Plastic response of four maritime pine (Pinus pinaster Aiton) families to controlled soil water deficit

Abstract

Key message

Separating the internal (ontogenetic) and external (environmental) components of maritime pine development during controlled soil water deficit helps to highlight the plastic response. The adjusted measurements reveal significant differences between families for their plastic response for several physiology and growth traits.

Context

Soil water deficit is and will be a growing problem in some regions. Pinus pinaster Ait. is a species of commercial interest and is recognized as a drought-avoiding species. It is thus of interest to evaluate the adaptation potential of P. pinaster to soil water deficit.

Aims

This paper aims to estimate the plastic response to the variation of water availability at the family level (half-sibs).

Methods

Two-year-old P. pinaster cuttings from four families were submitted during 6 weeks to two contrasting watering regimes. The experiment started in April 2011 shortly after sprouting. The photosynthesis and stomatal conductance to water vapor were measured on 1-year-old needles. Intrinsic water-use efficiency was calculated as the ratio of photosynthesis to stomatal conductance. Radial growth, length of terminal shoot, and total height were also measured. The ontogenetic component of tree development was estimated on the well-watered trees for all the traits. Then, this development effect was eliminated from the data collected on the trees submitted to the soil water deficit in order to keep only the effect of this soil water deficit.

Results

After 6 weeks of reduced watering, the value of all adjusted traits decreased. An average plastic response to the variation of water availability was found to be significant and variable at the family level for the six adjusted variables.

Conclusion

These results suggest that there is genetic variation of phenotypic plasticity to drought in P. pinaster for several traits, including stomatal conductance, which appears to be a promising variable for future selection for resistance to drought.

1 Introduction

With climate warming, plants face more frequently longer periods of increased water demand and decreased water availability. In this context, a better understanding of the processes involved in tree response to water availability is required (Pachauri et al. 2015). Maritime pine (Pinus pinaster Ait.) is recognized as a drought-avoiding species (Granier and Loustau 1994; Picon et al. 1996). Such species are often isohydric plants (Tardieu 1998), with a high stomatal sensitivity to soil water deficit. In response to mild water stress, these plants close their stomata to regulate water flux, and reduce transpiration and soil water uptake. This is followed by a reduction of stomatal conductance to water vapor (Schulze et al. 1987; Picon et al. 1996; Fernández et al. 1999; Fernández et al. 2000; Flexas 2002; Sánchez-Gómez et al. 2010) and of predawn leaf water potential (Aussenac and Granier 1978; Fernández et al. 2000; Fernández et al. 1999; Picon-Cochard and Guehl 1999). A restriction of the CO₂ transfers from the atmosphere to the chloroplast leads to decreased photosynthesis (Chaves 1991; Cornic 2000; Farquhar and Sharkey 1982; Lawlor and Cornic 2002).

Selecting fast growing water-saving genotypes is an important objective for P. pinaster breeders (de la Mata et al. de la Mata et al. 2014, 2012; Guehl et al. 1994; Plomion et al. 2016). Water-use efficiency (WUE) is an integrative trait related to plant water economy. WUE is defined at the whole-plant level as the ratio between biomass production and cumulative water losses by transpiration. While WUE is difficult to measure at the whole plant level, intrinsic water-use efficiency (Wi), defined as the ratio between net CO₂ assimilation and stomatal conductance to water vapor, is easily measured at the leaf level. QTL were detected for Wi in maritime pine in different water regimes by Brendel et al. (2002) and de Miguel et al. (2014). According to Marguerit et al. (2014), the lack of genetic link between WUE and growth allows simultaneous improvement of both traits. Our objective in this article is to contribute to the evaluation of the adaptation potential of P. pinaster to water stress. The adaptation potential depends, on one hand, on the genetic determinism (a long-term response at population level) and, on the other hand, on the phenotypic plasticity (a shorter-term response at individual level) of adaptive traits (Schlichting 1986). According to DeWitt and Scheiner (2004), the phenotypic plasticity is the ability of a genotype to change its phenotype in response to changes in the environment. This variation can be described as a norm of reaction (Ghalambor et al. 2007). Environment varies in space and time; hence, phenotypic plasticity can be considered both against space-related and time-related environmental variation (Scheiner 2013). Space-related variation is especially relevant for multi-trial common garden experiments and for controlled-conditions experiments with genetic entities and treatments. In consequence, phenotypic plasticity is estimated either with vegetative copies of a genotype distributed across spatially distinct environments, or with a single genotype observed across time-related environmental variation. Each dimension corresponds to a type of plastic response. In this article, we estimate the space-related phenotypic plasticity between the treatments of a controlled-condition experiment: we call it between-treatment plasticity. We call the effect of temporal environmental changes time-related plasticity. The genetic entities in our study are half-sib families where the individuals are genetically related genotypes. On rigorous application of the phenotypic plasticity definition of DeWitt and Scheiner (2004), we called average plastic response the plasticity studied at the general level and family plastic response the plasticity studied with norms of reaction at the family level. The average plastic response is the phenotypic response considering the pooled sample, while the family plastic response is the average phenotypic response of each family. Significant genotype × treatment and genotype × site interactions are evidences of significant genetic variation for space-related (between-treatment, between-site) phenotypic plasticity.

