Chapter 6

INTRODUCTION

New process-based models are presented in this thesis to simulate the potential effects of putative global warming on the flux of isoprene from leaves, which is a precursor of ozone, together with the effects of tropospheric ozone on photosynthesis. In order to predict the effects of atmospheric and climate change on vegetation, the new leaf level models were scaled up to the canopy level, and subsequently incorporated as sub-models within a larger model which already contained the mechanistic biochemical equations of photosynthesis (Farquhar et al., 1980; von Caemmerer and Farquhar, 1981; Humphries and Long, 1995) (Equations of Tables A and B, Appendix I, p196 and 197). The main effects of global warming are expected to be elevated concentrations of atmospheric CO2, together with concurrent increases in temperature (Watson et al. 1990). However, concentrations of tropospheric ozone are also rising, and, as emission rates of isoprene, a major pre-cursor of ozone, are highly temperature sensitive, the new process-based isoprene emission model is used to simulate the interactive effects of elevated [CO2] and temperature on isoprene emitted from forest canopies (Trainer et al., 1987; Chameides et al., 1988; Monson et al., 1991a; 1994; PORG, 1993). Meanwhile, the interactive effects of ozone and elevated [CO2] on photosynthesis also needs to be assessed, as tropospheric concentrations of both gases are expected to continue to increase (Watson et al., 1990; PORG, 1993). The phytotoxic effects of acute ozone exposure, that is, the episodes of relatively high concentrations of ozone for a few days, are manifest as a relatively rapid decrease in photosynthetic rates and stomatal closure in wheat (Farage et al., 1991). Chronic ozone exposure results from background concentrations over the period of the whole growing season, and the effects are most obvious as a decrease in productivity (Manning and Krupa, 1992). Bearing in mind the fundamental differences between acute and chronic O3 exposure and plant responses to these two ozone regimes, the interactive effects of the two types of ozone exposure with elevated [CO2] on leaf photosynthesis, canopy assimilation and productivity are examined separately. Model predictions that consider simultaneous changes in more than one factor demonstrate the overriding importance of accounting for interactions when considering the impacts of global atmospheric change on plant production processes (Long, 1991). Interactions cannot be accounted for simply by summing the individual effects of driving variables, since one can often modify the plants response to another.

 6.2 STUDY FINDINGS AND LIMITATIONS

6.2.1 Leaf isoprene emission rates

The leaf isoprene emission model of Chapter Two (Table 2.1, p.42) predicts the interactive effects of elevated [CO2] and 2oC rise in temperature, to be an increase in leaf isoprene emission rate of 26.4%, at a light level of 1000 µmol m-2 s-1. The interactive effects of rising temperature and elevated [CO2] are also predicted to cause a reduction in the quantum yield of isoprene emission, together with an increase in the light level at which light saturated maximum isoprene emission is reached (Figure 2.17, p.68). The significance of these findings is due to both the unexpected magnitude and direction of effects. For example, the predicted reduction in quantum yield is of particular relevance to the prediction of emissions from forest canopies, when proportion of shaded leaves is dependent on leaf area index, and can lead to lower canopy emissions than might be expected. The other significance is the fact that these predictions could be made at all, which is due to the high mechanistic content of this model. Previous models used to simulate rates of isoprene emission were based on empirical relationships, and, having no basis in the underlying physiology of isoprene synthesis, could not predict the interactive effects of changing atmospheric and climate conditions (Guenther et al. 1991; 1993). Meanwhile, the mechanistically based model presented here, produces across a wide range of environmental conditions, a better simulation of observations, and therefore makes a novel contribution towards predicting the feedback effects of climate and atmospheric change.

