The response of wheat crops to elevated CO₂ (eCO₂) was measured and modelled with the Australian Grains Free-Air CO₂ Enrichment experiment, located at Horsham, Australia. Treatments included CO₂ by water, N and temperature. The location represents a semi-arid environment with a seasonal VPD of around 0.5 kPa. Over 3 years, the observed mean biomass at anthesis and grain yield ranged from 4200 to 10 200 kg ha⁻¹ and 1600 to 3900 kg ha⁻¹, respectively, over various sowing times and irrigation regimes. The mean observed response to daytime eCO₂ (from 365 to 550 μmol mol⁻¹ CO₂) was relatively consistent for biomass at stem elongation and at anthesis and LAI at anthesis and grain yield with 21%, 23%, 21% and 26%, respectively. Seasonal water use was decreased from 320 to 301 mm (P = 0.10) by eCO₂, increasing water use efficiency for biomass and yield, 36% and 31%, respectively. The performance of six models (APSIM-Wheat, APSIM-Nwheat, CAT-Wheat, CROPSYST, OLEARY-CONNOR and SALUS) in simulating crop responses to eCO₂ was similar and within or close to the experimental error for accumulated biomass, yield and water use response, despite some variations in early growth and LAI. The primary mechanism of biomass accumulation via radiation use efficiency (RUE) or transpiration efficiency (TE) was not critical to define the overall response to eCO₂. However, under irrigation, the effect of late sowing on response to eCO₂ to biomass accumulation at DC65 was substantial in the observed data (~40%), but the simulated response was smaller, ranging from 17% to 28%. Simulated response from all six models under no water or nitrogen stress showed similar response to eCO₂ under irrigation, but the differences compared to the dryland treatment were small. Further experimental work on the interactive effects of eCO₂, water and temperature is required to resolve these model discrepancies.

Introduction

Crop simulation models are robust tools to extrapolate likely behaviour under circumstances that are difficult or impossible to experimentally test in the field, such as a changing climate. In agriculture, many studies have employed crop modelling with various future climatic scenarios (e.g. Hoogenboom et al., 1995; Boote et al., 1996; Semenov et al., 1996; Howden & O'Leary, 1997; Amthor, 2001; Howden et al., 2003; Asseng et al., 2004, 2013a; Power et al., 2004; Ludwig & Asseng, 2006; Anwar et al., 2007; Lobell & Ortiz-Monasterio, 2007; O'Leary et al., 2011). The range of crop models available for use in such studies is large because of differences in complexity and function. Models that include some realistic simulation of the effects of elevated atmospheric CO₂ (eCO₂), temperature and water stress on crop growth and development are considered a prerequisite in terms of model function for future climate change studies. Other subcomponents involving mineral uptake may also be critical for some applications (e.g. nitrogen in soil C sequestration studies).

Comparison against measured field data is fundamental to improving crop models (Boote et al., 1996). Such improvement offers assurance that subsequent simulation scenario studies are the best available. This study tests the performance of six wheat crop models against 3 years of measured crop growth and water use data from the Australian Grains Free-Air CO₂ Enrichment (AGFACE) experiment, located in a semi-arid (seasonal VPD ranging from 0.2 to 1.1 kPa with a mean around 0.5 kPa) environment in southern Australia. Previous studies from other environments that have compared simulations with FACE results have reported response to eCO₂ to be close to the observed response in biomass, yield and water use (Tubiello et al., 1999; Jamieson et al., 2000; Asseng et al., 2004; Ko et al., 2010). Seasonal growth responses to eCO₂ for wheat have been in the order of 5–40% (mean ~18%) (Ainsworth & Long, 2005; Sun et al., 2009) depending on other growth limiting factors. In general, C₃ cereal crops under water stress responded proportionally more to eCO₂ (~22%) than unstressed crops (~16%), showing the importance of water supply (Kimball, 2011). The lack of observed and consistent high responses (>25%) in field crops is puzzling (Long et al., 2006), but one reason proposed might be artefacts of the experimental environment (Long et al., 2006; Bunce, 2013). Another aspect needing consideration is the relative dryness of the atmosphere during the season that will affect transpiration efficiency. Previous FACE experiments are not considered representative of semi-arid regions because their annual potential evapotranspiration/rainfall (PET/R) ratio is either less than or close to unity or very high. The PET/R ratio at the AGFACE site in southern Australia is around 3 compared to around 10 in Arizona. Similarly, from a seasonal perspective and despite crops being irrigated in Arizona, the seasonal VPD also remains about three times the Horsham AGFACE site (1.4 kPa cf. 0.5 kPa). Because of the intermediate dryness of the Australian AGFACE experiment and the present view that the relative response to eCO₂ is much higher under conditions of water stress (Leakey et al., 2009; Sun et al., 2009; Kimball, 2011; Dias De Oliveira et al., 2012), this study offers a new opportunity to test the performance of a range of crop models under eCO₂ in a new unirrigated environment. In addition, to testing models in terms of the accumulation of crop biomass, grain yield and water use and subsequent water use efficiency (WUE) in response to eCO₂, we explore whether model components used to calculate these responses need improvement.

Materials and methods

FACE experiment

The AGFACE experiment at Horsham, Australia (36°45′S, 142°07′E, 128 m elev.) (Mollah et al., 2009), was used to measure the growth, yield and water use of wheat under various controlled conditions. The experiment included multiple cultivars, but we restricted our study to one popular local cultivar (cv. Yitpi) that was grown under two water regimes (rain-fed and supplemental irrigation), two nitrogen fertilisation regimes (0 and 53–138 kg N ha⁻¹) and two sowing dates (to create two growing season temperature environments) for both daytime ambient (365 μmol mol⁻¹) and elevated (550 μmol mol⁻¹) CO₂ atmospheric conditions. Injection of CO₂ was controlled by wind speed and direction whereby it was injected approximately 150 mm above the crop canopy upwind when wind speed was between 0.3 and 10 m s⁻¹ between 2 h after sunrise and 2 h before sunset. The primary design criteria of daytime 550 μmol mol⁻¹ CO₂ were achieved 86–94% of time ranging from median 495–605 μmol mol⁻¹ (within ±10%) in the centre of 12-m (2007 and 2008) and 16-m (2009) diameter octagonal rings (Mollah et al., 2009). Within 1-m downwind of the rings, CO₂ levels ranged from 500 to 1700 μmol mol⁻¹. The agronomic design with the two times of sowing (TOS) over 3 years (2007–2009) comprised a complete randomised block experimental design of four replicates. This equates to 48 crops over the 3 years (2CO₂ × 2TOS × 2 irrigation × 2 nitrogen × 3 years). In 2007, plots were split for TOS, where half of each plot was sown to each TOS. In subsequent years, plots were split for irrigation, in which a 0.8-m-deep buried plastic barrier separated the two rain-fed and supplemental irrigated half-plots. Each replication was one ‘bay’ with two aCO₂ experimental areas and two eCO₂ rings per bay. Sowing time alters biomass partitioning including yield as crops are forced to develop into warmer, less efficient conditions approaching summer (O'Leary et al., 1985).

Gravimetric soil water content was measured at sowing and harvest using a hydraulically operated soil sampler where sampling was in increments of 0–0.1, 0.1–0.2 and 0.2 m increments thereafter to 2 m from one core per plot (42 mm diameter). Soil mineral nitrogen (NO₃ and NH₄) was also measured from an additional core taken close to the sampling time of the soil water measurements. Site bulk density was measured from 70 mm diameter × 75 mm deep sampling rings from each octagonal area. We defined seasonal crop water use as the change in soil water content (mm) during the growing season (sowing to harvest) plus rainfall and irrigation between sowing and harvest. This therefore includes losses from soil evaporation, crop transpiration, deep drainage and run-off, with the latter two considered insignificant from an agronomic perspective at this location (i.e. within the error of measurement, after O'Leary & Connor, 1997). For example, seasonal drainage is known to be around 20 mm with typical errors of water use for these soils around 50 mm. We defined WUE as grain yield divided by water used (soil evaporation plus transpiration plus deep drainage plus run-off). Large soil mineral nitrogen content (~300 kg N ha⁻¹) at the site precluded any significant effects of applied N, so soil analyses were pooled across the N treatments.

There were 12 subplots per ring in 2007 and 2008, each 1.4 by 4 m. There were 24 subplots in 2009, the additional plots containing cultivars not analysed here. Each subplot was 8 plant rows wide. Row spacing was 0.217 m in 2007 and 0.19 m in 2008 and 2009. Within each subplot, there were three destructive harvests; at maximum tillering/initiation of stem elongation (DC31), anthesis (DC65) and maturity (DC90) (Zadoks et al., 1974). The centre rows were sampled, leaving the outer rows on each side as buffers. In 2007, destructive samples for the DC31 and DC65 samplings were randomly distributed within each subplot as 0.5-m and 1.0-m row segments at maturity (DC90). Beginning in 2008, all samples were collected as quadrats of four rows by 0.5 m (DC31 and DC65) and four rows by 1-m areas at DC90. The DC31 and DC65 samples were collected within a reserved area termed ‘growth’ subplot, and the DC90 sample was collected from a designated area termed ‘maturity’ subplot. Biomass samples were oven-dried at 70 °C, and leaf and stem area measurements were made from using an electronic planimeter from subsamples comprising approximately 25% of the collected fresh biomass. Mean sowing plant density measured by plant counts about 3 weeks after the emergence was 120 plants m⁻² and ranged from 60 to 175 plants m⁻². Agronomic management at both sites was according to local practices, including spraying fungicides and herbicides, as needed. Granular phosphorus and sulphur (as ‘superphosphate’) were incorporated into the soil at sowing at rates between 7–9 kg P ha⁻¹ and 8–11 kg S ha⁻¹ depending on year.

The experimental site had previously been laser-graded to assist in flood irrigation and had been irrigated with sewage water for over 20 years to assist bulking seed in various plant breeding programmes. This history resulted in extreme variability of critical soil properties at the site, and as a consequence for modelling the response to eCO₂, we pooled the initial soil water content at sowing across the ambient and eCO₂ treatments to be consistent with single soil type parameters for the site (Table S1). This option, as opposed to modelling each ring, provided an effective and simple solution to compare the experimental data with all its site variance, to the modelled data.