In the literature, the phenotypic plasticity of P. pinaster to water stress was studied for different types of genetic entities and traits. At the population level, P. pinaster presented significant phenotypic changes with water stress on biomass allocation (Aranda et al. 2010; Chambel et al. 2007) and carbon isotope discrimination (as a proxy of Wi) (Aranda et al. 2010; Corcuera et al. 2012). This plasticity was also observed on the molecular machinery involved in wood formation (Paiva et al. 2008) and for hydraulic traits such as xylem specific conductivity (Corcuera et al. 2011) and cavitation resistance (Corcuera et al. 2011; Lamy et al. 2014).

In some studies, no genetic difference was observed between populations of P. pinaster in their phenotypic plasticity. It was the case for plasticity to water stress (population × water stress), biomass allocation (Aranda et al. 2010; de la Mata et al. 2014; Sánchez-Gómez et al. 2010), height growth (Chambel et al. 2007; Corcuera et al. 2010), relative height growth rate (Sánchez-Gómez et al. 2010), and cavitation resistance (Lamy et al. 2014). But genetic variation of plasticity were found in other studies for biomass-related variables (Chambel et al. 2007), height growth (de la Mata et al. 2014; Lamy et al. 2014), cavitation resistance (Corcuera et al. 2011), and carbon isotope discrimination (Aranda et al. 2010; Corcuera et al. 2012, 2010).

Few studies were conducted at the family level on plastic response to soil water deficit for early growth and physiological traits. Significant differences were observed between P. pinaster families for growth (Corcuera et al. 2010; Fernández et al. 2006), biomass and Wi based on gas exchange measurements (Fernández et al. 2006), and δ¹³C (Corcuera et al. 2010). An average plastic response to drought was found at the family level for height (Corcuera et al. 2010; Fernández et al. 2006) and for δ¹³C by Corcuera et al. (2010). Nevertheless, no significant family × treatment interaction was found for δ¹³C and gas exchange (A, gs, Wi) by Fernández et al. (2006). At the clone level, a plastic response to drought was detected for P. pinaster for gs and Wi (de Miguel et al. 2012). Hence, according to the sample, the phenotypic trait, and the environmental variable, P. pinaster plastic response to water stress was not always found to be genetically variable.

Our aim is to contribute to the study of the family variation of the early-growing season phenotypic plasticity of young P. pinaster trees submitted to variable watering regimes in controlled conditions for several traits, including CO₂ assimilation, stomatal conductance, and growth.

During the early growing season, the ontogenetic development follows and partly overlaps the recovery of photosynthesis from winter inhibition. These internal effects coincide and are confused with the response to environmental variation: according to the trait, such response can be underestimated or overestimated (Maseda and Fernández 2006; Egea et al. 2011). Ignoring them may lead to biased results and misinterpretation. This could explain part of the diverging published results. In this article, we separated the internal ontogeny and recovery effects from the response to environmental variation. We studied the variation with time and between families of CO₂ assimilation (A), stomatal conductance (gs), Wi and of the growth traits (total height (height), length of terminal shoot (shoot), and diameter at root collar (diam) between well-watered and water-stressed P. pinaster. Then, we estimated the phenotypic plasticity (average and family plastic response) of the same traits to between-treatment and time-related variation of soil water deficit. Taking into account the combined recovery and ontogenetic effects as one, single internal effect was a key step in investigating the between-family variation for phenotypic plasticity.

