Content uploaded by William H Schlesinger
Author content
All content in this area was uploaded by William H Schlesinger
Content may be subject to copyright.
Biogeochemistry
Canopy N and P dynamics of a southeastern US
pine forest under elevated CO
2
ADRIEN C. FINZI
1,
*, EVAN H. DELUCIA
2
and WILLIAM H. SCHLESINGER
3
1
Department of Biology, Boston University, Boston, MA 02215, USA;
2
Department of Plant Biology,
University of Illinois, Urbana, IL 61801, USA;
3
Nicholas School of the Environment and Earth Sciences,
Duke University, Durham, NC 27708, USA; *Author for correspondence (e-mail: afinzi@bu.edu)
Received Publ.: Pl. provide the received date; accepted in revised form 11 September 2003
Key words: Elevated CO
2
, Foliar N:P ratios, Nitrogen, Phosphorus
Abstract. Forest production is strongly nutrient limited throughout the southeastern US. If nutrient
limitations constrain plant acquisition of essential resources under elevated CO
2
, reductions in the mass
or nutrient content of forest canopies could constrain C assimilation from the atmosphere. We tested this
idea by quantifying canopy biomass, foliar concentrations of N and P, and the total quantity of N and P in
a loblolly pine (Pinus taeda) canopy subject to 4 years of free-air CO
2
enrichment. We also used N:P
ratios to detect N versus P limitation to primary production under elevated CO
2
. Canopy biomass was
significantly higher under elevated CO
2
during the first 4 years of this experiment. Elevated CO
2
significantly reduced the concentration of N in loblolly pine foliage (5% relative to ambient CO
2
) but not
P. Despite the slight reduction foliage N concentrations, there were significant increases in canopy N and
P contents under elevated CO
2
. Foliar N:P ratios were not altered by elevated CO
2
and were within a
range suggesting forest production is N limited not P limited. Despite the clear limitation of NPP by N
under ambient and elevated CO
2
at this site, there is no evidence that the mass of N or P in the canopy is
declining through the first 4 years of CO
2
fumigation. As a consequence, whole-canopy C assimilation is
strongly stimulated by elevated CO
2
making this forest a larger net C sink under elevated CO
2
than under
ambient CO
2
. We discuss the potential for future decreases in canopy nutrient content as a result of
limited changes in the size of the plant-available pools of N under elevated CO
2
.
Introduction
The relationship between soil nutrient availability, foliage biomass production and
canopy nutrient content is a key determinant of C storage in terrestrial ecosystems
exposed to elevated CO
2
. Photosynthesis is strongly dependent upon the con-
centration of nitrogen (N) and phosphorus (P) in leaves (Pearcy et al. 1987; Pe-
terson et al. 1999) and gross primary production depends on the photosynthesis –
nutrient relationship throughout the plant canopy. If elevated CO
2
or limiting
quantities of soil nutrients alter the production of leaves or the distribution of N and
P within plant canopies, the magnitude of the C influx via photosynthesis will
change (Luo and Reynolds 1999).
Foliage biomass and net primary production (NPP) are highly correlated across a
broad range of forest types (Gholz 1982; Webb et al. 1983; Vose and Allen 1988;
Piatek and Allen 2000). In young forest ecosystems, foliage biomass increases
rapidly during stand development (Marks and Bormann 1972). With their high
69:
#2004 Kluwer Academic Publishers. Printed in the Netherlands.
363–378, 2004.
nutrient contents, leaves are an important sink for available nutrients in plants.
For example, the production of new foliage in young loblolly pine forests (15–25
years old) sequesters between 30 and 70% of the total annual nutrient
uptake (Switzer and Nelson 1972), ranging between 50–83 kg N ha
1
year
1
and 5–
11 kg P ha
1
year
1
(Piatek and Allen 2000 and references therein). This large
annual requirement for N and P accounts for widespread nutrient limitation to NPP
throughout southeastern US forests (Richter and Markewitz 2001).
The available data suggest a wide range of canopy responses to elevated CO
2
.
Most short-term experiments (<4 years) on tree seedlings and saplings demonstrate
increases in foliage biomass production and canopy N content under elevated CO
2
(e.g., Murray et al. 1996; Rey and Jarvis 1997; Norby et al. 1999; Jach et al. 2000).
Notable exceptions report that foliage production and canopy-N content do not
respond to elevated CO
2
in the absence of increases in soil nutrient availability
(Ha
¨ttenschwiler and Ko
¨rner 1998; Zak et al. 2000; Oren et al. 2001). These may
represent extreme cases of nutrient limitation where nutrient supply prohibits even
a transient response to elevated CO
2
. Very little data are available on plant P pools
under elevated CO
2
(Johnson et al. 1997; Walker et al. 2000).