Simulation models

Six crop simulation models were tested against the AGFACE data (Table 1). They represent two primary groups of popular models dealing with the effects of eCO₂ and final crop yield by the primary method of daily biomass accumulation: (1) those that primarily use radiation use efficiency (RUE) calculations and (2) those that primarily use transpiration efficiency (TE) calculations. Both approaches are approximations to the real photosynthetic mechanisms, but their utility across diverse environments is well known. Each model was run by the respective modelling groups/authors without adjusting parameters to fit the data apart from the data to define the cultivar phenology. In addition to the selected models, we analysed the median of all the models as an ensemble to explore whether an ensemble result was more accurate than any individual model.

Table 1. List of selected models to compare against measured FACE experiment showing the primary biomass accumulation method – radiation use efficiency (RUE) or transpiration efficiency (TE)

Model name	Biomass method	Source	Example climate change analyses
APSIM-Wheat v7.4	RUE	McCown et al. (1996)	Ludwig & Asseng (2006) and Asseng et al. (2013a)
APSIM-Nwheat v1.55	RUE	Asseng et al. (2004)	Asseng et al. (2011, 2013a,b)
CAT-Wheat v8.4.5	RUE	Weeks et al. (2008) and Christy et al. (2013)	O'Leary et al. (2011)
CROPSYST v 4.15.24	TE	Stöckle & Nelson (2001)	Anwar et al. (2007) and Asseng et al. (2013a)
OLEARY-CONNOR v7	TE	O'Leary & Connor (1996a,b)	Howden & O'Leary (1997) and Asseng et al. (2013a)
SALUS v1	RUE	Basso et al. (2010)	Asseng et al. (2013a) and Bassu et al. (2014)

These models have all been used in various climate change applications.

The experimental site is defined by its soil type (Table S1) and daily weather conditions that were measured mostly on site. When on-site measurements were unavailable, locally calibrated calculated values for temperature were made from two nearby meteorological stations approximately 10 km away. Other generic site parameters needed by some of the models are listed in Table S2.

Simulation models – the APSIM-Wheat model

The Agricultural Production Systems Simulator (APSIM) is a farming systems simulation framework that has been designed to allow evaluation of farm decision-making in the face of climatic risk, climate change or changes in policy (Keating et al., 2003). The Wheat model within APSIM has been broadly tested across Australia and internationally in a range of experimental (Holzworth et al., 2011; Zhang et al., 2012) and farm (Hochman et al., 2009; Carberry et al., 2013) conditions. Testing of simulated responses to increasing atmospheric CO₂ has been published for predecessors of the current APSIM-Wheat model (Reyenga et al., 1999; Asseng et al., 2004). Increased atmospheric CO₂ concentration impacts upon simulated crop growth and resource use via changes to RUE, TE and the critical nitrogen concentration (CNC) for crop growth (Reyenga et al., 1999). To capture CO₂ effects on RUE, and interactions with temperature, APSIM-Wheat (v 7.4) scales RUE using the ratio (ϕ_p) of the light-limited photosynthetic response at the enhanced CO₂ level to that at 350 μmol mol⁻¹:

$urn:x-wiley:13541013:media:gcb12830:gcb12830-math-0001$ (1)

where the temperature-dependent CO₂ compensation point (Г) is calculated as:

$urn:x-wiley:13541013:media:gcb12830:gcb12830-math-0002$ (2)

where t is the temperature (°C). The responses of transpiration efficiency (g m⁻² mm⁻¹) and leaf CNC (%) to increased CO₂ are assumed to be linear with changes of +37% and −7%, respectively, for a doubling of CO₂ concentration to 700 μmol mol⁻¹ (Fig. 1). Actual transpiration is indirectly reduced under eCO₂ through the gain in TE. Table S3 lists the salient cultivar parameters for APSIM-Wheat used to simulate the AGFACE experiments.

Details are in the caption following the image — **Figure 1**
Open in figure viewer PowerPoint

Correction factors applied to radiation use efficiency (growth ratio, a) for APSIM-Wheat and APSIM-Nwheat at 15 °C (○) and 30 °C (●) and OLEARY-CONNOR & CAT-Wheat (◊), CROPSYST (□) and SALUS (∆) and transpiration efficiency (TE ratio, b) for the APSIM-Wheat and APSIM-Nwheat (○), CROPSYST (□) and OLEARY-CONNOR (◊) and CAT-Wheat & SALUS (∆) models. The factors for atmospheric CO₂ of 550 μmol mol⁻¹ vary slightly between the models. Note: the various models apply RUE and TE in different ways, and as TE is related to RUE by growth, both correction factors are applied in all these models such that double accounting is avoided (Table 1).

Simulation models – the APSIM-Nwheat model

The APSIM-Nwheat model (v 1.55) was developed from an earlier version of APSIM. This model has been tested and applied in Australia (Asseng et al., 1998a,b, 2003; Asseng & Van Herwaarden, 2003) and many other growing environments in the world (Asseng et al., 2000, 2002, 2003, 2004; Wessolek & Asseng, 2006; Bassu et al., 2009). It uses the same CO₂ functions by Reyenga et al. (1999) as described above for APSIM v7.4 but has accelerated senescence functions and a physiological approach for water-soluble carbohydrate remobilisation, which is critical for terminal drought situations (Asseng & Van Herwaarden, 2003). Table S4 lists the salient cultivar parameters for APSIM-Nwheat used to simulate the AGFACE experiments.

Simulation models – the CAT-Wheat model

CAT-Wheat model (v 8.4.5) (Weeks et al., 2008) was selected as a simple model with respect to biomass accumulation (see below) but has high utility in spatial analyses across landscapes (O'Leary et al., 2011; Christy et al., 2013). It includes modules for phenological development, crop growth and yield, together with dynamics of water and nitrogen in the crop and soil. Photosynthetic area (includes leaf and stem) is determined by above-ground biomass. The model was based on a generic annual crop model to enable the simulation of any crop and is based on extensively used contemporary models (Williams et al., 1989; Littleboy et al., 1992). Biomass accumulation is determined from intercepted radiation, RUE, water and nutrient stress factors and a photoperiod factor (Williams et al., 1989). The minimum value of these factors (i.e. the most limiting) is always used to reduce RUE. The fertilisation effect of eCO₂ is achieved by multiplication of RUE by a factor (RUECC) from (Stöckle et al., 1992).

$urn:x-wiley:13541013:media:gcb12830:gcb12830-math-0003$ (3)

where CO₂ is the atmospheric CO₂ concentration (μmol mol⁻¹) such that RUECC = 1, 1.18 and 1.23 at 350, 550 and 650 μmol mol⁻¹, respectively (Fig. 1). Potential or actual crop transpiration is not reduced by eCO₂. Grain yield is determined using the harvest index approach that is modified by water and nitrogen stress. Components of the water balance include crop interception, soil evaporation, transpiration, run-off and drainage that provide daily soil water content. Table S5 lists the salient cultivar parameters for CAT-Wheat used to simulate the AGFACE experiments.

Simulation models – the CROPSYST model

The CROPSYST model (Stöckle et al., 1994, 2003) is a popular systems model that has been well tested and used in USA and Mediterranean environments in Europe and the Middle East (Stöckle et al., 1994, 1997; Pala et al., 1996; Donatelli et al., 1997; Stöckle & Debaeke, 1997; Pannkuk et al., 1998; Stöckle & Jara, 1998; Ferrer-Alegre & Stöckle, 1999; Tubiello et al., 2000; Sadras, 2002; Díaz-Ambrona et al., 2005).

We used CROPSYST (v 4.18.27) which calculates dry matter accumulation as a function of daily intercepted solar radiation and daily crop transpiration, using constant coefficients of RUE (Monteith 1981), and variable coefficients of TE (Kremer et al., 2008). The minimum of the two calculations each day of the growing season determines the dry matter production for the day. For unstressed wheat, a ratio (r) of 1.134 for a CO₂ change between 360 and 560 μmol mol⁻¹ was used in this study. The value of this ratio as a function of CO₂ is given by a nonrectangular hyperbola (NRH) fitted to three points: r = 1 at CO₂ = 360 μmol mol⁻¹, r = 1.134 at CO₂ = 560 μmol mol⁻¹ and r = 0.95 × NRH maximum ratio at CO₂ = 900 μmol mol⁻¹ (CO₂ approaching saturation for C3 species). The ratio for any given CO₂ of interest is applied directly to RUE. In the case of TE, the effect of CO₂ on canopy conductance for water vapour (and therefore on crop transpiration) is accounted for following Allen (1990). Thus, both TE (increase) and crop transpiration (reduction) are adjusted to obtain the expected value of r. Table S6 lists the salient cultivar parameters for CROPSYST used to simulate the AGFACE experiments.

Simulation models – the OLEARY-CONNOR wheat model

The OLEARY-CONNOR (v7) model calculates biomass primarily by transpiration and transpiration efficiency (TE, kg ha⁻¹ mm⁻¹), which is modified by temperature and nitrogen functions via an algebraic equivalent radiation use efficiency (RUE, g MJ⁻¹) (O'Leary & Connor, 1996a,b). It has been successfully tested and used in southern Australia in unirrigated conditions (O'Leary & Connor, 1998; Latta & O'Leary, 2003). The fertilisation effect of eCO₂ is achieved by multiplication of TE by a factor (TECC) that is derived from a RUE factor (RUECC) (Stöckle et al., 1992; Stöckle & Nelson, 2001) with simplifying assumptions of aerodynamic resistance of crop canopy 300 s m⁻¹ and canopy resistance of 36 s m⁻¹ at 350 μmol mol⁻¹ CO₂:

$urn:x-wiley:13541013:media:gcb12830:gcb12830-math-0004$ (4)

$urn:x-wiley:13541013:media:gcb12830:gcb12830-math-0005$ (5)

where CO₂ is the atmospheric CO₂ concentration (μmol mol⁻¹) and δ (kPa °C⁻¹) and γ (kPa °C⁻¹) are the psychrometric constant and slope of saturated vapour pressure–temperature curve, respectively, such that RUECC = 1, 1.18 and 1.23 at 350, 550 and 650 μmol mol⁻¹, respectively. Potential crop transpiration is reduced by eCO₂ in proportion to the ratio of RUECC/TECC that provides a TECC of 1.30 at 700 μmol mol⁻¹ CO₂ (Fig. 1). The RUE correction is almost identical to APSIM at 25 °C, but no temperature correction is applied (Fig. 1). Photosynthetic area (includes leaf and stem) is determined by above-ground biomass and its nitrogen concentration. Table S7 lists the salient cultivar parameters for OLEARY-CONNOR used to simulate the AGFACE experiments.

