論文ID: D-24-00027
In greenhouse cultivation, CO2 enrichment enhances the net photosynthetic rate (Pn) of the plant canopy. Null balance CO2 enrichment (NB-CE) maintains equal CO2 concentrations inside and outside the greenhouse (Cin and Cout), limiting CO2 leakage; however, there may be more cost-effective strategies. This study proposes an alternative strategy termed CO2 enrichment only for high photosynthetic photon flux density (PPFD) hours (HP-CE). In this strategy, Cin is maintained slightly above Cout when Pn is expected to be high, aiming to improve cost-effectiveness. We defined the cost-effectiveness of CO2 enrichment as the increase in integrated Pn over a photoperiod compared to non-CO2 enrichment, divided by the CO2 supply rate (S) integrated over the same period. Since CO2 leakage increases with the number of air exchanges per hour (N), we evaluated the cost-effectiveness of HP-CE at several N levels. In this study, the integrated S of HP-CE was set equivalent to that of NB-CE, and the CO2 supply period of HP-CE was adjusted based on the target Cin (450-700 µmol mol-1) at N of 4, 6, 8, and 10 h-1. Using an open-type assimilation chamber to reproduce ventilated greenhouse environments, we measured the PPFD-Pn and PPFD-S curves of cucumber seedlings. We then estimated the Pn and S over a 12-hour photoperiod for NB-CE and several HP-CE scenarios. The estimation revealed that HP-CE could be more cost-effective than NB-CE for Cin 450-700 µmol mol-1 at N of 4, 6, and 8 h-1. HP-CE was less cost-effective than NB-CE at N of 10 h-1 due to significant CO2 leakage at increased target Cin. These findings suggest that for ventilated greenhouses with relatively low N, CO2 enrichment can be more cost-effective than NB-CE by slightly elevating Cin above Cout only for high PPFD hours.
CO2 concentration control is essential for increasing crop growth, yield, and quality in greenhouses (Mortensen, 1987; Wang et al., 2022). Ventilation to prevent an excessive increase in air temperature in a greenhouse also contributes to the supply of atmospheric CO2 to the greenhouse. However, the CO2 concentration inside a greenhouse with a dense canopy drops below the atmospheric concentration, even with ventilation, thereby reducing crop yield (Sánchez-Guerrero et al., 2000; 2005; Vox et al., 2010). Accordingly, even in ventilated greenhouses, CO2 enrichment is necessary to increase the net photosynthetic rate (Pn) of plants (Mortensen, 1987; Osuka, 2003).
Slack and Hand (1985) reported a 22% increase in cucumber yield when CO2 concentration inside the greenhouse was elevated to 450 µmol mol-1 for a photoperiod, compared to no CO2 enrichment in a ventilated greenhouse. Kawashiro et al. (2009) compared cucumber yields under CO2 enrichment at 500 µmol mol-1 for 7 h after sunrise and at 1000 µmol mol-1 for 3 h after sunrise (until ventilation started). They observed that the cucumber yield was 15% higher than the former method, even though the CO2 supply rate (S, in mol s-1) integrated over a day was low. These CO2 enrichment strategies, which elevate the CO2 concentration inside the greenhouse to a level equivalent to or slightly higher than the atmospheric concentration, are called partial CO2 enrichment (Hand, 1984).
During ventilation, if the CO2 concentration inside the greenhouse exceeds atmospheric levels, CO2 is inevitably released outside the greenhouse due to a concentration gradient (Kläring et al., 2007). Therefore, null balance CO2 enrichment (NB-CE), a strategy of partial CO2 enrichment, has been proposed to dynamically control the S in a ventilated greenhouse, maintaining the CO2 concentration inside the greenhouse close to the atmospheric level (Ohyama et al., 2005; Kozai et al., 2015). In case of the NB-CE, S becomes equal to the CO2 absorption rate by the Pn of plants (Ohyama et al., 2005). NB-CE reportedly ensures that the supplied CO2 is not released outside the greenhouse and is often described as advantageous in terms of cost-effectiveness.
Generally, the Pn of plants increases with increasing photosynthetic photon flux density (PPFD) up to the light saturation point. Moreover, light-saturated Pn increases with higher CO2 concentrations (Pan et al., 2020). However, because NB-CE maintains the CO2 concentration inside the greenhouse at approximately 400 µmol mol-1 (Ohyama et al., 2005), the high PPFD on sunny days cannot be fully utilized, limiting the expected enhancement of Pn due to CO2 enrichment. On the other hand, when the CO2 concentration inside the greenhouse is elevated slightly above atmospheric levels only for high PPFD hours on sunny days, the increase in Pn through CO2 enrichment is expected to surpass that achieved with NB-CE, even if some of the supplied CO2 leaks outside. Therefore, we hypothesized that a CO2 enrichment strategy may be developed that is more cost-effective than NB-CE.
