Effects of elevated ozone and carbon dioxide on wheat crop yield … · 2015-02-17 · 2 Abstract Elevated levels of ground level ozone (O 3) and carbon dioxide (CO 2) may have a

Malin Broberg

Degree project for Master of Science (120 HEC)

Atmospheric Science with orientation towards Environmental Sciences 60 HEC

Department of Biological and Environmental Sciences University of Gothenburg

February 2015

Supervisor: Håkan Pleijel

Effects of elevated ozone and carbon dioxide on wheat crop yield

– Meta-analysis and exposure-response relationships

1

Effects of elevated ozone and carbon dioxide on wheat crop yield

– Meta-analysis and exposure-response relationships

…………………………………………….

Malin Broberg

Master degree thesis, 60 HEC, in Atmospheric Science with

orientation towards Environmental Sciences

Department of Biological and Environmental Sciences

University of Gothenburg

Supervisor: Håkan Pleijel

Department of Biological and Environmental Sciences

University of Gothenburg

2

Abstract

Elevated levels of ground level ozone (O3) and carbon dioxide (CO2) may have a significant

impact on crop yield production on a global scale. Since wheat is an important staple crop it is

central to evaluate possible changes in nutrient content, such as protein and nutrient minerals,

essential to the human body. Additionally, alterations of baking quality traits and other

properties affecting market-price are of great interest, such as grain mass and specific grain

mass. This project aims to examine the most important effects on quantity and quality of wheat

grains, by aggregation of experimental data available in published literature. Data analysis was

performed with meta-analysis and by deriving exposure-response functions.

Among the negative effects of ozone on wheat were the strong reductions in grain yield, grain

mass, harvest index, and total aboveground biomass, while effects on specific grain mass and

grain number were weaker but still significant. Also starch concentration and, in particular,

starch yield were significantly reduced. Moreover, a number of positive effects of ozone were

observed. There was a positive significant effect on concentrations of protein and several

nutrient minerals (K, Mg, Ca, P, P, Zn, Mn, Cu), while the areal yield of these compounds was

negatively affected. For other minerals (Fe, S, Na) results were non-significant or inconclusive.

Both concentration and yield of the potentially toxic element Cd was significantly reduced,

which also was true for experiments with elevated carbon dioxide. Additionally, all baking

quality traits included (Hagberg falling number, Zeleny value, wet and dry gluten content) were

significantly improved under ozone exposure.

The main effects of carbon dioxide were to significantly increase grain yield, grain number and

total aboveground biomass, while effects on harvest index, grain mass and specific grain mass

were very small and non-significant. Starch concentration was not significantly affected,

whereas starch yield strongly increased following the pattern of grain yield. Protein

concentration was significantly reduced while yield was increasing. Concentrations of a number

of nutrient minerals (S, Fe, P, Mg, Ca, Mn, Zn) significantly decreased, however for Na and K

effects were non-significant. Areal yield of a few elements significantly decreased (Fe, Mg, Mn)

while the effect was opposite for Zn. Effects on remaining mineral yields (S, P, Na, Ca, K) were

non-significant. Opposed to ozone, baking qualities (bread loaf volume, resistance breakdown,

peak resistance, dry and wet gluten, Zeleny value, and Hagberg falling number) were mainly

negative affected by elevated carbon dioxide, except for mixing time where the effect was

significantly positive.

From this study it can be concluded that elevated levels of both ozone and carbon dioxide have

significant effects on wheat grain yield, although with different response patterns. Not only is the

quantity of yield affected but also many qualitative variables. This point towards the importance

of incorporating quality and nutritional aspects into further assessment of ozone and carbon

dioxide impacts on global agriculture and food security.

3

Table of contents

1. INTRODUCTION ..................................................................................................................... 5

1.1 BACKGROUND ......................................................................................................................... 5

1.1.1 GLOBAL TRENDS IN CARBON DIOXIDE AND OZONE .................................................................................. 5

1.1.2 PLANT RESPONSES TO ELEVATED O3 AND CO2 ....................................................................................... 6

1.1.3 WHEAT QUALITY TRAITS .................................................................................................................... 7

1.2 AIM ...................................................................................................................................... 9

1.2.1 RESEARCH QUESTIONS ...................................................................................................................... 9

2. MATERIALS AND METHODS .................................................................................................. 10

2.1 DATABASE ............................................................................................................................ 10

2.2 DATA ANALYSIS ..................................................................................................................... 11

2.2.1 META-ANALYSIS ............................................................................................................................ 11

2.2.2 RESPONSE FUNCTIONS .................................................................................................................... 15

3. RESULTS .............................................................................................................................. 19

3.1 EFFECTS OF OZONE ................................................................................................................. 19

3.1.1 YIELD COMPONENTS ....................................................................................................................... 19

3.1.2 PROTEIN AND STARCH .................................................................................................................... 22

3.1.3 MINERALS .................................................................................................................................... 24

3.1.4 BAKING QUALITIES ......................................................................................................................... 26

3.2 EFFECTS OF CARBON DIOXIDE .................................................................................................... 28

3.2.1 YIELD COMPONENTS ....................................................................................................................... 28

3.2.2 PROTEIN AND STARCH ..................................................................................................................... 32

3.2.3 MINERALS .................................................................................................................................... 34

3.2.4 BAKING QUALITY ............................................................................................................................ 35

3.3 INTERACTIVE EFFECTS OF CARBON DIOXIDE AND OZONE ................................................................... 37

4. DISCUSSION ......................................................................................................................... 38

4.1 OZONE EFFECTS ..................................................................................................................... 38

4.2 CARBON DIOXIDE EFFECTS ........................................................................................................ 40

4.3 COMPARISON OF OZONE AND CARBON DIOXIDE EFFECTS .................................................................. 41

4.4 FOOD SECURITY AND CROP MODELING ......................................................................................... 42

4.5 METHODS – LIMITATIONS AND POTENTIAL DEVELOPMENTS .............................................................. 42

5. CONCLUSIONS ..................................................................................................................... 44

ACKNOWLEDGEMENTS ............................................................................................................... 45

REFERENCES ............................................................................................................................... 46

REFERENCES USED FOR THE GENERAL TEXT ............................................................................................ 46

4

REFERENCES USED IN DATA ANALYSIS ................................................................................................... 49

OZONE DATABASE ...................................................................................................................................... 49

CARBON DIOXIDE DATABASE ........................................................................................................................ 51

DATABASE FOR EXPERIMENTS WITH OZONE AND CARBON DIOXIDE INTERACTION ................................................... 53

APPENDIX 1. SUBGROUP ANALYSIS OF OZONE EFFECTS ............................................................... 54

APPENDIX 2. SUBGROUP ANALYSIS OF CARBON DIOXIDE ............................................................ 57

APPENDIX 3. INTERACTIVE EFFECTS OF OZONE AND CARBON DIOXIDE ........................................ 59

5

1. Introduction

1.1 Background

1.1.1 Global trends in carbon dioxide and ozone

Carbon dioxide (CO2) and ozone (O3) are both important greenhouse gases and concentrations of

these compounds have steadily increased since the industrial revolution. Due to anthropogenic

emissions atmospheric CO2 concentration has risen from the pre-industrial level of 280 ppm to

current 400 ppm, and where future scenarios predict concentrations reaching 421 ppm

(RCP2.6) to 1313 ppm (RCP8.5) in year 2100 (IPCC, 2013). CO2 is a well-mixed and long-lived

greenhouse gas, with an atmospheric lifetime up to 100 years.

Tropospheric O3 is semi-global air pollutant produced through photochemical reactions of

nitrogen oxides (NOx) and volatile organic compounds (VOC) (including methane and carbon

monoxide). These O3 precursors are generally emitted from anthropogenic sources, and

background average concentrations of O3 have risen from pre-industrial levels of about 10 ppb

to current 20-35 ppb (Vingarzan, 2004). However, there are both spatial and temporal

variations in O3 concentrations, and higher concentrations of O3 might occur for shorter time-

periods, ozone episodes. Figure 1 gives an example of the average diurnal variation over one

year and the growing season (May-August). O3 concentration is strongly linked to the presence

of its precursors and the meteorological conditions; hence it mainly peaks during summer when

there are high levels of sunlight and high-pressure weather conditions prevail (Pleijel, 2007).

Future O3 concentrations will mainly be determined by the emission of its precursors. It is also

suggested that global warming may act to decrease background concentration, while high

methane (CH4) levels (RCP8.5) can offset this decrease and by year 2100 raising concentration

by 8 ppb (25 % of current levels) (IPCC, 2013). On the other hand, higher surface temperatures

in polluted areas could trigger regional feedbacks in chemistry and local emissions that will

cause an increase in peak levels of O3.

A number of indices are used to estimate ozone exposure of plants, such as AOT40, POD6 and

ozone daytime concentration. AOT40 is the accumulated exposure over a threshold of 40 ppb

during daylight hours for a specific time-interval (e.g. growing season) (Fuhrer et al., 1997).

POD6 (phytotoxic O3 dose above a threshold of 6 nmol s-1 m-2 projected leaf area) is a flux based

exposure index and is estimated by a combination of model predicted stomatal conductance and

environmental factors, such as soil moisture, air temperature, air humidity, and solar radiation

(Mills et al., 2011). Daytime concentration [O3]day is basically the average O3 concentration for a

specific time-interval, i.e. 7-h seasonal average for 1000-1700.

6

Figure 1. Average O3 concentration for 40 European monitoring stations in 2007, data from The European Monitoring and Evaluation Programme (EMEP, www.emep.int).

1.1.2 Plant responses to elevated O3 and CO2

Key plant processes and its responses to environmental change are essential when it comes to

the understanding of how wheat yield will be affected by elevated levels of CO2 and O3. Changes

in concentrations of these gases may affect many plant processes, such as photosynthesis,

transpiration, nutrient uptake, and senescence, resulting in significant effects on crop growth

and yield.

Photosynthesis in C3 plants, such as wheat and rice, is directly stimulated by elevated CO2

whereas the conductance of CO2 and water vapor is reduced. Net photosynthesis immediately

increases with a rise in CO2 because ribulose-1,5-biphosphate carboxylase-oxygenase (Rubisco)

is not saturated at current atmospheric levels. Besides, photorespiration is reduced since rising

CO2 competitively inhibits the oxygenase reaction, leading to a higher net gain of carbon in the

plant (Ainsworth and McGrath, 2010). However, when plants are grown under elevated CO2 for

longer durations effects may be different. For example, Rubisco activity could decrease, which is

a mechanism that is believed to occur to optimize utilization of nitrogen (Ainsworth and Rogers,

2007). Despite the reduction in photosynthetic capacity, elevated CO2 is expected to result in a

net gain of carbon for C3 plants, which has been estimated to be about 13 % for wheat (Long et

al., 2006).

In addition to changes in photosynthetic rate, elevated levels of CO2 also cause a decreased

stomatal conductance of water vapor and CO2. This effect is observed for both C3 and C4 species,

but more strongly in C3 plants. Decrease in canopy transpiration generally leads to higher soil

moisture content and improved water use efficiency, which could be beneficial during drought

periods (Ainsworth and McGrath, 2010).

Ozone exposure causes oxidative stress to the plant, which damages plant tissues and affects

photosynthetic rate. Loss of photosynthetic capacity is an early symptom of O3 stress and may

cause damage or inhibit almost every step of the photosynthesis, from light capture to starch

accumulation (Farage et al., 1991). Ahead of all other effects on the photosynthetic apparatus, O3

mainly acts to reduce Rubisco activity, which is being caused by damage of existing Rubisco

7

rather than the Rubisco synthesis (McKee et al., 1995). High levels of O3 could damage stomatal

functioning, while lower exposure only results in an indirect effect where transpiration is

reduced only due to the decrease in photosynthesis (Long, 1985). In addition, other mechanisms

linked to transpiration will be affected such as uptake of minerals from the soil. Elevated levels

of ozone also promote accelerated senescence (Grandjean and Fuhrer, 1989) that results in a

shortened period for both growth and grain filling (Gelang et al., 2001), but also tends to favor

protein over starch deposition in cereal grains (Wang and Frei, 2011).

Most experimental setups separate the effects of CO2 and O3 but in reality they are both likely to

occur at elevated levels over wide areas. Looking at previously mentioned effects of these

compounds it can be hypothesized that elevated CO2 could act to inhibit O3 uptake due to a

decrease in stomatal conductance and stimulation of photosynthesis. However, it remains

unclear whether the negative effects of O3 on photosynthesis and plant growth are of same

magnitude as the positive effects of CO2.

1.1.3 Wheat quality traits

Elevated levels of CO2 and O3 may significantly affect both quantity and quality of wheat grain

yield. There are a number of properties describing the quality of wheat grains; some important

for the market price, such as protein content, grain mass, specific grain mass, and baking

properties, while other are external to the market but essential for human nutrition, like

concentrations of nutrient minerals and potentially toxic compounds.

Harvest index (HI) is the fraction of total aboveground biomass (TAB) that is represented by the

grain yield (GY). Grain number (GN) is the number of grains per plant or area unit and grain

mass (GM) (often given as the “1000-grain weight” in agronomic literature) is the average mass

of one grain. The relationship between these variables is following:

𝐺𝑌 = 𝑇𝐴𝐵 ∗ 𝐻𝐼 = 𝐺𝑁 ∗ 𝐺𝑀 (1)

Grain mass and specific grain mass (technically the density, often referred to as “volume-

weight”, “hectoliter weight” or “test weight”) are quality traits linked to the physical

characteristics of the grain and central to the market price. Higher values of these variables are

generally associated with larger flour yield, while low values could be an indication of a large

fraction of small and malformed grains (Manley et al., 2009; Weiss and Moreno-Sotomayer,

2006).

The next group of yield quality properties is the content of nutrient elements, including

concentrations of grain protein, macronutrients (Mg, K, Ca, P, S) and micronutrients (Zn, Fe, Mn,

Cu). In addition to these compounds it is also of interest to examine how minerals non-essential

to plants respond to elevated levels of CO2 and O3, therefore Na and Cd are included in the preset

study. Besides, Cd is a potentially toxic and bio-accumulating compound, which may have

serious health effects (EFSA, 2009; Saturug et al., 2010). Moreover, the effects on unit area yield

of these constituents are relevant for human nutrition, but also for the biogeochemical balance

of input/output in the agro-ecosystem, where a reduced uptake from plants potentially could be

mitigated with an increased input of minerals. In areas of food scarcity a reduction in areal yield

of nutrients might have serious effects on humans and cause malnutrition. Consequently effects

on both concentration and areal yield of protein and minerals were examined in this study.

8

Starch is the main source of energy in wheat grains and the content is often in the range 60-70%

of grain mass (Shewry, 2009). Since wheat is a major food crop and consequently an important

source of energy, the concentration and areal yield of starch was also included in the study.

