Tree Physiology Advance Access published January 6, 2012

Tree Physiology 00, 1–12

doi:10.1093/treephys/tpr133

Research paper

Increased resin flow in mature pine trees growing under elevated

CO2 and moderate soil fertility

Downloaded from [Link] at DigiTop USDA's Digital Desktop Library on January 10, 2012
K.A. Novick1,2,5, G.G. Katul1, H.R. McCarthy3 and R. Oren1,4

1NicholasSchool of the Environment and Pratt School of Engineering, Duke University, Box 90328, Durham, NC 27708, USA; 2Present address: USDA Forest Service, Southern
Research Station, Coweeta Hydrologic Laboratory, Otto, NC 28763, USA; 3Department of Botany and Microbiology, University of Oklahoma, Norman, OK 73019, USA;
4Department of Forest Ecology & Management, Swedish University of Agricultural Sciences (SLU), SE-901 83, Umeå, Sweden; 5Corresponding author: Coweeta Hydrologic

Laboratory, 3160 Coweeta Lab Road, Otto, NC 28763, USA (kimberlynovick@[Link])

Received April 12, 2011; accepted November 19, 2011; handling Editor Yann Nouvellon

Warmer climates induced by elevated atmospheric CO2 (eCO2) are expected to increase damaging bark beetle activity in pine
forests, yet the effect of eCO2 on resin production—the tree’s primary defense against beetle attack—remains largely
unknown. Following growth-differentiation balance theory, if extra carbohydrates produced under eCO2 are not consumed by
respiration or growth, resin production could increase. Here, the effect of eCO2 on resin production of mature pines is
assessed. As predicted, eCO2 enhanced resin flow by an average of 140% (P = 0.03) in canopy dominants growing in low-
nitrogen soils, but did not affect resin flow in faster-growing fertilized canopy dominants or in carbohydrate-limited sup-
pressed individuals. Thus, pine trees may become increasingly protected from bark beetle attacks in an eCO2 climate, except
where they are fertilized or are allowed to become overcrowded.

Keywords: bark beetles, carbon allocation, Free Air CO2 Enrichment, Pinus taeda, resin flow, resistance.

Introduction
evaporated, seals wounds in the bark (Ruel et al. 1998, Wilkens
Bark beetles (Curculionidae: Scolytinae) and their fungal associ- et al. 1998). Resin flow may be either constitutive or induced,
ates are among the most ecologically altering and economically where the former represents the flow of a standing pool of resin
damaging natural disturbance agents in coniferous forests formed in the tree prior to attack, and the latter represents the
(Price et al. 1998, Kurz et al. 2008, Raffa et al. 2008). Several de novo synthesis of resin following mass wounding events
studies predict increases in beetle activity in warmer climates associated with a high density of beetle attacks (Lombardero
or cite higher air temperatures as a co-factor in recent beetle et al. 2000). Results from previous studies suggest that both
outbreaks (Logan et al. 2003, Gan 2004, Berg et al. 2006, constitutive and induced resin flow play an important role in
Kurz et al. 2008). However, future forest-beetle dynamics also defending pine trees from attacking beetles (Reeve et al. 1995,
depend on the response of the host’s defense mechanisms to Lorio et al. 1995, Strom et al. 2002, Wallin et al. 2008). This
climate and atmospheric changes, and these have not yet been study is focused on the effects of atmospheric CO2 concentra-
determined (Negron et al. 2008). In particular, the effect of tion, as well as soil nutrient availability, canopy position and
elevated atmospheric CO2 on resin production in mature pine time of year, on constitutive resin flow in mature loblolly pine
trees—a primary defense against bark beetle attack—has not trees, noting that the mass wounding treatments necessary to
been quantified. produce induced resin flow were not permissible in the Free Air
Resin production is largely, though not exclusively, a mechan- CO2 Enrichment (FACE) setting, where this study took place.
ical defense strategy. Resin physically ejects or entombs attack- Our hypotheses and the interpretation of our results are framed
ing beetles and, when volatile components of resin have by the following theoretical considerations.

© The Author 2012. Published by Oxford University Press. All rights reserved. For Permissions, please email: [Link]@[Link]
 2 Novick et al.

Early work on the resistance of pines to bark beetle attacks

suggested that resin production was a reflection of tree vigor,
such that a scarcity of photosynthates from environmental- or
competition-caused stress decreases both tree growth and
resistance (Waring and Pitman 1985). Other studies demon-
strated that resin flow could be increased even as growth or
vigor decreased through manipulations of soil nutrients, mois-
ture and stand density, provided that reductions in photosyn-
thesis were less than reductions in growth (Lorio 1986, Lorio
et al. 1995, McDowell et al. 2007, Knebel et al. 2008). The
notion that growth and secondary processes (such as resin
production) are competing sinks for assimilated carbon is a

Downloaded from [Link] at DigiTop USDA's Digital Desktop Library on January 10, 2012
central component of the ‘growth-differentiation balance’
(GDB) hypothesis for plant defense, first applied to pines by
Lorio (1986) and later formalized by Herms and Mattson
(1992).
Foremost, the GDB hypothesis predicts that, when herbivory
levels are low, assimilated photosynthates are preferentially
allocated to maintenance and growth processes over second- Figure 1. The growth-differentiation balance hypothesis. (a)
ary processes such as resin synthesis, and thus constitutive Conceptual change in available photosynthates (A′, solid lines) and
resin production will be directly correlated to the balance growth processes (NPP, dashed lines) with a factor F, where F may be
any potentially limiting resource. Shaded areas show the pool of
between two quantities: (i) the rate of production of available excess carbohydrates available for secondary processes like resin syn-
carbohydrates (A′), here defined as gross primary production thesis (A′ − NPP), which is also shown separately in (b). In Phase I, A′
(GPP) less maintenance respiration (i.e., A′ = GPP − RM), and (ii) increases faster than NPP (i.e., ∂A′/∂F > ∂NPP/∂F) such that A′ − NPP
also increases with F. In Phase II, A′ approaches its maximum so that
the rate at which available carbohydrates are used in primary ∂A′/∂F < ∂NPP/∂F and A′ − NPP decreases with F. Thus, the sign of
growth processes, or net primary production (NPP). When A′ is ∂(A′ − NPP)/∂F dictates the sign of the change in resin flow (mR) with
limited by some factor (F), an increase in F will tend to increase F. The trends for ambient and elevated CO2 show that the A′ − NPP
balance can be larger under elevated CO2 for mid-range F. The figure
A′ more than NPP (Phase I of Figure 1). Here, F represents any
is adapted from Figure 1 in Herms and Mattson (1992).
variable that is potentially limiting to A′ and that changes in time
and/or space, including soil nutrient and moisture availability,
light availability and temperature. Thus, in Phase I, Based on this framework, it is possible to predict how spa-
∂(A′ − NPP)/∂F > 0, and the balance A′ − NPP widens with tial and temporal gradients in factors known to impact A′ may
increasing F (Figure 1a and b). As a result, in Phase I, positive affect resin production. Specifically, elevated CO2 should
relationships among A′, NPP, A′ − NPP and resin production are increase resin production in low- to mid-range fertility by
expected. increasing A′ while NPP is limited by nutrient availability
When F is no longer limiting to A′ (Phase II), A′ reaches satu- (Figure 1). However, as soil fertility increases, growth restric-
ration with increasing F (Figure 1a, ∂A′/∂F ≈ 0) while NPP con- tions are eased (Oren et al. 2001, Stamp 2004, Palmroth
tinues to increase (Palmroth et al. 2006) so that et al. 2006, McCarthy et al. 2007) and the eCO2-induced
∂(A′ − NPP)/∂F < 0. Thus, in Phase II, the GDB hypothesis sug- enhancement of resin production is expected to decline.
gests an inverse relationship between NPP and resin produc- Resin production should be lower in suppressed as compared
tion due to competition for limited carbohydrates (Herms and with dominant individuals as A′ is limited for the former by
Mattson 1992). Consequently, the sign of ∂(A′ − NPP)/∂F low light availability in the sub-canopy. Over the course of a
determines whether resin production increases or decreases growing season, a period exists in which light intensity, day
with changes in F. While the hypothesis is generally presented length, temperature and leaf area index are sufficiently high
and interpreted over a spatial or temporal gradient of a single and stable, shifting forest canopies from Phase I (when A′ is
resource, the framework may be generalized such that F = (F1, limited by these factors) to Phase II (when A′ is relatively high
F2, . . .) represents a vector of factors that vary in time and and stationary). During that stable A′ period, temporal varia-
space. In this case, resin production is correlated with the sign tion in factors such as soil moisture introduces variation in
of the aggregated sum of d(A′ − NPP), given as NPP. Thus, as the shift occurs, the sign of ∂(A′ − NPP)/∂F will
change, and the relationship between NPP and resin produc-
∂ ( A '− NPP ) ∂ ( A '− NPP )
d ( A '− NPP ) = dF1 + dF2 + …. tion should accordingly be positive in Phase I and negative in
∂F1 ∂F2 (1) Phase II (see Figure 1).

