florida. Castanea, 138–162.
Ross, D.S., 1995. Recommended Methods for Determining Soil Cation Exchange
Capacity. College of Agriculture, University of Delaware, Newark.
Schafale, M.P., 2012. Guide to the natural communities of North Carolina, Fourth
Approximation. March 2012. NC Department of Natural Resources, Natural
Heritage Program.
Schultz, R.P., 1999. Loblolly – the pine for the twenty-first century. New Forest. 17,
71–88.
Schwilk, D.W., Keeley, J.E., Knapp, E.E., McIver, J., Bailey, J.D., Fettig, C.J., Fiedler, C.E.,
Harrod, R.J., Moghaddas, J.J., Outcalt, K.W., Skinner, C.N., Stephens, S.L., Waldrop,
T.A., Yaussy, D.A., Youngblood, A., 2009. The national Fire and Fire Surrogate
study: effects of fuel reduction methods on forest vegetation structure and
fuels. Ecol. Appl. 19, 285–304.
Steen, D.A., Conner, L., Smith, L.L., Provencher, L., Hiers, J.K., Pokswinski, S., Helms,
B.S., Guyer, C., 2013. Bird assemblage response to restoration of fire-suppressed
longleaf pine sandhills. Ecol. Appl. 23, 134–147.
Varner, J.M., Gordon, D.R., Putz, E., Hiers, J.K., 2005. Restoring fire to long-unburned
Pinus palustris ecosystems: novel fire effects and consequences for long-
unburned ecosystems. Restor. Ecol. 13, 536–544.
Varner, J.M., Hiers, J.K., Ottmar, R.D., Gordon, D.R., Putz, F.E., Wade, D.D., 2007.
Overstory tree mortality resulting from reintroducing fire to long-unburned
longleaf pine forests: the importance of duff moisture. Can. J. For. Res. 37,
1349–1358.
Varner, J.M., Putz, F.E., O’Brien, J.J., Hiers, J.K., Mitchell, R.J., Gordon, D.R., 2009. Post-
fire tree stress and growth following smoldering duff fires. For. Ecol. Manage.
258, 2467–2474.
Wahlenberg, W.G., 1960. Loblolly Pine. Its Use, Ecology, Regeneration, Protection,
Growthand Management.Duke UniversitySchool of Forestry, Durham, NC, 603 pp.
Walker, J., Peet, R.K., 1984. Composition and species diversity of pine-wiregrass
savannas of the Green Swamp, North Carolina. Plant Ecol. 55, 163–179.
Walker, J.L., Silletti, A.M., 2006. Restoring the ground layer of longleaf pine
ecosystems. In: The Longleaf Pine Ecosystem. Springer, pp. 297–333.
Wear, D.N., Greis, J.G., 2002. Southern forest resource assessment: summary of
findings. J. Forest. 100, 6–14.
Wells, B.W., 1932. The Natural Gardens of North Carolina. University of North
Carolina Press, Chapel Hill, North Carolina, USA.
Woodwell, G.M., 1958. Factors controlling growth of pond pine seedlings in organic
soils of the Carolinas. Ecol. Monogr., 220–236.
10 S. Mitchell et al. / Forest Ecology and Management xxx (2015) xxx–xxx
Please cite this article in press as: Mitchell, S., et al. Patterns of vegetation composition and diversity in pine-dominated ecosystems of the Outer Coastal
Plain of North Carolina: Implications for ecosystem restoration. Forest Ecol. Manage. (2015), http://dx.doi.org/10.1016/j.foreco.2015.07.035
... The study sites spanned an elevation gradient from 13.5 to 19.7 m. Corresponding to these elevation differences, distinct vegetation gradients were evident, ranging from pocosins-like habitats dominated by Cyperaceae and pocosins indicator shrubs (Mitchell et al., 2015) Mitchell et al. (2015). ...
... The study sites spanned an elevation gradient from 13.5 to 19.7 m. Corresponding to these elevation differences, distinct vegetation gradients were evident, ranging from pocosins-like habitats dominated by Cyperaceae and pocosins indicator shrubs (Mitchell et al., 2015) Mitchell et al. (2015). ...