Primary production in southeastern US forests is commonly nutrient limited
(e.g., Piatek and Allen 2000; Richter et al. 2000). While much of the focus on
nutrient limitation in the temperate zone focuses of N (c.f. Vitousek and Howarth
1991), the soils of the piedmont region of the southeastern US are highly weathered
with the potential for P limitations due to the depletion of P-containing primary
minerals and the occlusion of P into unavailable forms for plant growth (Walker
and Syers 1976; Crews et al. 1995; Cross and Schlesinger 1995; Vitousek and
Farrington 1997). Foliar nutrient status is commonly used to detect the nature and
extent of soil nutrient limitations to NPP (e.g., Valentine and Allen 1990). Foliar
N:P ratios have been used to assess N versus. P limitation to primary production in
grassland and heathland ecosystems (Verhoeven et al. 1996) and N saturation in
ecosystems of the western United States (Fenn et al. 1996).
Biogeochemical models predict that the magnitude and direction of the CO
2
stimulation of NPP is controlled by the rate of nutrient mineralization from soil
organic matter or weathering from source pools (McMurtrie and Comins 1996;
Rastetter et al. 1997; Luo and Reynolds 1999). These models suggest that if soil
nutrients limit NPP under elevated CO
2
, then the initial increase in canopy biomass
and nutrient content under elevated CO
2
should attenuate through time, and
eventually no longer be significantly different from that under ambient CO
2
.We
have studied the productivity response of a rapidly aggrading pine forest grown at
elevated CO
2
in the Duke Forest, NC (DeLucia et al. 1999; Hamilton et al. 2002).
Both atmospheric CO
2
concentrations and soil N availability limit NPP in this
ecosystem (Finzi et al. 2002). The potential for P limitation to NPP has not been
explored in this ecosystem. Thus, the objectives of this study were to (1) assess the
effect of elevated CO
2
on the concentration of N and P throughout the loblolly pine
canopy, (2) use N:P ratios to assess the degree of N versus P limitation to NPP, and
(3) determine the temporal pattern in canopy biomass production and nutrient
content under ambient and elevated CO
2
.
364
Methods and materials
Site description
The FACE experiment in the Duke Forest (Orange County, North Carolina, USA) is
composed of six 30-m diameter plots. Three experimental plots are fumigated with
CO
2
to maintain the atmospheric CO
2
concentration 200 mll
1
above ambient.
Three control plots are fumigated with ambient air only. The experiment began 27
August 1996 and is continuous (24 h day
1
; 365 days year
1
). Details on FACE
operation can be found in Hendrey et al. 1999.
The forest is derived from 3-year-old loblolly pine (Pinus taeda) seedlings that
were planted in 1983 in a 2.4 2.4-m spacing. In 1996, the trees were approxi-
mately 14 m tall and accounted for 98% of the basal area of the stand. Since
planting, a deciduous understory layer has recruited from nearby hardwood forests
and stump sprouts. The most abundant understory tree species is sweet gum (Li-
quidambar styraciflua), with admixtures of red maple (Acer rubrum), red bud
(Cercis canadensis), and dogwood (Cornus florida). The 32-ha site contains an
elevation gradient of 15-m between the highest and lowest points, but topographic
relief is 18throughout. Soils are classified in the Enon Series (fine, mixed, active,
thermic Ultic Hapludalfs). Enon soils, derived from mafic bedrock, are slightly
acidic (0.1 M CaCl
2
pH ¼5.75), and have well-developed soil horizons with mixed
clay mineralogy. These soils are very deep (>15m) but organic C, N and P are
concentrated in the upper 15 cm of the soil profile (Schlesinger and Lichter 2001).
Additional site details can be found in Schlesinger and Lichter (2001) and Finzi
et al. (2001).
Canopy sampling and chemistry
The longevity of loblolly pine foliage in the Piedmont of NC is 19 months (Zhang
and Allen 1996) so that at any time there are needles of two different ages on a
single branch (e.g., 2 cohorts). We collected foliage samples from both needle
cohorts–those produced in the current year and those produced in the previous
year–at three heights within the canopy; the bottom 25%, the middle 50%, and the
top 25% of the crown. The canopy divisions were based on the large variation in the
specific-leaf area (SLA) of projected foliage with crown depth (DeLucia et al.
2002). SLA was 33.5 cm
2
g
1
at the top of the canopy, 38.2cm
2
g
1
in the middle of
the canopy and 45.6 cm
2
g
1
at the bottom (DeLucia et al. 2002).
We collected canopy samples above a randomly selected location on each arm of
a cross-shaped boardwalk that extends through each FACE ring to the north, south,
east, and west. At each of these locations and heights, we sampled a single branch
on each of four trees and collected 5–8 fascicles of current and year-old foliage
along a primary branch. Foliage samples were collected in June and September of
each year from 1997 through 2000. The June sample represents the initial con-
centration of N and P in the first fully expanded flush of new needles, and the
365
September sample represents the peak canopy biomass N and P pools (Zhang and
Allen 1996).
We compared the biomass of foliage predicted from pretreatment allometric
relationships with the biomass predicted from the leaves collected in litter baskets
(see below, Finzi et al. 2002). Foliage biomass predicted allometrically was con-
sistently smaller than that predicted from the litter baskets. We cannot harvest trees
from the replicated experiment at this time, leading to uncertainties in the allo-
metric relationship between tree diameter and foliage biomass under elevated CO
2
.