Tree Physiology Volume 00, 2012

 Elevated CO2 increases resin flow 3

In this study, components of A′, NPP and resin production sufficiently high to cause canopy class differentiation between
are measured on a monthly basis in plots of mature, field- suppressed, shaded individuals and dominant individuals pro-
grown pine trees experiencing both ambient and elevated vided with ample light. Hence, the Duke FACE facility provides
atmospheric CO2 concentrations and a wide range of soil fertil- a unique setting for evaluating GDB predictions of the effect of
ity. Results are interpreted within the context of these predic- eCO2, nutrient availability and canopy position on resin
tions from GDB theory (guided by Figure 1), and the trade-offs production.
between growth and defense processes are discussed with
consideration of forest management strategies in future climate Resin measurements
conditions. Resin mass flow (mR) was measured monthly from March to
October of 2009 in two canopy classes: dominant and sup-
Methods pressed trees. In each subplot, mR was measured in three
dominant trees and two suppressed trees with two subsam-

Downloaded from [Link] at DigiTop USDA's Digital Desktop Library on January 10, 2012
Setting ples per tree taken from opposite sides of the trunk. Following
Components of A′, NPP and resin production were measured in standard practice (Lombardero et al. 2000, Knebel et al.
the Duke FACE study (Hendrey et al. 1999), set in a 27-year- 2008), resin flow was measured by removing a 0.8 cm disk of
old Pinus taeda L. (loblolly pine) plantation in the Duke Forest, bark to the phloem–xylem interface at breast height, and
NC, USA. The pine plantation was established in 1983 follow- installing trays below the wound to direct the flow of resin into
ing a clear cut and a burn. Pinus taeda L. (loblolly pine) seed- pre-weighed collection vials for a 24-h period. Vials were
lings were initially planted at a 2.0 m by 2.4 m spacing, and installed on all trees between 08:00 and 12:00 a.m. to mini-
pine density was ~1100 trees hectare−1 in 1999 (Hendrey mize the effect of time-of-day variation on resin flow.
et al. 1999). Tree height approached 20 m at the time of this Measurements were repeatedly conducted on the same set of
study. The stand has not been managed after planting and a trees, and care was taken to ensure that new wounds were
diverse sub-canopy has developed consisting of Liquidambar laterally displaced at least 5 cm from previous wound sites.
styraciflua L. and nearly 40 other hardwood species. Using a lateral displacement distance of at least 5 cm is stan-
The FACE facility comprises eight circular, 30 m diameter dard practice in constitutive resin flow studies and reflects the
plots—four under ambient CO2 concentrations (aCO2) and four common assumption that the scarcity of radial resin duct con-
under elevated concentrations (eCO2; ambient + 200 ppm). One nections limits the effect of previous resin flow measurements
ambient plot and one elevated plot comprise the FACE prototype on subsequent sampling events (Ruel et al. 1998, Knebel et al.
complex, which has operated since 1994. The six remaining 2008). The resin flow measurements are normalized by the
FACE plots have operated since 1996 as blocked replicates. area of the wound and expressed in units of g cm−2 day−1. It is
Annual nitrogen (N) fertilization of half of each plot began sev- important to note that while the resin flow measurements are
eral years after fumigation was initiated (Oren et al. 2001, referenced to the area of the wound (~0.50 cm2), the longitu-
Palmroth et al. 2005). This split-plot amendment created 16 dinal resin canal lengths are on the order of ~10 cm (Lapasha
subplots—four each of unfertilized and fertilized aCO2 plots (AC and Wheeler 1990). Thus, for a wound diameter of 0.8 cm,
and AF, respectively), and four each of unfertilized and fertilized resin flows from an area of the phloem that is at least ~8 cm2,
eCO2 plots (EC and EF). The prototype and replicated FACE and probably larger if any radial duct connections exist.
plots are treated as one replicated experiment following initial
analysis that failed to show a significant block effect on resin Assimilation and NPP
flow (P = 0.62 from a split-plot analysis of variance (ANOVA) Gross ecosystem productivity (GEP) under ambient CO2
with block as a random effect) and prior results that strongly (GEPaCO2 ) was estimated using eddy-covariance measurements
reject the hypothesis that the prototype and replicated FACE of the net ecosystem exchange of CO2 (NEE) and was used as
complexes are not from the same study (McCarthy et al. 2007). a proxy for A′. Gross ecosystem productivity is defined as GPP
This study altered components of the A′ − NPP balance less the amount of CO2 that is internally recycled within the
experimentally with eCO2 and with fertilization, but also utilized ecosystem, either by re-assimilation within the leaf of CO2 pro-
natural resource gradients known to affect carbon assimilation duced by dark respiration, or by re-assimilation of respired CO2
and growth. Experimental plots cover a wide range of native N occurring within the canopy below the eddy-covariance instru-
availability (from 2.6 to 5.0 g mineral N m−2). Native N avail- mentation. Thus, GEP and GPP are analogous, but not identical
ability has been previously shown to be directly related to ini- (Goulden et al. 1997, Stoy et al. 2006). Measurements were
tial stand biomass (i.e., biomass before the onset of the FACE subjected to quality control procedures to remove excessive
experiment), to stand-level NPP in both aCO2 and eCO2 sensor noise and data collected under very stable conditions.
(McCarthy et al. 2010), and to canopy leaf area (McCarthy First, spikes in the data record were removed using the algo-
et al. 2007). Furthermore, tree density in the FACE plots is rithm recommended for FLUXNET (Papale et al. 2006). Net

