ICU ventilator Monitors to buy viagra without rx display ECG, capnography, oxygen saturation etc. It is important discount generic cialis to seek medical assistance to correct sexual concerns that can affect a relationship. Ultimately, it could be good alternate to switch the medicine, if they have budget issues. cialis canadian generic Organic reasons for impotence involve the physical health of the reproductive organs involved with 100mg viagra price an erection.
Invasions by two non-native insects alter regional forest species
composition and successional trajectories
Randall S. Morin
a
,
⇑
, Andrew M. Liebhold
b
a
USDA Forest Service, Northern Research Station, 11 Campus Blvd., Suite 200, Newtown Square, PA 19073, United States
b
USDA Forest Service, Northern Research Station, 180 Canfield St, Morgantown, WV 26505, United States
article info
Article history:
Received 17 July 2014
Received in revised form 17 December 2014
Accepted 18 December 2014
Keywords:
Adelges tsugae
Eastern hemlock
Tsuga canadensis
Beech bark disease
Fagus grandifolia
Growth and mortality rates
abstract
While invasions of individual non-native phytophagous insect species are known to affect growth and
mortality of host trees, little is known about how multiple invasions combine to alter forest dynamics
over large regions. In this study we integrate geographical data describing historical invasion spread of
the hemlock woolly adelgid,
Adelges tsugae
, and beech scale,
Cryptococcus fagisuga
, with regional forest
inventory data collected by the US Forest Service’s Forest Inventory and Analysis program to quantify
the individual and combined impacts of these pest species. This analysis indicates that regional impacts
of these insects on their hosts occur surprisingly slow but act to change regional forest succession path-
ways. Because beech and hemlock commonly co-occur in eastern North American forests, invasions by
the two pest species are altering the current and future composition of large forest regions through their
impacts on these two late-successional species. Such results demonstrate how forest insect invasions can
profoundly modify forest dynamic processes, resulting in long-term changes in forest ecosystems.
Published by Elsevier B.V.
1. Introduction
Invasions by non-native insects and pathogens are major causes
of disturbance, affecting the stability, productivity, and economic
value of forest ecosystems worldwide (
Liebhold et al., 1995;
Holmes et al., 2009; Aukema et al., 2011
). Most invading forest
insects and diseases are not particularly abundant and conse-
quently have negligible effects, but a few have altered forest eco-
systems in profound ways (
Niemela and Mattson, 1996; Aukema
et al., 2010
). Over the last century, a large number of invasive spe-
cies have become established in forests of eastern North America
(
Liebhold et al., 2013
) and some of these organisms, such as chest-
nut blight, emerald ash borer, and beech bark disease have caused
extensive tree mortality.
Forest insect and pathogen invasions can affect forest commu-
nities in a multitude of ways, acting both directly and indirectly
(
Lovett et al., 2006; Loo, 2009
). Such effects include changes in tree
species composition (
Fajvan and Wood, 1996; Jedlicka et al., 2004
),
tree age structure (
Garnas et al., 2011
), nutrient cycling (
Townsend
et al., 2004; Lovett et al., 2006
), carbon sequestration (
Peltzer et al.,
2010
), and the abundance of organisms such as aquatic inverte-
brates (
Smock and MacGregor, 1988
), large mammals (
Kendall
and Arno, 1990
) and birds (
Showalter and Whitmore, 2002;
Tomback and Achuff, 2010
).
While there is an extensive body of literature on the ecological
impacts of invasive species in forest ecosystems, most studies have
been limited to sampling from individual stands. The critical
importance of evaluating impacts of invaders across their entire
range has been recognized (
Parker et al., 1999
), but only a handful
of studies have taken a regional perspective to measuring impacts.
Here we use the concept of regional evaluation to quantify individ-
ual and combined impacts across the entire range of two invading
species. Given trends of continued accumulation of non-native
insects and diseases in forest ecosystems worldwide, there is a
serious need to assess the impacts of these species at the regional
level. The implementation of quarantine measures to exclude
future invasions can only be justified based on economic assess-
ments of area-wide impacts of past invasions (
Holmes et al.,
2009; Aukema et al., 2011
), thus highlighting the need for regional
estimation of pest impacts over their entire range.