Thus, we opted to calculate foliage biomass from collections of litterfall (cf. Finzi
et al. 2002).
Aboveground litterfall mass was collected from 5 June 1996 onward by placing
12 replicate 40 cm 40 cm baskets in each plot. Litterfall was collected once per
month between January and August and twice per month between September and
December to minimize leaching losses from leaf litter during the period of peak
litterfall (Finzi et al. 2001). The samples were brought to the laboratory, dried at
65 8C for 4 days, and weighed. The litter was sorted into seven categories, including
pine needles. Given the longevity of loblolly pine foliage, a new cohort of leaves
produced in 1 year does not abscise until the following year. Thus the peak biomass
of loblolly pine needles in the canopy in a given year is the sum of litterfall in that
year and in the following year. For example, the mass of loblolly pine needles in the
canopy in 1998 can be estimated from the sum of litterfall mass in 1998–the
needles initially produced in 1997 but present in the canopy during the 1998
growing season, and 1999–the needles produced in 1998 that did not abscise until
the end of the growing season in 1999. We used litterfall mass data from 1997
through 2001 to calculate leaf biomass for the period 1997–2000. Repeated mea-
sures ANOVA indicated that the difference in leaf mass per unit area (LMA,
mg cm
2
) between green leaf and litter samples of loblolly pine was not sig-
nificantly different during the first 4 years of this experiment (Finzi et al. 2002).
Rather than multiplying litterfall mass by the ratio of green LMA and litter LMA,
we used the simpler assumption that LMA was not different and that the mass of
leaves for a given year in the canopy is the same as that collected as litterfall.
We measured the N and P concentration of green leaves in a sulfuric–salicylic
acid Kjeldahl digestion (Lowther 1980) followed by colorimetric analysis on an
automated ion analyzer (Lachat QuickChem FIAþ8000 Series, Zellweger Ana-
lytics, Milwaukee, WI). We calculated the quantity of N and P throughout the entire
pine canopy by multiplying the concentration of N or P at a given canopy position
(bottom, middle, or top) by the foliage biomass in that canopy position. We as-
sumed that 25% of the foliage biomass was in the lower canopy, 50% was in the
mid-canopy and 25% was in the upper canopy.
Statistical analysis
We used repeated-measures ANOVA to determine the influence of atmospheric
CO
2
(2 levels: ambient, elevated), canopy position (3 levels: bottom, middle, top)
366
and needle age class (2 levels: new, old) on N and P concentrations and their ratio.
There was large pretreatment variation in the mass and the quantity of N and P in
the loblolly pine canopy across the six plots that comprise this experiment (Finzi
et al. 2001). Thus we used analysis of covariance to determine the influence of
atmospheric CO
2
on the mass of foliage and the content of N and P in the loblolly
pine canopy on a year-by-year basis. All data were assessed for normality and
homogeneity of variance following ANOVA or ANCOVA. We used Tukey’s test to
compare means among the different levels of the fixed effects.
Results
Pine needle N and P concentrations and ratios
Elevated CO
2
slightly but significantly decreased the concentration of N in loblolly
pine needles (Table 1). Concentrations of N were 10.6 0.1 mg g
1
under ambient
CO
2
and 10.1 0.1 mg g
1
under elevated CO
2
. Concentrations of N in loblolly
pine foliage under elevated CO
2
were consistently lower across all canopy positions
(Figure 1). There were large differences in N concentrations in the different age
classes, and there was a significant age class xposition-within-the-canopy inter-
action (Table 1). Concentrations of N were lower in 1-year-old needles than in
needles produced within the current year. Concentrations of N were lowest in the
bottom of the canopy and highest at the top of the canopy in the needles produced
within the current year (Figure 2). Conversely, concentrations of N were highest in
the bottom of the canopy and lowest in the top of the canopy in the 1-year-old
needles. Concentrations of N in foliage were significantly higher early in the
growing season (11.1 0.1 mg g
1
in June) than in the late growing season (Sep-
tember, 9.6 0.1 mg g
1
), when canopy N reaches its peak (Table 1). The con-
centration of N in foliage was significantly higher in 1997 and 2000 than in 1998
and 1999 (Table 1).
Elevated CO
2
had no effect on the concentration of P in pine needles (Table 1,
Figure 1). The concentration of P in foliage was significantly higher at the top of the
canopy (1.15 0.04 mg g
1
) than the middle of the canopy (1.05 0.03 mg g
1
)
but not significantly different than in the bottom of the canopy (1.06 0.03 mg g
1
,
Table 1). Concentrations of P in loblolly pine foliage were significantly lower in the
1-year-old needles (0.89 0.01 mg g
1
) than in the current year needles
(1.29 0.03 mg g
1
, Table 1). Concentrations of P were significantly higher early
in the growing season (1.26 0.03 mg g
1
in June) than late in the growth season
(0.92 0.01 mg g
1
in September, Table 1). Although P concentrations varied from
year-to-year and between years and needle age classes, seasons and canopy posi-
tions, there were no significant year xCO
2
or higher order interactions (Table 1).
