Effects of elevated CO2, warming and drought episodes on plant carbon uptake in a temperate heath ecosystem are controlled by soil water status

Authors


K. R. Albert. E-mail: kria@risoe.dtu.dk

ABSTRACT

The impact of elevated CO2, periodic drought and warming on photosynthesis and leaf characteristics of the evergreen dwarf shrub Calluna vulgaris in a temperate heath ecosystem was investigated. Photosynthesis was reduced by drought in midsummer and increased by elevated CO2 throughout the growing season, whereas warming only stimulated photosynthesis early in the year. At the beginning and end of the growing season, a T × CO2 interaction synergistically stimulated plant carbon uptake in the combination of warming and elevated CO2. At peak drought, the D × CO2 interaction antagonistically down-regulated photosynthesis, suggesting a limited ability of elevated CO2 to counteract the negative effect of drought. The response of photosynthesis in the full factorial combination (TDCO2) could be explained by the main effect of experimental treatments (T, D, CO2) and the two-factor interactions (D × CO2, T × CO2). The interactive responses in the experimental treatments including elevated CO2 seemed to be linked to the realized range of treatment variability, for example with negative effects following experimental drought or positive effects following the relatively higher impact of night-time warming during cold periods early and late in the year. Longer-term experiments are needed to evaluate whether photosynthetic down-regulation will dampen the stimulation of photosynthesis under prolonged exposure to elevated CO2.

Abbreviations
C

leaf carbon concentration

δ13C

carbon isotope ratio

Ci

intercellular CO2 concentration

C/N ratio

leaf carbon-to-nitrogen ratio

CO2

elevated CO2

D

experimental drought

delta (Pn, gs and SWC)

difference between the parameter mean for each treatment minus the mean for the unmanipulated control

delta T

difference between the temperature means for warming minus the mean for the unmanipulated control

FACE

free-air carbon enrichment

GDD

growing degree days

gs

stomatal conductance

Jmax

maximal velocity of RuBP regeneration

N

leaf nitrogen concentration

Pmax

light and CO2 saturated net photosynthesis

Pn

light saturated net photosynthesis

PWP

plant xylem water potential

SLA

specific leaf area

SWC

soil water content

T

passive night-time warming

Vcmax

maximal velocity of Rubisco carboxylation

WUE

photosynthetic water-use efficiency

INTRODUCTION

Climate is changing, and the anthropogenic forcing has been thoroughly documented (IPCC 2007). The level of atmospheric CO2 has increased from a pre-industrial level of 270 ppm to current values around 380 ppm, and is expected to increase to around 700 ppm (IPCC 2007). This will lead to temperature increases of 1.4–5.8 °C during the next 100 years, with more pronounced warming during night-time relative to daytime (IPCC 2001). Precipitation changes with prolonged summer droughts, heavy precipitation events and higher frequency of extremes are also expected (IPCC 2007). This has attracted focus on ecosystem responses to climatic changes and on ecosystem feedbacks to climate, and much effort is invested in understanding these complex impacts (Rustad 2006, 2008; Heimann & Reichstein 2008).

Photosynthetic carbon uptake is controlled directly by the factors that are predicted to change in the future, atmospheric CO2 concentration, temperature and water availability (Morison & Lawlor 1999; Sage & Kubien 2007; Lawlor & Tezara 2009), and will therefore almost certainly be affected by climate change. It is likely that the relative importance of the factors limiting carbon uptake may change as climate change affects various regulators of photosynthesis and ecosystems adjust to the new conditions. Carbon demand for storage, growth, metabolism or export also influences the residence time in the carbon pools and the flux rates between them (Körner 2006); this may interact with the factors that control carbon uptake. Elevated CO2 decreases stomatal conductance and increases intercellular CO2 concentration, light saturated net photosynthesis and plant WUE (Curtis & Wang 1998; Ainsworth & Long 2005; Ainsworth & Rogers 2007). Elevated CO2 often down-regulates photosynthetic capacity via the maximal velocity of ribulose 1·5-bisphosphate carboxylase/oxygenase (Rubisco) carboxylation, Vcmax and to lesser degree the maximal rate of RuBP regeneration, Jmax which may decrease the maximal light and CO2 saturated net photosynthesis, Pmax (Drake, Gonzalez-Meler & Long 1997; Moore et al. 1999; Ainsworth & Rogers 2007). This is in part caused by the build-up of carbon compounds produced in the Calvin cycle, and in part caused by limitations in the nitrogen supply (Drake et al. 1997; Ainsworth & Rogers 2007). Leaf carbon uptake may be sustained in elevated CO2 if photosynthetic down-regulation does not occur, as reported from grasslands (Hungate et al. 1997a) and forests (Körner et al. 2005). In the long term, however, nutrient availability is likely to be of increasing importance for sustained productivity in elevated CO2 (Oren et al. 2001; Lou et al. 2004; Reich et al. 2006; Menge & Field 2007).

Plant water status also affects the magnitude of plant carbon uptake in elevated CO2 (Körner 2000; Volk, Niklaus & Körner 2000; Knapp et al. 2002; Morgan et al. 2004). Elevated CO2 often results in reduced plant water consumption and leads to reduced soil water depletion, so-called water savings caused by reduced stomatal conductance (Morison & Gifford 1984; Hungate et al. 1997b; Leuzinger & Körner 2007; Robredo et al. 2007). This enables maintained plant carbon uptake in dry periods in elevated CO2 relative to ambient CO2 (Morison & Gifford 1984; Hungate et al. 1997b; Leuzinger & Körner 2007; Robredo et al. 2007).

Changes in plant water availability affect photosynthesis, particularly during dry conditions that limit photosynthesis because of metabolic impairments, as well as diffusion limitations (Jones 1985; Chaves 1991; Flexas & Medrano 2002; Lawlor 2002; Lawlor & Cornic 2002; Flexas et al. 2006a). The responsiveness is highly species specific (Knapp et al. 2002; Heisler & Weltzin 2006). However, the carbon balance of a plant enduring a water stress period not only depends on the degree of photosynthetic decline during water depletion, but also on the rate and degree of photosynthetic recovery, although information on these aspects is scarce (Flexas et al. 2006b; Galmes, Medrano & Flexas 2007; Lawlor & Tezara 2009).

Warming has been demonstrated to increase productivity (Arft et al. 1999; Rustad et al. 2001; Shaw et al. 2002; Dukes et al. 2005; Sage & Kubien 2007), but the impacts of this driver are complex; the confounding of direct and indirect effects can make it difficult to elucidate the underlying mechanisms of the responses of plants and ecosystems (Shaver et al. 2000; Peñuelas & Filella 2001). Warming can directly stimulate photosynthesis via provision of more optimal growth temperatures (e.g. Sage & Kubien 2007), and may be of particular importance in temperate ecosystems during periods of sub-optimal temperatures. Increased daytime respiration or changes in temperature acclimation can occur and may reduce the photosynthetic plant carbon uptake (Atkin & Tjoelker 2003). Indirect stimulation of daytime photosynthesis has been demonstrated in response to night-time warming, where an increased night-time plant respiration increased the carbon sink strength and in turn stimulated the following daytime net photosynthesis (Turnbull, Murthy & Griffin 2002; Turnbull et al. 2004). Warming may also affect photosynthesis and plant carbon uptake indirectly through increased length of the growing season (Menzel & Fabian 1999; White, Running & Thornton 1999; Walther et al. 2002; Cleland et al. 2006; Piao et al. 2007), plant phenology changes (Harte & Shaw 1995; Wan et al. 2005) and increased nutrient availability (Rustad et al. 2001). Some authors have argued that the indirect effects, via changes in nutrient and soil water availability, are more important than direct effects (Körner 2000; Shaver et al. 2000; Volk et al. 2000; Morgan et al. 2004; Wan et al. 2005; Lou 2007). Clearly, the effects of warming on plants and ecosystems are not straightforward and depend on the outcomes of the individual effects on many processes.

