Materials and methods

Two identical commercial greenhouses were used as the experimental test sites. Greenhouses were located in Delta, OH (lat. 41°34′N, long. 84°0′W, elevation 722 ft), and consisted of single-span, double polyethylene-glazed greenhouse (29 × 184 ft). Side walls were 5 ft tall, the height at the middle of the greenhouse was 14 ft, and the greenhouses were single peak design. In one greenhouse, a CO₂ controller was installed (model iGS-061; SpecialtyLights.com, Boca Raton, FL) along with a two-way solenoid (model SV122; Omega Engineering, Stamford, CT) to allow CO₂ to be injected into the greenhouse air ducts along each side of the greenhouse from a liquid CO₂ source. This also ensured that all injected CO₂ would be immediately mixed throughout the greenhouse. The day-time CO₂ set point was 500 ppm, while night-time CO₂ was not controlled. The second greenhouse (control house) did not have CO₂ control. Atmospheric CO₂ measurements made at a variety of places around the northern hemisphere indicate that outside CO₂ concentration was between 392 and 395 ppm during that winter (U.S. Dept. Commerce, 2011). Basic CO₂ response curves for C3 plants predict that the 0 to 500 ppm CO₂ range is within the linear response for photosynthesis, so a difference of even a few parts per million would directly influence the photosynthetic rate for lettuce (Taiz and Zeiger, 2002).

In addition to the CO₂ controller, which constantly measured ambient CO₂ concentrations to add additional CO₂, independent measurements were made. About three to four times per week, CO₂ concentrations were measured during the day with a hand-held CO₂ and temperature meter (Telaire 7001; GE Sensing, Billerica, MA) to verify the set point, monitor the control house CO₂ environment, and monitor air temperature in both greenhouses. Humidity was not monitored. The temperature set point in the CO₂-controlled greenhouse was 62 °F constant and that in the control house was 65 °F constant. In the Ohio area, calculations have indicated that for every 1 °F decrease in set point, a 3% savings in energy cost can be realized (P. Ling and C. Pasian, personal communication). A 3 °F would therefore result in nearly 10% fuel savings.

Lettuce seeds were sown in soilless rooting cubes (Oasis, Cleveland, OH) on 22 Nov. 2008. After 10 d (1 Dec.), 10 seedlings were transplanted individually into 6-inch pots (1.3 L) filled with sphagnum peat-based substrate (Sunshine Mix 1; Sun Gro Horticulture, Bellevue, WA). Five plants were grown in each greenhouse for 45 d. At the end of the growth period, leaves were removed, counted, and dried in a forced air oven (55 °C for 3 d) for dry weight measurement. The remaining area in the greenhouses was filled with stock geranium (Pelargonium ×hortorum) for cutting propagation. The large amount of actively growing plants in both greenhouses contributed to the CO₂ drawdown during the day and allowed the CO₂ differences to occur in the uncontrolled CO₂ greenhouse compared with the greenhouse that was actively maintained at a set point of 500 ppm CO₂.

Fuel use simulations were performed using Virtual Grower software (Frantz et al., 2010). A previous study (Frantz et al., 2010) evaluated this greenhouse as part of the verification of Virtual Grower's accuracy. Predicted fuel use and cost was found to be within 5.1% to 5.6% of actual costs based on the grower's heating bills. While it would have been more accurate to measure the actual fuel consumed by installing fuel meters on each greenhouse bay, the model's previous accuracy indicated that C, fuel, and cost predicted from Virtual Grower would fairly represent this commercial facility.

A single-bay greenhouse with double polyethylene covering with sizes matching the experimental greenhouses was used as the structure within the Virtual Grower software (Frantz et al., 2010), with Toledo, OH, weather as the weather database. A temperature set point of either constant 62 or 65 °F was simulated. Propane was used as the fuel source at $2.00 per gallon, which matched what the facility spent on fuel at that time. Infiltration through gaps was set to 0.5 air exchanges per hour to simulate the well-sealed structure. A heating efficiency of 52% was used to simulate older unit heaters with below bench heat distribution tubes. A winter production period of 3 months was simulated covering 1 Dec. through 28 Feb. (total of 90 d). This period matched the period of active CO₂ control during the experiment.

Lettuce mass and leaf number were averaged for each treatment and standard errors were calculated. A Student's paired t test of means was used to determine significance using statistical software (Statistix version 9.0; Analytical Software, Tallahassee, FL).