Pine needle N:P ratios were generally lower under elevated CO
2
(Table 2), but
the effect of elevated CO
2
was not statistically significant (Table 1). N:P ratios were
significantly higher in 1-year-old foliage (11.0 0.2) than in current year foliage
(9.2 0.2, Table 1), and significantly lower in June (9.5 0.2) than September
367
Table 1. The results of repeated measures ANOVA for N and P concentrations and the N:P ratio of loblolly pine foliage under ambient and elevated CO
2
(‘CO
2
’), at
three heights within the canopy (‘position’), across needle age classes (‘age class’), in June and September (‘season’) during the first 4 years of this experiment (‘year’).
Fstatistics with superscript symbols indicate statistical significance at * ¼P<0.05, ** ¼P<0.01, *** ¼P<0.001, **** ¼P<0.0001.
Source of variation df N concentration P concentration N:P ratio
MS FMS FMS F
Between subjects
CO
2
1 25.49 11.83** 0.007 0.09 38.37 2.26
Position 2 10.78 5.00** 0.258 3.27* 15.29 0.90
CO
2
position 2 0.49 0.33 0.004 0.06 0.23 0.01
Age class 1 265.69 123.33**** 11.749 148.74**** 241.26 14.21***
CO
2
age class 1 0.10 0.05 0.027 0.34 11.83 0.70
Age class position 2 10.21 4.74* 0.004 0.05 9.40 0.55
CO
2
age class position 2 0.15 0.07 0.018 0.22 0.36 0.02
Season 1 163.31 75.81**** 7.977 100.99**** 116.97 6.89*
CO
2
season 1 3.00 1.39 0.014 0.18 0.23 0.01
Season position 2 3.97 1.84 0.042 0.53 0.21 0.01
CO
2
season position 2 1.36 0.63 0.001 0.01 0.88 0.05
Age class season 1 0.11 0.05 3.597 45.54**** 157.24 9.26**
CO
2
age class season 1 1.04 0.49 0.002 0.03 2.98 0.18
Age class season position 2 1.02 0.47 0.003 0.03 0.62 0.04
CO
2
age class season position 2 0.05 0.02 0.005 0.06 0.04 0.01
Error 48 2.16 0.079 16.97
Within subjects
Year 3 26.41 76.27**** 0.394 67.86**** 31.30 85.79****
YearCO
2
3 0.64 1.85 0.001 0.11 0.56 1.53
Yearposition 6 0.15 0.42 0.035 6.03**** 1.18 3.24**
YearCO
2
position 6 0.06 0.16 0.001 0.24 0.17 0.48
368
year age class 3 5.11 14.76**** 0.158 27.27**** 2.37 6.48**
YearCO
2
age class 3 0.08 0.22 0.001 0.23 0.47 1.29
Yearage class position 6 0.40 1.15 0.011 1.86 0.13 0.35
YearCO
2
age class position 6 0.13 0.37 0.006 0.95 0.46 1.27
Yearseason 3 0.31 0.91 0.059 10.16**** 2.41 6.61***
YearCO
2
season 3 0.32 0.93 0.002 0.34 0.50 1.37
Yearseason position 6 0.24 0.68 0.019 3.22** 0.49 1.33
YearCO
2
season position 6 0.25 0.71 0.001 0.13 0.28 0.76
Yearage class season 3 1.74 5.01 0.083 14.26**** 2.35 6.44***
YearCO
2
age class season 3 0.08 0.23 0.001 0.14 0.18 0.50
Yearage class season position 6 0.16 0.45 0.009 1.61 0.35 0.95
YearCO
2
age class season position 6 0.15 0.42 0.002 0.32 0.23 0.64
Error 144 0.35 0.006 0.36
369
(10.7 0.2, Table 1). There was a significant age class xseason interaction (Table
1). Foliar N:P ratios were highest in 1-year-old needles in June (11.1 0.3) and
lowest in current year foliage in June (7.8 0.1). There was significant inter-annual
variation in foliar N:P ratios and significant variations between years and canopy
position and age class (Table 1). However, there were no significant year xCO
2
or
higher order interactions (Table 1).
Canopy biomass and N and P content
The mass of foliage in the loblolly pine canopy was significantly higher under
elevated CO
2
throughout the first 4 years of fumigation (Figure 3). The mass of N in
the loblolly pine canopy was significantly higher under elevated CO
2
in the second
and third years of fumigation but not during the first and fourth years of fumigation.
Figure 1. The average (1 S.E.) concentration of N and P at three different heights in the canopy under
ambient (open symbols) and elevated CO
2
(filled symbols). The September 1996 data represent pre-
treatment differences in the N and P concentration of loblolly pine needles.
370
Figure 2. The concentration of N and P in current year and 1-year old needles at three different heights
within the canopy. Bars of the same color with different superscript letters are significantly different from
one another at P<0.01.
Table 2. Foliar N:P ratios (weight:weight) by needle age class under
ambient and elevated CO
2
during the first 4 years of CO
2
fumigation.
