in open-top chambers during the entire growing season from 1989 through 199 1. Dominant .... Agricultural Research Service's Air Quality Field Lab- oratory at ...
BIOMASS PRODUCTION IN A TALLGRASS PRAIRIE ECOSYSTEM EXPOSED TO AMBIENT AND ELEVATED COz’ CL E N T O N E. OWENSBY Department qf Agronomy, Kansas State University, Manhattan. Kansas 66506
C’S,4
PATRICK I. COY-NE Fort Hays Branch Experiment Station, Kansas State University, Manhattan, Kansas 66506 C’S4 J AY M . HAM AND LI S A M . A UEN Department ofilgronornv, Kansas State Clniversitv, Manhattun, Kansas 66506
CT.%4
A LAN K. KNAPP Division qf Biology. Kansas State C’niversity. Manhattan, Kansas 66506 US’,4 Abstract. Responses to elevated CO2 have not been measured for natural grassland ecosystems. Global carbon budgets will likely be affected by changes in biomass production and allocation in the major terrestrial ecosystems. Whether ecosystems sequester or release excess carbon to the atmosphere will partly determine the extent and rate that atmospheric CO, concentration rises. Elevated CO, also may change plant community species composition and water status. We determined above- and belowground biomass production, plant community species composition. and measured and modeled water status ofa tallgrass prairie ecosystem in Kansas exposed to ambient and twice-ambient CO2 concentrations in open-top chambers during the entire growing season from 1989 through 199 1. Dominant species were Andropogon gerardii, .4. scoparius, and Sorghastrurn ntltans (C, metabolism) and Pou pratensis (C,). Aboveground biomass and leaf area were estimated by periodic sampling throughout the growing season in 1989 and 1990. In 199 I, peak biomass and leaf area were estimated by an early August harvest. Relative root production among treatments was estimated using root ingrowth bags which remained in place throughout the growing season. Latent heat tlux was simulated with and without water stress. Botanical composition was estimated annually. Compared to ambient COY levels, elevated CO, increased production ofC, grass species. but not of C, grass species. Species composition of C, grasses did not change. but Poa pratrnsis (C,) declined. and C, forbs increased in the stand with elevated CO, compared to ambient. Open-top chambers appeared to reduce latent heat flux and increase wateruse efficiency similar to the elevated CO, treatment when water stress was not severe, but under severe water stress. the chamber effect on water-use efficiency was limited. In natural ecosystems with periodic moisture stress. increased water-use efficiency under elevated CO, apparently would have a greater impact on productivity irrespective of photosynthetic pathway. Kc1, word Pea pratensis: IN
Karl ws.
TRODUCTION
Global atmospheric carbon dioxide concentration is increasing at an unprecedented rate (Boden et al. 1990). and the consequences of that increase are subject to much speculation (Dahlman 1984). Because rangelands occupy 470/o of the earth’s land area (Williams et al. 1968) and 54%) of the conterminous IJnited States (USDA 1974). responses of rangelands to this environmental change are potentially significant to the global carbon budget. Even though organism-level responses to CO, enrichment have been documented. rangcland
’ Manuscript received 70 M a r c h 1992: 1-c”ised 1992: accepted 6 November 1992: final bcrsion November 1992.
5 Octobci received 18
ecosystem-level responses remain unresolved (Strain and Cure 1985. Barr.at 1990). Mooney et al. (199 1) pointed out the urgent need for additional research on terrestrial ecosystem response to elevated CO,. They concluded that thcsc experiments must be at least a decade in length to allow a response trajectory to be established. Responses to CC& enrichment at the single plant level with respect to productivity have been dependent on photosynthetic pathway. Carbon fixation rates in plants with the C, pathway generally show a greater response to increasing CO levels than rates in c‘, plants (Kimball 1983. Wray and Strain 1986. Nijs et al. 1988. Reichers and Strain 1988). Enhanced wateruse efficiency associated with partial stomata1 closure in (X),-enriched environments results in enhanced productivity for both c‘, a n d CVJ p l a n t s (Tolley a n d
November 1993
TALLGRASS PRAIRIE AND ELEVATED CO1
Strain 1984) with greater water savings occurring in C, plants. Even though numerous controlled-environment experiments have demonstrated plant responses to CO2 enrichment, according to Drake and Leadley (199 1) only two studies of natural-ecosystems reactions to CO, enrichment have been published. Curtis et al. (1989) reported on growth and senescence in C?, C,, and mixed C, and C, saltmarsh communities. They concluded that communities dominated by Scirpus olneyi (C,) had greater productivity, and that senescence was delayed with CO? enrichment compared to ambient COZ. Production was not increased in Spartina patens (C,) communities with CO> enrichment. Oechel and Strain (1985) reported that a tussock sedge ecosystem in the arctic tundra initially responded to CO? enrichment with increased productivity, but the