SEARCH

SEARCH BY CITATION

Keywords:

  • Fagus sylvatica;
  • Quercus robur;
  • Castanea sativa;
  • elevated CO2;
  • stomatal conductance;
  • vapour pressure deficit;
  • water stress

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References

Stomatal conductance (gs) and photosynthetic rate (A) were measured in young beech (Fagus sylvatica), chestnut (Castanea sativa) and oak (Quercus robur) growing in ambient or CO2-enriched air. In oak, gs was consistently reduced in elevated CO2. However, in beech and chestnut, the stomata of trees growing in elevated CO2 failed to close normally in response to increased leaf-to-air vapour pressure deficit (LAVPD). Consequently, while gs was reduced in elevated CO2 on days with low LAVPD, on warm sunny days (with correspondingly high LAVPD) gs was unchanged or even slightly higher in elevated CO2. Furthermore, during drought, gs of beech and chestnut was unresponsive to [CO2], over a wide range of ambient LAVPD, whereas in oak gs was reduced by an average of 50% in elevated CO2. Stimulation of A by elevated CO2 in beech and chestnut was restricted to days with high irradiance, and was greatest in beech during drought. Hence, most of the additional carbon gain in elevated CO2 was made at the expense of water economy, at precisely those times (drought, high evaporative demand) when water conservation was most important. Such effects could have serious consequences for drought tolerance, growth and, ultimately, survival as atmospheric [CO2] increases.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References

Atmospheric [CO2] has been rising steadily over the past two centuries, largely due to the burning of fossil fuels and land-use changes such as deforestation (Houghton, Callander & Varney 1992). Even the most conservative estimates of future emissions suggest that [CO2] will continue to rise steadily through the next century; a ‘business as usual’ scenario could result in concentrations of over 800 p.p.m. by the year 2100 (Houghton et al. 1992). Many trees growing today are therefore likely to experience double the present ambient [CO2] within their natural life spans (Kerstiens et al. 1995). Such changes will have major impacts because of the central role of CO2 in plant metabolism—around 15% of the atmospheric carbon pool is exchanged annually by the photosynthesis, respiration and decomposition of terrestrial higher plants (Amthor 1995). It is expected that atmospheric CO2 enrichment will be generally beneficial to plant growth, both through increased carbon assimilation and improved water economy (resulting from reductions in gs). However, there are widespread differences between species in the response to elevated CO2, as well as numerous interactions with other environmental factors (Lee & Jarvis 1995). Therefore, as well as the direct effects of CO2 enrichment on plant physiology, it is necessary to consider interactions with climatic and environmental factors which may be altered as a result of the atmospheric ‘greenhouse effect’. In particular, water availability is likely to have a strong influence on growth, productivity and survival. For example, the annual growth of European beech within its natural range is primarily determined by soil moisture availability (Innes 1992a). Reduced summer rainfall and soil moisture availability is anticipated over much of Western Europe (including the UK) over the coming decades (Hadley Centre 1992); consequently, the effects of CO2 enrichment on plant water economy have great potential to influence the growth and distribution of species such as beech.

Although instantaneous water use efficiency is usually increased in elevated CO2, this does not necessarily lead to increased drought tolerance (Tschaplinski et al. 1995); indeed, actual water use per individual tree may increase if the stomatal response to CO2 is weak, particularly when CO2 enrichment causes increases in total leaf area (Eamus 1996). Such effects could have major consequences for tree water use, drought tolerance, growth and ultimately survival, especially given the possibility of increased frequency and severity of drought due to climate change (Beerling et al. 1996). In the case of beech, the health of trees has been declining over recent decades, both in continental Europe (Innes 1992b) and in southern Britain (Ling, Power & Ashmore 1993). One of the factors most strongly correlated with the decline of beech in the UK was the severity of drought experienced in 1976—and, probably, in subsequent years such as 1983–1984 (Ling et al. 1993). Peterken & Mountford (1996) found that following the 1976 drought, the growth rate of damaged beech trees was negligible for 12 years and never fully recovered; trees were still dying from drought-induced damage up to 15 years later. Consequently, even slight increases in water use or subtle changes in stomatal behaviour could affect the health and survival of European beech. For example, Linder et al. (1997) used two separate forest gap models to simulate vegetation composition across the German state of Brandenburg. They used three climate scenarios: (1) current climate; (2) annual mean temperature + 0·8 °C, annual precipitation + 8%; and (3) annual mean temperature + 1·4 °C, annual precipitation – 20%. In scenario 2 the relative abundance of beech was increased, but in scenario 3, with increased drought, both models predicted that the occurrence of beech would be ‘negligible’ while the more drought-tolerant species would become more abundant. The gradient of water availability (a result of increasing evapotranspiration from the maritime to subcontinental climates) is considered to be responsible for the distribution of beech in the Brandenburg region (Linder et al. 1997). Hence, irrespective of any changes in climate, the direct effects of CO2 enrichment on stomatal behaviour and evapotranspiration could become a major influence on the future distribution of beech in continental Europe.