The equations used in the construction of the isoprene model are based, wherever possible, on the underlying mechanisms of isoprene synthesis, and use the findings of the most recent research into isoprene synthesis and flux (Monson et al. 1991b; 1992; 1994; Sharkey et al., 1991b; Kuzma and Fall, 1993; Loreto and Sharkey, 1990; 1993a; Silver and Fall, 1991; 1995; Baldocchi et al., 1995; Sharkey and Singsaas, 1995; Sharkey, 1996). The model is based on the premise that the rate of isoprene emission from a leaf is determined by the availability of carbon, the availability of ATP and the temperature dependency of the activity of isoprene synthase rate reaction. Thus, the model equations (Table 2.1, p.42) were used to determine which of the three rate-determining processes underlying isoprene synthesis are the most limiting under any given environmental conditions. This method of representing the rate of isoprene synthesis by the minimum rate of the processes that limit synthesis, is akin to the method of predicting rates of CO2 assimilation of Farquhar et al. (1980) and von Caemmerer and Farquhar (1981). In the photosynthesis model, the rate of assimilation is determined by the minimum of the rate when limited by i). the activation state and kinetics of Rubisco; ii). ribulose bisphosphate regeneration and iii). the availability of organic phosphate, which limits ribulose bisphosphate regeneration (Farquhar et al., 1980). However, the parts of the isoprene sub-model that are less mechanistic include the ATP limitation equation, and the temperature dependency equation taken from Thornley and Johnson (1980). Whereas the biochemical equations of Farquhar are well-validated, the lack of data on Fiso for validation, together with the lack of knowledge of parts of the pathway of isoprene synthesis limits the extent that this model can be considered to be as well-validated and biochemically mechanistic as the model of Farquhar et al. (1980). Meanwhile, the isoprene model has a higher mechanistic content than the isoprene emission model of Guenther et al. (1991; 1993), which has been used to estimate isoprene emissions from woodlands of the USA (Geron et al., 1994; 1995). Although the 1993 version of the empirical Guenther et al. model simulates the effects of temperature and PFD on isoprene emission rates better than the 1991 version, it cannot simulate the effects of [CO2]. The model of Guenther et al. (1993) simulates the response of isoprene emission rate to light by an equation similar to that used for the light dependency of photosynthesis, by Harley and Tenhunen (1991). In contrast, the sub-model of isoprene emission presented in this study, simulates the effects of light on the synthesis of isoprene emission, via the effect of light on CO2 assimilation rate, and availability of carbon for the skeletal carbon of isoprene in the form of pyruvate, thus using the already well validated model equations of Farquhar et al. (1980). Guenther et al. (1993) use their two equations for light and temperature to determine empirical coefficients to scale a mean rate of isoprene emission.

Although the new process-based isoprene emission model, presented here, mimics phenomenon well, the mechanism of isoprene production is still uncertain. Limitations include the exact source of carbon for the molecular skeleton of isoprene, and the unknown biochemical mechanism of ATP limitation of isoprene synthesis. Of the five sources of carbon for isoprene suggested, pyruvate formed directly from Rubisco catalysed reactions is considered to be the most likely, but has not been established as such (Sharkey et al. 1991). Also, although ATP limitation of isoprene synthesis is apparently linked to a reversal of the sensitivity of CO2 assimilation to [CO2] and [O2], how it is linked is uncertain (Sharkey et al., 1991a). Meanwhile the method of model construction is apparently robust, and predicts the variation in leaf level isoprene emission rates with most changing environmental factors well.

6.2.2 Canopy isoprene emission

The canopy isoprene emission model of Chapter Three (Equations B.1-B.29, Table B, Appendix I, p.197), predicted the proportionate change in isoprene emissions rates from forest canopies with the global warming effects of rising [CO2] and rising temperatures to be a greater than 25% decrease in emissions from the deciduous forests of the UK and USA, and a decrease in excess of 35% from the tropical rain forest in the Belém region of Brazil (Table 3.4, p.92). This unexpected effect of concurrent higher temperatures and elevated [CO2] is partly a result of the relatively high proportion of shaded to sunlit leaves within forest canopies, which, due to the lower quantum yield predicted for leaves under low light conditions, results in lower overall emission rates. The relatively large reduction predicted for the tropical rain forest canopy emissions, due to the larger leaf area index, is a particularly significant result, as, at present most isoprene is estimated to be emitted from tropical forests, and global warming is expected to increase the amount of isoprene emitted by tropical rain forest canopies (Müller, 1992).

Although the limitations of the leaf isoprene emission model also apply to the canopy model, the limitations of greatest significance to the canopy model are the poor understanding of inter-specific variation, the species parameter is entirely empirical, and the lack of knowledge of intra-specific variation. Also long-term predictions could be altered by species and population change. Other estimates of forest emission rates from forests of the USA have included species dependent emissions, and species composition data of forested areas (Geron, et al., 1994;1995). However, the species dependent emission rates have been derived empirically, and the leaf emission model equations are also empirical (Guenther et al., 1991; 1993). An immediate improvement to the methods of estimating emission rates from American woodlands, would be to replace the empirical equations of Guenther et al. (1991; 1993) with the process-based model equations presented here. This would also allow predictions of future emissions under varying conditions of atmospheric and climate change, within the limitations imposed by the lack of knowledge of species specific rates, intra species variation and future changes in forest species composition. However, the underestimation of emission rates at low light levels (Figure 2.9, p.56) must also be borne in mind, which is of particular relevance to the prediction of emission rates from canopies having high LAI, such as the tropical rain forest.