increase disappeared by the fourth year. They reasoned that an acclimation response had occurred in the photosynthetic mechanism. Productivity in most temperate terrestrial ecosystems is usually limited by moisture and/or nutrient availability. In natural systems, essentially all nutrient resources are supplied by the system through nutrient cycling. Thus, changes in water-use efficiency or more efficient use of other nutrients would enhance ecosystem productivity. Curtis et al. (1989) reported increased productivity for Scirpus communities under elevated CO2 with the same nutrient resources as communities with ambient CO,, thereby indicating an increased nutrient-use efficiency. Most research on natural systems with elevated CO? has placed more emphasis on photosynthetic pathway (C, vs. C,) than on water relations. In natural systems that are exposed frequently to water stress, the major response is likely to be increased productivity from increased water-use efficiency. Eamus (199 1) indicated that substantially more data are required to predict ecosystem response to COz-induced changes in water-use efficiency. Numerous authors have hypothesized that because of the differential responses of C, and C, plants to CO1 enrichment, competitive relationships will be altered (Bazzaz 1990). Wray and Strain (1986) concluded that Aster pilosus (aster, C,) may outcompete Andropogon virginicus (broomsedge, C,) in natural systems, but competition may be mitigated by differential rooting patterns and canopy structure. Weaver and Albertson (1943) revealed that, in natural prairies, feeding roots of C, forbs and C, grasses seldom occupied the same soil layer. Also, canopy layering in prairies with an overstory of forbs may provide improved growing conditions for grasses (Launchbaugh and Owensby 1978). Thus, shifts in botanical composition among species of differing photosynthetic pathways in prairies may or may not occur. The objectives of this study were to assess the effect of ambient and elevated (2 x ambient) atmospheric CO, concentrations on biomass production, leaf area,
645
and species composition of a tallgrass prairie ecosystem and to develop a model to simulate latent heat flux for the CO, treatments. MATERIALS
AND
METHODS
Study site The experimental site was located in pristine tallgrass prairie north of Manhattan, Kansas (39.12” N, 96.35” W), 324 m above mean sea level. Vegetation on the site was a mixture of C, and C, species, dominated by big bluestem (Andropogon gerardii Vitman) and indiangrass (Sorghastrum nutans (L.) Nash). Subdominants included Kentucky bluegrass (Poa pratensis L.), sideoats grama (Bouteloua curtipendula (Michx.) Torr.), and tall dropseed (Sporobolus asper var. asper (Michx.) Kunth). Members of the sedge family made up 5-l 0% of the composition. Principal forbs included ironweed (Vernonia baldwinii var. interior (Small) Schub.), western ragweed (Ambrosiapsilostachya DC.), Louisiana sagewort (Artemisia ludoviciana Nutt.), and manyflower scurfpea (Psoralea tenuiflora var. jloribunda (Nutt.) Rydb.). Average peak total aboveground live biomass (dry mass) of 425 g/m* occurs in early August, ofwhich 35 g/m> is from forbs (Owensby and Anderson 1967). Soils in the area are transitional from Ustolls to Udolls (Tully series: fine, mixed, mesic, montmorillonitic, Pachic Argiustolls). Slope on the area is 5%. Fire has been infrequent, occurring 2-3 times in 10 yr. Past history has included primarily winter grazing by cow-calf pairs. The 30-yr average annual precipitation is 840 mm, with 520 mm occurring during the growing season. Nine circular plots were established in early May 1989. Treatments, replicated three times, were: ambient COZ, with no chamber (A); ambient CO,, with chamber (CA); and elevated (2 x ambient) CO,, with chamber (CE). The aluminum structural framework of the opentop chambers was a scaled-up version of a design described by Heagle et al. (1979) as modified by the USDA Agricultural Research Service’s Air Quality Field Laboratory at Raleigh, North Carolina. The chambers were 4.5 m in diameter by 3.25 m in height, with a conetop baffle that reduced the top opening to 3 m. The baffle added 0.75 m to the height ofthe open-top chamber for a total height of 4 m. The structural framework was covered by 1.5 mm thickness (6 mil), UV-resistant, polyethylene film. The cone-top baffle placed atop each chamber reduced the opening by 54%, thereby restricting the precipitation that entered the chamber. Within 24 h following each rainfall event, water equal to 54% of the rainfall amount for an unchambered plot was added using a rotating sprinkler adjusted to cover the diameter of the chamber. Aluminum edging was placed around the upslope bottom edge of the chamber to prevent runoff from entering the chamber. No edging was placed on the lower half of the chamber. One half of each chamber was used to estimate biomass pro-
Ecological
CLENTON E. OWENSBY ET AL.
646
Applications
Vol. 3, No. 4 TABLE 1. Monthly precipitation and average daily maximum temperatures and deviation from normal (current year minus 30-yr average) for the study site. 1989 Month Jan Feh
Ppt.
Dev.
