International Association for Ecology
Growth and Senescence in Plant Communities Exposed to Elevated CO? Concentrations on an
Estuarine Marsh
Author(s): P. S. Curtis, B. G. Drake, P. W. Leadley, W. J. Arp, D. F. Whigham
Reviewed work(s):
Source: Oecologia, Vol. 78, No. 1 (1989), pp. 20-26
Published by: Springer in cooperation with International Association for Ecology
Stable URL: http://www.jstor.org/stable/4218826 .
Accessed: 04/03/2012 10:38
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Springer and International Association for Ecology are collaborating with JSTOR to digitize, preserve and
extend access to Oecologia.
http://www.jstor.org
Oecologia (1989) 78:20-26 Oecogia 
? Springer-Verlag 1989 
Growth and senescence in plant communities 
exposed to elevated CO2 concentrations on an estuarine marsh 
P.S. Curtis*, B.G. Drake, P.W. Leadley, W.J. Arp, and D.F. Whigham 
Smithsonian Environmental Research Center, Box 28, Edgewater, MD 21037, USA 
Summary. Three high marsh communities on the Chesa- 
peake Bay were exposed to a doubling in ambient CO2 
concentration for one growing season. Open-top chambers 
were used to raise CO2 concentrations ca. 340 ppm above 
ambient over monospecific communities of Scirpus olneyi 
(C3) and Spartina patens (C4), and a mixed community 
of S. olneyi, S. patens, and Distichlis spicata (C4). Plant 
growth and senescence were monitored by serial, nondes- 
tructive censuses. Elevated CO2 resulted in increased shoot 
densities and delayed senescence in the C3 species. This 
resulted in an increase in primary productivity in S. olneyi 
growing in both the pure and mixed communities. There 
was no effect of CO2 on growth in the C4 species. These 
results demonstrate that elevated atmospheric CO2 can 
cause increased aboveground production in a mature, un- 
managed ecosystem. 
Key words: Elevated CO2 - Productivity - Salt marsh - 
Scirpus olneyi - Spartina patens 
The steady rise in atmospheric carbon dioxide concentra- 
tion has prompted considerable research concerning the lik- 
ely consequences of this anthropogenic change on plant 
growth (reviewed in Strain and Cure 1985). Most of this 
work has been conducted with agricultural species under 
laboratory or controlled field conditions. Despite our im- 
proved understanding of the physiology of the CO2 re- 
sponse, it has been difficult to extrapolate from this work 
to unmanaged plant communities. The great diversity in 
growth responses among annual species to elevated CO2 
(Carlson and Bazzaz 1980; Kimball 1983), the paucity of 
long term research, and the important influence of environ- 
mental stress in the CO2 response (Patterson and Flint 
1982; Bowman and Strain 1987) all make very uncertain 
any predictions concerning the response of a specific ecosys- 
tem to this global climate change. 
Results from studies of agricultural species and, to a 
lesser degree, wild species have led to several general hy- 
potheses regarding ecological responses to elevated CO2. 
Plants with the C3 pathway of photosynthesis usually in- 
crease carbon assimilation and growth in response to in- 
creases in CO2 concentration (Ford and Thorne 1967; 
Rogers et al. 1983; Downton et al. 1987) whereas C4 plants 
are more variable and generally respond less than C3 plants 
(Carlson and Bazzaz 1980; Potvin and Strain 1985; Smith 
et al. 1987). In communities containing C3 species, net pri- 
mary productivity should therefore increase, and C3 species 
may gain a competitive advantage over C4 species (Carter 
and Peterson 1983; Zangerl and Bazzaz 1984). Both C3 
and C4 plants show an increase in water use efficiency under 
elevated CO2 (Morison 1985). This could have a significant 
effect on water availability in arid and mesic environments 
(Wigley and Jones 1985). Low nutrient availability tends 
to decrease the relative response to C02, but the opposite 
is true for water stress. In environments where plant growth 
is strongly controlled by one of these limiting factors (e.g. 
coniferous forests, deserts), the magnitude of the response 
should vary accordingly (Oechel and Strain 1985). 
To date, only one study has involved an unmanaged 
plant community that was exposed to elevated CO2 in situ 
for an entire growing season (Oechel et al. 1984). In an 
arctic tussock sedge ecosystem, Oechel and co-workers 
found that canopy and single leaf photosynthesis increased 
substantially in the first year of exposure to a doubling 
of CO2 but that acclimation occurred and by the fourth 
year there was no detectible difference between elevated 
and control plots. There was no effect on net productivity 
although the sedge Eriophorum vaginatum showed an in- 
crease in tillering (Tissue and Oechel 1987). These results 
suggested that in the arctic, sustained community level re- 
sponses to increased atmospheric CO2 would not occur. 