Since wheat to a large extent is used for baking, it is relevant to examine how properties related

to baking quality are affected by CO2 and O3. Some frequently used indicators are Hagberg falling

number, Zeleny value, gluten content, mixing time, break loaf volume, resistance breakdown,

and peak resistance. Hagberg falling number is a measure of the α-amylase activity in the grain,

and its resistance to enzymatic degradation. Low falling number is associated with high α-

amylase activity and consequently more germinated grains, which gives poor baking quality,

such as poorly structured loaves and sticky dough (Kindred et al., 2005). A high Hagberg falling

number results in better baking quality, but also enables longer storage time for grains and flour

(Hruskova et al., 2004). The Zeleny value gives an indication of the protein quality and is

determined by a sedimentation test, where the sedimentation rate is measured of wheat

suspended in an acid solution. Due to swelling of glutenin, a major protein in wheat,

sedimentation occurs and a high sedimentation volume (Zeleny value) is related to a high

content of glutenin and thus good baking quality (Eckert et al., 1993; Reeves et al., 1978).

Gluten content is important for elasticity of the dough, where higher gluten concentration gives

more resistance, but also larger bread loaf volume. Here peak resistance is used as an indicator

and measured as the time (seconds) to reach peak dough resistance, while resistance

breakdown is the percentage drop in resistance after 3 minutes after peak (Rogers et al., 1998).

Bread loaf volume is also related to mixing time, where longer mixing time generally gives better

loaf volume (Finney and Shogren, 1972; Kimball et al. 2001), hence a higher baking quality.

In summary, most baking properties are related to either quantity or quality of grain protein

where higher grain protein concentration generally is results in better baking quality.

9

1.2 Aim

This master thesis aims to examine the most important effects of CO2 and O3 on the quantity and

quality of wheat grains, by aggregation of experimental data available in published literature.

Analysis will be limited to responses of quantitative and qualitative yield variables;

photosynthesis, growth rate, visual damage etc. will consequently not be included in this study.

Additional inclusion criteria for database are defined in materials and methods section.

1.2.1 Research questions

What are the main effects of elevated levels of CO2 on wheat grain quantity and quality?

What are the main effects of long term O3 exposure on wheat grain quantity and quality?

Are there any effects of current (ambient) O3 exposure? Comparison of charcoal filtered

and non-filtered treatments.

What are the yield responses for experiments with interaction of elevated O3 and CO2?

Are yield variable responses consistent or do they vary geographically? Comparison of

variables with abundant data with large geographical distribution.

Are there any differences between experimental setups? Comparison of fumigation

facilities and rooting environments.

10

2. Materials and methods

2.1 Database

Web of Science was used to obtain relevant paper for meta-analysis and response functions.

Following inclusion criteria had to be fulfilled:

At least one of the following yield variables was reported: grain yield, harvest index, total

aboveground biomass, grain number, specific grain mass, grain protein concentration,

grain starch concentration, grain mineral concentrations (P, K, Ca, Zn, Fe, Mn, Mg, Cu, S,

Na, Cd, Cu, Na), baking properties (Hagberg falling number, Zeleny value, gluten content,

mixing time, peak resistance, bread loaf volume, resistance breakdown). Parameter

values for yield variables were considered independent if they were made on different

cultivars or CO2/O3 concentrations, following the approach of earlier meta-analysis

(Curtis and Wang, 1998; Feng et al., 2008).

For O3 experiments additional requirements/constraints were:

Daytime ozone concentration was reported, [O3]day (7-h, 8-h, 12-h, 6-h or 4-h seasonal

average).

Ozone exposure was at least 14 days.

[O3]day for elevated ozone treatments was at least 30 ppb during exposure.

Plants were rooted in field soil, i.e. pot experiments were excluded. Since only a few

experiments were performed in pots, removing these did not have a significant effect on

overall analysis, while they were too few to use for subgroup analysis of rooting

environment.

For CO2 experiments there are a wide range of experimental setups, with plants rooted in both

field soil and pots and several types of fumigation facilities are being used (greenhouse (GH),

growth chamber (GC), temperature gradient tunnel (TGT), open-top chambers (OTC), free-air

concentration-enrichment (FACE)). For analysis of CO2 data all types of experiments were

included since excluding e.g. pot experiments would result in a significantly smaller database,

not being sufficient for further analysis. Including data from different rooting environments and

fumigation facilities also allowed for subgroup analysis, separating and comparing the different

experimental setups. However, the study by Wu et al. (2004) was excluded since some variable

responses were out of range and considered as outliers, which was detected during data

exploration. Also the paper by Mishra et al. (2013) was taken out from the analysis, since units

for several variables were missing.

Data from figures were extracted using software (Get Data Graph Digitizer 2.26; http://getdata-

graph-digitizer.com/). Additional data, not found in the open literature, were mineral

concentrations for the Gelang et al. (2000) experiment, starch for the Pleijel et al (1998)

experiment, and minerals and protein for the Zhu et al. (2011) experiment. Data for grain yield,

protein and minerals for the experiment described in Högy et al. 2013 (including experiments

from 2004-2006) was also added to the database.

http://getdata-graph-digitizer.com/)

http://getdata-graph-digitizer.com/)

11

2.2 Data analysis

2.2.1 Meta-analysis

Meta-analyses were conducted separately for CO2 and O3 experiments, and for a few variables

interactive effects of CO2 and O3 were examined which was possible only for three experiments.

All analyses were performed using a meta-analytical software package (MetaWin2.1.3.4, Sinauer

Associates, Inc. Sunderland, Ma USA) (Rosenberg et al., 2000). The natural log of the response

ratio (r, the ratio of the means of two groups, experimental and control) was used as effect size

for the analyses and reported as percentage change from the control ((𝑟 − 1) ∗ 100%) (Feng et

al., 2008; Koricheva et al., 2013; Rosenberg et al., 2000). Due to lack of data for computation of

sample variance (standard deviation or standard error with replication) all variables were

analyzed using an un-weighted approach. The variance of the effect size (ln r) was calculated

using a resampling method with 59,000 iterations, in line with previous meta-analysis (Feng et

al., 2008; Rosenberg et al., 2000), and confidence intervals (CI) were calculated using bootstrap

method (Rosenberg et al., 2000). Average effect size was considered to be significant if the 95%

CI did not overlap zero, and for subgroup analysis the different group where assumed to be

significantly different if the 95% CI did not overlap (Curtis and Wang, 1998).

For meta-analysis of experiments with CO2 treatments ambient concentration of CO2 was chosen

as control. Subgroup analyses were performed in order to explain variation in responses of the

different yield variables. Following categories were included: (1) continent, (2) wheat type

(spring vs. winter wheat), (3) rooting environment (pots, vs. field), (4) fumigation facility

(growth chamber, greenhouse, FACE, OTC, temperature gradient tunnel).

To examine the yield variables in O3 experiments two parallel analyses were performed, one

with charcoal filtered air (CF) as control and another with non-filtered air (NF) as control. It

should be noted that experiments conducted in FACE facilities never include a reduced O3

treatment (CF), consequently FACE experiments are only included the meta-analysis with NF as

reference. As for the meta-analysis of CO2 data, a subgroup analysis was also conducted for some

variables in the O3 dataset with the following categories: (1) wheat type (spring vs. winter

wheat), (2) fumigation facility (OTC vs. FACE).

Due to a limited amount of data, meta-analysis with interaction experiments (CO2*O3) were

conducted for only 3 experiments, using treatment with ambient CO2 and CF/NF as the

reference. In this case CF and NF treatments were considered equivalent since [O3]day for the

experiments included were in the same range (18-35 ppb). Yield data were limited to grain yield,

harvest index, grain mass and grain protein concentration. To compare the effects of elevated

CO2 and/or O3, subgroup analysis was performed with following groups: elevated CO2, elevated

O3, and elevated CO2+O3.

Overview table for the data included in meta-analyses for CO2 and O3 could be found in Table 1, 2

and 3, showing the number of comparisons, individual experiments, countries, continents, and

cultivars.

12

Table 1. Overview table for data included in meta-analysis of CO2 experiments, with number of comparisons, experiments, countries, and cultivars included in the analysis for each variable. Meta-analysis output is presented by the average change in percentage and the 95 % confidence interval (ns= not significant, + = positive significant effect, - = negative significant effect).

Variable Comparisons Experiments Countries Continents Cultivars Average change (%) 95% CI Direction of change

Grain yield 65 54 11 4 21 23.55 19.23 28.52 + Harvest index 47 32 7 2 14 0.20 -1.89 2.50 ns Grain number 23 16 7 4 11 7.08 2.52 12.45 + Grain mass 56 38 10 4 17 -0.39 -2.19 1.47 ns Specific grain mass 10 7 3 2 4 -1.73 -8.18 2.85 ns Starch concentration 12 9 4 2 6 -0.45 -1.76 0.74 ns yield 12 9 4 2 6 16.15 10.02 21.87 + Protein concentration 42 34 10 3 17 -11.94 -14.42 -9.60 - yield 37 30 9 3 16 9.33 4.69 14.48 + Zn concentration 16 14 3 2 7 -10.17 -14.85 -5.56 - yield 16 14 3 2 7 7.39 2.19 12.99 + Mn concentration 13 11 2 1 5 -5.84 -10.65 -1.70 - yield 13 11 2 1 5 -11.08 -19.81 -1.13 - Cd concentration 7 6 2 1 2 0.89 0.83 0.94 + yield 7 6 2 1 2 0.95 0.89 1.00 + K concentration 6 5 2 1 4 -4.18 -11.16 2.02 ns yield 6 5 2 1 4 -6.74 -23.49 20.06 ns Ca concentration 10 9 2 2 5 -14.76 -21.18 -8.11 - yield 10 9 2 2 5 -14.38 -29.64 4.58 ns Mg concentration 8 7 1 1 4 -6.02 -9.52 -2.65 - yield 8 7 1 1 4 -17.12 -26.27 -6.78 - Na concentration 3 3 1 1 2 3.34 0.00 10.38 + yield 3 3 1 1 2 0.44 -11.31 11.10 ns P concentration 6 5 2 1 4 -5.26 -10.13 -0.15 - yield 6 5 2 1 4 -9.24 -27.25 23.23 ns Fe concentration 12 11 2 2 6 -15.73 -21.62 -10.12 - yield 12 11 2 2 6 -17.35 -30.89 -0.70 - Cu concentration 3 3 1 1 2 -2.16 -10.07 5.76 - yield 3 3 1 1 2 61.09 -25.84 487.56 ns S concentration 10 9 2 2 5 0.84 0.74 0.94 + yield 10 9 2 2 5 0.88 0.73 1.07 + Hagberg 6 3 1 1 1 -5.80 -9.87 -1.71 - Zeleny 6 3 1 1 1 -21.24 -25.52 -16.87 - Dry gluten 6 3 1 1 1 -16.48 -22.04 -11.24 - Wet gluten 6 3 1 1 1 -17.03 -22.95 -11.54 - Mixing time 6 6 2 2 3 11.23 0.64 21.49 + Peak resistance 5 5 1 1 2 -11.42 -17.28 -2.96 - Resistance breakdown 5 5 1 1 2 -2.60 -12.51 9.07 ns Bread loaf volume 3 2 2 2 3 -11.86 -21.28 -2.28 -

13

Table 2. Overview table for data included in meta-analysis of O3 experiments using CF treatment as the reference, with number of comparisons, experiments, countries, and cultivars included in the analysis for each variable. Meta-analysis output is presented by the average change in percentage and the 95 % confidence interval (ns= not significant, + = positive significant effect, - = negative significant effect).


Grain yield 91 33 8 3 19 -25.04 -29.75 -20.37 - Harvest index 57 27 8 3 14 -9.83 -13.22 -6.70 - Grain number 66 23 6 3 13 -9.58 -12.58 -6.70 - Grain mass 85 31 8 3 17 -16.71 -20.44 -13.03 - Specific grain mass 10 7 3 2 4 -0.42 -1.56 0.45 ns Starch concentration 21 9 3 1 3 -5.03 -7.62 -2.62 - Protein concentration 47 21 6 3 10 6.2 4.1 8.58 + yield 47 21 6 3 10 -17.07 -22.7 -11.69 - Zn concentration 12 5 2 2 3 22.24 9.81 40.03 + yield 12 5 2 2 3 -11.36 -20.29 -2.08 - Mn concentration 13 6 2 2 4 15.47 2.98 32.05 + yield 13 6 2 2 4 -13.53 -24.15 -2.33 - Cd concentration 9 4 1 1 3 -6.68 -12.57 -2.36 - yield 9 4 1 1 3 -25.93 -38.5 -14.06 - K concentration 27 11 4 2 5 9.7 3.48 17.23 + yield 27 11 4 2 5 -22.3 -29.58 -15.07 - Ca concentration 21 9 4 2 5 12.23 3.34 23.48 + yield 21 9 4 2 5 -23.63 -32.17 -14.92 - Mg concentration 21 9 4 2 5 18.61 8.12 32.44 + yield 21 9 4 2 5 -19.28 -29.13 -9.1 - Na concentration 11 4 3 2 3 -18.97 -41.41 11.94 ns yield 11 4 3 2 3 -40.28 -56.3 -20.19 - P concentration 21 9 4 2 5 15.5 6.81 26.92 + yield 21 9 4 2 5 -21.4 -30.49 -11.98 - Fe concentration 10 3 2 2 2 19.87 10.83 31.07 + yield 10 3 2 2 2 -14.05 -27.13 1.3 ns Cu concentration 10 3 2 2 2 25.98 8.28 48.10 + yield 10 3 2 2 2 -10.6 -26.51 9.28 ns S concentration 11 4 3 2 3 6.67 -1.58 16.35 ns yield 11 4 3 2 3 -21.38 -34.44 -7.2 - Hagberg 9 6 2 1 3 15.61 7.62 28.26 + Zeleny 10 6 2 1 3 10.89 6.76 15.1 + Dry gluten 4 4 1 1 2 8.96 4.65 13.78 + Wet gluten 4 4 1 1 2 8.48 4.43 13.35 +

14

Table 3. Overview table for data included in meta-analysis of O3 experiments using NF treatment as the reference, with number of comparisons, experiments, countries, and cultivars included in the analysis for each variable. Meta-analysis output is presented by the average change in percentage and the 95 % confidence interval (ns= not significant, + = positive significant effect, - = negative significant effect).


