Post on 04-Jul-2020
transcript
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.
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).
Variable Comparisons Experiments Countries Continents Cultivars Average change (%) 95% CI Direction of change
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).
Variable Comparisons Experiments Countries Continents Cultivars Average change (%) 95% CI Direction of change
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.
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)
specific grain mass (6)
% 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
Zn conc. (20)Zn yield (20)
Mn conc. (20)Mn yield (20)
K conc. (24)K yield (24)
Ca conc. (12)Ca yield (12)
Mg conc. (20)Mg yield (20)
Na conc. (7)Na yield (7)
P conc. (12)P yield (12)
Fe conc. (7)Fe yield (7)
S conc. (7)S yield (7)
Cu conc. (15)Cu yield (15)
Cd conc. (10)Cd yield (10)
% 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)
specific grain mass (10)
% change from ambient CO2
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)
% change from ambient CO2
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
Zn conc. (16)Zn yield (16)
Mn conc. (10)Mn yield (10)
K conc. (6)K yield (6)
Ca conc. (10)Ca yield (10)
Mg conc. (8)Mg yield (8)
Na conc. (3)Na yield (3)
P conc. (6)P yield (6)
Fe conc. (12)Fe yield (12)
S conc. (10)S yield (10)
Cd conc. (7)Cd yield (7)
% change from ambient CO2
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)
% change from ambient CO2
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
protein concentration
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)
% change from ambient CO2
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
protein concentration
-30 -25 -20 -15 -10 -5 0 5 10
overall
Europe
Australia
spring
winter
pots
field
OTC
FACE
GH
% change from ambient CO2
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)