The proposed CO2 enrichment strategy, termed CO2 enrichment only for high PPFD hours (HP-CE), is designed to elevate the CO2 concentration inside the greenhouse slightly above atmospheric levels only for high PPFD hours when the Pn is expected to be high. For HP-CE, CO2 is not supplied into the greenhouse for a few hours immediately after sunrise or before sunset when Pn is expected to be low because of lower PPFD. Specifically, this strategy requires determining the CO2 enrichment conditions, such as the target CO2 concentration inside the greenhouse and the CO2 supply period, according to the number of air exchanges per hour (N) since CO2 leakage increases with higher N.
Technically, to evaluate the cost-effectiveness of CO2 enrichment, the increased yield of harvested crops should be compared to the costs associated with CO2 enrichment. However, these values, including market prices, fluctuate significantly depending on factors such as season and market demand, making them difficult to quantify. Instead, in this study, we considered it reasonable to adopt the integrated S over a photoperiod as a cost parameter and the increase in the integrated Pn over a photoperiod due to CO2 enrichment compared to non-CO2 enrichment as an effectiveness parameter. Consequently, the increase in the integrated Pn over a photoperiod due to CO2 enrichment compared to non-CO2 enrichment, divided by the integrated S over the same period, was defined as CO2 enrichment efficiency (%). This value was used as an index to evaluate the cost-effectiveness of CO2 enrichment.
In this study, we aimed to demonstrate a proof-of-concept of HP-CE as a more cost-effective strategy for CO2 enrichment than the conventional NB-CE at the laboratory scale. For HP-CE to be as effective as NB-CE at the same cost, the integrated S over a photoperiod for HP-CE must be equal to that of NB-CE. Therefore, the CO2 supply period for HP-CE was adjusted according to the target CO2 concentration inside the greenhouse, ranging from 450 to 700 µmol mol-1. Considering that CO2 leakage increases with higher N, we determined the suitable target CO2 concentration inside the greenhouse and CO2 supply period for HP-CE across various N levels (4, 6, 8, and 10 h-1). To achieve this, we utilized a laboratory-scale model greenhouse equipped with an open-type assimilation chamber (referred to as a model greenhouse), capable of simulating the environmental factors of a greenhouse with forced ventilation. This model greenhouse allowed us to theoretically evaluate the cost-effectiveness of different CO2 enrichment strategies across varying N levels.
Cucumber (Cucumis sativus L., ‘Hokushin,’ Takii & Co., Ltd., Kyoto, Japan.) seeds were sown on moistened tissue paper and placed in the dark at 25°C in a temperature-controlled growth chamber (MIR-554; SANYO Electric Co., Ltd., Osaka, Japan) equipped with a white light emitting diode (LED) panel (HMW120DC6; Stanley Electric Co., Ltd., Tokyo, Japan). After three days of sowing, cucumber seedlings were transplanted into urethane sponges in a cut 200-cell plug tray placed in a plastic container containing a nutrient solution (prescription A; OAT Agrio Co., Ltd., Tokyo, Japan) at an electrical conductivity of 0.13 S m-1 and a pH of 6.0. The seedlings were grown at an air temperature of 25°C/18°C (photo/dark) and PPFD 300 μmol m-2 s-1 at the top of the plant for the photoperiod of 16 h d-1. The air inside the growth chamber was ventilated using an air pump to introduce external air and maintain an internal CO2 concentration close to atmospheric levels. Ten days after sowing, four cucumber seedlings were transplanted into a hydroponic container containing 5 L of the same nutrient solution. The nutrient solution was aerated using an air pump to ensure adequate air supply to the roots. At 21 days after sowing, four cucumber seedlings (Fig. 1A) were moved to the model greenhouse (Fig. 1B), and Pn was measured to obtain the PPFD-Pn curves.