All climatic factors may influence photosynthesis, either positively or negatively; because they will change simultaneously in the future, and the overall effects need to be studied in combination. This is the goal of multifactor experimental manipulation studies (e.g. Norby & Lou 2004; Classen & Langley 2005; Dermody 2006; Rustad 2006; Heimann & Reichstein 2008), where interactions between drivers can be studied and quantified (Shaw et al. 2002; Henry et al. 2005; Norby et al. 2007; Luo et al. 2008), and where in-depth understanding of the causal mechanisms involved can be obtained (Heimann & Reichstein 2008; Rustad 2008).

Temperate heath is a semi-natural ecosystem and represents an important natural resource in the landscape of Northern and Western Europe, and the area covers about 70.000 ha in Denmark (Degn 2001). Heathland are sensitive to changes in environmental pressures as management practice, nitrogen deposition and natural succession (Heil & Bobbink 1993; Degn 2001; Terry et al. 2004).

In a temperate heath ecosystem, we therefore applied manipulations with elevated CO2, warming and summer drought in a full factorial design according to a scenario for climatic conditions in Denmark in 2075 (Boberg 2010; Christensen et al. 2010; Déqué & Somot 2010). Here, we report the impact on plant carbon uptake, plant water status and leaf characteristics. Considering the potential for complex interactive responses (i.e. via synergistic effects, increasing the combined response above what could be expected from single factors or antagonistic effects that would dampen the combined response), we constructed simple testable hypotheses according to the principle of Ockham's razor.

We hypothesized that:

  • • Elevated CO2 would increase plant carbon uptake and improve plant water status.
  • • Drought would decrease plant carbon uptake and negatively affect plant water status.
  • • Elevated CO2 would counteract the drought-reduced plant carbon uptake through additive effects.
  • • Warming would increase plant carbon uptake in spring and autumn, but have no effect on plant water status.
  • • Warming and elevated CO2 would in combination increase plant carbon uptake through additive effects.
  • • Responses in the full factorial combination to be additive from single-factor responses and the two-factor interactions.

MATERIALS AND METHODS

Site and experimental set-up

The study site is a dry heathland ecosystem on sandy soil in North Zealand, Denmark (55°53′N 11°58′E) dominated by the evergreen dwarf shrub heather (Calluna vulgaris L.) and wavy hairgrass (Deschampsia flexuosa L.). The ecosystem was subjected to the following experimental treatments: unmanipulated control (A), elevated CO2 (CO2), passive night-time warming (T), summer drought (D) and all combinations (TD, TCO2, DCO2, TDCO2) replicated in six blocks in a split-plot design. Each block consisted of two octagons of 6.8 m diameter each divided in four plots. In one octagon, CO2 was elevated to 510 ppm during daylight hours with the FACE technique, and in the other octagon CO2 level was ambient. Automated curtains covered one-half of each octagon during the night, impeding the loss of a large proportion of the daily incoming radiation energy, which increased the night air temperature by 1.4 °C on average. Another half of each octagon perpendicular to the warming curtains was periodically covered with a rain-excluding curtain, automatically activated by rain, thus creating an experimental drought period (further details, see Mikkelsen et al. (2008). In each experimental plot, hourly temperatures (+20 cm above surface, −2 and −5 cm depth; Pt100; Termokon, Mittenaar, Germany) and half hourly soil moisture (0–20 and 0–60 cm depth; TDR; Prenart Equipment APS, Frederiksberg, Denmark) were recorded. Climatic conditions (precipitation, air temperature, photosynthetically active radiation, wind directions and relative humidity) were recorded hourly by two automatic weather stations (see Mikkelsen et al. 2008 for details). The CO2 and warming treatments were initiated on 3 October 2005. The drought treatment started 3 July 2006 and was continued for 5 weeks until 4 August when SWC approached 5% in the top 20 cm of the soil. For further description of the site and experimental set-up, see Mikkelsen et al. (2008).

Leaf gas exchange

Measurements of leaf level CO2 and H2O fluxes were conducted in situ by using two LI-6400 (Li-Cor Biosciences, Lincoln, NE, USA) connected to standard 2 × 3 cm2 chambers with LED light source (6400-02B) and were carried out during six campaigns in 2006: 18–19 May, 13–14 June, 9–10 July, 17–21 August, 14–22 September and 13–20 October. Healthy shoots with consistent structure from the top of the canopy were selected for measurements. Pilot studies led to the following methodology, securing highly reproducible measurements: the shoots (one per plot) were flattened between transparent 0.2 mm nylon monofilament suspended in a metal frame with spacing 2.5–4 mm. The position of the shoots was upright and south facing. This approach facilitated uniform light doses to the shoots before and after the measurements. During measurement, the leaf cuvette was secured using flexible rods (Magic Arm 143; Manfrotto, Cassola, Italy) attached to the octagon scaffolds. The protocol for CO2–response curves was optimized and tested, and resulted in the following procedure: Samples were acclimated in the chamber for 2–4 min at the prevailing level of CO2 (380 ppm in non-FACE plots and 510 ppm in elevated CO2 plots), until net photosynthesis and stomatal conductance were stabilized (±1 CV over 30 s). The CO2 response curves were measured by stepping down CO2 concentration (from 380 to 300, 200, 100 and 50 ppm in non-FACE plots, and from 510 to 450, 380, 300, 200, 100 and 50 ppm in elevated CO2 plots). The CO2 concentration was then returned to the start level (380 ppm in non-FACE plots and 510 in elevated CO2 plots) for 4 min of re-acclimation until the initial state was again attained. The CO2 concentration was then stepped up to saturation at 1400 ppm (in non-FACE plots via 450, 510, 650, 800, 1000 and 1200 ppm, and in elevated CO2 plots via 650, 800, 1000 and 1200 ppm). The measurements were performed at a saturating light level of 1500 µmol photons m−2 s−1, using the Li-Cor Auto program ‘A/Ci curve’ (settings: minimum 45 s and maximum 60 s, reference CO2 stable in 10 s with CV < 1%, Ci stable in 10 s with CV < 1%, IRGA matching performed at each step). Block temperature was held constant at 25 °C, and relative humidity was stabilized at 45–55% during measurements. All measurements were area corrected (see below). Leak corrections were applied with the empty chamber approach (see e.g. Bernacchi et al. 2002; Bruhn, Mikkelsen & Atkin 2002; Manter & Kerrigan 2004). Gas exchange parameters (i.e. light saturated net photosynthesis, Pn; transpiration, Tr; stomatal conductance, gs; and intercellular CO2 concentration, Ci) were extracted from the CO2 response curves at CO2Ref = 380 ppm, CO2 in non-FACE plots and at CO2Ref = 510 ppm CO2 in FACE plots. The Pmax values were extracted at CO2Ref = 1200 ppm in all treatments. The WUE, WUE = PnTr−1, was calculated from the first point measured on the A/Ci curves with Pn and Tr values at CO2Ref = 380 ppm, CO2 in non-FACE plots and at CO2Ref = 510 ppm CO2 in FACE plots.