Grain yield 75 44 8 3 21 -24.89 -29.29 -20.54 - Harvest index 40 25 10 3 14 -11.35 -15.57 -7.49 - Grain number 59 36 8 3 18 -8.60 -11.49 -5.95 - Grain mass 67 42 10 3 21 -20.02 -23.23 -16.81 - Specific grain mass 6 3 1 1 2 -3.34 -6.12 -0.58 - Starch concentration 32 25 4 2 8 -4.24 -6.57 -2.21 - yield 32 25 4 2 8 -29.31 -36.22 -22.77 - Protein concentration 50 34 6 3 12 6.54 3.79 9.29 + yield 50 34 6 3 12 -17.88 -23.23 -12.67 - Zn concentration 20 13 2 2 6 7.41 -0.44 17.78 ns yield 20 13 2 2 6 -17.3 -23.53 -10.67 - Mn concentration 20 13 2 2 6 7.04 -0.82 17.13 ns yield 20 13 2 2 6 -17.58 -25.07 -9.59 - Cd concentration 10 4 1 1 2 11.30 0.76 24.94 + yield 10 4 1 1 2 -26.75 -37.46 -17.38 - K concentration 24 16 3 2 7 11.52 5.75 18.32 + yield 24 16 3 2 7 -23.9 -30.48 -17.42 - Ca concentration 12 6 3 2 3 22.98 11.7 37.22 + yield 12 6 3 2 3 -30.24 -39.44 -19.88 - Mg concentration 20 14 3 2 7 11.49 1.71 24.92 + yield 20 14 3 2 7 -26.41 -33.41 -19.14 - Na concentration 7 3 2 2 2 15.26 -6.74 42.15 ns yield 7 3 2 2 2 -28.7 -40.31 -12.72 - P concentration 12 6 3 2 3 20.85 7.39 38.9 + yield 12 6 3 2 3 -31.45 -40.77 -20.8 - Fe concentration 7 3 2 2 2 9.79 2.51 17.79 + yield 7 3 2 2 2 45.38 24.69 209.04 + Cu concentration 15 11 2 2 6 11.30 0.76 24.94 + yield 15 11 2 2 6 -19.53 -28.52 -8.77 - S concentration 7 3 2 2 2 5.35 -2 15.66 ns yield 7 3 2 2 2 -34.83 -46.09 -22.53 -

15

2.2.2 Response functions

Exposure-response relationships were derived through regression between relative effects of

yield variables and CO2/O3 concentration. For CO2 data the response were related to the effect

estimated at 350 ppm for each experiment, where the yield variable were set to the value 1 on a

relative scale with the assumption that there was no effect at 350 ppm. The same approach was

used to derive response functions for O3, but with the effects related to the effect at zero [O3]day.

For some of the CO2 experiments CO2 concentration was not reported for ambient treatments. To

be able to include this data in the response functions CO2 concentration was assumed to be equal

to the global mean for the year that the study was conducted. The Mauna Loa record was used as

reference for the global mean of CO2 concentration and retrieved from National Oceanic &

Atmospheric Administration (NOAA) (www.noaa.gov).

As already mentioned (section 1.1.1.) there are several indices used for ozone exposure. Even

though POD6 and AOT40 may give stronger relationships than [O3]day, they are only available for

a very limited number of experiments. Since the aim of the present investigation was to include

as many experimental observations as possible on a global scale, [O3]day was chosen for the

present study since it is reported in most of the studies.

Additional inclusion criteria were defined for ozone response functions:

1. Experimental treatments were excluded if [O3]day was larger than 100 ppb, to avoid

influence on regressions from a few very high ozone treatments.

2. Experiments were excluded if they contained only two treatments, where NF was the

control treatment and larger than 30 ppb, in order to avoid uncertainty in determination

of the intercept associated with large extrapolation in combination with only two

treatments.

3. Data were also excluded if the [O3]day range (difference between highest and lowest

treatment) was smaller than 15 ppb, since random effects become large in relation to the

difference in exposure.

Due criterion number 2, FACE experiments were excluded from response regressions for ozone.

The seasonal average of [O3]day was reported for different time-intervals (the number of hours

included in the average). In order to correct for this, a daily average was calculated for 40

European monitoring stations for the time period May-July for each time-interval (4-h, 6-h, 7-h,

8-h, 12-h). O3 data were retrieved from The European Monitoring and Evaluation Programme

(EMEP, www.emep.int ). Since [O3]7-h was the measurement interval reported most frequently it

was chosen as standard. [O3]12-h average for EMEP measurements was estimated to be 4.1%

lower than the 7-h standard, hence all [O3]12-h data was adjusted with a conversion factor of

1.041. The remaining time-intervals (4-h, 6-h, and 8-h) were estimated to differ less than 1%

and were consequently not corrected.

http://www.noaa.gov/

http://www.emep.int/

16

Table 4. Overview table for response functions of O3 experiments with number of treatments, experiments, countries, continents, and cultivars. Linear regressions for all yield variables are presented by coefficient of determination (R2), statistical significance (P), slope (% effect per 10 ppb), and intercept.

Variable Treatments Experiment Countries Continents Cultivars R2 P % effect per 10 ppb Intercept

Grain yield 121 37 10 3 20 0.525 <0.001 *** -4.68 1.00 Harvest index 92 30 9 3 15 0.458 <0.001 *** -2.53 1.01 Total aboveground biomass 92 30 9 3 15 0.493 <0.001 *** -3.77 1.03 Grain mass 113 35 9 3 17 0.538 <0.001 *** -3.84 1.01 Grain number 83 23 6 3 13 0.073 <0.05 * -1.02 0.97 Specific mass 23 9 3 2 5 0.354 <0.01 ** -1.13 1.01 Protein concentration 69 22 6 3 10 0.399 <0.001 *** 2.55 1.00 yield 69 22 6 3 10 0.503 <0.001 *** -4.10 1.02 Starch concentration 33 9 3 1 4 0.405 <0.001 *** -1.53 1.02 yield 33 9 3 1 4 0.810 <0.001 *** -7.37 1.04 Zn concentration 17 5 2 2 3 0.803 <0.001 *** 3.28 1.01 yield 17 5 2 2 3 0.203 0.070 ns -2.37 0.97 Mn concentration 19 6 3 2 4 0.204 0.052 ns 7.99 0.93 yield 19 6 3 2 4 0.000 0.985 ns 0.05 0.94 Cd concentration 17 5 1 1 3 0.107 0.200 ns -1.49 1.00 yield 17 5 1 1 3 0.487 <0.01 ** -5.49 0.97 K concentration 33 10 4 2 5 0.440 <0.001 *** 3.55 0.94 yield 33 10 4 2 5 0.647 <0.001 *** -4.92 0.98 Ca concentration 25 8 4 2 5 0.497 <0.001 *** 4.37 0.92 yield 25 8 4 2 5 0.680 <0.001 *** -5.18 0.98 Mg concentration 25 8 4 2 5 0.633 <0.001 *** 5.25 0.93 yield 25 8 4 2 5 0.574 <0.001 *** -4.89 0.99 Na concentration 11 3 3 2 3 0.005 0.837 ns -0.87 0.86 yield 11 3 3 2 3 0.199 0.170 ns -5.45 0.89 P concentration 25 8 4 2 5 0.620 <0.001 *** 4.36 0.94 yield 25 8 4 2 5 0.619 <0.001 *** -5.11 0.99 Fe concentration 9 2 2 2 2 0.565 <0.05 * 2.88 0.97 yield 9 2 2 2 2 0.230 0.191 ns -3.31 0.91 Cu concentration 9 2 2 2 2 0.299 0.128 ns 4.85 0.92 yield 9 2 2 2 2 0.042 0.599 ns -1.81 0.90 S concentration 11 3 3 2 3 0.179 0.195 ns 1.27 0.97 yield 11 3 3 2 3 0.397 <0.05 * -4.33 0.96 Hagberg falling number 14 6 2 1 3 0.309 <0.05 * 2.88 1.03 Zeleny value 16 6 2 1 3 0.717 <0.001 *** 3.80 1.00 Dry gluten 8 4 1 1 2 0.465 0.062 ns 2.99 1.00 Wet gluten 8 4 1 1 2 0.439 0.073 ns 2.78 1.01

Table 5. Overview table for response functions of CO2 experiments with number of treatments, experiments, countries, continents, and cultivars. Linear regressions for all yield variables are presented by coefficient of determination (R2), statistical significance (P), slope (% effect per 100 ppm), and intercept.

Variable Treatments Experiments Countries Continents Cultivars R2 P % Effect per 100 ppm CO2 Intercept

Grain yield 124 53 11 4 23 0.197 <0.001 *** 7.62 0.77 Harvest index 86 34 9 2 17 0.003 0.631 ns -0.27 1.02 Total aboveground biomass 86 34 9 2 17 0.490 <0.001 *** 9.07 0.69 Grain mass 100 41 11 4 20 0.006 0.455 ns 0.27 0.99 Grain number 94 38 11 4 20 0.176 <0.001 *** 5.77 0.82 Specific grain mass 17 7 3 2 4 0.048 0.400 ns -1.15 1.04 Protein concentration 63 28 7 3 15 0.159 <0.01 ** -1.91 1.04 yield 55 24 6 3 14 0.218 <0.001 *** 4.19 0.86 Starch concentration 23 9 4 2 6 0.037 0.376 ns 0.26 0.98 yield 23 9 4 2 6 0.549 <0.001 *** 5.37 0.83 Zn concentration 30 14 3 2 7 0.384 <0.001 *** -4.49 1.17 yield 30 14 3 2 7 0.043 0.271 ns 1.84 0.98 Mn concentration 22 10 2 1 6 0.076 0.213 ns -1.74 1.04 yield 22 10 2 1 6 0.004 <0.01 ** 4.58 0.83 Cd concentration 13 6 2 1 2 0.331 <0.01 ** -4.12 1.13 yield 13 6 2 1 2 0.000 0.980 ns 0.07 0.96 K concentration 11 5 2 1 4 0.051 0.506 ns -0.92 1.03 yield 11 5 2 1 4 0.820 <0.001 *** 10.79 0.60 Ca concentration 19 9 2 2 5 0.568 <0.001 *** -6.32 1.23 yield 19 9 2 2 5 0.005 0.775 ns 1.00 1.03 Mg concentration 15 7 1 1 4 0.498 <0.01 ** -2.72 1.10 yield 15 7 1 1 4 0.460 <0.01 ** 4.48 0.86 Na concentration 6 3 1 1 2 0.183 0.397 ns 2.96 0.89 yield 6 3 1 1 2 0.814 <0.05 * 12.26 0.57 P concentration 11 5 2 1 4 0.399 <0.05 * -2.03 1.08 yield 11 5 2 1 4 0.488 <0.05 * 9.77 0.64 Fe concentration 23 11 2 2 6 0.706 <0.001 *** -7.61 1.28 yield 23 11 2 2 6 0.000 0.931 ns -0.28 1.06 Cu concentration 6 3 1 1 2 0.029 0.748 ns -1.11 1.04 yield 6 3 1 1 2 0.451 0.144 ns 7.26 0.75 S concentration 19 9 2 2 5 0.602 <0.001 *** -5.59 1.21 yield 19 9 2 2 5 0.017 0.593 ns 2.08 1.00 Cr concentration 8 4 1 1 1 0.002 0.911 ns 0.40 0.99 yield 8 4 1 1 1 0.472 0.060 ns 8.00 0.72 Hagberg 9 3 1 1 1 0.046 0.581 ns -0.77 1.03 Zeleny 9 3 1 1 1 0.531 <0.05 * -5.69 1.20 Dry gluten 9 3 1 1 1 0.400 0.067 ns -4.16 1.15 Wet gluten 15 6 2 1 4 0.162 0.137 ns -2.50 1.03 Mixing time 14 6 2 2 3 0.353 <0.05 * 2.58 0.91 Peak resistance 12 5 1 1 2 0.411 <0.05 * -2.26 1.07 Resistance breakdown 12 5 1 1 2 0.003 0.869 ns -0.16 1.00

18

Bread loaf voulme 7 3 2 2 3 0.127 0.433 ns -2.03 1.07

3. Results

3.1 Effects of ozone

3.1.1 Yield components

Figure 2 a show the response function for grain yield, where there is a significant negative

relationship between [O3]day and relative grain yield. In Figure 2 b the same response function is

presented but separated into the three geographical regions where the experiments are

performed (Europe, North America, and Asia). Significant negative relationships are found for all

continents but with somewhat different values for R2, where data for North America is more

scattered and had a smaller slope for the O3 effect than European and Asian data, for which the

slope was very similar. Subgroup analysis of grain yield (Figure 3) did not show any statistical

differences between wheat types or fumigation facilities; however the average effect on OTC

tends to be stronger than for FACE, using NF as the reference.

y = -0.0047x + 1.00 R² = 0.53***

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0 20 40 60 80 100 120

rela

tive

gra

in y

ield

a)

20

Figure 2. Response functions for grain yield and average daytime O3 concentration, a) all data b) divided by continents (Europe, North America, and Asia).

Figure 3. Meta-analysis for grain yield a) using CF as the reference with subgroup analysis of wheat type (winter vs. spring) for grain yield and grain mass, b) using NF as the reference with subgroup analysis for wheat type (winter vs. spring) and fumigation facility (OTC vs. FACE) for grain yield and grain mass. Numbers within brackets are the number of comparisons included in the analysis for each variable.

y = -0.0060x + 1.02 R² = 0.76***

y = -0.0043x + 1.02 R² = 0.33***

y = -0.0052x + 0.99 R² = 0.82***

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0 20 40 60 80 100 120

rela

tive

gra

in y

ield

[O3] daytime, ppb

Europe

N. America

Asia

-40 -35 -30 -25 -20 -15 -10 -5 0

overall (83)

spring (34)

winter (49)

% change from CF

-40 -35 -30 -25 -20 -15 -10 -5 0

overall (75)

spring (24)

winter (51)

OTC (59)

FACE (16)

% change from NF

b)

a)

b)

21

O3 exposure also has a significant negative effect on grain mass as shown in Figure 4. Percentage

change per ppb (slope), statistical significance, intercept, and number of experiments for harvest

index, specific grain mass, grain number and total aboveground biomass can be found in Table 4.

When comparing the slopes of response regression of the main yield components it can be

concluded that the effects are largest on grain mass. As illustrated in Figure 5, the negative

effects of O3 on grain yield is dominated by the effects on grain mass, leaving the effects on grain

number and specific grain mass minor factors, although these are significantly affected.

Figure 4. Response function for grain mass and daytime O3 concentration.

Figure 5. Relative effect of O3 on grain yield vs. the corresponding effect on grain mass. The red dashed line represents the theoretical situation where the effect of O3 on grain yield is entirely explained by the effect on grain mass.

y = -0.0038x + 1.01 R² = 0.54***

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0 20 40 60 80 100 120

rela

tive

gra

in m

ass

[O3] daytime, ppb

y = 1.16x + 0.015 R² = 0.90

-0.2

0.0

0.2

0.4

0.6

0.8

-0.2 0.0 0.2 0.4 0.6 0.8

rela

tive

eff

ect

grai

n y

ield

relative effect grain mass

22

Meta-analysis output for specific grain mass, grain number, grain mass, harvest index is given in

Figure 6a and b, with CF and NF as control. All variables, except specific grain mass, significantly

decrease with ozone exposure using CF as reference, and subgroups of winter and spring wheat

are not significantly different for any of the variables (Appendix 1, Figure 27-29). The pattern

looks similar using NF as reference, but in this case also the negative effect on specific grain

mass is significant. No differences could be found for the subgroup analysis of spring and winter

wheat (Appendix 1, Figure 27-29).