Of particular importance is the need to understand how invad-
ing species alter regional trends in forest species composition and
thereby alter long-term forest dynamics. Given the extraordinarily
large number of damaging forest insect and pathogens species that
are accumulating worldwide, an immediate question is what is the
cumulative impact of these species on forest dynamics? Many of
these invading organisms are capable of causing extensive
http://dx.doi.org/10.1016/j.foreco.2014.12.018
0378-1127/Published by Elsevier B.V.
⇑
Corresponding author. Tel.: +1 610 557 4054; fax: +1 610 557 4250.
E-mail address:
rsmorin@fs.fed.us
(R.S. Morin).
Forest Ecology and Management 341 (2015) 67–74
Contents lists available at
ScienceDirect
Forest Ecology and Management
journal homepage: www.else
vier.com/loc
ate/foreco
mortality but little is known about how they may interact to alter
long-term trends in forest dynamics and consequently modify
long-term forest ecosystem processes.
To address this problem, we explore the individual and com-
bined effects of two major pest species invading the northeastern
US: the hemlock woolly adelgid (HWA),
Adelges tsugae
; and beech
scale (BS),
Cryptococcus fagisuga
, which is the causal agent of beech
bark disease (BBD). Their hosts, eastern hemlock (
Tsuga canadensis
(L.) Carr.), Carolina hemlock (
Tsuga caroliniana
Englem.), and Amer-
ican beech (
Fagus grandifolia
Ehrh.) were known to have dominated
large portions of presettlement northern forests (
Bürgi et al., 2000;
Thompson et al., 2013
).
Current forests in the northeastern US differ vastly from those
that existed prior to the time of European settlement (
Irland,
1999; Bürgi et al., 2000; Thompson et al., 2013
). Humans have
greatly altered forest composition via harvesting and conversion
to agricultural land use, followed by extensive agricultural aban-
donment. Other factors, such as alteration of presettlement fire
regimes and elevated deer populations have also greatly influenced
current forest composition (
Nowacki and Abrams, 2008; Horsley
et al., 2003
). Forests in the northeastern US are in flux and it is in
this context of a changing forest that the regional impacts of forest
insect and disease invasions should be considered. Specifically,
dominance by shade-tolerant hemlock and beech is increasing as
a result of successional processes (
Flinn and Vellend, 2005;
Thompson et al., 2013
) but it is not clear how this trend is altered
by insect and disease invasions.
Therefore, in the analysis presented here we focus on how HWA
and BS invasions combine to alter regional succession trajectories.
While this analysis is specific to the eastern United States, it pro-
vides insight into understanding the more general problem of
how alien forest pests affect forested ecosystems, a phenomenon
that is affecting forests worldwide.
1.1. Species backgrounds
Beech and hemlock dominate a large fraction of the late-succes-
sional forests of the eastern United States (
Fig. 1
A), and they both
fill distinctive roles in forest ecosystems. Mast produced by Amer-
ican beech is a critical source of food for various forest wildlife spe-
cies. Eastern hemlock is particularly common in riparian areas
where it plays a unique role in modifying microclimates, soil
chemistry, and stream temperatures. Both species are long-lived,
shade tolerant and compose a substantial proportion of the species
composition in late-successional forests in the maple/beech/birch
type (
online Supplement 1
).
BBD is an insect-fungus complex involving the non-native BS
which feeds on bark fluids from stems of American beech, provid-
ing an opportunity for the native canker fungi
Neonectria coccinea
var.
faginata
and
Neonectria ditissima
to invade the inner living bark
and cambium leading to dieback and mortality (
Mize and Lea,
1979; Houston, 1994
). While some trees survive infections for sev-
eral decades, one effect of the accumulation of cankers is reduced
growth (
Gavin and Peart, 1993; Gove and Houston, 1996
).
The BS was accidentally introduced with live plants imported to
Halifax, Nova Scotia from Europe, in the 1890s (
Houston, 1994
).
The scale insect has since slowly spread (
15 km/yr) into the
New England states, New York, Pennsylvania, and West Virginia
and several discontinuous ‘‘jumps’’ have transported it into North
Carolina, Tennessee, and Michigan (
Fig. 1
B) (
Morin et al., 2007;
Wieferich et al., 2013
). In 2004 the range of BBD comprised about
30% of the range of beech in the USA, but that area included about
50% of the total beech basal area (
Morin et al., 2005
).