Plant water potential (PWP)

The PWP was measured using the Scholander method (Scholander et al. 1965) using a pressure bomb (PWSC Research Model 3000; Soil Moisture Equipment Corp., Goleta, CA, USA). In each plot, two to four shoots were randomly sampled from the top of the canopy during the last campaign night in July, August, September and October, between 0200 and 0500 h (pre-dawn). The sampling was performed sequentially by block, and measurements were conducted immediately after sampling.

Leaf characteristics: weight, area, nitrogen, carbon and δ13C

The shoots were harvested after each campaign (see above), weighed and the area determined by digital pictures taken of the plant material flattened by transparent Plexiglas together with a square of known area and the projected shoot area determined by use of a pixel counting program (Bitmap, Danbæk, Department of Biology, University of Copenhagen). The specific leaf area (SLW) was estimated from weight and area. Shoots were split into photosynthetic active parts (‘green’) and the rest (‘brown’), dry weight was determined after oven drying at 80 °C, and the plant material analysed for C and N concentration and 13C natural abundance (δ13C), determined on a CN analyser (Eurovector, Milan, Italy) coupled to an isotope ratio mass spectrometer (Isoprime, Cheadle Hulme, UK). A reduction in δ13C is assumed to reflect improved WUE over a period (Farquhar, O'Leary & Berry 1982; Farquhar, Ehleringer & Hubick 1989; Ehleringer 1993).

Statistics

Analyses of variance (anova) were made in SAS using the proc mixed to test for effects of the fixed factors: time, T; D, CO2 and their interactions T × D, T × CO2, D × CO2, T × D × CO2. To take into account the split-plot design, D × octagon, T × octagon, block and octagon were included as random factors. Pre-treatment data were initially included in the full model as fixed factors (analogous to covariates). Backwards model selection was performed, reducing the model by progressively dropping non-significant fixed factors, starting with the higher-order interactions until only significant (P < 0.05) or near significant terms (P < 0.10) remained. Significance levels are reported as a tendency with † when P < 0.10 and as significant with * when P < 0.05, ** when P < 0.001 and *** when P < 0.0001. All data were normally distributed, and some parameters were transformed to fulfil the assumption of variance homogeneity.

RESULTS

Climate and experimental control measurements

The snow melted in late March to early April, and the mean daily temperature in 2 m height gradually increased to 25 °C in late July. The autumn in 2006 was the warmest ever recorded in Denmark with mean daily air temperatures above 11 °C in late November (Fig. 1). The warming treatment increased the canopy temperature at night by 1–2 °C (Fig. 1). Based on a GDD criteria, with a threshold for growth at 5 °C (Beier et al. 2004), the warming treatment increased the accumulated GDD by 33% from 1 April to 15 May, and annually by 7% (Mikkelsen et al. 2008). The CO2 concentration in the FACE plots was elevated to 510 ppm during daytime (monthly average daytime concentrations were 500–520 ppm CO2, see Mikkelsen et al. 2008). The SWC fluctuated from 7.6% (July) to 18.0% (August) over 0–20 cm depth, and from 8.3% (July) to 13.3% (October) over 0–60 cm depth (Fig. 2; Table 1). The experimental drought significantly decreased the mean SWC in July and August to 6.0 and 9.5%, respectively. When the experimental drought ended the SWC remained at depressed levels in the drought plots until rewetting occurred in late August (Fig. 3; Table 1). Warming reduced the SWC throughout the year. Elevated CO2 had a positive effect on SWC in early drought period (Fig. 2) and during the campaign days in July (Fig. 3; Table 1). Elevated CO2 partly compensated for the drought-induced SWC reduction at the peak of the drought in August (Figs 2 & 3). Drought and elevated CO2 (D × CO2) interacted to compensate for the negative effects of drought at the peak of the experimental drought (Figs 2 & 3; Table 1).

Figure 1.

Temperature. (a) Daytime air temperature at 2 m height (hourly means) at the experimental site is depicted as black line and 0 °C reference as dotted line. (b) Daytime temperature difference at 20 cm height for the warming minus control treatments. No significant effects on daytime temperature appeared, thus temperature was not directly affected when leaf gas exchange and fluorescence measurements were conducted; (c) Night-time temperature difference at 20 cm height for warming minus control treatments. Night-time corresponds to the period where the passive night-time warming treatment was active.

Figure 2.

Precipitation and SWC. (a) The daily averages of SWC in control plots over 0–20 cm and 0–60 cm depth and the daily accumulated precipitation. Each panel is divided into pre-drought, experimental drought and post-drought periods by vertical black lines. The vertical dashed line indicates the lag phase, still influenced by previous experimental drought period, because of sparse precipitation. (b) Change in percent of the SWC in 0–20 cm of treatments compared to control. Dotted line is the zero reference line for the unmanipulated control. The treatments are: elevated CO2 (CO2); experimental drought (D); passive night-time warming (T); and the combination of DCO2. Warming and elevated CO2 treatments were active in all periods, whereas the drought was only active in the experimental drought period. All treatments were replicated six times, in total 48 plots. In each period, the significant effects of the experimental factors T, D, CO2 and their interactions are indicated with *** for P < 0.001, ** for P < 0.01 and * for 0.01 < P < 0.05. Timing of gas exchange and leaf characteristic surveys are indicated with filled triangles. (c) Change in percent of the SWC in 0–60 cm of a treatment compared to control.

Table 1.  Soil water content in 0–20 and 0–60 cm
 AΔ%TΔ%DΔ%TDΔ%CO2Δ%TCO2Δ%DCO2Δ%TDCO2TDCO2D × CO2
  1. SWC over 0–20 and 0–60 cm from the days where leaf gas exchange campaign days were performed are shown. For the unmanipulated control (A), the means ± SE are given. For all other treatments, the change in % (Δ%) compared to control is given. Statistical significant effects and interactions are shown with **if P < 0.001, *if P < 0.05, †if P < 0.10 and ns if P > 0.10. The F values and response directions given: The ↓ indicates a single-factor decrease; the ↑ indicates a single-factor decrease, the ↑↑ a two-factor synergistic increase and ↑↓ a two-factor antagonistic effect. Treatments are (A) unmanipulated control (T) passive night-time warming (D) experimental drought (TD) warming and drought combined (CO2) elevated CO2 (TCO2) warming and elevated CO2 combined (DCO2) drought and elevated CO2 combined (TDCO2) warming, drought and elevated CO2 combined.

SWC 0–20 cm            
 May10.1 ± 2.0−6.7−8.1−8.7−17.4−15.3−4.9−15.98.20*0.24 ns0.37 ns0.64 ns
 June8.6 ± 2.7−8.7−11.6−18.9−18.8−18.8−8.4−11.815.00**0.56 ns0.11 ns2.88 ns
 July7.6 ± 1.3−16.4−26.3−35.4−24.1−21.8−17.2−27.24.95*18.13**0.07 ns2.99↑↓
 August18.0 ± 2.2−4.7−13.3−20.30.7−9.3−8.0−17.416.54*18.88**0.14 ns1.35 ns
 September12.7 ± 1.9−3.5−6.0−10.11.8−3.32.4−11.32.82 ns0.12 ns0.09 ns0.25 ns
 October15.3 ± 1.4−7.3−3.6−7.9−2.0−5.33.9−5.025.71**0.23 ns0.05 ns0.45 ns
SWC 0–60 cm            
 May11.4 ± 2.6−9.2−9.3−11.74.1−3.45.11.52.91 ns0.00 ns1.82 ns1.83 ns
 June10.6 ± 2.8−12.3−11.9−16.12.8−5.23.2−0.87.39*0.92 ns1.66 ns4.19*↑↓
 July8.3 ± 2.2−18.3−27.3−31.84.5−0.33.7−9.75.14*22.21**6.60*6.30*↑↓
 August13.0 ± 3.2−9.7−26.9−33.04.2−4.6−11.5−15.48.20*6.16*0.91 ns5.83*↑↓
 September11.6 ± 2.5−7.5−7.0−13.67.3−1.01.6−3.23.821.07 ns0.84 ns0.00 ns
 October13.3 ± 1.9−8.3−7.3−13.12.8−6.6−1.60.46.15*0.92 ns0.66 ns6.37*↑↓
Figure 3.

