• Season-long effects of elevated CO2 concentration ([CO2]) on the carbon balance of the rice (Oryza sativa) canopy are reported here.
• The experiment was conducted in six sunlit, semiclosed growth chambers for an entire growing season. Rice plants (cv. Nipponbare) were grown at 350 µmol mol−1[CO2] (ambient) in three chambers, or at 650 µmol mol−1 (elevated) in the other three chambers. Canopy net photosynthesis and night-time respiration were determined in the chambers by mass balance.
• Both canopy gross photosynthesis and total respiration, through the entire growing season, were increased by the CO2 enrichment. But CO2-induced variations in canopy carbon gain were mainly caused by changes in canopy photosynthesis. The enhancement of daily canopy gross photosynthesis by elevated [CO2] was 33.4% for the first 3 week, but it declined gradually and disappeared by heading. Enhancement of daily net carbon gain also decreased as rice plants grew.
• These results show that the increase in rice biomass by elevated [CO2] results more from the increase in carbon gain at early rather than later stages of growth.
Many studies have examined plant responses of photosynthesis and respiration to elevated [CO2] at the single-leaf and cellular levels (e.g. Stitt, 1991; Wolfe et al., 1998). However only a limited number of studies have examined the photosynthetic and respiratory responses at the whole-plant or canopy level, probably because of technical difficulties. Responses to elevated [CO2] at the whole-plant level may be different from those at single-leaf and cellular levels, because there are some variations of physiological (nitrogen content, sugar content, etc.) and environmental (light, temperature, etc.) factors between plant organs and even between single leaves in a canopy.
The response of canopy respiration to long-term elevated [CO2] has been reported in few experiments (Baker et al., 1992; Gifford, 1995; Baker et al., 2000). Baker et al. (1992) examined the effects of daytime CO2 enrichment on canopy respiration, and reported that canopy respiration rate on a ground area basis increased with increasing [CO2], but specific respiration rate decreased slightly under elevated [CO2]. Gifford (1995) conducted an experiment for wheat canopy and reported that maintenance and basal respiration coefficients decreased in elevated [CO2], but the respiration : photosynthesis ratio was unchanged. Clearly the current information is insufficient for understanding fully the effects of elevated [CO2] on canopy respiration.
Most canopy-level investigations of CO2 exchange rates addressed photosynthesis and respiration separately. It is necessary to address both photosynthetic and respiratory responses to elevated [CO2] because respiration is closely linked to photosynthesis (Amthor, 1989; Gifford, 1995), and canopy carbon gain is the balance between photosynthesis and respiration.
The objective of this study on rice was to examine the effects of long-term elevated [CO2] on canopy photosynthesis, respiration and carbon gain in relation to growth responses.
Materials and Methods
Controlled environment chambers
Rice plants (Oryza sativa L., cv. Nipponbare) of Japonica-type were grown in six naturally sunlit, semiclosed growth chambers for an entire growing season. Chamber dimensions were 4 × 3 × 2 m (L × W × H) with the space for plant growth 4 × 2 × 2 m. Each chamber housed two stainless-steel containers (1.5 × 1.5 × 0.3 m; L × W × D) filled with paddy soil. The frames, rear (north) walls, and floor of the chamber were made of stainless steel. The frames were glazed with 5-mm-thick tempered glass whose transmittance of visible light were > 80%. Air temperature and rh in each chamber were controlled by electrical resistive heaters (with bubbling system for humidification) and cold-water heat exchangers using PID (Proportional + Integral + Derivative) controllers (DB1000, CHINO, Tokyo, Japan). Air temperature and rh in each chamber were measured with temperature-humidity sensors (HN-Q500-1, CHINO, Tokyo, Japan) shielded against direct solar radiation and mounted above a rice canopy. In this experiment, air temperature was controlled to track ambient air temperature with the seasonal mean temperature being 23.4°C and a rh of 80 ± 1.9%. [CO2] was maintained at 353 ± 15/396 ± 23 µmol mol−1 (day/night) in three ambient [CO2] chambers and 667 ± 36/700 ± 41 µmol mol−1 (day/night) in three elevated [CO2] chambers. Daytime [CO2] was maintained by a computer-controlled pure CO2 injection system, which compensated for CO2 uptake by the rice canopy. During night-time , [CO2] increased due to plant respiration, but was kept below 100 µmol mol−1 higher than the daytime [CO2] by a computer-controlled air ventilation system, which introduced ambient air to reduce [CO2]. Ambient air temperature was measured with a platinum resistance thermometer which was shielded, aspirated and placed outside the chambers. Environmental data in each chamber and ambient air temperature were monitored every 10 s, and 5 min means were recorded. Ambient [CO2] data were provided by the laboratory of Micrometeorology of NIAES. They monitored ambient [CO2] every 10 s at several heights on an observation tower, which was located c. 50 m south of these chambers.