Here we report results from the first year of exposing 
a temperate salt marsh ecosystem to a doubling of atmo- 
spheric CO2 concentration. Three high marsh communities 
containing monospecific populations of C3 and C4 species, 
and these same species in combination were studied. The 
co-occurrence of C3 and C4 dominants and high system 
productivity make salt marshes ideal environments in which 
to test current theories of ecosystem responses to CO2 . Salt 
marshes also accrete large amounts of carbon annually 
(Haines and Dunn 1985) and may thus be important sinks 
for atmospheric CO2. 
Materials and methods 
Description of the study site 
The study site is located at 38053'N, 76033'W in the Rhode 
River, a subestuary of the Chesapeake Bay. It is typical 
of brackish high marshes in the Mid-Atlantic region of 
North America (Whigham et al. 1983). The marsh is infre- 
* Current address and address for offprint requests: School of Natu- 
ral Resources, The Ohio State University, Columbus, OH 43210, 
USA 
SCIRPUS 
c oo 
lom SPARTINA 10 m , ,, : ,bi 
Fig. 1. Map of the study site showing the Scirpus, Spartina and 
Mixed communities, the field laboratory (L), boardwalk (B), and 
permanent experimental plots (open circles). The treatments were 
Elevated chamber (E), Ambient chamber (A) and unchambered 
Control (C) 
quently flooded (Jordan et al. 1983) and is a mosaic of 
plant associations that are primarily dominated by Spartina 
patens (Ait.) Muhl., Scirpus olneyi Grey, Distichlis spicata 
(L.) Greene, Typha angustifolia L. or Iva fructescens L. 
Other common species are Spartina cynosuroides (L.) Roth, 
Scirpus robustus Pursh, Hibiscus moscheutos L. and Pani- 
cum virgatum L. Three communities on the marsh were 
selected for this study. One (Spartina) was dominated by 
the C4 grass S. patens, one (Scirpus) by the C3 sedge S. ol- 
neyi, and one (Mixed) by S. patens, S. olneyi, and D. spi- 
cata, also a C4 grass. 
Fifteen permanent circular plots 0.8 m in diameter were 
established along transect lines in each community (Fig. 1). 
A 20 cm deep cut was made into the substrate around the 
perimeter of each plot, severing all living rhizomes. Plastic 
garden edging was inserted 10 cm into this cut. Treatments 
within each community (described below) were assigned to 
plots according to a randomized block design. A prelimi- 
nary survey of all plots was conducted in late June of 1986 
prior to the start of the CO2 treatment. There were no 
significant differences in shoot densities among plots as- 
signed to the three treatments in each community. 
CO2 treatment 
Open top chambers were used to elevate CO2 within a plot. 
The chambers, and the CO2 control and monitoring system 
have been described in detail previously (Drake et al. 1987). 
The chambers were 1.2 m in height and 0.8 m in diameter, 
were covered with 300 ptm polyester film, and were sealed 
to the marsh surface by taping them to the plastic garden 
edging. Ambient air was introduced into the chambers by 
a high capacity blower and circulated with a second blower. 
In plots exposed to elevated C02, 100% CO2 was contin- 
uously injected into the input blower where it was thor- 
oughly mixed with ambient air before entering the chamber. 
CO2 levels were measured by an infra-red gas analyser (Bin- 
os 4B.2, Leybold-Heraeus, Hanau FRG) connected to an 
automatic gas sampling system. Light and temperature were 
monitored both inside and outside chambers at canopy 
height. 
Within each community, five plots were maintained with 
elevated CO2 concentrations (Elevated treatment), and five 
with chambers but exposed to ambient CO2 concentrations 
(Ambient treatment). Five plots in each community had 
no chambers but were otherwise treated identically to cham- 
bered plots (Control treatment). CO2 concentrations inside 
elevated chambers were allowed to vary diurnally in parallel 
with ambient variations in CO2 concentrations. Chambers 
were placed on the marsh and treatments begun on April 
23, 1987 and all chambers were removed from the marsh 
on November 15, 1987. 
Daily mean CO2 concentrations (sunrise to sunset) were 
350 + 22 (s.d.) ,ul l-' inside Ambient chambers and 
686 + 30 jl 1' inside Elevated chambers. Twenty four hour 
mean temperatures were 1.7 + 0.60 C higher inside Ambient 
chambers and 2.0 + 0.4? C higher inside Elevated chambers 
than temperatures outside chambers. Light intensity was 
reduced about 10% inside chambers but light quality was 
not affected (Drake et al. 1987). 