Year N:P ratio
1-Year-old needles Current-year needles
Ambient Elevated Ambient Elevated
1997 12.6 (0.6) 11.1 (0.7) 9.7 (0.5) 9.3 (0.5)
1998 11.8 (0.6) 10.7 (0.6) 9.8 (0.4) 9.8 (0.6)
1999 10.6 (0.6) 9.8 (0.5) 8.6 (0.4) 8.1 (0.4)
2000 11.3 (0.6) 10.0 (0.6) 9.3 (0.3) 8.9 (0.4)
371
The mass of P in the loblolly pine canopy was higher under elevated CO
2
throughout the experiment, but only significantly higher following the fourth year
of CO
2
fumigation (Figure 3).
Figure 3. The mass, N content and P content of the loblolly pine canopy under ambient (open symbols)
and elevated CO
2
(filled symbols). 1996 was the pre-treatment year. Significant differences between CO
2
treatments within a given year are denoted by: * ¼P<0.05, ** ¼P<0.01.
372
Discussion
The mass and nutrient content of forest canopies is highly correlated with NPP
(Gholz 1982; Vose and Allen 1988; Piatek and Allen 2000). If nutrient limitations
constrain plant acquisition of essential resources under elevated CO
2
, reductions in
the mass or nutrient content of forest canopies may constrain C assimilation from
the atmosphere. As a result, the mass and the nutrient content of forest canopies are
critical parameters for the forecasting of the productivity response of forests to
rising concentrations of atmospheric CO
2
.
Elevated concentrations of atmospheric CO
2
reduced needle N concentrations
but had no effect on the concentration of P during the first 4 years of this experi-
ment (Figure 1). At the same time, however, there was a significant increase in the
production of foliar biomass (Figure 3). This resulted in a significant increase in the
quantity of N and P in the loblolly pine canopy under elevated CO
2
(Figure 3). The
production of loblolly pine foliage in this ecosystem sequesters *53% of the total
annual nutrient requirement (Finzi et al. 2002). During the first 4 years of this study,
the quantity of N in the loblolly pine canopy increased by 3.7 gm
2
under elevated
CO
2
whereas it increased by 2.9 g m
2
under ambient CO
2
. Similarly, the quantity
of P in the loblolly pine canopy increased by 0.34 g m
2
under elevated CO
2
but
only 0.22 g m
2
under ambient CO
2
. This represents a 26 and 50% increase in the
quantity of N and P, respectively, in the loblolly pine canopy under elevated CO
2
.A
recently completed N budget for this ecosystems shows that increases in N-use
efficiency (NUE) and increases in N uptake from soils maintain the productivity
response of this ecosystem under elevated CO
2
(Finzi et al. 2002). We do not have a
complete budget for P in this ecosystem, however the same basic mechanisms are
likely to be in operation for P.
Four lines of evidence suggest that primary production is more strongly N
limited than P limited in this ecosystem. First, foliar N:P ratios suggest N limita-
tion. Foliar N:P ratios have been successfully used to demonstrate N versus. P
limitation of temperate grassland and heathland productivity (Verhoeven et al.
1996). Similarly, Valentine and Allen (1990) summarized diameter growth and
foliar nutrient responses to factorial additions of N and P fertilizers for loblolly pine
forests throughout the southeastern US. Based on (i) the initial concentration of N
and P in current year needles in the top third of the loblolly pine canopy, and (ii) the
increase in tree diameter growth following N and P fertilization, their results
suggest that N:P ratios <10.5 indicate N limitation, N:P ratios between 10.5 and
12.5 indicate joint limitation by N and P, and N:P ratios >12.5 indicate P limitation.
N:P ratios for current year needles averaged across the canopy during the first 4
years of this study were 9.8 suggesting that N limits forest NPP (Table 2). A
similar pattern was observed for the current year needles in the top third of the
canopy in this study (data not shown). Second, Finzi et al. (2002) reported a highly
significant positive relationship between NPP and the annual rate of net N miner-
alization at this site under ambient and elevated CO
2
in 1998, the second year of
CO
2
fumigation. There was no correlation between NPP and bicarbonate-ex-
tractable P pools in mineral soils at this site for the same time period (A. Finzi and
373
A. Gallardo, unpublished data), although base extractions of acid soils can under
recover plant-available pools of P. Third, foliar N concentrations were significantly
lower under elevated CO
2
(Table 1) and they showed a strong early-growing season
decline in the top third of the canopy whereas no such patterns were observed for P
(Figure 1). The reduction and seasonal lag in foliar N concentrations suggests that
demand for N early in the growing season exceeded the rate of N uptake from soils
and/or reallocation from storage pools to the canopy. These patterns do not appear
to be due to a dilution of N by carbohydrate buildup in leaves; if it were, the
dilution effect would have also manifested itself in foliar P concentrations but no
such effect was observed (Figure 1). Fourth, although N:P ratios were not statis-
tically significantly lower under elevated CO
2
, they were consistently lower than
those under ambient CO
2
(Table 2). This reduction in N:P ratio suggests that rapid
forest growth under elevated CO
2
drives the systems towards greater N limitation
than P limitation.