Three phases of BBD are generally recognized: (1) the ‘‘advanc-
ing front’’, which corresponds to areas recently invaded by scale
populations; (2) the ‘‘killing front’’, which represents areas where
fungal invasion has occurred (typically 3–5 years after the scale
insects appear, but sometimes as long as 20 years) and tree mortal-
ity begins; and (3) the ‘‘aftermath forest’’, which are areas where
the disease is endemic (
Shigo, 1972; Houston, 1994
).
The hemlock woolly adelgid, native to East Asia, may have been
introduced to the eastern US as early as 1911; however, the first
Fig. 1.
Maps of species distributions in the eastern United States. (A) Distribution of
American beech,
Fagus grandifolia
, and hemlock,
Tsuga
spp. derived from interpo-
lated maps of species volume density from FIA plots (
Wilson et al., 2012
); (B)
historical spread of the beech scale (2006); and (C) historical spread of hemlock
woolly adelgid (2006).
68
R.S. Morin, A.M. Liebhold/Forest Ecology and Management 341 (2015) 67–74
report of its presence was in Richmond, VA in 1951 (
Havill and
Montgomery, 2008
). Since then, it has slowly expanded its range
at 8–30 km/yr (
Fig. 1
C) (
Evans and Gregoire, 2007; Morin et al.,
2009
). Further spread of the adelgid into northern New England
is unlikely under current climates because of its inability to toler-
ate cold winter temperatures (
Trotter and Shields, 2009
). In areas
where the species has established, populations often reach high
densities, causing widespread defoliation and sometimes mortality
of hemlock (
McClure et al., 2001; Orwig et al., 2002
).
Limited information exists about the long-term effects of BBD
and HWA on forest composition. It appears that in some stands,
the advent of BBD results in significant decreases in the proportion
of beech, but in other stands, beech is able to persist because of its
often prolific regeneration through sprouts and seedlings
(
Houston, 1994; Twery and Patterson, 1984; Runkle, 1990;
Houston, 2001
). Despite regional increases in beech mortality fol-
lowing invasion, considerable amounts of live beech remain in
invaded areas. Additionally, the volume of beech is still increasing
in most areas, although this increase is generally smaller than for
associated tree species (
Morin et al., 2007
).
Impacts of HWA vary within the range of the infestation.
Observed rates of hemlock loss within individual infested stands
have ranged from near 0 to more than 95% (
Orwig and Foster,
1998; Paradis et al., 2008
). To date, studies evaluating the impact
of HWA on forest structure have focused on individual stands
(
Eschtruth et al., 2006; Orwig and Foster, 1998
) or regions within
a state (
Orwig et al., 2002
). Despite the mortality induced by
HWA, hemlock volume is still generally increasing across the range
of the infestation (
Morin et al., 2011; Trotter et al., 2013
).
Since seedlings are an important indicator of future overstory
species composition, impacts on the species distribution of seed-
lings are also important for the trajectory of stands into the future.
Mortality of overstory beech results in a proliferation of basal
sprouts from surviving stumps (
Shigo, 1972; Houston, 1994
) creat-
ing dense ‘‘beech brush’’ conditions (
Horsley and Bjorkbom, 1983
)
that can interfere with regeneration of other hardwood species
such as sugar maple (
Hane, 2003
).
2. Methods
The Forest Inventory and Analysis (FIA) program of the US
Department of Agriculture (USDA) Forest Service conducts an
inventory of forest attributes nationwide (
Bechtold and
Patterson, 2005
). The current FIA sampling design is based on a tes-
sellation of the United States into hexagons approximately 2,458
hectares in size with at least one permanent plot established in
each hexagon. The population of interest is stratified and plots
are assigned to strata to increase the precision of estimates. Tree
and site attributes are measured for forested plots established in
each hexagon. Plots consist of four 7.2-m fixed-radius subplots
on which standing trees and various other environmental charac-
teristics are inventoried.
Prior to 1999, FIA collected data regionally using a periodic
measurement system with sample designs that varied slightly
through time and by region. Generally, inventories were conducted
in each state every 6–18 years, depending on the state and region.
Since 1999, FIA has adopted an annual inventory system, where
some plots are surveyed using a consistent plot design in each state
every year, across the ranges of BBD and HWA. This system pro-
vides a statistically robust sampling program for estimation of
mortality and net growth rates (
Bechtold and Patterson, 2005
).
