Ozone – a significant threat to future world food production?


(*Author for correspondence:tel +44 1904 434070; fax +44 1904 432998;email ma512@york.ac.uk)

Air pollution and climate change are both recognised as significant threats to food production. Ozone was established as the most important regional pollutant in terms of its impact on agriculture in North America and Europe two decades ago, but more recently, it has become clear that global background concentrations of ozone are also increasing (Vingarzan, 2004). Assessing the impact of changes in both regional and global background ozone concentrations on food production in the context of other global atmospheric and climatic changes is a major challenge (Ashmore, 2005).

However, current assessments of the effects on crop yield from changes in ozone concentrations are based on experiments carried out in open-top chambers. These chambers modify environmental conditions such as temperature, evapo-transpiration and irradiance, and hence there is uncertainty over how well they represent the real effects of ozone under field conditions. The study by Morgan et al. reported in this issue of New Phytologist (pp. 333–343) instead used free-air gas concentration enrichment (FACE) to increase ozone exposures under field conditions. This is the first large-scale study to use FACE to show the effects of ozone in reducing the yield of a major arable crop (soybean) under field conditions. Importantly, the study shows losses in soybean yield under field conditions that are at least as large as those predicted from chamber studies.

‘… there is a need to consider plant adaptation strategies to increased ozone exposure alongside climate change.’

Global implications of ozone effects on crop yield

Morgan et al.'s study was carried out in North America. A number of models have now been developed to predict future changes in global ozone concentrations, based on scenarios of precursor emissions and climate. These predictions can be linked to IPCC scenarios, so that the impacts of ozone can be considered in the context of wider global change; one example of recent predictions of change in ozone concentrations for the period 1990–2020 is shown in Fig. 1 (Dentener et al., 2005). The results are based on a ‘Current Legislation’ scenario, which incorporates expected economic development and planned emission controls in individual countries.

Figure 1.

Predicted differences in decadal annual mean surface ozone concentrations from the 1990s to the 2020s, for two global chemistry-transport models, under a ‘Current Legislation’ scenario. The upper diagram presents predictions for the TM3 model and the lower diagram presents predictions for the STOCHEM model. This figure is reproduced from fig. 11(a) of Dentener et al. (2005), with kind permission of Frank Dentener and David Stevenson.

The results predict increases in annual mean surface ozone concentrations in all major agricultural areas of the northern hemisphere. The modelled increases show a large spatial variation; they are low in the areas of North America where Morgan et al.'s study was conducted, but are high in south and east Asia. Morgan et al. reported that an increase of 13 p.p.b. in mean daytime ozone concentration caused a 20% decrease in soybean seed yield, and compared this to projected ozone concentration increases for the USA by 2050. However, the predictions of Dentener et al. (2005) indicate that this increase in ozone concentration could be reached as soon as 2020 in south Asia.

It is thus vital to consider the implications of these findings for food production outside North America and Europe. While evidence is limited, significant effects of air pollution on crop yield have been shown in Asia, Africa and Latin America (Emberson et al., 2003). In Pakistan, field studies using a chemical protectant and open-top chambers (Wahid et al., 2001) showed yield losses of about 50% in a local cultivar of soybean, at ozone levels comparable to those of Morgan et al. Although soybean is not a staple crop in south Asia, several other bean species are important, especially in India, as components of a largely vegetarian diet. In this context, effects on crop quality, which were not considered by Morgan et al., may also be significant.

In assessing the wider implications for food production and security in regions with high projected increases in ozone concentration and increasing populations, it is important to note that soybean is among the most sensitive to ozone of the major crops. The effects on wheat and rice may be lower, although there is potential for significant yield reductions of major cereal crops in east and south Asia due to future ozone exposures (Emberson et al., 2003; Wang & Mauzerall, 2004). Although maximum technically feasible reduction scenarios for precursor emissions have been identified which result in reductions in ozone exposure by the 2020s (e.g. Dentener et al., 2005), in practice these scenarios are unlikely, and there is a need to consider adaptation strategies to increased ozone exposure alongside climate change. For example, in identifying and selecting genetic traits associated with increased tolerance of drought or high temperatures, ozone tolerance should also be considered.