(mm)
(mm) -21
Temp. (“C)
Dev. (“C)
(mm)
Ppt.
1991
1990 Temp. (mm) (“0
Dev. (“Cl
Ppt.
Dev.
(mm)
(mm)
Temp. (“Cl
Dev. (“0
Dev.
Apr May Jun jul Aug Sep .
0 0 6 8 40 74 30 141 175
-24 -46 -62 -74 -60 -71 60 72
9.8 0.2 13.1 22.5 25.9 28.7 33.5 32 25
6.1 -6.8 0.6 2.6 0.7 ~ 1.4 0.3 -0.5 ~2.6
27 22 105 23 100 124 79 180 20
6 -2 52 -48 -14 -10 -22 100 ~83
11.3 10.4 15.5 19.3 22.8 32.1 33.2 31.7 30.7
8.2 3.5 3 -0.6 -2.4 1.9 0.1 -0.8 3.1
34 1 35 107 130 51 47 56 44
13 -23 -18 36 16 ~83 -54 -25 -59
2.4 14.3 18.2 21.1 26.9 31.8 35.5 33.7 28.9
-0.7 7.4 5.7 1.1 1.7 1.7 2.3 1.3 1.3
Ott Nov Dee
1430 2
-37 70 -21
20.9 14.4 1.7
- 01.8 .9 -4.5
27 52 26
- 414 6 3
23.1 17.9 5.4
5.3 1.4 -0.8
83 33 48
- 446 0 25
23.6 8.7 8.1
298 1.9
Total
619
-214
785
-50
669
-166
Mar
duction, and the remaining half was grazed by esophageally fistulated sheep to determine forage quality differences among treatments. Data on forage quality are not reported here. CO, treatment Initially, pure CO, was added to the CE treatments through rotameters, which were manually adjusted using needle valves to maintain twice-ambient CO2 concentration inside a chamber. Wind velocity and air temperature both affected the flow rate required to maintain the desired CO, concentration, requiring many manual corrections during a 24-h day. During 1990 and 199 1 computer-controlled mass flow controllers were used to regulate CO, flow rate automatically based on real-time measurement of CO2 in a chamber. Fumigation of the enriched chambers began on 19 May 1989 and on 1 April in both 1990 and 1991. Carbon dioxide enrichment and environmental data acquisition were continuous until late October in 1989, 1990, and 199 1. The polyethylene film covering the chambers was removed in late October and replaced in late March of each year. All treatments were sustained on the same plots over the 3-yr period. Chamber
environment
During high photosynthetic activity, ambient CO, concentrations with and without chambers measured at 1 m were 330-340 mL/L, but during nighttime hours, CO2 levels reached ~400 mL/L. Photosyntheticallyactive-radiation (PAR) sensors placed at a height of 1 m inside a chamber and in an unchambered plot indicated that an 11 O/o reduction in PAR occurred inside the chamber. Soil temperature was measured at a 10 cm depth, and air temperature at 30 cm, 100 cm, and 300 cm above the soil surface. No difference (P < 10) occurred in soil temperature (- 10 cm) between chambered and unchambered plots. Likewise, the temperatures at 30 cm were similar in all plots. At 100 cm the temperature inside chambers was significantly
higher for the period from 1000 to 1500 central standard time (CST). The maximum difference was slightly more than 2°C on the hottest days. Again on the hottest days, the temperature at 300 cm was higher from 0800 to 1500 CST. The maximum temperature at 300 cm was some 5°C higher in chambered than in unchambered plots. The air delivery system for chambered plots provided for temperatures at plant canopy height that were equal to ambient conditions. The dew-point temperature averaged 1°C higher in chambered plots than in unchambered plots from 1200 to 1500 CST (P < . 10). Even though that difference was slight, higher humidities inside the chambers may have reduced evapotranspiration and thereby indirectly affected soil moisture levels. Soil moisture levels in chambered plots were significantly higher than those in unchambered plots from mid-June to late August each year (P < . 10). CE plots had higher soil moisture levels than CA plots under drought conditions during the sampling period (P < . 10). Chamber design, data acquisition and control, and data on environmental conditions are reported in detail in Owensby et al. (1989, 1990). Meteorological
conditions
Conditions at the experimental site in 1989 were extremely dry (Table 1). Precipitation was much below normal from January until late August, but temperatures averaged near normal during the growing season. In 1990, precipitation was slightly below normal except in August, and temperatures averaged slightly above normal during the growing season. Precipitation in 199 1 was much below normal during June through October, and temperatures were above normal. Air velocity and turbulent intensity Heat balance anemometers (HBA) of the Simmonstype design (Kanemasu and Tanner 1968) were constructed in our laboratory and calibrated in a wind tunnel. The anemometers were checked for agreement by comparing sensor output while all the instruments
November I993
TALLGRASS
PRAIRIE