Germinated seeds of rice were sown on 20th April in 1998. Seedlings were grown in plug pots at 23°C, 80% rh and 350 or 650 µmol mol−1[CO2]. On 15th May, they were transplanted in the containers in chambers with 3 seedlings per hill at 20 × 20 cm spacing. Plants were fertilized with 5 g N, 15 g P2O5, and 15 g K2O per m2 just before transplanting, and 3 g N per m2 on 56 d after transplanting (DAT). The amount of fertilizer was based on local agronomic practices. The containers were flooded with water at 1–5 cm depth throughout the season. When leaf area index of rice canopy reached 3, shading nets (50% light transmittance) were installed at canopy height along the outside of each chamber to make a light environment similar to that in a field. The rice plants were harvested on 15th October (153 DAT).
Growth and yield measurement
Three rice hills were destructively sampled from each chamber at 20, 40, 67, 98, 122 and 153 DAT. The gaps of the sampled plants were filled with potted plants grown outdoors with the same nitrogen application as those in the chambers. These pot-grown plants were not included in any further sampling. After leaf area was measured for each sampled plant, plants were detached into leaf blade, leaf sheath + stem, root, ear and dead leaf blade. Then each sample was oven-dried for > 48 h at 80°C and d. wt was determined. Samples of one chamber from each [CO2] treatment were used for carbon (C) and nitrogen (N) analysis. After grinding samples, C and N concentration were determined by a CN coder (MT-700, Yanako, Kyoto, Japan). At harvest, 12 rice hills were destructively sampled from each chamber and the yield was determined for each chamber.
Canopy CO2 exchange rate
[CO2] in each chamber was monitored every 10 s by an infra-red CO2 controller (ZFP9GD11, Fuji-denki, Tokyo, Japan) and recorded every 5 min as a 5-min average. For a more precise measurement of [CO2] in the chambers than the CO2 controllers, sample air from each of the six chambers was taken to the control house, and [CO2] was determined by an infra-red gas analyser (IRA-107, Shimadzu, Kyoto, Japan), which was automatically calibrated three times a day against nitrogen (zero) gas and standard CO2 gas (700 µmol mol−1). It took 5 min to determine [CO2] in each chamber and 30 min to scan all the six chambers. The [CO2] thus determined was recorded and used to describe the CO2 regimes in the chambers. The rate of pure CO2 injection to maintain [CO2] in each chamber constant at the target level was controlled and measured by a mass flow controller (SEC-400MARK3, STEC, Kyoto, Japan), and recorded every 5 min for each chamber.
The net photosynthetic rate of the rice canopy on a ground area basis was determined as:
( Eqn 1)
(Pnet, canopy net photosynthesis rate (mg CO2 m−2 min−1); Cin, carbon injection rate to keep [CO2] in a chamber (mg CO2 m−2 min−1); L, chamber leakage rate (mg CO2 m−2 min−1); Rsoil, CO2 flux out of the paddy water and soil (mg CO2 m−2 min−1).) Canopy dark respiration rate on a ground area basis in night-time was determined as:
( Eqn 2)
(Rnight, canopy dark respiration rate in night-time (mg CO2 m−2 min−1); ΔC, increase of [CO2] in a chamber (mg CO2 m−2 min−1) while the air-ventilation is closed.) The air-ventilation system was programmed to allow for the measurement of ΔC, while maintaining the night-time [CO2] at desired level.