Results and discussion

For the 3-month experimental period, the temperature was maintained to within 1 °F of set points in each greenhouse. This variation was low and compares favorably to research-quality greenhouses. During the experimental period, there was no active ventilation to the outside; outside temperatures were cold from December to February (below freezing average temperature). Heaters and passive cooling were the primary methods for temperature control. It is important to note that no measurements were made during the night. CO₂ in the control greenhouse was between 200 and 300 ppm on sunny or partly sunny days and always at least 100 ppm lower during the day than the cooler greenhouse receiving supplemental CO₂. Given the conditions during this experiment, it is necessary to determine if this decrease in CO₂ concentration would be reasonable to expect. A simple model driven by light availability can help predict how quickly this drawdown can occur (Table 2; Monteith, 1977; Volk et al., 1995). Using some typical values for the model in this greenhouse (Table 2), it can be calculated that after 1 h, 9.0 mol CO₂ were removed with photosynthesis leaving only 15.0 mol of CO₂ remaining or 258 ppm CO₂ if no ventilation, leakage, or CO₂ supplementation occurs. Drawdown would be slower if ambient light was reduced or fewer plants were present and faster with more or larger plants present, perhaps later in the production cycle. The low CO₂ measured on sunny days is a clear indication that there was little air leakage or mixing between inside and outside the greenhouse. Similarly, CO₂ addition was more frequent on sunny days, based on the audible operation of the solenoid indicating that plants were actively drawing down the CO₂ available at those times.

Table 2.Model parameters used in calculations for carbon dioxide (CO₂) reduction within a greenhouse in 1 h. This is a light-driven model that uses measurements or assumptions about how that light is absorbed, used photosynthetically, and the fixed carbon is subsequently respired in carbon use efficiency. Combined with measurements of greenhouse area, volume, and interior CO₂ concentration, an estimate of the decrease in CO₂ because of photosynthesis can be made in a time step (1 h in this example).

View Fullscreen

Forty-five days after planting, harvested lettuce plants were larger in the cooler, CO₂-controlled house (Fig. 1). Dry weight was substantially greater in plants from the cooler, CO₂-controlled house (Fig. 2A; P < 0.0001). Jie and Kong (1998) documented no change in plant dry weight for lettuce over a range of temperatures from 59 to 77 °F when grown in ambient CO₂ conditions. Thompson et al. (1998) found increased lettuce dry weight per plant from 62.6 to 75.2 °F because of more, larger leaves in warmer conditions. In the current study, the cooler, CO₂-supplemented lettuce plants averaged 2.8 leaves more per plant than the warmer, uncontrolled CO₂ plants (Fig. 2B; P < 0.0001). This suggests that development in the cooler environment, when it was predicted to decrease, was compensated by higher CO₂. Lettuce growers are not as concerned about leaf count or development as floriculture producers, who rely on days to first flower to meet their market demands. It is unknown if leaf development rate in lettuce would correspond to days to flower in ornamental crops. Even though the lettuce was harvested, CO₂ additions continued for an additional 45 d to allow the commercial grower to continue producing vegetative cuttings from stock plants in consistent environments.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Growers may encounter sink limitation when using supplemental CO₂. This occurs when the assimilated C no longer can be used in the construction of new roots, leaves, or fruit, so photosynthesis is downregulated making plants appear to no longer “benefit” from supplemental CO₂ (Arp, 1991). The plants grown in these greenhouses were vegetative. Large pots were used for the lettuce to reduce the possibility of encountering a sink limitation over the course of the 45-d growth period. Production in smaller containers (i.e., plug trays) or longer production of nonfruiting/nonharvested crops may encounter sink limitations and therefore not benefit from supplemental CO₂ (Arp, 1991). In some of those cases, then, slower development and growth because of lower temperatures is not an issue.

The differences between the control greenhouse and cooler greenhouse receiving supplemental CO₂ in cost and C consumed from heating and supplemental C are reported in Table 1. The cost of propane was nearly $1300 (8.4%) more, and the control greenhouse consumed over 7000 lb more C than the cooler greenhouse. This difference is consistent with widespread “rules of thumb” that ≈3% of costs can be saved for every 1 °F that temperature is reduced. This rule of thumb, however useful, ignores potential productivity and development losses at cooler temperatures (Lopez and Runkle, 2004). There was a total of 1200 lb of C consumed from supplemental CO₂. Even when that was included, the cooler, CO₂-controlled greenhouse consumed less C from combustion and supplemental CO₂ than the warmer, unsupplemented house. Liquid CO₂ is expensive, as is the cost of monitoring and controlling CO₂ injection. When these costs were combined, including the accompanying instrumentation, the cooler, CO₂-controlled greenhouse was less expensive to operate than the warmer, uncontrolled CO₂ greenhouse. This cost difference would be greater if similar sized plants were produced because it would have taken longer for the warmer, control house to produce lettuce of equal size as the CO₂-controlled house.