were positioned at the same height above an open rangeland site. Results showed that agreement among the sensors was approximately ? 0.1 m/s over the range of wind speeds observed. Horizontal air velocities within a single, open-top chamber (OTC) were continuously measured at three horizontal locations using six HBAs. Two anemometers were positioned at each location to account for the directional sensitivity of the sensors. One pair of sensors was positioned near the center of the chamber, and a second pair was located 0.5 m from the outside wall. The position of the third pair of anemometers bisected the position of the center and outside locations. Two additional anemometers were installed in an unchambered ambient-CO, plot. All anemometers were positioned 0.4 m above the soil surface, which was 5-10 cm above canopy height. Sensors were sampled every 2 s using a datalogger (Model 21X, Campbell Scientific, Inc., Logan, Utah), and data were reduced to 1 0-min averages and standard deviations (SD). Turbulent intensity was computed every 10 min as the ratio between the sample SD and mean air speed (i.e., coefficient of variation for the 1 0-min air speed mean). Data were collected continually for a 3-wk period starting on 7 May 1991 (day 127 of the year). The height of the sensors was adjusted every week to keep them near the top of the canopy. Chamber and CO, efect modeling Once environmental conditions inside the chamber had been quantified, a numerical model was developed to simulate the energy balance of a single leaf under chambered and ambient environmental conditions. The model was also designed to simulate the effect of elevated CO? on stomata1 conductance, transpiration, and leaf water status. The purpose of the model was to provide a diagnostic tool for evaluating the climatological impact of the chamber on the energy and water economy of the chambered surface. Modeling the leaf energy balance required the development of a mechanistic heat balance equation for the leaf that accurately represented the biophysical interaction between rangeland-type leaves and their environment. A numerical approach was then used to find the unique surface temperature of the leaf that satisfied the energy budget of the surface. This approach to energy balance simulation of leaves is widely used in plant-environment simulations (e.g., Bristow 1987, Kitano and Eguchi 1989). The energy balance of a single leaf can be represented as R, + LE + H = 0,
(1)
where R,, is net radiation, LE is latent heat flux, and H is sensible heat flux, all with units of watts per square metre. This equation assumes that energy storage and metabolic energy generation in the leaf are negligible. The net radiation of the leaf was computed as the sum of absorbed shortwave and net longwave radiation,
AND
ELEVATED CO2
647
R,, = ~u,R,(l + p,) + t,(~uT, CA > A. 1990. -Ange biomass and leaf area averaged over all clipping dates were the same for CE and CA plots, but greater than that from A plots (Fig. 3). Popr, forb, and total biomass and leaf area averaged over all clipping dates were similar for all treatments. Ange peak biomass from CE and CA plots was similar, but greater than that from A plots (Fig. 2). Peak biomass of Popr was greater in CE plots than in CA and A plots. Peak total biomass in CE and CA plots was greater than that in A plots. 1991. -Biomass and leaf area were sampled in early August only. Ange and total biomass were greater from CE plots than that from CA and A plots (Fig. 2). Popr peak biomass occurred in early June and was not estimated by the August sampling date. Root ingrowth biomass In 1990, root ingrowth biomass from the nylon bags in CE plots was essentially double that from bags in CA and A plots (Fig. 4). In 199 1, CE plots also had greater root ingrowth biomass than CA plots, which had greater biomass than A plots. Plant species composition Essentially no change occurred in the species composition and basal cover of the major warm-season dominants among the CO, treatments over the 3-yr period. However, Popr basal cover was significantly reduced during that period in CE and CA plots compared to A plots, which changed little in composition and cover of Popr during the study period (Fig. 5). Sporobolus asper, a C, species, had a greater reduction in basal cover and percentage of composition during the 3-yr period in A plots than in CE and CA plots. Forbs constituted a greater portion of the total composition and increased in basal cover over the 3-yr period in all treatments, but more in CE than in CA and A plots. Xylem pressure potential Midday xylem pressure potentials (XPP) indicated less moisture stress for Ange plants in CE plots compared to those in CA or A plots (Fig. 6). The seasonal
November
649
TALLGRASS PRAIRIE AND ELEVATED CO,
1993
a
Chamber Enriched -
5 -p -
.-
:
1.0
a !? Q 't;; 2
Ji 't;; 0.5
2 I 0 Ange
2
Popr
Forbs
Total
Species
0.0 Ange
Forbs
Popr
Total
Species FIG. 1. Mean biomass and leaf area index for indicated species and species groups averaged over nine growing-season sampling dates in 1989 for native tallgrass prairie exposed to twice-ambient and ambient CO, concentrations. Means within species or species groups with a common letter do not differ (P < 10). Ange = Andropogon gerardii, Popr = Poa pratensis.