2 Materials and methods

2.1 Experimental design, plant material, and water stress application

The experiment was conducted in 2011 in an automatically ventilated greenhouse for overheat control at INRA Val de Loire, Orléans, France. The plant material were P. pinaster cuttings of October 2009 coming from a seed orchard from Monfero, Galicia, Spain, potted in early February 2011 in 10-L cylindrical pots filled with compost (Orga Agrumes et Rosiers de Fertil’Aquitaine®). Each family of 16 trees was produced at TRAGSA, Maceda, Spain, by vegetative propagation of half-sibs. In the standard procedure, TRAGSA mixes the propagated half-sibs of each family, retaining their family identity but not their individual clonal identity. Hence, each family is a mixture of genetically related (half-sibs) clones of unknown identity. The experimental design is formed of two complete replicates with eight randomly distributed families, corresponding to two treatments. After 2 months at field capacity, the trees were subjected to two watering treatments during 6 weeks. At the beginning of the experiment, from April 4 to 8, the trees in both treatments were watered to field capacity every 2 days. The water status of all pots was adjusted daily by weighing 10 randomly chosen pots. The pots were supplemented with mineral elements (1 g l⁻¹ of 18:12:18 NPK) using hydro soluble fertilizer. Fertilization was applied until the beginning of the water stress period then was stopped in both treatments. Then, on April 8, watering was stopped on the water-stressed (WS) treatment. Until the end of the experiment (May 20), half of the trees were well-watered (WW treatment) and maintained to field capacity, while the other half was water-stressed (WS treatment) till 40% of field capacity by stopping watering. The amount of water to maintain the 40% water status (since May 9) was adjusted by daily weighing of 10 randomly chosen pots. Volumetric soil water content (SWC; %) was measured each day for each of the 64 trees during all the experiment with a soil moisture sensor (ThetaProbe, type ML2x, Delta T, Cambridge, UK). The water-stressed trees were at 60% of field capacity on April 18 and reached 40% of field capacity on May 9. We maintained a 40% field capacity until May 20. Then, the response to water stress was measured on 32 individuals per treatment and 16 individuals per family (four families).

The greenhouse evaporative cooling automatically regulated the temperature and humidity during the experiment period, in order to avoid heat peaks. Average day/night air temperature and hygrometry in the greenhouse were 17.9/10.8 °C and 65.3/86.5%, respectively, during the experiment and were found to be stable during the 6-week time lag of the experiment (variation of air temperature and hygrometry in the greenhouse during the experiment shown in Fig. 3 in the supplementary data).

2.2 Water relations

In addition to volumetric SWC (%), needle predawn water potential (Ψb, − MPa) was measured with a Scholander pressure chamber (Scholander et al. 1965) at the end of week 6 on the well-watered (n = 4) and water-stressed trees (n = 8).

2.3 Growth traits

The total height (height, cm), diameter at root collar (5 cm, diam, cm), and length of terminal shoot (shoot, cm, issued from the annual bud) from each cutting were measured each week during the experiment at a 1-mm precision.