Tree Physiology Online at [Link]

 4 Novick et al.

ecosystem exchange of CO2 fluxes was then filtered to remove sum of the production of stems, branches and coarse roots.
data collected during stable atmospheric conditions using a u* The stem and branch biomass values of individual trees were
filtering method (Reichstein et al. 2005). Finally, fluxes were calculated allometrically using recently developed, site-specific
rejected if they exceeded a conservative window of −0.5 to relationships that are driven by both tree diameter and height
0.85 mg C m−2 s−1 in accordance with data quality control pro- (McCarthy et al. 2010). Monthly diameter increments were
cedures previously employed at this site (Stoy et al. 2006). estimated by interpolating bi-weekly dendrometer measure-
Note that a positive flux denotes assimilation by the ecosys- ments and determining the change in diameter from the begin-
tem. Net ecosystem exchange of CO2 was partitioned into ning to the end of the month, noting that resin measurements
GEPaCO2 and ecosystem respiration using the non-rectangular were performed at the end of each month. Tree height incre-
hyperbola model described in Stoy et al. (2006). ment measured on an annual basis was interpolated to monthly
The eddy-covariance estimates of GEPaCO2 are stand-scale esti- values using the same seasonal pattern of diameter growth.
mates, and no attempt was made to use these measurements to Plot-level coarse root biomass (>2 mm) was calculated as a

Downloaded from [Link] at DigiTop USDA's Digital Desktop Library on January 10, 2012
derive tree-level assimilation or to quantify the magnitude of function of aboveground biomass (Johnsen et al. 2004).
stand-scale assimilation in eCO2 plots. This time series was used Because NPPW was estimated monthly, we accounted for the
only to identify the temporal pattern of assimilation. The dynam- transition between earlywood and latewood, where latewood
ics of GEP are assumed to be similar under all treatments but has twice the specific gravity of earlywood. The date of transi-
vary in absolute magnitude with CO2, and N and light availability. tion was estimated for each tree as the date at which the
This assumption is supported by the results of earlier studies cumulative diameter increment, normalized by the total annual
conducted at the Duke FACE facility, showing that monthly mea- diameter increment, exceeds the treatment-specific earlywood
surements of net leaf-level assimilation rate (Anet) are strongly fraction previously determined from analysis of growth rings.
correlated in aCO2 and eCO2 trees (r = 0.93, Figure 2a; Ellsworth
1999) and trees from unfertilized and fertilized sectors (r = 0.86, Nitrogen availability
Figure 2b; Maier et al. 2008). Coupled with the additional similar- Nitrogen losses from the FACE study site have been shown to
ity among treatments of the temporal dynamics of total leaf area be negligible (Phillips et al. 2001, Finzi et al. 2006); thus, min-
(McCarthy et al. 2007), it is not surprising that Schäfer et al. eral N availability was estimated as the sum of N mineraliza-
(2003) found that the seasonal pattern of canopy-scale GPP esti- tion, N deposition, N fixation and N addition via fertilization.
mated with the multi-layer, canopy conductance-constrained car- Annual N mineralization was measured over 3 separate years
bon assimilation model (4C-A) was similar under eCO2 and aCO2; (1998, 2003 and 2005) using a buried-bag approach (Finzi
furthermore, the pattern was consistent with eddy-covariance et al. 2006, McCarthy et al. 2010). Nitrogen fixation rates were
estimates of GEP. Thus, we conclude that the seasonal pattern of measured on five separate occasions between 2001 and
GEPaCO2 was a reasonable proxy of A′ of all treatments and both 2003 by applying the acetylene reduction assay to soil cores
size classes, and can be used to partition the study period and forest floor samples from each subplot (Hofmockel and
between periods of stable A′ and periods of variable A′. Schlesinger 2007). Nitrogen deposition was estimated by
Net primary productivity of woody biomass (NPPW), used as Sparks et al. (2008) using a combination of eddy-covariance
a proxy for NPP, was estimated on a monthly time step as the measurements of the flux of NO, local- and regional-scale

Figure 2. Temporal correlation of CO2 assimilation rates (Anet) derived from leaf-level gas-exchange measurements. (a) Bi-weekly measurements
performed in ambient CO2 (aCO2) and elevated CO2 (eCO2). Data are from Ellsworth (1999). (b) Monthly estimates of Anet from control and fertil-
ized sectors of both CO2 treatments. Data from Maier et al. (2008). Both datasets are from measurements made on current-year sunlit foliage. The
dotted lines are 1:1 lines, shown for reference.

Tree Physiology Volume 00, 2012

 Elevated CO2 increases resin flow 5

concentration measurements of other relevant N-containing Results

species, and modeling techniques. The quantity of N added by
The study period and characteristics of the time series
fertilization is uniform across subplots (Oren et al. 2001), and
mineral N availability in fertilized subplots was assumed to The year 2009 was a relatively cool, wet year for the Duke
increase by ~5.5 g N m−2 over native N availability, consistent FACE facility. Mean annual temperature was 14.4 °C, which is
with previous studies (McCarthy et al. 2007). less than the long-term mean annual temperature (15.5 °C).
Total annual precipitation at the Duke FACE facility in 2009
Statistical analyses was 1231 mm, which is greater than the long-term mean of
Unless otherwise noted in the text, the significance of the 1146 mm. Mean monthly temperature during the study period
difference in m R between any two populations was evaluated was highest in August (Figure 3a).
using a two-tailed T-test assuming unknown means and Canopy assimilation (i.e., GEPaCO2 ) was relatively low in
equal variances. The significance of eCO2, fertilization and March, April, September and October in response to reduced

Downloaded from [Link] at DigiTop USDA's Digital Desktop Library on January 10, 2012
time of year effects, and their interactions, on m R was also leaf area and unfavorable climatic conditions. Gross ecosystem
evaluated separately for dominant and suppressed popula- productivity under ambient CO2 was relatively high and station-
tions using a repeated-measures, split-plot ANOVA with CO2 ary during the peak of the growing season (i.e., May–August,
concentration as the main effect and fertilization treatment Figure 3b). In all treatments, NPPW in both dominant and sup-
as the split effect. pressed trees peaked in May or June and declined to a local

Figure 3. Temporal variation of key variables. Shown are mean monthly values of (a) temperature (T) and day length (DL), and (b) gross ecosys-
tem production estimated over an ambient portion of the stand with the eddy-covariance technique (GEPaCO2 ). (c–f) Resin flow and NPP of woody
biomass (mR and NPPW, respectively) for dominant trees from the four treatment classes: aCO2 control (AC), aCO2 fertilized (AF), eCO2 control (EC)
and eCO2 fertilized (EF). (g–j) The same variables for suppressed trees. Symbols in (c–j) show the monthly averages from n = 4 subplots in each
treatment class, with bars representing one standard error of the mean.