2.2. Laboratory-scale model greenhouse Structure of the model greenhouseThe Pn of cucumber seedlings was measured using the model greenhouse equipped with an open-type assimilation chamber, reproducing environmental factors of a forced-ventilated greenhouse, including air temperature, PPFD, and N (Fig. 1B). This model greenhouse was composed of a growth chamber (MIR-554; SANYO Electric Co., Ltd., Osaka, Japan), a white LED panel (ENB01-NHSD7-F1; Toyoda Gosei Co., Ltd., Aichi, Japan) (manufactured by Kuji Seisakusho), an open-type assimilation chamber, CO2 concentration control systems for air entering and inside the open-type assimilation chamber, and a PPFD control system. The open-type assimilation chamber inside the growth chamber was divided into top and bottom spaces for the shoots and roots of the plants. The shoot space was referred to as a cultivation container in this study. An LED panel was installed at the top of the cultivation container. Two AC fans were installed on its sides to prevent excessive temperature increase in the region surrounding the LED panel. In addition, various devices related to the CO2 concentration and PPFD control systems were installed on the exterior of the growth chamber, as described below.
Fig. 1. Cultivation container (A) as a laboratory-scale model greenhouse (B) to reproduce environmental conditions of a greenhouse with forced ventilation, such as air temperature, photosynthetic photon flux density (PPFD), and the number of air exchanges per hour (N). The cultivation container was placed inside a growth chamber.
The cultivation container (55.5 L) was acrylic and equipped with an air inlet and outlet. To replicate the forced ventilation greenhouse, atmospheric air was drawn into the cultivation container using an air pump and forced out through the outlet port. The root space under the cultivation container contained a nutrient solution that was aerated to provide oxygen to the roots. Because the air exchange between the shoots and roots of the cultivation container was considered too small, this air exchange was disregarded in this experiment. DC fans were installed inside the cultivation container to achieve a uniform CO2 distribution. Temperature and humidity sensors (THT-B4T; Shin-ei Technology Co., Ltd., Tokyo, Japan) were installed to measure the air temperature and relative humidity inside the cultivation container. The parameters measured using these sensors were recorded on a data logger (midi LOGGER GL800; GraphTech Co., Ltd., Tokyo, Japan) at 1-s intervals. All output voltages from the sensors related to the model greenhouse, such as temperature, humidity, and CO2 concentration, were recorded using the same data logger.
Control and measurement of Cin and CoutA schematic of the experimental setup for determining CO2 enrichment efficiency is shown in Fig. 2. An air pump (MX808-W; Tokyo Electric Co., Ltd., Tokyo, Japan) supplied the atmosphere (laboratory air) to the cultivation container, simulating a forced ventilated greenhouse. Water vapor and CO2 were removed from the atmosphere using desiccants and CO2 absorbents, respectively. The airflow rates of the water vapor- and CO2-free atmosphere and the 100% CO2 gas in the cylinder were controlled by two mass flow controllers (8500MC; KOFLOC Corp., Kyoto, Japan) and then mixed to achieve an atmospheric CO2 concentration of 400 µmol mol-1. This CO2 concentration is referred to as the CO2 concentration outside the cultivation container (Cout). A portion of this mixed air was passed through an infrared gas analyzer (ZRH; Fuji Electric Co. Ltd., Tokyo, Japan) using a mass flow meter (3810DS; KOFLOC Corp.) to measure Cout. The remaining mixed air flowed into the cultivation container, and the N was determined by measuring the airflow rate of entering air using a mass flow meter (3810DS; KOFLOC Corp.). The N of the cultivation container was controlled by adjusting two mass flow controllers for the water vapor- and CO2-free atmosphere and the 100% CO2 gas.
Fig. 2. Schematic diagram of the experimental set-up for determining the CO2 enrichment efficiency for a laboratory-scale model greenhouse with various numbers of air exchanges. MFC: Mass flow controller; MCM: Mass flow meter; Cin and Cout: CO2 concentration inside and outside the cultivation container, respectively.
To replicate the CO2 enrichment system of the greenhouse, 100% CO2 gas was supplied to the cultivation container from another CO2 cylinder via a mass flow controller (8500MC; KOFLOC Corp.). The CO2 concentration inside the cultivation container (Cin) was controlled by adjusting the S using a PID controller (E5CC-CX2ASM-007; Omron Corp., Kyoto, Japan). To measure Cin, the air inside the cultivation container was pumped using an air pump (LVM-10), and the water vapor was removed using a Peltier cooler module (DH-209; Komatsu Electronics Co., Komatsu, Japan). Subsequently, Cin was measured using a CO2 analyzer (ZRH; Fuji Electric Co., Ltd.) after passing through a mass flow meter (3810DS; KOFLOC Corp.).