Simulation models – the SALUS model

The Systems Approach to Land Use Sustainability (SALUS) (Basso et al., 2006, 2010) is an evolution of the CERES models (Ritchie et al., 1988). When perceived improvements in CERES models became available after 2002, those were included in SALUS. The primary purpose of SALUS is to do long-term simulation cropping systems with emphasis on simulating yields along with environmental consequences of various the management strategies. SALUS has been tested for crop yield (e.g. Basso et al., 2007, 2009, 2010, 2012; Asseng et al., 2013a; Rosenzweig et al., 2013; Dzotsi et al., 2013), soil C dynamics (e.g. Senthilkumar et al., 2009), plant N uptake and phenology (e.g. Basso et al., 2010, 2011a), nitrate leaching (e.g. Giola et al., 2012; Syswerda et al., 2012), and tillage effects on soil properties (Basso et al., 2006, 2011b). Growth is primarily determined by the RUE approach and is then reduced based on the transpiration and nitrogen limitations. The effects of eCO₂ are simulated by adjusting RUE as a function of CO₂ with a corresponding increase in the amount of transpiration use efficiency (Fig. 1). The effect of temperature on RUE and transpiration is made by applying a multiplier to both whereby the factor varies nonlinearly from 1.25 to unity below or above the optimum temperature for photosynthesis. The advantage of this model is its utility in soil management across diverse landscapes. Table S8 lists the salient cultivar parameters for SALUS used to simulate the AGFACE experiments.

Statistical analyses

Comparisons of simulated data against the observed data were made by simple regression of treatment means (arithmetic) of the form y = a + bx where y is the observed and x is the simulated data (n = 48). The regression coefficients (a, b), their standard errors (SE), 95% lower and upper confidence intervals of b, residual degrees of freedom (df), coefficient of determination (R²), significance of regression (P) and root mean square error (RMSE) are reported and listed in Table S9.

The response of a number of variables to eCO₂ was analysed by simple regression of treatment means (arithmetic) of the eCO₂ treatments against the ambient treatments, forced through zero. The classical method of determining the response to eCO₂ by dividing the elevated measurement (e.g. grain yield) by the ambient measurement is, by default, a fit forced through the origin with the slope (eCO₂/aCO₂ ratio) above unity indicating positive response to the level of CO₂ elevation. One of the problems of measuring response to eCO₂ for each treatment is the problem of increasing or decreasing response as the absolute growth or yield approaches zero (Amthor, 2001). The method of fitting a single slope that we employed weights the response against such sensitivity because the mean slope is forced through zero over the wide range of absolute growth or yield. This relative approach provided a robust and simple way to reduce experimental noise and obtain gross estimates of the effects of eCO₂ on observed and simulated yield, water use and WUE rather than by differences through classical anovas. This method then allows both graphical and statistical comparison of the experimental data against the simulated data in an efficient way. For this comparison, the resultant degrees of freedom are halved (n = 24).

Results

FACE experiment

FACE experiment – wheat response to elevated CO₂

The AGFACE field data have shown that early growth, yield and WUE of wheat increased under eCO₂. At stem elongation (DC31) and anthesis (DC65), biomass under ambient CO₂ was 1200 and 6200 kg ha⁻¹, respectively, but increased by 21% and 23% under eCO₂ (Table 2). There was also a corresponding increase in crop leaf area index of 21% and photosynthetic area index of 25% due to eCO₂. The increases in vegetative growth under eCO₂ resulted in a 26% increase in grain yield. For these crops, the WUE of producing biomass increased 36% (21.3–28.9 kg ha⁻¹ mm⁻¹) and for grain increased 31% (7.2–9.6 kg ha⁻¹ mm⁻¹) under eCO₂ with a corresponding small reduction (6%, P = 0.1) in total accumulated water use from 320 to 301 mm during the growing season.

Table 2. Observed gross response to elevated carbon dioxide (b) of wheat (cv. Yitpi) biomass at stem elongation (DC31) and anthesis (DC65), leaf area index at DC65 (LAI65), leaf + stem area index at DC65 (PAI65), yield, water use and water use efficiency for biomass (WUEb) and grain (WUEg) under FACE (free-air carbon dioxide enrichment) conditions

	Ambient CO₂ (365 μmol mol⁻¹)	Elevated CO₂ (550 μmol mol⁻¹)	b	SE	Low b	High b	R ²
DC31 (kg ha⁻¹)	1248 (91)	1504 (122)	1.209	0.0357	1.135	1.283	0.85
DC65 (kg ha⁻¹)	6169 (405)	7762 (452)	1.232	0.0391	1.151	1.313	0.69
LAI65 (m² m⁻²)	1.3 (0.1)	1.6 (0.2)	1.213	0.0730	1.065	1.360	0.67
PAI65 (m² m⁻²)	2.0 (0.2)	2.6 (0.3)	1.254	0.0630	1.124	1.384	0.73
Grain yield (kg ha⁻¹)	2289 (154)	2872 (214)	1.255	0.0372	1.178	1.332	0.82
Water use (mm)	320 (9)	301 (10)	0.936	0.0243	0.886	0.986	0.36
WUEb (kg ha⁻¹ mm⁻¹)	21.3 (1.3)	28.9 (2.0)	1.357	0.0407	1.273	1.441	0.79
WUEg (kg ha⁻¹ mm⁻¹)	7.2 (0.5)	9.6 (0.7)	1.311	0.0483	1.211	1.411	0.72

Data are from the AGFACE experiment for 3 years (2007–2009) and pooled across time of sowing and watering regime treatments. The fitted slope (b) of the model y = bx where y is the parameter under eCO₂ and x is the parameter under ambient CO₂ is shown together with its standard error (SE) and 95% lower and upper confidence intervals of b and its coefficient of determination (R²). The total residual degree of freedom was 23.

FACE experiment – wheat response to water availability

Early growth of wheat up to stem elongation was not affected by irrigation, but by anthesis, supplemental irrigation increased biomass and LAI by 10% and 37%, respectively, compared with the rain-fed crops (Table 3) showing the more sensitive nature of LAI to water stress. This translated through to a 19% increase in yield and 17% greater water use under irrigation treatments. There was no effect of water regime on wheat WUE for biomass or grain.

Table 3. The observed treatment-mean performance of wheat (cv. Yitpi) growth at stem elongation (DC31) and anthesis (DC65), leaf area index at DC65 (LAI), yield, water use and water use efficiency for biomass (WUEb) and grain (WUEg) under ambient (aCO₂) and elevated carbon dioxide (eCO₂) under FACE (free-air carbon dioxide enrichment) conditions

	TOS	Rain-fed	Irrigated	Rain-fed	Irrigated
	TOS	aCO₂		eCO₂
Biomass (DC31) (kg ha⁻¹)	N	1028 (132)	1004 (131)	1219 (182)	1159 (190)
Biomass (DC31) (kg ha⁻¹)	L	1393 (140)	1567 (226)	1815 (219)	1823 (279)
Biomass (DC65) (kg ha⁻¹)	N	7901 (370)	8059 (392)	9119 (581)	10 170 (531)
Biomass (DC65) (kg ha⁻¹)	L	4171 (113)	4546 (276)	5383 (122)	6376 (269)
LAI (DC65) (m² m⁻²)	N	1.7 (0.2)	2.1 (0.1)	1.9 (0.2)	2.6 (0.4)
LAI (DC65) (m² m⁻²)	L	0.6 (0.1)	0.8 (0.1)	0.7 (0.1)	1.2 (0.1)
Grain yield (kg ha⁻¹)	N	2777 (109)	3046 (202)	3391 (222)	3972 (269)
Grain yield (kg ha⁻¹)	L	1565 (188)	1768 (172)	1704 (180)	2420 (274)
Water use (mm)	N	316 (6)	354 (18)	280 (12)	357 (9)
Water use (mm)	L	272 (10)	339 (20)	278 (23)	290 (13)
WUEb (kg ha⁻¹ mm⁻¹)	N	26.4 (2.0)	26.1 (2.3)	38.9 (4.1)	33.7 (2.4)
WUEb (kg ha⁻¹ mm⁻¹)	L	16.7 (0.8)	16.2 (1.1)	19.1 (0.6)	23.9 (0.9)
WUEg (kg ha⁻¹ mm⁻¹)	N	8.8 (0.5)	8.8 (0.9)	12.4 (1.4)	11.2 (1.0)
WUEg (kg ha⁻¹ mm⁻¹)	L	5.8 (0.8)	5.5 (0.9)	6.2 (0.6)	8.3 (0.9)

Data are from the AGFACE experiment for 3 years (2007–2009) for normal (N) and late time (L) of sowing (TOS) and watering regime (rain-fed and irrigated). Standard errors of mean are in parentheses.

FACE experiment – wheat response to time of sowing

Early growth up to stem elongation was greater (50%) with late sowing, but this pattern reversed at anthesis where late sowing caused a 42% and 60% reduction in biomass and leaf area compared with early-sown wheat (Table 3). This translated to a 43% reduction in yield where crops were sown late due to crop exposure to increased temperature. Additionally, late sowing reduced water use and WUE of grain production by 10% and 37%, respectively, compared with early-sown crops.

FACE experiment – combined effect of elevated CO₂, water and time of sowing

For early growth of wheat up to stem elongation, the increase in growth (relative and absolute) due to eCO₂ was greatest (30%) for crops sown late under rain-fed conditions (Table 3). At anthesis, this pattern changed to the greatest growth response due to eCO₂. Late-sown crops under supplemental irrigation had the greatest absolute increases in growth due to eCO₂ (40%) (Table 4). For grain yield, both least and greatest response to eCO₂ occurred for rain-fed and irrigated crops, respectively, when crops were late sown, although the greatest absolute increases in yield due to eCO₂ occurred in combination when wheat was sown early under supplemental irrigation. The WUE of grain production was reduced under eCO₂ (−50%) when crops were sown late under rain-fed condition; however, when these crops were irrigated, there was a 34% increase in WUE due to eCO₂. Increases in WUE due to eCO₂ of early-sown wheat were 41% and 27% for rain-fed and irrigated crops, respectively. There were no effects of applied N on growth due to the high levels of soil mineral nitrogen content at the site (see 2).