SWC during gas exchange campaign surveys. SWC in 0–20 and 0–60 cm. Shown is the difference between the means for each treatment minus unmanipulated control (delta SWC) ± SEs on the days where leaf gas exchange campaigns were performed. Treatments (T) passive night-time warming (D) experimental drought (TD) warming and drought combined (CO2) elevated CO2 (TCO2) warming and elevated CO2 combined (DCO2) drought and elevated CO2 combined (TDCO2) warming, drought and elevated CO2 combined. The significant main effects T, D, CO2 and their interactions are indicated with *** for P < 0.0001, ** for P < 0.001 and with * for P < 0.05. The mean ± SEs in the unmanipulated control (0–20 cm; vol%) were: 10.1 ± 2.0 in May, 8.6 ± 2.7 in June, 7.6 ± 1.3 in July, 18.0 ± 2.2 in August, 12.7 ± 1.9 in September and 15.3 ± 1.4 in October. In 0–60 cm mean ± SE were: 11.4 ± 2.6 in May, 10.6 ± 2.8 in June, 8.3 ± 2.2 in July, 13.0 ± 2.3 in August, 11.6 ± 2.5 in September and 13.3 ± 1.9 in October.

PWP

In June, the mean xylem PWP in the controls was similar to −3.5 bar, decreased to −5.0 bar in August and then increased to −3.0 bar in September and to above −1 bar in October (Tables 2 & 3). The experimental drought significantly decreased the PWP in July and August by 42 and 39%, respectively. In September, warming significantly reduced the PWP by 13%.

Table 2.  Statistics of plant eco – physiological responses
Calluna vulgaris TDCO2T × DT × CO2D × CO2T × D × CO2
  1. F values and significance levels ( *P<0.05; **P<0.01; ***P<0.0001) for the main effects night-time warming (T), drought (D), elevated CO2 (CO2) and their interactions on light-saturated net photosynthesis (Pn), intercellular CO2 concentration (Ci), stomatal conductance (gs), water use efficiency (WUE), maximal light and CO2-saturated net photosynthesis (Pmax), plant water potential (PWP), leaf carbon-to-nitrogen ratio (C/N), specific leaf weight (SLW) and leaf δ13C by a linear mixed model anova. Degrees of freedom (df). Additive effects: single-factor increase (↑) and decrease (↓). Synergistic effects where the response in the combination exceeded what was expected from the single factor [i.e. an increase (↑↑) or decrease (↓↓)]. Antagonistic effects where response in the combination was less than expected from the single-factor responses (↑↓).

Numerator df 1111111
Denominator df 1010510101010
MayPn8.29*0.025.51*0.630.010.160.04
JunePn1.232.4410.73**2.070.661.360.09
JulyPn0.050.879.41***1.520.410.000.31
AugustPn9.17*23.72***19.69***1.611.244.70*↑↓0.01
SeptemberPn0.030.561.132.301.661.822.16
OctoberPn0.320.4111.57**1.068.21**↑↑1.080.70
Maygs0.080.030.711.190.840.080.02
Junegs0.411.510.027.410.010.010.50
Julygs7.18*0.000.0022.17***↑↓1.772.030.07
Augustgs1.1024.99***2.290.080.885.49**↑↓2.23
Septembergs0.001.113.080.900.831.660.44
Octobergs0.640.700.010.005.48*↑↑1.190.59
MayWUE0.150.197.41*1.510.010.971.07
JuneWUE1.140.2546.42*0.150.000.260.87
JulyWUE1.530.085.61*3.63↑↓0.111.370.08
AugustWUE3.364.78*41.43***0.011.090.002.29
SeptemberWUE0.070.3211.46*0.071.120.072.14
OctoberWUE0.190.056.68*0.851.622.500.04
MayCi0.130.35125.55***3.012.300.010.04
JuneCi0.250.0424.31***0.400.030.030.83
JulyCi1.710.025.48*0.070.040.860.02
AugustCi5.42*15.11***135.54***2.550.003.573.39
SeptemberCi0.280.1232.28**0.630.910.060.67
OctoberCi0.270.0034.37**0.281.291.450.43
MayPmax5.26*0.200.250.770.161.970.02
JunePmax7.35**0.500.070.310.900.110.67
JulyPmax0.030.720.020.430.030.000.04
AugustPmax0.6424.67***5.04*2.210.348.92**↑↓2.17
SeptemberPmax0.040.480.650.680.570.681.36
OctoberPmax3.47 †↑0.340.112.211.542.190.33
JulyPWP1.7713.83**0.110.520.721.260.00
AugustPWP1.3810.35**1.130.020.010.130.11
SeptemberPWP5.58*0.060.010.011.801.341.01
OctoberPWP0.462.791.040.511.000.010.43
MayC/N0.900.0014.29*0.140.017.40*0.48
JuneC/N0.172.546.66*0.010.002.360.36
JulyC/N0.030.001.660.190.220.030.57
AugustC/N0.010.7017.11*0.541.090.060.00
SeptemberC/N0.230.251.750.682.830.130.16
OctoberC/N0.031.341.110.210.170.090.48
Mayδ13C1.300.0261.40***1.997.35*↓↓0.090.04
Juneδ13C6.24*0.01248.20***0.102.131.661.00
Julyδ13C1.071.49310.17***0.760.011.841.77
Augustδ13C1.081.701055.35***0.521.040.761.57
Septemberδ13C0.384.25169.55***0.730.085.48*↑↑0.84
Octoberδ13C0.725.58*242.19***0.2013.02*↓↓12.18*↑↑2.05
Table 3.  Averages and standard errors of plant eco – physiological responses
Calluna vulgaris ATDTDCO2DCO2TCO2TDCO2
  1. For each month, the mean and SE (n = 6) for the treatments night-time warming (T), drought (D), elevated CO2 (CO2) and their combinations on light-saturated net photosynthesis (Pn; µmol CO2 m−2 s−1), intercellular CO2 concentration (Ci; ppm CO2), stomatal conductance (gs; mmol H2O m−2 s−1), WUE (µmol CO2 mmol−1 H2O), maximal light and CO2-saturated net photosynthesis (Pmax; µmol CO2 m−2 s−1), PWP (bar), leaf carbon-to-nitrogen ratio (C/N) and leaf δ13C (‰). A is ambient control plots.