Chamber
Chamber
sid
“L?%
“s!Y
FIG. 3. Mean biomass and leaf area index for indicated species and species groups averaged over nine growing-season sampling dates in 1990 for native tallgrass prairie exposed to twice-ambient and ambient CO1 concentrations. Means within species or species groups with a common letter do not differ (P c .lO).
midday Ange XPP (mean k 1 SE) was - 1.89 MPa + 0.03 for the CE plots, -2.18 MPa + 0.03 for the CA plots, and -2.38 MPa k 0.04 for the A plots. The diurnal XPP ofAnge was less negative in CE plots than in CA and A plots from 1200 to 1400 CST, and Ange XPP was less negative at 2000 in CE and CA plots than in A plots (Fig. 7). Predawn Ange XPP did not differ among treatments. Chamber and CO, efect modelling Horizontal air velocities within the chamber ranged from 0.4 m/s near the center of the chamber to 1.2 m/s near the outside wall. Areally averaged air speed within the chamber was steady at 0.75 m/s, whereas outside wind speeds typically fluctuated between 0.25 and 3.0 m/s. Chamber air velocities tended to be greater than
1989
1990
s-
1991
_ 3 0 0 ,
300
E I
-- 250 n 2 E 2
m
5 2
2oo
150 100
1989
1990
Year FIG. 2. Peak biomass for indicated years and species and species groups for native tallgrass prairie exposed to twiceambient and ambient CO? concentrations. Means with a common letter do not differ (P < . 10).
1990
Year
1991
FIG. 4. Root biomass in ingrowth bags to a 1 S-cm depth in tallgrass prairie exposed to twice-ambient and ambient CO, concentrations. Data are means of four bags per plot in 1990 and eight bags per plot in 199 1, replicated three times. Means with a common letter within year do not differ (P < .lO).
6.50
Ecological Applrcations Vol. 3, No. 4
CLENTON E. OWENSBY ET AL.
Simulations without water shortage indicated that both the chamber and high CO, had significant influences on leaf energy balance and transpiration (Fig. 8a). The environmental influences of the chamber alone decreased transpiration by 14%; augmentation of CO, to the chamber resulted in an additional 15% reduction in water use. Thus, the climatological effect of the chamber and the biological influence of elevated CO, caused a similar response in transpiration. These sim-
b 1
Popr
spos
Forbs
Species F IG. 5. Mean differences in botanical composition and
basal cover of indicated species from pre-treatment levels and after 3 yr with exposure to twice-ambient and ambient CO2 levels. Means with a common letter do not differ (P < .lO). Popr = Poa pratensis, Spas = Sporobolus asper.
ulations suggested that soil water depletion and plant stress would develop first in the A plot, then in the CA plot, and finally in the CE plot. These results are consistent with the diurnal patterns of leaf water potential at the site (Fig. 7). Simulation of CO, enrichment outside the chambers showed that stomata1 adjustment would reduce latent heat flux by 21% (not shown), which is greater than the 15% reduction in latent heat flux simulated inside the chamber. Results suggested that the antitranspirant action of elevated CO2 might be moderated or masked by the climatological effect of the chamber. Simulated leaf water stress increased leaf temperatures and decreased latent heat fluxes (Fig. 8b). During water stress, however, there was no significant difference in latent heat flux between the A and CA plots. Evaluation of the energy balances showed that, although net radiation was reduced inside the chamber, reduced air speeds caused leaf temperatures and, in
outside wind speeds at night, but were much lower than
outside wind speed during the day. For example, data from days 13 1 and 132 of the year showed that average wind speed outside the chamber was 1.4 m/s, almost twice that observed inside the chamber. Turbulent intensity inside the chamber was lower than that measured in the ambient unchambered plot. The average turbulent intensities for the chambered and unchambered plots were 0.07 and 0.20 m/s, respectively. Lower air speeds and reduced turbulence within the chambers indicated that convective transport coefficients for heat, water vapor, and CO, also would be lower for the chambered than for the unchambered plots. The energy balance model was used to simulate a leaf near the top of the plant canopy in both chambered and unchambered plot environments. The effects of
turn, transpiration to rise. Model results indicated that reduced irradiance and lower air speeds had offsetting effects inside the chamber during water shortage, and resulted in a latent heat flux similar to that in the Andropogon gerardii O-O r
-0.5-
2x
CO2
+
chamber
0-m
A m b i e n t CO2
W-----H
Ambient
+ chamber
C02,
no
chamber
m ‘G s ‘; a
high CO, and water stress were also evaluated. The
climatological impact of the chamber was simulated by reducing irradiance by 10% and increasing maximum air and dew-point temperatures by 2.0 and 1 .S”C, respectively. Air velocities in the chambered and unchambered plots were simulated at 0.7 and 1.5 m/s, respectively. The effect of CO2 enrichment was evaluated by increasing CO, from 350 to 700 mWL. For ambient CO, levels, simulated midday stomata1 conductances (single side) for the well-watered and water-
stressed simulations were 12.5 and 4.8 mm/s, respectively. For the high-CO, simulations, midday stomata1 conductances for the well-watered and water-stressed simulations were 8.2 and 2.7 mm/s, respectively.