Vegetation sampling 
Plant growth in each plot was followed by serial, nondes- 
tructive censuses of shoot number, shoot weight and above- 
ground biomass. Sampling methods were designed to mini- 
mize destructive changes to the plant canopy while provid- 
ing sufficient material and demographic information to de- 
scribe treatment responses. Approximately five days were 
required to census one community. Net primary productivi- 
ty (NPP) was calculated using the method of Smalley (1959) 
for Spartina and Distichlis, and cummulative mortality for 
Scirpus (Hopkinson et al. 1980). All other measures of 
aboveground biomass, shoot numbers and shoot weight are 
for green tissue only. 
Scirpus. Aboveground biomass of Scirpus consists solely 
of erect photosynthetic shoots. Scirpus was censused in each 
plot by measuring each shoot to the nearest 1 cm. Regres- 
sion equations relating shoot height to shoot biomass were 
calculated from destructive harvests of shoots outside of 
the experimental plots. Aboveground biomass per plot was 
calculated as the sum of estimated individual shoot dry 
weights. Separate regressions were calculated for the Scir- 
pus and Mixed communities at each census. All harvested 
shoots were dried at 600 C and weighed. 
Three to five shoots were also harvested from within 
each plot at each census, measured, and compared to the 
confidence limits of the regression equations. This compari- 
son showed that the allometric relationship between shoot 
length and dry weight was not affected by treatment so 
single equations were sufficient to estimate shoot dry 
weights for all plots in a community. Shoots harvested with- 
in plots were also used for calculating specific leaf weights 
(SLW= g/cm2). Leaf area, i.e. green shoot area, was esti- 
mated by measuring the base width, apex width, and height 
of one rhomboidal face of each shoot. 
Spartina and Distichlis. Because of the high density of Spar- 
tina and Distichlis shoots, shoot number, biomass, and leaf 
area were estimated by subsampling each plot. Each plot 
in the Spartina and Mixed communities was divided into 
permanent 100 cm2 quadrats using monofilament nylon 
line. Five quadrats per plot were randomly selected for sub- 
sampling at the beginning of the season. Combined, these 
21 
22 
five quadrats represented 10% of the total plot area. All 
shoots were counted within each quadrat at each census. 
Shoot density per plot was estimated by extrapolation 
from the mean density in the 5 quadrats. Shoot biomass 
and leaf area were estimated from limited destructive har- 
vests in each plot at each census. All living shoots within 
three 25 cm' areas located 2 cm from quadrats in each plot 
were harvested. Typically, 25-40 stems were collected per 
plot per census. Senescent material was measured separately 
from green tissue and no area within a plot was harvested 
more than once during the season. Leaf area was measured 
with an electronic leaf area meter. Mean dry weight per 
shoot was multiplied by shoot density to estimate above- 
ground biomass per plot. 
At peak standing biomass (late August), the area sub- 
sampled within each plot was expanded to 10 quadrats 
(20% of the plot area) and 80-100 shoots harvested. Esti- 
mates of shoot density and dry weight were compared using 
both the original and expanded methods. There were no 
significant differences between methods for within treat- 
ment estimates of growth (mean of five plots, t-test). 
Plant growth analysis 
The relative increases in aboveground biomass (Biomass 
RGR), shoot number (Shoot Density RGR) and shoot dry 
weight (Shoot Weight RGR) were calculated after the meth- 
ods of Hunt (1982). Cubic polynomials were fit to the ln 
transformed data (1) from each census for each plot by 
least squares regression. First derivatives were evaluated 
at the date of census. 
RGR = d(ln Y)/dx = 1/ Y dy/dx. 
Derivatives were not evaluated at the ends of the fitted 
curves (first and last censuses). 
Statistical analysis 
Treatment means within a census were analysed by analysis 
of variance (Anova) based on five replicates per treatment 
arranged in a randomized block design. Variance estimates 
for aboveground biomass, shoot density, and shoot weight 
were based on among plot variance only. Pairwise compari- 
son of means was by least significant difference (a priori 
comparisons: Elevated vs Ambient, Ambient vs Control) 
or minimum significant difference (a posteriori compari- 
sons) (Sokal and Rohlf 1981). Percentages were arc-sin 
transformed before analysis by ANOVA. 
Relative growth rates were compared using Friedman's 
method for randomized blocks (Sokal and Rohlf 1981). 
This nonparametric test uses the ranking of variates within 
blocks and therefore does not require the estimation of 
variance components. For significant treatment effects to 
be inferred, the ranking of variates must be identical within 
all five blocks. 
Results 
Shoot density 
Shoots density of Scirpus was higher in plots with elevated 
CO2 in both Scirpus and Mixed communities (Fig. 2A, B). 