Understanding the role of rising atmospheric [CO2] in the stomatal regulation of water use in trees is of great importance for two main reasons. Firstly, the direct effects on patterns of water use will affect the growth and survival of trees. Secondly, there will be important feedbacks on global climate brought about by any changes in evapotranspiration. At the level of the individual tree, the short-term control of transpiration rate during periods of high evaporative demand or drought is vital in preventing damage due to leaf water deficits (Liang, Zhang & Wong 1996) or xylem embolism (Cochard, Breda & Granier 1996; Lu et al. 1996). Over a longer time-scale, total seasonal water use (which is mainly dependent on stomatal behaviour and canopy leaf area) will determine the likelihood of exposure to drought, both for the individual and nearby trees, given any particular precipitation pattern. In addition to [CO2], stomata respond to a variety of environmental and endogenous factors, such as light quality and quantity, leaf-to-air vapour pressure deficit (LAVPD), temperature, air pollutants and leaf and soil water status. There may be complex interactions between [CO2] and any one or a number of these variables, which could influence the degree of stomatal response to elevated CO2, and therefore the water economy and drought tolerance of the plant. Any such interactions, and the mechanisms underlying them, must be better understood if accurate predictions of plant responses are to be made. As well as having direct consequences for tree growth and survival, the maintenance of favourable water relations is essential if forest ecosystems are to respond positively to atmospheric CO2 enrichment, thereby providing a terrestrial ‘sink’ for anthropogenic CO2 emissions. This in itself has major implications for the modelling of future climate. Forest ecosystems are of major importance in the global carbon cycle, because the large standing biomass and long life span of trees provides great potential for long-term carbon storage (Grace et al. 1995). Although the role of tropical forests may be predominant, it is possible that temperate forest ecosystems also contribute significantly to the so-called ‘missing sink’, responsible for the annual removal of an estimated 2·0–3·4 Gt of carbon from the atmosphere (Tans, Fung & Takashaki 1990; Jarvis 1995). In addition, changes in evapotranspiration at the ecosystem level will have direct feedbacks on the regional and global climate. Any suppression of transpiration could result in reduced relative humidity, cloud cover and rainfall, and increasing temperature (Lockwood 1995). Sellers et al. (1996) concluded, with the use of a coupled biosphere–atmosphere model, that the reduction in evapotranspiration accompanying a doubling of the present ambient [CO2] would result in substantial amplification of temperature increase resulting from the atmospheric ‘greenhouse effect’. Thus, in order to model global environmental change, it is important to be able to predict accurately the magnitude of changes in evapotranspiration across a variety of ecosystems, and to be able to attach a realistic time-scale to these predictions.

Crops and herbaceous species are generally believed to display large reductions in gs in elevated CO2, but many experiments have been of short duration and there are some uncertainties about the capacity for acclimation which may moderate the response (Morison 1998). Furthermore, in trees the response appears to be generally smaller and more variable. Field, Jackson & Mooney (1995) found that elevated CO2 caused an average reduction in stomatal conductance of only 23% in a survey of 23 tree species, compared with an average of 40% in herbaceous species (Morison 1987). In a survey of 83 experiments involving 41 woody species, Curtis (1996) concluded that there was no overall significant reduction in gs in elevated CO2; however, the response was very variable, and reductions in gs tended to be greatest for unstressed plants in experiments lasting over 100 d. A more recent analysis showed no overall significant change in tree gs in elevated CO2; responses ranged from strongly negative to strongly positive with no consistent modification by other cultural or stress factors (Curtis & Wang 1998). In a review of the literature from 1993 to 1997, Saxe, Ellsworth & Heath (1998) found that elevated CO2 caused an average reduction in gs of just 18% in deciduous broadleaves and 13% in conifers. The relative lack of stomatal response in many species may be partly due to the emergence of data from longer-term experiments involving larger trees rooted directly in the ground (Eamus 1996), thus representing a more realistic scenario. Morison (1998) suggests that the relative insensitivity of many tree species is likely to be a result primarily of growing conditions, and points to the lack of side-by-side comparative studies. However, the present study offers such a side-by-side comparison, and highlights fundamental differences in the stomatal response to CO2, both between species and according to daily environmental conditions.

A more careful consideration of the interactions between [CO2] and other environmental, endogenous and stress factors could drastically alter our assumptions about stomatal behaviour and tree/forest water use in a future, CO2-enriched atmosphere. A more detailed knowledge of such effects would certainly improve the efficacy of regional-scale models of vegetation responses to environmental change (Berryman, Eamus & Duff 1994). Furthermore, an understanding of the physiological basis for these responses would allow the development of mechanistic models, enabling a more reliable extrapolation to future environmental scenarios (Bunce 1997). Interspecific variations in the response may determine which species (or genotypes) are best suited to the predicted future climate, and therefore represent the best choices for forestry and conservation purposes. The species involved in the present study are of great importance in the UK and throughout much of Europe. Indeed, oak and beech represent the two most widespread broadleaved species in the UK, where they are particularly important in terms of landscape and recreation, as well as providing valuable sources of high quality timber (Evans 1984).

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References

Experimental conditions

Investigations into the responses of young deciduous trees to elevated CO2 were conducted using the Solardome facilities at Lancaster University. The Solardomes—large, hemispherical greenhouses; see Townend (1993) for details—were well ventilated (2·5–3 air changes per minute), either with ambient air or with air enriched by 250 p.p.m. CO2. Variations in ambient [CO2] were therefore tracked by the elevated CO2 treatment. Use of the Solardomes allowed for a close control of water supply in order to study drought responses; at the same time, the trees were subject to near-natural variations in ambient light, temperature and humidity.

Young trees (2-year-old seedlings) of beech (Fagus sylvatica), chestnut (Castanea sativa) and oak (Quercus robur) were transplanted into 30-cm-deep, 4 L pots and placed in the Solardomes in May 1997. The potting medium was a 4:1 v/v mix of Sphagnum peat/granite grit with slow release fertiliser (Osmocote plus, 8–9 month formulation, 16% N, 8% P2O5, 12% K2O, 2% MgO + trace elements) added at the rate of 1 g L–1. The medium was judged to provide an adequate but not excessive supply of nutrients and resulted in soil pH values within the range experienced in natural beech and oak woodlands in the UK (see Heath & Kerstiens 1997 for details). The trees developed a second flush of leaves during June; these leaves (which had developed entirely within the respective CO2 treatments) were used for subsequent gas exchange measurements. The trees were kept well watered except for a short period of imposed drought in August (see below).

Gas exchange measurements

Over a 10 d period in July 1997, measurements of stomatal conductance (gs) and carbon assimilation rate (A) were taken on 6 d of contrasting ambient light, temperature and humidity. Plants were watered to saturation on the evenings prior to taking gas exchange measurements, in order to prevent water stress during measurements. Water was withheld from the same plants for an 8 d period in August 1997, during which changes in A and gs were followed closely. All measurements were taken using a portable IRGA (LCA-3, ADC Ltd, Hoddesdon, Hertfordshire, UK). There were two Solardomes each receiving ambient or CO2-enriched air, with a total of nine plants per species per Solardome. On each day, there was one measurement per plant, made between 1000 h and 1600 h. The measurements were divided between three visits to each Solardome in order to minimise chamber effects caused by diurnal variations in physiological or environmental factors. Values of photosynthetic photon fluence density (PPFD), relative humidity, air and leaf temperature were recorded. This enabled the calculation of values of LAVPD corresponding to every individual measurement of stomatal conductance.