In this model greenhouse, Cout served as the baseline for establishing the setpoint of the PID controller. The process variable of the PID controller represented the actual Cin resulting from CO2 enrichment. For CO2 enrichment strategies, the setpoint of the PID controller was configured to 50 µmol mol-1, 100 µmol mol-1, or a specified amount higher than Cout. This adjustment ensured that Cin was maintained at a level higher than Cout. In the case of NB-CE, the setpoint of the PID controller was set to zero, so that Cin was equal to Cout.
Control of PPFDA DC power supply unit (PAS 80-9; Kikusui Electronics Corp., Yokohama, Japan) applied current to the LED panel for PPFD control of the leaves of the cucumber seedlings. A voltage generator consisting of a digital-to-analog converter (MCP4922-E/P; Microchip Technology, Chandler, AZ, USA), an operational amplifier (NJM13404; Japan Radio Corp., Tokyo, Japan), and an Arduino (Uno R3) was employed to control the current applied to the LED panel based on the external analog voltage controlled via a PC.
2.3. Determining the HP-CE conditionsTarget Cin for HP-CE was set at 50, 100, 200, and 300 μmol mol-1 higher than the NB-CE (400 μmol mol-1). These HP-CE strategies were referred to as +50HP-CE, +100HP-CE, +200HP-CE, and +300HP-CE, respectively. For each HP-CE with the same integrated S as NB-CE, we estimated the CO2 supply period based on the PPFD-Pn and PPFD-S curves of cucumber seedlings at each target Cin. The CO2 supply period for each HP-CE strategy was determined according to N levels of 4, 6, 8, and 10 h-1. To evaluate the cost-effectiveness of all CO2 enrichments, the photoperiod was set to 12 hours, and integrated Pn and S were calculated over a 12-hour photoperiod.
2.3.1. PPFD-Pn and PPFD-S curves Pn measurement conditionsBefore measuring Pn to obtain the PPFD-Pn curves, the four cucumber seedlings were acclimated to the environment of the model greenhouse at a PPFD of 300 µmol m-2 s-1 for 3 h.
Pn was then measured at PPFD values of 0, 100, 200, 300, 400, 500, 600, and 700 µmol m-2 s-1 (Supplemental Fig. 1A). The Cin for CO2 enrichment was set at 400 (reproduced NB-CE), 450, 500, 600, and 700 µmol mol-1 (referred to as Cin-400, Cin-450, Cin-500, Cin-600, and Cin-700, respectively) at N levels of 4, 6, 8, and 10 h-1. Additionally, non-CO2 enrichment conditions were tested in which CO2 gas was not supplied (referred to as non-CE). Supplemental Fig. 1B-G shows the recorded values of Cin and Cout during Pn measurements; although N levels of 4, 6, 8, and 10 h-1 were evaluated, only measurements at 8 h-1 are shown owing to the extensive amount of data. Different plant groups (one group of four cucumber seedlings) were used when measuring Pn at different N levels, and Pn at each PPFD point was measured in 2-4 replicates.
The air temperature inside the cultivation container was controlled indirectly by adjusting the temperature of the growth chamber. To simulate typical diurnal changes of air temperature and sunlight PPFD in an actual greenhouse, the temperature of the growth chamber was set to 20°C (PPFD 0 to 199 µmol m-2 s-1), 21°C (PPFD 200 to 399 µmol m-2 s-1), 22°C (PPFD 400 to 599 µmol m-2 s-1), and 23°C (PPFD 600 to 700 µmol m-2 s-1). The set temperature was switched five minutes before changing the PPFD to 0, 200, 400, and 600 µmol m-2 s-1. When the set temperature of the growth chamber was 23°C, the actual air temperature inside the cultivation container was approximately 28°C.
Pn calculationAfter setting the PPFD, four cucumber seedlings were acclimated to the PPFD for 35 minutes, and then Pn measurements were conducted for 10 minutes. The Pn of the cucumber seedlings was calculated based on the CO2 balance of the cultivation container, assuming that the CO2 concentration within the cultivation container was spatially uniform. The temporal variation of Cin (dCin/dt) during the Pn measurement was considered negligible and ignored, and Pn [mol s-1] was calculated using the following equation:
where N is the number of air exchanges per hour [h-1]; V is the volume of the cultivation container [mol], Cout is the CO2 concentration outside the cultivation container [mol mol-1]; Cin is the CO2 concentration inside the model cultivation container [mol mol-1]; and S is the CO2 supply rate [mol s-1]. The values of each parameter mentioned above were measured at 1-second intervals and averaged over 10 minutes.