In both cases the effects of CO2 first became significant 
at peak density in August and extended through the end 
1000 A SCIRPUS 
800 - 
600 v 
200 
C 1000 B MIXED-ScirDus 
E 
800 
0 
C/) 600 
-C SPARTIN A 
6200 - I 
o~~~~~~~~~~~~- 
o 00 
.~ 
cl~~~~~~~~~~~~~~~~~~~E 
2000 
M J J A S 0 N 
Month 
Fig. 2A-C. The change in shoot density in Scirpus (A), Mixed- 
Scirpus (B), and Spartina (C) plots. Treatments were Elevated (-), 
Ambient (o), and Control (o). Vertical bars are the LSD (P <0.05) 
and are included where significant differences occur (A and B) 
or at the second and fourth censuses to indicate variability (C) 
of the season. There was also a significant difference be- 
tween shoot densities of Scirpus from Ambient and Control 
plots in the Scirpus community (Fig. 2A). This chamber 
effect was not, however, found in the Mixed community 
(Fig. 2B). 
The relative rate of change in shoot density (Shoot Den- 
sity RGR) was consistently higher in Scirpus community 
Elevated plots than Ambient plots but this difference was 
only significant in July, immediately preceeding peak densi- 
ties (Fig. 3A). In the Mixed community, the effect of CO2 
on Scirpus Shoot Density RGR was seen later in the season, 
with significant differences between Elevated and Ambient 
plots in August and September (Fig. 3 B). These results indi- 
cate both a greater relative allocation of carbon into new 
shoots and a slower senescence of existing shoots under 
elevated CO2. 
Shoot densities showed a much more gradual increase 
over time in the Spartina community (Fig. 2 C). Shoot emer- 
gence occurred slightly earlier than in Scirpus, with a large 
number of shoots appearing in mid to late April. There 
were no significant differences in shoot densities or Shoot 
Density RGR (data not shown) among Elevated, Ambient, 
or Control plots at any time. 
Shoot weight 
CO2 had no effect on mean shoot weight in the Scirpus 
community (Fig. 4A). Shoots of Scirpus in the Mixed com- 
SCIRPUS MIXED-ScirDus 
7 A B 
0.02 O\ X 
>s0.00 - \ 
o~~~~~~~~~~ 
0 -0.02 , 
0 
C D 
_> 0.04 
*0~~~~~~~~~~~~~~~~~~~ 
0.0 
m - 
-0.04 r- . 8 0 
>2 
~~~ 0.04~~~~~~ 
(8 0~~~~~~~~~~~~~~~~~ 
-0.04 
J J A S 0 N J J A S 0 N 
Month 
Fig. 3 A-F. Relative change in Shoot Density, Shoot Weight, and 
Aboveground Biomass from Scirpus (A, C, E) and Mixed-Scirpus 
(B, D, F) plots exposed to Elevated (e) or Ambient (o) CO2 treat- 
ments. Asterixes denote a significant difference (P<0.05) between 
RGR means within a census 
munity were less than 50% of the size of shoots in the 
Scirpus community and there was a significant increase in 
shoot weight due to CO2 beginning in late August and 
extending through the end of the season (Fig. 4B). There 
was a significant effect of CO2 on Shoot Weight RGR in 
the Scirpus and Mixed communities in late August and 
September (Fig. 3 C, D). This response was particularly evi- 
dent in the Mixed community where shoot weight declined 
very little through November. There was a significant 
chamber effect on shoot weight in the Scirpus community 
in September and October and in the Mixed community 
in late October (Fig. 4A, B). There were no CO2 effects 
on shoot weight in Spartina (Fig. 4C). There were also no 
effects of CO2 or chamber on SLW from any of the study 
species (Table 1). 
Aboveground biomass 
Aboveground live biomass in the Scirpus community in- 
creased rapidly between shoot emergence in late-April and 
the end of July, reaching a maximum of between 600 and 
900 g/m2 in early August (Fig. SA). Biomass was signifi- 
cantly higher in Elevated plots in September and October. 
Peak standing biomass in Scirpus from the Mixed commun- 
ity was less than 20% of that from the Scirpus community 
and there was also a significant response to elevated CO2 
(Fig. SB). As with shoot density there was a significant 
A SCIRPUS 
1200 - 
800 
400 - 
_ - B NIIXED-Scirpus 
S 1200 
V) 0 
bO E 800- 
150 C SPARTINA I 
BC ~ ~ ~ ~ 0 1 00 --~-.-- 
50 - 
0 
1;, 
50A 0 N 
M J J MotA S 0 N 
Month 
Fig. 4A-C. The change in shoot weight with time in Scirpus (A), 
Mixed-Scirpus (B), and Spartina (C) plots. Treatments were Ele- 
vated (-), Ambient (o), and Control (o). Vertical bars are the 
LSD (P < 0.05) and are included where significant differences occur 
(A and B) or at the second and fourth censuses to indicate variabili- 
ty (C) 
Table 1. Specific leaf weights at peak standing biomass from Ele- 
vated, Ambient, and Control plots in three marsh communities. 
