Interactive Effect of Elevated CO2 and Reduced Summer Precipitation on Photosynthesis Is Species-Specific: The Case Study with Soil-Planted Norway Spruce and Sessile Oak in a Mountainous Forest Plot

We investigated how reduced summer precipitation modifies photosynthetic responses of two model tree species—coniferous Norway spruce and broadleaved sessile oak—to changes in atmospheric CO2 concentration. Saplings were grown under mountainous conditions for two growing seasons at ambient (400 μmol CO2 mol–1) and elevated (700 μmol CO2 mol–1) CO2 concentration. Half were not exposed to precipitation during the summer (June–August). After two seasons of cultivation under modified conditions, basic photosynthetic characteristics including light-saturated rate of CO2 assimilation (Amax), stomatal conductance (GSmax), and water use efficiency (WUE) were measured under their growth CO2 concentrations together with in vivo carboxylation rate (VC) and electron transport rate (J) derived from CO2-response curves at saturating light. An increase in Amax under elevated CO2 was observed in oak saplings, whereas it remained unchanged or slightly declined in Norway spruce, indicating a down-regulation of photosynthesis. Such acclimation was associated with an acclimation of both J and VC. Both species had increased WUE under elevated CO2 although, in well-watered oaks, WUE remained unchanged. Significant interactive effects of tree species, CO2 concentration, and water availability on gas-exchange parameters (Amax, GSmax, WUE) were observed, while there was no effect on biochemical (VC, J) and chlorophyll fluorescence parameters. The assimilation capacity (Asat; CO2 assimilation rate at saturating light intensity and CO2 concentration) was substantially reduced in spruce under the combined conditions of water deficiency and elevated CO2, but not in oak. In addition, the stimulatory effect of elevated CO2 on Amax persisted in oak, but completely diminished in water-limited spruce saplings. Our results suggest a strong species-specific response of trees to reduced summer precipitation under future conditions of elevated CO2 and a limited compensatory effect of elevated CO2 on CO2 uptake under water-limited conditions in coniferous spruce.


Introduction
Forest ecosystems play dominant roles in the ecology and evolution of life on the Earth. Forests have a critical function in regulating the global climate through their effects on the biogeochemical cycles, especially the cycles of carbon and water [1]. It is well known that, Figure 1. Photograph of cultivation domes with adjustable lamellar windows (top panel) and schematic drawing of the experiment layout (bottom panel). Each dome was divided into six blocks. Individual blocks (repetition; n = 3) are bounded by a thin black line. Open blocks = well-watered treatment (gular watering of 20 mm per week during vegetation period); colored blocks = waterlimited treatment (no irrigation for 5 weeks before the measurements); AC = ambient CO 2 concentration (~400 µmol CO 2 mol -1 ); EC = elevated CO 2 concentration (~700 µmol CO 2 mol -1 ).
Each dome (10 × 10 m in size) was split into six blocks ( Figure 1). Each block (approximately 16.5 m 2 ) contained 12 saplings of sessile oak and 12 saplings of Norway spruce. Three random blocks (replications; n = 3) were regularly irrigated (WW) while the remaining 3 blocks were left without irrigation during the period of July-August (WL) in two consecutive years. WW plants received regular watering of 20 mm per week during vegetation period, while WL plants had no irrigation for 5 weeks before the measurements. To prevent soil water movement, the blocks were separated by a PVC board (thickness of 5 mm) placed in the soil up to the native bedrock (depth of ca 0.50 m). An ML3 Theta Probe sensor (Delta-T Devices, UK) was used to measure soil water content in the depths of 5-10 cm and 30-40 cm. In topsoil layer, WL treatment led to a decrease of soil water content from 23.5 to 13%, while it remained high (22.5-23.5%) under WW treatment. However, soil water content was the same, 17.5%, under the both water treatments in the deeper soil layer.

Physiological Measurements
In total, 36 oaks and 36 spruces were measured at the end of August. Leaf-and shoot-level measurements were made on three representative saplings from each block (replication). The average from three sapling measurements per block was used for the consequent statistical analyses. Current-year spruce shoots were examined. All measurements were conducted between 10:00-15:00 CET on intact fully-expanded leaves. A Scholander type of pressure chamber was used to measure leaf water potential (Ψ) at pre-dawn (05:00) and noon (12:00).