The rate of CO2 leakage (L) out of a chamber was estimated by modification of the method of Acock & Acock (1989) and Kimball (1990). Every few weeks pure N2O was injected into each chamber, and the decay of N2O concentration was measured by using the air sampling system described above and an infra-red gas analyser (ZRC1ZC11, Fuji-denki, Tokyo, Japan) for N2O. L was calculated from the measured leakage rate and the [CO2] gradient between ambient air and chambers by mass balance.
The CO2 flux out of the paddy water and soil was measured at air temperatures between 15 to 35°C with a 5°C step under flooded conditions, after all aboveground plant material had been removed at the end of the growing season.
The canopy dark respiration rate during the daytime (Rday) was calculated by assuming the same rate as that at night-time and at a corresponding temperature (Baker et al., 1997; Monje & Bugbee, 1998). Gross photosynthesis rate (Pgross) was estimated as Pnet plus Rday. The daily total respiration (Rtotal) was estimated as Rday plus Rnight and the daily carbon gain of a canopy (Cgain) was calculated as:
( Eqn 3)
The canopy net photosynthesis rate per unit leaf area (Pnet/Leaf area; Pleaf) and specific respiration rate (Rtotal/Total d. wt; Rdw) were calculated from destructive sampling date and a Pnet and Rtotal of few days before and after sampling.
Daily values of canopy photosynthesis, respiration and carbon gain were compared between the two [CO2] treatments for five periods: 11–40, 41–70, 71–100, 101–130, 131–153 DAT, and the entire growing season by one-way ANOVA with the variance between the chambers as the error variance.
Destructive sampling data, Pleaf and Rdw were also tested for significant effects of the [CO2] treatment by one-way ANOVA, but data of C and N analysis were tested by one-way ANOVA with the variance between the plants as error.
Graphing and smoothing of the daily measurement data were done with the computer software ‘Kaleida Graph’ (SYNERGY SOFTWARE, Reading, Pennsylvania, USA).
Plant growth and yield
Total d. wt was increased by elevated [CO2], but difference between elevated and ambient [CO2] treatment was not statistically significant before harvest, probably due to the limited number of samples. At harvest, total d. wt was significantly increased by 10.0% (P = 0.008; Fig. 1a) by the elevated [CO2] treatment. Brown rice yield was 467 ± 33 and 568 ± 19 g m−2 under ambient and elevated [CO2], respectively, and significantly increased by the elevated [CO2] treatment (P = 0.0001). Leaf area did not respond to elevated [CO2] in the early stages of growth, but decreased significantly after heading (P = 0.04; Fig. 1b) under the elevated [CO2] treatment. The number of tillers increased under elevated [CO2] by 22% (P = 0.006) at the maximum tiller number stage (45DAT) and 8% (P = 0.03) at harvest.
Carbon and nitrogen concentration
The C and N concentration in leaves and the whole plant decreased with time as rice plants grew under both ambient and elevated [CO2]. Compared with ambient [CO2], C concentrations were c. 2% higher at elevated [CO2] for both leave and whole plants, almost through the growing season, whereas N concentrations gradually decreased under elevated [CO2] with the decrease being greatest (12% for leaf, 9% for whole plant) in the middle of the grain filling stage.
Leaf N contents per unit leaf area were significantly decreased by elevated [CO2] (Fig. 2) with the decrease being greatest (23%, P = 0.008) in the middle of grain filling stage.
Measured and estimated carbon gain
The final Cgain of the rice canopy under ambient and elevated [CO2] was calculated (from harvest biomass and C concentration) and compared with Cgain estimated by integrating the daily canopy Cgain in Eqn 3 (Fig. 3). Differences between the measured and estimated Cgain were small (RMSE = 1.9%).
Canopy CO2 exchange rate
Daily Pgross, Pnet, Rtotal and Rnight gradually increased until 50 DAT, and after a 50-d period of minor fluctuation, began to decrease around heading (100DAT) under both ambient and elevated [CO2] (Fig. 4b,c).
Pgross and Pnet were significantly increased by elevated [CO2] for the periods before heading (Table 1), but their enhancement by CO2 enrichment was diminished at heading (Fig. 4d). At the late grain filling stage, Pgross and Pnet were significantly decreased by CO2 enrichment (Table 1). Seasonally-integrated Pgross and Pnet were increased 8.9 and 11.3% respectively, by the elevated [CO2]. Pleaf was significantly enhanced by elevated [CO2], but its enhancement rate gradually decreased through the growing season (Fig. 5).