Tree Physiology Online at [Link]

 6 Novick et al.

minimum in August (Figure 3c–j), concurrent with the develop- classes. In this presentation, mR, GEPaCO2 and NPPW were
ment of a moderate soil moisture deficit which is typical in late normalized in reference to the study period maximum in each
summer. Conversely, the time series for mR shows a local mini- subplot to accommodate local differences in edaphic condi-
mum in May and a local maximum later in the growing season tions and CO2 concentration (mR,N, GEPaCO2 ,N and NPPW,N,
(Figure 3c–j). Monthly variation in mR was significant (P < 0.01, respectively). Conversely, when GEPaCO2 was relatively high and
Table 1). The time series of mR for all aCO2 and eCO2 individu- stable (May–August), inverse relationships between mR,N and
als pooled is shown in Figure 4. NPPW,N were observed for both size classes (P < 0.0001,
Figure 5e and f).
Temporal correlation among GEPaCO , NPPW and mR
2 The similarity of the temporal response of mR,N, GEPaCO2 ,N and
When GEPaCO2 was relatively low (March–April, September– NPPW,N (Figure 5a–f) among treatments was evaluated by first
October), resin flow was directly related to both GEPaCO2 and formulating linear relationships between these two variables
NPPW (P < 0.0001, Figure 5a–d) in all treatments and size within each treatment using subplot-scale averages from all four

Downloaded from [Link] at DigiTop USDA's Digital Desktop Library on January 10, 2012
subplots in each treatment. Then, for any two treatments, the
residual error was estimated for a simplified model assuming
the same slope and intercept for both treatments, and a full
model assuming a separate slope and intercept for each treat-
ment. The significance level of the reduction in error achieved
by using the full versus simplified model was assessed via an
F-test. In all cases, we found no differences in the relationships
between mR,N and GEPaCO2 ,N or NPPW,N among any of the treat-
ments within each canopy class (P = 0.46–0.74).

Canopy size class effects

Throughout the study period, resin flow was significantly lower
(P = 0.01) in suppressed compared with dominant trees
(〈mR〉 ± SD of 1.6 ± 0.7 and 2.5 ± 1.3 g cm−2 day−1, respec-
Figure 4. Effect of eCO2 on resin flow time series. Monthly resin flow tively) when averaged across CO2 and fertilization treatments.
(mR) in dominant, unfertilized trees growing in aCO2 (AC, open sym-
bols) and eCO2 (EC, closed symbols), showing that the eCO2 enhance-
Net primary productivity of woody biomass was also signfi-
ment of mR in these individuals persists throughout the growing cantly lower (P < 0.0001) in suppressed compared with domi-
season. Error bars show one standard error of the mean. nant trees (〈NPPW〉 ± SD of 14 ± 4.5 and 60.2 ± 14.0 g
C tree−2 day−1, respectively). Gross ecosystem productivity was
Table 1. Probability values for the impacts of atmospheric CO2 con- not measured separately for the different size classes. Temporal
centration (C), fertilization (N) and sampling month (M) on resin flow variation in mR was significant in suppressed as well as domi-
(mR) from a repeated-measures, split-plot ANOVA with CO2 concentra-
tion as the main effect and fertilization treatment as the split effect.
nant individuals (Table 1).
The ANOVA was run independently for dominant and suppressed
trees, and significant effects are shown in bold type. Elevated atmospheric CO2 and N fertilization effects
Effect Df mR The eCO2 treatment significantly increased mR among domi-
nant individuals (P = 0.03, repeated-measures ANOVA, Table
Dominant (n = 48)
1). This enhancement was entirely limited to trees growing in
C 1 0.03
N 1 0.78 unfertilized subplots (Figure 6a), where eCO2 increased mR by
C×N 1 0.13 140% (〈mR〉 = 1.42 ± 0.52 and 3.44 ± 1.31 g cm−2 day−1 for AC
M 7 <0.01 and EC subplots, respectively). The eCO2 enhancement per-
C×M 7 0.48 sisted throughout the growing season (Figure 4). Atmospheric
N×M 7 0.47 CO2 concentration had no effect on mR among fertilized indi-
C×N×M 7 0.01
viduals (〈mR〉 = 2.50 ± 0.90 and 2.50 ± 0.93 g cm−2 day−1 for
Suppressed (n = 32)
AF and EF subplots, respectively, P = 0.92) and no effect
C 1 0.97
N 1 0.60 among suppressed individuals regardless of N availability
C×N 1 0.63 (Figure 6).
M 7 <0.01 Spatial variation in mean mR with N availability was evaluated
C×M 7 0.45 for unfertilized and fertilized dominant and suppressed popula-
N×M 7 0.45 tions (Figure 6a and b). In each case, a linear model allowing
C×N×M 7 0.02
unique slopes and intercepts for aCO2 and eCO2 populations

Tree Physiology Volume 00, 2012

 Elevated CO2 increases resin flow 7

Downloaded from [Link] at DigiTop USDA's Digital Desktop Library on January 10, 2012
Figure 5. Relationships between normalized resin flow (mR,N) and normalized eddy-covariance-based GEP (GEPaCO2 ,N ) and normalized NPP of
woody biomass (NPPW,N). During Phase I (March–April and September–October; four left panels), mR,N is directly related to GEPaCO2 N and to NPPW,N
for dominant (upper panels) and suppressed (lower panels) individuals. In Phase II (May–August), m R,N was inversely related to NPPW,N (two right
panels). The variables were normalized by the study period maximum in each subplot to accommodate local differences in edaphic conditions and
CO2 concentration. The variable GEPaCO2 represents GEP of untreated conditions measured using the eddy-covariance technique over an ambient
plot only. Symbols show the mean value in each treatment, and error bars show one standard error of the mean. Some data points are obscured
by others, particularly in (a) and (b).

was compared with a simpler model assuming a single slope

and intercept across CO2 treatments. For unfertilized dominants,
the full model offered a significant reduction in error
(P = 0.0002, r2 = 0.92) and suggested a decrease in mR with
N availability in both AC and EC populations (Figure 6a). In all
other cases, the two models were statistically indistinguishable
and none of the simple model slopes were statistically different
from zero. While mR decreased with native N availability among
both AC and EC populations, the decrease was sharper for EC
trees such that the eCO2-induced enhancement decreased in
unfertilized subplots as native N availability increased (from
2.6 to 4.5 g N m−2, Figure 6a).
No significant variation in mR with N availability or CO2 treat-
ment was observed for suppressed individuals (Table 1, Figure 6).
Additionally, although mR of all unfertilized dominant individuals
pooled is significantly related to N availability (i.e., Figure 6a),
the effect of fertilization on mR is not significant (Table 1).
Both eCO2 and fertilization increased NPPW among our study
trees, though the effects were not significant. However, the
trends are consistent with other previous results from the FACE
study showing significant enhancement to NPPW under eCO2
and with increasing N (McCarthy et al. 2010).