One major assumption in scaling nutrient concentrations from individual leaves
at selected heights in a forest canopy to an estimate of the nutrient content of the
entire canopy is the distribution of foliage biomass throughout the forest canopy. In
our initial analysis we assumed that 25% of the foliar biomass was in the lower
third of the canopy, 50% of the foliar biomass was in the middle third of the canopy
and the remaining 25% of the foliar biomass was in the top third of the canopy. We
then calculated the mass of N and P in the forest canopy by multiplying the N and P
concentration of needles at each height in the canopy by the mass of foliage at these
heights. We performed a sensitivity analysis of this assumption by evenly redis-
tributing foliar biomass across the canopy heights (i.e., 33, 34, and 33% of the foliar
biomass in the bottom, middle and top of the canopy, respectively). There was no
evidence that an even distribution of foliar biomass altered our estimate of the
quantity of N (Figure 4(A)) or P (data not shown) in the plant canopy. The in-
sensitivity of total N and P content in the canopy reflects the inverse pattern in the
N concentration of different age-class needles throughout the forest canopy and the
very subtle variation in P concentrations (Figure 2).
A second major assumption in our scaling is that the morphology of needles (i.e.,
LMA) is not significantly different between abscised needles and green needles.
Theoretically the LMA of green needles should be greater than that of abscised
needles because the retranslocation of carbohydrates and nutrients from senescing
tissues decreases leaf mass (assuming no shrinkage of needles area). Interestingly,
our data show that the LMA of abscised needles was slightly higher than that of
green needles in 1997 (þ5%), 1999 (þ6%) and 2000 (þ7%) whereas it was lower
in 1998 (5%). We propagated the difference between green needle LMA and
abscised needle LMA through our calculation of the content of canopy N (Figure
4(B)). This propagation shows that the absolute quantity of N estimated for the
canopy is highly sensitive to our assumption of no change in LMA. However,
repeated measures ANOVA shows that the difference in LMA between green and
abscised needles is not statistically significant under ambient or elevated CO
2
(analysis not shown). This ANOVA suggests that there is little reason to use the
difference in LMA between green and abscised needles when calculating canopy
374
nutrient content. Moreover, the relative difference in LMA between green and
abscised leaves is smaller than the relative difference in foliar biomass production
under ambient and elevated CO
2
. Therefore, while the absolute content of the N in
the canopy would change with a different assumption for LMA, this difference
would not fundamentally alter our conclusion regarding the effect of elevated CO
2
on canopy nutrient content.
Biogeochemical models predict a progressive decline in forest production under
elevated CO
2
in nutrient-limited ecosystems (Rastetter et al. 1997; Luo and Rey-
nolds 1999). This occurs because rapid plant growth under elevated CO
2
im-
mobilizes nutrients in long-lived plant tissues and soils more rapidly than their
mineralization from soil pools or weathering from soil minerals. Chronosequence
studies show that aggrading forests actively assimilate N from mineral soil horizons
and that this N is stored in woody biomass and O horizon pools (Gholz et al. 1985;
Magill et al. 2000; Richter et al. 2000; Hooker and Compton 2003). If the majority
of the N is stored in woody biomass, then the reduction in the size of labile N pools
in mineral soils could result in a rapid attainment of severe N limitation under
elevated CO
2
(cf. Richter et al. 2000). Conversely, if the majority of the N is
returned to the O horizon of soils in litterfall, then the rate of N release from the
decomposition of the O horizon relative the rate at which labile N is removed from
mineral soil will dictate the severity of nutrient limitation under elevated CO
2
.In
particular, if the O horizon decomposes rapidly, then the reduction in N uptake as a
Figure 4. A sensitivity analysis of (A) the distribution of foliage biomass throughout the canopy, and
(B) the effect of leaf-mass to area ratios for the content of N in the canopy. In (A) the X-axis is the
predicted canopy N content given our original assumption that 25, 50, and 25% of the foliage biomass is
distributed in the bottom, middle and top of the canopy, respectively. The Y-axis is the predicted canopy
N content assuming that foliage biomass is evenly distributed across the plant canopy. There are 10
points total reflecting the predicted canopy biomass from 1996 through 2000 in the ambient (open
symbols) and elevated CO
2
(filled symbols) plots. In (B) the X-axis is the predicted canopy N content
assuming that there is no difference in the LMA of green needles and abscised needles. The Y-axis is the
predicted canopy N content assuming that the mass of foliage per unit area differs between green and
abscised needles (see text for details). There are 8 points total reflecting the predicted canopy biomass
from 1997 through 2000 in the ambient (open symbols) and elevated CO
2
(filled symbols) plots. In both
plots the solid line is the 1:1 line.
375
result of the mining of labile soil N from mineral soil horizons by plants could be
offset by the uptake of N from the O horizon.
Despite the clear limitation of NPP by N at this site (Finzi et al. 2002; Oren et al.