Table 4. The observed and simulated treatment-mean performance of wheat (cv. Yitpi) growth at anthesis (kg ha⁻¹) (DC65) under ambient and elevated carbon dioxide under FACE (free-air carbon dioxide enrichment) conditions

	TOS	Ambient	Elevated	%	Ambient	Elevated	%
	TOS	Rain-fed			Irrigated
Observed	N	7901 (370)	9119 (581)	15.4	8059 (392)	10 170 (531)	26.2
Observed	L	4171 (113)	5383 (122)	29.1	4546 (276)	6376 (269)	40.3
APSIM-Wheat	N	7145 (1002)	8758 (1124)	22.6	8031 (689)	9831 (660)	22.4
APSIM-Wheat	L	5342 (710)	6463 (848)	21.0	6421 (158)	7931 (142)	23.5
APSIM-Nwheat	N	7086 (2558)	8622 (2372)	21.7	8042 (159)	9843 (410)	22.4
APSIM-Nwheat	L	3892 (915)	4480 (1156)	15.1	5244 (259)	6445 (258)	22.9
CAT-Wheat	N	6232 (654)	7449 (722)	19.5	7310 (426)	8787 (463)	20.2
CAT-Wheat	L	4435 (559)	5301 (652)	19.5	6074 (316)	7241 (433)	19.2
CROPSYST	N	6405 (929)	8202 (1032)	28.0	7080 (338)	9012 (343)	27.3
CROPSYST	L	4857 (482)	6195 (637)	27.5	5273 (94)	6755 (104)	28.1
OLEARY-CONNOR	N	6768 (1143)	8332 (1213)	23.1	6979 (716)	8818 (960)	26.3
OLEARY-CONNOR	L	4742 (744)	5841 (1192)	23.2	5438 (667)	6815 (516)	25.3
SALUS	N	7087 (201)	8181 (199)	15.5	7056 (275)	8199 (478)	16.2
SALUS	L	5679 (395)	6451 (453)	13.6	6031 (800)	7038 (800)	16.7

Data are from the AGFACE experiment for 3 years (2007–2009) for normal (N) and late time (L) of sowing (TOS) and watering regime (rain-fed and irrigated). Standard errors of mean of the observed data and standard deviation of simulated data are in parentheses.

The timing of emergence, anthesis and maturity were not affected by eCO₂ (Table 5) despite the expectation of accelerated development through raised canopy temperatures under eCO₂ (Kimball, 2011). In simulating phenological development, there were, however, larger than expected unexplained errors (6−8 days) in all of the models in simulating time to anthesis and maturity (Table 5). Despite this variation in phenological performance, all the models were able to simulate the response of biomass, yield and water use to eCO₂.

Table 5. The observed mean time to emergence, stem elongation, anthesis and maturity and root mean square error (RMSE) for six crop models

	Sowing to emergence (days)	Sowing to stem elongation (days)	Sowing to anthesis (days)	Anthesis to Maturity (days)
Observed	13.3	68.7	109.0	40.2
RMSE
APSIM-Wheat	3.3	11.1	8.7	8.3
APSIM-Nwheat	3.7	12.9	8.8	9.1
CAT-Wheat	3.3	6.5	7.8	6.6
CROPSYST	1.7	na	6.4	11.2
OLEARY-CONNOR	1.6	10.6	7.9	9.2
SALUS	3.0	7.7	8.6	13.7

No differences were observed under elevated CO_2, so data were pooled across CO₂ levels. CROPSYST does not simulate stem elongation.
na, not available.

Simulation performance

Simulation performance – biomass and yield

Despite large variation in the data, the simulated biomass and grain yield showed no serious bias, and all models provided a similar mean error (Table S9). There were, however, some differences between the models in the absolute simulation of early biomass and LAI. Comparison of modelled growth over time at the early growth stage of DC31 is somewhat difficult without observations from multiple sampling dates, because the cause of mismatch may not be clear from a single DC31 measurement. Nevertheless, the overall response to eCO₂ was similar among the models (Table S10).

For crop growth up to anthesis, RMSE between observed and simulated biomass ranged from 1400 to 1500 kg ha⁻¹ across the six models (Table S9). All the models tended to simulate biomass until anthesis reasonably well, despite over prediction of LAI. Both APSIM-Wheat (R² = 0.24 and RMSE = 0.70 m² m⁻²) and APSIM-Nwheat (R² = 0.39 and RMSE = 1.18 m² m⁻²) simulated LAI closer to the observed data than the other models simulating PAI. No differences in model performance in simulating biomass to anthesis occurred between the ambient and eCO₂ treatments. For crop yield (Table S9), comparison of RMSE between observed and predicted ranged from 600 to 1300 kg ha⁻¹. The SALUS model had the highest R² (0.69) and the lowest RMSE (600 kg ha⁻¹) with CAT-Wheat and OLEARY-CONNOR models exhibiting typical means errors around 750 kg ha⁻¹ and the APSIM-Wheat and APSIM-Nwheat >1000 kg ha⁻¹. This variation among the models highlights the extent of accuracy that can be expected using such models without fitting the models to the data and is considered sufficiently accurate to test the relative response of eCO₂.

Simulation performance – water use and water use efficiency

The average water use across all treatments was 310 mm during the growing season and ranged from 272 to 357 mm (Table 3). All the models simulated water use with a relatively high degree of accuracy and consistency with R² ranging from 0.58 to 0.31 and RMSE from 31.3 to 67.6 mm (Table S9). No differences in model performance in simulating water use occurred between the ambient and eCO₂ treatments. There was some variation between the models in simulating WUE for biomass and grain yield. For the accumulation of biomass, the water use efficiency (WUEb) was simulated with varied accuracy by APSIM-Wheat, APSIM-Nwheat, CAT-Wheat, CROPSYST and SALUS with R² ranging from 0.18 to 0.84 and RMSE from 6.63 to 9.69 kg ha⁻¹ mm⁻¹). For grain yield, the water use efficiency (WUEg) R² ranged from 0.03 to 0.51 and RMSE from 2.48 to 4.12 kg ha⁻¹ mm⁻¹ (Table S9).

Biomass and leaf area response to eCO₂

The observed increase in growth of wheat, up to stem elongation, due to eCO₂ was 21% (using regression) (Table S10). This early growth response was not captured well by all the models where APSIM-Wheat, APSIM-Nwheat, CAT-Wheat, CROPSYST, OLEARY-CONNOR and SALUS produced a predicted response to eCO₂ of 29%, 18%, 25%, 40%, 45% and 9%, respectively (Fig. 2 and Table S10). The models: APSIM-Wheat, CAT-Wheat and SALUS substantially oversimulated early biomass (stem elongation), whereas the OLEARY-CONNOR model undersimulated the amount of biomass but oversimulated the response to eCO₂ for the same period. Wheat growth up to anthesis increased on average (using regression) by 23% due to eCO₂ (Fig. 3 and Table S10). In comparison, simulated response was 22%, 21%, 20%, 28%, 25% and 16% for the APSIM-Wheat, APSIM-Nwheat, CAT-Wheat, CROPSYST, OLEARY-CONNOR and SALUS models, respectively. All the six models performed within or close to the 95% confidence limits of observed slope. However, in contrast to the oversimulation of early growth, SALUS was not able to capture the observed range of biomass. The simulated increase in LAI at DC65 from eCO₂ was 19% for both APSIM-Wheat and APSIM-Nwheat and 13%, 7%, 24% and 8% for PAI for CAT-Wheat, CROPSYST, OLEARY-CONNOR and SALUS models, respectively (Fig. 4 and Table S10). For both APSIM models, the range of both observed and simulated data matched, whereas the CAT-Wheat, CROPSYST, OLEARY-CONNOR and SALUS did not match well the observed range of PAI. The observed and simulated biomass and LAI response to eCO₂ at DC65 (50% anthesis) under irrigation was generally increased despite irrigation not fully preventing stress (Table S10). Under irrigation, the effect of late sowing on response to eCO₂ to biomass accumulation at DC65 was substantial in the observed data (~40%), but the simulated response ranged from 17% from SALUS to 28% from CROPSYST (Table 4). Simulated response from all the six models under no water or nitrogen stress showed similar response to eCO₂ under irrigation, but the differences compared to the dryland treatment were small.

Yield response to eCO₂

The observed mean wheat yield response to eCO₂ was 26% (using regression) (Fig. 5 and Table S10). Similarly, simulations match well the response with simulated increases of 19%, 28%, 20%, 27%, 21% and 25% for the APSIM-Wheat, APSIM-Nwheat, CAT-Wheat, CROPSYST, OLEARY-CONNOR and SALUS models, respectively. There was a tendency for APSIM-Wheat to slightly oversimulate the range of yields, but all the other models simulated the response to eCO₂ and range well (Fig. 5).

Water use and water use efficiency response to eCO₂

The water use during the growing season of wheat was slightly reduced by eCO₂, but simulated water use from all the models showed no consistent reduction (Table S10). Some models actively reduce daily transpiration under eCO₂ conditions (e.g. CROPSYST and OLEARY-CONNOR), but the accumulated WU (Table S10) and T (data not shown) were near 1 : 1 with respect to eCO₂. Of all the parameters, simulated water use gave the most consistent nonresponse to eCO₂ (Table S10). The observed water use efficiency of biomass (WUEb) and wheat yield (WUEg) increased by 36% and 31% (using regression) under eCO₂, respectively (Tables 2 and 4). In comparison, simulations tended to underestimate the effect of eCO₂ on WUEb and WUEg but were closest for APSIM-Nwheat (22% and 27%) and CROPSYST (29% and 29%), respectively. Given the large variance in WUEb and WUEg data, differences between the models are not considered significant (Tables 4 and S9).

Discussion

The measured response of wheat cv. Yitpi to eCO₂ over the 3 years was strong and consistent for all above-ground biomass at DC31 and DC65, and LAI/PAI at DC65 and grain yield. The response is considered large but not necessarily indicative of some future climate because the experimental conditions evaluated CO₂ response under ambient temperature and VPD conditions and did not consider likely future elevated temperature or increased VPD. Nevertheless, an interactive effect of temperature was seen in the observed data where the late-sown crops consistently responded proportionally greater to eCO₂ than the normal time of sowing (Table 4). All the six models did not match this observed interaction, and a closer focus on such response is needed in future experiments.

