Source DocumentTable Of ContentsNext Page

The effect of elevated carbon dioxide on the growth and yield of wheat in the Australian Grains Free Air Carbon dioxide Enrichment (AGFACE) experiment

Rob Norton1, Glenn Fitzgerald2 and Chris Korte3

1 The University of Melbourne, Private Bag 260, Horsham, Vic, Email
Department of Primary Industries, 110 Natimuk Rd., Private Bag 260, Horsham, Vic,
Department of Primary Industries, 110 Natimuk Rd., Private Bag 260, Horsham, Vic,


Current predictions indicate that Australia is likely to be particularly challenged by the impacts of rising atmospheric carbon dioxide and the consequent perturbations in climate. The Australian Grains Free Air Carbon dioxide Enrichment (AGFACE) project in Horsham, Victoria was designed to simulate predicted atmospheric carbon dioxide levels in the year 2050. The experiment measures the interacting effects of carbon dioxide (ambient aCO2 ~380 ppm, elevated eCO2 ~550 ppm), irrigation (rainfed, irrigated), higher temperatures during grain fill (time of sowing), nitrogen (0, +), and variety (Yitpi, Janz) on wheat growth and production. Carbon dioxide was injected over the crop in open-air 12 m rings from emergence (July) until maturity (December) in 2007. Crop development was not affected by eCO2. The effect of eCO2 was to increase crop biomass at maturity by 20% (P<0.001) and anthesis root biomass increased by 49% (P=0.004). Harvest index was not affected but mean grain yield across all treatments increased from 2.68 t/ha under aCO2 to 3.23 t/ha under eCO2. Both sowing time and additional water affected growth and yield but there were no significant interactions among these factors and eCO2. The effect of higher carbon dioxide was to slightly increase the number of kernels per spikelet (P=0.055). Water use, the sum of rainfall and change in soil water from sowing to maturity was 387 mm with no differences among the treatments other than irrigation. There were no significant interactions between carbon dioxide and genotype or nitrogen treatment on growth or yield. These data will be used to calibrate crop simulation models to assist with developing strategies to assist the grains industry adapt to the changing climate.

Key Words

Climate change, grain protein, water use, root depths.


The 2050 IPCC emissions scenario A1B indicates that atmospheric carbon dioxide will reach 550 ppm (Carter et al. 2007), and current climate models suggest that annual rainfall in the grain production regions of Australia will decline by 50–100 mm and annual mean surface temperatures rise by 1-2C (Whetton 2001).

The effects of elevated carbon dioxide are to increase photosynthetic rates and decrease transpiration, which should result in higher transpiration efficiency (Gifford 2004). However, the concurrent rise in temperature and decline in water availability with climate change are likely to impact significantly on any gains due to the CO2 fertilisation effect. While the response of cereals to elevated carbon dioxide has been investigated in glasshouse, open-topped chambers and open field (Free Air Carbon Dioxide Enrichment - FACE) experiments, most of this research has been done in relatively high yielding environments and not with Australian germplasm (eg Kimball et al. 1995). As a result, the Australian Greenhouse Office (now the Department of Climate Change) commissioned the establishment of a research facility in Australia to identify the impact of the higher atmospheric CO2, rising temperatures and declining water availability on wheat growth and yield in field conditions in Australia.

This short paper reports the first results of an experiment at Horsham for wheat grown in a FACE facility.


In 2007, a FACE facility was established at Horsham (3645’S, 14206’E), in the grain growing region of southeastern Australia. The facility consists of 8 elevated carbon dioxide (eCO2) rings each 12 m in diameter, with equivalent ambient carbon dioxide (aCO2) experimental areas spread over a 5 ha site. Treatments imposed aimed to develop a range of temperature and water regimes during crop growth under aCO2 and eCO2 conditions. Rainfall was adjusted using irrigation to give decile 5 (D5 - 225 mm Apr-Nov) and decile 7 (D7 - 275 mm Apr-Nov) conditions. Two sowing times (TOS1 June 18 and TOS2 August 22) were used to give a range of temperatures during crop growth. In the 10 days after anthesis, mean temperatures were 15C and 21C for June 18 and August 22 respectively. Two cultivars (Yitpi and Janz), one with two rates of nitrogen fertiliser were also compared as sub-treatments within the rings. Four replications were used with CO2 and water supply combined factorially at the ring level. Each ring was split for sowing time and then within each sowing time, the plot level factors were combined.

