Diurnal and Seasonal Variations in the Photosynthetic Characteristics and the Gas Exchange Simulations of Two Rice Cultivars Grown at Ambient and Elevated CO2

Investigating the diurnal and seasonal variations of plant photosynthetic performance under future atmospheric CO2 conditions is essential for understanding plant adaptation to global change and for estimating parameters of ecophysiological models. In this study, diurnal changes of net photosynthetic rate (Anet), stomatal conductance (gs), and photochemical efficiency of PSII (Fv′/Fm′) were measured in two rice cultivars grown in the open-top-chambers at ambient (∼450 μmol mol–1) and elevated (∼650 μmol mol–1) CO2 concentration [(CO2)] throughout the growing season for 2 years. The results showed that elevated (CO2) greatly increased Anet, especially at jointing stage. This stimulation was acclimated with the advance of growing season and was not affected by either stomatal limitations or Rubisco activity. Model parameters in photosynthesis model (Vcmax, Jmax, and Rd) and two stomatal conductance models (m and g1) varied across growing stages and m and g1 also varied across (CO2) treatments and cultivars, which led to more accurate photosynthesis and stomatal conductance simulations when using these cultivar-, CO2-, and stage- specific parameters. The results in the study suggested that further research is still needed to investigate the dominant factors contributing to the acclimation of photosynthetic capacity under future elevated CO2 conditions. The study also highlighted the need of investigating the impact of other environmental, such as nitrogen and O3, and non-environmental factors, such as additional rice cultivars, on the variations of these parameters in photosynthesis and stomatal conductance models and their further impacts on simulations in large scale carbon and water cycles.


INTRODUCTION
Under the influence of human activities, atmospheric CO 2 concentration [(CO 2 )] has been increasing to 415 µmol mol −1 in 2019 since the Industrial Revolution, before which the (CO 2 ) was steadily maintained at about 280 µmol mol −1 (National Oceanic and Atmospheric Administration NOAA, 2019). It would possibly reach up to 936 µmol mol −1 in 2100 under RCP8.5 scenario (IPCC, 2014). Rice (Oryza sativa L.) has been one of the most important crops for human beings since the 1900s, especially in Asia and in the tropical and subtropical regions of Africa. Rice has the third largest planting area in the world and has been the main source of daily food for nearly half of the world's population. With Global (CO 2 ) rising and the increase of global population, it is critical to investigate and predict the responsive changes of the physiology and growth of rice to the increasing (CO 2 ).
As the essential substrate for photosynthesis, the increase of (CO 2 ) will stimulate photosynthesis and increase the biomass production and yield (Ainsworth and Long, 2005;Ainsworth and Rogers, 2007;Brito et al., 2020). Soil-Plant-Atmosphere-Research (SPAR) showed that rising (CO 2 ) from 350 to 700 µmol mol −1 increased rice growth, grain yield and canopy photosynthesis and increased the final aboveground biomass by 29% with sufficient irrigation (Baker and Allen, 2005). The increase of air temperature at future elevated (CO 2 ) made the CO 2 effect on photosynthesis more complicated. Leaf photosynthetic rate (A) of two species, rice and soybean, were both increased under (CO 2 ) enrichment, but this enhancement was reduced when the air temperature increased by 3 • C (Vu et al., 1997). Rising (CO 2 ) from 370 to 700 µmol mol −1 could offset the negative effect on photosynthesis due to 3.0-3.9 • C warming, but had larger negative effect on photosynthetic carboxylation capacity in warming condition compared with ambient air temperature (Lamba et al., 2018). The combination of rising (CO 2 ) and temperature increased the photosynthesis but decreased yield for rice (Cai et al., 2016). The yield reduction under the combination of 200 µmol mol −1 above ambient (CO 2 ) and 1 • C warming was reported in another research and the decrease of spikelet density might be the dominant factor (Wang W. et al., 2018).