The overall observed response to eCO₂ (~550 μmol mol⁻¹) of around 21–23% in biomass, LAI and yield was slightly higher than reported meta-analysis means around ~18% (Ainsworth & Long, 2005; Sun et al., 2009), but near identical expected for dry environments (cf. ~22% in Arizona, USA, Kimball, 2011). The assumption that there is a larger response under water-stressed conditions of the experiment is, however, not supported by the experimental data nor modelling data. Indeed, the measured response to eCO₂ of biomass accumulation by anthesis was higher under irrigation (28% cf. 18%). This response was contrary to that observed in Arizona whereby unstressed conditions had a lower response to eCO₂ (cf. ~16% Kimball, 2011). The reasons are unclear and warrant further investigation.

This magnitude of response to eCO₂ was also close to the findings of Yang et al. (2007a) that reported an average increase of 24.8% in wheat grain yield across various N levels over 2 years in Nanjing, China. These two FACE experiments were conducted under near-identical CO₂ concentrations (365 vs. 550 μmol mol⁻¹ for AGFACE; 350 vs. 550 μmol mol⁻¹ for China FACE). The response in biomass to eCO₂ was reported to be an average of 13.6% across key growth stages (Yang et al., 2007b). The extent of the increase in biomass from China FACE is relatively small compared with our findings. This was partly due to their reported negative impact of eCO₂ on biomass accumulation from heading to maturity (−5.5%) even though positive effects were found in the pre-anthesis growth phases ranging from 11% to 41% (Yang et al., 2007b). It is possible that excess pre-anthesis storage and translocation resulted in greater yield gain.

The modelled gross response to eCO₂ was similar among all models despite their different primary mechanism of computing daily growth. None of the models were specifically fitted to the data and thus show a robustness that should allow confident application to climate change scenarios involving eCO₂ in combination with varying water and N supply and growing season temperatures. Despite the good overall performance of each model with respect to response to eCO₂ against the measured data at Horsham, differences were evident. This was particularly noted by growth stage DC31 where without temperature corrections to the TE-RUE functions OLEARY-CONNOR and CROPSYST appear to overpredict the response to eCO₂ (45% and 40%, respectively, Fig. 2 and Table S10), but this is a relatively small error occurring so early in the season that this apparent discrepancy did not affect later growth. APSIM-Nwheat followed the early observed response data more closely than all the other models. We concur with Ewert (2004) regarding the need to improve LAI simulation to achieve the correct CO₂ response in all the models except the APSIM-Nwheat model. Nevertheless, despite the poorly modelled early growth by anthesis, all the models matched the seasonal biomass production quite well. Because the grain yield response to eCO₂ was simulated well by all models, the consequences of poor early simulation were not detectable by maturity at this location. This is a particularly critical finding for the applicability of these models to other similar dryland areas where incomplete cover is a dominant feature of the growing season.

Another potential uncertainty is the artefacts of high frequency changes in CO₂ concentration in FACE raised by Bunce (2013). As the modelled response does not contain any physiological mechanisms that might mimic such processes, and the modelled crop response at 550 μmol mol⁻¹ was within the experimental error for all models for biomass at DC65, we consider the pulse CO₂ injection artefact to likely be small and within the experimental error of our experiment. However, whilst the mean response to eCO₂ was consistent across the range of treatments and years, results from individual crops did exhibit large variability. The effect of such variability was reduced by the forcing of the regression through the origin (Table S10). Similarly, the 6-model ensemble generally reduced the modelled variability, but some models matched the observed data better than the ensemble (e.g. SALUS in grain yield). This is not unexpected because the ensemble represents a statistical outcome rather than an explanatory outcome represented by each model. Careful application of such ensemble data is therefore needed.

Despite the modelled response being within or close to the experimental error, an obvious issue of FACE experiments is the spatial and temporal variation in CO₂ concentrations. The destructive samples for the three growth stages were collected at least 1 m away from CO₂ delivery rings, thus avoiding the areas of highest CO₂ concentrations, at times over 1200 μmol mol⁻¹ within 1 m of the ring (Mollah et al., 2012). Median plot CO₂ concentrations varied in 2009 from 524 to 640 μmol mol⁻¹, depending on prevailing winds, and were as high as 722 μmol mol⁻¹ in one plot in 2007. Whilst the equipment maintained its designed 550 μmol mol⁻¹ in the centre, no crops were grown at the centre and we suggest that a 650-μmol mol⁻¹ level is probably more realistic for majority of the area within the rings for the measured crops, as noted in the standard deviation ring map in Mollah et al. (2012).

The lack of a phenological response to eCO₂ is an interesting result because of the strong growth response; however, the differences of <2 days may not have been noted in the AGFACE experiment. At Nanjing in China, eCO₂-accelerated phenological development was reported by 1.3 days to heading and another 1.3 days between heading and maturity (Yang et al., 2007a). Such very small changes in phenology are unlikely to gain any large benefit in terms of water savings that might ultimately occur in high-VPD environments. Future FACE experiments should consider measuring canopy temperatures throughout the day and night to better explain water fluxes with canopy temperature and crop development. The performance of all the models in terms of predicting anthesis date with RMSE's over 6 days provides room for improvement; however, notably only one crop in 2008 was responsible for the large error. Interestingly, uncertainties in simulating the impact of increasing global temperature and temperature by eCO₂ interactions have been recently pointed out as the largest shortcomings in wheat models for simulating future climate change impacts, with additional experimental research required for targeted model improvements (Asseng et al., 2013a). Our analyses support that view.

The simulated WUE was quite poor particularly for grain yield despite significant performance in WU and biomass and yield (Table S9). This probably reflects the problem of compounding of errors in both the numerator and denominator in calculating WUE. A way to dampen such effects would be to fit regressions through the prime data with various sources of variance reduced such as removing soil evaporation and examining the TE dynamics. Nevertheless, the modelled response matches the observed first-order processes sufficiently well with acceptable absolute errors (Table S10).

The gross response to eCO₂ was simulated satisfactorily; however, on close examination, there are aspects requiring model improvement. For APSIM-Wheat, CAT-Wheat, OLEARY-CONNOR and SALUS models, the primary focus should be on early growth in terms of biomass and leaf and stem components and water relations under drought and unstressed conditions. Despite simple temperature reduction via RUE in CROPSYST and OLEARY-CONNOR, a temperature interactive term with eCO₂ might be needed. This would allow very high responses to eCO₂ to be reduced early in the season when mean temperatures are typically below 10 °C and TE is high because of the low VPD. Similarly, the failure of all models to simulate the eCO₂ response under late sowing might be indicative of a poorly understood eCO₂ response under such conditions. For CAT-Wheat and SALUS, the simple RUE approach with no TE or temperature reduction and a simple harvest index method simulated the gross response reasonably well, but its interaction of eCO₂ and water supply conditions warrants further experimentation because of the relatively high early growth. New field experiments are needed focusing on early growth and water relations. The phenological development should also be scrutinised in future experiments with a focus on canopy heating and cooling as water fluxes change under FACE conditions.

The FACE experiment was conducted under ambient temperature and VPD conditions that are unlikely to be directly representative of future climates. It remains to be seen what the effect of high temperatures and VPD when combined with eCO₂ will be on leaf area development, tiller number, above-ground biomass and grain yield. Certainly, the large response to eCO₂ under late sowing shows that the interactive effects will be of continuing interest. It was recently shown that biomass and particularly grain yield of vigorous and nonvigorous wheat genotypes increased when grown under eCO₂ and a mean temperature of 1.0 °C above the mean ambient temperature (Dias De Oliveira et al., 2012). However, when grown under eCO₂ and mean temperatures of 2.0 and 3.0 °C above the mean ambient temperature, grain yields tended to decrease linearly, mainly because the combination of eCO₂ and high temperature reduced the time between anthesis and physiological maturity (Dias De Oliveira et al., 2012). Consequently, it is expected that the predicted increments in the temperature for the Mediterranean climatic region of Australia (Christensen et al., 2007) will reduce the grain yield response of wheat to eCO₂ unless wheat cultivars with high tolerance to temperature are developed. We suggest field experimentation using open air heating for heating the air around field-grown crops be explored (Kimball & Conley, 2009).

This experiment at Horsham has allowed for the first time the testing of several crop models to eCO₂ under semi-arid unirrigated conditions. At this Mediterranean-type climatic site, the response to eCO₂ was higher under irrigation. However, no elevated temperature treatments were explicitly applied in our FACE experiment, and the models did not fully match the observed interaction of the late-sown crops; therefore, some caution is required when applying the models to some future climate. These findings can therefore be cautiously used as representative of semi-arid environments, and the models used here can equally be applied in eCO₂ impact and adaptation studies for other similar semi-arid regions of the world.

Overall, without any amendment, six different crop models were able to simulate gross response to eCO₂ at Horsham, Australia – notwithstanding some second-order deviations at the early growth stage for APSIM-Wheat, CAT-Wheat, CROPSYST, OLEARY-CONNOR and SALUS models. Whilst APSIM-Nwheat performed well across many variables, it was not the better performer for all of them. General improvements in all the models are expected with future local calibration, but the RUE and TE methods of biomass accumulation represented in the models were sufficient in reproducing the main responses in the FACE experiment and therefore applicable for future climate change analyses in similar dry environments where increased water supply increases the response to eCO₂ at around 550 μmol mol⁻¹.

Acknowledgements

The research at the Australian Grains Free-Air Carbon Dioxide Enrichment (AGFACE) facility was jointly managed by the Victorian State Government Department of Environment and Primary Industries and The University of Melbourne with additional funding from the Australian Grains Research and Development Corporation and the Australian Government Department of Agriculture, Fisheries and Forestry. Technical and engineering support was provided by Mahabubur Mollah, Russel Argall, Peter Howie, Janine Fitzpatrick, Justine Ellis, Jennifer Briggs and Jason Ellifson. At the outset of the project, we invited guest collaborators to join the team (O'Leary & Anwar, 2008; O'Leary et al., 2010). Participation of CS and CROPSYST simulations was supported by the project of Regional Approaches to Climate Change for Pacific Northwest Agriculture (REACCH-PNA) which is funded through award #2011-68002-30191 from the US National Institute for Food and Agriculture. We acknowledge helpful discussions during the project from Doug Arbrect, Jeanphilippe Aurambout, Craig Beverly, Hamish Brown, Deli Chen, Kithsiri Bandara Dassanayake, Andrew Fletcher, Peter Grace, Bruce Kimball, Robert Norton, Eduardo Dias de Oliveira, David Stephens, Humaira Sultana, Michael Tausz, Gerard Wall and Jeff White. The experimental data set comprising meteorological observations and treatment crop and soil data from the 48 crops presented here are available upon request from the senior author. Helpful comments were made by an anonymous referee on an earlier manuscript.