2.4 Gas exchange measurements

Net CO₂ assimilation rate (A, μmol CO₂ m⁻² s⁻¹) and stomatal conductance to water vapor (gs, mmol H₂O m⁻² s⁻¹) were measured using a LI-6400 XT portable gas exchange system (Li-Cor Biosciences Inc., Lincoln, NE, USA, standard 2 × 3-cm clear top chamber). The measurements were performed on three 1-year-old leaf fascicles (two needles per pseudophylls) placed across the width of the chamber. Needle length (l) and diameter (d) were measured in order to estimate the total external photosynthetic surface, calculated as (1 + π/2)ld × 6 needles × 1/2, the plane surface of one needle and the semicylindrical surface of the other needle of a fascicle being illuminated with the LED light source (6400–02 LED). The carbon dioxide assimilation rate was related to this total external needle surface. The measurements were completed with the LED light source set at 1200 μmol photons m⁻² s⁻¹, which corresponds to a saturated photosynthetic photon flux density (PPFD) for P. pinaster, a constant flow rate of 500 μmol s⁻¹, a leaf vapor pressure deficit of 1.33 ± 0.18 kPa, and a reference CO₂ concentration of 400 μmol mol⁻¹. The value of 1200 μmol photons m⁻² s⁻¹ was determined in preliminary measurements. We measured light-saturated net CO₂ assimilation rate and stomatal conductance at ambient CO₂ concentration at steady state conditions. The measurements were performed each day on weeks 0, 2, 5, and 6 after the beginning of the application of the stress on 16 randomly chosen pine trees so that all the 64 pines were characterized in 4 days per week. Three repeated measurements were averaged per plant (eight trees per family, four families per treatment, and two treatments).

2.5 Water-use efficiency

Intrinsic water-use efficiency (Wi, μmol CO₂ mmol H₂O⁻¹) was calculated from gas exchange measurements on individual plants as the ratio of A to gs.

2.6 Data analysis

The data was analyzed using the R software (version 2.8.0, R Development Core Team 2008). The data was found to meet the assumptions of homoscedasticity and of normal distribution of the residuals. The statistical tests were considered significant at P ≤ 0.05.

The genetic variation and the effect of time on the study traits were analyzed in the WW treatment using the following model:

$$ {X}_{ij}=\mu +{week}_i+{F}_j+\left({week}_i\ \mathrm{x}\ {F}_j\right)+{\varepsilon}_{ij} $$

(1)

where X_ij is the value of the trait, μ is the general mean, week is the time effect, F is the family effect, (week × F) is their corresponding interactions, and ε_ij is the residual. We observed that during the 6 weeks of the experiment, the air temperature and the air humidity did not vary significantly in the greenhouse (Fig. 3 in supplementary material). This assumption is consistent with our measurements of water content in the tree containers of the WW treatment. In this case, the week effect estimated on the WW trees measures the combined recovery and ontogenetic effect, consequence of tree development on the variables (called ontogenetic effect in the following).

Then, the data measured on the WS treatment was adjusted of this ontogenetic effect.

$$ {X}_{ij}=\mu +{week}_i+{\varepsilon}_{ij} $$

(2)

was used to estimate the week effect on the dataset of the WW treatment.

$$ {X}_{ij}-{week}_i=\mu +{\varepsilon}_{ij} $$

(3)

was used to adjust of the week effect on the dataset of the WS treatment.

The between-treatment plastic response between both treatments was studied using the following model of ANOVA on the pooled adjusted data of week 0 and 2 on the one hand and of weeks 5 and 6 on the other hand (weeks during which the difference between the treatments for SWC was the highest)

$$ {X}_{ij}=\mu + SWC+{F}_i+\left( SWC\ \mathrm{x}\ {F}_i\right)+{\varepsilon}_{ij} $$

(4)

where X_ij is the value of the trait, μ is the general mean, SWC is the soil water content measured in each pot of each tree, F_i is the family effect, SWC × F_i is the interaction between these two factors, and ε_ij is the residual. In this case, the linear relationship estimated by the SWC effect corresponds to the average plastic response, while the SWC × F_i interaction estimates the family effect on this relationship and thus the family plastic response.

Finally, the time-related plastic response was studied using Eq. (4) applied to a different dataset: the adjusted data of the WS treatment of weeks 0 to 6. In this way, we test the effect of the temporal variation of SWC on the value of the traits. Different transformations of the SWC variable were tested to meet the assumptions of the linear model.

The general relationship between SWC and the variable X_ij describes the temporal average plastic response of P. pinaster when SWC decreases with time in the stressed treatment. The (SWC × F_i) interaction tests the differences between the four families for this time-related plastic response.