Stomatal function is an important factor affecting photosynthesis and transpiration of plants. Lower stomatal conductance (g s ) and transpiration rate (TR), therefore higher water use efficiency (WUE) due to elevated (CO 2 ) were reported in varieties of species (Wand et al., 1999;Medlyn et al., 2001;Ainsworth and Rogers, 2007;Shimono and Bunce, 2009). Free air concentration enrichment (FACE) experiments showed an average increase of 23% in yield and a 25% reduction in g s when (CO 2 ) increased from 360 to 627 µmol mol −1 (Ainsworth, 2008). FACE experiments in Iwate, Japan, indicated that raising (CO 2 ) by 200 µmol mol −1 significantly decreased g s by 23% on average and doubling (CO 2 ) made transpiration of rice reduced by 45 and 41% at full heading stage and mid-ripening stage, respectively (Shimono and Bunce, 2009;Shimono et al., 2010). Another research showed that lower g s of wheat grown under ambient (CO 2 ) was observed compared with that under elevated (CO 2 ) when ammonium fertilization was supplied as the sole N source, and this phenomenon disappeared when plants were set under nitrate nutrition (Torralbo et al., 2019). And EucFACE in 2017 found no significant influence on g s for one dominant C 3 grass and two sympatric C 3 forbs due to increasing (CO 2 ) by 150 µmol mol −1 (Pathare et al., 2017). Therefore, the regulation of stomatal conductance by elevated CO 2 could vary greatly depending on other environmental factors.
Responses of plant photosynthesis to short-term changes in environmental factors can be reflected through diurnal variations in photosynthesis. The maximum of A net could be observed at 9 a.m. in tea [Camellia sinensis (L.) O. Kuntze] (Mohotti and Lawlor, 2002). An Conviron environmental chamber experiment for drought-tolerant plant Elaeagnus umbellata showed that A net and g s both reached the maximum rates early in the day (around 10 a.m.) and the daily average levels decreased significantly during times of drought stress (Naumann et al., 2010). Research for maize showed that the maximum of A net and g s occurred from 12 p.m. to 15 p.m. among two N fertilized and nonfertilized treatments, and the difference of A net was significant due to N treatment when the light intensity was high, while the influence was found minimal on g s and C i (Peng et al., 2014). FACE research focused on maize found no significant difference in the diurnal variation of A net at ambient and elevated (CO 2 ) in the early growing stage, but the difference could be observed in the late growing stage (Markelz et al., 2011). Depression of net CO 2 assimilation occurred around noon due to stomatal closure and the decrease of intercellular CO 2 (C i ) and elevated (CO 2 ) could either significantly enhance the assimilation ability or weaken the midday depression (Nijs et al., 1992;Špunda et al., 2005). However, when C 3 plants were exposed to elevated (CO 2 ) for an extended period of time, the accelerated rate of photosynthesis often cannot be maintained, a phenomenon called photosynthetic acclimation to elevated (CO 2 ) (Van Oosten and Besford, 1996). Ignoring this acclimation would possibly overestimate the magnitude of photosynthetic stimulation of elevated (CO 2 ) concentrations compared with what was expected with the influence of short-term elevated (CO 2 ) (Dusenge et al., 2019). Photosynthetic acclimation was usually associated with the down-regulation of stomatal conductance, Rubisco amount or activity, nitrogen concentrations or sink strength under elevated CO 2 conditions (Ainsworth et al., 2004;Leakey et al., 2009;Rogers et al., 2017). For plants holding nitrogen-fixing bacteria, reduction of photosynthesis will be effectively alleviated under long-term elevated (CO 2 ), especially under nitrogen limitation (Rho et al., 2020). Few studies have investigated the diurnal variations of photosynthetic characteristics of rice throughout the growing season. Whether the effects varied across different cultivars and growing stages requires systematic measurements of the diurnal variation of photosynthesis and stomatal conductance throughout the growing season to shed lights on both the shortand long-term effect of CO 2 on photosynthetic characteristics.
Understanding and predicting large-scale carbon, water, and energy cycles requires accurate simulations in leaf photosynthesis and stomatal activities (Laisk et al., 2005). The Farquhar-von Caemmerer-Berry model (FvCB model), Ball, Woodrow and Berry (BWB model), and Medlyn model (MED model) are widely used photosynthetic and stomatal models to simulate the photosynthesis and stomatal conductance at the leaf level (Farquhar et al., 1980;Ball et al., 1987;Medlyn et al., 2012). The parameters of these models (V cmax , J max , and R d from the FvCB model, m and g 0 from BWB model, g 1 and g 0 from MED model) are valuable for large-scale simulations and represent important physiological traits that determine plant photosynthetic potential and water-use efficiency. How do photosynthetic parameters of V cmax , J max , and R d and stomatal slope parameters vary among different rice cultivars and under different CO 2 conditions require further study and analysis. Whether using cultivar-specific or environment-specific, instead of generic, model parameters can increase the accuracy of both photosynthesis and stomatal conductance simulations needs further investigation too.