*\ ~,
io T:
5
-3.5
150
I 170
190
I 210
‘, . ..--w’
/ 230
250
J
270
Day of Year Midday xylem pressure potentials for Andropogon gerardii exposed to ambient and twice-ambient CO, concentrations. Data for weekly estimates from 30 May to 25 September 199 1. Only the largest standard error for each treatment is plotted. F IG. 6.
November
1993
0.0 O-O
z -0.5-
%
2x
CO2 +
relations. No difference occurred in aboveground biomass accumulation between CE and CA plots, but both treatments resulted in substantially more biomass production than unchambered plots. Obviously, CO2 enrichment and chambers have similar effects. Further support for the CO, effect on productivity being primarily through better water relations came with the 199 1 season, when, as a result of low rainfall, response to CO1 treatment and chamber effect were similar to those in 1989. These results are contrary to those of Curtis et al. (1990) who worked with both C, and C, saltmarsh communities and indicated that only C,dominated communities had increased biomass with CO? enrichment. Most tallgrass prairie forbs have the C, photosynthetic pathway, and in the absence of severe compe-
chamber
0-O
Ambient
M-----W
A m b i e n t CO2
CO2
+ chamber , no chamber
,
29 July 1991
A
-3.01
651
TALLGRASS PRAIRIE AND ELEVATED CO,
I 0600
1
,B 1200
;w
, 1600
I 2000
600
Time of Day FIG. 7. Diurnal xylem pressure potentials (XPP) for Andropogon gerardii exposed to ambient and twice-ambient CO, concentrations. Data for 29 July 199 1 under water stress conditions Only the largest standard error for each treatment is plotted. XPP means within a sampling time with a common letter do not differ (P < .05)
700 2 600 2
500
5 E 4 0 0
unchambered plot. Hence, simulation showed that the water balance of the unchambered plot, and the chamber ambient treatment would be similar when drought occurred. These results are compatible with biomass data showing that chamber effects were minimal during low-rainfall years. Simulations of water stress showed that CO2 enrichment reduced daily latent heat flux by 25% (Fig 8b), a reaction that could offset or delay the response to drought more than the presence of the chamber would. DISCUSSION
AND
I
3
3 0 0
E s 2 0 0 Y 100
0000
tion was substantially greater in CE plots (chamber, with elevated [2 x ambient] CO>) than in CA (chamber, with ambient CO?) and A (ambient COz, no chamber) plots. Because A. gerardii and the majority of the
0000
1200
1600
2000
2400
600
(b)
700
CONCLUSIONS
Clearly, the differential response (C, vs. C,) of plant productivity to CO, enrichment in this northeast Kansas tallgrass prairie was related to changes in moisture relationships. During an extremely dry year (1989) (Table 1), Andropogon gerardii and total biomass produc-
0400
WF 2 -
600
500
5 ii 400
other species in the plant community are C, species, the increased biomass under elevated CO, was probably not due to greater carbon fixation as a result of
the higher internal CO1 (Kirkham et al. 199 l), though increased photosynthetic efficiency may have occurred as well. The increased biomass in CA plots likely was due to improved water relations associated with altered environmental conditions. The reaction to improved precipitation in 1990 supported the conclusion that the primary effect of CO, enrichment was improved water
Time of Day FIG. 8. Simulated latent heat flux for tallgrass prairie exposed to 350 mL/L of CO, with and without a chamber and 700 mL/L of CO, with a chamber, (a) without water shortage and (b) with water stress.
652
CLENTON E. OWENSBY ET AL.