Figure 6. Spatial variation in subplot-scale m R with mineral N avail- Discussion

ability is shown for unfertilized and fertilized dominant (a) and sup-
pressed (b) populations. Among unfertilized dominants, a significant Recently, large-scale bark beetle outbreaks have caused
eCO2 enhancement of m R was observed (P = 0.03), and resin flow extensive tree mortality in several distinct infestations, includ-
decreased significantly with native N availability (P = 0.0002). In all
panels, error bars show one standard error of the mean. Symbols are ing the ongoing mountain pine beetle (Dendroctonus
as in Figure 4. ponderosae) outbreak in British Columbia, Canada that has

Tree Physiology Online at [Link]

 8 Novick et al.

affected over 100,000 km2 of lodgepole pine stands (Kurz GEPaCO2 , NPPW and mR largely conform to expectations from
et al. 2008), the recent outbreak of pinyon ips (Ips confusus) GDB theory.
beetles in pinyon pine woodlands in southwestern North
America (Raffa et al. 2008), and the outbreak of spruce bee- The use of GPP and NPPW as proxies for A′ and NPP
tles (Dendroctonus rufipennis) in white pine forests in Alaska Previously, direct estimates of Anet and the different compo-
in the 1990s (Berg et al. 2006). These outbreaks drew politi- nents of NPP have been made for the AC and EC plots (Schäfer
cal and scientific attention to the effect of bark beetles on et al. 2003, McCarthy et al. 2010), permitting direct estimation
their host species (Negron et al. 2008) and the potential for of A′ and NPP. However, monthly estimates of these variables
increased severity of such damage in future climates (Gan are not available for our study period; therefore, we rely on
2004, Kurz et al. 2008). Bark beetle attacks are predicted to GEPaCO2 and NPPW as proxies for A′ and NPP. In this section, we
affect even larger areas in the future as warmer temperatures discuss how temporal and spatial patterns of these proxies may
associated with climate change may be beneficial to beetle differ from expected temporal and spatial trends in A′ and NPP.

Downloaded from [Link] at DigiTop USDA's Digital Desktop Library on January 10, 2012
voltinism (Logan et al. 2003, Gan 2004, Berg et al. 2006, Gross ecosystem productivity is a gross assimilation flux,
Kurz et al. 2008). whereas A′ is gross assimilation less maintenance respiration
An important contributor to increased temperature is the (i.e., A′ ~ GEP − RM). Therefore, treatment effects on the tem-
rising concentration of atmospheric CO2, and our study is the poral evolution of RM, which is primarily controlled by tempera-
first to quantify the impact of eCO2 on resin flow in mature ture and the amount of respiring biomass, may produce
pine trees. We show that eCO2 signficantly enhances resin differences in the temporal pattern of A′ across treatments.
flow (by 140%) in dominant trees growing in low- to mid- Temperature does not differ significantly among the treatment
range, but not high, fertility (Figure 6). We also show that plots, and as we demonstrate in Figure 3c–j, patterns of NPPW
while this enhancement persists throughout much of the year (and hence biomass accumulation) are also similar across
(Figure 4), the magnitude of resin flow varies significantly treatments. We note that foliage biomass, the plant component
from month to month (Figure 3–5). In the following sections, contributing most to RM, was shown to have similar dynamics
we link these results to expectations from GDB theory and to among aCO2 and eCO2 treatments (Schäfer et al. 2003,
previous observations of the effect of eCO2 and fertilization on McCarthy et al. 2007). The effect of fertilization on the season-
carbon assimilation and allocation in the study site. We also ality of leaf area production is not yet known for the study site,
discuss the limitations of using GEP and NPPW as proxies for though previous work has shown that fertilization does not
A′ and NPP. affect the seasonal pattern of shoot and foliage development in
loblolly pine (Zhang et al. 1997). Thus, we have no reason to
Linking observed temporal trends and correlations to expect that the treatments promoted significant differences in
GDB expectations the dynamics of RM.
As presented in Figure 1, the GDB hypothesis proposes that A′, Treatment effects on leaf phenology may also influence the
NPP and mR will all be directly correlated provided A′ is limited temporal patterns of GEP, which we have assumed to be similar
by some factor F (i.e., Phase I). When A′ saturates with F, the across treatments in this analysis. Again, we do not expect
former becomes stationary while growth continues to increase large differences in the seasonality of leaf area production
such that an inverse relationship between NPP and mR is across sites. Furthermore, as most leaf area production occurs
expected (Phase II). Often, F is interpreted to represent spatial during May–July (i.e., Phase II), it is unlikely that plot-specific
gradients of nutrient availability, though it may represent any differences in leaf area expansion would affect the delineation
potentially limiting resource, including factors that vary in time between Phase I (March–April, September–October) and
(i.e., temperature, soil moisture content or even leaf area). Phase II (May–August). Similarly, it is also unlikely that site-
Measured GEPaCO2, which we use as a proxy for A′, was rela- specific differences in leaf area expansion would strongly
tively low at the beginning and the end of the growing season affect Phase I temporal relationships between mR,N and
(i.e., March–April, September–October, Figure 2b), and with GEPaCO2 ,N shown in Figure 5a and b. Variations in A′ are not
respect to temporal gradients of F, the study sites should be explicitly considered in the formulation of the relationships
conceptually located in Phase I during these months (see between mR,N and NPPW,N (Figure 5c–f), and our interpretation
Figure 1). As expected, direct correlations between GEPaCO2, mR of the results relies on the practical assumption that A′ is high
and NPPW were observed during this time (Figure 5a–d). and stationary during Phase II months. Clearly, uncertainty
Conversely, when GEPaCO2 was relatively high and stationary attributed to temporal variation in A′ during Phase II months
(May–August, Figure 3b), the study sites should be conceptu- (i.e., Figure 3b), as well as site-specific differences in A′, may
ally located in Phase II. And as expected, inverse relationships contribute to some of the unexplained variation between mR,N
between NPPW and mR were observed during these months and NPPW,N.
(Figure 5e and f). Thus, observed temporal variability in

Tree Physiology Volume 00, 2012

 Elevated CO2 increases resin flow 9

Because NPP = NPPW + NPP foliage + NPP fine-root, the dynamics (e.g., ∂(A′ − NPP)/∂F < 0, where F is nutrient availability)
of NPPW might differ from that of NPP if leaf biomass produc- because carbon investment in leaf area results in diminishing
tion (NPP foliage) and fine root production (NPP fine-root) are a returns of absorbed light (Palmroth et al. 2006). Thus, when F
large proportion of total NPP and follow a different seasonal represents spatial gradients of nutrient availability, the plots of
pattern from NPPW. Developing an estimate of NPP for this the study site fit conceptually in Phase II of Figure 1, and GDB
analysis is difficult for a number of reasons. First, some of the theory would predict a decrease in mR with increasing soil fer-
relevant variables (including leaf and fine root expansion) are tility. Indeed, resin flow of dominant trees grown in both eCO2
not available for 2009. Second, NPP must be specified at the and aCO2 subplots was inversely related to native N availability
same scale as resin flow (i.e., the tree scale) and indepen- (Figure 6a).
dently for dominant and suppressed populations, which The observed effects of eCO2 on mR also agree with predic-
requires assumptions regarding the extent to which patterns in tions from the GDB hypothesis. Namely, eCO2 increased resin
leaf and woody biomass production differ with tree size. flow in plots of low- to mid-range, but not high, soil fertility