Statistical analysis

Gas exchange measurements taken while the plants were well watered were analysed separately for each species and each day. Analysis of variance was carried out using a split-plot design (Solardome nested within CO2, n = 2) in order to take account of potential chamber effects. These data were also used to investigate the stomatal response to LAVPD: regression analysis was carried out for each species and CO2 treatment, with gs as the dependent variable and LAVPD as the independent variable. Analysis of covariance was also carried out for each species, with gs as the independent variable, CO2 as the treatment and LAVPD as the covariate. This was used to evaluate the significance of any differences between the slope of gs versus LAVPD according to CO2 treatment (a significant treatment–covariate interaction indicates that the slopes are significantly different between the two treatments). In the case of the drought treatment, a repeated measures analysis of variance (SYSTAT for Windows, version 5·03) was carried out on values of gs and A, again with a split-plot design to take into account chamber effects. The repeated measures analysis gives an indication of how treatment effects change over time, in the form of treatment–time interactions; it also takes into account the fixed variation between individuals used repeatedly for the same measurements.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References

Photosynthesis and stomatal conductance of well-watered trees

Net photosynthetic rate was substantially increased under elevated CO2 in all three species, but only on days of high incident PPFD (Fig. 1). Only on one of the cloudy days was there a significant stimulation of A by elevated CO2 (+ 29% in oak). The data in Fig. 1 can be divided distinctly into days of high and low PPFD; average stimulation of A by elevated CO2 for each was as follows:

image

Figure 1. . Photosynthetic rate (A) of well-watered beech, chestnut and oak on 6 d in July 1997. Open bars: ambient air; solid bars: elevated CO2; error bars represent 1 SE; each point represents nine plants in each of two Solardomes. Significance levels: *0·050 ≥P > 0·010; **0·010 ≥P > 0·001; ***0·001 ≥P. Measurements were taken in situ in the Solardomes (i.e. at the CO2 concentration in which the trees were growing) on days of contrasting ambient irradiance, temperature and humidity. However, environmental conditions during each day's measurements were relatively constant. The upper section of the figure shows average values (± 1 SE) of photosynthetic photon fluence density (PPFD, dashed line) and leaf-to-air vapour pressure deficit (LAVPD, solid line) for each day.

Download figure to PowerPoint

Low PPFD:

beech: + 11%; chestnut: + 11%; oak: + 20%;

High PPFD:

beech: + 82%; chestnut: + 100%; oak: + 41%.

Thus, in terms of carbon assimilation, the benefits of growth in CO2-enriched air were restricted to days of high incident sunlight and ambient temperature. This was especially true of beech and chestnut; in oak, although the stimulation of A by elevated CO2 was smaller overall, there was less variation from day to day. For all three species, increased A was associated with increased stem basal cross-sectional area; in beech and oak, stem height was also increased (data not shown).

The effects of CO2 enrichment on stomatal conductance were also strongly modified according to day-to-day variations in light, temperature and humidity. Again, this was most especially true of beech and chestnut. Oak generally showed large and significant reductions in gs in elevated CO2 (Fig. 2). However, the response was always weaker on sunny days, when PPFD and LAVPD were highest; on day 6, the 14% reduction in the gs of oak was not significant (Fig. 2). In beech and chestnut, the contrast between sunny and cloudy days was much more extreme. On cloudy days, there were large and significant reductions in gs in elevated CO2, just as for oak. However, on sunny days, there were never any significant reductions in gs in elevated CO2. In fact, on sunny days, the gs of chestnut was on average higher in elevated CO2 than in ambient air (although not significantly so); in beech, gs was equal in both CO2 treatments on sunny days.

image

Figure 2. . Stomatal conductance (gs) of well-watered beech, chestnut and oak on 6 d in July 1997. Open bars: ambient air; solid bars: elevated CO2; error bars represent 1 SE; each point represents nine plants in each of two Solardomes. Significance levels: *0·050 ≥P > 0·010; **0·010 ≥P > 0·001; ***0·001 ≥P. Measurements were taken in situ in the Solardomes (i.e. at the CO2 concentration in which the trees were growing) on days of contrasting ambient irradiance, temperature and humidity. However, environmental conditions during each day's measurements were relatively constant. The upper section of the figure shows average values (± 1 SE) of photosynthetic photon fluence density (PPFD, dashed line) and leaf-to-air vapour pressure deficit (LAVPD, solid line) for each day.

Download figure to PowerPoint

Interactions between CO2 and LAVPD

In attempting to identify the causes of the day-to-day variations in the stomatal responses to CO2, it was impossible to separate the effects of PPFD, temperature and humidity, since they all varied concurrently in the Solardomes. However, rather than increasing with incident PPFD, as is commonly observed, gs was, on average, significantly lower on sunny days than on cloudy days (Fig. 2). This suggested that the most important factor controlling day-to-day variations in gs might be leaf-to-air vapour pressure deficit, which was always very much higher on days of high PPFD, and so would tend to cause stomatal closure. A closer examination of the complete data set confirmed that, for all three species, gs did indeed decrease steadily as LAVPD increased. However, the most important finding was that the relationship between LAVPD and gs was drastically altered by [CO2], particularly in beech and chestnut. At ambient [CO2], gs was clearly reduced with increasing LAVPD in all three species (Fig. 3). The response was strongest in oak, followed by chestnut, then beech (this is reflected by the slopes of the regressions given with Fig. 3). However, at elevated CO2 the response of gs to changes in LAVPD was very much weaker (in fact, in beech it had disappeared altogether), hence the much lower values of R2 and slope in the regression analysis (Fig. 3). In elevated CO2, only oak retained a sufficiently strong sensitivity to changes in LAVPD to ensure that gs was always, on average, lower than in ambient air across the whole range of LAVPD to which it was exposed (Fig. 3). For beech and chestnut in elevated CO2, the relatively small degree of stomatal closure with increasing LAVPD meant that gs tended to become equal or higher than in ambient air through the upper range of LAVPD (Fig. 3). This explains why there was apparently no stomatal response to [CO2] on hot, dry, sunny days (Fig. 2).