Figure 6. Meta-analysis of specific grain mass, grain number, grain mass, and harvest index, a) with CF as the reference and b) with NF as the reference. Numbers within brackets are the number of comparisons included in the analysis for each variable.

An additional comparison was made between CF and NF treatments for grain yield and grain

mass. For grain yield the average percentage change was -6.83 (CI -11.25 -1.05) and for grain

mass –3.85 (CI -6.5 -1.56) and the number of comparisons for each variable being 33 and 29,

respectively. This analysis shows that present ambient O3 concentrations are sufficiently high

cause statistically significant reductions in the response variables.

3.1.2 Protein and Starch

Regardless of reference treatment used (CF or NF), grain protein concentration was significantly

positively affected by O3 exposure, whereas the effect on grain protein yield was significantly

negative (Figure 7a and b). Moreover, subgroup analysis (Appendix 1, Figure 30) of wheat type

and exposure system (only NF as control) did not show any significant differences between

groups. O3 had a significant negative effect on grain starch concentration, both using CF and NF

as reference, however only with borderline significance for spring wheat with NF as control

(Appendix 1, Figure 31). Starch yield was strongly reduced by O3 using NF as the reference

(Figure 7b). Response function for protein concentration (Figure 8 a) shows that there is a

significant positive response during O3 exposure, while for starch the concentration decreases

-25 -20 -15 -10 -5 0 5

harvest index (57)

grain mass (75)

grain number (66)

specific grain mass (10)

% change from CF

-25 -20 -15 -10 -5 0

harvest index (40)

grain mass (67)

grain number (59)


% change from NF

a)

b)

23

with O3 concentration (Figure 8 b). Response function for starch yield showed the strongest

reduction among all variables included in the O3 dataset, with -7.37 % per 10 ppb and R2 = 0.81.

Also meta-analysis for starch yield resulted in strongly significant negative effects of about 20-

30%.

Figure 7. Meta-analysis for protein concentration, protein yield, starch concentration, and starch yield a) using CF as the reference and subgroup analysis of wheat type for protein concentration and yield, b) using NF as reference and with subgroup analysis of both wheat type (spring vs. winter wheat) and fumigation facility (OTC vs. FACE) for protein concentration and yield, and for starch concentration. Numbers within brackets are the number of comparisons included in the analysis for each variable.

-40 -30 -20 -10 0 10 20

protein conc. (47)

protein yield (47)

starch conc. (21)

starch yield (21)

% change from CF

-50 -40 -30 -20 -10 0 10 20

protein conc. (50)

protein yield (50)

starch conc. (32)

starch yield (32)

% change from NF

a)

b)

24

Figure 8. Response functions for a) grain protein concentration b) grain starch concentration, with average daytime O3 concentration.

3.1.3 Minerals

Mineral concentrations of plant macronutrients (P, K, Mg, Ca) and micronutrients (Zn, Mn, Fe,

Cu) were significantly positively affected by O3 using CF as the reference (Figure 9a), and with

marginal significance for S concentration. Yields were significantly negatively affected for all

mineral nutrients except Fe and Cu. The pattern is similar when using NF as the reference

(Figure 9b) but the positive effects on concentrations were not significant for S, Zn, and Mn in

this case. Contrary to the essential mineral compounds, the effect was significantly negative on

both concentration and yield of the non-essential and potentially toxic compound Cd. It should

be noted that the highest levels of Cd in the experiment were in the range of 55-60 µg kg-1 (data

not shown), which is below the EU limit value of 200 µg kg-1 for Cd concentration in wheat grain

(EC Regulation No 1881/2006). However, dietary Cd exposure depends on both concentration

and intake and cereals is a major contributor (EFSA, 2009).

y = 0.0026x + 1.00 R² = 0.40***

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0 20 40 60 80 100 120

rela

tive

pro

tein

co

nce

ntr

atio

n

[O3] daytime, ppb

y = -0.0015x + 1.02 R² = 0.40***

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0 20 40 60 80 100

rela

tive

sta

rch

co

nce

ntr

atio

n

[O3] daytime, ppb

a)

b)

25

Figure 9. Meta-analysis for mineral concentration and yield: plant micronutrients (P, K, Mg, Ca, S) and plant micronutrients (Zn, Mn, Fe, Cu), and elements non-essential to plants (Cd, Na), a) using CF as the reference and b) using NF as the reference. Number within brackets gives the number of comparisons included in the analysis for each variable. Note the different scales on the x-axis.

-80 -60 -40 -20 0 20 40 60

Zn conc. (12)Zn yield (12)

Mn conc. (13)Mn yield (13)

K conc. (27)K yield (27)

Ca conc. (21)Ca yield (21)

Mg conc. (21)Mg yield (21)

Na conc. (11)Na yield (11)

P conc. (21)P yield (21)

Fe conc. (10)Fe yield (10)

S conc. (11)S yield (11)

Cu conc. (10)Cu yield (10)

Cd conc. (9)Cd yield (9)

% change from CF

-100 -50 0 50 100 150 200 250










Cu conc. (15)Cu yield (15)


% change from NF

a)

b)

26

For concentration of Na no firm conclusions can be drawn, but yield was significantly reduced

using both CF and NF as the reference. Response relationships for P concentration and yield is

presented in Figure 10, and the pattern looks similar for all macronutrients included and the

micronutrient Zn. Slope, intercept and statistical significance for regressions of all mineral

concentrations and yields are presented in Table 4.

Figure 10. Response functions for phosphorus a) concentration and b) yield with average daytime O3 concentration.

3.1.4 Baking qualities

Due to limited amount of data for baking quality variables, meta-analysis where restricted to

only using CF as the reference (Figure 11). All variables included, Zeleny value, Hagberg falling

number, dry and wet gluten content, were significantly positively affected by O3. Response

y = 0.0044x + 0.94 R² = 0.62 ***

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0 20 40 60 80 100 120

rela

tive

P c

on

cen

trat

ion

[O3] daytime, ppb

y = -0.0051x + 0.99 R² = 0.62 ***

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0 20 40 60 80 100 120

rela

tive

P y

ield

[O3] daytime, ppb

b)

a)

27

functions for Zeleny value (Figure 12) and Hagberg falling number (Table 4) were significant,

whereas for dry and wet gluten significance were marginal with p values of 0.062 and 0.073

respectively.

Figure 31. Meta-analysis for baking properties, Zeleny value, Hagberg falling number, wet and dry gluten content, using CF as the reference. Number within brackets gives the number of comparisons included in the analysis for each variable.

Figure 12. Response function for Zeleny value with average daytime O3 concentration.

0 5 10 15 20 25 30

Hagberg (9)

Zeleny (10)

dry gluten (4)

wet gluten (4)

% change from CF

y = 0.0038x + 1.00 R² = 0.72 ***

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0 20 40 60 80 100

rela

tive

Zel

eny

valu

e

[O3] daytime, ppb

28

3.2 Effects of carbon dioxide

3.2.1 Yield components

The relative grain yield is significantly positively affected by elevated CO2 (Figure 13), but there

is a large scatter and the coefficient of determination for the linear regression is rather low

(R2=0.20). Fitting a polynomial trend line gives a slightly better fit (R2=0.25), which is suggesting

that the effect seems to level off at higher CO2 concentrations. Comparing data geographically

(Figure 14) showed that data from Australian and North American experiments were somewhat

more scattered than European, however it should be noted that more data was available for

Europe. Additionally, data was split up by rooting environment and fumigation facility (Figure

15). It could be concluded that experiments performed in “field, chambers” and “field, FACE”

were most scattered, whereas “field, OTC”, “pots, OTC”, “pots, GH” and “pots, chamber” had

strong correlations (using polynomial fit). Note that polynomial fit was only used where it

improved the correlation coefficient. Meta-analysis for grain yield is presented in Figure 16 and

the overall effect is positively significant with a percentage increase of 23.55 (CI 19.23 28.52).

Subgroup analysis of rooting environment, fumigation facility, continent and measuring unit did

not show any significant differences between groups, except for “greenhouse, GH” and

“temperature gradient tunnel, TGT”. Hence it can be concluded that the positive effects on grain

yield are robust, regardless of experimental setup or geography.

Figure 13. Response functions for grain yield and average CO2 concentration, with linear regression and polynomial trend line.

y = 0.00080x + 0.77 R² = 0.20***

y = -3E-06x2 + 0.0040x - 0.034 R² = 0.25

0.0

0.5

1.0

1.5

2.0

2.5

3.0

250 350 450 550 650 750 850 950

rela

tive

gra

in y

ield

[CO2], ppm

29

Figure 44. Response functions for grain yield and CO2 concentration divided by continents (Australia, Europe, North America). Only two data points available for Asia therefore no response function could be derived.

Figure 55. Response functions for grain yield and CO2 concentration, data separated into rooting environment and fumigation facility.

y = 0.00050x + 0.84 R² = 0.48

y = 0.0014x + 0.50 R² = 0.31

y = -5E-06x2 + 0.0054x - 0.33 R² = 0.55

y = 0.0019x + 0.40 R² = 0.20

y = -6E-06x2 + 0.0071x - 0.73 R² = 0.67

y = -3E-06x2 + 0.004x - 0.11 R² = 0.63

y = -8E-06x2 + 0.0095x - 1.38 R² = 0.74

0.0

0.5

1.0

1.5

2.0

2.5

3.0

250 350 450 550 650 750 850 950

rela

tive

gra

in y

ield

[CO2], ppm

field TGTfield FACEfield OTCfield chamberpots GHpots OTCpots chamber

y = 0.00050x + 0.90 R² = 0.16*

y = 0.00070x + 0.75 R² = 0.40***

y = 0.0019x + 0.36 R² = 0.29*

0.0

0.5

1.0

1.5

2.0

2.5

3.0

250 350 450 550 650 750 850 950

rela

tive

gra

in y

ield

[CO2], ppm

Australia

Europe

Asia

N. America

30

Figure 66. Meta-analysis for grain yield, with subgroup analysis of measuring unit, fumigation facility, rooting environment, and continent. Number within brackets presents the number of comparisons included for each variable.

Regressions between CO2 and total aboveground biomass and number of grains were both

significantly positive (Figure 17), while no significant relationships were found for grain mass,

specific grain mass and harvest index (Table 5). Results from meta-analysis supports this

conclusion, where grain number is the only yield component that is significantly affected (Figure

18). Figure 19 shows the relative effect of CO2 on grain yield vs. the corresponding effect on

grain number, showing that the effect on grain yield is dominated by the effect on grain number.

Removing the two “outliers” did not change the slope of trend line or coefficient of

determination very much (slope= 1.01 and R2=0.74 when excluded) and were consequently not

removed.

0 10 20 30 40 50 60 70 80 90

overall (65)

N. America (10)

Europe (42)

Australia (12)

spring (45)

winter (5)

pots (15)

field (50)

GC (4)

TGT (8)

OTC (37)

FACE (9)

GH (7)

g plant-1 (5)

g pot-1 (10)

g m-2 (50)

% change from ambient CO2

31

Figure 7. Response function for a) total aboveground biomass and b) grain number and CO2 concentration.

Figure 8. Meta-analysis for specific grain mass, grain number, grain mass and harvest index. Number within brackets presents the number of comparisons included for each variable.

y = 0.00090x + 0.69 R² = 0.49***

0.0

0.5

1.0

1.5

2.0

2.5

250 350 450 550 650 750 850

rela

tive

to

tal a

bo

vegr

ou

nd

bio

mas

s

[CO2], ppm

y = 0.00060x + 0.82 R² = 0.18***

0.0

0.5

1.0

1.5

2.0

2.5

3.0

250 350 450 550 650 750 850 950

rela

tive

gra

in n

um

ber

[CO2], ppm

-15 -10 -5 0 5 10 15 20 25

harvest index (47)

grain mass (56)

grain number (23)



a)

b)

32

Figure 9. Relative effect of CO2 on grain yield vs. the corresponding effect on grain number. The red dashed line represents the theoretical situation where the effect of CO2 on grain yield is entirely explained by the effect on grain number.

3.2.2 Protein and starch

Elevated CO2 had a significant negative effect on protein concentration (Figure 20), while

protein yield was significantly positive. For subgroup analysis (Appendix 2, Figure 33a) of

protein there were no significant differences between groups, except for “growth chamber, GC”

and “OTC” when comparing fumigation facility. Response functions for protein concentration

(Figure 21 a) and protein yield (Figure 21b) resulted in strongly significant relationships but

with large scatter. The derived response function for starch concentration was not significant,

while starch yield significantly increased with CO2 concentration (Table 5) and the same was

true for meta-analysis output (Figure 20).

y = 1.07x - 0.033 R² = 0.88

-2.0

-1.5

-1.0

-0.5

0.0

0.5

-2.0 -1.5 -1.0 -0.5 0.0 0.5

rela

tive

eff

ect

grai

n y

ield

relative effect grain number

33

Figure 10. Meta-analysis for concentration and yield of starch and protein, number within brackets presents the number of comparisons included for each variable.

Figure 11. Response functions for protein concentration and protein yield with CO2 concentration.

-15 -10 -5 0 5 10 15 20 25

protein conc. (42)

protein yield (37)

starch conc. (12)

starch yield (12)


y = -0.00020x + 1.04 R² = 0.16**

0.0

0.2

0.4

0.6

0.8

1.0

1.2

250 350 450 550 650 750 850 950

rela

tive

pro

tein

co

nce

ntr

atio

n

[CO2], ppm

y = 0.00040x + 0.86 R² = 0.22***

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

250 350 450 550 650 750 850 950

rela

tive

pro

tein

yie

ld

[CO2], ppm

a)

b)

34

3.2.3 Minerals

The general effect of elevated CO2 on mineral concentration was either negative or non-

significant (Table 5). Response functions for concentrations of nutrient minerals Fe, Zn, Ca, Mg,

P, and S, were strongly significant (Fe and Zn presented in Figure 23). Also the effect on the

potentially toxic compound Cd was significantly negative. For concentrations of other mineral

compounds (Mn, K, Na, Cu, and Cr) response regressions were non-significant. For mineral

yields the effects were either positive or insignificant (Table 5) with the strongest positive

relationship found for Na and K yield, with the percentage effect per 100 ppm CO2 being 12.26

and 10.79 respectively. Output from meta-analysis (Figure 22) is in line with response

regressions regarding mineral concentrations. For mineral yield, there was a significant negative

effect on Cd, Fe, Mg, and Mn, while Zn yield was positively affected. Results for yields of S, P, and

K were inconclusive.