Gas-Exchange Measurements
Basic photosynthetic characteristics-CO 2 assimilation rate (A), intercellular [CO 2 ] (Ci), stomatal conductance (G S ), and transpiration rate (E)-were measured using an open gas-exchange system Li-6400 (LI-COR Biosciences, USA) equipped with an LED light source 6400-02B. The A-Ci response curves were determined at saturating light intensity (1400 µmol m −2 s −1 ) and starting at growth [CO 2 ], i.e., at 400 µmol CO 2 mol -1 for AC and 700 µmol CO 2 mol -1 for EC plants. In this way, instantaneous light-saturated rates of CO 2 assimilation (A max ) and stomatal conductance (G Smax ) at growth [CO 2 ] were determined. Then, the following set of [CO 2 ] in the assimilation chamber was applied: 1500, 700, 400, 250, 150, 90, and 50 µmol CO 2 mol -1 . Plants were acclimated approximately 10 min to each [CO 2 ]. Such measuring protocol enabled us to derive CO 2 assimilation capacity (A sat ), CO 2 assimilation rate at saturating light intensity and saturating CO 2 concentration (1500 µmol CO 2 mol −1 ), as well as the in vivo Rubisco carboxylation rate (V C ) and electron transport (J) at 25 • C. See [24] for the calculation details. Moreover, water use efficiency was calculated under saturating light conditions (WUE, defined as the ratio between A max and E max ). During all measurements, the leaf temperature was kept at 25 • C (±0.5 • C) using the Peltier thermo-electric cooler mounted on the assimilation chamber. Relative air humidity was maintained during the measurement at 50% (±1%), thus keeping vapor pressure deficit (VPD) at 1.1-1.5 kPa.

Chlorophyll Fluorescence
The efficiency of primary photochemical reactions was determined by chlorophyll fluorescence (Chl-F) following the protocols described in details by Urban et al. [25]. In brief, the emission of the Chl-F signal at the red band (near 690 nm) was measured by a fluorometer PAM-2500 (H. Walz, Effeltrich, Germany). Night-time Chl-F measurement was carried out to estimate maximum quantum yield of photosystem (PS) II photochemistry (Fv/Fm). Leaves adapted to saturating light intensity (1400 µmol m −2 s −1 ) were used to assess an actual quantum yield of PSII photochemistry (Φ PSII ) and thermal energy dissipation (D) according to Demmig-Adams et al. [26]. All measurements were done on intact leaves in situ at growth [CO 2 ].

Statistical Data Processing
Three-way analysis of variance (ANOVA) was used to analyze effects of tree species (TS), CO 2 concentration ([CO 2 ]), water availability (WA), and their interactions (×) on physiological parameters. Tukey's HSD post-hoc test was used to analyze significant differences between means at p ≤ 0.05. Statistical analyses were conducted using the software STATISTICA 12 (StatSoft, Tulsa, CA, USA). The bar graphs representing means with standard deviations were developed with the software SigmaPlot 11.0 (Systat Software, San Jose, CA, USA).

Growth Conditions
Pre-dawn and noontime leaf water potentials were significantly influenced by tree species (TS) and water availability (WA), but CO 2 treatment had significant effect only on Ψ pre-dawn (Table 1). No interactive effects of these factors on Ψ values were observed. Exclusion of precipitation led to mild decreases in leaf Ψ at pre-dawn and particularly noon hours. However, the Ψ noon values between −1.0 and −1.3 MPa measured in this experiment for WL plants are not very severe. On average, the noon Ψ decrease amounted up to 22% in oak and up to 25% in spruce. Although there were mostly no significant differences between AC and EC counterparts, lower (more negative) Ψ values were observed in EC plants ( Figure 2). Table 1. Summary of significance levels (p-values of the three-way ANOVA) for the factors of trees species (TS), CO 2 concentration ([CO 2 ]), water availability (WA), and their interactions (×) on physiological parameters (Ψ pre-dawn -pre-dawn leaf water potential, Ψ noon -noon leaf water potential, A max -light-saturated CO 2 assimilation rate, G Smax -light-saturated stomatal conductance, E max -light-saturated transpiration rate, and WUE-water use efficiency, V C -in vivo Rubisco carboxylation rate, J-in vivo electron transport rate, Φ PSII -actual quantum yield of PSII photochemistry under saturating light intensity, D-thermal energy dissipation under saturating light intensity). Bold p-values indicate significant effect at p ≤ 0.05.  2. Effect of CO 2 concentration and water availability on pre-dawn leaf water potential (Ψ pre-dawn ) and noon leaf water potential (Ψ noon ) in sessile oak (left) and Norway spruce (right). Means (columns) and standard deviations (error bars) are presented (n = 3). Different letters indicate significant differences (p ≤ 0.05) between means based on Tukey's ANOVA post-hoc test. AC-ambient CO 2 concentration (~400 µmol CO 2 mol -1 ), EC-elevated CO 2 concentration (~700 µmol CO 2 mol-1), WW-well-watered (regular watering of 20 mm per week during vegetation period), WL-water-limited (no irrigation for 5 weeks before the measurements).