tition for light and nutrients from the C, perennial grass dominants, would probably increase under CO, enrichment. Indeed, that was the case in this study. Roots of most perennial forbs in tallgrass prairie are essentially unbranched in the upper 60 cm of the soil profile, and extract water and nutrients below that level (Weaver and Albertson 1943) thereby reducing competition from the C, dominants. Contrary to expectations, Poa pratensis (Popr), a C, species, did not respond to CO, enrichment with increased aboveground productivity. The normal growth period for Popr begins in mid-March, with slow growth until early April. The warm-season perennial, tallgrass dominants begin growth in late April and grow rapidly, providing substantial competition for light and nutrients. Because mineralization of N does not occur at a rapid rate until the soil warms in late April, Popr growth is limited by N availability (Owensby et al. 1969). Therefore, the lack of response by Popr to CO, enrichment was probably initially due to nutrient limitation and subsequently to competition for light from the rapidly developing C, tallgrass canopy. The plant census data support that conclusion, in that Popr declined in basal cover and relative amount in the stand, particularly in chambered plots with their enhanced productivity. Under grazed conditions with normal precipitation, Popr normally increases in unburned plant communities, where the tallgrass canopy is reduced by herbivory (Anderson et al. 1970). Popr is relatively short compared to the C, dominants and, consequently, in ungrazed prairie would be shaded from early May throughout the growing season, and subject to severe competition for water as well as nutrients. Also, Popr normally is reduced in tallgrass prairie during drought periods (Weaver and Albertson 1943). During our study, precipitation was below normal in 2 of the 3 yr. Because the reductions in Popr were greater in chambered plots than in unchambered ones, the increased production in chambered plots, with their increased canopy cover, likely provided greater competition for light and nutrients. Wray and Strain (1987) indicated that competitive relationships between a C, perennial grass and a C, perennial forb were altered by time of planting. In our study, differences in season of growth, plant height, and canopy cover likely accounted for the differential responses in population dynamics among the grass and forb species. Therefore, it is difficult to distinguish differential CO2 effects from competitive effects. Based on the reactions of species in this study, we conclude that, in ungrazed tallgrass prairie ecosystems, plant species that were substantially shorter than the dominants would not respond to CO, enrichment regardless of photosynthetic pathway. However, those C, species that were not subject to intense competition from the C, perennial tallgrass dominants because of height and/or rooting depth would likely increase in the stand. Ifcontinuous grazing were applied, the short-
Ecological
Applications Vol. 3, No. 4
er C, grasses, such as Popr, would probably also respond to CO, enrichment. Even though no difference occurred in aboveground biomass accumulation between CE and CA plots in 1990, root ingrowth biomass in CO,-enriched plots was essentially double that of ambient-CO, plots whether chambered or not. However, in 1991, root ingrowth biomass in CE plots was only 17% greater than that of CA plots. Apparently in years with nearnormal precipitation and elevated CO,, root growth is proportionally greater than in drought years. The explanation for increased root growth in 1990 compared to 199 1 likely centers around the timing of precipitation. Root growth for C, perennial tallgrasses occurs after early June (McKendrick et al. 1975) and, in 199 1, essentially no rainfall occurred from mid-June to October. We conclude that ungrazed and unburned tallgrass prairie plant communities, and probably most C,dominated plant communities with periodic moisture stress, will have greater aboveground and belowground productivity with elevated CO,-if annual precipitation amount and distribution remain near current levels. That increased productivity results from higher water-use efficiency related to the reduction in stomata1 conductance of plants exposed to elevated CO,. We further suggest that in environments with frequent strong winds and warm temperatures, exposure chambers will produce environments that mimic elevatedCO2 effects, and, therefore, may obscure responses to elevated COz. We found no evidence for a drastic shift in competitive advantage for the C, perennial grasses of the tallgrass prairie. C, forbs may increase in productivity, but probably will not increase in population at the expense of the C,, perennial tallgrass species. ACKNOWLEDGMENTS
Dale Strickler, Denise Garrett, Dean Larson, Garry Hatter,
and Neal Adam were extremely valuable in accomplishing this research. Terry Bolger helped in the initial setup of the study. This research was supported by the U.S. Department of Energy, Carbon Dioxide Research Division, via a grant to C. E. Owensby. Contribution number 92-480-J from the Kansas Agricultural Experiment Station. L ITERATURE
C ITED
Anderson, K. L., C. E. Owensby, and E. F. Smith. 1970. Burning bluestem range. Journal of Range Management 23: 8 1-92.