Figure 12. Meta-analysis for mineral concentration and yield, number within brackets presents the number of comparisons included for each variable.

-40 -30 -20 -10 0 10 20 30












35

Figure 13. Response functions for concentrations of a) iron and b) zinc with CO2 concentration.

3.2.4 Baking quality

Baking quality properties were mainly negatively affected by elevated CO2. Significant negative

effects from meta-analysis were found for bread loaf volume, peak resistance, wet and dry

gluten content, Zeleny value, and Hagberg falling number (Figure 24). For resistance breakdown

result was insignificant while effect on mixing time was significantly positive. The strongest

effect was found for Zeleny value, response function presented in Figure 25. Response

regressions for other baking quality properties are given in Table 5.

y = -0.00080x + 1.28 R² = 0.71***

0.0

0.2

0.4

0.6

0.8

1.0

1.2

250 350 450 550 650 750

rela

tive

Fe

con

cen

trat

ion

[CO2], ppm

y = -0.00040x + 1.17 R² = 0.38***

0.0

0.2

0.4

0.6

0.8

1.0

1.2

250 350 450 550 650 750

rela

tive

Zn

co

nce

ntr

atio

n

[CO2], ppm

a)

b)

36

Figure 14. Meta-analysis for baking properties bread loaf volume, resistance breakdown, peak resistance, mixing time, dry and wet gluten, Zeleny value, and Hagberg falling number. Number within brackets presents the number of comparisons included for each variable.

Figure 15. Response function for Zeleny value and CO2 concentration.

-30 -20 -10 0 10 20 30

hagberg (6)

zeleny (6)

dry gluten (6)

wet gluten (6)

mixing time (6)

peak resistance (5)

resistance breakdown (5)

bread loaf volume (3)


y = -0.00060x + 1.20 R² = 0.53*

0.0

0.2

0.4

0.6

0.8

1.0

1.2

250 350 450 550 650 750 850

rela

tive

Zel

eny

valu

e

[CO2], ppm

37

3.3 Interactive effects of carbon dioxide and ozone

Meta-analysis of the factorial designed interaction experiments is presented in Figure 26. For

grain yield there was a significant effect only with CO2 treatment, but with borderline

significance for the positive effect of CO2*O3 treatments. For protein concentration all treatments

were significantly different from each other, where a negative effect was observed for the

CO2*O3 and CO2 treatments while O3 effects were positive. Analyses were also conducted for

grain mass and harvest index but with no significant effects being observed, expect for ozone

treatment in grain mass (Appendix 3, Figure 34).

§

Figure 16. Meta-analysis for a) grain yield and b) grain protein concentration. Subgroup analysis of treatment types: elevated CO2*O3, elevated CO2, and elevated O3 compared to control treatment with ambient CO2 and O3. Number within brackets presents the number of comparisons included for each treatment type.

-20 -10 0 10 20 30 40

O3 (5)

CO2 (5)

CO2*O3 (6)

grain yield

-20 -10 0 10 20 30 40

O3 (4)

CO2 (3)

CO2*O3 (4)

% change from control

protein concentration

38

4. Discussion

4.1 Ozone effects

In line with previous reviews (Feng et al 2008, Wang and Frei, 2011) this study shows that there

are a number of significant effects of O3 on wheat grain yield and its quality traits, although of

different magnitude and direction. However, the present study contains the most

comprehensive investigation for effects of O3 on wheat quality (Broberg et al., 2015)

All yield components were negatively affected by O3 exposure (Equation 2). The strongest effect

was found for grain yield, where most of the decrease in grain yield can be attributed to the

decrease in grain mass (Figure 5), while the number of grains was not affected to the same

extent but still with a significant decrease. As shown in earlier studies (Grandjean and Fuhrer,

1989) reduction in grain mass could be explained by the post-anthesis senescence-promoting

effect of O3 that acts to shorten the growing period for wheat grains, hence there is a shorter

time for grain filling (Gelang et al., 2000). The number of grains is mainly determined in an

earlier stage of plant growth when O3 sensitivity is lower (Soja et al., 2000); therefore this

variable is not affected very strongly. As stated earlier, grain mass is an important variable for

the market value of wheat grains and high O3 exposure could possibly have significant effects on

market price.

𝐺𝑌 ↓= 𝑇𝐴𝐵 ↓∗ 𝐻𝐼 ↓= 𝐺𝑁(↓) ∗ 𝐺𝑀 ↓ (2)

The meta-analysis comparing CF and NF treatments for grain yield and grain mass resulted in

significant negative effect for both variables, which implies that current O3 exposure already has

a negative impact on wheat crop yields. This is a novel result which is of large importance in an

environmental policy context, since current effects are more convincing when it comes to

decisions concerning reductions of O3 precursors than effects from elevated O3 treatments.

Another important observation, not highlighted to any large extent in earlier reviews and meta-

analyses, is the significant negative effect on starch concentration. The research synthesis by

Wang and Frei (2011) found this effect on some crops, although not in cereals. They suggested

that key enzymes for starch synthesis might be inhibited during environmental stress, such as O3

exposure, resulting in a hampered conversion from starch to sugars. The decrease in grain mass

and specific grain mass also indicates that smaller and/or malformed grains are produced where

a smaller fraction of starch is a logical consequence. Since there are significant reductions of

both grain yield and starch concentration, this result in an even larger effect on starch yield,

which is the single strongest effect found among the variables included in this study. The areal

yield of starch has previously been neglected, but in terms of food energy supply it may have

serious impacts on food security.

Regardless of wheat type or exposure system (Appendix 1, Figure x), there was a significant

positive effect on protein concentration while the effect was significant but negative for protein

yield. The same pattern was observed for most of the nutrient minerals using CF as the

reference; however the limited amount of data did not allow any subgroup analysis of wheat

type or exposure system. With NF as the control pattern look similar but effects on mineral

concentration were not entirely consistent, with S, Zn, Mn and Cu not showing any significant

39

effects. Similar effects on protein and mineral concentrations have been found in some previous

individual studies, e.g. of Ca, Mg, K, and P by Fuhrer et al. (1990), K and P by Vandermeiren et al.

(1992), K and Ca by Feng et al. (2008); and Mg, P, and K by Pleijel et al. (2006). Yet, combing the

results gives stronger statistical power and shows that the pattern is similar regardless of

cultivar and growing environment. The increase in concentration of protein and nutrient

minerals is considered as an improvement of grain quality; while the negative effect on areal

yield means that the total amount of nutrients accumulated per area unit is reduced. In areas of

food scarcity this might be a serious issue for human nutrition, in addition to the overall

decrease in grain yield and starch that result in reduction of food energy supply.

There are several factors that might contribute to the observed pattern where protein and

mineral yield is reduced while concentrations increase. Since the period for growth and grain

filling is reduced by O3 so will the time for uptake of nutrient from the soil be. It has also been

observed that roots are disproportionally negatively affected by O3 (Cooley and Manning, 1987),

which consequently causes a decrease in nutrient uptake. Moreover, O3 tends to decrease overall

plant growth and nutrient uptake will then decrease regardless of the partitioning effects.

According to common crop physiology there is a general anti-correlation between nitrogen

(protein) concentration and crop yield, because nitrogen uptake is often strongly source limited

(Kibite and Evans, 1984). This means that crops grown under environmental stresses, like O3,

will maintain nitrogen uptake to a larger extent than biomass accumulation resulting in an

increased protein concentration (Wang and Frei, 2011). Looking at the results of nutrient

minerals in this study they seem to behave in a similar way, thus the same principle can be used

to explain the observed pattern of increasing mineral concentration. Additionally, the decrease

in protein yield, corresponding to a reduction in nitrogen uptake, results in a reduced efficiency

in the utilization of applied nitrogen. This could cause an increase in nitrate leaching and

emission of nitrous oxide (N2O), and therefore being an important issue for assessment of

eutrophication.

For the non-essential element Cd both concentration and yield were decreasing under O3

exposure, which is considered as positive effects due to the potential toxicity of this compound.

Uptake and transport of Cd is known to depend on transpiration (Salt et al., 1995) thus the

observed pattern could be explained by a reduction in stomatal conductance caused by O3

exposure, which is a well-known fact (Mulholland et al., 1997). Concentration of Na, another

non-essential element, was not significantly affected by O3; however there was a significant

negative effect on Na yield. These observations point towards a general pattern where passive

transport with the transpiration stream through the plant is a strong feature for non-essential

compounds, while specific uptake mechanisms are an important function for the essential

nutrient minerals.

Other positive effects on wheat grain quality were the improvement of baking quality traits.

Both gluten quality (Zeleny value) and quantity (wet and dry gluten concentration) was

significantly positively affected, which could be explained by the overall increase in protein

concentration. Also Hagberg falling number was significantly increased, reflecting a lower α-

amylase activity that may be explained by a premature ripening of the grain under O3 exposure.

Low falling numbers, thus high α-amylase activity, is usually associated with late ripening, which

is lowering the quality and may in worst case (under humid conditions) result in grains

germinating before harvest (Lunn et al., 2001).

40

The comparison between wheat types (spring vs. winter wheat) and fumigation facilities (OTC

vs. FACE) did not result in any statistical differences between groups. For most of the variables

included in response regression, correlation coefficient was strong and data not very scattered.

This suggests that the effects of O3 are fairly consistent regardless of wheat type and growing

environment. However, it was not attempted to examine the differences between cultivars, since

each cultivar was only used in one or just a few experiments. There are also important agro-

environmental differences between experimental sites that probably have a large effect on how

the cultivars perform. But, there are examples of earlier studies comparing cultivars, where

different cultivars have been grown under the same conditions to examine differences in yield

quantity and quality (Pleijel et al., 2006; Sarkar and Agrawal, 2010).

4.2 Carbon dioxide effects

From this study it can be concluded that elevated CO2 acts to stimulate photosynthesis of wheat

plants, resulting in a significant increase of several yield components (Equation 3). There is a

larger effect on total aboveground biomass than grain yield, suggesting that other plant

structures, such as leafs, are more stimulated than crop yield. The negative effect on harvest

index is however not significant. Additionally, grain number significantly increases under

elevated CO2. As shown in Figure x the increase in grain yield can mainly be attributed to the

increase in grain number while grain mass is not significantly affected, which are observations

in line with previous studies (Pleijel and Uddling, 2012).

𝐺𝑌 ↑= 𝑇𝐴𝐵 ∗ 𝐻𝐼 (𝑛𝑠) = 𝐺𝑁 ↑∗ 𝐺𝑀(𝑛𝑠) (3)

Response regression of grain yield shows a large scatter of data, which implies that other factors

than CO2 are of importance for grain yield stimulation. In a recent study by Pleijel and Högy

(2015) it was shown that there is a strong relationship between CO2 stimulation of wheat grain

yield and nitrogen status of the plant, and subsequently the rate of nitrogen-fertilization will

have importance for effects on overall crop yield. However, the impact of fertilization rates was

not within the scope of this study. Temperature is another factor that possibly could be

correlated to CO2 stimulation of grain yield, which has been tested in some factorial experiments

with CO2 and temperature (Bencze et al., 2004; Hakala 1998; Rawson, 1995; van Oijen et al.,

1999), although with inconclusive results. Due to the fact that elevated CO2 may increase water

use efficiency of plants it could also be hypothesized that CO2 effects on grain yield would be

larger under water limited conditions. Even though no significant difference was found for grain

yield when comparing wheat types (spring vs. winter wheat) there may still be cultivar

differences, which partly could explain the variation in CO2 response. Still it remains unclear

whether the response to elevated CO2 is determined by environmental conditions or the genetic

properties of the plant, or maybe more likely: a combination.

A significant reduction in protein concentration and several nutrient minerals (S, Fe, P, Mg, Ca,

Mn, Zn) was observed, which is a pattern already observed in previous studies (Myers et al.,

2014; Taub et al., 2008). The response regression for protein concentration shows a significant

effect, but data is very scattered (R2= 0.16) proposing that other factors than CO2 concentration

controls the content of grain protein, such as nitrogen-fertilization rates. The decrease in protein

concentration could be a result of a reduced nitrogen uptake, caused by a slower uptake in roots

and a decreased transpiration-driven mass flow of nitrogen (Myers et al., 2014). Carbohydrate

dilution has also been suggested as an explanation (Gifford et al., 2000), but the hypothesis

41

could not be confirmed in this study since starch concentration was not significantly affected.

However, concentrations of other carbohydrates, e.g. sugars and dietary fibers, were not

examined here. Moreover, the increase in starch yield can only be attributed to the overall

increase in grain yield. In line with O3 experiments, also elevated CO2 significantly decreased

both concentration and yield of Cd. This is not a surprising observation since transpiration rates

generally decrease under higher CO2 concentrations.

The baking properties associated with protein content, Zeleny value, wet and dry gluten content,

bread loaf volume, and peak resistance, was negatively affected when grown under elevated CO2.

This observation is not surprising since overall protein concentration is decreasing and gluten

being the major protein in wheat grains. Hagberg falling number is also significantly decreasing,

while there was a significant improvement of mixing time, but the physical understanding of this

effect remains unclear. Finally, it can be concluded that elevated CO2 acts to decrease overall

baking quality of wheat grain and thus reduces the market price.

4.3 Comparison of ozone and carbon dioxide effects

For many of the yield variables included in data-analysis effects of CO2 and O3 appears to be

opposite, e.g. for grain yield, total aboveground biomass, concentration of protein and many

nutrient minerals, and baking qualities. Due to these observations it is logic to assume that the

negative impacts of O3 could be offset to a larger or smaller extent by the positive effects of CO2.

Data for interaction experiments is very limited, only 3 included here, but they give an indication

of possible effects for some variables. For grain yield the average change is higher for CO2

treatment than for CO2*O3, but with borderline significance for the later. All the treatments for

protein concentration are significantly different from each other, negative for CO2 and CO2*O3

and positive for O3. Here the combined treatment of CO2 and O3 still results in a negative effect,

suggesting that the negative effect of O3 is stronger than the positive effect of CO2. For harvest

index and grain mass the results are inconclusive. With more experiments included confidence

intervals probably would decrease for some variables and effects would be easier to distinguish,

hence there is a need to further investigate this issue and to quantify the extent to which O3

effects can be counteracted by elevated CO2.