Because different cucumber seedlings were used in each measurement plot, Pn was measured under the same condition after generating one plot, and that value of Pn was used for calibration. That condition was at PPFD 300 µmol m-2 s-1 without CO2 enrichment.
where Pn′ is the calibrated Pn [mol s-1]; A is Pn measured under the calibration condition [mol s-1]; Ā is the average A for all plots [mol s-1].
PPFD-Pn curveUnder each N and target Cin, PPFD-Pn curves were fitted by a non-rectangular hyperbolic function proposed by Marshall and Biscoe (1980). The model is as follows:
where ϴ and Pmax are the convexity of the light response curve and the maximum Pn′, respectively; Ф is the apparent quantum yield, equal to the initial slope of the linear regression; І is the PPFD on the leaf surface; and Rd is the dark respiration rate.
PPFD-S curveWhen measuring Pn under each condition, the S, which allows Cin to reach the target level for CO2 enrichment, was also measured. Each S value was calibrated using the same parameters used for calibrating the Pn, as shown in Equation (4):
where S′ is the calibrated S [mol s-1]; A is Pn measured under the calibration condition [mol s-1]; Ā is the average A for all plots [mol s-1]. The recorded S′ values at N level of 8 h-1 are presented in Supplemental Fig. 1H-L. Subsequently, the PPFD-S curve was also fitted in the same manner as the PPFD-Pn curve.
2.3.2. Determining the CO2 supply period for each HSEFig. 3 shows the time course of PPFD over a 12-hour photoperiod, which was assumed to follow a sinusoidal curve with a peak value of 700 µmol m-2 s-1. First, using the PPFD-Pn and PPFD-S curves of cucumber seedlings, Pn and S were estimated for each time point in Fig. 3 for various target Cin and N. Estimated Pn and S are referred to as PnE and SE, respectively. Next, the integrated SE for NB-CE, the standard for all HP-CEs in this study, was calculated at each N level.
Based on the SE values over the 12-hour photoperiod, the CO2 supply periods of +50HP-CE, +100HP-CE, +200HP-CE, and +300HP-CE, centered around the peak PPFD time, were determined to ensure the integrated SE matched that of NB-CE at each N level. As the next step, for each HP-CE scenario, the PnE value at each target Cin was used during the determined CO2 supply period, while the PnE value for non-CE was used for the remaining period when CO2 was not supplied.
Through the above processes, it was possible to estimate the time course of PnE during the 12-hour photoperiod of +50HP-CE, +100HP-CE, +200HP-CE, and +300HP-CE.
Fig. 3. Time course of photosynthetic photon flux density (PPFD) above cucumber seedlings over a 12-hour photoperiod. The time course follows a sinusoidal curve with a maximum PPFD of 700 μmol m-2 s-1.
The cost-effectiveness of CO2 enrichment was defined as the CO2 enrichment efficiency (η), as expressed in Equation (5).
In Equation (5), ∫ ∆PnE dt is the increase in integrated PnE over a photoperiod when CO2 enrichment is applied compared with that when CO2 enrichment is not applied [mmol d-1]; ∫ SE dt is integrated SE over a photoperiod [mmol d-1].
PPFD-Pn curves for the four cucumber seedlings were generated at Cin-400, Cin-450, Cin-500, Cin-600, and Cin-700 at N levels of 4, 6, 8, and 10 h-1 (Fig. 4). Additionally, to calculate ΔPnE (the increase in integrated PnE due to enrichment compared to non-CE), PPFD-Pn curves were also generated without CO2 enrichment (non-CE) at each N level. PPFD-S curves for each condition were also generated. In the model greenhouse replicating the CO2 enrichment system of a ventilated greenhouse, the calculation of Pn (as shown in Equation 2) is significantly influenced by the value of S required to increase Cin, which is decreased by cucumber seedling photosynthesis, to the target Cin. Consequently, the PPFD-Pn and PPFD-S curves exhibited similar trends (Supplemental Fig. 2).
Fig. 4. Photosynthetic photon flux density (PPFD)-net photosynthetic rate (Pn) curves for four cucumber seedlings placed in the cultivation container (laboratory-scale model greenhouse). Pn was measured at PPFD values of 0, 100, 200, 300, 400, 500, 600, and 700 µmol m-2 s-1, and the curve was fitted by a nonrectangular hyperbola model. The number of air exchanges per hour (N) was set to 4 (A), 6 (B), 8 (C), and 10 h-1 (D). Under each N, Pn of four cucumber seedlings was determined under five CO2 concentrations in the cultivation container (Cin): 400 (null balance CO2 enrichment, Cin-400), 450 (Cin-450), 500 (Cin-500), 600 (Cin-600), and 700 (Cin-700) µmol mol-1. The condition without CO2 enrichment is referred to as non-CE. Pn values were calibrated using Equation (2), as different cucumber seedlings were used in each measurement plot, and the calibrated Pn is referred to as Pn′.