Mean + (s.e.) 
Community Elevated Ambient Control 
glcm2 
Scirpus 0.0274 (0.0016) 0.0260 (0.0008) 0.0274 (0.0004) 
Mixed-Scirpus 0.0288 (0.0013) 0.0268 (0.0013) 0.0251 (0.0019) 
Spartina 0.0233 (0.0036) 0.0198 (0.0003) 0.0217 (0.0006) 
Mixed-Spartina 0.0198 (0.0010) 0.0210 (0.0005) 0.0204 (0.0004) 
Mixed-Distichlis 0.0141 (0.0011) 0.0142 (0.0005) 0.0147 (0.0006) 
chamber effect on aboveground biomass only in the Scirpus 
community. 
Although elevated CO2 had no significant effect on 
aboveground biomass in the Scirpus community until Sep- 
tember, there were small but significant increases in Bio- 
mass RGR due to CO2 in both July and August (Fig. 3 E). 
Scirpus in the Mixed community showed similar, although 
non-significant, differences in Biomass RGR at these times 
and much greater differences during September and Oc- 
tober (Fig. 3 F). The CO2 effects on aboveground biomass 
were therefore due in part to an increase in the efficiency 
of new growth (principally through new shoot production) 
23 
24 
A SCIRPUS I 
1000 - 
800 - 
600 - - -E. 
400 - 
0 
0 
200 - 0~~~~~~~~~~~~ 
E 1000 - B MIXED-Scirpus 
800 
E 600 
.Lo 20e 
1000 C SPARTINA 
800 
600 I 
400 ~- - - 
oo~ I,- 
I, 
200 - 
M J i A S 0 N 
Month 
Fig. 5A-C. The change in aboveground biomass with time in Scir- 
pus (A), Mixed-Scirpus (B), and Spartina and Mixed-C4 (C) plots. 
Treatments were Elevated (.), Ambient (o), and Control (o). Verti- 
cal bars are the LSD (P<0.05) and are included where significant 
differences occur (A and B) or at the second and fourth censuses 
to indicate variablity (C) 
Table 2. Percentage of total biomass (live+ senescent) which was 
senescent at the final census in November 1987 in Elevated, Ambi- 
ent, and Control plots in three marsh communities. Mean + (s.e.) 
Community Elevated Ambient Control 
Scirpus 35.5 (4.6) a 45.7 (5.6)b 79.3 (6. 1) 
Mixed-Scirpus 37.8 (4.6)a 80.1 (2.4)b 68.7 (6.3)b 
Spartina 45.3 (4.1 )a 44.9 (6.0)' 53.1 (6.5)a 
Mixed-Spartina 51.8 (9.0)a 56.3 (6.6)a 69.6 (9.6)a 
Mixed-Distichlis 66.7 (9.3)a 64.3 (12.7)a 57.2 (11.3)a 
* similar superscript denotes no significant difference within a com- 
munity, P<0.05, except Scirpus Elevated vs Ambient where P< 
0.10 
and in part to a delay in the loss of dry weight through 
senescence. 
There were no treatment effects on aboveground bio- 
mass in Spartina (Fig. 5C). Shoot emergence began in mid 
April and peak biomass of about 500 g/m2 was reached 
in late August. Peak aboveground biomass in the C4 com- 
ponent of the Mixed community also showed no effect of 
CO2 and was very similar (479 + 27 g m- 2, pooled across 
treatments) to the Spartina community. Analysis of Dry 
Weight RGR also showed no treatment effects or consistent 
trends in either community (data not shown). 
Table 3. Net primary productivity from Elevated, Ambient, and 
Control plots in three marsh communities. Mean + (s.e.) 
Community Elevated Ambient Control 
g/m2 
Scirpus 539 (47)' 463 (44)b 345 (21)c 
Mixed-Scirpus 139 (25)a 78 (15)b 63 (11)b 
Spartina 645 (22)a 668 (61)a 650 (58)a 
Mixed-C4 732 (49)a 694 (47)' 660 (74)a 
* similar superscript denotes no significant difference within a com- 
munity, P<0.05 
The percentage of total Scirpus biomass present as dead 
tissue at the final census in November was significantly 
lower under elevated CO2 in both the Scirpus and Mixed 
communities (Table 2). Again, there was a significant 
chamber effect in the Scirpus but not the Mixed community. 
Senescence of the two C4 species appears to have progressed 
somewhat more rapidly in the Mixed than in the Spartina 
community but there was no effect of CO2 in either case. 