image

Figure 3. . Plots of stomatal conductance (gs) against leaf-to-air vapour pressure deficit (LAVPD) in beech, chestnut and oak at ambient and elevated CO2 concentrations. The data are taken from Fig. 2 (i.e. six separate measurements on each of 18 plants per species/CO2 treatment). At ambient CO2 concentration, all three species showed clear reductions in gs with increasing LAVPD. However, at elevated CO2, stomatal sensitivity to changes in LAVPD was much reduced. Analysis of covariance confirmed that the slope of the relationship between gs and LAVPD was significantly altered by growth at elevated CO2 in all three species (P = 0·001). Note that, for beech and chestnut at elevated CO2, when LAVPD was high, gs was on average equal or higher in elevated CO2 compared to ambient air. In contrast, oak maintained a significantly lower gs in elevated CO2 at all times. Regression analysis gave the following values of R2, slope (gs, dependent, vs LAVPD, independent), and P (significance of slope from zero):

Download figure to PowerPoint

The significance of the effect of [CO2] on the relationship between gs and LAVPD was confirmed using analysis of covariance, with [CO2] as the treatment and LAVPD as the covariate. A significant interaction between treatment and covariate indicates a significant difference between the slopes of covariate (LAVPD) versus dependent variable (gs) according to treatment ([CO2]). For all three species, this difference was highly significant (P < 0·001). Thus, without doubt, elevated CO2 drastically altered the stomatal response to LAVPD in all three species. The analysis of covariance gave F-ratios for CO2–LAVPD interactions of 44 (beech), 20 (chestnut) and 18 (oak), showing that the effect of [CO2] on the stomatal response to LAVPD was greatest in beech and least in oak.

LAVPD was always slightly higher at elevated CO2, as a result of reduced transpiration rates (if only from oak) in the Solardomes receiving CO2-enriched air. However, the effect was only statistically significant on two days, and was extremely small: the average increase was just 0·19 kPa, and the greatest increase only 0·25 kPa. This was judged to be of little significance compared with the absolute range of around 4 kPa, and insufficient to influence the observed patterns of gs significantly. In any case, an increase in LAVPD would be expected to encourage further stomatal closure, not less.

Photosynthesis and stomatal conductance during drought

Only beech was able to maintain an increased photosynthetic rate in elevated CO2 during drought, by an average of + 88% through the whole period (Fig. 4). In beech, after 8 d of drought, A was just 22% of its original value in ambient air, but had only declined to 47% of its original value in elevated CO2. In contrast, photosynthesis in chestnut and oak had declined almost to zero by the end of this period (Fig. 4). In chestnut, although there was initially a very large stimulation (+ 108%) of A by elevated CO2, this was lost after the first day; [CO2] had no significant effect whatsoever on the rate of photosynthesis in oak during drought (Fig. 4).

image

Figure 4. . Photosynthetic rate of beech, chestnut and oak over an 8 d period of drought in August 1997. Open circles: ambient air; closed circles: elevated CO2; error bars represent 1 SE; each point represents nine plants in each of two Solardomes. Significance levels: *0·050 ≥P > 0·010; **0·010 ≥P > 0·001; ***0·001 ≥P. Measurements were taken in situ in the Solardomes (i.e. at the CO2 concentration in which the trees were growing) on days of contrasting ambient irradiance, temperature and humidity. However, environmental conditions during each day's measurements were relatively constant. The upper section of the figure shows average values (± 1 SE) of photosynthetic photon fluence density (PPFD, dashed line) and leaf-to-air vapour pressure deficit (LAVPD, solid line) for each day.

Download figure to PowerPoint

Stomatal conductance declined rapidly during the drought in all three species (Fig. 5). The apparently more severe effect on chestnut was simply due to the larger size of the trees. Oak was the only species to show consistently lower values of gs in elevated CO2 throughout the drought (– 50% on average). [CO2] had no effect on the gs of chestnut throughout the whole drought period (Fig. 5). In beech, gs was initially 23% lower in elevated CO2; towards the end of the drought, there was a tendency for gs to remain higher in elevated CO2 than in ambient air, by an average of nearly 50% over the last 2 d. Hence, there was no overall significant effect of [CO2] on the gs of beech during drought, but there was a highly significant CO2–time interaction (Fig. 5).

image

Figure 5. . Stomatal conductance of beech, chestnut and oak over an 8-d period of drought in August 1997. Open circles: ambient air; closed circles: elevated CO2; error bars represent 1 SE; each point represents nine plants in each of two Solardomes. Significance levels: *0·050 ≥P > 0·010; **0·010 ≥P > 0·001; ***0·001 ≥P. Measurements were taken in situ in the Solardomes (i.e. at the CO2 concentration in which the trees were growing) on days of contrasting ambient irradiance, temperature and humidity. However, environmental conditions during each day's measurements were relatively constant. The upper section of the figure shows average values (± 1 SE) of photosynthetic photon fluence density (PPFD, dashed line) and leaf-to-air vapour pressure deficit (LAVPD, solid line) for each day.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References

Large increases in leaf photosynthetic rate were seen in all three species in elevated CO2 (Fig. 1), although only beech was able to maintain this increase during drought (Fig. 4). While increased A is not necessarily a good predictor of the incorporation of carbon into tissues (Lee & Jarvis 1995), stem height and basal area were substantially increased after one season in elevated CO2 (data not shown). The increases in A observed in beech, chestnut and oak were of a comparable magnitude to those previously observed in beech and oak (Heath & Kerstiens 1997), when substantially higher growth rates (both above- and below-ground) were maintained in elevated CO2 in trees grown from seed for 3 years in the Solardomes and the same soil medium. Therefore, at least for young beech, chestnut and oak trees, we can conclude that atmospheric CO2 enrichment will potentially result in much higher growth rates.