Prior to the availability of the remeasured annualized FIA plot sys-
tem, it was not possible to directly estimate mortality and growth
across large areas. As a result, previous studies that used FIA data
to quantify regional impacts of HWA and BBD (
Morin et al., 2007,
2011; Trotter et al., 2013
) relied on estimates of standing dead tree
volume to quantify impacts rather than directly quantifying mor-
tality rates.
We used these historical FIA surveys to examine changes in host
tree species density over time. Since FIA plot data collected prior to
the 2000s were not collected annually, we were unable to examine
growth and mortality rates over time for the entire study region.
Therefore, we used periodic and annual inventory data to estimate
host species basal area over time in two states that are partially
infested by BBD and HWA, West Virginia and Pennsylvania. By
extracting basal areas from successive surveys and standardizing
basal area relative to the first estimate, we were able to character-
ize temporal trends in host species density. Therefore, the first esti-
mate is scaled to one and subsequent estimates represent
proportional change. To elucidate the potential impact of BBD
and HWA on these trends, we also estimated the time series in
two categories of BBD and HWA historical presence: infested
greater than 15 years (i.e., infested prior to 1999) and uninfested
or infested less than or equal to 15 years.
This study represents the first analysis of annualized FIA data
for regional estimation of mortality and growth rates in relation
to BBD and HWA invasion. The study utilized FIA plot data from
22 states where the ranges of American beech and eastern hemlock
overlap: Alabama, Connecticut, Delaware, Georgia, Kentucky,
Maine, Maryland, Massachusetts, Michigan, New Hampshire,
New Jersey, New York, North Carolina, Ohio, Pennsylvania, Rhode
Island, South Carolina, Tennessee, Vermont, Virginia, West Virginia,
and Wisconsin.
Remeasured plots (originally surveyed 2001–2005 and remea-
sured 2006–2010) were used to compute annual net growth and
mortality rates as proportions of live volume at the time of the ini-
tial survey (i.e., annual mortality volume/live volume at time 1).
Annualized net growth and mortality rates for trees 12.7-cm in
diameter and greater were calculated for eastern hemlock and
American beech, as well as for commonly associated species, sugar
maple (
Acer saccharum
Marsh.) and red maple (
Acer rubrum
L.)
based upon changes measured between successive forest invento-
ries. A detailed description of methods for computing mortality
and net growth are provided in
Bechtold and Patterson (2005)
.
Numbers of seedlings per hectare were calculated for eastern
hemlock, American beech, sugar maple, red maple, and birch spe-
cies (
Betula
spp.). Only data from large-diameter stands in the FIA
maple/beech/birch forest-type group (measured 2006–2010) were
included in order to avoid data from young, regenerating stands
overwhelming the estimates. It is important to note that FIA only
counts hardwood seedlings that are at least 30.5 cm tall and coni-
fer seedlings that are at least 15.2 cm inches tall. Seedlings must
also be less than 2.5 cm in root collar diameter. To be classified
as large-diameter stands, FIA protocols require that the predomi-
nant (based on stocking) diameter class of live trees is at least
27.9 cm for hardwoods and at least 22.9 cm for conifers.
Historical county-level records of the year of initial BS insect
and HWA establishment through 2006 were compiled by the US
Forest Service, Northeastern Area State and Private Forestry, Mor-
gantown, WV, and are available online (BS –
http://na.fs.fed.us/
fhp/bbd/infestations/infestations.shtm
;HWA–
http://na.fs.fed.us/
fhp/hwa/infestations/infestations.shtm
). These data were not
based upon systematic surveys and therefore slight inconsistencies
may exist among years and regions in how adelgid and BS popula-
tions were detected. Detection surveys were spatially crude and
variation can exists within counties, but these represent the best
available spatial distribution data available for the present of these
two invaders. Although these records are based on establishment
of the BS insect, we generally refer to BBD throughout the remain-
der of the paper.