... This community is characterized by an overstorey of longleaf pine and an understorey dominated by bunchgrasses but also containing a diverse array of grasses, sedges, forbs, and firedwarfed shrubs (Frost, 2001). Mitchell et al. (2015) identified three pine-dominated, natural community types on Camp Lejeune -dry pine/oak savanna, wet pine savanna, and high pocosin -which are arranged along a moisture gradient from xeric to hydric, respectively. ...
Article
Question Longleaf pine (Pinus palustris) restoration is an important management objective throughout the southeastern U.S. Site preparation prior to planting longleaf pine seedlings is often required to overcome challenges to seedling establishment on hydric sites. We know site preparation improves growth and survival of planted longleaf pine on hydric sites, but what are the impacts of site preparation on the understory plant community? Location Marine Corps Base Camp Lejeune in Onslow County, North Carolina, U.S. Methods This study tested eight site preparation treatments, including an untreated control, six combinations of two vegetation control treatments (chopping or herbicide) with three soil manipulation treatments (mounding, bedding, or flat-planting [no treatment]), and a chopping-herbicide-bedding treatment. We collected data on understory plant abundance, diversity, and composition through the first 3 years and at 15 years after plantation establishment. We used ANOVA procedures and multivariate ordination to test for differences in understory responses among the study treatments and through time. Results Site preparation had lasting impacts on the understory plant community. The chop-only treatment was the only treatment that resulted in dominance of herbaceous vegetation and was among the treatments with the highest diversity and greatest abundance of Aristida stricta (wiregrass). Herbicide produced lasting reductions in A. stricta abundance and apparent shifts in community composition. Reduced graminoid abundance was particularly pronounced when herbicide was combined with soil manipulation treatments. Conclusions Understory responses to site preparation through 3 years differed from responses observed in year 15, highlighting the importance of long-term monitoring. For example, herbicides controlled shrubs early, but by year 15, herbicide treatments were dominated by shrubs. Restoration scenarios often must balance ecological objectives (e.g., maintenance of a diverse understory community) with forestry objectives (e.g., timely establishment of overstory trees), and our findings can be used as a guide to balance these tradeoffs.
... The Terrestrial Module focused on management strategies to restore and enhance RCW habitat, enhance carbon storage, as well as maintain conditions suitable for military training. The results of field studies and model simulations informed MCBCL management prescriptions to meet these goals (Mitchell et al., 2015;Walters et al., 2017). Furthermore, such management was shown to be ideal for maintaining overall avian species diversity (Walters et al., 2013). ...
Article
The Defense Coastal/Estuarine Research Program (DCERP) was a 10-year multi-investigator project funded by the Department of Defense to improve understanding of ecosystem processes and their interactions with natural and anthropogenic stressors at the Marine Corps Base Camp Lejeune (MCBCL) located in coastal North Carolina. The project was aimed at facilitating ecosystem-based management (EBM) at the MCBCL and other coastal military installations. Because of its scope, interdisciplinary character, and duration, DCERP embodied many of the opportunities and challenges associated with EBM, including the need for explicit goals, system models, long-term perspectives, systems complexity, change inevitability, consideration of humans as ecosystem components, and program adaptability and accountability. We describe key elements of this program, its contributions to coastal EBM, and its relevance as an exemplar of EBM.
... Differences between plant community compositions were determined using a multi-response permutation procedure (MRPP) with the Sørensen or Bray-Curtis distance method. MRPP is a nonparametric procedure used for testing hypotheses between two or more groups (Mitchell et al., 2015). Differences in community compositions were analyzed for all plant communities, and pairwise comparisons were generated. ...
Article
Full-text available
Black-tailed prairie dogs (Cynomys ludovicianus) have been described as a keystone species and are important for grassland conservation, yet many concerns exist over the impact of prairie dogs on plant biomass production and consequently livestock production. The ability to map plant communities in pastures colonized by prairie dogs can provide land managers with an opportunity to optimize rangeland production while balancing conservation goals. The aim of this study was to test the ability of random forest (RF) models to classify five plant communities located on and off prairie dog towns in mixed-grass prairie landscapes of north central South Dakota, assess the stability of RF models among different years, and determine the utility of utilizing remote sensing techniques to identify prairie dog colony extent. During 2015 and 2016, Pleiades satellites were tasked to image the study site for a total of five monthly collections each summer (June–October). Training polygons were mapped in 2016 for the five plant communities and used to train RF models. Both the 2015 and 2016 RF models had low (1 %) out-of-bag error rates. However, comparisons between the predicted plant community maps using the 2015 imagery and one created with the 2016 imagery indicate over 32.9 % of pixels changed plant community class between 2015 and 2016. The results show that while RF models may predict with a high degree of accuracy, overlap of plant communities and interannual differences in rainfall may cause instability in fitted models. A final RF model combining both 2015 and 2016 data yielded the lowest error rates and was also highly accurate in determining prairie dog colony boundaries.
