Plant material and exposure conditions 

Cuttings of the two Euramerican poplar genotypes “Carpaccio” and “Robusta” were grown in growth chambers as already described in Dusart et al. (2019b) with slight modifications. Cuttings were planted in ten-liter pots filled with a sand/peat mixture (1/1, v/v) and fertilised by adding 15 g of slow-release nutritive granules (Nutricot T-100) and 1 g.L-1 CaMg(CO3). For both genotypes, forty-eight plants were randomly distributed in eight phytotronic chambers, i.e. twenty-four plants in control chambers (charcoal-filtered air), and twenty-four plants in chambers set for O3 treatment (80 nmol.mol-1 for 13 hours, from 09:00 to 22:00). For reasons of space in the culture chambers and length of measurement times, the experiment was duplicated separately for each genotype. After a 7-day-long acclimation period, the O3 treatment started while control saplings were exposed to charcoal-filtered air for 13 days. After 13 days (d) of fumigation, the total cumulative sum of O3 flux (SUM00), the cumulative O3 dose above a threshold of 40 ppb (AOT 40), and the phytotoxic O3 dose above a threshold flux of 0 nmol.m-2.s-1 (POD0) (based on measured stomatal conductance, see Bagard et al., 2015) were determined (Tableau 13). At the end of the O3 exposure period, half of the saplings were submitted to a moderate water deficit for 7 d. Soil Water Content (SWC) was determined with 24 wireless Time Domain Reflectometry (TDR) probes (CWS655E, Campbell Scientific Ltd, Antony, France). A calibration between volumetric SWC measured by TDR and pot weight was performed. The biological available water was expressed as relative extractable water (REW), as described by Wildhagen et al. (2018) for the same soil. Poplars were watered with a known volume of water several times a day to maintain the level of REW stable. For the wellwatered treatment, poplars were irrigated at 75 % (±10%) of REW, whereas for the water deficit treatment, irrigation was set to 45% (± 2%) of REW until the end of the experiment (Figure 53). A cumulative sum of the amount of water added for each treatment for 21 d is presented in Tableau 13. 

 Plant growth 

The number of leaves and the diameter at the collar and height were recorded twice a week until the end of the experiment for each individual. At the end of the experiment, leaves, stems and roots were oven-dried at 60 °C until they reached a constant dry mass. 

Gas exchanges and photosynthetic pigment kinetics 

Gas exchanges (An, net CO2 assimilation, and gs, stomatal conductance to water vapour) were measured using a Li-6200 (Li Cor, Lincoln, NE, USA) as described in Dusart et al. (2019b). Non-destructive determination of the chlorophyll pigment content was performed with a Dualex (Force-A, Orsay, France). For all non-destructive leaf measurements, the same leaf was used, i.e., the first fully expanded leaf (the 10th leaf from the apex) at the beginning of the O3 treatment. 

Stomatal response to irradiance and vapour pressure deficit 

Gas exchange measurements

 Gas exchange measurements were performed with a Li-6400 system, as described in Durand et al. (2019) with some minor modifications. Parameters of the leaf cuvette were for light: PAR: 800 μmol.m-2.s-1 with 30 μmol.m-2.s-1 of blue irradiance and VPD: 0.8 kPa, until gs reached a steady state (g0, defined as a variation lower than 5% over 5 minutes). Then light was turned off (as well as in the phytotronic chamber) until gs got to a new steady state (g1), then turned on to 800 μmol.m-2.s-1 until stomatal conductance reached the last steady state (g2). The 800 μmol.m-2.s-1 value was chosen to avoid photoinhibition due to excess light (Niinemets & Kull, 2001). A similar procedure was used to monitor gs response to a change of VPD: it was switched to 3 kPa instead of 0.8 kPa (for a fixed PAR: 800 μmol.m-2.s-1). VPD from leaf tissues to air was controlled with a dew point generator as described in Vialet-Chabrand et al. (2013). 

Modelling 

The obtained stomatal response curves were fitted using the following sigmoidal model (Vialet-Chabrand et al., 2013): gs = g0 + (G − g0 ) e −e ( λ−t τ ) where gs is the fitted stomatal conductance, g0 and G are the steady-state values of gs (mol.m2 .s-1), respectively at the start and at the end of the curve, τ is a time constant (s), λ is the lag time (s), and t is time (s). The speed of the stomatal response was estimated by calculating the maximum slope (SLmax), as follows: SLmax = G − g0 τ. e where (G-g0) represents the amplitude of the stomatal response and e is Euler’s number (e ≈ 2.718). Further information regarding the model parameters and fitting procedure can be found in Gérardin et al. (2018) and Durand et al. (2019). Chapitre V : Réponse du peuplier soumis à une succession ozone & sécheresse

Statistical analyses 

Statistical analyses were performed using R 3.1.0 (R Development Core Team) open-source software. Linear models created from the nlme package (Pinheiro et al., 2018) were used to study growth parameters with ANOVA, including the effects of water deficit, O3 and genotype. The growth chamber was also tested and excluded from the models because the effect was not significant for all the parameters tested throughout the whole experimental period. Model parameters, G, g0, λ, τ, SLmax were explored in the same way. The lme4 package (D. Bates et al., 2015) was used to fit a linear mixed-effect model on gas exchange and chlorophyll content data with fixed variables (water deficit and O3 data) whereas biological replicas were random variables. Residual plots of the model were used to assumed heteroscedasticity and variance homogeneity. The emmeans package (Lenth, 2016) was used to perform multiple comparisons. To determine if O3 and water deficit had an additive, synergistic or antagonistic impact on gs, we compared the observed effects to the expected additive effects for the saplings exposed to O3 and then to water deficit (Methods 1 available in Supplementary data).