MayPn5.89 ± 0.077.68 ± 0.175.48 ± 0.216.76 ± 0.128.14 ± 0.109.13 ± 0.459.11 ± 0.219.75 ± 0.64
JunePn4.06 ± 0.753.62 ± 0.593.53 ± 1.094.53 ± 0.525.74 ± 1.076.37 ± 0.986.42 ± 0.709.27 ± 2.07
JulyPn4.60 ± 0.304.49 ± 0.754.75 ± 0.834.67 ± 0.535.81 ± 1.857.51 ± 0.726.42 ± 0.965.93 ± 0.97
AugustPn6.33 ± 0.476.60 ± 0.811.44 ± 0.752.94 ± 0.579.59 ± 0.866.13 ± 0.697.51 ± 0.374.57 ± 0.64
SeptemberPn6.33 ± 0.876.17 ± 1.637.44 ± 1.065.02 ± 1.468.85 ± 0.495.20 ± 0.377.99 ± 1.247.86 ± 1.17
OctoberPn10.81 ± 0.855.63 ± 1.1312.53 ± 1.648.41 ± 1.2312.23 ± 1.5913.86 ± 0.6417.73 ± 1.4016.31 ± 1.58
Maygs0.23 ± 0.010.28 ± 0.010.24 ± 0.010.22 ± 0.000.24 ± 0.000.23 ± 0.010.27 ± 0.010.22 ± 0.01
Junegs0.10 ± 0.020.08 ± 0.020.09 ± 0.040.12 ± 0.030.09 ± 0.020.08 ± 0.020.08 ± 0.020.12 ± 0.03
Julygs0.05 ± 0.010.06 ± 0.010.08 ± 0.010.06 ± 0.010.05 ± 0.010.08 ± 0.020.07 ± 0.030.04 ± 0.01
Augustgs0.27 ± 0.040.24 ± 0.030.05 ± 0.020.08 ± 0.020.19 ± 0.030.12 ± 0.030.17 ± 0.020.08 ± 0.02
Septembergs0.17 ± 0.020.15 ± 0.060.17 ± 0.030.13 ± 0.040.13 ± 0.020.06 ± 0.010.13 ± 0.040.10 ± 0.01
Octobergs0.27 ± 0.050.16 ± 0.020.29 ± 0.050.24 ± 0.050.20 ± 0.060.22 ± 0.030.30 ± 0.040.32 ± 0.04
MayWUE4.50 ± 0.093.70 ± 0.033.35 ± 0.113.91 ± 0.095.03 ± 0.125.02 ± 0.174.73 ± 0.095.11 ± 0.16
JuneWUE2.21 ± 0.412.80 ± 0.462.37 ± 0.542.73 ± 0.414.16 ± 0.544.24 ± 0.274.73 ± 0.414.19 ± 0.36
JulyWUE4.05 ± 0.234.00 ± 0.613.19 ± 0.404.41 ± 0.585.59 ± 1.405.47 ± 0.945.51 ± 1.137.29 ± 0.80
AugustWUE1.90 ± 0.352.03 ± 0.421.56 ± 0.411.64 ± 0.284.32 ± 0.723.18 ± 0.893.21 ± 0.143.05 ± 0.22
SeptemberWUE3.13 ± 0.423.37 ± 0.433.87 ± 0.353.02 ± 0.495.43 ± 0.435.15 ± 0.955.18 ± 1.126.59 ± 1.61
OctoberWUE4.45 ± 0.333.44 ± 1.114.54 ± 0.834.28 ± 0.745.61 ± 0.905.65 ± 0.656.35 ± 0.875.82 ± 0.94
MayCi309 ± 4313 ± 4318 ± 2305 ± 10397 ± 5418 ± 3421 ± 5417 ± 6
JuneCi297 ± 13271 ± 14289 ± 18279 ± 16352 ± 15336 ± 17332 ± 15349 ± 13
JulyCi224 ± 15229 ± 14265 ± 15212 ± 16261 ± 17284 ± 15270 ± 17217 ± 10
AugustCi310 ± 7304 ± 8350 ± 10332 ± 10388 ± 7404 ± 7379 ± 6392 ± 2
SeptemberCi295 ± 8287 ± 6282 ± 4295 ± 9366 ± 6359 ± 12356 ± 17348 ± 17
OctoberCi292 ± 8305 ± 10284 ± 16295 ± 12374 ± 12380 ± 13376 ± 11376 ± 11
MayPmax14.7 ± 0.219.1 ± 0.615.2 ± 0.515.6 ± 0.214.2 ± 0.116.8 ± 0.618.6 ± 0.518.5 ± 0.9
JunePmax11.5 ± 1.912.4 ± 1.49.7 ± 1.213.3 ± 1.97.7 ± 0.58.8 ± 0.912.3 ± 2.015.1 ± 1.8
JulyPmax11.2 ± 1.312.2 ± 1.512.6 ± 1.412.7 ± 0.711.3 ± 2.713.1 ± 1.112.1 ± 2.012.3 ± 1.9
AugustPmax13.0 ± 0.916.9 ± 1.27.1 ± 2.211.0 ± 1.514.3 ± 1.010.5 ± 2.112.6 ± 0.39.5 ± 1.0
SeptemberPmax14.4 ± 1.814.8 ± 2.315.8 ± 1.513.3 ± 2.314.7 ± 0.210.4 ± 2.113.4 ± 3.211.0 ± 2.0
OctoberPmax21.3 ± 3.021.3 ± 3.627.4 ± 2.018.2 ± 2.419.5 ± 2.225.8 ± 3.024.4 ± 2.825.7 ± 2.1
JulyPWP−3.6 ± 0.6−4.9 ± 0.5−4.5 ± 0.5−6.2 ± 0.5−3.8 ± 0.8−5.3 ± 0.7−3.4 ± 0.6−6.2 ± 0.7
AugustPWP−5.0 ± 1.2−4.4 ± 1.0−5.9 ± 1.6−5.4 ± 1.3−3.1 ± 0.8−5.3 ± 1.7−2.7 ± 0.7−4.6 ± 1.4
SeptemberPWP−3.1 ± 0.5−3.9 ± 0.5−4.1 ± 0.6−4.2 ± 0.4−3.4 ± 0.5−4.7 ± 0.4−4.8 ± 0.5−4.5 ± 0.6
OctoberPWP−0.8 ± 0.1−0.9 ± 0.1−0.7 ± 0.1−0.8 ± 0.1−0.8 ± 0.1−0.8 ± 0.1−0.8 ± 0.1−0.7 ± 0.1
MayC/N23.1 ± 1.022.1 ± 1.123.8 ± 1.524.0 ± 1.026.6 ± 1.025.5 ± 0.726.3 ± 1.224.8 ± 1.3
JuneC/N29.4 ± 1.930.4 ± 1.429.8 ± 1.429.8 ± 1.834.9 ± 1.831.0 ± 1.234.6 ± 2.332.1 ± 1.4
JulyC/N37.3 ± 2.039.6 ± 1.338.8 ± 2.137.6 ± 1.540.7 ± 3.040.6 ± 1.439.1 ± 1.039.9 ± 2.8
AugustC/N23.1 ± 0.824.8 ± 1.323.2 ± 1.223.9 ± 1.027.1 ± 2.226.8 ± 1.126.6 ± 1.825.3 ± 1.1
SeptemberC/N22.3 ± 0.821.4 ± 1.022.5 ± 1.220.9 ± 0.322.4 ± 0.722.5 ± 1.724.1 ± 1.422.2 ± 1.7
OctoberC/N22.9 ± 1.223.1 ± 1.122.0 ± 1.222.6 ± 1.323.9 ± 1.023.5 ± 1.024.5 ± 1.222.5 ± 1.2
Mayδ13C−28.06 ± 0.61−27.52 ± 0.30−27.54 ± 0.44−27.72 ± 0.37−31.23 ± 0.39−31.02 ± 0.24−32.08 ± 0.97−32.42 ± 0.53
Juneδ13C−27.43 ± 0.13−27.56 ± 0.31−27.59 ± 0.17−28.00 ± 0.34−31.02 ± 0.50−30.93 ± 0.16−32.31 ± 0.53−31.70 ± 0.29
Julyδ13C−27.35 ± 0.48−27.69 ± 0.38−27.25 ± 0.27−27.91 ± 0.38−33.46 ± 1.06−33.08 ± 0.29−34.86 ± 0.73−32.90 ± 0.69
Augustδ13C−27.93 ± 0.30−28.07 ± 0.31−27.95 ± 0.29−27.82 ± 0.16−34.37 ± 0.74−33.29 ± 0.22−34.34 ± 0.27−34.23 ± 0.28
Septemberδ13C−27.53 ± 0.24−27.74 ± 0.40−27.64 ± 0.33−27.81 ± 0.39−35.40 ± 0.92−33.35 ± 0.83−35.26 ± 0.81−34.48 ± 0.89
Octoberδ13C−28.42 ± 0.27−28.02 ± 0.42−28.42 ± 0.24−28.67 ± 0.32−36.34 ± 0.89−35.26 ± 0.69−37.78 ± 0.77−35.45 ± 0.53