Assimilation Capacity (A sat ) and Light-Saturated CO 2 Assimilation Rate (A max )
Assimilation capacity (A sat ) was significantly influenced by tree species (TS) and water availability (WA), but not by [CO 2 ] (Table 1). Generally, we found higher assimilation capacity in broadleaved oak compared to coniferous spruce ( Figure 3A,B). WL treatment led to an increase of A sat in oak saplings treated under both [CO 2 ] conditions. However, the assimilation capacity was substantially reduced in spruce saplings under the combined conditions of WL and EC. Strong interactive effect of all factors investigated (TS × [CO 2 ] × WA) on A sat was also found ( Table 1). Figure 3. Effect of CO 2 concentration and water availability on CO 2 assimilation capacity (A sat ; (A,B)), light-saturated CO 2 assimilation rate at growth [CO 2 ] (A max ; (C,D)) and light-saturated stomatal conductance (G Smax ; (E,F)) in sessile oak (left) and Norway spruce (right). The measurements were done during the second growing season of cultivation under manipulated conditions: AC-ambient CO 2 concentration (~400 µmol CO 2 mol -1 ), EC-elevated CO 2 concentration (~700 µmol CO 2 mol -1 ), WW-well-watered (regular watering of 20 mm per week during vegetation period), WL-waterlimited (no irrigation for 5 weeks before the measurements). Means (columns) and standard deviations (error bars) are presented (n = 3). Different letters indicate significant differences (p ≤ 0.05) between means based on Tukey's ANOVA post-hoc test.
Generally, tree species and [CO 2 ] had significant effects on A max , while the effect of water availability was statistically non-significant (Table 1). A higher stimulatory effect of EC on A max was found in broadleaved oak compared to coniferous spruce. EC conditions tended to increase A max values in oak saplings across all water treatments; however, this effect fell short of being statistically significant ( Figure 3C). Noticeably, the stimulatory effect of EC on A max completely diminished in WL-treated spruce saplings ( Figure 3D). Analysis of variance revealed significant TS × WA on A max , while no significant interactive effect of [CO 2 ] and WA was evident (Table 1).

Light-Saturated Stomatal Conductance (G Smax ) and Water Use Efficiency (WUE)
Generally, the only significant effect on G Smax and E max (light-saturated rate of transpiration) was the tree species (TS); rather surprisingly, [CO 2 ] and water availability themselves had no effect (Table 1). There were, however, some interactions. While G Smax decreased under WL conditions in spruce saplings at both [CO 2 ] and oak saplings at EC, an opposite response to WL was observed in broadleaved oak saplings under AC conditions. EC reduced G Smax in spruce irrespective of water availability, but it was true for oak only under WL conditions ( Figure 3E,F). The effect of EC was, however, not statistically significant at any water availability. Since the G Smax and E max are tightly interconnected, the same findings were noted for E max ( Figure 4A,B). WUE, defined as A max /E max ratio, was significantly influenced by all factors investigated-tree species, [CO 2 ], and water availability (Table 1). Moreover, significant interactive effects of tree species and [CO 2 ], [CO 2 ] and water availability, as well as tree species, [CO 2 ], and water availability on WUE were found. High [CO 2 ] substantially increased WUE in all saplings under both water regimes, with the exception of the well-watered broadleaved oak ( Figure 4C,D). . Effect of CO 2 concentration and water availability on transpiration rate (E max ; (A,B)) and water use efficiency (WUE; (C,D)) at saturating light intensity in sessile oak (left) and Norway spruce (right) saplings. The measurements were made during the second growing season of cultivation under manipulated conditions: AC-ambient CO 2 concentration (~400 µmol CO 2 mol -1 ), EC-elevated CO 2 concentration (~700 µmol CO 2 mol -1 ), WW-well-watered (regular watering of 20 mm per week during vegetation period), WL-water-limited (no irrigation for 5 weeks before the measurements). Means (columns) and standard deviations (error bars) are presented (n = 3). Different letters indicate significant differences (p ≤ 0.05) between means based on Tukey's ANOVA post-hoc test.