Elevated CO2 caused a significant increase in net prima- 
ry productivity (NPP) in Scirpus from both the Scirpus 
and Mixed communities (Table 3). Although peak live bio- 
mass in the Scirpus community was not significantly higher 
in Elevated plots, sustained growth later in the season led 
to greater NPP under elevated CO2. Senescent Scirpus 
shoots weighed less per cm than did living shoots which 
resulted in lower NPP than peak aboveground live biomass 
(Fig. 5 A, SB). Net primary productivity in the C4 species 
was greater than in Scirpus but was unaffected by elevated 
CO2 . 
Discussion 
The most pronounced effect of the doubling in ambient 
CO2 concentration on these salt marsh communities was 
an increase in shoot numbers (Fig. 2) and decrease in the 
rate of senescence in the C3 sedge, Scirpus olneyi (Fig. 3, 
Table 3). This resulted in a significant increase in live, 
aboveground biomass in the latter half of the season (Fig. 5) 
and greater net primary productivity (Table 4) in Scirpus 
from both the Scirpus and Mixed communities. These re- 
sults support the prediction that plant growth in mature, 
unmanaged ecosystems containing C3 species will increase 
in response to increasing atmospheric CO2 concentrations 
(Bazzaz et al. 1985). We found no growth response in the 
Spartina community or the C4 component of the Mixed 
community. 
Our estimates of net primary productivity were based 
solely on aboveground dry matter and therefore do not 
take into account the substantial amounts of carbon trans- 
located belowground in perennial marsh species (Good 
et al. 1982). An increase in carbon allocation to roots or 
belowground storage organs is a commonly observed re- 
sponse to elevated CO2 (Ford and Thorne 1967; Bhatta- 
charya et al. 1985). Since shoot number and size early in 
the season were largely a function of previously stored car- 
bon in Scirpus americanus (Giroux and Bedard 1987), a 
steadily increasing growth response in subsequent years 
with continuing exposure to elevated CO2 may be likely 
in Scirpus olneyi. This was not observed, however, in the 
arctic sedge Eriophorum vaginatum where photosynthetic 
acclimation to elevated CO2 occurred within a single season 
(Tissue and Oechel 1987). Productivity in the arctic tundra 
is low and growth is strongly nutrient limited (Shaver et al. 
1986). The highly productive brackish marsh may be more 
analogous to C3 agricultural ecosystems in which growth 
almost always increases in response to elevated CO2 (Kim- 
ball 1983). 
Scirpus shoots arise from axillary buds on the below- 
ground stem, or rhizome, and are morphologically analo- 
gous to tillers in grasses (Esau 1977). An increase in tillering 
in response to elevated CO2 has been observed previously 
in wheat (Gifford 1977; Sionit et al. 1981), the sedge Erio- 
phorum vaginatum (Tissue and Oechel 1987) and the C4 
grass Andropogon glomeratus (Bowman and Strain 1987). 
Increased tillering may be a general response to increasing 
carbon or nutrient supply in monocots producing axillary 
buds (Fletcher and Dale 1974). In clonal salt marsh grasses, 
the rate of rhizome and tiller growth is an important factor 
determining the outcome of competition for open space 
following disturbance (Bertness and Ellison 1987). The in- 
creased shoot growth by Scirpus in the Mixed community 
did not have any detectible negative effect on Spartina and 
Distichlis but the long term consequences of a sustained 
growth response by Scirpus in this community are difficult 
to predict. Regions of the marsh with vigorous Scirpus pop- 
ulations have very little Spartina or Distichlis present. Com- 
petition as well as edaphic conditions are probably impor- 
tant in determining local species abundances (Snow and 
Vince 1984). 
The slower rate of senescence and continued production 
of new shoots in Scirpus under elevated CO2 resulted in 
a greater number of green shoots present in September and 
October (Fig. 2), a slower relative rats of decline in above- 
ground biomass (Fig. 3), and a lower percentage senescent 
tissue present in November (Table 3). Previous studies of 
elevated CO2 effects on whole plant senescence have pro- 
duced conflicting results. Bhattacharya et al. (1985) found 
early leaf senescence in sweet potato grown at 675 ppm 
CO2 and St. Omer and Horvath (1983) reported early senes- 
cence in two California annuals at 2100 ppm but not 
700 ppm CO2 . Carter and Peterson (1983) observed delayed 
senescence in Sorghum at 600 ppm CO2. The mechanism 
by which CO2 might affect senescence is not clear. High 
levels (>2000 ppm) inhibit the action of ethylene, a senes- 
cence promoting hormone (Nooden 1980). Early senescence 
under elevated CO2 may be correlated with the timing of 
other phenological events such as flowering (St. Omer and 
Horvath 1983) or tuber maturation (Bhattacharya et al. 
1985). 