Author contribution

The various contributions from the authors were as follows: concept, primary author: GOL; modelling data contributions: GOL, BC, JN, NH, DC, CS, BB, IS and SA; and discussion contributions with respect to other published data: GF, QL, IFC and JP.

Supporting Information

Filename

Description

gcb12830-sup-0001-TableS1-S10.docWord document, 452 KB

Table S1. Soil profile data describing the mean air dry, crop lower limit (LL), drained upper limit (DUL), saturation (SAT), bulk density (BD), pH, organic carbon (OC) and organic nitrogen (ON) content of the experimental site used by all models. APSIM-Wheat used additional water availability coefficients (KL and XF).

Table S2. Additional generic site specific and soil type parameters defining the experimental site including some APSIM-Wheat specific parameters (APSIM).

Table S3. Key parameters used to define wheat growth in APSIM-Wheat for cultivar Yitpi.

Table S4. Key parameters used to define wheat growth in APSIM-Nwheat for cultivar Yitpi.

Table S5. Key parameters used to define wheat growth in CAT-Wheat for cultivar Yitpi.

Table S6. Key parameters used to define wheat growth in CROPSYST for cultivar Yitpi.

Table S7. Key parameters used to define wheat growth in the OLEARY-CONNOR model for cultivar Yitpi.

Table S8. Key parameters used to define wheat growth in SALUS for cultivar Yitpi.

Table S9. Comparison of simulated and observed data showing regression coefficients (a, b), their standard errors (SE), 95% lower and upper confidence intervals of b, coefficient of determination (R²), significance of regression (P) and root mean square error (RMSE) for the linear model y = a + bx where y is the observed and x is the simulated data of biomass at DC31 (BIO31 kg ha⁻¹), biomass at DC65 (BIO65 kg ha⁻¹), leaf area index (LAI at DC65 m² m⁻²), leaf and stem area index (PAI at DC65 m² m⁻²), grain yield (GY kg ha⁻¹), water use (WU mm), water use efficiency for biomass (WUEb kg ha⁻¹ mm⁻¹) and water use efficiency for grain (WUEg kg ha⁻¹ mm⁻¹) for the APSIM-Wheat, APSIM-Nwheat, CAT-Wheat, CROPSYST, OLEARY-CONNOR, SALUS wheat models and the 6-model ENSEMBLE. The residual degrees of freedom was 46. ^aCROPSYST does not simulate stage DC31 but simulated biomass was output on the observed date of DC31 for comparison.

Table S10. Regression coefficients of the mean response to elevated CO₂ on biomass at DC31, biomass at DC65, leaf area index at DC65 (LAI at DC65), leaf + stem area index (PAI at DC65), grain yield, water use and water use efficiency for biomass and grain for the observed experimental data (OBS) and APSIM-Wheat, APSIM-Nwheat, CAT-Wheat, CROPSYST, OLEARY-CONNOR, SALUS models and the 6-model ENSEMBLE. The response under dryland and irrigated conditions and simulated unstressed conditions showing the varied general effect of added water to the response to eCO₂. The fitted slope (b) of the model y = bx where y is the parameter under eCO₂ (550 μmol mol⁻¹) and x is the parameter under ambient CO₂ (365 μmol mol⁻¹) is shown together with its standard error (SE) and 95% lower and upper confidence intervals of b for the observed data. The residual degrees of freedom was 23. ^aCROPSYST does not simulate stage DC31 but simulated biomass was output on the observed date of DC31 for comparison.

Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

References

Ainsworth E, Long S (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.
10.1111/j.1469-8137.2004.01224.x
PubMed Web of Science® Google Scholar
Allen R (1990) REF-ET Reference Evapotranspiration Calculator V. 2.0. Operations Manual, 1990. Utah State University, Logan, UT.
Google Scholar
Amthor J (2001) Effects of atmospheric CO₂ concentration on wheat yield: review of results from experiments using various approaches to control CO₂ concentration. Field Crops Research, 73, 1–34.
10.1016/S0378-4290(01)00179-4
Web of Science® Google Scholar
Anwar M, O'Leary G, McNeil D, Hossain H, Nelson R (2007) Climate change impact on rainfed wheat in south-eastern Australia. Field Crops Research, 104, 139–147.
10.1016/j.fcr.2007.03.020
Web of Science® Google Scholar
Asseng S, Van Herwaarden A (2003) Analysis of the benefits to wheat yield from assimilates stored prior to grain filling in a range of environments. Plant and Soil, 256, 217–229.
10.1023/A:1026231904221
CAS Web of Science® Google Scholar
Asseng S, Anderson G, Dunin F, Fillery I, Dolling P, Keating B (1998a) Use of the APSIM wheat model to predict yield, drainage, and NO₃ leaching for a deep sand. Australian Journal of Agricultural Research, 49, 363–378.
10.1071/A97095
CAS Web of Science® Google Scholar
Asseng S, Keating B, Fillery I et al. (1998b) Performance of the APSIM-wheat model in Western Australia. Field Crops Research, 57, 163–179.
10.1016/S0378-4290(97)00117-2
Web of Science® Google Scholar
Asseng S, Van Keulen H, Stol W (2000) Performance and application of the APSIM Nwheat model in the Netherlands. European Journal of Agronomy, 12, 37–54.
10.1016/S1161-0301(99)00044-1
Web of Science® Google Scholar
Asseng S, Turner N, Ray J, Keating B (2002) A simulation analysis that predicts the influence of physiological traits on the potential yield of wheat. European Journal of Agronomy, 17, 123–141.
10.1016/S1161-0301(01)00149-6
Web of Science® Google Scholar
Asseng S, Turner N, Botwright T, Condon A (2003) Evaluating the impact of a trait for increased specific leaf area on wheat yields using a crop simulation model. Agronomy Journal, 95, 10–19.
10.2134/agronj2003.1000b
Web of Science® Google Scholar
Asseng S, Jamieson P, Kimball B, Pinter P, Sayre K, Bowden J, Howden S (2004) Simulated wheat growth affected by rising temperature, increased water deficit and elevated atmospheric CO₂. Field Crops Research, 85, 85–102.
10.1016/S0378-4290(03)00154-0
Web of Science® Google Scholar
Asseng S, Foster I, Turner N (2011) The impact of temperature variability on wheat yields. Global Change Biology, 17, 997–1012.
10.1111/j.1365-2486.2010.02262.x
Web of Science® Google Scholar
Asseng S, Ewert F, Rosenzweig C et al. (2013a) Uncertainty in simulating wheat yields under climate change. Nature Climate Change, 3, 827–832.
10.1038/nclimate1916
CAS Web of Science® Google Scholar
Asseng S, Travasso M, Ludwig F, Magrin G (2013b) Has climate change opened new opportunities for wheat cropping in Argentina? Climatic Change, 117, 181–196.
10.1007/s10584-012-0553-y
Web of Science® Google Scholar
Basso B, Ritchie J, Grace P, Sartori L (2006) Simulating tillage impacts on soil biophysical properties using the SALUS model. Italian Journal of Agronomy, 4, 688.
Google Scholar
Basso B, Bertocco M, Sartori L, Martin EC (2007) Analyzing the effects of climate variability on spatial pattern of yield in a maize-wheat-soybean rotation. European Journal of Agronomy, 26, 82–91.
10.1016/j.eja.2006.08.008
Web of Science® Google Scholar
Basso B, Cammarano D, Chen D et al. (2009) Landscape position and precipitation effects on spatial variability of wheat yield and grain protein in Southern Italy. Journal of Agronomy and Crop Science, 195, 301–312.
10.1111/j.1439-037X.2008.00351.x
Web of Science® Google Scholar
Basso B, Cammarano D, Troccoli A, Chen D, Ritchie J (2010) Long-term wheat response to nitrogen in a rainfed Mediterranean environment: field data and simulation analysis. European Journal of Agronomy, 33, 132–138.
10.1016/j.eja.2010.04.004
Web of Science® Google Scholar
Basso B, Ritchie J, Cammarano D, Sartori L (2011a) A strategic and tactical management approach to select optimal N fertilizer rates for wheat in a spatially variable field. European Journal of Agronomy, 35, 215–222.
10.1016/j.eja.2011.06.004
Web of Science® Google Scholar
Basso B, Sartori L, Bertocco M, Cammarano D, Grace P (2011b) Economic and environmental evaluation of site-specific tillage in a maize crop in NE Italy. European Journal of Agronomy, 35, 83–92.
10.1016/j.eja.2011.04.002
Web of Science® Google Scholar
Basso B, Sartori L, Cammarano D, Grace P, Sorensen C, Fountas S (2012) Environmental and economic evaluation of N fertilizer rates in a maize crop in Italy: a spatial and temporal analysis using crop models. Biosystems Engineering, 113, 103–111.
10.1016/j.biosystemseng.2012.06.012
Web of Science® Google Scholar
Bassu S, Asseng S, Motzo R, Giunta F (2009) Optimising sowing date of durum wheat in a variable Mediterranean environment. Field Crops Research, 111, 109–118.
10.1016/j.fcr.2008.11.002
Web of Science® Google Scholar
Bassu S, Brisson N, Durand J et al. (2014) How do various maize crop models vary in their responses to climate change factors. Global Change Biology, 20, 2301–2320.
10.1111/gcb.12520
CAS PubMed Web of Science® Google Scholar
Boote K, Jones J, Pickering N (1996) Potential uses and limitations of crop models. Agronomy Journal, 88, 704–716.
10.2134/agronj1996.00021962008800050005x
Web of Science® Google Scholar
Bunce J (2013) Effects of pulses of elevated carbon dioxide concentration on stomatal conductance and photosynthesis in wheat and rice. Physiologia plantarum, 149, 214–221.
10.1111/ppl.12026
CAS PubMed Web of Science® Google Scholar
Carberry P, Liang W-L, Twomlow S et al. (2013) Scope for improved eco-efficiency varies among diverse cropping systems. Proceedings of the National Academy of Sciences of the United States of America, 110, 8381–8386.
10.1073/pnas.1208050110
CAS PubMed Web of Science® Google Scholar
Christensen J, Hewitson B, Busuioc A et al. (2007) Regional climate projections. In: Climate Change, 2007: The Physical Science Basis. Contribution of Working group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. (eds S Solomon, D Qin, M Manning, Z Chen, M Marquis, K Averyt, M Tignor, H Miller), pp. 847–940. Cambridge University Press, Cambridge, UK and New York, NY.
Web of Science® Google Scholar
Christy B, O'Leary G, Riffkin P, Acuna T, Potter T, Clough A (2013) Long-season canola (Brassica napus L.) cultivars offer potential to substantially increase grain yield production in south-eastern Australia compared with current spring cultivars. Crop and Pasture Science, 64, 901–913.
10.1071/CP13241
Web of Science® Google Scholar
Dias De Oliveira E, Bramley H, Siddique K, Henty S, Berger J, Palta J (2012) Can elevated CO₂ combined with high temperature ameliorate the effect of terminal drought in wheat? Functional Plant Biology, 40, 160–171.
10.1071/FP12206
CAS Web of Science® Google Scholar
Díaz-Ambrona C, O'Leary G, Sadras V, O'connell M, Connor D (2005) Environmental risk analysis of farming systems in a semi-arid environment: effect of rotations and management practices on deep drainage. Field Crops Research, 94, 257–271.
10.1016/j.fcr.2005.01.008
Web of Science® Google Scholar
Donatelli M, Stöckle C, Ceotto E, Rinaldi M (1997) Evaluation of CropSyst for cropping systems at two locations of northern and southern Italy. European Journal of Agronomy, 6, 35–45.
10.1016/S1161-0301(96)02029-1
Web of Science® Google Scholar
Dzotsi KA, Basso B, Jones J (2013) Development, uncertainty and sensitivity analysis of the simple SALUS crop model in DSSAT. Ecological Modeling, 260, 62–76.
10.1016/j.ecolmodel.2013.03.017
Web of Science® Google Scholar
Ewert F (2004) Modelling plant responses to elevated CO₂: How important is leaf area index? Annals of botany, 93, 619–627.
10.1093/aob/mch101
PubMed Web of Science® Google Scholar
Ferrer-Alegre F, Stöckle C (1999) A model for assessing crop response to salinity. Irrigation Science, 19, 15–23.
10.1007/s002710050067
Web of Science® Google Scholar
Giola P, Basso B, Pruneddu G, Giunta F, Jones J (2012) Impact of manure and slurry applications on soil nitrate in a maize-triticale rotation: field study and long term simulation analysis. European Journal of Agronomy, 38, 43–53.
10.1016/j.eja.2011.12.001
Web of Science® Google Scholar
Hochman Z, Holzworth D, Hunt J (2009) Potential to improve on-farm wheat yield and WUE in Australia. Crop and Pasture Science, 60, 708–716.
10.1071/CP09064
Web of Science® Google Scholar
Holzworth D, Huth N, Devoil P (2011) Simple software processes and tests improve the reliability and usefulness of a model. Environmental Modelling & Software, 26, 510–516.
10.1016/j.envsoft.2010.10.014
Web of Science® Google Scholar
Hoogenboom G, Tsuji G, Pickering N, Curry R, Jones J, Singh U, Godwin D (1995) Decision support system to study climate change impacts on crop production. In: Climate Change and Agriculture: Analysis of Potential International Impacts (eds C Rosenzweig, L Allen, L Harper, S Hollinger, J Jones), pp. 51–75. American Society of Agronomy, Madison, WI.
Web of Science® Google Scholar
Howden S, O'Leary G (1997) Evaluating options to reduce greenhouse gas emissions from an Australian temperate wheat cropping system. Environmental Modelling & Software, 12, 169–176.
10.1016/S1364-8152(97)00007-8
Web of Science® Google Scholar
Howden SM, Ash AJ, Barlow EWR et al. (2003) An overview of the adaptive capacity of the Australian agricultural sector to climate change – options, costs and benefits. Report to the Australian Greenhouse Office. Canberra, Australia. p. 157.
Google Scholar
Jamieson P, Berntsen J, Ewert F et al. (2000) Modelling CO₂ effects on wheat with varying nitrogen supplies. Agriculture, Ecosystems and Environment, 82, 27–37.
10.1016/S0167-8809(00)00214-0
CAS Web of Science® Google Scholar
Keating B, Carberry P, Hammer G et al. (2003) An overview of APSIM, a model designed for farming systems simulation. European Journal of Agronomy, 18, 267–288.
10.1016/S1161-0301(02)00108-9
Web of Science® Google Scholar
Kimball B (2011) Lessons from FACE: CO₂ effects and interactions with water, nitrogen and temperature. In: ICP Series on Climate Change Impacts, Adaption and Mitigation: Volume 1 Handbook of Climate Change and Agroecosystems: Impacts, Adaptation, and Mitigation (eds D Hillel, C Rozenzweig), pp. 87–107. Imperial College Press, London UK.
Google Scholar
Kimball B, Conley M (2009) Infrared heater arrays for warming field plots scaled up to 5-m diameter. Agricultural and Forest Meteorology, 149, 721–724.
10.1016/j.agrformet.2008.09.015
Web of Science® Google Scholar
Ko J, Ahuja L, Kimball B et al. (2010) Simulation of free air CO₂ enriched wheat growth and interactions with water, nitrogen, and temperature. Agricultural and Forest Meteorology, 150, 1331–1346.
10.1016/j.agrformet.2010.06.004
Web of Science® Google Scholar
Kremer C, Stöckle C, Kemanian A, Howell T (2008) A reference canopy transpiration and photosynthesis model for the evaluation of simple models of crop productivity. In: Response of Crops to Limited Water (eds L Ahuja, V Reddy, S Saseendran, Q Yu), pp. 165–190. Advances in Agricultural Systems Modeling. American Society of Agronomy, Inc., Crop Science Society of America, Inc., Soil Science Society of America, Inc., Madison, WI.
Google Scholar
Latta J, O'Leary G (2003) Long-term comparison of rotation and fallow tillage systems of wheat in Australia. Field Crops Research, 82, 173–190.
10.1016/S0378-4290(03)00073-X
Web of Science® Google Scholar
Leakey A, Ainsworth E, Bernacchi C, Rogers A, Long S, Ort D (2009) Elevated CO₂ effects on plant carbon, nitrogen, and water relations: six important lessons from FACE. Journal of Experimental Botany, 60, 2859–2876.
10.1093/jxb/erp096
CAS PubMed Web of Science® Google Scholar
Littleboy M, Silburn D, Freebairn D et al. (1992) Impact of soil erosion on production in cropping systems. I. development and validation of a simulation model. Australian Journal of Soil Research, 30, 757–774.
10.1071/SR9920757
Web of Science® Google Scholar
Lobell D, Ortiz-Monasterio J (2007) Impacts of day versus night temperatures on spring wheat yields. Agronomy Journal, 99, 469–477.
10.2134/agronj2006.0209
Web of Science® Google Scholar
Long S, Ainsworth E, Leakey A, Nösberger J, Ort D (2006) Food for thought: lower-than-expected crop yield stimulation with rising CO₂ concentrations. Science, 312, 1918–1921.
10.1126/science.1114722
CAS PubMed Web of Science® Google Scholar
Ludwig F, Asseng S (2006) Climate change impacts on wheat production in a Mediterranean environment in Western Australia. Agricultural Systems, 90, 159–179.
10.1016/j.agsy.2005.12.002
Web of Science® Google Scholar
McCown R, Hammer G, Hargreaves J, Holzworth D, Freebairn D (1996) APSIM: a novel software system for model development, model testing and simulation in agricultural systems research. Agricultural Systems, 50, 255–271.
10.1016/0308-521X(94)00055-V
Web of Science® Google Scholar
Mollah M, Norton R, Huzzey J (2009) Australian grains free-air carbon dioxide enrichment (AGFACE) facility: design and performance. Crop and Pasture Science, 60, 697–707.
10.1071/CP08354
CAS Web of Science® Google Scholar
Mollah M, Partington D, Fitzgerald G (2012) Understand distribution of carbon dioxide to interpret crop growth data: Australian grain free air carbon dioxide enrichment experiment. Crop and Pasture Science, 62, 883–891.
10.1071/CP11178
CAS Web of Science® Google Scholar
Monteith JL (1981) Evaporation and surface temperature. Quarterly Journal of the Royal Meteorological Society, 107, 1–27.
10.1002/qj.49710745102
Web of Science® Google Scholar
O'Leary G, Anwar M (2008) Simulation of wheat grain yield under elevated CO₂ conditions in the Victorian Wimmera. In: Global Issues, Paddock Action: Proceedings of the 14th Australian Agronomy Conference (ed. M Unkovich), Australian Society of Agronomy/The Regional Institute, Adelaide, South Australia.
Google Scholar
O'Leary G, Connor D (1996a) A simulation model of the wheat crop in response to water and nitrogen supply: I. Model construction. Agricultural Systems, 52, 1–29.
10.1016/0308-521X(96)00003-0
Web of Science® Google Scholar
O'Leary G, Connor D (1996b) A simulation model of the wheat crop in response to water and nitrogen supply: II. Model validation. Agricultural Systems, 52, 31–55.
10.1016/0308-521X(96)00002-9
Web of Science® Google Scholar
O'Leary G, Connor D (1997) Stubble retention and tillage in a semi-arid environment: 3. Response of wheat. Field Crops Research, 54, 39–50.
10.1016/S0378-4290(97)00036-1
Web of Science® Google Scholar
O'Leary G, Connor D (1998) A simulation study of wheat crop response to water supply, nitrogen nutrition, stubble retention and tillage. Australian Journal of Agricultural Research, 49, 11–19.
10.1071/A97020
Web of Science® Google Scholar
O'Leary G, Connor D, White D (1985) Effect of sowing time on growth, yield and water-use of rain-fed wheat in the Wimmera, Vic. Crop and Pasture Science, 36, 187–196.
10.1071/AR9850187
Web of Science® Google Scholar
O'Leary G, Christy B, Weeks A et al. (2010) Modelling the growth and gain yield response of wheat crops under Free Air Carbon Dioxide Enrichment at Horsham and throughout Victoria, Australia. In: Food Security form Sustainable Agriculture. Proceedings of the 15th ASA Conference (eds H Dove, R Culvenor), pp. 15–19. Australian Society of Agronomy, Lincoln, New Zealand.
Google Scholar
O'Leary G, Christy B, Weeks A, Nuttall J, Riffkin P, Beverly C, Fitzgerald G (2011) Downscaling global climatic predictions to the regional level: a case study of regional effects of climate change on wheat crop production in Victoria, Australia. In: Crop Adaptation to Climate Change (eds S Shyam, RJR Yadav, JL Hatfield, H Lotze-Campen, AE Hall), pp. 12–26. John Wiley & Sons, Ltd., Chichester, West Sussex, UK.
10.1002/9780470960929.ch2
Google Scholar
Pala M, Stöckle C, Harris H (1996) Simulation of durum wheat (Triticum turgidum ssp. durum) growth under different water and nitrogen regimes in a Mediterranean environment using CropSyst. Agricultural Systems, 51, 147–163.
10.1016/0308-521X(95)00043-5
Web of Science® Google Scholar
Pannkuk C, Stöckle C, Papendick R (1998) Validation of CropSyst for winter and spring wheat under different tillage and residue management practices in a wheat-fallow region. Agricultural Systems, 57, 121–134.
10.1016/S0308-521X(97)00076-0
Web of Science® Google Scholar
Power B, Meinke H, Devoil P, Lennox S, Hayman P (2004) Effects of a changing climate on wheat cropping systems in northern New South Wales. In: New Directions for a Diverse Planet: Proceedings of the 4th International Crop Science Congress (ed R Fischer), Australian Society of Agronomy, Brisbane, Australia.
Google Scholar
Reyenga P, Howden S, Meinke H, Mckeon G (1999) Modelling global change impacts on wheat cropping in south-east Queensland, Australia. Environmental Modelling & Software, 14, 297–306.
10.1016/S1364-8152(98)00081-4
Web of Science® Google Scholar
Ritchie J, Godwin D, Otter-Nacke S (1988) CERES-wheat. A Simulation Model of Wheat Growth and Development. Texas A&M University Press, College Station, TX.
Google Scholar
Rosenzweig C, Jones JW, Hatfield JL et al. (2013) The Agricultural Model Intercomparison and Improvement Project (AgMIP): Protocols and Pilot Studies. Agricultural and Forestry Meteorology, 170, 166–182.
10.1016/j.agrformet.2012.09.011
Web of Science® Google Scholar
Sadras V (2002) Interaction between rainfall and nitrogen fertilisation of wheat in environments prone to terminal drought: economic and environmental risk analysis. Field Crops Research, 77, 201–215.
10.1016/S0378-4290(02)00083-7
Web of Science® Google Scholar
Semenov M, Wolf J, Evans L, Eckersten H, Iglesias A (1996) Comparison of wheat simulation models under climate change. II: application of climate change scenarios. Climate Research, 7, 271–281.
10.3354/cr007271
Web of Science® Google Scholar
Senthilkumar S, Basso B, Kravchenko A, Robertson G (2009) Contemporary evidence of soil carbon loss in the U.S. corn belt. Soil Science Society of America Journal, 73, 2078–2086.
10.2136/sssaj2009.0044
CAS Web of Science® Google Scholar
Stöckle C, Debaeke P (1997) Modeling crop nitrogen requirements: a critical analysis. Developments in Crop Science, 25, 217–225.
10.1016/S0378-519X(97)80022-0
Google Scholar
Stöckle C, Jara J (1998) Modeling transpiration and soil water content from a corn (Zea Maize L.) field: 20 min vs. daytime integration step. Agricultural and Forest Meteorology, 92, 119–130.
10.1016/S0168-1923(98)00086-0
Web of Science® Google Scholar
Stöckle C, Nelson R (2001) Cropping System Simulation Model User's Manual. Biological Systems Engineering Department, Washington State University, Pullman, WA. Available at: http://modeling.bsyse.wsu.edu/CS_Suite_4/CropSyst/manual/index.htm. (accessed 19 December 2014).
Google Scholar
Stöckle C, Williams J, Rosenberg N, Jones C (1992) A method for estimating the direct and climatic effects of rising atmospheric carbon dioxide on growth and yield of crops: Part I. Modification of the EPIC model for climate change analysis. Agricultural Systems, 38, 225–238.
10.1016/0308-521X(92)90067-X
Web of Science® Google Scholar
Stöckle C, Martin S, Campbell G (1994) CropSyst, a cropping systems simulation model: water/nitrogen budgets and crop yield. Agricultural Systems, 46, 335–359.
10.1016/0308-521X(94)90006-2
Web of Science® Google Scholar
Stöckle C, Cabelguenne M, Debaeke P (1997) Comparison of CropSyst performance for water management in southwestern France using submodels of different levels of complexity. Developments in Crop Science, 25, 113–122.
10.1016/S0378-519X(97)80013-X
Google Scholar
Stöckle C, Donatelli M, Nelson R (2003) CropSyst, a cropping systems simulation model. European Journal of Agronomy, 18, 289–307.
10.1016/S1161-0301(02)00109-0
Web of Science® Google Scholar
Sun J, Yang L, Wang Y, Ort D (2009) FACE-ing the global change: opportunities for improvement in photosynthetic radiation use efficiency and crop yield. Plant Science, 177, 511–522.
10.1016/j.plantsci.2009.08.003
CAS Web of Science® Google Scholar
Syswerda S, Basso B, Hamilton S, Tausig J, Robertson G (2012) Long-term nitrate loss along an agricultural intensity gradient in the upper Midwest USA. Agriculture, Ecosystems and Environment, 149, 10–19.
10.1016/j.agee.2011.12.007
CAS Web of Science® Google Scholar
Tubiello F, Rosenzweig C, Kimball B et al. (1999) Testing CERES-wheat with free-air carbon dioxide enrichment (FACE) experiment data: CO₂ and water interactions. Agronomy Journal, 91, 247–255.
10.2134/agronj1999.00021962009100020012x
Web of Science® Google Scholar
Tubiello F, Donatelli M, Rosenzweig C, Stöckle C (2000) Effects of climate change and elevated CO₂ on cropping systems: model predictions at two Italian locations. European Journal of Agronomy, 13, 179–189.
10.1016/S1161-0301(00)00073-3
Web of Science® Google Scholar
Weeks A, Christy B, Lowell K, Beverly C (2008) The Catchment Analysis Tool: demonstrating the benefits of interconnected biophysical models. In: Landscape Analysis and Visualisation (eds C Pettit, W Cartwright, I Bishop, K Lowell, D Pullar, D Duncan), pp. 49–71. Springer-Verlag, Berlin.
10.1007/978-3-540-69168-6_4
Google Scholar
Wessolek G, Asseng S (2006) Trade-off between wheat yield and drainage under current and climate change conditions in northeast Germany. European Journal of Agronomy, 24, 333–342.
10.1016/j.eja.2005.11.001
Web of Science® Google Scholar
Williams J, Jones C, Kiniry J, Spanel D (1989) The EPIC crop growth model. Transactions of the American Society of Agricultural and Biological Engineers, 32, 497–511.
10.13031/2013.31032
Web of Science® Google Scholar
Yang L, Li S, Wang Y et al. (2007a) Effects of free-air CO₂ enrichment (FACE) on yield formation of wheat]. Chinese Journal of Applied Ecology, 18, 75–80.
PubMed Web of Science® Google Scholar
Yang L, Wang Y, Li S et al. (2007b) Effects of free-air CO₂ enrichment (FACE) on dry matter production and allocation in wheat. Chinese Journal of Applied Ecology, 18, 339–346.
CAS Google Scholar
Zadoks JC, Chang TT, Konzak CF (1974) A decimal code for the growth stages of cereals. Weed Research, 14, 415–421.
10.1111/j.1365-3180.1974.tb01084.x
Web of Science® Google Scholar
Zhang Y, Feng L, Wang E, Wang J, Li B (2012) Evaluation of the APSIM-Wheat model in terms of different cultivars, management regimes and environmental conditions. Canadian Journal of Plant Science, 92, 937–949.
10.4141/cjps2011-266
Web of Science® Google Scholar