Leaf gas exchange

Net photosynthesis, Pn, in the control plots ranged from 3.6 ± 0.3 to 10.8 ± 1.4 µmol CO2 m−2 s being highest at the early and late season (May and October), and smallest at midsummer (July; Fig. 4). Elevated CO2 generally increased net photosynthesis throughout the growing season (Fig. 4). Warming increased photosynthesis (May and August; Fig. 4) and drought reduced photosynthesis at the peak of the drought (August; Fig. 4). An antagonistic interaction (D × CO2) dampened the response of net photosynthesis in the combination of experimental drought and elevated CO2, thus net photosynthesis decreased less than expected from single factors in August (Fig. 4). A T × CO2 interaction synergistically increased net photosynthesis to even higher levels in the TCO2 and TDCO2 treatments relative to non-CO2 treatments in October (Fig. 4).

Figure 4.

Light-saturated net photosynthesis of Calluna. Light-saturated photosynthesis was extracted from A/Ci curves with Pn at reference CO2 (CO2Ref) = 380 ppm CO2 in non-FACE plots and at CO2Ref = 510 ppm CO2 in FACE plots. Leaf cuvette conditions were: saturating light intensity, 1500 µmol photons m−2 s−1; temperature, 25 °C; relative humidity, 45–55%. Shown are the differences between means for each treatment minus the unmanipulated control (delta Pn) ± SE. Treatments are (T) passive night-time warming (D) experimental drought (TD) warming and drought combined (CO2) elevated CO2 (TCO2) warming and elevated CO2 combined (DCO2) drought and elevated CO2 combined (TDCO2) warming, drought and elevated CO2 combined. The significant main effects T, D and CO2, and their interactions are indicated with *** for P < 0.0001, ** for P < 0.001 and with * for P < 0.05. Unmanipulated control mean ± SE (µmol CO2 m−2 s−1) were: 5.9 ± 0.1 in May, 4.1 ± 0.8 in June, 4.1 ± 0.3 in July, 6.3 ± 0.4 in August, 0.62 ± 0.9 in September and 10.8 ± 0.9 in October.

Stomatal conductance decreased by 40–60% as the SWC was reduced from May through July, and strongly increased to the highest levels in August and October after rewetting (Fig. 5). The experimental drought had a strong negative effect on gs at the peak of the drought in August (Fig. 5). In July, warming increased stomatal conductance, but when combined with drought, an antagonistic effect (T × D) dampened the response. In August, both drought and elevated CO2 decreased stomatal conductance. When combined, an antagonistic effect (D × CO2) reduced stomatal conductance less than expected from single factors. In October, there were no effects of warming and elevated CO2, but when combined (T × CO2), the stomatal conductance increased synergistically (Fig. 5).

Figure 5.

Stomatal conductance of Calluna. Stomatal conductance was extracted as the first measured point on the A/Ci curves. Shown are the differences between the treatment means minus the unmanipulated control (delta gs) ± SE. Treatments are (T) passive night-time warming (D) experimental drought (TD) warming and drought combined (CO2) elevated CO2 (TCO2) warming and elevated CO2 combined (DCO2) drought and elevated CO2 combined (TDCO2) warming, drought and elevated CO2 combined. The significant main effects T, D and CO2, and their interactions are indicated with *** for P < 0.0001, ** for P < 0.001 and with * for P < 0.05. Unmanipulated control mean ± SE (mmol H2O m−2 s−1) were: 0.23 ± 0.01 in May, 0.10 ± 0.02 in June, 0.05 ± 0.01 in July, 0.27 ± 0.04 in August, 0.17 ± 0.02 in September and 0.27 ± 0.05 in October.

The WUE varied from about 2.5 to 5 µmol CO2 mmol−1 H2O, with a seasonal low in August and highest WUE in May and October (Tables 2 & 3). Elevated CO2 generally stimulated WUE. Except for a marginal reduction of WUE by drought and by warming in August, the other climate factors had little or no effects (Tables 2 & 3). Elevated CO2 increased the intercellular CO2 in all months, while drought increased only following the peak of the drought in early August (Tables 1 & 2).

The maximal light and CO2 saturated net photosynthesis (Pmax) was stimulated by warming in the early summer (May and June; Tables 1 & 2). In combination with elevated CO2, at the peak of the experimental drought, there was an antagonistic D × CO2 interaction that resulted in Pmax being reduced less than expected from the single factors alone. Net photosynthesis increased with higher stomatal conductance (Fig. 6), and this relationship was similar in both elevated and ambient CO2. Given a similar magnitude of stomatal conductance, net photosynthesis was higher in elevated CO2 compared to ambient CO2. This indicates that a major part of the stimulation of net photosynthesis by elevated CO2 was caused by higher intercellular CO2 concentrations.

Figure 6.

The relationship between stomatal conductance, net photosynthesis and leaf δ13C. Shown are the (a) average net photosynthesis versus average stomatal conductance, and (b) leaf δ13C versus average stomatal conductance from each gas exchange campaign. In (c) average stomatal conductance versus average SWC over 0–60 cm. Data were divided into ambient CO2 (A, T, D, TD; closed circles) and elevated CO2 (CO2, TCO2, DCO2 and TDCO2; open circles) and n = 24 for each data set.

Leaf and shoot characteristics

The leaf and shoot characteristics generally showed only small changes in response to the treatments with some minor exceptions such as reduced leaf-to-shoot ratio by drought at the peak of the drought and a stimulated leaf-to-shoot ratio by elevated CO2 in September and October (data not shown). In addition, the specific leaf weight (SLW) (g m−2) showed no or only minor and single effects of the treatments on top of the natural growth pattern; significant reductions in SLW were observed from May to August as leaf area increased at a higher rate than the weight in this period. In September and October, the SLW was similar to the level in August. Elevated CO2 increased the SLW in July, while drought reduced the SLW in July and August (data not shown).

The C/N ratio of the leaves increased significantly each month from May through July, but dropped to a significantly lower level in the period from August to October (Tables 2 & 3). Elevated CO2 increased the observed C/N ratio in May, June and August, while C/N was unaffected by other treatments (Tables 2 & 3).