... Many factors would restrain vegetation restoration of degraded wetlands, including water supplement (Guo et al., 2017), nutrient availability (Mitchell et al., 2015), cultivation period (Barrett and Watmough, 2015), seed banks (Liu et al., 2005;Bai et al., 2014), restoration time and so on (Lemein et al., 2017). Usually, vegetation would be more and more similar with target wetlands as restoration time increasing. ...
Article
Deyeuxia angustifolia wetlands were widely distributed in the Sanjiang Plain in Northeast China. Due to strong demand for food production, large-area wetlands were reclaimed to farmlands, which threatened regional ecological security greatly. Since the 21th century, returning farmlands to wetlands was widely adopted for natural restoration in the Sangjiang Plain. As the first reflection of wetland restoration, vegetation succession of restored D. angustifolia wetlands should be fully assessed. In this study, vegetation investigation was carried out in three restored D. angustifolia wetlands with 5, 8 and 12 yr restoration, respectively. Meanwhile, a natural D. angustifolia wetland was selected as reference wetland. Results showed that community composition changed greatly and there was visible community succession. Community dominant species changed from composite to gramineae as restoration time increasing. At first, weeds community appeared in the restored wetlands, especially the xerophytes developed to the pioneer species rapidly. And then, mesophytes and wetland species became the dominant species in the restored wetlands. Finally, wetland species, especially D. angustifolia, occupied the dominant position of restored community. Shannon-wiener index (H) and Simpson index (D) both decreased to close to natural D. angustifolia wetlands. Compared with natural D. angustifolia wetland, species composition and diversity in restored wetlands were more complex and higher. As restoration time increasing, there were not significant differences between community characteristics of restored wetlands and natural wetland. All these suggested that vegetation in reclaimed D. angustifolia wetland could be restored naturally, but its restored period is 10 yr at least. From another angle, it is important to protect current natural wetlands.
... Many factors would restrain vegetation restoration of degraded wetlands, including water supplement ( Guo et al., 2017), nutrient availability ( Mitchell et al., 2015), cultivation period ( Barrett and Watmough, 2015), seed banks ( Liu et al., 2005;Bai et al., 2014), restoration time and so on ( Lemein et al., 2017). Usually, vegetation would be more and more similar with target wetlands as restoration time increasing. ...
Article
Deyeuxia angustifolia wetlands were widely distributed in the Sanjiang Plain in Northeast China. Due to strong demand for food production, large-area wetlands were reclaimed to farmlands, which threatened regional ecological security greatly. Since the 21th century, returning farmlands to wetlands was widely adopted for natural restoration in the Sangjiang Plain. As the first reflection of wetland restoration, vegetation succession of restored D. angustifolia wetlands should be fully assessed. In this study, vegetation investigation was carried out in three restored D. angustifolia wetlands with 5, 8 and 12 yr restoration, respectively. Meanwhile, a natural D. angustifolia wetland was selected as reference wetland. Results showed that community composition changed greatly and there was visible community succession. Community dominant species changed from composite to gramineae as restoration time increasing. At first, weeds community appeared in the restored wetlands, especially the xerophytes developed to the pioneer species rapidly. And then, mesophytes and wetland species became the dominant species in the restored wetlands. Finally, wetland species, especially D. angustifolia, occupied the dominant position of restored community. Shannon-wiener index (H) and Simpson index (D) both decreased to close to natural D. angustifolia wetlands. Compared with natural D. angustifolia wetland, species composition and diversity in restored wetlands were more complex and higher. As restoration time increasing, there were not significant differences between community characteristics of restored wetlands and natural wetland. All these suggested that vegetation in reclaimed D. angustifolia wetland could be restored naturally, but its restored period is 10 yr at least. From another angle, it is important to protect current natural wetlands.