Citing Literature

Volume21, Issue7

July 2015

Pages 2670-2686

Response of wheat growth, grain yield and water use to elevated CO₂ under a Free-Air CO₂ Enrichment (FACE) experiment and modelling in a semi-arid environment

Abstract

Introduction

Materials and methods

FACE experiment

Simulation models

Simulation models – the APSIM-Wheat model

Simulation models – the APSIM-Nwheat model

Simulation models – the CAT-Wheat model

Simulation models – the CROPSYST model

Simulation models – the OLEARY-CONNOR wheat model

Simulation models – the SALUS model

Statistical analyses

Results

FACE experiment

FACE experiment – wheat response to elevated CO₂

FACE experiment – wheat response to water availability

FACE experiment – wheat response to time of sowing

FACE experiment – combined effect of elevated CO₂, water and time of sowing

Simulation performance

Simulation performance – biomass and yield

Simulation performance – water use and water use efficiency

Biomass and leaf area response to eCO₂

Yield response to eCO₂

Water use and water use efficiency response to eCO₂

Discussion

Acknowledgements

Author contribution

Supporting Information

References

Citing Literature

Figures

References

Information

About Wiley Online Library

Help & Support

Opportunities

Connect with Wiley

Response of wheat growth, grain yield and water use to elevated CO2 under a Free-Air CO2 Enrichment (FACE) experiment and modelling in a semi-arid environment

Abstract

Introduction

Materials and methods

FACE experiment

Simulation models

Simulation models – the APSIM-Wheat model

Simulation models – the APSIM-Nwheat model

Simulation models – the CAT-Wheat model

Simulation models – the CROPSYST model

Simulation models – the OLEARY-CONNOR wheat model

Simulation models – the SALUS model

Statistical analyses

Results

FACE experiment

FACE experiment – wheat response to elevated CO2

FACE experiment – wheat response to water availability

FACE experiment – wheat response to time of sowing

FACE experiment – combined effect of elevated CO2, water and time of sowing

Simulation performance

Simulation performance – biomass and yield

Simulation performance – water use and water use efficiency

Biomass and leaf area response to eCO2

Yield response to eCO2

Water use and water use efficiency response to eCO2

Discussion

Acknowledgements

Author contribution

Supporting Information

References

Citing Literature

Figures

References

Related

Information

Response of wheat growth, grain yield and water use to elevated CO₂ under a Free-Air CO₂ Enrichment (FACE) experiment and modelling in a semi-arid environment

FACE experiment – wheat response to elevated CO₂

FACE experiment – combined effect of elevated CO₂, water and time of sowing

Biomass and leaf area response to eCO₂

Yield response to eCO₂

Water use and water use efficiency response to eCO₂