However, almost all of the increase in carbon assimilation was restricted to days with high incident PPFD, especially in beech and chestnut (Fig. 1). It was on these warm, sunny days that the stomata of beech and chestnut failed to close in response to elevated CO2, and even the response of oak was noticeably weaker (Fig. 2). Thus, for beech and chestnut, the cost of increased carbon gain was the loss of the potential for improved water economy in elevated CO2. This represents a very considerable cost, because the stomata were failing to close at precisely those times when a short-term control over transpirational water loss is most important in preventing the occurrence of harmfully low tissue water potentials and/or xylem embolism. Although gs was, on average, never significantly higher in elevated CO2, the trend could clearly result in such an increase. Indeed, it was found that elevated CO2 consistently increased gs of European beech growing in growth cabinets by an average of 20% (J. Dick, personal communication).

As well as stomatal behaviour during short periods of high evapotranspirational demand, the stomatal response to soil water deficits is of prime importance in determining drought tolerance and survival. The stimulation of A by elevated CO2 was even greater in droughted than in well-watered beech; again, this was at the expense of water conservation, since there was no reduction in gs in elevated CO2, either in beech or chestnut—in fact, elevated CO2 caused an increased gs in beech towards the end of the drought. In contrast, oak reduced gs by an average of 50% in elevated CO2 during drought (Fig. 5), representing an even greater improvement in water economy than when the plants were well watered (Fig. 2). Although the effects of withholding water were made artificially rapid and severe by the relatively small soil volumes, there can be no doubt about the contrast in stomatal behaviour between oak and beech/chestnut. Over a 4-week period of drought during the hot, dry summer of 1995, young beech trees (growing in the Solardomes in the same soil medium but with twice the soil volume per plant) actually maintained a significantly higher gs in elevated CO2 than in ambient air (Heath & Kerstiens 1997). This stomatal response alone, even without any increase in total leaf area, was sufficient to cause substantially increased rates of soil drying in elevated CO2 (Heath & Kerstiens 1997). These findings raise the possibility that CO2 enrichment could result in more frequent exposure to harmful soil water deficits, regardless of any future changes in seasonal precipitation/evaporative demand. The increase in leaf area that commonly occurs in elevated CO2 would exacerbate such effects (Eamus 1996). Indeed, given a higher nutrient supply, increased leaf area was the main cause of increased soil moisture depletion by beech in elevated CO2 during drought (Heath & Kerstiens 1997).

There is clearly a difference between the responses of beech and chestnut on the one hand, and oak on the other, to CO2 enrichment and drought. Roden & Ball (1996) found a similar, although less extreme, difference between Eucalyptus rossii and E. macrorhyncha. Both showed large reductions in gs in elevated CO2 when well watered, but during drought the stomatal response of E. rossii to [CO2] was much weaker. In combination with a strong stimulation of growth rate by CO2 enrichment, this resulted in more rapid rates of soil moisture depletion in elevated CO2 than in ambient air. In contrast, rates of soil drying were unaffected by [CO2] in E. macrorhyncha, since the stomatal response was stronger and the growth response somewhat weaker than that of E. rossii. An understanding of the physiological basis for such interspecific differences in stomatal behaviour and growth response would allow improved predictions of future tree drought tolerance, forest growth, and feedbacks on global climate. Certainly, these interspecific differences in their ability to benefit from CO2 enrichment through improved water economy may have direct implications for the survival of seedlings and patterns of forest regeneration; but will mature trees behave in a similar manner? It has been shown that branches of mature beech growing in an open position failed to reduce their transpiration rate per unit leaf area over three seasons in elevated CO2 (Dufrene, Pontailler & Saugier 1993; Pontailler et al. 1994). This was in spite of a 30% reduction in stomatal density in elevated CO2 in the third growing season (Pontailler et al. 1994), suggesting that stomatal apertures must have been increased in elevated CO2. Therefore, it is certainly possible that the stomata of mature beech trees were behaving in a similar manner to that reported here. It should be noted that all the leaves in the present study would have been sun-adapted, and that shade-adaptation would almost certainly have resulted in different patterns of photosynthetic and stomatal response to [CO2] and light. For example, Kubiske & Pregitzer (1997) found that gs of Betula papyrifera (measured in full sunlight) was significantly increased in elevated CO2 in sun-adapted plants, but in shade-adapted plants the increase was smaller and not significant. There was a very similar pattern for sun- and shade-adapted plants of Acer rubrum, except that the CO2 effect was not significant in either case (Kubiske & Pregitzer 1997). However, in a forest canopy, it is the response of those leaves exposed to long periods of direct sunlight that will be most important in protecting the trees against excessive and potentially harmful rates of water loss during critical periods of high evapotranspirational demand. For the same reason, trees growing individually or in open or exposed positions will be generally more at risk from drought damage. This certainly appears to be the case for beech, for example in the drought of Summer 1995 in the UK (personal observation), and may account for the generally poorer health of beech trees growing in open or disturbed sites (Ling et al. 1993; Stribley 1993).

What are the possible mechanisms that could account for the lack of stomatal closure of beech and chestnut in elevated CO2 during periods of drought and high evapotranspirational demand? Stomatal responsiveness to LAVPD was much reduced at elevated CO2, particularly in beech and chestnut (Fig. 3), accounting for the apparent loss of CO2 response on hot, sunny days (Fig. 2). Can we be sure, though, that CO2 enrichment is actually modifying the stomatal response to changes in LAVPD, or is the effect only apparent? It is unlikely that changes in Ci (intercellular CO2 concentration) could explain the day-to-day variations in stomatal response, since the relative increase in Ci at elevated CO2 was no greater on cloudy days than on sunny days. It was shown previously that beech and oak growing in the Solardomes at elevated CO2 did not alter their stomatal density (Heath & Kerstiens 1997).