6.2.3 Effects of acute ozone exposure

The interactive effects of [CO2] and acute O3 on wheat photosynthesis were predicted to be minimal, both under the acute ozone concentrations measured at Sibton on a day in June during the year 1992 (DOE, 1994), and under increased [O3] (Figures 4.13, p.132 and 4.14, p.133). The change in assimilation rates observed were almost entirely due to the change in [CO2]. However, it must be borne in mind that a limitation of this model is the lack of knowledge of recovery times between days of acute ozone exposure, which restricts the model to predicting one continuous period of acute ozone exposure at a time, that is, during one day. It must also be borne in mind that the duration of exposure to acute [O3] can last a few days at a time, with longer periods of lower background concentrations in between (PORG, 1993). Therefore recovery times need to be measured and a mechanism for recovery sought.

Meanwhile, the two important findings that arise from the acute ozone model are: 1. that the use of a threshold based on flux, rather than concentration, allows the effects of acute ozone exposure at different concentrations to be predicted (Figure 4.8, p.119), and 2. that the decline in stomatal conductance can be predicted from the effects of the ozone-induced decline in Vcmax on Ci (Figure 4.1, p.109). The first advantage of the model based on the effects above a threshold flux, FO3(0), which is associated with the maximum rate of delivery of ozone into the leaf that the protective mechanisms can cope with, are that it provides a mechanism upon which to build a process-based model of ozone damaging effects, and so overcomes the difficulty of relating ozone dose to response. Relating the magnitude of plant responses to ozone uptake often produces a curvilinear pattern for each ozone concentration, as seen for the data of Farage et al. (1991) in Figure 4.5, p.115 (Nouchi and Aoki, 1979). But this is overcome by basing the effect on a dose above a threshold flux (Figure 4.8, p.119). The advantage of the second finding, that the decline in stomatal conductance can be predicted from the effects of the ozone-induced decline in Vcmax on Ci, is that it simplifies the modelling of ozone effects, as it overcomes the complication of predicting the reduced uptake of ozone by ozone-induced stomatal closure, which is solved iteratively within the model (Figure 4.9, p.121). As a result, the only new input variables required to predict the effects of acute [O3] are the ambient ozone concentration, and control values of stomatal conductance.

The method of construction of the model to predict the effect of ozone on Vcmax is akin to the method used in the well established, phenomenological model of stomatal conductance, of Ball et al. (1987). The ozone model has a lower mechanistic content than the isoprene model presented in this thesis. This reflects the complexity of the suite of biochemical anti-oxidant defence mechanisms to ozone, which can result in a variety of different sequences of response, that can vary from species to species, population to population and can be affected by growth conditions (Reiling and Davison, 1992; Pearson et al., 1995; Wellburn and Wellburn, 1996). The ozone model, like the model of Ball et al. (1987), uses empirical coefficients to represent both the rate of change within a relationship, and the intercept, or threshold, to wit, g(1) and Kz, and g(0) and FO3(0), for the stomatal conductance and ozone models, respectively.

6.2.4 Effects of chronic ozone exposure

The interactive effects of elevated [CO2] and chronic O3 exposure on wheat productivity is such that the damaging effects of chronic ozone at levels measured at Sibton, during the growing season of 1992, are predicted to be counteracted by the effects of [CO2] elevated to 700 µmol mol-1 (Figure 5.9, p.163) (DOE, 1994). This, perhaps, unexpected result, is partly a result of using background concentrations that are not excessively high, but indicate the need to assess interactive results using mechanistic models. The finding that the acute ozone model can be adapted to fit data of chronic ozone exposure effects at the leaf level, by re-parameterising the model coefficients, is of great significance, as once again it supports the finding that ozone-induced stomatal closure is a result of the effects of the decline in Vcmax, via effects on Ci, as suggested by McKee et al. (1995), and provides a mechanism for predicting the effects of chronic ozone exposure. The over-riding limitation of this model is the small amount of data that it is based on, and emphasises the need for data from further experimental research for model validation.