Growth, yield and water extraction was measured at stem elongation, anthesis and maturity for all treatments, and yield components, screenings (<2 mm) and grain protein content (at 0% moisture) were also measured. Only the results for the maturity harvests are presented here. Because some plots had poor emergence with the later sowing, a ReML model was used to compare means with the main factors of carbon dioxide enrichment, sowing time, supplementary water and “plots”. The latter term refers to the three within half-ring treatments of Yitpi with no N, Yitpi with additional N and Janz with no additional N. The ReML analysis was done with Treatment Structure as CO2*Irrig*ToS*PlotTrt (assumed as fixed effects) and Block Structure as (Bay*Row)/HalfRing/Plot (assumed as random effects). Terms that turned out to have negative variance components in the Block Structure were removed from the model and ReML re-run using rest of the Blocking Terms.

Crop root biomass and root length was estimated from two soil cores taken within the Yitpi with no N plots in each of the main treatment blocks at anthesis. Cores were taken and washed to remove the roots which were then measured using a WinRhizo scanner and software, and the roots dried. Root length density (cm of root per cm3 of soil volume) and root mass (kg/ha) were derived from those measurements. These data were balanced and were compared using analysis of variance.

Results and Discussion

The treatments selected provided a range of environments for wheat growth and development. The two sowing times and the two water treatments provided a contrast in both temperature and water supply (Table 1) to test the interaction among these factors on the response to eCO2. Figure 1 shows the grain yields for the two sowing times, two water treatments and the two carbon dioxide levels.

Figure 1. The effect of elevated carbon dioxide on wheat grain yields for two sowing times (Jun 18, Aug 22) and two levels of water supply (D5 -225 mm, D7 – 275 mm) in the Horsham FACE experiment, 2007. SEd (C02) = 0.165 (df = 5.3), SEd (water supply) = 0.163 (df = 5.0), SEd (ToS) = 0.145 (df = 44.6).

When averaged across all treatments, elevated CO2 increased biomass at maturity by 22% and grain yield by 20%, and there was no significant effect on harvest index or on crop development. When averaged across all treatments, grain yield went from 2.68 t/ha for aCO2 to 3.23 t/ha in response to eCO2. The higher yield from eCO2 was a result of a combination of non-significant individual yield components (see Table 1), with slightly higher kernel weights (aCO2 31.3 mg; eCO2 32.9 mg), slightly more kernels per spike (aCO2 25.3; eCO2 27.3) and a small increase in spikes per plant (aCO2 4.0; eCO2 4.6). There were no significant interactions among CO2 and the other factors for grain yield and biomass at maturity (Table 1). The higher temperatures during early grain filling for the August sowing (+6C) did not alter the “fertilisation” effect of higher carbon dioxide in this experiment.

In a meta-analysis of a large number of FACE experiments, Ainsworth and Long (2005) showed that wheat yield increases in response to higher CO2 levels were in the order of 0 to 30%, depending on water and N supply. The results from this AGFACE experiment are within this general range and it may be that under even lower water supply, a larger response may have occurred (Wall 2001).

Grain protein under eCO2 was 13.2% which was significantly lower than the protein content under aCO2 (13.8%). Nitrogen supply at this site was high (~150 kg N/ha to 50 cm) and no significant effects were noted on any parameter due to the applied N. Lower grain protein under high CO2 has been noted in other studies (Blumenthal et al. 1996) and this appears to be a consequence of lower N content in the plant during growth as Rubisco biosynthesis is down regulated as less enzyme is required to maintain photosynthetic rates (Ainsworth and Rogers 2007). As a consequence, at maturity there is somewhat less N available for remobilization to the grain.

Table 1. F statistic probability from ReML analysis of two CO2 treatments (CO2), two sowing times (TOS), two levels of water supply (Water) and three subplot treatments (Plots – Yitpi 0N, Yitpi +N, Janz 0N) and the significant interactions among those main factors. Significant effects (P<0.05) are shown in bold.





































Kernel Wt






% Screeings





CO2*Plots, TOS*Plots, CO2*TOS*Plots

Grain Protein %





Water*TOS, TOS*Plots

At anthesis, roots were found to 1.6 m although ~90% of the root mass in all treatments was in the top 60 cm of the soil profile although there were no significant treatment effects. Irrigation and sowing time showed a significant interaction for root biomass to 60 cm, with relatively more roots for the first sowing compared to the second sowing under the wetter conditions (D7), although the root biomass was similar for the D5 conditions. Under eCO2, there was 46% higher root biomass than under aCO2 (87653 kg/ha versus 59755 kg/ha) (P=0.004). The higher root biomass at anthesis was reflected in a higher root length density of the crops grown under eCO2 (1.820.13 compared to crops grown under aCO2 (1.360.13 in the top 60 cm of the soil (P=0.024).