In this study, diurnal changes of net photosynthetic rate (A net ), stomatal conductance (g s ), intercellular CO 2 concentration (C i ) and chlorophyll a fluorescence characteristics (photochemical efficiency of PSII, F v /F m ; actual photochemical efficiency of PSII, PSII; photochemical quenching, qP) were measured in two rice cultivars grown in the open-top-chambers at ambient (∼450 µmol mol −1 ) and elevated (∼650 µmol mol −1 ) CO 2 concentrations throughout the growing season for 2 years. Seasonal variations of V cmax , J max , and R d and stomatal slope parameters were determined by A net -C i measurements at different growing stages. The objectives of this study were: (1) to investigate the diurnal and seasonal variations of rice photosynthetic performance under future atmospheric CO 2 conditions; (2) to compare the diurnal and seasonal variations across two rice cultivars; (3) to test whether the photosynthetic and stomatal conductance models are effective under elevated CO 2 conditions and whether using cultivar-, CO 2 -and stage-specific parameters can improve the accuracy of the photosynthesis and stomatal conductance simulations. Specifically, we hypothesized that (1) the positive effects of elevated (CO 2 ) observed in the earlier growing season will decrease in the later growing season; (2) the extent of photosynthetic acclimation will be smaller for the more CO 2responsive cultivar; (3) the cultivar-, CO 2 -, and stage-specific parameters will increase the accuracy of the photosynthetic and stomatal conductance models.

Site Description
The study site was located in the agrometeorological experimental station of Nanjing University of Information Science and Technology, in Nanjing, Jiangsu province of China (32 • 16 N, 118 • 86 E). The climate in this region characterizes subtropical monsoon season, with the annual average precipitation of 1,100 mm, the average temperature in recent years of 15.6 • C and the average annual frost-free period of 237 days. The soil texture in the tillage layer was loamy clay and the clayey content was 26.1%. The bulk density of 0-20 cm soil was 1.57 g·cm −3 and the pH (H 2 O) value were 6.3. The organic carbon and total nitrogen content were 11.95 and 1.19 g·kg −1 , respectively.

Experimental Design
Open top chambers (OTC) were used in the experiment to simulate elevated (CO 2 ) treatments and description was provided in Supplementary Material 1. The OTCs are regular octagonal prisms with a diameter of 3.75 m, a height of 3 m, and a base area of 10 m 2 . There were two (CO 2 ) treatments, ambient [a(CO 2 )] and elevated [e(CO 2 )], each with four replicative chambers. The treatment of elevated (CO 2 ) started from the turning green stage and lasted to the end of growing season. The (CO 2 ) concentration in the OTCs was controlled with an automatic control platform, composed of CO 2 sensors, gas-supplying devices and automatic control system. Three wind-blowing fans were placed in each chamber to make the CO 2 gas in the chamber evenly distributed and the top of the OTC is designed with an opening inclined 45 • inward to avoid the rapid loss of CO 2 gas. The CO 2 sensor fed back the surrounding CO 2 concentration information in the chambers to the automatic control system every 2 s. Our experiment was performed from 2019 to 2020 to make sure that the trend we observed was not a random impact. The daytime (CO 2 ) averaged over the growing season was 641 ± 43 and 631 ± 39 µmol mol −1 in elevated (CO 2 ) chambers and 478 ± 34 and 485 ± 33 µmol mol −1 in ambient (CO 2 ) chambers in 2019 and 2020, respectively. The nighttime (CO 2 ) was 667 ± 27 and 673 ± 24 µmol mol −1 in elevated (CO 2 ) chambers and 537 ± 49 and 550 ± 36 µmol mol −1 in ambient (CO 2 ) chambers.