3 Results

3.1 Variation with time of the physiological and growth traits in the WW and WS treatments and estimation of an ontogenetic effect

Figure 1 and Table 1 show that in the WW treatment, the SWC in the pots is stable with time (average 26.6%) while it decreases in the WS treatment from 27.88 to 12.08% from week 0 to week 6. In the WW treatment, the raw values of A, gs, diam, shoot, and height increase from week 0 to week 5, then levels off (Table 1). Needle predawn water potential (Ψb) is almost zero for the WW trees (n = 4) on week 6. Indeed, for all the WW data (Table 2a), we find a significant week (time) effect for A, gs, Wi, diam, shoot, and height with no significant week × family interaction (week × F), meaning that the week effect is the same for all the families. Using the daily weather data automatically collected by the greenhouse meteorological station, we find that there was neither significant temperature nor relative air humidity variation during the 6-week time lag of the experiment (results shown in Fig. 3 in the supplementary data). Thus, we attribute the phenotypic variation observed in the WW trees to the ontogenetic, developmental effect. We used this ontogenetic effect estimated on the WW trees to adjust the data collected on the WS trees. The unadjusted and adjusted data (A, gs, Wi, diam, shoot, and height) are shown in Table 1. The WW and WS trees have different variation patterns: the adjusted WS data shows a significant decrease of A, gs, diam, shoot, and height and a significant increase of Wi when SWC lowers (SWC is approximately 12% while Ψb of the WS trees was − 0.54 MPa ± 0.12 on week 6). The time-related SWC effect is significant on all the adjusted WS variables (Table 2d).

Fig. 1

Volumetric soil water content (SWC; %) of the 64 trees during the 6 weeks of the experiment. Week 0 (April 4 to 8), week 2 (April 18 to 22), week 5 (May 9 to 13), week 6 (May 16 to 20). The well-watered trees are represented by circles and the water-stressed trees by squares. The level of significance of the SWC effect is indicated by asterisks: ***P < 0.001, **0.001 ≤ P < 0.01, *0.01 < P < 0.05. Different letters (a, b, and c) indicate significant differences (p = 0.05 by Student’s t test) between weeks

Table 1 For each date (week), mean values of soil water content (SWC; %), net CO₂ assimilation (A; μmol CO₂ m⁻² s⁻¹), stomatal conductance to water vapor (gs; mmol H₂O m⁻² s⁻¹), intrinsic water-use efficiency (Wi = A/gs), diameter at 5 cm (diam; cm), length of terminal shoot (shoot; cm), and total height (height; cm) of the well-watered (WW) P. pinaster trees and the water-stressed (WS) P. pinaster trees for each measurement week

Table 2 Proportion of the variance due to different factors: time (week), family (F), time by family interaction (week × F), soil water content (SWC; %), and soil water content by family interaction (SWC × F) on all variables (net CO₂ assimilation (A, μmol CO₂ m⁻² s⁻¹), stomatal conductance to water vapor (gs, mmol H₂O m⁻² s⁻¹), intrinsic water-use efficiency (Wi = A/gs), diameter at 5 cm (diam, cm), length of terminal shoot (shoot, cm), and total height (height, cm)) for (a) the 32 WW trees (all the WW data), (b) the 64 trees of the pooled WW and WS treatments adjusted from the ontogenetic effect for weeks 0–2 on one hand (WW and WS data for weeks 0–2 pooled), (c) for weeks 5–6 on the other hand (WW and WS data for weeks 5–6 pooled), and (d) the 32 WS trees adjusted from the ontogenetic effect discovered in the WW treatment (WS data for weeks 0 to 6)

3.2 Variation between the four families for the physiological and growth traits

There is a significant family effect for A, gs, diam, shoot, and height in the WW treatment (Table 2a). There is also a significant family effect for Wi estimated in well-watered conditions for the WW and WS trees on the weeks 0–2 (Table 2b, F).

3.3 Average and family plastic response to the spatial and temporal variations of water availability

At the beginning of the experiment (Table 2b), there is no effect of SWC on A, gs, Wi, diam, and height for the 64 trees in both treatments. Conversely, at the end of the 6-week experiment, there is a significant SWC effect (Table 2c) on all the traits: in other words, there is a significant between-treatment phenotypic relationship between SWC and all the variables at the tree level in the pooled WW and WS treatments during weeks 5 and 6. This relationship measures the average between-treatment plastic response of the grouped families to the variation of water availability between both treatments. We note a significant family effect for gs, Wi, diam, shoot, and height (Table 2c, F) but no SWC × family interaction (Table 2c). This means that there is no significant difference between the four families for their between-treatment plastic response to SWC variation.