Moreover, for wheat plants grown under elevated CO2 a number of variables were not affected,

such as harvest index, grain mass, while the effects of O3 were strong. The same was true for a

few minerals (Ca and K) where concentration was significantly increased under O3 exposure but

not affected while grown under elevated CO2. It is also important to clarify that the negative

effects of O3 and the positive effects of CO2 are not always a result of the same mechanisms. As

shown earlier, the reduction in grain yield under O3 exposure is mainly explained by reduced

grain mass, whereas the positive effect by CO2 can be attributed to an increase in number of

grains. Therefore it could be hypothesized that the combined effect of O3 and CO2 would result in

a larger number of grains but with a reduced mass, possibly giving an enhanced grain yield but

with impaired quality.

Another general observation is that data tend to be more scattered in response functions for CO2

than for O3, protein concentration and yield being a good example. This means that the

damaging effect of O3 is highly correlated to the average daytime concentration regardless of

other influences, while for CO2 the yield stimulation could be limited by e.g. nutrient and water

42

availability, radiation, or temperature. However, further research is necessary to understand

why there is such a large variability in response of CO2 stimulation.

4.4 Food security and crop modeling

The impact of O3 and CO2 on wheat crop yield may have significant effects on a global scale. Also

climate change, in terms of changes in temperature and precipitation, may strongly affect crop

production. This means that projections of future crop yields depend both on the accuracy of up-

scaling the estimated effects of O3 and CO2 as well as the uncertainties that lies within current

climate models and emission scenarios.

The global threat on food security from O3 pollution and its interaction with climate change was

recently quantified by Tai et al. (2014). Results from their modeling suggested that the yield of

four major food crops, including wheat, would be significantly affected by the year 2050 with an

overall decrease in crop production estimated to more than 10%. However, these projections

only consider changes in the quantity of crop yield, while the present study shows that also

many qualitative variables are significantly affected. In order to obtain a more comprehensive

picture of future impacts on food security, both quantity and quality should be considered. The

reduction in yields of protein and many nutrient minerals may have serious implication for

human nutrition in areas of food scarcity where cereals are the main source of energy and

nutrients.

The effects of elevated CO2 on grain yield could be regarded as positive in terms of an increase in

food energy supply, while the decrease in concentration of many nutrients may cause

malnutrition, also referred to as “hidden hunger”. In a recent research synthesis of CO2 effects on

nutrient content in crops Myers et al. (2014) argue that the negative effects on Fe and Zn

concentration is an issue for global public health, since about two billion people suffer from

dietary deficiencies of these nutrients. They also suggest breeding of less sensitive cultivars as a

strategy to mitigate the issue of nutrient loss under high CO2 concentrations. In line with

previous work the present study shows that many quality variables are significantly affected,

but with a more extensive amount of variables included than before, e.g. mineral yields. This

gives an opportunity to now incorporate a larger number of quality traits in projections of future

crop yields.

4.5 Methods – limitations and potential developments

The methods of meta-analysis and response regressions used in this study have been applied in

several peer-reviewed papers (references given in materials and methods section), but they do

however have some limitations. Since data variance (standard deviation or standard error with

replication) was poorly reported in the available data an un-weighted approach was applied for

the meta-analysis, which have implications for calculation of confidence intervals and therefore

the statistical significance. Using a weighted approach would be preferable since then the

studies with more replicates and smaller variance would have a larger impact on the estimation

on average effect size and its confidence intervals, giving more accurate results assuming that

more replicates gives a more “true” value.

Another concern with the current meta-analysis is issue of non-independence of data within

studies. For example, data have been treated independently even if experiments were

43

performed at the same place and the same cultivar but in two consecutive years. A way to deal

with this issue would be to calculate an average of the two years or use the data with highest

replicate size. For some data, the same control treatment has been compared with several

treatments of elevated O3/CO2, where a solution could be to only include one of the elevated

treatments or an average. However, these approaches would drastically decrease the amount of

data available for further analysis and has consequently not been applied in this study.

For some yield variables there are very few data included in analysis and it is possible that a

larger dataset would give significantly different outcomes. If data variance is available (using a

weighted meta-analysis) fail-safe numbers could be calculated (Koricheva et al., 2013), which

gives a measure of how robust the results are and an approximation of how many studies that

are required to change the direction and significance of the average effect.

Concerning regressions for exposure response relationships, there are also some

methodological limitations. The method is based on extrapolation of regression trend line to

estimate the effect at zero O3/ambient CO2 (already described in materials and methods

section), and uncertainties with this method becomes larger for experiments with fewer

treatments. For O3 database additional inclusion criteria was defined to deal with this issue, but

for CO2 database a similar approach could not be applied since the majority of experiments only

had two treatments. This could also explain why CO2 data is more scattered than O3 when put on

a relative scale. Although the difference between ambient and elevated treatment is larger for

CO2 than O3 data thus the extrapolation uncertainties will be smaller. In contrary to meta-

analysis, response functions are generally more sensitive to outliers, especially for variable with

limited amount of data.

44

5. Conclusions

Main effects of O3 on wheat crop yield included:

O3 strongly decreases grain yield, grain mass, harvest index, and total aboveground

biomass, while grain number and specific grain mass more weakly affected, although

significantly decreasing.

Concentration of protein and a number of nutrient minerals (P, Mg, K, Ca, Zn, Mn) were

significantly enhanced by O3 exposure, whereas the areal yield of these constituents was

strongly reduced.

O3 significantly reduced both concentration and yield of the potentially toxic compound

Cd.

O3 significantly reduced starch concentration and yield, and food energy (starch yield) is

even more negatively affected by O3 than grain yield – a novel finding.

O3 positively affects several baking properties; due to enhanced protein concentration

and possibly by O3 induced premature senescence.

The effects of O3 on grain yield and grain mass did not vary between geographical

regions (continents).

There were no significant differences of effects when comparing fumigation facilities

(OTC vs. FACE).

Main effects of elevated CO2 on wheat crop yield were:

CO2 strongly increased grain yield, grain number, and total aboveground biomass, but

harvest index, grain mass and specific grain mass were not significantly affected.

CO2 significantly reduced concentration of protein and a number of nutrient minerals (S,

Fe, P, Mg, Ca, Mn, Zn). Protein yield was enhanced due to the overall increase in grain

yield, which was also true for the yield of Zn yield, while the yield of other elements was

significantly reduced (Fe, Mg, and Mn).

CO2 significantly reduced both concentration and yield of the potentially toxic compound

Cd.

CO2 mainly had a negative effect on baking properties, as a result of the reduction in

overall protein concentration.

CO2 effects on grain yield and grain number did not vary between geographical regions

(continents), while for grain mass the effect was positive for North America but

insignificant for Europe and Australia.

Comparing experimental setup for effects on grain yield showed no difference between

rooting environments (pots vs. field) but significant difference between fumigation

facility “greenhouse” (strong effect) and “temperature gradient tunnel” (weak effect).

For grain mass effects were insignificant for field experiments but significantly negative

for potted plants.

45

Acknowledgements

Sincere thanks to my supervisor Håkan Pleijel (Department of Biological and Environmental

Sciences, University of Gothenburg) for great guidance and support during this project. I also

would like to thank Zhaozhong Feng (State Key Laboratory of Urban and Regional Ecology,

Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences), Petra Högy

(Institute for Landscape and Plant Ecology, Universität Hohenheim) and Lisbeth Mortensen

(National Environmental Research Institute, Department of Atmospheric Environment,

Denmark)for and providing additional data to this study.

46

References

References used for the general text

Ainsworth, E.A., McGrath, J.M., 2010. Direct Effects of Rising Atmospheric Carbon Dioxide and Ozone on Crop Yields, in: Lobell, D.B., Burke, M. (Eds.), Climate Change and Food Security - Adapting Agriculture to a Warmer World. Springer.

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

Bencze, S., Veisz, O., Bedo, Z., 2004. Effects of high atmospheric CO2 and heat stress on phytomass, yield and grain quality of winter wheat. Cereal Research Communications 32, 75-82.

Cooley, D.R., Manning, W.J., 1987. The Impact of Ozone on Assimilate Partitioning in Plants - a Review. Environmental Pollution 47, 95-113.

Curtis, P.S., Wang, X.Z., 1998. A meta-analysis of elevated CO2 effects on woody plant mass, form, and physiology. Oecologia 113, 299-313.

Eckert, B., Amend, T., Belitz, H.D., 1993. The Course of the Sds and Zeleny Sedimentation Tests for Gluten Quality and Related Phenomena Studied Using the Light-Microscope. Zeitschrift Fur Lebensmittel-Untersuchung Und-Forschung 196, 122-125.

EFSA, 2009. Cadmium in Food, The EFSA Journal 980, pp. 1-139.

Farage, P.K., Long, S.P., Lechner, E.G., Baker, N.R., 1991. The sequence of change within the photosynthetic apparatus of wheat following short-term exposureto ozone Plant Physiology 95, 529-535.

Feng, Z.Z., Kobayashi, K., Ainsworth, E.A., 2008. Impact of elevated ozone concentration on growth, physiology, and yield of wheat (Triticum aestivum L.): a meta-analysis. Global Change Biology 14, 2696-2708.

Finney, K.F., Shogren, M.D., 1972. 10-G mixograph for determingin and predicting functional properties of wheat flours. Cereal Science Today 17, 262-&.

Fuhrer, J., Lehnherr, B., Moeri, P.B., Tschannen, W., Shariatmadari, H., 1990. Effects of Ozone on the Grain Composition of Spring Wheat Grown in Open-Top Field Chambers. Environmental Pollution 65, 181-192.

Fuhrer, J., Skarby, L., Ashmore, M.R., 1997. Critical levels for ozone effects on vegetation in Europe. Environmental Pollution 97, 91-106.

Gelang, J., Pleijel, H., Sild, E., Danielsson, H., Younis, S., Sellden, G., 2000. Rate and duration of grain filling in relation to flag leaf senescence and grain yield in spring wheat (Triticum aestivum) exposed to different concentrations of ozone. Physiologia Plantarum 110, 366-375.

Gelang, J., Sellden, G., Younis, S., Pleijel, H., 2001. Effects of ozone on biomass, non-structural carbohydrates and nitrogen in spring wheat with artificially manipulated source/sink ratio. Environmental and Experimental Botany 46, 155-169.

Gifford, R.M., Barrett, D.J., Lutze, J.L., 2000. The effects of elevated [CO2] on the C : N and C : P mass ratios of plant tissues. Plant and Soil 224, 1-14.

Grandjean, A., Fuhrer, J., 1989. Growth and Leaf Senescence in Spring Wheat (Triticum-Aestivum) Grown at Different Ozone Concentrations in Open-Top Field Chambers. Physiologia Plantarum 77, 389-394.

47

Hakala, K., 1998. Growth and yield potential of spring wheat in a simulated changed climate with increased CO2 and higher temperature. European Journal of Agronomy 9, 41-52.

Hruskova, M., Skodova, V., Blazek, J., 2004. Wheat sedimentation values and falling number. Czech Journal of Food Sciences 22, 51-57.

Högy, P., Brunnbauer, M., Koehler, P., Schwadorf, K., Breuer, J., Franzaring, J., Zhunusbayeva, D., Fangmeier, A., 2013. Grain quality characteristics of spring wheat (Triticum aestivum) as affected by free-air CO2 enrichment. Environmental and Experimental Botany 88, 11-18.

IPCC, 2013. Summary for Policymakers. In: Climate Change 2013: The physical Science Basis. Contribution of Working group I to the Fifth Assessment Report of Intergovermental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Bouschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Camebridge University Press, Cambridge, United Kingdom and New York, USA

Kibite, S., Evans, L.E., 1984. Causes of Negative Correlations between Grain-Yield and Grain Protein-Concentration in Common Wheat. Euphytica 33, 801-810.

Kimball, B.A., Morris, C.F., Pinter, P.J., Wall, G.W., Hunsaker, D.J., Adamsen, F.J., LaMorte, R.L., Leavitt, S.W., Thompson, T.L., Matthias, A.D., Brooks, T.J., 2001. Elevated CO2, drought and soil nitrogen effects on wheat grain quality. New Phytologist 150, 295-303.

Kindred, D.R., Gooding, M.J., Ellis, R.H., 2005. Nitrogen fertilizer and seed rate effects on Hagberg failing number of hybrid wheats and their parents are associated with alpha-amylase activity, grain cavity size and dormancy. Journal of the Science of Food and Agriculture 85, 727-742.

Koricheva, J., Gurevitch, J., Mengersen, K., 2013. Handbook of Meta-analysis in Ecology and Evolution. Princeton University Press, Princeton and Oxford.

Long, S.P., 1985. Leaf gas exchange, in: Barber, J., Baker, N.R. (Eds.), Photosynthetic Mechanisms and the Environment. Elsevier, Amsterdam.

Long, S.P., Ainsworth, E.A., Leakey, A.D.B., Nosberger, J., Ort, D.R., 2006. Food for thought: Lower-than-expected crop yield stimulation with rising CO2 concentrations. Science 312, 1918-1921.

Lunn, G.D., Major, B.J., Kettlewell, P.S., Scott, R.K., 2001. Mechanisms leading to excess alpha-amylase activity in wheat (Triticum aestivum, L) grain in the UK. Journal of Cereal Science 33, 313-329.

Manley, M., Engelbrecht, M.L., Williams, P.C., Kidd, M., 2009. Assessment of variance in the measurement of hectolitre mass of wheat, using equipment from different grain producing and exporting countries. Biosystems Engineering 103, 176-186.

McKee, I.F., Farage, P.K., Long, S.P., 1995. The interactive effects of elevated CO2 and O-3 concentration on photosynthesis in spring wheat (vol 45, pg 111, 1995). Photosynthesis Research 46, 479-479.

Mills, G., Pleijel, H., Braun, S., Buker, P., Bermejo, V., Calvo, E., Danielsson, H., Emberson, L., Fernandez, I.G., Grunhage, L., Harmens, H., Hayes, F., Karlsson, P.E., Simpson, D., 2011. New stomatal flux-based critical levels for ozone effects on vegetation. Atmospheric Environment 45, 5064-5068.

Mishra, A.K, Rai, R., Agrawal, S.B., 2013. Differential response of dwarf and tall tropical wheat cultivars to elevated ozone with and without carbon dioxide enrichment: growth, yield and grain quality. Field Crops Research 145, 21-32.

Mulholland, B.J., Craigon, J., Black, C.R., Colls, J.J., Atherton, J., Landon, G., 1997. Impact of elevated atmospheric CO2 and O-3 on gas exchange and chlorophyll content in spring wheat (Triticum aestivum L.). Journal of Experimental Botany 48, 1853-1863.

Myers, S.S., Zanobetti, A., Kloog, I., Huybers, P., Leakey, A.D.B., Bloom, A.J., Carlisle, E., Dietterich, L.H., Fitzgerald, G., Hasegawa, T., Holbrook, N.M., Nelson, R.L., Ottman, M.J., Raboy, V., Sakai, H., Sartor, K.A., Schwartz, J., Seneweera, S., Tausz, M., Usui, Y., 2014. Increasing CO2 threatens human nutrition. Nature 510, 139-+.