Downloaded from [Link] at DigiTop USDA's Digital Desktop Library on January 10, 2012
Despite these challenges, we produced estimates of monthly (Figure 6). Previous work at the Duke FACE experiment has
NPP for each treatment class during our study period and shown that eCO2 stimulates photosynthesis by ~40% in unfer-
found strong correlation between NPP and NPPW within domi- tilized plots (Schäfer et al. 2003, Maier et al. 2008) with no
nant (r = 0.96) and suppressed (r = 0.94) populations. further enhancement expected with N addition (Palmroth et al.
Nevertheless, we elected to use NPPW instead of NPP in this 2006, Maier et al. 2008), yet enhances NPP by only 18% in
investigation because this estimate of growth contains much nutrient-poor plots compared with 30% in richer plots
less uncertainty both in time and among populations than NPP. (McCarthy et al. 2010). Thus, the observation that dominant
Furthermore, it can be argued that the most direct competition trees in the EC subplots have greater mR is consistent with a
with resin production is the local wood production sink for car- larger A′ − NPP balance in these trees relative to AC
bohydrates. Resin synthesis occurs in the stem, and the avail- populations.
ability of carbohydrates for resin production and stem growth We note that among eCO2 trees, N amendment did not
is controlled by both the size of the source and transport pro- increase mR relative to trees in the plots with the highest native
cesses moving carbohydrates from leaves to the stem. N availability (P = 0.58 of EF versus highest EC). Yet, resin flow
Treatment effects on the relative magnitude of assimilation increased in dominant AF versus AC trees in high native N sub-
and growth are not explicitly considered in this analysis, though plots (Figure 6a, P = 0.007)—an observation that does not
they are important for the interpretation of the results, and in agree with the GDB predictions. There are two possible expla-
particular the discussion presented in the next section. nations for the observed discrepancy. The first is that fertiliza-
Increases in GEP related to eCO2 or fertilization should also tion increased both assimilation and growth in AF subplots, but
enhance A′, provided that these treatments do not produce an with a relatively larger increase in assimilation. Earlier work has
enhancement to RM of similar magnitude. This is most certainly shown that fertilization tends to increase leaf area production
the case for eCO2, which enhances net canopy assimilation by in aCO2 plots by ~20–25%, though the effect is not significant
400–1000 g C m−2 year−1 (Schäfer et al. 2003) through (McCarthy et al. 2007). Fertilization has also been shown to
increases in both leaf-level assimilation capacity (Maier et al. significantly increase leaf-level assimilation capacity (i.e., Anet)
2008) and total leaf area (McCarthy et al. 2007), while enhanc- in aCO2 plots, with an 18% increase for current-year foliage
ing RM by <100 g C m−2 year−1 (Schäfer et al. 2003). (Figure 2; see also Maier et al. 2008). Consequently, increases
Enhancements to NPPW in EC as compared with AC subplots in leaf area and leaf-level assimilation capacity may have
should also be accompanied by similar enhancement to NPP, as resulted in ∂(A′ − NPP)/∂F > 0 by increasing ∂A′ /∂F with
previous work has shown that eCO2 does not affect the parti- increasing N. In eCO2 subplots, canopy leaf area was already
tioning of assimilated carbon among plant biomass pools 25% higher than aCO2 subplots and substantial gains in leaf
(McCarthy et al. 2010). On the other hand, the ratio of NPPW/ area with fertilization were not possible. Thus, in eCO2 sub-
NPP may not be constant between fertilized and control sectors plots, little additional gain in ∂A′ /∂F is expected with increas-
as fertilization affects the ratio of belowground to aboveground ing N.
carbon allocation (Palmroth et al. 2006). Such a shift in parti- The second explanation is that under fertilization, partition-
tioning may explain the increase in resin flow observed in AF as ing of carbon was redirected from belowground processes to
compared with AC plots as discussed in the following section. aboveground processes (Palmroth et al. 2006), and could thus
support increases of both growth and mR in apparent contra-
Linking observed spatial variability in mR to GDB diction of the GDB scheme. However, if the latter is the domi-
predictions nant reason, the effect should have been noticeable under
In forests of dense, light-limited canopies, increased nutrient eCO2 as well, and thus the behavior seems more consistent
availability typically enhances growth more than photosynthesis with our first explanation.

Tree Physiology Online at [Link]

 10 Novick et al.

Table 2. Synthesis of published studies that observed the impact of fertilization on resin flow.

Species Location Age Other treatments Fertilization effect Reference

Pinus sylvestris Harjavalta, Finland 60–90 ... No effect Kyto et al. (1999)
Pinus sylvestris Harjavalta, Finland 60–90 Defoliated No effect Kyto et al. (1999)
Pinus taeda Louisiana, USA 25 ... No effect Lombardero et al. (2000)
Pinus taeda Louisiana, USA 25 Thinned No effect Lombardero et al. (2000)
Pinus taeda Louisiana, USA 15 ... No effect Ruel et al. (1998)
Pinus taeda Louisiana, USA 13 Thinned mR decreased Wilkens et al. (1998)
Pinus taeda Louisiana, USA 13 ... No effect Wilkens et al. (1998)
Pinus taeda North Carolina, USA 16 ... No effect Klepzig et al. (2005)
Pinus taeda North Carolina, USA 16 Irrigated No effect Klepzig et al. (2005)
Pinus taeda North Carolina, USA 6 ... mR increased Knebel et al. (2008)
Pinus taeda North Carolina, USA 12 ... mR increased Knebel et al. (2008)

Downloaded from [Link] at DigiTop USDA's Digital Desktop Library on January 10, 2012
Pinus taeda North Carolina, USA 11 ... mR decreased Warren et al. (1999)
Pinus taeda Virginia, USA 24 ... No effect Matson et al. (1987)
Pinus taeda Virginia, USA 24 Thinned No effect Matson et al. (1987)