The significant effect of SWC on adjusted A, gs, Wi, diam, shoot, and height for the 32 trees in the WS treatment during the 6 weeks of the experiment (Table 2d) reveals an average plastic response to the temporal variation of SWC. We found a significant family effect for gs, shoot, and height and a significant SWC × family interaction for gs and Wi (Table 2d, F, SWC × F). This interaction corresponds to a significant family effect for the time-related plastic response to SWC for these two variables. Conversely, we found no significant SWC × family interaction for A and the three growth traits.

The corresponding plots are shown in Fig. 2a–f for A, gs, Wi, diam, shoot, and height, respectively. The fitted linear relationships are family norms of reaction that measure their time-related plastic response (Fig. 2). We could fit significant family norms of reaction for one to three of the four families, according to the trait, for A, gs, Wi, diam, shoot, and height. For the range of variation of SWC measured in the experiment, these norms of reaction are linear models. These family norms of reaction measure the time-related family plasticity to SWC variation in the WS treatment: there is a significant family variation for these norms of reaction for gs and Wi (Fig. 2b, c).

Fig. 2

Plot of the family norms of reaction (family plastic response) for the net CO₂ assimilation rate (A, μmol CO₂ m⁻² s⁻¹) (a), the stomatal conductance to water vapor (gs, mmol H₂O m⁻² s⁻¹) (b), Wi (=A/gs, μmol CO₂ mmol H₂O⁻¹) (c), the diameter at 5 cm (diam, cm) (d), the length of the terminal shoot (shoot, cm) (e), and the total height (height, cm) (f) as a function of the temporal variation of the log of the soil water content (log(SWC)) in the WS treatment. Each family (1062, 1058, 1046, 2082) is represented by a geometric symbol (o, Δ, +, ×). Only the lines corresponding to the significant norms of reaction are drawn. Solid, dashed, and dotted lines represent the norm of reaction for families 1062, 1058, and 1046, respectively. The significant correlation coefficient (R) for the pooled data are 0.27 for A (P = 0.0019), 0.56 for gs (P < 0.0001), − 0.56 for Wi (P < 0.0001), 0.34 for diam (P < 0.0001), 0.21 for shoot (P = 0.0185), and 0.21 for height (P = 0.0177)

3.4 Correlations between variables

Table 3 shows the correlations between the physiological and growth variables. As expected, A and gs are positively correlated. Wi is highly negatively correlated with gs in the WW and WS conditions. The growth variables positively correlate all together in all cases and are positively correlated with Wi in the WS conditions.

Table 3 Linear correlations (Spearman’s coefficients) computed for the well-watered (WW; below the diagonal line) and the water-stressed (WS; above the diagonal line) treatments on week 6 of the experiment between the six variables: net CO₂ assimilation (A, μmol CO₂ m⁻² s⁻¹), stomatal conductance to water vapor (gs, mmol H₂O m⁻² s⁻¹), intrinsic water-use efficiency (Wi = A/gs), diameter at 5 cm (diam, cm), length of terminal shoot (shoot, cm), and total height (height, cm) with 32 trees in each treatment

4 Discussion

4.1 Variation in the WW and WS treatments and estimation of an ontogenetic effect

The photosynthesis increase observed during the first 2 weeks after bud flush on 1-year-old P. pinaster needles in the well-watered treatment (WW, weeks 0 and 2; Table 1) is in accordance with other results on 1-year-old Pinus sylvestris, Picea abies, and Pinus contorta needles (Strand et al. 2002; Strand and Lundmark 1995). The significant week effect observed on all the physiological traits in the well-watered treatment (WW; Table 2a) is due to both the early spring recovery and the ontogenetic development effect (Slaney 2006). The ontogenetic development is observed on growth traits (Table 1) in the WW conditions as in other studies (Chambel et al. 2007; Fernández et al. 2006). It follows photosynthesis recovery with a gradual transition of unknown duration: we call their combination the ontogenetic effect. This seasonal change in A was also observed in spring on other species like Quercus douglasii (Xu and Baldocchi 2003). As them, we observe a similar pattern for gs and A that increase then level off (Table 1; weeks 5 and 6, WW, measurements at saturated PAR).