Annual net growth rates, annual mortality rates, and numbers
of seedlings per hectare for each species were estimated from
R.S. Morin, A.M. Liebhold/Forest Ecology and Management 341 (2015) 67–74
69
inventory data as described above. These were estimated for spe-
cific ranges of pest invasion duration based on inventory plots that
fell within counties grouped according to specific invasion dura-
tion classes. The classes used for length of BBD establishment were
0, 1–15, 16–40 and greater than 40 years and for HWA classes were
0, 1–15, 16–25, and greater than 25 years. Note that the two high-
est HWA classes were combined for the seedling analysis due to
low sample sizes in the individual classes. The BBD classes were
selected to correspond to the recognized phases of beech bark dis-
ease invasion: (1) the advancing front, (2) the killing front, and (3)
the aftermath forest (
Houston, 1994
). Invasion by HWA does not
appear to occur in discrete phases like BBD so the use of such ter-
minology does not seem appropriate. However, we chose similar
classes for HWA to determine if mortality progression would be
comparable to BBD, but the largest class is a shorter duration
due to HWA’s more recent invasion (i.e., HWA has not been present
in many areas for more than 40 years).
Additionally, in order to ascertain the potential combined BBD
and HWA impacts, estimates are presented for the counties where
the ranges of infestations of both pests overlap for comparison
with the counties that are only infested by one pest (
Fig. 1
B and
C). We tested the statistical significance of the difference between
net growth rate, mortality rate, and seedlings per hectare estimates
within each species by pest duration categories with two-tailed
t
–
tests (
a
= 0.05). The false detection rate adjustment (FDR) was
employed to control experiment-wide error levels within each
family of comparisons (
Benjamini and Hochberg, 1995
). This
resulted in a reduced critical threshold for determining signifi-
cance between estimates in each multiple comparison family. All
associated statistics are listed in
online Supplement 2
.
Finally, linear regression analyses were employed to model the
relationship of county level estimates of net growth and mortality
as a function of duration of HWA and BBD invasions and the inter-
action between them to account for spatial overlap. The durations
were not divided into classes for the regression analyses. Plots of
residuals versus predicted values were examined to ensure that
the assumptions of linearity and homoscedasticity were met, and
normal probability plots of residuals were inspected to test for nor-
mally distributed errors. The intent of the regression analyses is to
examine the significance and direction of the relationships, not to
predict mortality or growth from duration of infestation.
3. Results
Eastern hemlock and American beech co-occur across large for-
ested regions (
Fig. 1
A). More than half of the volume of eastern
hemlock and beech occurs in the FIA maple/beech/birch forest-
type group, with the majority of the remaining hemlock occurring
in FIA pine groups and beech in the FIA oak/hickory group (
online
Supplement 1
). Within the maple/beech/birch group, both hemlock
and beech comprise substantial proportions of volume across all
diameter classes though hemlock is particularly dominant in large
diameters (
online Supplement 3
).
Live basal area of both hemlock and beech has generally been
increasing over the last three decades in areas uninfested and in
areas infested less than 15 years by BBD or HWA in West Virginia
and Pennsylvania. By contrast, basal area of beech has begun to
decrease in areas that have been infested by BBD for more than
15 years in West Virginia and Pennsylvania, and basal area of hem-
lock has begun to decrease in areas of West Virginia that have been
infested by HWA for more than 15 years. Although hemlock basal
area has not started to decrease in areas of Pennsylvania that have
been infested by HWA for more than 15 years, basal area increase
is much less than in areas where the insect has been present for a
shorter period or is absent (
Fig. 2
).
Evaluation of the remeasured annual inventories indicates that
annual beech mortality rates increase with increasing duration of
BBD infestation but this pattern is not seen for associated species
(
Fig. 3
A). Beech mortality rates appear to increase markedly after
15 years, but from the 16–39 to >40 year BBD infestation durations
mortality rates did not increase significantly (
online Supplement
2
). It is important to recognize that these are not increases in
cumulative mortality, but in annual rates. This means that beech
mortality rates increase following BBD establishment (
Fig. 3
A),
but then settle in to sustained, constant annual mortality rates
after 15 years of BBD presence (
online Supplement 2
). Additionally,
linear regression analyses indicated that the annual mortality rate
of American beech was positively associated with duration of BBD
infestation (
p
< 0.0001) and negatively associated with the interac-
tion between BBD and HWA establishment duration (
p
= 0.0259)
(
online Supplement 4
).
Mortality can be compensated for by growth of surviving trees
and ingrowth of young trees into the minimum (12.7 cm and
greater) diameter class. The combined effect of growth and
ingrowth with mortality is described by net growth. Although
the means indicate that the annual net growth rate of American
beech generally decreases with increasing BBD duration (
Fig. 3
B),
statistical comparisons of the means did not reveal significant dif-
ferences between the duration categories (
online Supplement 2
).