Crop water use, as measured by the change in soil water between sowing and harvest plus rainfall and irrigation, was not affected by carbon dioxide treatment, but sowing time and irrigation showed a significant interaction for crop water use (Table 2). The second time of sowing did not extract all the additional water supplied as irrigation while the first sowing time extracted 77 mm more water when irrigated, which was approximately the same as the amount of additional water supplied to this treatment.

Table 2. Crop water use (mm) for crops sown on June 18 or August 22 and the two water supply levels in the FACE experiment.


June 18

August 22

Rainfed (D5)



Irrigated (D7)



There were no particular interactions with the cultivars tested and carbon dioxide levels. In these experiments, Janz had a lower kernel weight than Yitpi (29.1 mg; 33.8 mg respectively, P<0.001), but yields were not different as Janz had more kernels per spike than Yitpi (28.1; 25.2 respectively, P<0.001). These cultivars were chosen for this experiment as they are widely grown commercially in southeastern Australia and are also genetically quite different (Ogbonnaya et al. 2006). Despite these differences, in this study there is no indication of differential responses to elevated carbon dioxide, which suggests that a wide gene pool needs to be investigated to see if particular cultivars are able to respond more to eCO2.

The growth, yield and water use data collected in this experiment will be used to develop an understanding of the response of wheat to eCO2 under high temperatures and with low water availability, and to assist in calibrating existing simulation models. Once we have confidence in the models, various adaptation scenarios can be tested to assist the grains industry adapt to the impacts of climate change.


This experiment has demonstrated the “fertilisation” effect of elevated carbon dioxide, as reported from areas with higher yield potentials than at Horsham. The main yield components affected by eCO2 in this experiment were kernels per spike and kernel weight, however these were at a low level of statistical confidence (P=0.07). The actual growth and yield outcomes from this experiment are in the general range of the meta-analyses reported for eCO2 impacts, and there were few significant interactions among the factors tested here between carbon dioxide concentration, time of sowing, water supply and cultivar.


This project has been supported by the Department of Climate Change and the Grains Research and Development Corporation (UM00027), with additional support from BOC Ltd. The authors acknowledge the technical support of R Argall, J Fitzpatrick and P Howie. Dr J Smith and Dr S Chandra provided assistance with biometrics and the grain analyses were done by Dr J Panozzo (all Victorian Department of Primary Industries).


Ainsworth EA, Long SP (2005). What have we learned from 15 years of free-air CO2 enrichment (FACE)? A meta-analytic review of the responses of photosynthesis, canopy properties and plant production to rising CO2. New Phytologist 165, 351-372.

Ainsworth EA, Rogers A (2007). The response of photosynthesis and stomatal conductance to rising [CO2]: mechanisms and environmental interactions. Plant Cell and Environment 30, 258-270.

Blumenthal C, Rawson HM, McKenzie E, Gras PW, Barlow EWR, Wrigley CW (1996). Changes in wheat grain quality due to doubling the level of atmospheric CO2. Cereal Chemistry 73, 762-766.

Carter TR, Jones RN, Lu X, Bhadwal S, Conde C, Mearns LO, O’Neill BC, Rounsevell MDA, Zurek MB (2007) New Assessment Methods and the Characterisation of Future Conditions. Climate Change 2007: Impacts, Adaptation and Vulnerability. IPCC, Cambridge University Press, Cambridge, UK, 133-171.

Gifford RM (2004). The CO2 fertilising effect – does it occur in the real world? New Phytologist 163, 221-225.

Kimball BA, Pinter Jr PJ, Garcia RL, LaMorte RL, Wall GW, Hunsaker DJ, Wechsung G, Wechsung F, Kartschall Th (1995). Productivity and water use of wheat under free-air CO2 enrichment. Global Change Biol. 1, 429–442.

Ogbonnaya FC, Imtiaz M, DePauw RM (2006). Haplotype diversity of preharvest sprouting QTL’s in wheat. Genome 50, 107-116.

Steffen W, P Canadell (2005). Carbon dioxide fertilization and climate change policy. Australian Greenhouse Office.

Whetton P (2001). Climate change: projections for Australia. CSIRO Atmospheric Research Report, Aspendale, Victoria.

Wall GW (2001). Elevated atmospheric CO2 alleviates drought stress in wheat. Agriculture, Ecosystems and Environment 87, 261–271.

Top Of PageNext Page