... Our literature review indicates that mid-rotation management (e.g., prescribed fire, selective herbicide) can extend the duration of high-quality open pine woodlands. Soils, previous land use, and disturbance history continue to be important factors that can facilitate or inhibit open pine restoration and conservation efforts on private and public lands (Hedman et al., 2000;Mitchell et al., 2015;Brudvig et al., 2013). Prioritization of landscapes for open pine restoration and maintenance efforts should consider proximity to neighboring open pine woodlands, legacy of agricultural land use on soils and plant communities, degree of subdivision, ownership type, and management goals. ...
Article
Pine forests managed for economic return comprise 19% of southeastern U. S. forests and can provide vegetation structure and composition necessary for many wildlife species adapted to open pine conditions. However, in the absence of frequent fire and other disturbance in southeastern pine (Pinus spp.) forests, midstory hardwoods create closed canopy conditions limiting value of pine stands for many endemic, disturbance-adapted wildlife species. Limitations of past research investigating exact mechanisms for producing and maintaining open pine conditions in managed pine systems have led to a paucity of information regarding ecoregion applications and outcomes. This study evaluated (1) to what extent managed pine stands replicate open pine structural conditions and (2) to what extent managed pine stands provide habitat for the gopher tortoise, an open pine keystone species. We used summary statistics from existing literature to identify structural metrics provided by managed loblolly pine (P. taeda) stands and compared them to disturbance-maintained open pine woodlands. We used spatially-explicit forest management simulations to determine amount and configuration of open pine woodland conditions within two large (>9000 ha), economically-feasible managed pine landscapes in the East and West Gulf Coastal Plains ecoregions. As a case study, we estimated availability and connectivity of gopher tortoise (Gopherus polyphemus) habitat, in managed pine stands of the East Gulf Coastal Plains. Pine stands planted at ≤1310 stems/ha (≤530 stems/ac) provide up to four years of “excellent” open pine condition for wildlife following commercial thinning, based on desired conditions for southern pine forests, with temporal availability depending on site quality. Mid-rotation applications of prescribed fire and selective herbicide can extend “excellent” open pine condition initiated by commercial thinning up to an additional four years. Landscape metrics indicated that, under typical management practices, open pine stands can be spatially and temporally connected to the extent necessary to facilitate among-patch movement for many open pine-adapted species of conservation concern. Our landscape analyses indicated tortoises using primarily regenerating and thinned stands potentially can have a relatively even flow of suitable habitat conditions and opportunities to move to adjacent or nearby open stands when currently-occupied stands reach canopy closure. The disturbance regimes inherent in pine plantation management create diverse landscapes that conserve plant and wildlife biodiversity while meeting economic objectives. Therefore, pine plantations may complement current conservation efforts for open pine woodland species dependent on or tolerant of large-scale disturbance and shifting habitat conditions while offering refuge for high vagility species.
Article
Longleaf pine (Pinus palustris) savanna characterized by open‐canopy, diverse herbaceous vegetation, and high amounts of bare soil once covered much of the southeastern United States Coastal Plain. The unique structural and vegetative conditions of this ecosystem support endemic reptiles and amphibians that have declined as longleaf pine forests have been lost or degraded. Private working pine (Pinus spp.) forests managed for timber production now occur throughout the southeastern United States and have replaced much of the historical longleaf pine savanna. The examination of herpetofaunal (reptile, amphibian) communities in private working loblolly pine (P. taeda) landscapes, particularly in the western Gulf Coastal Plain is lacking. Using repeated field surveys and hierarchical community occupancy models, we examined occupancy and species richness of herpetofauna across 81 sites spanning gradients of management practices, vegetative conditions, and soil composition in northwestern Louisiana, USA, 2017–2019. Young pine stands (<6 yr) exhibited structural characteristics most similar to mature longleaf pine reference sites (>30 yr), while mid‐aged stands (13–26 yr) often featured closed canopy and dense midstory. Vegetation conditions varied widely depending on landscape characteristics and site‐specific disturbance regimes. We documented 43 species of herpetofauna, including 9 open‐pine‐associated species. Occupancy of open‐pine‐associated herpetofauna was positively associated with open‐canopy and understory conditions, and sandy soil area. Sites providing open‐canopy conditions were often occupied by open‐pine‐associated species regardless of overstory type and disturbance method. Overall richness of herpetofauna was greatest at sites with moderate canopy cover outside of sandy soil regions. Working pine landscapes in the western Gulf Coastal Plain can support diverse herpetofaunal assemblages, including open‐pine‐associated species, when management practices maintain open‐canopy conditions on sandy, upland soils. More broadly, our results provide insight into how forest management practices affect herpetofauna and may guide practices that can contribute to conservation value of working pine forests. This study highlights the potential and limitations of pine forests under varying management regimes and featuring different landscape characteristics to support a range of herpetofaunal species. This research will guide management decisions targeting herpetofaunal conservation on a regional and site‐specific basis by illustrating the influence of site‐specific structural conditions and larger‐scale landscape characteristics on an area's ability to support a variety of species.