Meanwhile, by scaling the model up to the canopy level and combining it with equations to predict the above ground biomass of wheat produced over a growing season (Table B, p.197, and Table 5.1, p.146). This model provides a means of predicting the interactive effects of elevated [CO2] and chronic O3 exposure on wheat productivity. Such predictions were not previously available, due to the lack of a process-based model, for example, the empirical model of Amthor et al. (1994). To simulate the relative inhibition of photosynthesis by ozone Amthor et al. (1994) use a coefficient to represent the uptake of ozone into leaves, which calculated by eddy correlation, per unit Rubisco in the canopy. However, the coefficient only relates to the instantaneous rate of ozone uptake, and does account for any damage to the photosynthetic machinery that might have resulted from previous (hours to days) exposure of the mesophyll to ozone. This omission of accumulated dose effects, which is accounted for in the model presented here, could account for the overprediction, by their model, of stomatal conductance in the afternoon.

6.3 MODEL PREDICTION

From the models presented the interacting effects of elevated [CO2] and rising temperature on rates of isoprene emission are predicted. Similarly, the interacting effects of elevated [CO2] and elevated tropospheric [O3] on rates of CO2 assimilation. An overview of the predicted effects, at the atmospheric level (Figure 6.1) are summarised as follows: Increasing atmospheric [CO2] leads to rising air temperatures, via radiative forcing (Watson et al., 1990), which increases leaf temperature, affecting the rates of isoprene emission, via increased isoprene synthase activity (Silver and Fall, 1995; Sharkey, 1996). However, the direction of change in isoprene emission rates is dependent upon the concurrent [CO2], (Figure 2.15, p.66). As isoprene is a precursor of tropospheric ozone, varying rates of isoprene emission would be expected to alter the potential for ozone formation (Trainer et al., 1987; Chameides et al., 1988).

Meanwhile, at the level of internal leaf dynamics, the predicted effects can be summarised as follows: Increasing atmospheric [CO2] causes an increase in Ci and the maximum capacity for Rubisco carboxylation, Vcmax, depending on the interactive effects of temperature (Long, 1991; Woodrow, 1994). Elevated [CO2] allows stomatal closure against water loss due to transpiration (Morison, 1985).

Due to a decrease in cooling by transpiration, stomatal closure increases leaf temperature, causing an increase in rates of isoprene emission (Sharkey, 1996). Thus, the potential for tropospheric ozone formation is increased (Trainer et al., 1987; Chameides et al., 1988). Ozone reduces Vcmax, also causing stomatal closure, leading to further increases in leaf temperature (Farage et al., 1991). The interactive effects of elevated [CO2] and [O3], as predicted by the models of this study, are that the stomatal closure caused by elevated [CO2] will reduce the harmful effects of ozone, by reducing the uptake of ozone into the leaf (Figure 4.1, p.109). Evidence to support this prediction can be found in the recent research by McKee et al. (1995) and Fangmeier et al. (1996). However, a more complete model validation will need to be carried out at a later date, as, at present, there are no comparable studies with which direct comparisons can be made.

 6.4 FUTURE WORK

As the model presented has one main linking factor missing, the highest priority of future work that comes from this study would be to correct this omission, namely, to incorporate an atmospheric chemistry model to determine the quantitative effect of changes in isoprene emission rates on tropospheric ozone formation. The change in ozone formation will not be linearly related to the change in rates of isoprene emission, as the formation of ozone depends on a complex suite of reactions (Tables 1.1, 1.2 and 1.3, p.12-14) (Chameides and Lodge, 1992). Atmospheric chemistry models would be needed to simulate the formation of ozone from known concentrations of ozone precursors, such as isoprene and nitrogen oxides, as well as topographic features of the underlying terrain and prevailing meteorological conditions, such as wind speed (Lefohn, 1992; Chameides et al., 1994). Such models have been developed over the last two decades or so, and can now include the most recent knowledge of the main photooxidation mechanisms of isoprene's reactions with O3, OH·, O(3P) and NO3, as well as the OH· reactions with the products of the reaction of isoprene with OH·, methacrolein and methyl vinyl ketone (Hough and Derwent, 1987; MacKenzie et al., 1991; Paulson and Seinfeld, 1992). The incorporation of such a model could be used to simulate the effects of changing rates of isoprene emission on ozone formation, and would answer the question of how much rates of ozone formation would be expected to change under atmospheric and climate change, including the effects from changing rates of isoprene emission.