However, county-level linear regression analyses indicated that
the annual net growth rate of American beech was inversely
related with duration of BBD (
p
< 0.0001) and positively associated
with the interaction between BBD and HWA duration (
p
= 0.0529)
(
online Supplement 4
). Nevertheless, uninfested areas have net
growth rates that are nearly four times higher than regions
infested for more than 40 years but even in areas where BBD has
been present the longest (and mortality is greatest), net beech vol-
ume continues to increase (i.e., net growth > 0).
Evaluation of the remeasured annual inventories indicate that
annual mortality of eastern hemlock increased strongly with
increasing numbers of years of HWA presence (
Fig. 4
A). Annual
mortality rates of hemlock appear to increase most markedly after
15 years, but beyond 25 years mortality rates did not increase sig-
nificantly relative to the 16–24 year class (
online Supplement 2
).
Additionally, annual mortality of eastern hemlock was positively
correlated with duration of HWA infestation (
p
< 0.0001) and neg-
atively correlated with the interaction between BBD and HWA
duration (
p
= 0.0007) (
online Supplement 4
).
Similarly, hemlock net growth rates decrease as duration of
HWA invasion increases (
Fig. 4
B). Means indicate a lag in these
declines; net growth rates in regions with recent infestations are
similar to those in uninfested counties but once invasion duration
surpasses 15 years, net growth rates drop significantly and actually
become negative when HWA is present for more than 25 years.
Although statistical comparisons of the means do not indicate a
significant difference between the 16–25 and >25 year classes
(
online Supplement 2
), linear regression analysis indicated that
annual net growth of eastern hemlock was negatively correlated
with duration of HWA infestation (
p
< 0.0001) (
online Supplement
4
).
Interestingly, beech appears to benefit from hemlock mortality
associated with HWA and hemlock benefits from beech mortal-
ity related to BBD. This compensatory growth and offsetting
mortality is illustrated by examining the region where the ranges
of HWA and BBD overlap. American beech and hemlock mortality
rates are lower and net growths are higher when the other species’
pest duration is greater than 15 (
Fig. 5
A and B).
Seedling densities were at least partially associated with the
duration of BBD and HWA invasions. Beech seedlings densities
are much higher in BBD infested areas while the opposite trend
occurs for sugar maple (
Fig. 6
A). Densities of eastern hemlock
70
R.S. Morin, A.M. Liebhold/Forest Ecology and Management 341 (2015) 67–74
seedlings are similar across all HWA infestation classes (
Fig. 6
B).
Densities of seedlings of both maple species decrease with years
of HWA infestation, and the number of birch and beech seedlings
are similar across categories. Densities of beech seedlings are much
higher in counties that have been infested by BBD for more than
15 years regardless of whether HWA is also present (
Fig. 7
).
4. Discussion
The impacts of these two invaders on growth and mortality of
host species have previously been studied at specific sites (BBD –
Mize and Lea, 1979; Jones and Raynal, 1987; Gavin and Peart,
1993; Gove and Houston, 1996; Kasson and Livingston, 2011
;
HWA –
Orwig and Foster, 1998; Orwig et al., 2002; Eschtruth
et al., 2006; Rentch et al., 2009
). These studies documented very
high levels of mortality associated with infestations as well as
declines in the dominance of host species in affected stands. How-
ever, such studies generally have not captured the range of impacts
seen as these invasive insects persist for many years (
Fitzpatrick
et al., 2012
). While it is not surprising that we found mortality
and decreases in net growth associated with invasions, it is note-
worthy that mortality rates increase and net growth rates decrease
in both hemlock and beech over several decades following initial
invasions by these insects. There may be several explanations for
these protracted increases in damage. One reason may be that an
area (e.g., county) may be designated as ‘‘infested’’ but it may take
several years before all stands in the area become infested and the
severity of infestations may be quite variable. Another factor is that
populations of both HWA and BS may fluctuate following estab-
lishment (
McClure, 1991; Garnas et al., 2012
). Such recurring out-
break episodes may have cumulative adverse impacts on trees and
Fig. 2.
Standardized host species basal area by species, state, and pest infestation duration category: (A) beech basal area in West Virginia; (B) beech basa
l area in
Pennsylvania; (C) hemlock basal area in West Virginia; and (D) hemlock basal area in Pennsylvania.