The chambers had a significant effect on growth in the 
Scirpus community although there was no effect on Scirpus 
from the Mixed community or on the C4 species (Figs. 2, 
5). The 2? C temperature increase, protection of shoots 
from mechanical damage, and possibly higher humidity in- 
side chambers could have contributed to the observed ef- 
fects on growth. In a review of the literature on plant 
growth in open top chambers, Drake et al. (1985) found 
no consistent pattern, with both positive and negative ef- 
fects reported. They concluded that open top chambers 
were the best available technology for field exposure of 
plants to elevated CO2. 
This is the first demonstration that elevated atmospheric 
CO2 can lead to an increase in growth and productivity 
in an unmanaged ecosystem within a single year. The poten- 
tial impact of this response on net carbon storage will de- 
pend both on the degree to which this response is sustained 
and the effects of elevated CO2 on ecosystem carbon loss. 
With no change or a decline in decomposition rate, an in- 
crease in primary productivity could result in greater rates 
of carbon accretion. This suggests that the direct effects 
of elevated CO2 on terrestrial vegetation could be an impor- 
tant consideration in the global carbon budget. 
Acknowledgements. We would like to acknowledge the field assis- 
tance of Suzanne Hill, Sarah Chamberlain and Lisa Balduman. 
John Craig and Jim Johnson provided valuable technical help. 
This research was jointly supported by the Smithsonian Institution 
and the United States Department of Energy. 
References 
Bazzaz FA, Garbutt K, Williams WE (1985) Effect of increased 
atmospheric carbon dioxide concentration on plant communi- 
ties. In: Strain BR, Cure JD (eds) Direct effects of increasing 
carbon dioxide on vegetation. United States Department of 
Energy, Carbon Dioxide Research Division, DOE/ER-0238, 
Office of Energy Research, Washington, DC, pp 155-170 
Bertness MD, Ellison AM (1987) Determinants of pattern in a 
New England salt marsh community. Ecol Monogr 57:129-147 
Bhattacharya NC, Biwas PK, Bhattacharya S, Sionit N, Strain 
BR (1985) Growth and yield response of sweet potato to atmo- 
spheric CO2 environment. Crop Sci 25:975-981 
Bowman WD, Strain BR (1987) Interaction between CO2 enrich- 
ment and salinity stress in the C4 non-halophyte Andropogon 
glomeratus (Walter) BSP. Plant Cell Environ 10:267-270 
Carlson RW, Bazzaz FA (1980) The effects of elevated carbon 
dioxide concentrations on growth, photosynthesis, transpira- 
tion, and water use efficiency of plants. In: Singh JJ, Deepak 
A (eds) Environmental and climatic impact of coal utilization. 
Academic Press, New York, pp 609-622 
Carter DR, Peterson KM (1983) Effects of a C02-enriched atmo- 
sphere on the growth and competitive interactions of a C3 and 
a C4 grass. Oecologia (Berlin) 58:188-193 
Downton WJS, Grant WJR, Loveys BR (1987) Carbon dioxide 
enrichment increases yield of valencia orange. Aust J Plant 
Physiol 14:493-501 
Drake BG, Rogers HH, Allen LH Jr (1985) Methods of exposing 
plants to elevated carbon dioxide. In: Strain BR, Cure JD (eds) 
Direct effects of increasing carbon dioxide on vegetation. Unit- 
ed States Department of Energy, Carbon Dioxide Research 
Division, DOE/ER-0238, Office of Energy Research, Washing- 
ton, DC, pp 11-32 
Drake BG, Arp W, Craig J, Curtis PS, Leadley PW, Whigham 
D (1987) Effects of elevated CO2 on Chesapeake Bay wetlands. 
II. Gas exchange and microenvironment in open-top chambers. 
United States Department of Energy, Carbon Dioxide Research 
Division Report Number 038, Office of Energy Research, 
Washington, DC 
Esau K (1977) Anatomy of seed plants. J Wiley and Sons, New 
York 
Fletcher GM, Dale JE (1974) Growth and tiller buds in barley: 
effects of shade treatment and mineral nutrition. Ann Bot 
38:63-76 
Ford MA, Thorne GN (1967) Effect of carbon dioxide concentra- 
tion on growth of sugar-beet, barley, kale, and maize. Ann 
Bot 31:629-644 
Gifford RM (1977) Growth pattern, carbon dioxide exchange and 
dry weight distribution of wheat growing under differing photo- 
synthetic environments. Aust J Plant Physiol 4:99-110 
Giroux J-F, Bedard J (1987) Effects of simulated feeding by snow 
geese on Scirpus americanus rhizomes. Oecologia (Berlin) 
74:137-143 
Good RE, Good NF, Frasco BR (1982) A review of primary pro- 
25 
26 
duction and decomposition dynamics of the belowground 
marsh component. Est Comp 1982:139-157 
Haines BL, Dunn EL (1985) Coastal marshes. In: Chabot BF, 
Mooney HA (eds) Physiological ecology of North American 
plant communities. Chapman and Hall, NY, pp 323-347 
Hopkinson CS, Gosselink JG, Parrondo RT (1980) Production 
of coastal Louisiana marsh plants calculated from phenometric 
techniques. Ecology 61:1091-1098 
Hunt R (1982) Plant growth curves: An introduction to the func- 
tional approach to plant growth analysis. Edward Arnold, Lon- 
don 
Jordan TE, Correll DL, Whigham DF (1983) Nutrient flux in the 
Rhode River: tidal exchange of nutrients by brackish marshes. 