Table 1.Costs of fuel and carbon (C) consumed during the 90-d winter production period for the control and carbon dioxide (CO₂)-elevated greenhouses that were maintained at a constant 65 or 62 °F (18.3 or 16.7 °C), respectively. The 90-d period spanned from 1 Dec. to 28 Feb.

View Fullscreen

The entire cost of the CO₂ controller, solenoid, and piping was factored into these costs even though their lifetime is far greater than a 3-month production system. Most commercial facilities would factor those costs over the lifetime of the materials and equipment. Additionally, liquid CO₂ from a laboratory supply company in small batches was used, which artificially increased the cost. Given differences in how different facilities calculate amortization of costs and variability in cost of CO₂, the full costs are presented herein as a worst-case scenario. Any changes in those costs per unit would further favor the cooler greenhouse receiving supplemental CO₂. For more recent cost estimates for the use of CO₂ as either liquid form or from combustion, see Blom et al. (2009). Liquid CO₂ was also used as a way to ensure no ethylene problems were encountered as can occur in incomplete combustion from CO₂ generators. Cleaning ethylene and other off-gases from the combustion used for heating the greenhouse under normal operating conditions would open up the possibility of using the CO₂ generated during this step (Armor, 1992; Critten and Bailey, 2002). This would enhance C conservation further and potentially provide an on-site source for CO₂ supplementation.

In this study, elevated CO₂ (500 ppm) was compared with below “ambient CO₂” (typically 200 to 300 ppm on sunny days within the greenhouse). It is not known if such a large benefit (plant size and leaf number) would be observed if the comparison was made between plants grown at 400 ppm (close to “ambient”) and a set point of 700 to 800 ppm at two different temperatures. Certainly, growth differences observed were due in part to the control plants grown nearly at their CO₂ compensation point during sunny days making their growth negligible on those days. Their growth would have improved considerably even if they had been grown in an environment with “ambient” CO₂ conditions (i.e., 400 ppm CO₂). Still, in well-sealed greenhouses with no active ventilation, CO₂ is likely to be below ambient, particularly in partly cloudy or sunny conditions.

How much would leakage or ventilation help control CO₂ concentration within these greenhouses? The following equation, similar to exponential decay equations, describes the amount of original air remaining in a volume given a flow into that volume over a period of time: percent volume remaining = e^{–[(flow/volume) × time]} × 100. To use the equation for a greenhouse, the volume of the greenhouse and the air infiltration must be known. Greenhouse air infiltration is typically reported as “air exchanges per hour” and is a relative volume specific to that greenhouse. For example, if a greenhouse has a 50,000 ft³ volume, an air infiltration rate of 1 means that 50,000 ft³ of air flows into that greenhouse (or out) per hour. A tight, newly constructed, well-sealed greenhouse often has an air exchange rate of 0.5 to 1.0, while a poorly sealed or well-ventilated greenhouse may have an air exchange rate of 5.0 or more (Aldrich and Bartok, 1994).

Assuming the same environmental conditions and greenhouse as above, but an air exchange rate of 5.0 air exchanges per hour, the drawdown is less severe but still significant. After one hour, the CO₂ concentration is calculated to be 373 ppm. In a well-constructed, well-insulated greenhouse (0.5 air exchanges per hour), the CO₂ concentration is calculated to be just under 290 ppm. In other words, ventilation would help alleviate CO₂ drawdown, but even in well-ventilated greenhouses, CO₂ would still be less than outside conditions. Using Virtual Grower software (Frantz et al., 2010) to simulate the cost of heating between a well-sealed greenhouse as described above to a well-ventilated greenhouse with 5.0 air exchanges per hour, the costs of heating in Toledo, OH for December through February rises from $15,147 to $27,436. Clearly, it is more expensive to ventilate for CO₂ control than add supplemental CO₂.

These results illustrate that it is possible to save money, heating energy, and carbon from heating in certain situations where supplemental CO₂ is used. The use of CO₂ enrichment should, therefore, be considered in sustainable systems when its use can counteract the plant growth and development effects brought on by lowered temperatures. More work is needed, however, that development can accelerate in supplemental CO₂ environments compared with subambient CO₂ environments. Further, it would be of interest to determine if sink limitations are encountered in containerized plant production systems when supplemental CO₂ is used in longer term crops.

Units

View Fullscreen