Two rice cultivars, Yangdao6, a more CO 2 -responsive indica cultivar and Wuyunjing30, a less CO 2 -responsive japonica cultivar, as indicated in previous studies (Zhu et al., 2015;Tao and Wang, 2021;Wang et al., 2021), were selected in this study. Seedlings of each rice cultivar were transplanted into the eight OTCs on June 19, 2019 and June 16, 2020, and harvested on October 25, 2019 and October 28, 2020, respectively. The spacing of the hills was 16.7 cm × 25 cm (equivalent to 24 hills m −2 ). During the whole growing season, sufficient supplies of water and fertilizer were maintained.

Photosynthetic Measurements
Gas exchange characteristics (including net photosynthetic rate and stomatal conductance) were measured with a portable infrared gas analyzer (LI-COR 6400LCF; LI-COR, Lincoln, NE, United States) on one randomly selected and fully expanded healthy leaf of each cultivar from each chamber. The measurement was taken at jointing, booting, heading, grainfilling and maturity stage on July 24, August 18, September 8, September 28 and October 13 in 2019, and July 25, August 24, September 7, September 24 and October 20 in 2020, respectively, which were confirmed by observation and previous research (Counce et al., 2000;IRRI, 2013). During diurnal measurements, the (CO 2 ) of the leaf chambers were set according to the OTC conditions and the temperature and light levels were set as the environmental conditions. Photochemical efficiency of PSII in light-adapted leaves (F v /F m ) and photochemical quenching (qP) were measured using a Licor 6400-40 Leaf Chamber Fluorometer. The photosynthesis-CO 2 response (A net -C i ) curves were measured on the next day after the diurnal measurements were taken. During measurements, leaves were acclimated for 30-60 min before adjusting the CO 2 concentrations. Thereafter, CO 2 concentration was decreased in six steps (400,300,230,150,90, and 50 µmol mol −1 CO 2 ) and then increased in four steps (400, 600, 800, and 1,000 µmol mol −1 CO 2 ). Leaf temperature was controlled at 35, 35, 35, 30, and 25 • C when A net -C i curve measurement was conducted at jointing, booting, heading, grain-filling, and maturity stage, respectively, and photosynthetic photon flux density (PPFD) was maintained at 2,000 µmol m −2 s −1 all the time because the light level is close to the midday light level in the region.
where C i is the intercellular (CO 2 ), O is the oxygen concentration, * is the photosynthetic CO 2 compensation point without dark respiration, K C and K O are the Michaelis-Menten constants for CO 2 and O 2 and can be found in Bernacchi et al. (2001). Equation 1 can be used if A net is limited by Rubisco activity [low (CO 2 )] and if A net is limited by the regeneration of ribulose-1,5-bisphosphate (RuBP) at high (CO 2 ), it can be determined as follows: where a and b are model parameters and are given as 4.0 and 8.0, respectively (Bernacchi et al., 2003); J is the photosynthetic electron transport rate and can be used to estimate J max with the non-rectangular hyperbolic model (von Caemmerer, 2000). Values of both V cmax and J max would be corrected to those at 25 • C after the estimation by using Arrhenius equation provided from Bernacchi et al. (2003). For each leaf, a linear least squares regression of Equations 3 or 4 was used to estimate the intercept and slope parameters of the BWB and MED model, respectively. Biologically, the slope parameter of each model represents the sensitivity of g s to changes in A net , C a , and atmospheric water status and will be the focus of this analysis. A term for the y intercept of each model algorithm (g 0 ) can be used to describe variation in minimum g s . Only simulations that provided a regression between modeled and observed stomatal conductance with an R 2 > 0.8 were included in further analyses (Wolz et al., 2017).
where g s is stomatal conductance (mol m −2 s −1 ), A is the net rate of photosynthetic CO 2 uptake (µmol m −2 s −1 ), h is atmospheric relative humidity (unitless), C a is the atmospheric CO 2 concentration at the leaf surface (µmol mol −1 ), g 0 is the y-axis intercept and m is the slope of the line.
where D is the atmospheric vapor pressure deficit (kPa) and g 1 is the model parameter related to the slope of the line. Previous study showed that whether the value of g 0 is 0 does not significantly affect the linear and non-linear relationship in Equations 3, 4 (Leakey et al., 2006), so we set the value of g 0 to 0 in order to focus on the variations of the stomatal slope. The diurnal measurements of A net and g s were used to validate the coupled FvCB photosynthetic model and BWB and MED stomatal conductance models (Wolz et al., 2017). Specific (cultivar-, CO 2 -and stage-specific) photosynthetic and stomatal parameters and generic parameters (the average values across all the groups) were used to test whether the specific parameters will increase the accuracy of the models.