Resin flow in suppressed individuals plant defense properties and the number of attacking beetles
(Waring and Pitman 1985, Coops et al. 2009). Thus, the
In the sub-canopy, light limitations significantly reduce GPP in observed eCO2-induced enhancement of resin should be con-
suppressed individuals (Schäfer et al. 2003) such that light sidered when making predictions about the frequency and
may be the primary limiting resource to assimilation in these severity of bark beetle attacks in future climates, at least in
individuals. In this study, NPPW was also lower for suppressed forests with low- to mid-range fertility, which currently
trees. However, during the beginning and end of the growing represent the majority of southern pine forests (Fox et al.
season (i.e., March–April, September–October), mR scaled 2007). Indeed, previous studies have shown that even more
with GEPaCO2 and NPPW among suppressed individuals (Figure modest increases in resin flow (i.e., enhancements ≤ 100%)
5b and d), suggesting that ∂(A′ − NPP)/∂F is nonetheless pos- significantly increase the survival probability of pine trees
itive. During the peak of the growing season (May–August), experiencing bark beetle attack (Reeve et al. 1995, Strom
when leaf area (and light penetration in the canopy) is rela- et al. 2002).
tively stationary, inverse relationships between GEPaCO2 and mR Under current atmospheric CO2 concentrations, fertilization
were observed (Figure 5f). These results follow expectations may increase resin flow (as observed in the ambient plots in
from GDB theory if GEPaCO2 was also stationary among sup- this study) or may have no effect on resin flow (as demon-
pressed individuals during this portion of the study period. In strated in the majority of studies investigating the response of
space, the fertilization and eCO2 treatments had no effect on mR to fertilization) (see Table 2). However, under future condi-
mR in suppressed individuals (Figure 6b), which is consistent tions when protection from increased attacks is needed, fertil-
with GDB predictions if light is indeed more limiting to their ization may reduce pine resistance to beetles. Moreover, any
photosynthesis than N availability or atmospheric CO2 changes in weather patterns that enhance NPP (e.g., longer
concentration. growing season) or that limit A′ (e.g., severe drought) will likely
do so at the expense of resin production. In unmanaged, dense
stands, which cover vast areas of temperate and boreal for-
Implications
ests, many individuals are effectively suppressed with a small
The Duke FACE experiment permitted us to quantify for the A′ − NPP balance. This study suggests that resistance to bark
first time the effect of elevated atmospheric CO2 on resin flow beetle attack in suppressed trees will not increase with CO2,
of mature, field-grown pine trees in the context of carbon regardless of soil fertility.
source–sink interactions, utilizing variation in the balance of
A′ − NPP induced spatially by both N availability and tree
Acknowledgments
position in the canopy, and temporally by seasonal dynamics.
Following predictions based on the GDB scheme, eCO2 We acknowledge Eric Ward and Sari Palmroth for feedback on the
enhanced resin flow in dominant trees. The absolute enhance- analysis, and Jeff Pippen, Andrea Casanova and Javier Luscarain
ment was most clear during periods when growth was sea- for data collection assistance. We also acknowledge three anony-
sonally low (Figures 3d–j and 4) or limited by N availability mous reviewers for insightful comments on the manuscript. We
(Figure 6a). Prior studies have shown that tree mortality dur- thank Fred Hain and John Strider for methodological and material
ing bark beetle attacks follows threshold dynamics driven by assistance.

Tree Physiology Volume 00, 2012

 Elevated CO2 increases resin flow 11

Funding Logan, J.A., J. Regniere and J.A. Powell. 2003. Assessing the impacts
of global warming on forest pest dynamics. Front. Ecol. Environ.
1:130–137.
This research was sponsored by the Office of Science (BER), Lombardero, M.J., M.P. Ayres, P.L. Lorio and J.J. Ruel. 2000.
US Department of Energy (FG02-95SER62083), and by the Environmental effects on constitutive and inducible resin defences
NSF Graduate Research Fellowship Program (DGE1106401) of Pinus taeda. Ecol. Lett. 3:329–339.
and the James B. Duke Fellowship Program. Lorio, P.L. 1986. Growth-differentiation balance—a basis for under-
standing southern pine-beetle tree interactions. For. Ecol. Manag.
14:259–273.
Lorio, P.L., F.M. Stephen and T.D. Paine. 1995. Environment and ontog-
References eny modify loblolly-pine response to induced acute water deficits
Berg, E.E., J.D. Henry, C.L. Fastie, A.D. De Volder and S.M. Matsuoka. and bark beetle attack. For. Ecol. Manag. 73:97–110.
2006. Spruce beetle outbreaks on the Kenai Peninsula, Alaska, and Maier, C.A., S. Palmroth and E. Ward. 2008. Short-term effects of fer-
Kluane National Park and Reserve, Yukon Territory: relationship to tilization on photosynthesis and leaf morphology of field-grown lob-
summer temperatures and regional differences in disturbance lolly pine following long-term exposure to elevated CO2