Est Coast Shelf Sci 17:651-667 
Kimball BA (1983) Carbon dioxide and agricultural yield: an as- 
semblage and analysis of 430 prior observations. Agr J 
75:779-788 
Morison JIL (1985) Sensitivity of stomata and water use efficiency 
to high CO2 . Plant Cell Environ 8:467-474 
Nooden LD (1980) Senescence in the whole plant. In: Thimann 
KV (ed) Senescence in plants. CRC Press, Boca Raton, pp 219- 
258 
Oechel WC, Strain BR (1985) Native species responses to increased 
carbon dioxide concentration. In: Strain BR, Cure JD (eds) 
Direct effects of increasing carbon dioxide on vegetation. Unit- 
ed States Department of Energy, Carbon Dioxide Research 
Division, DOE/ER-0238, Office on Energy Research, Washing- 
ton, DC, pp 117-154 
Oechel WC, Hastings S, Hilbert D, Lawrence W, Prudhomme T, 
Reichers G, Tissue D (1984) The response of arctic ecosystems 
to elevated CO2 regimes. United States Department of Energy, 
Carbon Dioxide Research Division, Report Number 019, Office 
of Energy Research, Washington 
Patterson DT, Flint EP (1982) Interacting effects of CO2 and nu- 
trient concentration. Weed Sci 30:389-394 
Potvin C, Strain BR (1985) Photosynthetic response to growth 
temperature and CO2 enrichment in two species of C4 grasses. 
Can J Bot 63:483-487 
Rogers HH, Thomas JF, Bingham GE (1983) Responses of agron- 
omic and forest species to elevated atmospheric carbon dioxide. 
Science 220:428-429 
Sionit N, Mortensen DA, Strain BR, Hellmers H (1981) Growth 
responses of wheat to carbon dioxide enrichment with different 
levels of mineral nutrition. Agr J 73:1023-1027 
Shaver GR, Chapin FS III, Gartner BC (1986) Factors limiting 
growth and biomass accumulation in Eriophorum vaginatum 
L in Alaskan tussock tundra. J Ecol 74:257-278 
Smalley AE (1959) The role of two invertebrate populations, Lit- 
torina irrorat and Orchelinmum fidicinum in the energy flow 
of a salt marsh ecosystem. PhD thesis, Univ Georgia, Athens 
Smith SD, Strain BR, Sharkey TD (1987) Effects of CO2 enrich- 
ment on four Great Basin grasses. Funct Ecol 1:139-143 
Snow AA, Vince SW (1984) Plant zonation in an Alaskan salt 
marsh. II. An experimental study of the role of edaphic condi- 
tions. J Ecol 72:669-684 
Sokal RR, Rohlf FJ (1981) Biometry. WH Freeman and Co, San 
Francisco 
St Omer L, Horvath SM (1983) Elevated carbon dioxide concentra- 
tions and whole plant senescence. Ecology 64:1311-1314 
Strain BR, Cure JD (1985) (eds) Direct effects of increasing carbon 
dioxide on vegetation. United States Department of Energy, 
Carbon Dioxide Research Division, DOE/ER-0238, Office of 
Energy Research, Washington DC 
Tissue DT, Oechel WC (1987) Response of Eriophorum vaginatum 
to elevated CO2 and temperature in the Alaskan tussuck tun- 
dra. Ecology 68:401-410 
Whigham DF, McCormick J, Good RE, Simpson RL (1978) Bio- 
mass and primary production in freshwater tidal wetlands of 
the middle Atlantic coast. In: Good RE, Whigham DF, Simp- 
son RL (eds) Freshwater wetlands. Ecological processes and 
management potential. Academic Press, New York, pp 3-21 
Wigley TML, Jones PD (1985) Influences of precipitation changes 
and direct CO2 effects on streamflow. Nature 314:149-152 
Zangerl AR, Bazzaz FA (1984) The response of plants to elevated 
CO2. II. Competitive interactions among annual plants under 
varying light and nutrients. Oecologia 62:412-417 
Received February 22, 1988