Statistical Analysis
Four-way analysis of variance (ANOVA) was used to test the fixed effects of years, cultivars, growing stages, CO 2 treatments and their interactions on A net , g s , F v /F m , V cmax , J max , R d and stomatal slope parameters. Post hoc Tukey's HSD tests were conducted on specific contrasts to examine significant treatment effects among groups. General linear models (GLM) were used to assess the relationship between observed and modeled parameters. For all the analysis, the normality of the residuals was tested using the Shapiro-Wilk test. All statistical testes were considered significant at P ≤ 0.05. Mean values of each variable were expressed with their standard error (SE). All analyses were conducted in R with package "plantecophys" and all figures were drawn with the "ggplot" function in package "tidyverse" (R 3.6.3 1 ).

Diurnal and Seasonal Variations in Environmental Parameters
The photosynthetic active radiation (PAR) usually reached its maximum from 10 a.m. to 13 p.m. (Figure 1)

Diurnal and Seasonal Variations in A net and g s
In most cases, A net increased to the maximum around noon and decreased in the afternoon and the maximum value appeared no later than 14 p.m. (Figure 2). The midday depression of A net was observed on some of the measuring dates for both the cultivars and (CO 2 ) treatments. On average, elevated (CO 2 ) significantly stimulated A net of both cultivars by 34.19% in 2019 and 47.65% in 2020 in the whole growing season, which was consistent with the impact of elevated (CO 2 ) on intercellular (CO 2 ) (Supplementary Material 2). The greatest stimulation of elevated (CO 2 ) on A net could be observed around midday at jointing stage (54.35% in 2019 and 56.12% in 2020 for Wuyunjing30; 57.95% in 2019 and 61.50% in 2020 for Yangdao6), and the enhancement of A net of each cultivar due to elevated (CO 2 ) at the following four growing stages were much lower than that at this stage in 2 years (Supplementary Material 3). A net decreased significantly with the advance of growing season and there were significant (CO 2 ) × growing stage effect on A net (Table 1 and Supplementary Material 4). The more CO 2responsive cultivar Yangdao6 had higher A net than Wuyunjin30 and the two cultivars responded similarly to elevated (CO 2 ) (Supplementary Material 5).
The values of g s remained relatively high before 15 p.m. and started to decrease after that (Figure 3). In 2019, g s under all (CO 2 ) × cultivars tended to decrease significantly with the advance of growing season, while a slight recovery of g s beginning at heading stage was observed for Wuyunjing30 under both (CO 2 ) treatments (Supplementary Material 4). In 2020, g s of Yangdao6 under both (CO 2 ) treatments slightly increased from jointing to booting stage, while that of Wuyunjing30 decreased or remained stable during this period, and g s always decreased from heading to maturity stage. Seasonal variations in g s synchronized well with those in A net , especially in 2020. Elevated (CO 2 ) had no significant effect on g s (Table 1 and Supplementary Material 3). The values of g s of cultivar Yangdao6 were higher than those of Wuyunjing30 (Supplementary Material 5).

Diurnal and Seasonal Variations in F v /F m , PSII, and qP
The values of F v /F m , PSII, and qP decreased to the minimum around noon and recovered to normal or even higher levels under each treatment (Figure 4 and Supplementary Materials 6, 7). F v /F m increased with the advance of growing season and reached the maximum at heading stage or grain-filling stage, and decreased at maturity stage in 2019 (Supplementary Material  4). In 2020, F v /F m decreased to the minimum value at booting stage at first, then increased to the maximum value at heading or grain-filling stage, and decreased till the end of the growth period. On average, elevated (CO 2 ) significantly increased F v /F m of both cultivars by 6.92% in 2019 and 6.41% in 2020 across the growth season, while qP was decreased by 4.30% in 2019 and 3.97% in 2020, respectively ( Table 1). The greatest stimulation influences on F v /F m due to elevated (CO 2 ) at midday were 21.39% at heading stage in 2019 and 26.80% at the same stage in 2020 for Wuyunjing30, and 16.18% at maturity stage in 2019 and 29.27% at grain-filling stage in 2020 for Yangdao6 (Supplementary Material 3). The values of F v /F m of cultivar Yangdao6 were significantly higher than those of Wuyunjing30 (Supplementary Material 5).