Downloaded from [Link] at DigiTop USDA's Digital Desktop Library on January 10, 2012
regimes. For. Ecol. Manag. 227:219–232. concentration. Tree Physiol. 28:597–606.
Coops, N.C., R.H. Waring, M.A. Wulder and J.C. White. 2009. Prediction Matson, P.A., F.P. Hain and W. Mawby. 1987. Indices of tree susceptibil-
and assessment of bark beetle-induced mortality of lodgepole pine ity to bark beetles vary with silvicultural treatment in a loblolly pine
using estimates of stand vigor derived from remotely sensed data. plantation. For. Ecol. Manag. 22:107–188.
Remote Sens. Environ. 113:1058–1066. McCarthy, H.R., R. Oren, A.C. Finzi, D.S. Ellsworth, H.S. Kim, K.H.
Ellsworth, D.S. 1999. CO2 enrichment in a maturing pine forest: are Johnsen and B. Millar. 2007. Temporal dynamics and spatial variabil-
CO2 exchange and water status in the canopy affected? Plant Cell ity in the enhancement of canopy leaf area under elevated atmo-
Environ. 22:461–472. spheric CO2. Glob. Change Biol. 13:2479–2497.
Finzi, A.C., D.J.P. Moore, E.H. DeLucia, et al. 2006. Progressive N limi- McCarthy, H.R., R. Oren, K.H. Johnsen, A. Gallet-Budynek, S.G.
tation of ecosystem processes under elevated CO2 in a warm tem- Pritchard, C.W. Cook, S.L. LaDeau, R.B. Jackson and A.C. Finzi.
perature forest. Ecology 87:15–25. 2010. Re-assessment of plant carbon dynamics at the Duke free-air
Fox, T.R., E.J. Jokela and H.L. Allen. 2007. The development of pine CO2 enrichment site: interactions of atmospheric CO2 with nitrogen
plantation silviculture in the southern United States. J. For. and water availability over stand development. New Phytol.
105:337–347. 185:514–528.
Gan, J.B. 2004. Risk and damage of southern pine beetle outbreaks McDowell, N.G., H.D. Adams, J.D. Bailey and T.E. Kolb. 2007. The role of
under global climate change. For. Ecol. Manag. 191:61–71. stand density on growth efficiency, leaf area index, and resin flow in
Goulden, M.L., B.C. Daube, S.-M. Fan, D.J. Sutton, A. Bazzaz, J.W. Munger southwestern ponderosa pine forests. Can. J. For. Res. 37:343–355.
and S.C. Wofsy. 1997. Physiological responses of a black spruce for- Negron, J.F., B.J. Bentz, C.J. Fettig, et al. 2008. US Forest Service bark
est to weather. J. Geophys. Res. Atmos. 102:28987–28996. beetle research in the western United States: looking toward the
Hendrey, G.R., D.S. Ellsworth, K.F. Lewin and J. Nagy. 1999. A free-air future. J. For. 106:325–331.
enrichment system for exposing tall forest vegetation to elevated Oren, R., D.S. Ellsworth, K.H. Johnsen, et al. 2001. Soil fertility limits
atmospheric CO2. Glob. Change Biol. 5:293–309. carbon sequestration by forest ecosystems in a CO2-enriched atmo-
Herms, D.A. and W.J. Mattson. 1992. The dilemma of plants—to grow sphere. Nature 411:469–472.
or defend. Q. Rev. Biol. 67:283–335. Palmroth, S., C.A. Maier, H.R. McCarthy, A.C. Oishi, H.S. Kim, K.H.
Hofmockel, K.S. and W.H. Schlesinger. 2007. Carbon dioxide effects Johnsen, G.G. Katul and R. Oren. 2005. Contrasting responses to
on heterotrophic dinitrogen fixation in a temperate pine forest. Soil. drought of forest floor CO2 efflux in a loblolly pine plantation and a
Sci. Soc. Am. J. 71:140–144. nearby oak-hickory forest. Glob. Change Biol. 11:421–434.
Johnsen, K., B. Teskey, L. Samuelson, J. Butnor, D. Sampson, F. Palmroth, S., R. Oren, H.R. McCarthy, K.H. Johnsen, A.C. Finzi, J.R.
Sanchez, C. Maier and S. McKeand. 2004. Carbon sequestration in Butnor, M.G. Ryan and W.H. Schlesinger. 2006. Aboveground sink
loblolly pine plantations: methods, limitations, and research needs strength in forests controls the allocation of carbon below ground
for estimating storage pools. In Southern Forest Science: Past, and its [CO2] - induced enhancement. Proc. Natl Acad. Sci. USA
Present, and Future. GTR-SRS-75. Eds. M.H. Rauscher and K. 103:19362–19367.
Johnsen. USDA Forest Service, Southern Research Station, Asheville. Papale, D., M. Reichstein, M. Aubinet, et al. 2006. Towards a standard-
Klepzig, K.D., D.J. Robison, G. Fowler, P.R. Minchin, F.P. Hain and H.L. ized processing of net ecosystem exchange measured with eddy
Allen. 2005. Effects of mass inoculation on induced oleoresin covariance technique: algorithms and uncertainty estimation.
response in intensively managed loblolly pine. Tree Physiol. Biogeosciences 3:571–583.
25:681–688. Phillips, R.L., S.C. Whalen and W.H. Schlesinger. 2001. Influence of
Knebel, L., D.J. Robison, T.R. Wentworth and K.D. Klepzig. 2008. Resin atmospheric CO2 enrichment on nitrous oxide flux in a temperate
flow responses to fertilization, wounding and fungal inoculation in lob- forest. Glob. Biogeochem. Cy. 15:741–752.
lolly pine (Pinus taeda) in North Carolina. Tree Physiol. 28:847–853. Price, T., C. Doggett, J. Pye and B. Smith. 1998. A history of southern
Kurz, W.A., C.C. Dymond, G. Stinson, G.J. Rampley, E.T. Neilson, A.L. pine beetle outbreaks in the southeastern United States. Georgia
Carroll, T. Ebata and L. Safranyik. 2008. Mountain pine beetle and Forestry Commission, Atlanta, GA.
forest carbon feedback to climate change. Nature 452:987–990. Raffa, K.F., B.H. Aukema, B.J. Bentz, A.L. Carroll, J.A. Hicke, M.G. Turner
Kyto, M., P. Niemela, E. Annila and M. Varama. 1999. Effects of forest and W.H. Romme. 2008. Cross-scale drivers of natural disturbances
fertilization on the radial growth and resin exudation of insect-defo- prone to anthropogenic amplification: the dynamics of bark beetle
liated Scots pines. J. Appl. Ecol. 36:763–769. eruptions. Bioscience 58:501–517.
Lapasha, C.A. and E.A. Wheeler. 1990. Resin canals in Pinus taeda: Reeve, J.D., M.P. Ayres, P.L. Lorio, Jr, N. Cappuccino and P.W. Price.
longitudinal canal lengths and interconnections between longitudi- 1995. Host suitability, predation, and bark beetle population dynam-
nal and radial can. IAWA Bull. 11:227–238. ics. In Population Dynamics: New Approaches and Synthesis. Eds. N.

Tree Physiology Online at [Link]

 12 Novick et al.

Cappuccino and P.W. Price. Academic Press, San Diego, pp Strom, B.L., R.A. Goyer, L.L. Ingram, Jr, G.D.L. Boyd and L.H. Lott.
339–357. 2002. Oleoresin characteristics of progeny of loblolly pines that
Reichstein, M., E. Falge, D. Baldocchi, et al. 2005. On the separation of escaped attack by the southern pine beetle. For. Ecol. Manag.
net ecosystem exchange into assimilation and ecosystem respira- 157:169–178.
tion: review and improved algorithm. Glob. Change Biol. Wallin, K.F., T.E. Kolb, K.R. Skov, and M. Wagner. 2008. Forest manage-
11:1424–1439. ment treatments, tree resistance, and bark beetle resource utiliza-
Ruel, J.J., M.P. Ayres and P.L. Lorio. 1998. Loblolly pine responds to tion in ponderosa pine forests of northern Arizona. For. Ecol. Manag.
mechanical wounding with increased resin flow. Can. J. For. Res. 255:3263–3269.
28:596–602. Waring, R.H. and G.B. Pitman. 1985. Modifying lodgepole pine stands
Schäfer, K.V.R., R. Oren, D.S. Ellsworth, C.T. Lai, J.D. Herrick, A.C. Finzi, to change susceptibility to mountain pine-beetle attack. Ecology
D.D. Richter and G.G. Katul. 2003. Exposure to an enriched CO2 66:889–897.
atmosphere alters carbon assimilation and allocation in a pine forest Warren, J.M., H.L. Allen and F.L. Booker. 1999. Mineral nutrition, resin
ecosystem. Glob. Change Biol. 9:1378–1400. flow and phloem phytochemistry in loblolly pine. Tree Physiol.
Sparks, J.P., J. Walker, A. Turnipseed and A. Guenther. 2008. Dry nitro- 19:655–663.
gen deposition estimates over a forest experiencing free air CO2 Wilkens, R.T., M.P. Ayres, P.L. Lorio, Jr and J.D. Hodges. 1998.

Downloaded from [Link] at DigiTop USDA's Digital Desktop Library on January 10, 2012
enrichment. Glob. Change Biol. 14:768–781. Environmental effects on pine tree carbon budgets and resistance
Stamp, N. 2004. Can the growth-differentiation balance hypothesis be to bark beetles. In The Productivity and Sustainability of Southern
tested rigorously? Oikos 107:439–448. Forest Ecosystems in a Changing Environment. Eds. R.A. Mickler
Stoy, P.C., G.G. Katul, M.B.S. Siqueira, J.Y. Juang, K.A. Novick, J.M. and S. Fox. Springer, New York, pp 591–616.
Uebelherr and R. Oren. 2006. An evaluation of models for partition- Zhang, S.S., H.L. Allen and P.M. Dougherty. 1997. Shoot and foliage
ing eddy covariance-measured net ecosystem exchange into photo- growth phenology of loblolly pine trees as affected by nitrogen fer-
synthesis and respiration. Agric. For. Meteorol. 141:2–18. tilization. Can. J. For. Res. 27:1420–1426.

Tree Physiology Volume 00, 2012