Using the PPFD-S curves for Cin-400, we calculated the time courses of SE over the 12-hour photoperiod for NB-CE at each N level (Fig. 5A) to determine the integrated SE for HP-CE. Also, using the PPFD-Pn curves for non-CE and Cin-400, the time courses of PnE over the 12-hour photoperiod for non-CE (Fig. 5B) and NB-CE (Fig. 5C) were calculated, respectively, at each N level. During the 12-hour photoperiod, for both NB-CE and non-CE, the maximum Pn value at peak PPFD of 700 µmol m-2 s-1 tended to increase as N increased. Thongbai et al. (2010) reported that increasing air circulation increases the Pn of tomato seedlings, similar to the effect of increasing CO2 concentrations. In the NB-CE, as the N of the cultivation container increased, the internal air circulation also increased, leading to a higher Pn of the cucumber seedlings. In the case of the non-CE, sufficient CO2 for photosynthesis was supplied through ventilation, explaining the increase in Pn with higher N. However, since the PnE of non-CE was still lower than that of NB-CE regardless of N level, CO2 enrichment appears necessary even in a ventilated greenhouse with high N.
Fig. 5. Time courses of the estimated CO2 supply rate (SE) for null balance CO2 enrichment (NB-CE; A) and estimated net photosynthetic rate (PnE) for non-enrichment (non-CE; B) and NB-CE (C) over a 12-hour photoperiod. The number of air exchanges per hour was set to 4, 6, 8, and 10 h-1. SE and PnE were estimated using the photosynthetic photon flux density (PPFD)-Pn curves in Fig. 4, with the PPFD set as shown in Fig. 3.
Figure 6 shows the integrated SE for NB-CE calculated using data from Fig. 5A. In the NB-CE, the integrated SE over the 12-hour photoperiod, representing the amount of CO2 supplied, gradually increased as N increased (Fig. 6). The integrated SE at N of 10 h-1 was 30% higher than that at 4 h-1. This increase might be due to higher Pn with increasing air circulation, but it could also result from increased CO2 leakage at a high N level (Mortensen, 1987).
Fig. 6. Estimated CO2 supply rate (SE) integrated over a 12-hour photoperiod when applying the null balance CO2 enrichment. Values were calculated using data shown in Fig. 5C.
Using the PPFD-Pn and PPFD-S curves for Cin-450, Cin-500, Cin-600, and Cin-700, the PnE and SE time courses over the 12-hour photoperiod were estimated, assuming that Cin was elevated to each target level over the 12-hour photoperiod (Supplemental Fig. 3). Based on these estimations, CO2 supply periods for +50HP-CE, +100HP-CE, +200HP-CE, and +300HP-CE were determined to ensure the integrated SE matched that of NB-CE at each N level (Fig. 7A-D). As a result, the CO2 supply periods for +50HP-CE, +100HP-CE, +200HP-CE, and +300HP-CE were approximately 7 h, 6 h, 5 h, and 4 h, respectively, regardless of N levels. For the PnE time course of HP-CE, the PnE values under non-CE were applied during a period when CO2 enrichment was not activated, and the PnE values obtained from Supplemental Fig. 3E-H were applied during periods when CO2 enrichment was activated. Accordingly, the PnE time courses over the 12-hour photoperiod for +50HP-CE, +100HP-CE, +200HP-CE, and +300HP-CE were estimated in accordance with the SE time courses, as shown in Fig. 7E-H.
Fig. 7. Time courses of the estimated CO2 supply rate (SE; A, B, C, and D) and estimated net photosynthetic rate (PnE; E, F, G, and H) for a null balance CO2 enrichment (NB-CE) and CO2 enrichment only for high photosynthetic photon flux density (PPFD) times (HP-CE) over a 12-hour photoperiod. The number of air exchanges per hour were set at 4 (A and E), 6 (B and F), 8 (C and G), and 10 h-1 (D and H). The CO2 concentration inside the cultivation container (laboratory-scale model greenhouse) for NB-CE was maintained at 400 µmol mol-1, while for HP-CE it was set at 50 (+50HP-CE), 100 (+100HP-CE), 200 (+200HP-CE), and 300 μmol mol-1 (+300HP-CE) higher than the NB-CE. SE and PnE were estimated using the PPFD- Pn curves in Fig. 4, with the PPFD set as shown in Fig. 3.