Fig. 3.
Annual percent mortality (A) and net growth (B) of American beech,
hemlock, red maple, and sugar maple by years of BBD infestation (error bars
represent 68% confidence intervals. Remeasured plots (originally surveyed 2001–
2005 and remeasured 2006–2010) were used to compute annual net growth and
mortality as proportions of live volume at the time of the initial survey (i.e., annual
mortality volume/live volume at time 1).
R.S. Morin, A.M. Liebhold/Forest Ecology and Management 341 (2015) 67–74
71
result in tree mortality that increases over many years. Other envi-
ronmental factors such as climate, soils and land use also likely
affect growth, mortality, and seedling density and the duration of
infestation only explains a fraction of the variation in these proper-
ties (e.g.,
online Supplement 4
). Furthermore, geographical varia-
tion in these other environmental factors may be confounded
with the duration of pest invasion but nevertheless it is still possi-
ble to characterize consistent regional impacts of these invasions
(
Figs. 3 and 4
).
In addition to the differences in net growth and mortality
observed in beech and hemlock in relation to BBD and HWA infes-
tation history, distinct trends were also seen in densities of seed-
lings among species in relation to infestation history. Since
seedlings are an important indicator of future overstory species
composition, impacts on the species distribution of seedlings are
important for the trajectory of stands into the future. Beech seed-
lings densities are higher in areas where beech scale is present
(
Fig. 6
A), but hemlock seedlings densities did not vary across
HWA infestation history classes (
Fig. 6
B). The trend in beech seed-
ling densities reflects a phenomenon commonly observed in stands
heavily affected by BBD; mortality of overstory beech results in a
proliferation of basal sprouts from surviving stumps and roots
(
Shigo, 1972; Houston, 1994
) creating dense ‘‘beech brush’’ condi-
tions (
Horsley and Bjorkbom, 1983
) that can interfere with regen-
eration of other hardwood species such as sugar maple (
Hane,
2003
). Indeed, we found that sugar maple seedling densities
decrease as the density of beech seedlings increases (
Fig. 6
A).
These results indicate that prolific beech sprouting may adversely
affect sugar maple regeneration even at the regional scale, which
may further exacerbate the impacts on long-term successional
dynamics given that sugar maple is another shade tolerant species.
The lack of an observed impact of HWA presence on hemlock seed-
Fig. 4.
Annual percent mortality (A) and net growth (B) of American beech,
hemlock, red maple, and sugar maple years of HWA infestation (error bars
represent 68% confidence intervals). Remeasured plots (originally surveyed 2001–
2005 and remeasured 2006–2010) were used to compute annual net growth and
mortality as proportions of live volume at the time of the initial survey (i.e., annual
mortality volume/live volume at time 1).
Fig. 5.
Annual percent mortality (A) and net growth (B) of American beech,
hemlock, red maple, and sugar maple in overlapping and non-overlapping HWA and
BBD infested areas (error bars represent 68% confidence intervals). Remeasured
plots (originally surveyed 2001–2005 and remeasured 2006–2010) were used to
compute annual net growth and mortality as proportions of live volume at the time
of the initial survey (i.e., annual mortality volume/live volume at time 1).
Fig. 6.
Number of seedlings per hectare of American beech, hemlock, red maple,
sugar maple, and birch by years of (A) BBD infestation and (B) HWA infestation
(error bars represent 68% confidence intervals).
72
R.S. Morin, A.M. Liebhold/Forest Ecology and Management 341 (2015) 67–74
ling densities suggests that hemlock may persist, at least at low
levels, in the region over the coming decades.
Another interesting result is the apparent increase in beech net
growth in response to HWA induced overstory hemlock mortality.
Similarly, hemlock net growth responds favorably to BBD induced
beech mortality (
Figs. 3
B and
4
B). Previous studies have identified
compensatory growth of small-diameter beech in response to the
loss of overstory beech (
Gravel et al., 2011
) but not growth in other
species (
Garnas et al., 2011
). Since beech and hemlock frequently
co-occur in the same stands, it is not surprising that we observed
that each compensate for losses of the other species, though such
compensation was surprisingly not seen in sugar or red maple.
As HWA and BS continue to expand their ranges, the area where
both species are present is likely to expand and this may decrease
the potential for compensatory growth.