The evaluation of the ontogenetic effect in the WW treatment gave us the possibility to separate the internal ontogenetic effect from the plastic response in the WS treatment. Few studies have tried to separate the role of ontogenetic and water stress on photosynthesis (Maseda and Fernández 2006; Egea et al. 2011). The effect of this separation is obvious in Table 1: the developmental effect dominates and masks the response to the SWC decrease of all the phenotypic variables. Thanks to this separation, their response to environmental variation becomes apparent, statistically significant, and consistent with what is expected. As an example, Wi, diam, and total height are not significantly different between WW and WS during weeks 5 and 6 (Table 1), whereas a significant SWC effect is detected on these traits when data is corrected from the ontogenetic effect (Table 2c and d; SWC).

The high positive correlation between A and gs (Table 3) shows, in accordance with the literature (Galmés et al. 2007; Lawlor and Cornic 2002; Xu and Baldocchi 2003), that the stomatal closure during drought induces a decrease of A (adjusted WS data during the weeks 5 and 6; Table 1) (Cornic 2000; Flexas 2002; Lawlor and Cornic 2002; Limousin et al. 2010; Picon et al. 1996; Xu and Baldocchi 2003). The restriction of the CO₂ diffusion through the closing stomata during mild water stress explains most of this decrease (Dietz and Heber 1983; Lawlor and Cornic 2002; Limousin et al. 2010). The same type of response was observed on P. pinaster families (de Miguel et al. 2012; Fernández et al. 2006) and on P. pinaster provenances (Fernández et al. 2000; Picon-Cochard and Guehl 1999). The higher negative correlation between Wi and gs than between Wi and A (Table 3) shows that the Wi increase with drought (Limousin et al. 2010) (Table 1; Ad data) is mainly explained by gs. This is in accordance with previous works on P. pinaster (de Miguel et al. 2012), Populus (Monclus et al. 2006), and the fact that the reduction of stomatal aperture is one of the first protection reactions of isohydric plants like Pinus trees (Picon et al. 1996).

4.2 Variation between the four families

The family variation in the WW treatment for A and gs is in agreement with previous works on P. pinaster showing between-family variation of A (Fernández et al. 2006) and between-clone variation of gs (de Miguel et al. 2012). Such significant effect was also observed between 25 populations of Pseudotsuga menziesii for A, gs, Wi, and carbon isotopic discrimination (Zhang et al. 1993) and between five families of western larch trees for carbon isotopic discrimination and gs (Zhang et al. 1994). Nevertheless, in other studies, no significant difference was found for A and gs between five provenances of P. pinaster (Fernández et al. 2000) and between two varieties of Pinus nigra (Lebourgeois et al. 1998), neither for A and Wi between five families of western larch (Zhang et al. 1994). According to Fernández et al. (2000), limitations in the measurement equipment could explain that no genetic variation was found for gas exchange rate in some studies, while according to de Miguel et al. (2012), accurate measurement of gas exchange rates increases statistical power and allows to observe significant between-family difference. But the adjustment of the ontogenetic could be the main component of this improvement. We found a significant family effect for all the variables except Wi in the WW treatment during the 6 weeks of the experiment. When we pool the data of the WW and WS treatments for weeks 0–2 (no water stress), the family effect is significant for all variables except diam and height.

Our results demonstrate that there is significant variation between the four families. Of course, we are aware that the low number of families does not permit a robust quantitative estimate of the genetic variation. However, it proves that there is genetic variation for the studied traits.