The CO2 treatments generally depleted the leaf δ13C because of the lower δ13C signature of the CO2 gas added to the FACE plots (c. −28.1 δ13C as opposed to −8.0 in ambient air) (Tables 2 & 3). In June, warming decreased δ13C, mainly in elevated CO2 plots. The observed δ13C was synergistically decreased when warming was combined with elevated CO2 (T × CO2) early and late in the growing season (May and October). The experimental drought influenced δ13C after the drought treatment ended, in September and October, including synergistic effects (D × CO2) on δ13C when drought was combined with elevated CO2.

DISCUSSION

Drought

Calluna vulgaris is a slow-growing evergreen shrub considered to be a stress tolerant competitor (Grime, Hodgson & Hunt 1988) with small in-rolled leaves with waxy surfaces and trichomes sheltering the stomates, long leaf lifespan, low nitrogen content, low nutrient loss and deep roots (Aerts 1993, 1995). Despite these adaptive traits for dry and nutrient poor habitats, the progressively lower water availability from May to July in the control plots decreased photosynthesis and stomatal conductance simultaneously, while photosynthetic capacity (Pmax) was maintained. The experimental drought decreased SWC even further in July and August. This reduced the PWP, leaf-to-shoot ratio and SLW in July, as well as all photosynthesis-related parameters (gs, Pn and WUE) at the peak of the experimental drought in early August. These results are in agreement with previous studies showing a decrease in PWP, Pn, Tr and gs in relation to drought (Gordon et al. 1999a; Jackson, Irvine & Grace 1999). In contrast, other studies found little to no effect of water availability on WUE in Calluna growing under dry conditions (Gordon et al. 1999a,b; Llorens et al. 2004). At the end of the experimental drought period, water shortage was severe and the photosynthetic capacity (Pmax) was down-regulated. Thus, our results from the control and drought plots demonstrate that mild to moderate reductions in soil moisture reduce stomatal conductance, but not WUE. In contrast, under severe water shortage, as during the peak of the drought experiment, down-regulation of the photosynthetic capacity will lead to decreased photosynthesis and consequently a decrease in WUE. These responses indicate that reduced stomatal conductance controls photosynthesis at mild/moderate water shortage, whereas metabolic impairment by reduced photosynthetic capacity dominates the limitation of photosynthesis when water shortage is severe. After rewetting and increased water availability, none of the negative effects of the previous experimental drought on photosynthetic performance persisted. This demonstrates a strong capacity of Calluna to recover photosynthetic performance, even after severe drought.

Elevated CO2

Elevated CO2 increased photosynthesis and WUE during most of the growing season. The increased photosynthesis was associated with, or driven by, increased intercellular CO2 concentration (Fig. 5) generating higher substrate availability for Rubisco, in line with, for example Ainsworth & Long (2005) and Ainsworth & Rogers (2007), and thus stimulation of photosynthesis also led to increased leaf C/N ratios. Surprisingly, no general reduction in stomatal conductance was seen in elevated CO2, although this is often reported (e.g. Ainsworth & Long 2005; Ainsworth & Rogers 2007). This suggests that the improved WUE observed in our elevated CO2 plots was caused by increased photosynthesis.

The soil water savings observed under elevated CO2 were not clearly coupled to reductions in stomatal conductance, which, together with LAI adjustments, have been the primary water-saving mechanisms in other elevated CO2 studies (Niklaus, Spinnler & Körner 1998; Morgan et al. 2004; Leuzinger & Körner 2007). However, the conservation of soil water in our study may have been associated with reduced stomatal conductance, as shown by the marginally significant effects seen in August and October. The robust detection of these responses may have been hindered by the lesser statistical power provided by the monthly measurements of leaf gas exchange, compared to the greater statistical power provided by the half-hourly measurements of SWC. Further, heterogeneous structural conditions may prevail within the Calluna canopy, and we cannot exclude that shoots, other than the uppermost shoots selected for measurement, may have responded to the elevated CO2 by reducing stomatal conductance. This demonstrates that more frequent measurements of leaf gas exchange may be necessary to fully monitor changes in stomatal conductance through periods of water scarcity, but also that Calluna seems to take advantage of the soil water savings occurring to sustain photosynthesis in elevated CO2 during dry periods. Because rewetting clearly increased photosynthesis, leaf-to-shoot ratio and the C/N ratio in elevated CO2 plots, it seems that the photosynthetic stimulation is closely dependent on water availability, with low photosynthesis stimulation during dry periods and high photosynthesis stimulation when water availability is high.

Interestingly, soil water savings were not detected over the upper 0–20 cm, but only over 0–60 cm. This suggests that intensive competition for water in the upper 0–20 cm occurred and that the species with the largest capacity for water uptake in the 20–60 cm compartment, for example via a deeper and more extensive roots system, are likely to be the primary species causing the water saving. Reduced water consumption, via stomatal conductance or biomass reduction in the co-occuring grass Deschampsia flexuosa, could potentially also influence the SWC. In Calluna, the rooting systems may extend down to 84 cm (Gimingham 1960), while the co-occuring Deschampsia extends down to 58 cm (Scurfield 1954). On the CLIMAITE field site, total root biomass in 0–15 cm was reported not to differ between species, but fine root biomass were magnitudes higher in Deschampsia (Andresen et al. 2009). This may indicate larger capacity for water uptake in Calluna in the deeper soil layers, but more information taking into account both above- and below-ground biomass distributions, as well as stomatal conductance, are needed to investigate these issues.

Night-time warming

The night-time warming generally had limited effects on the leaf surface temperature, but did under some conditions increase vegetation surface temperature for up to 3–5 h after sunrise, although normally much less (Mikkelsen et al. 2008). This could potentially induce short-term stimulation of photosynthesis via more optimal growth temperatures (Sage & Kubien 2007). However, these direct effects were not detected in this study, for various reasons. Too few measurements were conducted during the early daytime hours where the night-time warming effects persisted. Further, such effects would most likely not influence the overall treatment effects reported here as the sampling procedure with a large number of plots and the several hours of measuring time for a complete campaign would have constrained this variation to the experimental blocks, causing it not to be detected as a treatment effect in the anova. Instead, the potential effect of the night-time warming on photosynthesis must be indirect. Warming increased the amount of GDDs (Mikkelsen et al. 2008), and increased photosynthesis and photosynthetic capacity in the early season (Pn in May, and Pmax in May and June). This suggests that the warming treatment may have caused an earlier onset of plant growth or faster development of the photosynthetic machinery in the spring in accordance with other studies demonstrating earlier growth season start-up in response to warming (Harte & Shaw 1995; Menzel & Fabian 1999; Wan et al. 2005; Menzel et al. 2006). Later in the season, the maturation of photosynthetic capacity in the non-warmed treatments caught up and warming effects did not translate into significant effects on photosynthesis, except for marginal reductions of net photosynthesis in warmed plots in August. Furthermore, the night-time warming reduced SWC in most months, and reduced PWP in July and September, but only influenced net photosynthesis negatively in the dry midsummer (August). The reduced SWC indicates increased plant water consumption per ground area or increased soil evaporation in response to night-time warming, but apart from the midsummer, this reduction was not strong enough to induce a general reduction in photosynthesis. Our finding of a small response of photosynthesis to warming in this relatively dry site may be further supported by previous findings from passive night-time warming studies across a European gradient indicating that photosynthetic plant carbon uptake may be more responsive on wet sites compared to dry sites (Llorens et al. 2004; Peñuelas et al. 2007). Thus, the primary effect of night-time warming on photosynthesis in our study was associated with earlier seasonal onset and maturation of the photosynthetic capacity.