48

Pleijel, H., Eriksen, A.B., Danielsson, H., Bondesson, N., Selldén, G., 2006. Differential ozone sensitivity in an old and a modern Swedish wheat cultivar—grain yield and quality, leaf chlorophyll and stomatal conductance. Environmental and Experimental Botany 56, 63-71.

Pleijel, H., Danielsson, H., Gelang, J., Sild, E., Sellden, G., 1998. Growth stage dependence of the grain yield response to ozone in spring wheat (Triticum aestivum L.). Agriculture Ecosystems & Environment 70, 61-68.

Pleijel, H., Uddling, J., 2012. Yield vs. Quality trade-offs for wheat in response to carbon dioxide and ozone. Global Change Biology 18, 596-605.

Pleijel, H.E., 2007. Transboundary Air Pollution. Studentlitteratur.

Rawson, H.M., 1995. Yield Responses of 2 Wheat Genotypes to Carbon-Dioxide and Temperature in-Field Studies Using Temperature-Gradient Tunnels. Australian Journal of Plant Physiology 22, 23-32.

Reeves, S.G., Baxter, E.D., Martin, H.L., Wainwhright, T., 1978. Prediction of the malting quality of barley by a modified zeleny sedimentation test. J. Inst. Brew. 85, 141-143.

Rogers, G.S., Gras, P.W., Batey, I.L., Milham, P.J., Payne, L., Conroy, J.P., 1998. The influence of atmospheric CO2 concentration on the protein, starch and mixing properties of wheat flour. Australian Journal of Plant Physiology 25, 387-393.

Rosenberg, M.S., Adams, D.C., Gurevitch, J., 2000. MetaWin: Statistical Software for Meta-Analysis, Version 2.0 ed. Sinauer Associates, Inc, Sunderland, MA, USA.

Salt, D.E., Prince, R.C., Pickering, I.J., Raskin, I., 1995. Mechanisms of Cadmium Mobility and Accumulation in Indian Mustard. Plant Physiology 109, 1427-1433.

Sarkar, A., Agrawal, S.B., 2010. Elevated ozone and two modern wheat cultivars: An assessment of dose dependent sensitivity with respect to growth, reproductive and yield parameters. Environmental and Experimental Botany 69, 328-337.

Satarug, S., Garrett, S.H., Sens, M.A., Sens, D.A., 2010. Cadmium, Environmental Exposure, and Health Outcomes. Environmental Health Perspectives 118, 182-190.

Shewry, P. R., 2009. Wheat. Journal of Experimental Botany 60, 1537-1553

Soja, G., Barnes, J.D., Posch, M., Vandermeiren, K., Pleijel, H., Mills, G., 2000. Phenological weighting of ozone exposures in the calculation of critical levels for wheat, bean and plantain. Environmental Pollution 109, 517-524.

Tai, A.P.K., Val Martin, M., Heald, C.L., 2014. Threat to future global food security from climate change and ozone air pollution. Nature Climate Change Advance Online Publication.

Taub, D.R., Miller, B., Allen, H., 2008. Effects of elevated CO2 on the protein concentration of food crops: a meta-analysis. Global Change Biology 14, 565-575.

Van Oijen, M., Schapendonk, A.H.C.M., Jansen, M.J.H., Pot, C.S., Maciorowski, R., 1999. Do open-top chambers overestimate the effects of rising CO2 on plants? An analysis using spring wheat. Global Change Biology 5, 411-421.

Vandermeiren, K., Detemmerman, L., Staquet, A., Baeten, H., 1992. Effects of Air Filtration on Spring Wheat Grown in Open-Top Field Chambers at a Rural Site .2. Effects on Mineral Partitioning, Sulfur and Nitrogen-Metabolism and on Grain Quality. Environmental Pollution 77, 7-14.

Wang, Y.X., Frei, M., 2011. Stressed food - The impact of abiotic environmental stresses on crop quality. Agriculture Ecosystems & Environment 141, 271-286.

Vingarzan, R., 2004. A review of surface ozone background levels and trends. Atmospheric Environment 38, 3431-3442.

49

Weiss, A., Moreno-Sotomayer, A., 2006. Simulating grain mass and nitrogen concentration in wheat. European Journal of Agronomy 25, 129-137.

Wu, D.X., Wang, G.X., Bai, Y.F., Liao, J.X., 2004. Effects of elevated CO2 concentration on growth, water use, yield and grain quality of wheat under two soil water levels. Agriculture Ecosystems & Environment 104, 493-507.

Zhu, X., Feng, Z., Sun, T., Liu, X., Tang, H., Zhu, J., Guo, W., Kobayashi, K., 2011. Effects of elevated ozone concentration on yield of four Chinese cultivars of winter wheat under fully open-air field conditions. Global Change Biology 17, 2697-2706.

References used in data analysis

Ozone database

De Temmerman, L., Vandermeiren, K., Guns, M., 1992. Effects of air filtration on spring wheat grown in open-top field chambers at a rural site. 1. Effect on growth, yield and dry-matter partitioning. Environmental Pollution 77, 1-5.

Fangmeier, A., De Temmerman, L., Mortensen, L., Kemp, K., Burke, J., Mitchell, R., van Oijen, M., Weigel, H.J., 1999. Effects on nutrients and on grain quality in spring wheat crops grown under elevated CO2 concentrations and stress conditions in the European, multiple-site experiment 'ESPACE-wheat'. European Journal of Agronomy 10, 215-229.

Feng, Z.Z., Yao, F.F., Chen, Z., Wang, X.K., Meng, Q.W., Feng, Z.W., 2007. Response of gas exchange and yield components of field-grown Triticum aestivum L. to elevated ozone in China. Photosynthetica 45, 441-446.

Fuhrer, J., Egger, A., Lehnherr, B., Grandjean, A., Tschannen, W., 1989. Effects of ozone on the yield of spring wheat (Triticum-aestivum L., cv. Albis) grown in open-top field chambers. Environmental Pollution 60, 273-289.

Fuhrer, J., Grimm, A.G., Tschannen, W., Shariatmadari, H., 1992. The response of spring wheat (Triticum aestivum L.) to ozone at higher elevations. 2. Changes in yield, yield components and grain quality in response to ozone flux. New Phytologist 121, 211-219.

Fuhrer, J., Lehnherr, B., Moeri, P.B., Tschannen, W., Shariatmadari, H., 1990. Effects of ozone on the grain composition of spring wheat grown in open-top field chambers. Environmental Pollution 65, 181-192.

Gelang, J., Pleijel, H., Sild, E., Danielsson, H., Younis, S., Sellden, G., 2000. Rate and duration of grain filling in relation to flag leaf senescence and grain yield in spring wheat (Triticum aestivum) exposed to different concentrations of ozone. Physiologia Plantarum 110, 366-375.

Gelang, J., Sellden, G., Younis, S., Pleijel, H., 2001. Effects of ozone on biomass, non-structural carbohydrates and nitrogen in spring wheat with artificially manipulated source/sink ratio. Environmental and Experimental Botany 46, 155-169.

Hu, H.J., 2011. Effects of free air ozone concentration enrichment on absorption and distribution of nutrient element in wheat. Master thesis (in Chinese). Yangzhou University.

Kohut, R.J., Amundson, R.G., Laurence, J.A., Colavito, L., Vanleuken, P., King, P., 1987. Effects of ozone and sulfur-dioxide on yield of winter-wheat. Phytopathology 77, 71-74.

Kress, L.W., Miller, J.E., Smith, H.J., 1985. Impact of ozone on winter-wheat yield. Environmental and Experimental Botany 25, 211-228.

Mortensen, L., Engvild, K.C., 1995. Effects of ozone on C-14 translocation velocity and growth of spring wheat (Triticum-aestivum L) exposed in open-top chambers. Environmental Pollution 87, 135-140.

50

Mortensen, L., Egelov, A.H., Grundahl, L., Mansher, O.H., 1997. Final report of the ESPACE-Wheat open-top chamber experiment 1994-1996 at Roskilde, Denmark. National Environmental Research Institute, Department of Atmospheric Environment, Roskilde, Denmark: Technical report, 1-37.

Mulchi, C., Rudorff, B., Lee, E., Rowland, R., Pausch, R., 1995. Morphological responses among crop species to full-season exposures to enhanced concentrations of atmospheric CO2 and O3. Water Air and Soil Pollution 85, 1379-1386.

Mulholland, B.J., Craigon, J., Black, C.R., Colls, J.J., Atherton, J., Landon, G., 1998. Growth, light interception and yield responses of spring wheat (Triticum aestivum L.) grown under elevated CO2 and O3 in open-top chambers. Global Change Biology 4, 121-130.

Ojanpera, K., Leinonen, P., Yläranta, T., 1993. Effect of ozone on the grain yield of spring wheat in an open-top chamber experiment in Finland, in: Fuhrer, J., Achermann, B. (Eds.), Critical levels for ozone - a UN-ECE workshop report. UN-ECE Convention on Long-Range Transboundary Air Pollution Bern, Switzerland, pp. 248-251.

Ojanperä, K., Pätsikkä, E., Yläranta, T., 1993. Effects of low ozone exposure of spring wheat on net CO2 uptake, Rubisco, leaf senescence and grain filling. New Phytologist 138, 451-460.

Piikki, K., De Temmerman, L., Ojanpera, K., Danielsson, H., Pleijel, H., 2008. The grain quality of spring wheat (Triticum aestivum L.) in relation to elevated ozone uptake and carbon dioxide exposure. European Journal of Agronomy 28, 245-254.

Pleijel, H., Danielsson, H., 2009. Yield dilution of grain Zn in wheat grown in open-top chamber experiments with elevated CO2 and O3 exposure. Journal of Cereal Science 50, 278-282.

Pleijel, H., Danielsson, H., Gelang, J., Sild, E., Sellden, G., 1998. Growth stage dependence of the grain yield response to ozone in spring wheat (Triticum aestivum L.). Agriculture Ecosystems & Environment 70, 61-68.

Pleijel, H., Eriksen, A.B., Danielsson, H., Bondesson, N., Selldén, G., 2006. Differential ozone sensitivity in an old and a modern Swedish wheat cultivar - grain yield and quality, leaf chlorophyll and stomatal conductance. Environmental and Experimental Botany 56, 63-71.

Pleijel, H., Gelang, J., Sild, E., Danielsson, H., Younis, S., Karlsson, P.E., Wallin, G., Skarby, L., Sellden, G., 2000. Effects of elevated carbon dioxide, ozone and water availability on spring wheat growth and yield. Physiologia Plantarum 108, 61-70.

Pleijel, H., Ojanpera, K., Mortensen, L., 1997. Effects of tropospheric ozone on the yield and grain protein content of spring wheat (Triticum aestivum L) in the Nordic countries. Acta Agriculturae Scandinavica Section B-Soil and Plant Science 47, 20-25.

Pleijel, H., Skarby, L., Wallin, G., Sellden, G., 1991. Yield and grain quality of spring wheat (Triticum-aestivum L., cv. Drabant) Exposed to Different Concentrations of Ozone in Open-Top Chambers. Environmental Pollution 69, 151-168.

Rai, R., Agrawal, M., Agrawal, S.B., 2007. Assessment of yield losses in tropical wheat using open top chambers. Atmospheric Environment 41, 9543-9554.

Rudorff, B.F.T., Mulchi, C.L., Daughtry, C.S.T., Lee, E.H., 1996a. Growth, radiation use efficiency, and canopy reflectance of wheat and corn grown under elevated ozone and carbon dioxide atmospheres. Remote Sensing of Environment 55, 163-173.

Rudorff, B.F.T., Mulchi, C.L., Fenny, P., Lee, E.H., Rowland, R., 1996b. Wheat grain quality under enhanced tropospheric CO2 and O3 concentrations. Journal of Environmental Quality 25, 1384-1388.

Rudorff, B.F.T., Mulchi, C.L., Lee, E.H., Rowland, R., Pausch, R., 1996c. Effects of enhanced O3 and CO2 enrichment on plant characteristics in wheat and corn. Environmental Pollution 94, 53-60.

51

Sarkar, A., Agrawal, S.B., 2010. Elevated ozone and two modern wheat cultivars: An assessment of dose dependent sensitivity with respect to growth, reproductive and yield parameters. Environmental and Experimental Botany 69, 328-337.

Sild, E., 1999. Impact of increasing concentrations of O3 and CO2 on wheat, clover and pasture. Institute of Botany and Ecology, University of Tartu, Estonia, pp. 5-122.

Slaughter, L.H., Mulchi, C.L., Lee, E.H., Tuthill, K., 1989. Chronic ozone stress effects on yield and grain quality of soft red winter wheat. Crop Science 29, 1251-1255.

Tomer, R., Bhatia, A., Kumar, V., Kumar, A., Singh, R., Singh, B., Singh, S.D., 2014. Impact of Elevated Ozone on Growth, Yield and Nutritional Quality of Two Wheat Species in Northern India. Aerosol and Air Quality Research 14: 1392-1401.

Vandermeiren, K., De Temmerman, L., Staquet, A., Baeten, H., 1992. Effects of air filtration on spring wheat grown in open-top field chambers at a rural site. 2. Effects on mineral partitioning, sulfur and nitrogen-metabolism and on grain quality. Environmental Pollution 77, 7-14.

Wang, X., Zhang, Q., Zheng, F., Zheng, Q., Yao, F., Chen, Z., Zhang, W., Hou, P., Feng, Z., Song, W., Feng, Z., Lu, F., 2012. Effects of elevated O3 concentration on winter wheat and rice yields in the Yangtze River Delta, China. Environmental Pollution 171, 118-125.

Zhang, R.B., 2012. Effects of free air ozone concentration enrichment on grain quality and activity of related enzyme in wheat plants. Master thesis (in Chinese). Yangzhou University.

Zheng, F., Wang, X., Zhang, W., Hou, P., Lu, F., Du, K., Sun, Z., 2013. Effects of elevated O(3) exposure on nutrient elements and quality of winter wheat and rice grain in Yangtze River Delta, China. Environmental Pollution 179, 19-26.

Zhu, X., Feng, Z., Sun, T., Liu, X., Tang, H., Zhu, J., Guo, W., Kobayashi, K., 2011. Effects of elevated ozone concentration on yield of four Chinese cultivars of winter wheat under fully open-air field conditions. Global Change Biology 17, 2697-2706.

Carbon dioxide database

Bencze, S., Veisz, O., Bedo, Z., 2004. Effects of high atmospheric CO2 and heat stress on phytomass, yield and

grain quality of winter wheat. Cereal Research Communications 32, 75-82.