Orwig (2002)
reported a
concurrent increase in maple growth associated with a decrease
in hemlock growth in south-central Connecticut. Although this
phenomenon may occur in stands that have been severely
impacted by HWA, the results of this study do not indicate that
compensatory growth has occurred in associated species at the
regional scale even though hemlock growth has decreased
(
Fig. 4
B). Several studies have reported that black birch (
Betula len-
ta
L.) is the species that benefits the most from declines in hemlock
based on prolific regeneration and re-establishment (
Orwig and
Foster, 1998; Orwig, 2002; Stadler et al., 2005; Sullivan and
Ellison, 2006
), but our results show that the number of birch seed-
lings is similar across HWA infestation duration categories.
Increases in black birch regeneration may be occurring in some
areas, but this is not evident at the regional scale (
Fig. 6
B).
To fully understand the impact of BBD and HWA invasion at the
regional scale, it is necessary to consider the observed variation in
growth rates, mortality rates and seedling densities with respect to
long-term successional trends in the region.
Online Supplement 5
illustrates how the most common shade-tolerant species (includ-
ing beech and hemlock) found in northeastern deciduous forests
have been steadily increasing in dominance over the last half cen-
tury. This trend reflects well-documented successional processes
operating in the region; extensive forest clearing prior to 1900, fol-
lowed by regrowth and agricultural abandonment, and has placed
forests in the region in a long-term trend of increasing dominance
by shade-tolerant species such as hemlock and beech (
Flinn and
Vellend, 2005; Thompson et al., 2013
).
Results presented here provide some indication of how inva-
sions by two phytophagous insect species may likely modify spe-
cies composition trends into the future. Because impacts on net
growth of their hosts are delayed over several decades, the reversal
of trends of increasing beech and hemlock dominance is currently
evident only in stands where BBD and HWA have been the longest
(
Figs. 3 and 4
). The full impact of these invading pests is likely to
play out very slowly in the future. Hemlock volume is continuing
to increase in many areas that are already infested with HWA
despite the adverse effects of this insect (
Fig. 2
D,
Trotter et al.,
2013
). Forest succession across the region is favoring increases in
hemlock ingrowth from the understory and may be capable of off-
setting declines following HWA-caused tree mortality. However, it
appears that after 25 years of HWA presence, such ingrowth offsets
are not sufficient and total hemlock density begins to decrease at
the regional scale (
Figs. 4 and 5
). These decreases may be rein-
forced by a trend of decreased hemlock regeneration in areas with
long histories of HWA though we did not detect such an effect
(
Fig. 6
B). As HWA persists in areas over many years, we can thus
anticipate a slow-motion decline in hemlock density across the
entire infested region in the future. It should, however, be pointed
out that HWA is not expected to expand its current range substan-
tially to the north due to climatic unsuitability (
Trotter and Shields,
2009
), and hemlock in these northern portions of its range are
likely to escape this future downward trend. Nevertheless, warm-
ing due to climate change could make the northern range of hem-
lock more suitable for HWA in the future.
Long-term trends in beech, however, can be anticipated to be
slightly different. The pattern of increased beech regeneration in
response to BBD documented here (
Fig. 6
A) and several other stud-
ies (
Shigo, 1972; Houston, 1994
) means that beech is likely to per-
sist following scale presence over many years. Nevertheless,
increases in beech density (
Fig. 2
) are likely to be dampened over
the long-term. The persistence of BBD over future decades is also
likely to result in a shift in the distribution of beech volume to
lower diameter classes (
Garnas et al., 2011
).
The loss of hemlock as a result of HWA invasion has already had
impacts on ecosystem properties such as stream temperatures and
soil chemistry (
Stadler et al., 2005; Orwig et al., 2008
). Similarly,
beech is an important source of mast and losses of large beech have
been implicated in declines of certain wildlife species (
Storer et al.,
2005
). The future regional trends in hemlock and beech volume are
likely to translate into changes in other regional properties, such as
water chemistry (
Eshleman et al., 1998
), hydrologic budgets (
Ford
and Vose, 2007
), and wildlife densities (
Storer et al., 2005
).
Acknowledgements
We would like to thank Susan Stout, Songlin Fei, and three
anonymous reviewers for their constructive reviews of earlier ver-
sions of this manuscript.
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.2014.12.
018
.
Refoliation in the Appalachian Plateau. Forest Ecol. Manage. 89 (1), 79–88
.