Biochemical Parameters Derived from A-Ci Response Curves
In vivo rates of Rubisco carboxylation (V C ) and electron transport (J) were derived from A-Ci response curves measured at saturating light intensity and a temperature of  (Table 1). Moreover, significant interactive effects of tree species and water availability on V C and J were found. In contrast, no interactive effects of [CO 2 ] with other investigated factors were observed (Table 1).
WL treatment increased V C and J values under both [CO 2 ] conditions in sessile oak saplings, but not in Norway spruce ( Figure 5). EC treatment tended to reduce V C and J in oak as well as spruce saplings under both water regimes. These changes, however, are not statistically significant (p > 0.05). The J/V C ratio for EC-WW (4.24) and AC-WL (6.42) were the lowest and highest, respectively, for sessile oak saplings, whereas EC-WL (3.98) and AC-WW (4.73) were the lowest and highest, respectively, for Norway spruce. Values of V C and J were linearly correlated in both tree species irrespective of [CO 2 ] and water availability (Figure 6), suggesting a relatively constant balance between carboxylation and regeneration of RuBP. Both parameters were derived from A-Ci response curves estimated at saturating light intensity and temperature of 25 • C. The measurements were made during the second growing season of cultivation under manipulated conditions: AC-ambient CO 2 concentration (~400 µmol CO 2 mol -1 ), EC-elevated CO 2 concentration (~700 µmol CO 2 mol -1 ), WW-well-watered (regular watering of 20 mm per week during vegetation period), WL-water-limited (no irrigation for 5 weeks before the measurements). Means (columns) and standard deviations (error bars) are presented (n = 3). Different letters indicate significant differences (p ≤ 0.05) between means based on Tukey's ANOVA post-hoc test.

Chl-F Parameters
Maximum quantum yield of PSII photochemistry (Fv/Fm) ranged between 0.790 and 0.810 (data not shown). The Fv/Fm ratio was not influenced by any investigated factor and/or their interactions. Under saturating light conditions, species-specific effects of [CO 2 ] and water availability on Φ PSII and D were observed (see statistically significant interactions TS × [CO 2 ] and TS × WA in Table 1). While Φ PSII values increased with increasing [CO 2 ] and reduced water availability in oak, opposite trends were found in spruce ( Figure 7A,B). These differences were, however, statistically not significant (p > 0.05). In contrast to Φ PSII changes, EC-induced differences in D were mostly statistically significant (p ≤ 0.05; Figure 7C,D). Significant interactive effect of TS × [CO 2 ] × WA on investigated fluorescence parameters was not observed.
[CO2] and water availability on ΦPSII and D were observed (see statistically significa interactions TS × [CO2] and TS × WA in Table 1). While ΦPSII values increased w increasing [CO2] and reduced water availability in oak, opposite trends were found spruce ( Figure 7A,B). These differences were, however, statistically not significant (p 0.05). In contrast to ΦPSII changes, EC-induced differences in D were mostly statistica significant (p ≤ 0.05; Figure 7C,D). Significant interactive effect of TS × [CO2] × WA investigated fluorescence parameters was not observed.

Evaluation of Experimental Conditions
The experiment was carried out on a mountainous site receiving usually a lar amount of precipitation (100-150 mm) during the summer months (July-August). T regular watering (20 mm per week) was designed to ensure a common total amount natural precipitations, while water-limited treatment allowed no moisture into the soil f two months. Such experimental design agrees with expectations of increased frequen duration, and severity of summer drought periods in the central European region [2 Applied water regimes led to the desiccation of topsoil layers (5-10 cm), while the s water content in deeper soil layers (30-40 cm) was unaffected by exclusion precipitation. Although the experimental conditions led to a reduction of leaf wa Figure 7. Effect of CO 2 concentration and water availability on actual quantum yield of PSII photochemistry (Φ PSII; upper panels) and thermal energy dissipation ((D); bottom panels) under saturating light intensity in sessile oak (left) and Norway spruce (right). Means (columns) and standard deviations (error bars) are presented (n = 3). Different letters indicate significant differences (p ≤ 0.05) between means based on Tukey's ANOVA post-hoc test. AC-ambient CO 2 concentration (~400 µmol CO 2 mol -1 ), EC-elevated CO 2 concentration (~700 µmol CO 2 mol -1 ), WW-well-watered (regular watering of 20 mm per week during vegetation period), WL-water-limited (no irrigation for 5 weeks before the measurements).