Blumenthal, C., Rawson, H.M., McKenzie, E., Gras, P.W., Barlow, E.W.R., Wrigley, C.W., 1996. Changes in

wheat grain quality due to doubling the level of atmospheric CO2. Cereal Chemistry 73, 762-766.

Deepak, S.S., Agrawal, M., 1999. Growth and yield responses of wheat plants to elevated levels of CO2 and

SO2, singly and in combination. Environmental Pollution 104, 411-419.

Dijkstra, P., Schapendonk, A.H.M.C., Groenwold, K., Jansen, M., Van de Geijn, S.C., 1999. Seasonal changes in

the response of winter wheat to elevated atmospheric CO2 concentration grown in Open-Top Chambers

and field tracking enclosures. Global Change Biology 5, 563-576.

Fangmeier, A., De Temmerman, L., Mortensen, L., Kemp, K., Burke, J., Mitchell, R., van Oijen, M., Weigel,

H.J., 1999. Effects on nutrients and on grain quality in spring wheat crops grown under elevated CO2

concentrations and stress conditions in the European, multiple-site experiment 'ESPACE-wheat'.

European Journal of Agronomy 10, 215-229.

Fangmeier, A., Gruters, U., Hertstein, U., SandhageHofmann, A., Vermehren, B., Jager, H.J., 1996. Effects of

elevated CO2, nitrogen supply and tropospheric ozone on spring wheat .1. Growth and yield.

Environmental Pollution 91, 381-390.

Fangmeier, A., Gruters, U., Hogy, P., Vermehren, B., Jager, H.J., 1997. Effects of elevated CO2 nitrogen supply

and tropospheric ozone on spring wheat .2. Nutrients (N, P, K, S, Ca, Mg, Fe, Mn, Zn). Environmental

Pollution 96, 43-59.

52

Fernando, N., Panozzo, J., Tausz, M., Norton, R., Fitzgerald, G., Seneweera, S., 2012. Rising atmospheric CO2

concentration affects mineral nutrient and protein concentration of wheat grain. Food Chemistry 133,

1307-1311.

Hakala, K., 1998. Growth and yield potential of spring wheat in a simulated changed climate with increased

CO2 and higher temperature. European Journal of Agronomy 9, 41-52.

Hogy, P., Brunnbauer, M., Koehler, P., Schwadorf, K., Breuer, J., Franzaring, J., Zhunusbayeva, D., Fangmeier,

A., 2013. Grain quality characteristics of spring wheat (Triticum aestivum) as affected by free-air CO2

enrichment. Environmental and Experimental Botany 88, 11-18.

Kimball, B.A., Morris, C.F., Pinter, P.J., Wall, G.W., Hunsaker, D.J., Adamsen, F.J., LaMorte, R.L., Leavitt,

S.W., Thompson, T.L., Matthias, A.D., Brooks, T.J., 2001. Elevated CO2, drought and soil nitrogen

effects on wheat grain quality. New Phytologist 150, 295-303.

Manderscheid, R., Bender, J., Jager, H.J., Weigel, H.J., 1995. Effects of Season Long Co2 Enrichment on

Cereals .2. Nutrient Concentrations and Grain Quality. Agriculture Ecosystems & Environment 54, 175-

185.

Mulchi, C., Rudorff, B., Lee, E., Rowland, R., Pausch, R., 1995. Morphological responses among crop species to

full-season exposures to enhanced concentrations of atmospheric CO2 and O-3. Water Air and Soil

Pollution 85, 1379-1386.

Mulholland, B.J., Craigon, J., Black, C.R., Colls, J.J., Atherton, J., Landon, G., 1998. Growth, light interception

and yield responses of spring wheat (Triticum aestivum L.) grown under elevated CO2 and O-3 in open-

top chambers. Global Change Biology 4, 121-130.

Piikki, K., De Temmerman, L., Ojanpera, K., Danielsson, H., Pleijel, H., 2008. The grain quality of spring wheat

(Triticum aestivum L.) in relation to elevated ozone uptake and carbon dioxide exposure. European

Journal of Agronomy 28, 245-254.

Pleijel, H., Danielsson, H., 2009. Yield dilution of grain Zn in wheat grown in open-top chamber experiments

with elevated CO2 and O3 exposure. Journal of Cereal Science 50, 278-282.

Pleijel, H., Gelang, J., Sild, E., Danielsson, H., Younis, S., Karlsson, P.E., Wallin, G., Skarby, L., Sellden, G.,

2000. Effects of elevated carbon dioxide, ozone and water availability on spring wheat growth and yield.

Physiologia Plantarum 108, 61-70.

Rawson, H.M., 1995. Yield Responses of 2 Wheat Genotypes to Carbon-Dioxide and Temperature in-Field

Studies Using Temperature-Gradient Tunnels. Australian Journal of Plant Physiology 22, 23-32.

Rogers, G.S., Gras, P.W., Batey, I.L., Milham, P.J., Payne, L., Conroy, J.P., 1998. The influence of atmospheric

CO2 concentration on the protein, starch and mixing properties of wheat flour. Australian Journal of Plant

Physiology 25, 387-393.

Rudorff, B.F.T., Mulchi, C.L., Daughtry, C.S.T., Lee, E.H., 1996a. Growth, radiation use efficiency, and canopy

reflectance of wheat and corn grown under elevated ozone and carbon dioxide atmospheres. Remote

Sensing of Environment 55, 163-173.

Rudorff, B.F.T., Mulchi, C.L., Fenny, P., Lee, E.H., Rowland, R., 1996b. Wheat grain quality under enhanced

tropospheric CO2 and O-3 concentrations. Journal of Environmental Quality 25, 1384-1388.

Rudorff, B.F.T., Mulchi, C.L., Lee, E.H., Rowland, R., Pausch, R., 1996c. Effects of enhanced O-3 and CO2

enrichment on plant characteristics in wheat and corn. Environmental Pollution 94, 53-60.

Saebo, A., Mortensen, L.M., 1996. Growth, morphology and yield of wheat, barley and oats grown at elevated

atmospheric CO2 concentration in a cool, maritime climate. Agriculture Ecosystems & Environment 57,

9-15.

Thompson, G.B., Woodward, F.I., 1994. Some Influences of Co2 Enrichment, Nitrogen Nutrition and

Competition on Grain-Yield and Quality in Spring Wheat and Barley. Journal of Experimental Botany

45, 937-942.

53

Van Oijen, M., Schapendonk, A.H.C.M., Jansen, M.J.H., Pot, C.S., Maciorowski, R., 1999. Do open-top

chambers overestimate the effects of rising CO2 on plants? An analysis using spring wheat. Global

Change Biology 5, 411-421.

Weigel, H.J., Manderscheid, R., 2012. Crop growth responses to free air CO2 enrichment and nitrogen

fertilization: Rotating barley, ryegrass, sugar beet and wheat. European Journal of Agronomy 43, 97-107.

Weigel, H.J., Manderscheid, R., Jager, H.J., Mejer, G.J., 1994. Effects of Season-Long Co2 Enrichment on

Cereals .1. Growth-Performance and Yield. Agriculture Ecosystems & Environment 48, 231-240.

Wolf, J., 1996. Effects of nutrient supply (NPK) on spring wheat response to elevated atmospheric CO2. Plant

and Soil 185, 113-123.

Wroblewitz, S., Huther, L., Manderscheid, R., Weigel, H.J., Watzig, H., Danicke, S., 2013. The effect of free air

carbon dioxide enrichment and nitrogen fertilisation on the chemical composition and nutritional value of

wheat and barley grain. Archives of Animal Nutrition 67, 263-278.

Ziska, L.H., 2008. Three-year field evaluation of early and late 20th century spring wheat cultivars to projected

increases in atmospheric carbon dioxide. Field Crops Research 108, 54-59.

Ziska, L.H., Morris, C.F., Goins, E.W., 2004. Quantitative and qualitative evaluation of selected wheat varieties

released since 1903 to increasing atmospheric carbon dioxide: can yield sensitivity to carbon dioxide be a

factor in wheat performance? Global Change Biology 10, 1810-1819.

Database for experiments with ozone and carbon dioxide interaction

Mulchi, C., Rudorff, B., Lee, E., Rowland, R., Pausch, R., 1995. Morphological responses among crop species to

full-season exposures to enhanced concentrations of atmospheric CO2 and O-3. Water Air and Soil

Pollution 85, 1379-1386.

Mulholland, B.J., Craigon, J., Black, C.R., Colls, J.J., Atherton, J., Landon, G., 1998. Growth, light interception

and yield responses of spring wheat (Triticum aestivum L.) grown under elevated CO2 and O-3 in open-

top chambers. Global Change Biology 4, 121-130.

Pleijel, H., Danielsson, H., 2009. Yield dilution of grain Zn in wheat grown in open-top chamber experiments

with elevated CO2 and O3 exposure. Journal of Cereal Science 50, 278-282.

Pleijel, H., Gelang, J., Sild, E., Danielsson, H., Younis, S., Karlsson, P.E., Wallin, G., Skarby, L., Sellden, G.,

2000. Effects of elevated carbon dioxide, ozone and water availability on spring wheat growth and yield.

Physiologia Plantarum 108, 61-70.

Rudorff, B.F.T., Mulchi, C.L., Daughtry, C.S.T., Lee, E.H., 1996a. Growth, radiation use efficiency, and canopy

reflectance of wheat and corn grown under elevated ozone and carbon dioxide atmospheres. Remote

Sensing of Environment 55, 163-173.

Rudorff, B.F.T., Mulchi, C.L., Fenny, P., Lee, E.H., Rowland, R., 1996b. Wheat grain quality under enhanced

tropospheric CO2 and O-3 concentrations. Journal of Environmental Quality 25, 1384-1388.

Rudorff, B.F.T., Mulchi, C.L., Lee, E.H., Rowland, R., Pausch, R., 1996c. Effects of enhanced O-3 and CO2

enrichment on plant characteristics in wheat and corn. Environmental Pollution 94, 53-60.

54

Appendix 1. Subgroup analysis of ozone effects

Figure 27. Meta-analysis of harvest index, a) using CF as the reference and b) using NF as the reference, both with subgroup analysis of wheat type (spring vs. winter). Number within brackets presents the number of comparisons included for each variable.

Figure 28. Meta-analysis of grain mass, a) using CF as the reference and subgroup analysis of wheat type (spring vs. winter) and b) using NF as the reference with subgroup analysis of fumigation facility (OTC vs. FACE) and wheat type (spring vs. winter). Number within brackets presents the number of comparisons included for each variable.

-30 -25 -20 -15 -10 -5 0

overall (57)

spring (34)

winter (23)

% change from CF

harvest index

-30 -25 -20 -15 -10 -5 0

overall (40)

spring (24)

winter (16)

& change from NF

-30 -25 -20 -15 -10 -5 0

overall (85)

spring (40)

winter (45)

% change from CF

grain mass

-30 -25 -20 -15 -10 -5 0

overall (77)

spring (34)

winter (43

OTC (55)

FACE (12)

% change from NF

a)

b)

a)

b)

55

Figure 29. Meta-analysis of grain number, a) using CF as the reference and b) using NF as the reference, both with subgroup analysis of wheat type (spring vs. winter). Number within brackets presents the number of comparisons included for each variable.

Figure 30. Meta-analysis of protein concentration, a) using CF as the reference and subgroup analysis of wheat type (spring vs. winter) and b) using NF as the reference with subgroup analysis of fumigation facility (OTC vs. FACE) and wheat type (spring vs. winter). Number within brackets presents the number of comparisons included for each variable.

-16 -14 -12 -10 -8 -6 -4 -2 0

grain number (66)

spring (19)

winter (47)

% change from CF

grain number

-30 -25 -20 -15 -10 -5 0

grain number (59)

winter (15)

spring (44)

% change from NF

0 5 10 15 20

overall (47)

spring (31)

winter (16)

% change from CF


0 5 10 15 20

overall (50)

spring (22)

winter (28)

OTC (34)

FACE (16)

% change from NF

a)

b)

a)

b)

56

Figure 31. . Meta-analysis of starch concentration, using NF as the reference with subgroup analysis of fumigation facility (OTC vs. FACE) and wheat type (spring vs. winter). Number within brackets presents the number of comparisons included for each variable. Not enough data for each subgroup conduct subgroup analysis of starch using CF as the reference.

-20 -15 -10 -5 0 5

overall (32)

spring (14)

winter (18)

OTC (16)

FACE (16)

% change from NF

starch concentration

57

Appendix 2. Subgroup analysis of carbon dioxide

-10 -5 0 5 10 15 20

overall (47)

N. America (10)

Europe (37)

spring (41)

winter (5)

pots (11)

field (36)

OTC (36)

FACE (4)

GC (7)

harvest index

-10 -5 0 5 10 15 20

overall (56)

N. America (9)

Europe (36)

Australia (7)

spring (41)

winter (3)

pots (14)

field (39)

GC (7)

TGT (5)

OTC (37)

GH (3)

grain mass

-10 -5 0 5 10 15 20

overall (23)

N. America (2)

Europe (18)

Australia (2)

winter (3)

spring (16)

pots (13)

field (10)

GH (3)

OTC (19)


grain number

a)

b)

c)

58

Figure 32. Meta-analysis for a) harvest index, b) grain mass and c) grain number with subgroup analysis of fumigation facility (OTC, FACE, GC, GH, TGT), rooting environment (field vs. pot), continent (Australia, Europe, North America). Number within brackets presents the number of comparisons included for each subgroup.

Figure 33. Meta-analysis of a) protein concentration and b) starch concentration, with subgroup analysis of fumigation facility (OTC, FACE, GC, GH, TGT), rooting environment (field vs. pot), continent (Australia, Europe, North America). Number within brackets presents the number of comparisons included for each subgroup.

-30 -25 -20 -15 -10 -5 0 5 10

overall

N. America

Europe

Australia

winter

spring

pots

field

GC

TGT

OTC

FACE

GH


-30 -25 -20 -15 -10 -5 0 5 10

overall

Europe

Australia

spring

winter

pots

field

OTC

FACE

GH


starch concentration

a)

b)

59

Appendix 3. Interactive effects of ozone and carbon dioxide

Figure 34. Meta-analysis for a) harvest index and b) grain mass. Subgroup analysis of treatment types: elevated CO2*O3, elevated CO2, and elevated O3 compared to control treatment with ambient CO2 and O3. Number within brackets presents the number of comparisons included for each treatment type.

-10 -5 0 5 10 15 20

O3 (5)

CO2 (5)

CO2*O3 (6)

harvest index

-10 -5 0 5 10 15 20

O3 (5)

CO2 (5)

CO2*O3 (6)

% change from control

grain mass

a)

b)

Documents

Effects of elevated ozone and carbon dioxide on wheat crop yield … · 2015-02-17 · 2 Abstract Elevated levels of ground level ozone (O 3) and carbon dioxide (CO 2) may have a