A more serious consideration is that other environmental variables (most importantly light and temperature) vary alongside LAVPD in the Solardomes. While this represents a more realistic scenario, it means that separating the effects of either variable is not easy. It was concluded that LAVPD was more important than light in controlling daily variations in gs because on hot, sunny days, at least in ambient air, gs decreased rather than increased. Thus, of the opposing stimuli (high PPFD causing stomatal opening and high LAVPD causing stomatal closure), increased LAVPD was clearly the overriding influence. The hypothesis that LAVPD, and not PPFD, was the most important factor influencing day-to-day variations in the stomatal response to CO2 is supported by other studies. The stomatal sensitivity of Pinus radiata and Pseudotsuga menziesi to LAVPD was decreased in elevated CO2, with a constant temperature of 20 ± 1 °C and PPFD of 700–800 μmol m–2 s–1 (Hollinger 1987). Will & Teskey (1997) found that the absolute decrease in gs due to growth at elevated CO2 was reduced at LAVPD of 2·5 compared to 1·0 kPa in four tree species; in Cercis canadensis the relative decrease in gs was also reduced at high LAVPD, with a leaf temperature of 26 °C and PPFD of 1590 μmol m–2 s–1. Preliminary results from the Duke Forest FACE experiment (North Carolina, USA) also suggest that high LAVPD (at constant leaf temperature and saturating PPFD) may reduce the magnitude of the stomatal response to [CO2] in three hardwood species (Heath, Ellsworth & Naumburg, unpublished results). All these findings are consistent with a direct role of high LAVPD, rather than PPFD, in reducing the stomatal response to [CO2] on hot, sunny days in the Solardomes. In contrast, high PPFD actually increased stomatal responsiveness to [CO2] (Will & Teskey 1997).

The most important factor that cannot be accounted for satisfactorily is temperature, since changes in LAVPD closely followed changes in temperature. It is possible that temperature may have a direct effect on the response to [CO2], independent of changes in humidity. For example, Berryman et al. (1994) found that elevated CO2 reduced the stomatal sensitivity to temperature, but not to LAVPD, in Eucalyptus tetrodonta. However, in that study, trees were removed from outdoor growth chambers to the laboratory 1 d prior to measurements, whereas in the Solardomes gs was measured in situ with minimal disturbance of the microclimate surrounding the leaf, representing changing stomatal responses with naturally occurring variations in light, temperature and humidity. In Bellis perennis and Cardamine pratensis, elevated CO2 inhibited stomatal opening at higher temperatures but actually stimulated opening at lower temperatures (Honour 1994). Thus, the temperature-dependency of the CO2 response was the opposite of that found by Berryman et al. (1994).

Without further work, the exact mechanisms underlying the interactions between CO2, LAVPD and drought cannot be reliably identified. However, given the relative lack of stomatal closure in response both to soil drought and to high LAVPD, it would seem likely that growth at elevated CO2 is directly affecting the stomatal regulation of leaf water relations in beech and chestnut. The common factor may well be the involvement of the hormone ABA in the stomatal responses both to soil drought and to LAVPD. There is now considerable evidence that the stomatal response to LAVPD is mediated by changes in transpiration, which alter the rate of delivery of ABA to stomatal guard cells (Bunce 1997, 1998). It is well established that during water stress, the delivery of ABA from the roots via the xylem acts as a signal for stomatal closure even before any reduction in leaf water potential (e.g. in Acacia confusa and Litsea glutinosa; Liang et al. 1996). Could growth at elevated CO2 affect the functioning of the ABA signal during periods of drought and high LAVPD? It has been shown that elevated CO2 did not reduce the stomatal sensitivity to exogenous ABA in excised shoots of beech (Heath & Kerstiens 1997) or Maranthes corymbosa (Berryman et al. 1994). Using the same method, the relationship between exogenous [ABA] and gs in excised shoots was found to be the same as that between xylem [ABA] and gs in intact plants (Liang et al. 1996). However, the concentration of ABA in the xylem sap cannot always explain the observed stomatal response to drought (Thompson et al. 1997), and modifications to apoplastic concentrations of ABA occur upstream of the xylem (Wilkinson & Davies 1997). It is therefore possible that elevated CO2 might affect the processing of the ABA signal upstream of the xylem, or possibly even signal transduction in the guard cell itself. It is now certain that the signal transduction pathways for ABA and for CO2 share common elements (Webb & Hetherington 1997) and that both involve an increase in cytosolic free calcium ions in the guard cell (Webb et al. 1996; McAinsh, Brownlee & Hetherington 1997). In an ABA-insensitive poplar hybrid, stomatal sensitivity to soil drought declined with leaf age, and this age-dependency was mirrored by the stomatal sensitivity to [CO2] and, in detached leaves, to exogenous calcium ions (Ridolfi et al. 1996). CO2 enrichment resulted in reduced stomatal sensitivity to calcium ions in Maranthes corymbosa, even though sensitivity to leaf water status was increased (Berryman et al. 1994). Therefore, although highly speculative, it is possible that growth at elevated CO2 causes reduced stomatal sensitivity to soil drought and to LAVPD through the common role of calcium ions in the signal transduction pathways for ABA and for CO2.

Consistent with this hypothesis was the demonstration that, for beech grown in elevated CO2, the relative lack of stomatal closure during drought was not due to altered growth patterns which might have improved plant water relations, such as increased fine root-to-leaf area ratio (Heath & Kerstiens 1997). The plants also did not show any increase in nutrient allocation to the roots (J. Heath, unpublished results), which might have increased fine root activity, thus allowing a more efficient uptake of water from the soil (Jackson & Reynolds 1996). An increase in stem hydraulic conductance could also potentially improve the supply of water to the leaves, allowing the maintenance of higher stomatal conductance (Kubiske & Pregitzer 1997); however, the stem hydraulic conductance of beech was unchanged by [CO2] (Heath, Kerstiens & Tyree 1997). The lack of stomatal responsiveness to [CO2] during drought (Fig. 5) was not simply a result of high LAVPD brought about by reduced transpiration rates (this might be expected to create the same effect as seen on hot, sunny days when the plants were well watered). On days with low LAVPD during drought, there was still no stomatal response to CO2 in beech or chestnut (Fig. 5) or in the previous experiment with beech (Heath & Kerstiens 1997). On such days when the plants were well watered, there would have been a highly significant stomatal closure in elevated CO2 (Fig. 2). Thus, it would appear that the effects of soil drought and dry air were additive in diminishing the stomatal response to [CO2]. Therefore, the generally higher values of LAVPD experienced in the previous drought experiment with beech (J. Heath, unpublished results) could explain why, on that occasion, gs was significantly increased in elevated CO2 (Heath & Kerstiens 1997) rather than simply becoming equal to that in ambient air (Fig. 5).