Seasonal Variations of Photosynthetic and Stomatal Parameters
For most treatments, V cmax and J max reached the maximum at heading stage and decreased to the minimum at maturity stage ( Table 2), which could also be observed in A net -C i curves (Supplementary Material 8). Values of V cmax at booting stage were higher than those at jointing stage only for Yangdao6 in 2019. Values of J max at booting stage were lower than those at jointing stage for Wuyunjing30 grown at ambient and elevated (CO 2 ) and Yangdao6 grown at elevated (CO 2 ) in 2019. The values of R d increased to the maximum at grain-filling or maturity stage. There is no significant influence of (CO 2 ) or cultivar or the interaction on V cmax , J max , and R d ( Table 2), and A net -C i curves showed no differences of three parameters among four (CO 2 ) × cultivar treatments, either (Supplementary Material  9). The values of m and g 1 were highest at maturity stage across (CO 2 ) × cultivar treatments. On average, elevated (CO 2 ) significantly decreased m and g 1 of both cultivars by 7.62 and 9.82% respectively. The values of m and g 1 of more CO 2responsive cultivar Yangdao6 was 10.75 and 10.87% lower than those of less CO 2 -responsive cultivar Wuyunjing30, respectively.

The Photosynthesis and Stomatal Conductance Simulations
The predicted values of A net and g s correlated well with the measured ones across all treatments no matter whether BWB or MED was used or specific or generic model parameters were selected (Figures 5, 6). The square of Pearson correlation coefficient (r 2 ) between predicted and measured values was used to describe the simulation accuracy in our study. We found that the accuracy of photosynthetic simulations was higher when using the specific parameters (r 2 = 0.83) than using the generic ones (r 2 = 0.66). The moderate improvement of g s simulations was achieved using specific parameters (r 2 = 0.45 for BWB and 0.47 for MED model), compared with using generic parameters (r 2 = 0.41 for BWB and 0.37 for MED models).

DISCUSSION
Individual studies and meta-analysis have investigated the general tendency of elevated CO 2 effects on plant physiology and production (Ainsworth and Long, 2005;Ainsworth, 2008;Wang et al., 2012). However, it remains unclear how the effects of elevated CO 2 varied under short-and long-term treatment and across different rice cultivars. In this study, we took the diurnal and A net -C i photosynthetic measurements throughout the growing season for 2 years. Overall, we found that (1) the diurnal variation of net photosynthetic rate was mainly affected by the dynamic of photosynthetic active radiation and air temperature and the midday depression of net photosynthetic rate observed in most cases was caused by both stomatal and

non-stomatal factors;
(2) CO 2 had a positive effect on net photosynthetic rate and the effects decreased after the booting stage and the degree of the acclimation did not vary between the two tested cultivars; (3) the photosynthetic and stomatal conductance models proved to be effective under elevated CO 2 conditions and using specific parameters greatly improved the accuracy of the photosynthetic simulations and moderately improved that of stomatal conductance simulations. In general, values of A net were lower in the early morning and the late afternoon and the maximum occurred around noon. These unimodal or bimodal patterns were consistent with the dynamics of PAR and T air across growing stages. In contrary, although the decrease of g s was synchronous with that of A net in the afternoon, the variation of g s in the morning could not reflect the dynamic of A net in this period. Therefore, light intensity as well as air temperature, rather than stomatal functions, were the dominant factors affecting the diurnal variation of A net on daily scale (Wen et al., 1998). The midday depression of A net was observed in most cases in these 2 years and the synchronous decline of g s could also be found when midday depression of A net occurred, which was consistent with previous studies (Farquhar and Sharkey, 1982;Panda, 2011;Koyama and Takemoto, 2014). Short-term changes of C i should be considered when determining whether stomatal or non-stomatal limitation caused the midday depression (Farquhar and Sharkey, 1982). Stomatal limitation would be the major factor if the reduction of C i and g s was synchronously observed when the midday depression of A net occurred. In contrary, if C i kept at the stable value or even increased, non-stomatal limitation should be taken into account. In our research, for