Evaluation of Experimental Conditions
The experiment was carried out on a mountainous site receiving usually a large amount of precipitation (100-150 mm) during the summer months (July-August). The regular watering (20 mm per week) was designed to ensure a common total amount of natural precipitations, while water-limited treatment allowed no moisture into the soil for two months. Such experimental design agrees with expectations of increased frequency, duration, and severity of summer drought periods in the central European region [27]. Applied water regimes led to the desiccation of topsoil layers (5-10 cm), while the soil water content in deeper soil layers (30-40 cm) was unaffected by exclusion of precipitation. Although the experimental conditions led to a reduction of leaf water potentials in both tree species studied (Figure 2), changes in water potentials indicate a rather mild drought stress in both tree species investigated and particularly in oak. Species-specific differences are likely caused by the different architecture of the root system typical for these two species. We assume that the gradual desiccation of topsoil layer influenced particularly shallow-rooting spruce, while deep-rooting oak saplings remained less affected by the exclusion of precipitation. This assumption is supported by higher (less negative) Ψ pre-dawn values found in oak then spruce saplings (Table 1).

Species Specificity in Response to Elevated CO 2 and Water Availability
Elevated [CO 2 ] induces many adaptive responses in plants, including reduced stomatal conductance, increased WUE, and stimulated root growth as well as accumulation of osmolytes and phyto-hormones [28][29][30][31]. Since these responses can increase resistance of plants to drought, enhanced resistance of future plants grown under EC conditions is likely. To test this hypothesis, tree species with distinct drought sensitivity and carbon sink capacity were selected. While oaks species are known as drought-tolerant species [17,32], spruce is significantly more drought-sensitive [15,16,33]. Higher sensitivity of spruce to water-limited conditions was indeed evident in the current study by the more pronounced reduction of leaf water potential (Figure 2), CO 2 assimilation rate, transpiration, and increased WUE (Figures 3 and 4). In general, measured values of physiological parameters are not so much reduced as observed in other studies on related tree species [15][16][17][18]34]. Exclusion of summer precipitation resulted even in slight increases of V C and J values in oak saplings. The stimulatory effect of mild stress conditions was observed also in other plant species [33,35].
Our results further show significant effect of tree species ( Table 1) on most of the gas-exchange, chlorophyll fluorescence, and biochemical parameters investigated. In accordance with previous studies [6,25,28,[35][36][37], we have found higher CO 2 assimilation rate and capacity in well-watered broadleaved than coniferous trees under ambient [CO 2 ] (Figure 3). Such findings are in agreement with higher values of V C , J, and G Smax in broadleaved saplings (Figures 3 and 5) and document higher strength of carbon sinks in oak as compared to spruce.
Analysis of variance revealed the significant interactive effects of tree species and water availability on almost all parameters investigated (Table 1), while the significant interactive effects of tree species and [CO 2 ] were found only for A sat , WUE, and both fluorescence parameters. Similarly, TS × [CO 2 ] × WA had significant effects only on the assimilation capacity and parameters associated with WUE. Generally, well-watered oak saplings had higher stimulation of A max by EC as compared to spruce saplings.
Our experiment showed that photosynthesis of Q. petraea was stimulated by elevated [CO 2 ] (an increase of the A max ), but the stimulatory effect of elevated [CO 2 ] diminished in P. abies under the WL conditions. Others have also shown that photosynthesis of sessile oak could be stimulated when subjected to mild drought (simulated or natural) and CO 2 [38]. Nevertheless, most of physiological processes associated with photosynthetic carbon uptake tend to decline under severe drought conditions irrespective of [CO 2 ] treatment [18,34]. As expected, water use efficiency improved too, especially in spruce saplings. However, enhancement of WUE in oak saplings was observable only under the combined effect of WL and elevated [CO 2 ]. Such variability could be explained by a stronger CO 2 effect on stomatal closure in broadleaved than in coniferous tree species [28,39].