Ozone studies in a broader environmental context: the importance of FACE

Ozone impacts on vegetation cannot be considered in isolation, because they interact with various other environmental factors including temperature, light, water, atmospheric CO2, nutrients, pathogens and pests (Ashmore, 2005). The FACE study of Morgan et al. shows how an extreme climatic event (hailstorm) can affect yield loss due to elevated ozone. FACE is very suitable for the study of such plant–climate and other interactions with ozone since it can closely approach natural field conditions. The large scale and long-term nature of the FACE experiments also facilitates studies on ecosystem level processes such as pathogen and pest outbreaks, intraspecific and interspecific competition, and carbon and nutrient cycling. The SoyFACE study, of which the work reported by Morgan et al. is part, involves several interacting factors, i.e. elevated CO2, herbivory, drought and genotypic differences, which will yield a wealth of future information on ozone impacts to crops (e.g. Miyazaki et al., 2004; Hamilton et al., 2005).

The importance of such interactions is demonstrated by other ozone FACE studies in a young temperate tree ecosystem (Karnosky et al., 2005) and a sub-alpine semi-natural grassland community (Volk et al., 2006). In the AspenFACE experiment, elevated CO2 reduced the effects of ozone on photosynthesis and the above-ground growth of trembling aspen (Populus tremuloides) (Karnosky et al., 2003, 2005). This may have resulted from decreased stomatal conductance, an increase in detoxification capacity, or changes in other interacting factors. Carbon sequestration in soils at elevated CO2 levels was also affected at elevated ozone in this experiment (Loya et al., 2003). After 4 years of exposure to elevated ozone and CO2, soil carbon formation was reduced by 50% compared to ambient ozone and elevated CO2. These reductions most likely resulted from decreased plant litter inputs (Karnosky et al., 2003) and the enhanced microbial respiration of recent carbon inputs.

Studies on pest and pathogen outbreaks are also facilitated by minimal impediment to movement to and from the plots. An early FACE study indicated that foliar pathogens differed in their response to sulphur dioxide in winter barley and winter wheat (McLeod, 1988). Infection of a foliar pathogen on trembling aspen increased under elevated ozone in the Aspen FACE experiment, probably due to the changes in leaf surface properties (Karnosky et al., 2002). Elevated ozone also affected the performance of forest pests in the same experiment, which may be related to changes in plant chemistry or the abundance of natural enemies (Percy et al., 2002; Karnosky et al., 2003). Differential responses of tree genotypes and species were observed for photosynthesis and above-ground growth (Karnosky et al., 2005), while a FACE study in an old regularly harvested grassland showed that the relative biomass contributions of functional groups (grasses, herbs and legumes) were affected by elevated ozone (Volk et al., 2006). Cumulative ozone effects occurred over several years in both studies, emphasising the need for long-term studies, which are only possible with FACE, on whether ozone causes major shifts in species and genetic diversity in sensitive ecosystems.

The impacts of ozone need to be considered in combination with major global change factors. FACE is an excellent tool for improving such understanding, but it has some limitations. In contrast to open-top chambers, FACE systems cannot be used for ozone impact studies that include levels below ambient. Volk et al. (2003) suggested that the relatively high temporal fluctuations in the ratio of elevated ozone concentration to ambient under FACE conditions may influence biological responses to ozone. Significant spatial gradients can also occur within the large FACE plots, although this can be dealt with by careful experimental design including subsampling, subplots and the randomisation of plant genotypes and species (Karnosky et al., 2003; Volk et al., 2003; Morgan et al., 2006).


Morgan et al. provide important new evidence of the effects of ozone on crop yield under field conditions. The impacts of ozone on future food security need to be considered as an important component of global change, especially in regions with rapid economic development. However, our knowledge both of the impacts of ozone in these regions and of its interactions with other elements of global change remains very limited.