example, rapid decrease of C i could be observed under the situation of photosynthetic midday depression for Wuyunjing30 under both ambient and elevated (CO 2 ) at heading stage in 2019. But C i changed little or remained unchanged, for Wuyunjing30 and Yangdao6 under ambient (CO 2 ) at jointing stage in 2020. Meanwhile, we noticed the decrease of F v /F m and qP around noon for each cultivar at both the (CO 2 ) treatments and they recovered to normal or even higher values at late afternoon, proving the existence of reversible photoinhibition in rice throughout the growth season Demmig-Adams and Adams, 1992). Liu et al. (2020) demonstrated that the midday depression of A net in leaves of Drepanostachyum luodianense was mainly caused by non-stomatal limitation in April, but caused by stomatal limitation in July. And in September, stomatal limitation was the dominant factor leading to the A net depression from 10 a.m. to 12 p.m. while non-stomatal limitation dominated the depression from 12 p.m. to 13 p.m. Therefore, the midday depression of photosynthesis was caused by both the stomatal and nonstomatal limitations such as photoinhibition, and the dominant factor changed depending on the surrounding environment. While the dynamics of major environmental factors such as PAR and air temperature affected the overall trend of the diurnal variation of rice photosynthesis, stomatal closure and photoprotection of plants responding to high light intensity and other non-stomatal limitations were the dominant factors affecting the midday decrease of the photosynthesis.
The stimulation of elevated (CO 2 ) on plant photosynthesis has been widely reported in recent decades, but the magnitude of the stimulation varied greatly depending on species, plant functional types (PFTs) and other environmental factors (Ainsworth and Long, 2005;Wang et al., 2012). In this study, elevated (CO 2 ) enhanced A net of rice by 34.19 and 47.65% on average throughout the growing season in 2019 and 2020, respectively, which was comparable to what was reported 33% increase in the metaanalysis studies (Wand et al., 1999;Ainsworth and Long, 2005). Higher intercellular CO 2 concentration (C i ) and photochemical efficiency of PSII (F v /F m ) were the main factors for the stimulation on A net of rice at elevated (CO 2 ). Photosynthetic acclimation to elevated (CO 2 ) was found in crop varieties with limited sink size (Thilakarathne et al., 2015;Ruiz-Vera et al., 2017), but not in those with many large leaves (Ruiz-Vera et al., 2017) or high tillering potentials . The enhancement of A net due to elevated (CO 2 ) was greatest at jointing stage in 2 years, but was weakened afterward and even disappeared at maturity stage in 2020, which suggested the occurrence of photosynthetic acclimation at elevated (CO 2 ) and it could happen at the early growing season. We did not observe the less enhancement by elevated (CO 2 ) on F v /F m and qP at the following growing stages compared with jointing stage or any significant influence of elevated (CO 2 ) on g s , V cmax and J max which could explain the early season photosynthetic acclimation. Therefore, the down-regulation of stomatal and Rubisco enzyme functions might not be the only reasons for the photosynthetic acclimation of rice at elevated (CO 2 ) (Chen et al., 2005). A more frequent and detailed measurement in the early growing season was required to fully understand the evident but decreased elevated CO 2 effect on the net photosynthesis. The elevated CO 2 effect on rice roots growth, morphology and physiology was also under investigation in this study to further discover whether the photosynthetic acclimation of rice was the added-up effects FIGURE 6 | Relationships between the measured and simulated daily average stomatal conductance for the two rice cultivars at ambient and elevated (CO 2 ) throughout the growing season. Stomatal conductance was simulated with BWB (A,B) or MED models (C,D), using either specific model parameters (A,C) or generic model parameters (B,D). p value of the correlation analysis and the square of Pearson correlation coefficient (r 2 ) are given.
from both the above-and below-ground physiological changes. Besides, there was no significant difference of the photosynthetic acclimation between more CO 2 -responsive cultivar Yangdao6 and less CO 2 -responsive cultivar Wuyunjing30 in terms of the occurrence and the extent of the acclimation, even though higher A net and g s were observed in Yangdao6 compared with Wuyunjing30 throughout the growing season, which was consistent with previous studies (Zhu et al., 2015).