2001), this limitation has not precluded a response to elevated CO
2
during the first 4
years of this experiment (Finzi et al. 2002). Similarly, there is no evidence that the
mass of N in the canopy is declining through the first 4 years of CO
2
fumigation
(Figure 2). As a consequence, whole-canopy C assimilation is strongly stimulated
by elevated CO
2
making this forest a larger net C sink under elevated CO
2
than
under ambient CO
2
(DeLucia et al. 2002; Hamilton et al. 2002). However, there is
little evidence that the rate of N supply for plant growth has increased under
elevated CO
2
and N is being immobilized in woody biomass and the O horizon
more rapidly under elevated than ambient CO
2
(Finzi and Schlesinger 2003). Given
that NPP in this ecosystem is N limited (Finzi et al. 2002) and that foliar N:P ratios
are consistently lower under elevated CO
2
and declining in a direction suggesting
greater N limitation (Table 2), we hypothesize that limited soil N availability will
eventually curtail the initial increase in productivity under elevated CO
2
(see Oren
et al. 2001). The time scale for such a down-regulation, however, is not known. One
of the first indications of such a response, however, should be a convergence in the
pool of N in the loblolly pine canopy under ambient and elevated CO
2
.
Acknowledgements
George Hendrey, John Nagy and Keith Lewin were instrumental in construction and
maintenance of the FACE facilities. We would also like to thank Heather Hemric,
Jeffrey Pippen, Anthony Mace, Shawna Naidu, Damon Bradbury, Ariana Sutton,
Meredith Zaccherio, and Bridgid Curry for their extremely valuable field and lab
assistance. This study was supported by the US Department of Energy, with an-
cillary support from the National Science Foundation (DEB 98-15350). Additional
support for A. Finzi was provided by an appointment as an Alexander Hollaender
Distinguished Postdoctoral Fellow, sponsored by the US Department of Energy,
Office of Biological and Environmental Research, and administered by the Oak
Ridge Institute for Science and Education.
References
Crews T., Kitayama K. et al. 1995. Changes in soil phosphorus fractions and ecosystem dynamics across
a long chronosequence in Hawaii. Ecology 76: 1407–1424.
Cross A. and Schlesinger W. 1995. A literature review and evaluation of the Hedley fractionation:
applications to the biogeochemical cycle of soil phosphorus in natural ecosystems. Geoderma 64:
197–214.
DeLucia E.H., Hamilton J.G. et al. 1999. Net primary production of a forest ecosystem with experi-
mental CO
2
enrichment. Science 284: 1177–1179.
DeLucia E.H., George K. and Hamilton J.G. 2002. Radition-use efficiency of a forest exposed to elevated
concentrations of atmospheric carbon dioxide. Tree Physiol. 22: 1003–1010.
376
Fenn M., Poth M. et al. 1996. Evidence for nitrogen saturation in the San Bernadino mountains in
southern California. Forest Ecol. Manage. 82: 211–230.
Finzi A.C. and Schlesinger W.H. 2003. Soil-N cycling in a pine forest exposed to five years of elevated
CO
2
. Ecosystems. 6: 444–456.
Finzi A.C., Allen A.S. et al. 2001. Forest litter production, chemistry and decomposition following two
years of free-air CO
2
enrichment. Ecology 82: 470–484.
Finzi A., DeLucia E. et al. 2002. The nitrogen budget of a pine forest under free-air CO
2
enrichment.
Oecologia 132: 567–578.
Gholz H.L. 1982. Environmental limits on aboveground net primary production, leaf area, and biomass
in vegetation zones of the Pacific Northwest. Ecology 63: 469–481.
Gholz H.L., Fisher R.F. and Pritchett W.L. 1985. Nutrient dynamics in slash pine plantation ecosystems.
Ecology 66: 647–659.
Hamilton J.G., DeLucia E.H., George K., Naidu S.L., Finzi A.C. and Schlesinger W.H. 2002. Forest
carbon balance under elevated CO
2
. Oecologia 131: 250–260.
Ha
¨ttenschwiler S. and Ko
¨rner C. 1998. Biomass allocation and canopy development in spruce model
ecosystems under elevated CO
2
and increased N deposition. Oecologia 113: 104–114.
Hendrey G.R., Ellsworth D.S. et al. 1999. A free-air enrichment system for exposing tall forest vege-
tation to elevated atmospheric CO
2
. Global Change Biol. 5: 293–309.
Herbert D. and Fownes J. 1995. Phosphorus limitation of forest leaf area and net primary production in a
highly weathered soil. Biogeochemistry 29: 223–235.
Hooker T.D. and Compton J.E. 2003. Forest ecosystem carbon and nitrogen accumulation during the first
century after agricultural abandonment. Ecological Appl. 13: 299–313.
Jach M., Laureysens I. et al. 2000. Above- and below-ground production of young scots pine (Pinus
sylvestris L.) trees after three years of growth in the field under elevated CO
2
. Ann. Botany 85:
789–798.
Johnson D.W., Ball J.T. and Walker R.F. 1997. Effects of elevated CO
2
and nitrogen fertilization on
vegetation and soil nutrient content in juvenile ponderosa pine. Plant Soil 190: 29–40.
Lowther J.R. 1980. Use of a single sulfuric acid–hydrogen peroxide digest for the analysis of Pinus
radiata needles. Communications. in Soil Sci. Plant Anal. 11: 175–188.