Synergistic effects of T × CO2 and antagonistic influence of D × CO2

Across a range of ecosystems, including forests, grasslands and heathlands, the interactive effects between warming and precipitation on net primary production have been predicted by modelling approaches to be positive when soil water availability is high, and negative when low (Luo et al. 2008); however, there are limited experimental data to support these model results. In a recent experiment where precipitation was increased in combination with warming by infrared heating, interactive effects were present after 3 years of experimentation (Shaw et al. 2002), but not after 5 years (Dukes et al. 2005). Until now, such effects have not been investigated using experimental drought via water exclusion and passive night-time warming, as in the CLIMAITE experimental set-up (Mikkelsen et al. 2008). Instead, a range of responses to experimental drought and warming have been tested in single-factor experiments (e.g. Beier et al. 2004; Schmidt et al. 2004; Peñuelas et al. 2007), and the response to warming was stimulatory on photosynthesis and biomass, especially during cold periods and sufficient water availability conditions (Llorens et al. 2004; Peñuelas et al. 2007; Prieto et al. 2009a,b), whereas experimental drought decreased photosynthesis and biomass (Prieto et al. 2009a,b). In our study, an antagonistic effect was seen in the combination of warming and experimental drought (T × D), and only in August. However, this did not influence photosynthesis and no other T × D interactive effects were found. This indicates that the combined effects of passive night-time warming and experimental drought are mainly additive, and thus short-term photosynthetic effects may be inferred from the findings based on the single-factor experiments.

Ecosystem responses to the combination of warming and elevated CO2 are complex; multifactor experiments do not provide clear and interpretable results on many processes in natural ecosystems (Norby & Lou 2004; Dermody 2006). The direct effects from the combination of warming and elevated CO2 potentially cause interactions (Morison & Lawlor 1999), and in general, elevated CO2 is expected to stimulate the net photosynthesis when combined with direct warming (Long 1991; Tjoelker, Oleksyn & Reich 1998; Morison & Lawlor 1999; Turnbull et al. 2004; Sage & Kubien 2007). Modelling predicts that elevated CO2 and warming in combination amplify net primary production (Luo et al. 2008). However, unexpected experimental results have shown that elevated CO2 can partly suppress the stimulatory effects of warming, precipitation and nitrogen on NPP after 2 years of experimentation (Shaw et al. 2002), but not in the following 3 year period (Dukes et al. 2005).

The night-time warming treatment in our study does not directly elevate the temperature of leaf surfaces in the daytime and thus differ from studies using IR-lamps or other techniques with active daytime direct warming. Thus, in line with this, no direct effects of warming on photosynthesis were observed in our study; the effects observed were most likely mediated via indirect effects on seasonality causing earlier maturity of photosynthetic tissues and higher photosynthesis in early season. Based on the photosynthetic stimulation by elevated CO2 in periods with high stomatal conductance, one might expect stimulation when elevated CO2 is combined with warming in early and late season where stomatal conductance was observed to be high. The synergistically reduced δ13C in the combination of warming and CO2 in the early and late season clearly indicate increased stomatal conductance (Farquhar et al. 1982, 1989; Ehleringer 1993), indeed did stimulate net photosynthesis (Fig. 6). This interpretation is supported by the photosynthetic responses in early and late season.

Despite favourable conditions for photosynthesis in early season (May) (high stomatal conductance, high intercellular CO2 concentration in the elevated CO2 treatment, a high photosynthetic capacity in the warming treatment), no interactions between elevated CO2 and warming were present on stomatal conductance and net photosynthesis. Therefore, it was primarily the extended period with improved photosynthetic metabolism in the warming treatment that caused the accumulated synergistic response that was reflected in the δ13C in the combination of warming and elevated CO2. Late in the growing season (October), the combination of warming and elevated CO2 synergistically stimulated photosynthesis, but this was caused by the combination of the higher intercellular CO2 concentration in elevated CO2 and the synergistically increased stomatal conductance in the combination of elevated CO2 and warming.

Although several studies have found elevated CO2 to alleviate the negative effects of drought on growth and improve plant water status (Volk et al. 2000; Leuzinger & Körner 2007; Robredo et al. 2007), the underlying photosynthetic responses are complex and the interactions are not well understood (e.g. Tschaplinski et al. 1996; Centritto, Lee & Jarvis 1999; Tezara et al. 2002; Dermody et al. 2007). In our study, the direction of response to drought and elevated CO2 was different, leading to complex reactions when combined. Mild to moderate drought in the early phase of the experimental drought (July) did not influence the photosynthetic parameters (Pn, gs or Pmax) in the DCO2 combination, which generally showed responses at a similar level as in elevated CO2. However, at peak drought (early August), experimental drought and elevated CO2 (D × CO2) antagonistically reduced Pn, gs and Pmax, but the response was dampened compared to single-factor responses. These responses were generally related to SWC with small and additive effects at small to moderate reductions in SWC, and significant responses and antagonistic (D × CO2) effects at peak drought. These antagonistic (D × CO2) effects suggest a pronounced capacity of elevated CO2 to mitigate the negative effects of drought on photosynthetic plant carbon uptake, which has been demonstrated by several studies (Morgan et al. 2004; Robredo et al. 2007), may not be experienced during extreme droughts.

Responses to the combination of warming, drought and elevated CO2

The combined treatment with warming, elevated CO2 and drought represents the predicted future climatic scenario for Denmark in 2075 (Mikkelsen et al. 2008; Boberg 2010; Christensen et al. 2010). Generally, the combination of all treatments (TDCO2) showed small responses compared to the ambient plots, and no T × D × CO2 interactions were detected. Hence, the effects of all climate change factors together in TDCO2 could be explained by the impact of the single factors (T, D, CO2) and the two-factor interactions (D × CO2, T × CO2, T × D). This is in agreement with findings from a wide range of processes in the same experiment, demonstrating a clear dominance of antagonistic effects over synergistic and simple additive effects, when treatments were combined across a large number of nitrogen cycle-related parameters (Larsen et al. 2010).

Immediate responses in photosynthesis parameters were seen; however, after only 1 year of treatment, the ecosystem is likely in a transition phase. At this time, the plants may not yet be influenced by the potential longer-term changes in resource supply, for example leading to progressive nutrient limitation in elevated CO2 (PNL, Lou et al. 2004; Reich et al. 2006) or long-term changes in growth and biomass allocation, which influence the plant water availability (Körner 2003, 2006). If these phenomena were to occur, they could either dampen the stimulatory effect on photosynthesis, for example via down-regulation under elevated CO2 in the case of PNL or increase photosynthesis through improved water relations. This demonstrates the importance of performing long-term multifactorial experiments to give better insight into these responses.

ACKNOWLEDGMENTS

The work was carried out in the CLIMAITE project financially supported by the Villum Kann Rasmussen Foundation. Support was also received from Air Liquide and DONG energy. The authors would like to thank Christel Barker, Esben Vedel Nielsen, Gosha Sylvester, Niels Bruun, Karna Heinsen and Karin Larsen who helped with field work, and leaf chemical analyses. Thanks to Sven Danbæk for his tireless efforts with the collection of the climatic data. Svend Danbæk, Poul T. Sørensen, Preben Jørgensen, Nina Wiese and Andreas Fernquist are thanked for their timeless effort in keeping the site running. Two anonymous reviewers are thanked for providing constructive feedback to this paper.

Ancillary

Advertisement