It is also possible to identify another variable that is of particular relevance to both isoprene emission rates and the effects of ozone at the leaf level, via it's affects on stomatal conductance, and that is the effects of water stress. Fiso is affected by water stress, probably via the increase in leaf temperature induced by stomatal closure against water loss (Fang et al. 1996; Sharkey, 1996). Stomatal closure also decreases ozone uptake into the leaf (Fangmeier et al., 1994; Heath, 1994). However, elevated [CO2] is expected to relieve the effects of water stress, allowing stomatal closure, and thereby reducing water loss (Morison, 1985). As drought conditions are expected to occur with greater frequency in the regions of the UK that are already the driest areas, and which are also important agricultural areas: the Southeast of England, the interactions of elevated [CO2], ozone and water stress will need to be incorporated into models of atmospheric and climate change (DOE, 1996). The interactions between water stress, ozone exposure, elevated [CO2] and isoprene emission will also be of interest in forested regions, where an increase in drought conditions are also predicted (DOE, 1996).

6.5.1 Interactive effects of ozone, isoprene emission, elevated [CO2] and water stress

In the developed areas of the northern hemisphere, under increasing temperatures and frequency of drought conditions, in addition to gradually increasing [CO2], and increasing [O3], the hypothesis that the harmful effects of chronic ozone exposure on wheat productivity will be alleviated by elevated [CO2] and drought conditions, needs to be tested. The interactive effects of increasing temperatures and elevated [CO2] on isoprene emission are a reduction in forest emission, but an increase in leaf emission rates (Table 3.4, p.92 and Figure 2.17, p.68). Therefore, in urban areas, where trees occur in a more isolated pattern than those within forests, an increase in isoprene emission might be expected, enhanced by the stomatal closure induced by water stress. This increase in isoprene emission might be expected to enhance the production of ozone. Once a model that simulates the reactions of ozone production has been incorporated to link isoprene emission with ozone, the model could be used to test theoretically if the hypothesis that the interactions of elevated [CO2] and water stress would compensate for the effects of ozone.

6.5.2 Relative ozone sensitivity, and the values of Kz and FO3(0)

A second hypothesis to test, that arises from this study, is that the effects of ozone on other cultivars of wheat can be represented by the ozone model. If this hypothesis proves to be true, the next question to be answered would be: can the relative ozone sensitivity of wheat cultivars be related to values of Kz and FO3(0)? This could be tested by subjecting various wheat cultivars, of known relative sensitivity to ozone, to the experimental procedures of Farage et al. (1991), and using the results to determine their values of Kz and FO3(0). The findings of such a study could help to formulate hypotheses concerning the biochemical paths of ozone damage and protective mechanisms, by comparing changes in the protein synthesis and anti-oxidant contents of leaves of different cultivars of wheat exposed to ozone, with the values found for the coefficients Kz and FO3(0), and their relative sensitivity to ozone.

6.6 CONCLUSION

The overall aim of this study was to use process-based models to predict some of the interactive effects of atmospheric and climate change on vegetation response. The interactive effects of concurrently changing variables are needed to identify potential feedback effects on climate and atmosphere. By attaining the objectives of each chapter, the overall aim has been achieved, predicting that the effects of elevated [CO2] on isoprene emission rates from forest canopies would counteract the higher rates predicted to be caused by rising temperatures alone, and predicting the phytotoxic effects of ozone exposure on photosynthesis to be alleviated by elevated [CO2]. However, to predict the effects of changing rates of isoprene emission on tropospheric ozone formation, will require a model to simulate atmospheric chemical reactions. The reactions of ozone formation are complex, and the rate of ozone formation will be affected by the concentrations of nitrogen oxides, isoprene and other hydrocarbons, and meteorological conditions (Chameides and Lodge, 1992; Chameides et al. 1992). The next interactive variable that should be included within the model of atmospheric and climate change is that of water stress, which affects both isoprene emission and ozone uptake by leaves. Further experimental work on the effects of ozone on cultivars of wheat with varying sensitivity to ozone could help increase our understanding of the anti-oxidant protective mechanisms, for example, by evaluating the applicability of the model to other wheat cultivars, and by investigating a potential relationship between the values of Kz and FO3(0) for wheat cultivars and their known relative sensitivity to ozone.

[ Home ] [ Up ]

Send mail to humph@essex.ac.uk with questions or comments about this web site.
Copyright © 1998 WIMOVAC Ltd.
Last modified: August 19, 1997