Luo Y. and Reynolds J. 1999. Validity of extrapolating field CO
2
experiments to predict carbon se-
questration in natural ecosystems. Ecology 80: 1568–1583.
Magill A.H., Aber J.D., Berntson G.M., McDowell W.H., Nadelhoffer K.J., Melillo J.M. and Steudler P.
2000. Long-term nitrogen additions and nitrogen saturation in two temperate forests. Ecosystems 3:
238–253.
Marks P.L. and Bormann F.H. 1972. Revegetation following forest cutting: mechanisms for returning to
steady-state nutrient cycling. Science 176: 914–915.
McMurtrie R.E. and Comins H.N. 1996. The temporal response of forest ecosystems to doubled at-
mospheric CO
2
concentration. Global Change Biol. 2: 49–57.
Murray M., Leith L. et al. 1996. The effect of long-term CO
2
enrichment on the growth, biomass
partitioning and mineral nutrition of Sitka spruce (Picea sitchensis (Bong.) Carr.). Trees 10:
393–402.
Norby R., Wullschleger S. et al. 1999. Tree responses to rising CO
2
in field experiments: implications for
the future forest. Plant Cell Environ. 22: 683–714.
Oren R., Ellsworth D.S. et al. 2001. Soil fertility limits carbon sequestration by forest ecosystems in a
CO
2
-enriched atmosphere. Nature 411: 469–472.
Pearcy R., Bjorkman O. et al. 1987. Carbon gain by plants in natural environments. Bioscience 37:
21029.
Peterson A.G., Ball J.T. et al. 1999. The photosynthesis–leaf nitrogen relationship at ambient and
elevated atmospheric carbon dioxide: a meta-analysis. Global Change Biol. 5: 331–346.
Piatek K. and Allen H. 2000. Site preparation effects on foliar N and P use, retranslocation, and transfer
to litter in 15-year old Pinus taeda. Forest Ecol. Manage. 129: 143–152.
Rastetter E.B., A
˚gren G.I. et al. 1997. Responses of N-limited ecosystems to increased CO
2
: a balanced-
nutrition, coupled-element-cycles model. Ecol. Appl. 7: 444–460.
377
Rey A. and Jarvis P. 1997. Growth response of young birch trees (Betula pendula Roth.) after four and a
half years of CO
2
exposure. Ann. Botany 80: 809–816.
Richter D. and Markewitz D. 2001. Understanding Soil Change: Soil Sustainability Over Millenia,
Centuries, and Decades. Cambridge University Press, New York, pp.
Richter D., Markewitz D. et al. 2000. Legacies of agriculture and forest regrowth in the nitrogen of old-
field soils. Forest Ecol. Manage. 138: 233–248.
Schlesinger W.H. 1997. Biogeochemistry: An Analysis of Global Change. Academic Press, San Diego,
CA, pp. 588.
Schlesinger W.H. and Lichter J. 2001. Limited carbon storage in soil and litter of experimental forest
plots under increased atmospheric CO
2
. Nature 411: 466–469.
Switzer G. and Nelson L. 1972. Nutrient accumulation and cycling in loblolly pine (Pinus taeda L.)
plantation ecosystems: the first twenty years. Soil Sci. Soc. Am. J. 36: 143–147.
Valentine D. and Allen H. 1990. Foliar responses to fertilization identify nutrient limitation in loblolly
pine. Can. J. Forest Res. 20: 144–151.
Verhoeven J., Koerselman W. et al. 1996. Nitrogen- and phosphorus limited growth in herbaceous, wet
vegetation: relations with atmospheric inputs and management regimes. Trends in Ecol. Evol. 11:
494–497.
Vitousek P. and Farrington H. 1997. Nutrient limitation and ecosystem development: experimental test of
a biogeochemical theory. Biogeochemistry 37: 63–75.
Vitousek and Howarth 1991. Nitrogen limitation on land and in the sea: how can it occur? Bio-
geochemistry 13: 87–115.
Vose J.M. and Allen H.L. 1988. Leaf area, stemwood growth, and nutrition relationships in loblolly pine.
Forest Sci. 34: 547–563.
Walker T. and Syers J. 1976. The fate of phosphorus during pedogensis. Geoderma 15: 1–19.
Walker R.F., Johnson D.W. et al. 2000. Growth, nutrition and water relations of ponderosa pine in a field
soil as influenced by long-term exposure to elevated atmospheric CO
2
. Forest Ecol. Manage. 137:
1–11.
Webb R., Lauenroth W. et al. 1983. Primary production and abiotic controls in forests, grasslands, and
desert ecosystems in the United States. Ecology 64: 134–151.
Zak D.R., Pregitzer K.S. et al. 2000. Atmospheric CO
2
, soil N availability, and allocation of biomass and
nitrogen by Popolus tremuloides. Ecol. Appl. 10: 34–46.
Zhang S. and Allen H.L. 1996. Foliar nutrient dynamics of 11-year-old loblolly pine (Pinus taeda)
following nitrogen fertilization. Can. J. Forest Res. 26: 1426–1439.
378