Using data from Fig. 7, the integrated PnE, ΔPnE, and CO2 enrichment efficiency were calculated (Fig. 8). In all CO2 enrichment scenarios, except for non-CE, although the integrated SE was the same, the integrated PnE varied depending on the target Cin of CO2 enrichment and N (Fig. 8A-D). At N levels of 4, 6, and 8 h-1, the integrated PnE of the HP-CE was higher than that of NB-CE, with the maximum value of integrated PnE observed for +100HP-CE. For CO2 enrichment in ventilated greenhouses, if the total amount of CO2 supplied is the same, the increase in Pn is significantly greater in the afternoon with higher PPFD than in the morning with lower PPFD (Hao et al., 2008). Notably, NB-CE is a low-environmental-impact technology that prevents CO2 leakage outside the greenhouse because almost all supplied CO2 is absorbed by plants through photosynthesis (Ohyama et al., 2005). Consistent with this, our results showed that for NB-CE, the integrated SE and PnE were the same at all N levels. In contrast, for HP-CE, at each N level except 10 h-1, the integrated PnE was higher than the integrated SE, suggesting that there was no leakage of supplied CO2 to the outside of the cultivation container. Additionally, external CO2 introduced through ventilation for low PPFD hours, when CO2 enrichment is not activated, was likely absorbed through photosynthesis. Therefore, despite having the same integrated SE, our results indicate the HP-CE, which elevates Cin by 50-300 µmol mol-1 above Cout only for high PPFD hours, effectively improves the integrated PnE over that of NB-CE. Moreover, HP-CE can be considered a low-environmental-impact technology, as it absorbs external CO2 without leaking supplied CO2 to the outside, thereby increasing integrated Pn compared to NB-CE while maintaining environmental sustainability.
Fig. 8. Integrated estimated net photosynthetic rate (PnE) for non-enrichment (non-CE) and several CO2 enrichment conditions, with the number of air exchanges per hour (N) set to 4 (A), 6 (B), 8 (C), and 10 h-1 (D). The difference in integrated PnE between non-CE and each CO2 enrichment (ΔPnE; E, F, G, and H) and CO2 enrichment efficiency for each CO2 enrichment (I, J, K, and L) were calculated at N values of 4 (E and I), 6 (F and J), 8 (G and K), and 10 h-1 (H and L). The CO2 concentration inside the cultivation container (laboratory-scale model greenhouse) for null balance CO2 enrichment (NB-CE) was maintained at 400 µmol mol-1, while the levels for CO2 enrichment only for high photosynthetic photon flux density (PPFD) times (HP-CE) were set at 50 (+50HP-CE), 100 (+100HP-CE), 200 (+200HP-CE), and 300 μmol mol-1 (+300HP-CE) higher than the NB-CE.
At N of 10 h-1, no HP-CE conditions resulted in an integrated PnE higher than that for NB-CE conditions. In ventilated greenhouses with high N, the rate of CO2 outflow increases due to the CO2 concentration gradient between the inside and outside of the greenhouse. Wang et al. (2022) emphasized the need for precise control strategies in CO2 distribution and concentration maintenance to mitigate the negative effects of high N on the CO2 concentration inside the greenhouse.
For non-CE and all CO2 enrichment conditions, the integrated PnE tended to increase with increasing N. The increase in Pn with higher N levels could be attributed to improved internal air circulation, which enhances CO2 distribution within the cultivation container (Thongbai et al., 2010). Therefore, the effect of CO2 enrichment on Pn might vary depending on the N levels in the greenhouse.
Since the integrated SE was consistent across all CO2 enrichment scenarios, the CO2 enrichment efficiency mirrored the trend of ΔPnE (Fig. 8E-L). At each N level, except for 10 h-1, the CO2 enrichment efficiencies of all HP-CE scenarios were higher than that of NB-CE. Among HP-CE scenarios, +100HP-CE had the highest CO2 enrichment efficiency, 14-19% higher than that of NB-CE. This result indicates that the +100HP-CE, which elevated Cin to 500 µmol mol-1 only for high PPFD times, achieved higher CO2 enrichment efficiency compared to NB-CE, which maintained Cin at 400 µmol mol-1 throughout the photoperiod, despite both having the same integrated SE. However, for +200HP-CE and +300HP-CE with higher target Cin, the CO2 enrichment efficiency tended to be lower than +100HP-CE due to increased CO2 leakage. This highlights the need to carefully determine the target concentration and CO2 supply period for CO2 enrichment to balance the benefits against potential losses due to leakage.