4.3 Average plastic response to the between-treatment and time-related variation of water availability

The adjustment of the ontogenetic effect gave us the opportunity of a double approach of plasticity: between-treatment (between the two treatments in the greenhouse) and time-related (along the 6 weeks of increased drought stress) related. We observed a significant average plastic response of P. pinaster to decreasing water availability for both approaches (between-treatment (Table 2c) and time-related (Table 2d)) with a significant effect of SWC on the six traits studied. A significant between-treatment plastic response for the same variables was often found in the form of a significant treatment effect in water stress experiments (Fernández et al. 2006; Aranda et al. 2010; Sánchez-Gómez et al. 2010; de la Mata et al. 2014; Chambel et al. 2007; Corcuera et al. 2010) or along a rainfall gradient (Martin-StPaul et al. 2013). However, to our knowledge, there is no published result about time-related plasticity for these variables in this type of water-stressed experiment for P. pinaster or other conifer species, mostly because the time-related plastic response is generally confounded with the ontogenetic development.

4.4 Family effect for plastic response

We found no family effect for the between-treatment plastic response for any variable (no significant SWC × family interaction). Genetic variation of plastic response against spatial environmental variation is often studied in the form of genotype × environment interaction in drought-stress experiments. As we did, Fernández et al. (2006) found no significant family effect on between-treatment plasticity for water availability for gs, A, and Wi. Others (Aranda et al. 2010; Corcuera et al. 2010, 2012) found significant population or family effect for Wi using carbon isotope discrimination as a proxy. For growth traits, the results are very diverse: significant (de la Mata et al. 2014) or non-significant (Corcuera et al. 2010; Chambel et al. 2007) population plastic response for height; similar behavior of populations in mesic and xeric sites for total height at age 1 after planting (Gaspar et al. 2013); and non-significant population plastic response for relative height and diameter growth rate (Sánchez-Gómez et al. 2010), but significant for diameter (Chambel et al. 2007). Finally, Fernández et al. (2006) found a significant family plastic effect for dry weight while Aranda et al. (2010) did not. Correia et al. (2008) found significant differences between P. pinaster populations for total height and carbon isotope discrimination used as a proxy of Wi.

While we found no family effect for the between-treatment plastic response for any variable, we found a significant family effect for the time-related plastic response for gs and Wi. The adjustment of the ontogenetic effect allowed us to access to this time-related plasticity and contributed to the identification of this significant family effect. To our knowledge, this is the first time that this type of time-related plastic response to water availability is observed in P. pinaster.

The variety of results observed suggests that time-related and between-treatment plasticity are not alike. This is consistent with predictions using a theoretical model showing that variation in time and variation in space have different effects on plasticity (Scheiner 2013). Accurate comparison of between-treatment and time-related plasticity in fixed organisms like plants requires large-scale controlled condition experiments with a strict monitoring of ontogenetic effects. Our study suggests that such monitoring is possible.

In conclusion, we found variation of physiological and growth traits with water-stressed intensity in young P. pinaster trees studied few weeks after bud flush, along time as well as between treatments. We were able to quantify an ontogenetic effect for the six variables studied. We used this ontogenetic effect to better estimate the plastic response to water availability. We found a significant average plastic response to the spatial (between-treatment) and temporal (time-related) variation of water availability for the six variables. In the spatial analysis, the trees of each family in the WW and WS treatments are different individuals of the same half-sib families: intra-family variation is added to between-treatment variation. In the temporal analysis, the same trees of each family in the WS treatment are observed at different times during the 6-week experiment. Therefore, the time-related plastic response is more precisely estimated than the between-treatment plastic response. This is probably why we found a significant family effect for the time-related plastic response for gs and Wi, while this effect was not significant for the between-treatment plastic response. The observed plastic response to SWC is a convenient integrated measure of P. pinaster response to drought. Stomatal conductance (gs) and shoot are the only traits showing a significant family effect at all time in both treatments. Stomatal conductance was found to be more variable than A and more strongly correlated with Wi. All this suggests that the family plastic response for gs could be a suitable integrated selection criterion for improving Wi in this species. This result, if verified at early and adult age, with more trees and families for a precise estimation of genetic variation, could be used as a tool for early selection and prediction of future performance under water limitation conditions. At the same time, the growth traits were found to be mostly independent from the physiological traits. It suggests that it may be possible to minimize the negative impact of drought on growth when improving maritime pine drought tolerance.

Data availability

Data generated or analyzed during this study are included in this published article and its supplementary information files.