Article
Full-text available
Biodiversity hotspots are conservation priorities. We identify the North American Coastal Plain (NACP) as a global hotspot based on the classic definition, a region with > 1500 endemic plant species and > 70% habitat loss. This region has been bypassed in prior designations due to misconceptions and myths about its ecology and history. These fallacies include: (1) young age of the NACP, climatic instability over time and submergence during high sea-level stands; (2) climatic and environmental homogeneity; (3) closed forest as the climax vegetation; and (4) fire regimes that are mostly anthropogenic. We show that the NACP is older and more climatically stable than usually assumed, spatially heterogeneous and extremely rich in species and endemics for its range of latitude, especially within pine savannas and other mostly herbaceous and fire-dependent communities. We suspect systematic biases and misconceptions, in addition to missing information, obscure the existence of similarly biologically significant regions world-wide. Potential solutions to this problem include (1) increased field biological surveys and taxonomic determinations, especially within grassy biomes and regions with low soil fertility, which tend to have much overlooked biodiversity; (2) more research on the climatic refugium role of hotspots, given that regions of high endemism often coincide with regions with low velocity of climate change; (3) in low-lying coastal regions, consideration of the heterogeneity in land area generated by historically fluctuating sea levels, which likely enhanced opportunities for evolution of endemic species; and (4) immediate actions to establish new protected areas and implement science-based management to restore evolutionary environmental conditions in newly recognized hotspots.
Article
Full-text available
In this study, we used a hierarchical, multifactor ecological classification system to examine how spatial patterns of biodiversity develop in one of the most species-rich ecosystems in North America, the fire-maintained longleaf pine-wiregrass ecosystem and associated depressional wetlands and riparian forests. Our goal was to determine which landscape features are important controls on species richness, to establish how these constraints are expressed at different levels of organization, and to identify hotspots of biological diversity for a particular locality. We examine the following questions: 1) How is the variance in patterns of plant species richness and diversity partitioned at different scales, or classification units, of the hierarchical ecosystem classification developed for the study area? 2) What are the compositional similarities among ecosystem types? 3) For our study area, what are the sites expected to harbour highest species richness? We used a spatially explicit map of biodiversity to project abundance of species-rich communities in the landscape based on a previously developed ecological classification system for a lower Gulf Coastal Plain landscape. The data indicate that high species richness in this ecosystem was found in sites with frequent fire and high soil moisture. Sites in fire-maintained landscapes with lower frequency of fire were associated with geomorphological characteristics, suggesting a dependence of the diversity-disturbance relationship with soil type. With more frequent fire on some sites, high diversity shifts from canopy component to ground flora, with an overall increase in total species richness. Our approach demonstrates how potential species richness can be identified as a restoration goal and that multiple vegetation endpoints may be appropriate vegetation objectives. We identify basic management needs for the maintenance of biodiversity in this ecosystem that can be derived from an understanding of the combination of factors that most strongly predict diverse plant communities.
Article
Plant communities vary from grassland and savanna through shrubland to needle- and broadleaved sclerophyllous woodland and rich mesophytic forest, differences which can occur within a few hundred metres and an elevational shift of only 10m. Dramatic gradients in physical and chemical characteristics of soil and water, together with fire, lead to such a mosaic pattern. Vegetation history is indicated and the physical background outlined. Upland pine forest vegetation is described - northern pine barrens, xeric sand communities and mesic pine communities. Upland hardwoods are noted, and the vegetation of wetlands is discussed, distinguishing between alluvial and non-alluvial types. Interspersion and gradation with maritime vegetation is shown. Community and gap dynamics are reviewed, and the significance of succession is stressed. -P.J.Jarvis