Bazzaz, F. A. 1990. The response of natural ecosystems to the rising global CO1 levels. Annual Review of Ecology and Systematics 21: 167-196. Boden, T. A., P. Kanciruk, and M. P. Farrell. 1990. Trends ‘90: A compendium of data on global climate change. Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, Oak Ridge, Tennessee, USA. Bristow, K. L. 1987. On solving the surface energy balance equation for surface temperature. Agriculture and Forest Meteorology 39:49-54. Campbell, G. S. 1977. An introduction to environmental biophysics. Springer Verlag, New York, New York, USA. Coyne, P. I., and J. A. Bradford. 1985. Comparison of leaf
Novemhcr I993
TALLGRASS PRAIRIE AND ELEVATED CO,
gas exchange and water-use efficiency in two eastern gamagrass accessions. Crop Science 25:65-75. Curtis. P. S., L. M. Balduman, B. G. Drake. and D. F. Whigham. 1990. Elevated atmospheric CO, effects on belowground processes in C, and C, estuarine marsh communities. Ecology 71:200 l-2006. Curtis. P. S., B. G. Drake, P. W. Leadley, W. J. Arp. and D. F. Whigham. 1989. Growth and senescence in plant communities exposed to elevated CO, concentrations on an estuarine marsh. Oecologia (Berlin) 78:20-29. Dahlman, R. C. 1984. Vegetation responses to carbon dioxide research plan. DOE/ER-0187. U.S. Department of Energy. Carbon Dioxide Research Division. Office of Basic Science, Washington, D.C., USA. Drake. B. G., and P. W. Leadley. 1991. Canopy photosynthesis of crops and native plant communities exposed to long-term elevated CO,. Plant, Cell and Environment 14: 853-860. Eamus. D. 1991. The interaction of rising CO, and temperature with water use efficiency. Plant. Cell and Environment 14:843-852. Heagle, A. S.. R. B. Philbeck. H. H. Rogers, and M. B. Letchworth. 1979. Dispensing and monitoring o7onc in open top field chambers for plant effects studies. Phytopathology 69: 15-20. Kanemasu. E. T.. and C. B. Tanner. 1968. A note on heat transport. Bioscience 18:327-329. Kimball. B. A. 1983. Carbon dioxide and agricultural yield: an assemblage and analysis of 430 prior observations. Agronomy Journal 75:779-788. Kirkham, M. B.. H. He. T. P. Bolger. D. J. Lawlor. and E. T. Kanemasu. 1991. Leaf photosynthesis and water use of big bluestem under elevated carbon dioxide. Crop Science 31:1589-1594. Kitano. M.. and H. Eguchi. 1989. Dynamic analysis of stomatal responses by an improved method of leaf heat balance. Environmental and Experimental Botany 29:175-185. Launchbaugh. J. L.. and C. E. Owensby. 1978. Kansas rangelands: their management based on a half century of research. Kansas .Agricultural Experiment Station Bulletin 622. Lund. Z. F. R. W. Pearson. and G. A. Buchanan, 1970. An implanted soil mass technique to study herbicide effects on root growth. Weed Science 18:?79-28 1, McKcndrick. J. D., C. E. Owensby, and R. M. Hyde. 1975. Big bluestem and indiangrass vegetative reproduction and annual reserve carbohydrate and nitrogen cycles, Agro-Ecosystems 2175-93. Mooney. H. A., B. G. Drake. R. J. Luxmore, W. C. Oechel, and L. F. Pitelka. 1991. Predicting ecosystem responses to elevated CO, concentrations, Bioscience 41:96-104. Nijs. I., I. Impens. and T. Behaeghe. 1988. Effects of rising
653
atmospheric carbon dioxide concentration on gas exchange and growth of perennial ryegrass. Photosynthetica 22:4450. Oechel, W. C., and B. R. Strain. 1985. Native species responses to increased carbon dioxide concentration. Pages 117-154 in B. R. Strain and J. D. Cure, editors. Direct effects of increasing carbon dioxide on vegetation, DOE/ ER-0238. U.S. Department of Energy. Washington, D.C., USA. Owensby. C. E., and K. L. Anderson. 1967. Yield responses to time of burning in the Kansas Flint Hills. Journal of Range Management 20: 12-I 6. Owensby, C. E., P. I. Coyne, and L. M. Auen. 1989. Response ofvegetation to carbon dioxide, Number 54. Rangeland-plant response to elevated CO,. U.S. Department of Energy, Washington, D.C., USA. Owensby, C. E., P. I. Coyne, L. M. Auen, and N. Sionit. 1990. Response of vegetation to carbon dioxide, Number 59. Rangeland-plant response to elevated COY. U.S. Department of Energy, Washington, D.C., IJSA. Owensby. C. E., R. M. Hyde, and K. L. Anderson. 1969. Effects of clipping and supplemental nitrogen and water on loamy upland bluestem range. Journal of Range Management 23:341-346. Reichers. G. H., and B. R. Strain. 1988. Growth of blue grama (Routeloua gruczlrs) in response to atmospheric CO, enrichment. Canadian Journal of Botany 66: 1570-l 573. Strain, B. R., and J. D. Cure. 1985. Direct effect ofincreasing carbon dioxide on vegetation. Carbon Dioxide Research. State of the Art. U.S. Department of Energy, Washington, D.C., LISA. Tolley. L. C., and B. R. Strain. 1984. Effects ofatmospheric CO, enrichment and water stress on growth of Llquidumher st!,racrflua L. and Pinus taeda L. seedlings. Canadian Journal of Botany 10:2 135-2 139. USDA [United States Department of Agriculture]. 1972. The nation’s range resources-a forest-range environmental study. Forest Resources Report 19. Weaver. J. E., and F. W. Albertson. 1943. Resurvey of grasses. forbs. and underground plant parts at the end of the great drought. Ecological Monographs 13:63-l 18. Williams. R. E.. 8. W. Allred, R. M. DeNio, and H. E. Paulsen. Jr. 1968. Conservation. development, and use of the world’s rangelands. Journal of Range Management 21:355360. Wray, S. M., and B. R. Strain. 1986. Response of two old field perennials to interactions of CO, enrichment and drought stress. American Journal of Botany 73: 1486-149 1. Wray. S. M., and B. R. Strain. 1987. Interaction of age and competition under CO, enrichment. Functional Ecology 1: 145-149.