Model parameters in FvCB model (V cmax , J max , and R d ) and stomatal slope parameters in BWB model (m) and MED model (g 1 ) varied across growing stages. The decrease of V cmax and J max could be the reason for the decline of photosynthetic capacity with the advance of growth period. V cmax , J max , and R d varied not only among different species (Wullschleger, 1993), but also within single species in different environments (Bunce, 2000;Medlyn et al., 2002a;Medlyn et al., 2002b). Elevated (CO 2 ) did not lead to significant changes of V cmax and J max in our research. The response of V cmax , J max , and R d in C 3 plants grown at elevated (CO 2 ) were reported to be quite different and both stimulation and inhibition of elevated (CO 2 ) effects had been reported in recent years (Sims et al., 1998a;Sims et al., 1998b;Ellsworth et al., 2004). Elevated (CO 2 ) treatment as short as 7 days increased V cmax and J max for cucumber by 12.0 and 14.7% without drought stresses, but the enhancement was no longer significant under mild and severe drought stresses . From our ANOVA analysis, V cmax , J max , and R d of rice did not vary significantly across cultivars, suggesting that photosynthetic simulations could use similar parameters at least between the two tested cultivars. For stomatal conductance models, previous research has reported significant correlation between stomatal slope parameters and environmental factors such as temperature, moisture and season and plant physiological factors such as wood density (Lin et al., 2015). And recent research has also indicated the variability of stomatal slopes in plants at species level (Wolz et al., 2017). FACE research reported no stomatal acclimation for soybean at elevated (CO 2 ) (Leakey et al., 2009). However, the Ball-Berry slope was significantly different between ambient and elevated (CO 2 ) grown wheat (Tausz-Posch et al., 2013). In our research, stomatal slopes (m and g 1 ) varied across both (CO 2 ) treatments and cultivars, demonstrating that variability of m and g 1 was caused not only by the diversity of plant function groups but also by the diversity of physiological characteristics among cultivars within a single species.
Higher accuracy of simulation using specific model parameters in both photosynthesis model (V cmax , J max , and R d ) and stomatal conductance model (m and g 1 ) compared with using generic parameters was achieved in this study, suggesting that future simulations in large scale carbon and water cycles should take account of the variations of model parameters across environmental and non-environmental factors. Though we did not observe a significant impact of elevated (CO 2 ) on V cmax and J max , it should be noted that not considering the impact of mesophyll conductance (g m ) might lead to the potential increased systematic errors of determining V cmax and J max (Singsaas et al., 2003). And the possibility that the differences of model parameters of rice will change under other environmental conditions should not be ruled out too. Therefore, further research focusing on the diversity of model parameters are needed to confirm the existence of (CO 2 ), cultivar and growing stage variation of photosynthesis and stomatal conductance model parameters of rice and other plant species.
In conclusion, diurnal variations of net photosynthetic rate of rice showed unimodal or bimodal patterns, which was mainly influenced by light intensity and air temperature. Meanwhile, the closure of stomata and the photoinhibition around noon were dominant factors for the short-term midday depression of net photosynthetic rate. Elevated (CO 2 ) greatly increased net photosynthetic rate at jointing stage. This stimulation was acclimated with the advance of growing season and was not affected by either stomatal limitation or Rubisco activity. Model parameters in photosynthesis model (V cmax , J max , and R d ) and two stomatal conductance models (m and g 1 ) varied across growing stages and m and g 1 also varied across (CO 2 ) treatments and cultivars, which led to more accurate photosynthesis and stomatal conductance simulations when using these specific parameters. The results in the study suggested that further researches are still needed to investigate the dominant factors contributing to the acclimation of photosynthetic capacity under future elevated CO 2 conditions. The study also highlighted the need of investigating the impact of other environmental, such as nitrogen and O 3 , and non-environmental factors, such as additional rice cultivars, on the variations of the model parameters in photosynthesis and stomatal conductance models and the further impacts on simulations in large scale carbon and water cycles.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

AUTHOR CONTRIBUTIONS
DW conceived the idea and led the study. YM and YC designed the experiment. YM, YC, and HW measured all the data used for this study. YM analyzed the data and wrote the manuscript with the critical suggestions by DW. All authors contributed to the article and approved the submitted version.