Table 1. Average Pgross, Pnet, Rtotal, Rnight, Cgain, air temperature and incident PAR at intervals during the growing season of rice (Oryza sativa L.) under elevated and ambient [CO2] treatments. Growing season was divided into five stages: early vegetative [11–40 DAT], late vegetative [41–70 DAT], panicle formation [71–100 DAT], early grain filling [101–130 DAT], and late grain filling [131–152 DAT]
Pgross (gCO2 m−2 day−1)
Pnet (gCO2 m−2 day−1)
Rtotal (gCO2 m−2 day−1)
Rnight (gCO2 m−2 day−1)
Cgain (gCO2 m−2 day−1)
Air temp (°C)
PAR (mol m−2 day−1)
, Significantly different from ambient at P = 0.01.
Rtotal and Rnight were significantly increased by elevated [CO2] before panicle initiation (70 DAT) and after heading, but were unaffected between these stages (Table 1). Seasonally-integrated Rtotal and Rnight were increased by 7.1 and 4.5%, respectively. Specific respiration rate (Rdw) was significantly enhanced by elevated [CO2] early in the growing season, but decreased significantly around heading (Fig. 5).
Canopy carbon balance
Daily Cgain increased gradually after transplanting and reached a maximum at about 80 DAT under both ambient and elevated [CO2] and declined thereafter (Fig. 6).
In the early growth stages, daily Cgain was increased by elevated [CO2], but the increase diminished by heading (100 DAT), and then Cgain decreased under elevated [CO2] at the late grain filling stage (Table 1).
Carbon gain under elevated [CO2]
Canopy carbon gain was increased by elevated [CO2] before heading (Fig. 6, Table 1), however the initial enhancement rate of daily carbon gain by elevated [CO2] gradually decreased with time. This indicates that the increase in rice biomass by elevated [CO2] resulted more from the increase in carbon gain during the early growth stages than at late stages.
No enhancement or even decline in canopy carbon gain after heading was observed under elevated [CO2] (Fig. 6), although yield was increased by elevated [CO2]. These changes are apparently in conflict, since a major portion of the carbohydrate in rice grains comes from the carbon assimilated after heading. It must be noted, however, that rice yield has been related to the level of carbohydrates stored in the stem before heading, and to the ability of the plant to translocate this storage to the ear (Yoshida, 1972). In this experiment, accumulated starch content in the stem and leaf sheath just before heading was about two times higher under elevated than ambient [CO2] (H. Sakai et al., unpublished). The contribution of the stored starch to starch accumulation in the ear at harvest could be up to 17% under elevated [CO2] and 5% under ambient [CO2]. Grüters (1999) reported that the contribution of stem carbohydrate reserves to yield was enhanced by elevated [CO2] in wheat.
Long-term response of canopy photosynthesis
At the single leaf level photosynthetic acclimation to long-term CO2 enrichment has often been reported (Stitt, 1991; Wolfe et al., 1998). Few studies have examined the response of canopy photosynthesis to long-term elevated [CO2]. Baker et al. (1990b) reported that differences in the net photosynthesis of the rice canopy between ambient and elevated [CO2] treatments had disappeared after flowering (Fig. 4 in Baker et al., 1990b). However, Baker et al. (1997) found no photosynthetic down-regulation when they conducted [CO2] cross-switching experiments with the rice canopy. Continued enhancements of canopy photosynthesis by long-term CO2 enrichment have been reported in wheat (Monje & Bugbee, 1998), and soybean (Jones et al., 1984). The elevated rates of photosynthesis seem to be maintained in CO2-enriched plants when the sink capacity becomes sufficiently large to prevent feed-back inhibition (Monje & Bugbee, 1998). In our study, we found higher starch accumulation in CO2-enriched plants throughout the season (H. Sakai et al., unpublished), although d. wt and tiller number were enhanced by elevated [CO2]. Rice plants in this experiment may, therefore, be sink-limited.