The CO2 enrichment efficiency decreased as N increased, regardless of the type of CO2 enrichment. For all CO2 enrichments, the average CO2 enrichment efficiency at N of 10 h-1 was approximately 35% lower than that at N of 4 h-1. This trend occurred because higher N levels provided sufficient CO2 into the greenhouse, which increased the integrated PnE for non-CE and thus reduced the effect of CO2 enrichment on the increase in Pn. This suggests that while CO2 enrichment can be beneficial, its efficiency is significantly influenced by N levels. At higher N levels, the cost-effectiveness of CO2 enrichment is less pronounced due to the CO2 supplied by ventilation and the increased CO2 leakage to the outside. In summary, the estimation revealed that HP-CE could be more cost-effective than NB-CE for Cin levels of 450-700 µmol mol-1 at N levels of 4, 6, and 8 h-1, despite having the same integrated S. HP-CE was less cost-effective than NB-CE at N of 10 h-1 due to significant CO2 leakage at higher target Cin levels than Cout.
Additionally, in this study, when N was 10 h-1, the CO2 enrichment efficiency under all HP-CE scenarios was similar to or slightly lower than that of NB-CE (Fig. 8L). The N levels in actual greenhouses vary depending on factors such as the greenhouse height, area, and type, and external environmental conditions, including wind speed and seasons. For example, in the case of Venlo-type glass greenhouses with a height of 3 m, the maximum N is approximately 20 h-1 when the wind speed is 1 m s-1 (Sase and Ishii, 2022). When N is 10 h-1 or higher, it is highly likely that all vents of the greenhouse are opened; in this case, it may be more cost-effective to utilize external CO2 through ventilation or to maintain the Cin at a level similar to Cout using NB-CE. However, under lower N conditions, the HP-CE strategy demonstrated in this study is expected to be more cost-effective than NB-CE.
This study evaluated the cost-effectiveness of CO2 enrichment for the 12-hour photoperiod using several estimated values based on the Pn of cucumber seedlings measured under each condition. For the practical application of the newly demonstrated method, several considerations must be taken into account. We used four cucumber seedlings 21 days after sowing for Pn measurements, and there may be slight differences in CO2 enrichment efficiency depending on the planting density and long-term treatments until harvest. Nevertheless, our results suggest that the CO2 enrichment efficiency of HP-CE is higher than that of NB-CE when N levels are 4, 6, and 8 h⁻¹. In our study, the temporal change in PPFD over the 12-hour photoperiod was modeled, demonstrating a sinusoidal curve with a peak PPFD of 700 µmol m-2 s-1; however, in actual greenhouses, PPFD fluctuations are more dynamic, and peak PPFD can exceed 700 µmol m-2 s-1 on sunny days. Therefore, we emphasize that utilizing HP-CE, which raises Cin up to 100 µmol mol⁻¹ above Cout on sunny days, has the potential to be more cost-effective than NB-CE. Additionally, although we set the integrated S of HP-CE to be the same as that of NB-CE, predicting the daily integrated S for NB-CE in actual ventilated greenhouses is challenging. Considering these points, future studies should explore methods for determining the CO2 supply period and the target Cin for CO2 enrichment in actual greenhouses. Empirical evaluations of these CO₂ enrichment strategies over several weeks or months in model or actual ventilated greenhouses would be valuable.
Based on the PPFD-Pn curves of cucumber seedlings, we demonstrated that HP-CE for Cin levels of 450-700 µmol mol-1 is more cost-effective than the conventional NB-CE at N of 4, 6, and 8 h-1. The +100HP-CE, which elevated Cin to 500 µmol mol-1 for 6 h at high PPFD, exhibited the highest CO2 enrichment efficiency. At N of 10 h-1, HP-CE was less cost-effective than NB-CE. These findings suggest that for ventilated greenhouses with relatively low N, CO2 enrichment can be more cost-effective than NB-CE by slightly elevating CO2 concentration inside the greenhouse above the atmospheric levels only for high PPFD times.
The authors are grateful to Shinya Tozawa and Kaiyan Zhang for their assistance in constructing and improving the experimental set-up for determining the CO2 enrichment efficiency and conducting the preliminary experiments. This work was financially supported in part by JSPS KAKENHI (Grant No. 23H00348) and the Konno & Lester Foundation Public Interest Incorporated Foundation.