The available evidence strongly suggests a direct effect of CO2 enrichment on the stomatal regulation of water loss during periods of high evaporative demand and drought in certain tree species. A possible cause may be the shared elements of the signal transduction pathway for ABA and for CO2, and in particular, the role of calcium ions may be worth further investigation. An understanding of the mechanistic basis for these responses, and of their interspecific variations, would enable the development of more reliable predictive models for forest growth and productivity, species distribution, and regional and global climate. Whatever the mechanisms, it would appear that in certain species, such as beech and chestnut, the benefits of atmospheric CO2 enrichment will be far less than expected, both in terms of seasonal water use and the short-term control of transpiration rate. Indeed, in extreme cases (very high LAVPD and soil drought) transpiration, and rates of soil moisture depletion, may be increased in elevated CO2 (Heath & Kerstiens 1997). In contrast, oak showed consistently improved water economy in elevated CO2, while maintaining increased rates of photosynthesis and growth. The stomata of beech and chestnut only became totally unresponsive to [CO2] on days with very high values of LAVPD (Fig. 2). Although the high values of LAVPD experienced in the Solardomes may not be typical of the present climate of northern Britain (due to some elevation of air temperature), the natural range of beech does not extend north of southeast England; it is here, and in continental Europe, that beech is a far more important species in terms of forestry, recreation and landscape. Furthermore, it is precisely during more extreme conditions (which may occur only relatively infrequently) that widespread and lasting damage is most likely due to excessive rates of transpiration. In addition, the frequency with which such conditions (as well as soil drought) are experienced may increase in many regions with climate change over the coming decades. Consequently, atmospheric CO2 enrichment may have adverse effects on the water economy and drought tolerance of certain tree species, resulting in changes in forest productivity and species distribution.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References

This work was funded by the UK Department of the Environment under the research programme entitled ‘Carbon Sequestration in Vegetation and Soils’; additional funding and assistance was previously provided by the Forestry Commission. Thanks also for the help and advice of T.A. Mansfield, G. Kerstiens and P.H. Freer-Smith.