The reduction of active rubisco has been frequently reported as an acclimation response to elevated [CO2] (Lawlor & Mitchell, 1991; Rowland-Bamford et al., 1991; Sage, 1994; Nakano et al., 1997; Vu et al., 1997; Wolfe et al., 1998). A decrease of leaf N content has also been reported (Rowland-Bamford et al., 1991; Nakano et al., 1997; Wolfe et al., 1998). Nakano et al. (1997) reported no difference between ambient and elevated [CO2] treatments in the relationship between rubisco and leaf N content of rice and that a decrease of rubisco was the result of the decrease in leaf N content. Makino et al. (1997, 2000) reported that the decrease in leaf N content is not due to dilution of N caused by relative increases in leaf area or plant mass, but the result of a change in N allocation at the morphogenetic level of the whole-plant. In this study, the partitioning rate of N to leaf blade was also decreased throughout the growing season (data not shown). A decrease in leaf N content by elevated [CO2] was also found (Fig. 2), and the ratio of [elevated : ambient] gradually decreased with rice development. Because the relationship between leaf N content and photosynthesis rate under elevated [CO2] is not available, we could not estimate the effect of a decrease in leaf N content on canopy photosynthesis under elevated [CO2]. A decrease in the enhancement of canopy photosynthesis under elevated [CO2] might be partly due to a decrease in leaf N, and it is suggested that canopy photosynthetic responses to elevated [CO2] are closely related to plant resource partitioning within a canopy.
Canopy net photosynthesis rate per unit leaf area (Pleaf) was enhanced by elevated [CO2] at heading (Fig. 5) when the enhancement of canopy photosynthesis had diminished. While the number of tillers and total d. wt of rice plants were increased, leaf area development did not respond to elevated [CO2] at the growth stage before heading. This phenomenon has been observed in many other studies with rice (Imai et al., 1985; Baker et al., 1990a; Kim et al., 1996; Ziska et al., 1997). At heading and after, leaf senescence was enhanced by elevated [CO2] (Fig. 1b), a response also reported by Baker et al. (1990a) and Sicher (1998). This enhanced leaf senescence should have contributed to the decrease in canopy photosynthesis under elevated [CO2] after heading. The mechanisms of the enhanced leaf senescence under elevated [CO2] may involve acceleration of plant development due to the higher plant temperatures resulting from partial stomatal closure under elevated [CO2]. The increased number of ears in elevated [CO2] should have required leaf N to be re-translocated to ears at a higher rate, which should have also contributed to the enhanced leaf senescence. However the detailed analyses of these mechanisms are beyond the scope of this paper.
Long-term response of canopy respiration
Knowledge about the effects of elevated [CO2] on plant respiration is quite limited, especially at the whole-plant level. In this study, canopy respiration rates on a ground area basis were increased by elevated [CO2] (Table 1) because of increased biomass, but specific respiration rate (Rdw) was significantly decreased in the middle of growing season (Fig. 5). This result is consistent with Baker et al. (1992), who reported that the differences in the specific respiration rate among [CO2] treatments were influenced by differences in the N concentration of aboveground biomass. In this study, however, the relationship between specific respiration rate and N concentration of total biomass was different among [CO2] treatments (data not shown). Gifford (1995) reported that the respiration : photosynthesis ratio was not affected by elevated [CO2]. The respiration : photosynthesis ratio through the growing season averaged 0.43 and 0.40 under ambient and elevated [CO2], respectively. While total respiration was up to 40% of canopy gross photosynthesis, the change in respiration due to elevated [CO2] did not contribute significantly to CO2-induced variations in canopy carbon gain (Table 1).
In conclusion, this study reports the long-term response of canopy carbon gain to elevated [CO2]. CO2-induced variations in canopy carbon gain were mainly due to changes in canopy photosynthesis. The long-term response of canopy photosynthesis may be caused by both a decline in carbon assimilation per unit leaf area (Pleaf) and the loss of leaf area due to enhanced leaf senescence.
This work was supported by the Rice-FACE project under the CREST program of Japan Science and Technology Corporation. We appreciate the laboratory of Micrometeorology of NIAES for providing us with the data of ambient [CO2] which were needed for chamber leakage calculation. The authors thank anonymous reviewers for valuable suggestions.