Fluorescence Parameters
Primary photochemical reactions measured as a quantum yield of PSII photochemistry (Φ PSII ) and thermal energy dissipation (D) show substantial species-specific response to a combined treatment of EC and WL. While Φ PSII tended to increase in oak under combined EC and WL conditions, substantial decrease was found in spruce ( Figure 7). Such findings are consistent with species-specific responses in G Smax . Higher availability of intercellular [CO 2 ] and faster A max in oak represents higher sink strength for primary photochemical products ATP and NADPH, consequently leading to a higher rate of electron transport measured as Φ PSII [40]. In contrast, reduced carbon sink capacity in spruce saplings resulted in increased demands for alternative pathways of absorbed radiation energy, including non-radiative thermal dissipation (D). Such phenomena have been observed in many stressed plant species having closed stomata [12] and/or having reduced rate of photosynthetic CO 2 uptake [41]. Reduced efficiency of photochemical reactions may, vice versa, lead to a deepening of photosynthetic down-regulation as reported in plants exposed to combined heat and drought stress [42] or elevated [CO 2 ] when the stomata remain closed [5,43].

Gas-Exchange Parameters
We have found evidence of substantial photosynthetic down-regulation in spruce saplings exposed to combined impact of EC and WL, but not in oak which probably has access to water in the lower layers of the soil. Down-regulation was associated with decline in both CO 2 assimilation capacity (A sat ) as well as light-saturated rate of CO 2 assimilation (A max ; Figure 3). These findings are in agreement with reduced values of J and V C ( Figure 5) reflecting the reduction of RuBP-and Rubisco-limited rate of photosynthesis, respectively [44]. Such phenomena are often observable in sink-limited coniferous trees, while EC usually stimulates CO 2 uptake in deciduous trees over long time periods [13,45].
Here we have shown, in addition, that down-regulation of photosynthesis is accelerated by insufficient water availability. Such interactive effects have received only limited attention so far [4,46]. Reduced J is in accordance with reduced Φ PSII value as described above. Suppressed V C in spruce saplings has been earlier linked to a decline in the total amount of Rubisco enzyme, and/or an increase of the proportion of inactive and active Rubisco forms [44]. Generally, biochemical adjustment, including phosphorus redistribution, nitrogen dilution, and/or inhibition of photosynthetic genes that are associated with photosynthetic down-regulation, are induced by an accumulation of non-structural assimilates in leaves [13,29].
Species-specific responses of trees to combined effects of elevated [CO 2 ] and limited water availability could arise from their biological and environmental diversity [39], different feedback mechanisms, such as their ability to export assimilates from leaves to roots and avoiding photosynthetic down-regulation [47]. Moreover, plant metabolic activities can play a key role in the modulation of photosynthetic down-regulation through an accumulation of certain compounds under stress conditions. Drought, among others, triggers increases in contents of some monosaccharides, disaccharides, and their derivatives. These sugars serve as osmolytes, anti-oxidative protectants, and provide energy during drought stress to minimize negative impacts of drought on plants [31]. Accordingly, we assume that an over-accumulation of these assimilates protecting plants against stress conditions may consequently lead to a transient down-regulation of photosynthesis under elevated [CO 2 ].
On the other hand, long-term exposure to elevated [CO 2 ] is known to stimulate root growth [30,48], rooting depth, and proportion of fine roots [49]. Development of the root system could lead to a better acquisition of water and mineral nutrients having a positive effect on stomatal conductance and photosynthesis [50,51]. Accordingly, such morphological adjustment results in a positive effect of elevated [CO 2 ] on photosynthesis in arid regions [4,52], and/or periods of insufficient water availability [13]. In accordance with these findings, we observed an increase of WUE under elevated [CO 2 ] in deep-rooting oak only under WL conditions, while it remained unchanged in well-watered oaks ( Figure 4). In contrast, [CO 2 ]-enhanced WUE was found in shallow-rooting spruce irrespective of water regime. Similar species-specific responses were reported by Picon et al. [17] in drought-tolerant Q. petraea and drought-avoiding Pinus pinaster.

Conclusions
Limited research has explored responses of mountainous trees to future climate conditions. Accordingly, we have investigated species-specific responses of two widespread tree species of the temperate zone-Norway spruce and sessile oak-exposed to a combined impact of elevated [CO 2 ] and excluded summer precipitation. The experimental setup resulted in a desiccation of the topsoil layers, but not deeper ones. Limited water availability influenced particularly shallow-rooting spruce, while deep-rooting oak saplings remained less affected as documented by higher Ψ pre-dawn values. In general, changes in leaf water potential revealed rather mild drought stress. Under water-limited conditions, down-regulation of photosynthesis by elevated [CO 2 ] was observed in droughtsensitive coniferous spruce, but it remained stimulated in drought-tolerant deciduous oak. Down-regulation was associated with reduction of RuBP-as well as Rubisco-limited rate of photosynthesis. We conclude that elevated [CO 2 ] increases resistance to reduced summer