References

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References
  • 1
    Amthor J.S. (1995) Terrestrial higher-plant response to increasing atmospheric [CO2] in relation to the global carbon cycle. Global Change Biology 1, 243274.
  • 2
    Beerling D.J., Heath J., Woodward F.I., Mansfield T.A. (1996) Drought–CO2 interactions in trees: observations and mechanisms. New Phytologist 134, 235242.
  • 3
    Berryman C.A., Eamus D., Duff G.A. (1994) Stomatal responses to a range of variables in two tropical tree species grown with CO2 enrichment. Journal of Experimental Botany 45, 539546.
  • 4
    Bunce J.A. (1997) Does transpiration control stomatal responses to water vapour pressure deficit? Plant, Cell and Environment 20, 131135.
  • 5
    Bunce J.A. (1998) Effects of humidity on short-term responses of stomatal conductance to an increase in carbon dioxide concentration. Plant, Cell and Environment 21, 115120.
  • 6
    Cochard H., Breda N., Granier A. (1996) Whole tree hydraulic conductance and water loss regulation in Quercus during drought—evidence for stomatal control of embolism. Annales Des Sciences Forestieres 53, 197206.
  • 7
    Curtis P.S. (1996) A meta-analysis of leaf gas exchange and nitrogen in trees grown under elevated carbon dioxide. Plant, Cell and Environment 19, 127137.
  • 8
    Curtis P.S. & Wang X. (1998) A meta-analysis of elevated CO2 effects on woody plant mass, form and physiology. Oecologia 113, 299313.
  • 9
    Dufrene E., Pontailler J.Y., Saugier B. (1993) A branch bag technique for simultaneous CO2 enrichment and assimilation measurements on beech (Fagus sylvatica L.). Plant, Cell and Environment 15, 11311138.
  • 10
    Eamus D. (1996) Responses of field grown trees to CO2 enrichment. Commonwealth Forestry Review 75, 3947.
  • 11
    Evans J. (1984) Silviculture of broadleaved woodland. Forestry Commission Bulletin62. HMSO, London.
  • 12
    Field C.B., Jackson R.B., Mooney H.A. (1995) Stomatal responses to increased CO2: implications from the plant to the global scale. Plant, Cell and Environment 18, 12141225.
  • 13
    Grace J., Lloyd J., McIntyre J., Miranda A.C., Meir P., Miranda H.S., Nobre C., Moncrieff J., Massheder J., Malhi Y., Wright I., Gash J. (1995) Carbon dioxide uptake by an undisturbed tropical rain-forest in southwest Amazonia, 1992–93. Science 270, 778780.
  • 14
    Hadley Centre (1992) The Hadley Centre Transient Climate Change Experiment. Meteorological Office, Bracknell.
  • 15
    Heath J. & Kerstiens G. (1997) Effects of elevated CO2 on leaf gas exchange in beech and oak at two levels of nutrient supply: consequences for sensitivity to drought in beech. Plant, Cell and Environment 20, 5767.
  • 16
    Heath J., Kerstiens G., Tyree M.T. (1997) Stem hydraulic conductance of European beech (Fagus sylvatica L.) and pedunculate oak (Quercus robur L.) grown in elevated CO2. Journal of Experimental Botany 48, 14871489.
  • 17
    Hollinger D.Y. (1987) Gas exchange and dry matter allocation responses to elevation of atmospheric CO2 concentration in seedlings of three tree species. Tree Physiology 3, 193202.
  • 18
    Honour S.J. (1994) Effects of CO2 enrichment, temperature, water deficit and abscisic acid on the growth and stomatal behaviour of three herbaceous species from different altitudes. PhD, Lancaster University, England.
  • 19
    Houghton J.T., Callander B.A., Varney S.K. (1992) Climate Change 1992. The supplementary report to the IPCC Scientific Assessment. Cambridge University Press, Cambridge.
  • 20
    Innes J.L. (1992a) Observations on the condition of beech (Fagus sylvatica L.) in Britain in 1990. Forestry 65, 3560.
  • 21
    Innes J.L. (1992b) Forest decline. Progress in Physical Geography 16, 164.
  • 22
    Jackson R.B. & Reynolds H.L. (1996) Nitrate and ammonium uptake for single- and mixed-species communities grown at elevated CO2. Oecologia 105, 7480.
  • 23
    Jarvis P.G. (1995) The role of temperate trees and forests in CO2 fixation. Vegetatio 121, 157174.
  • 24
    Kerstiens G., Townend J., Heath J., Mansfield T.A. (1995) Effects of water and nutrient availability on physiological responses of woody species to elevated CO2. Forestry 68, 303315.
  • 25
    Kubiske M.E. & Pregitzer K.S. (1997) Ecophysiological responses to simulated canopy gaps of two tree species of contrasting shade tolerance in elevated CO2. Functional Ecology 11, 2432.
  • 26
    Lee H.S.J. & Jarvis P.G. (1995) Trees differ from crops and from each other in their responses to increases in CO2 concentration. Journal of Biogeography 22, 323330.
  • 27
    Liang J., Zhang J., Wong M.H. (1996) Stomatal conductance in relation to xylem sap abscisic acid concentrations in two tropical trees, Acacia confusa and Litsea glutinosa. Plant, Cell and Environment 19, 93100.
  • 28
    Linder M., Bugmann H., Lasch P., Flechsig M., Cramer W. (1997) Regional impacts of climatic change on forests in the state of Brandenbug, Germany. Agricultural and Forest Meteorology 84, 123135.
  • 29
    Ling K.A., Power S.A., Ashmore M.R. (1993) A survey of the health of Fagus sylvatica in southern Britain. Journal of Applied Ecology 30, 295306.
  • 30
    Lockwood J.G. (1995) The suppression of evapotranspiration by rising levels of atmospheric CO2. Weather 50, 304308.
  • 31
    Lu P., Biron P., Granier A., Cochard H. (1996) Water relations of adult Norway spruce (Picea abies (L.) Karst) under soil drought in the Vosges mountains: whole-tree hydraulic conductance, xylem embolism and water loss regulation. Annales Des Sciences Forestieres 53, 113121.
  • 32
    McAinsh M.R., Brownlee C., Hetherington A.M. (1997) Calcium ions as second messengers in guard cell signal transduction. Physiologia Plantarum 100, 1629.
  • 33
    Morison J.I.L. (1987) Intercellular CO2 concentration and stomatal responses to CO2. In Stomatal Function (eds E. Zeiger, G.D. Farquhar & I.R. Cowan), pp. 229–251. Stanford University Press, Stanford.
  • 34
    Morison J.I.L. (1998) Stomatal response to increased CO2 concentration. Journal of Experimental Botany 49, 443452.
  • 35
    Peterken G.F. & Mountford E.P. (1996) Effects of drought on beech in Lady Park Wood, an unmanaged mixed deciduous woodland. Forestry 69, 125136.
  • 36
    Pontailler J.Y., Dufrene E., Saugier B., Samain E. (1994) Long term CO2 enrichment and gas exchange measurement on branches of mature beeches (Fagus sylvatica L.). In Responses of Trees and Forests to Global Change. ECOCRAFT/ICAT International Symposium, Dourdan, France. Abstracts, p. 41.
  • 37
    Ridolfi M., Fauveau M.L., Label P., Garrec J.P., Dreyer E. (1996) Responses to water stress in an ABA-unresponsive hybrid poplar (Populus koreana× trichocarpa cv. Peace). I. Stomatal function. New Phytologist 134, 445454.
  • 38
    Roden J.S. & Ball M.C. (1996) The effect of elevated [CO2] on growth and photosynthesis of two Eucalyptus species exposed to high temperatures and water deficits. Plant Physiology 111, 909919.
  • 39
    Saxe H., Ellsworth D.S., Heath J. (1998) Tansley Review: Tree and forest functioning in an enriched CO2 atmosphere. New Phytologist (Tansley Review No. 98) 139, 395436.
  • 40
    Sellers P.J., Bounoua L., Collatz G.J., Randall D.A., Dazlich D.A., Los S.O., Berry J.A., Fung I., Tucker C.J., Field C.B., Jensen T.G. (1996) Comparison of radiative and physiological effects of doubled atmospheric CO2 on climate. Science 271, 14021406.
  • 41
    Stribley G.H. (1993) Studies on the health of beech trees in Surrey, England: relationship between winter canopy assessment by Roloff's method and twig analysis. Forestry 66, 1926.
  • 42
    Tans P.P., Fung I.Y., Takashaki T. (1990) Observational constraints on the global atmospheric CO2 budget. Science 247, 14311438.
  • 43
    Thompson D.S., Wilkinson S., Bacon M.A., Davies W.J. (1997) Multiple signals and mechanisms that regulate leaf growth and stomatal behaviour during water deficit. Physiologia Plantarum 100, 303313.
  • 44
    Townend J. (1993) Effects of elevated carbon dioxide and drought on the growth and physiology of clonal sitka spruce plants (Picea sitchensis (Bong.) Carr.). Tree Physiology 13, 389399.
  • 45
    Tschaplinski T.J., Stewart D.B., Hanson P.J., Norby R.J. (1995) Interactions between drought and elevated CO2 on growth and gas exchange of seedlings of three deciduous tree species. New Phytologist 129, 6371.
  • 46
    Webb A.A.R. & Hetherington A.M. (1997) Convergence of the abscisic acid, CO2, and extracellular calcium signal transduction pathways in stomatal guard cells. Plant Physiology 114, 15571560.
  • 47
    Webb A.A.R., McAinsh M.R., Mansfield T.A., Hetherington A.M. (1996) Carbon dioxide induces increases in guard cell cytosolic free calcium. The Plant Journal 9, 297304.
  • 48
    Wilkinson S. & Davies W.J. (1997) Xylem sap pH increase: a drought signal received at the apoplastic face of the guard cell that involves the suppression of saturable abscisic acid uptake by the epidermal symplast. Plant Physiology 113, 559573.
  • 49
    Will R.E. & Teskey R.O. (1997) Effect of irradiance and vapour pressure deficit on stomatal response to CO2 enrichment of four tree species. Journal of Experimental Botany 48, 20952102.