Species climate range influences hydraulic and stomatal traits in Eucalyptus species

Abstract

Background and Aims Plant hydraulic traits influence the capacity of species to grow and survive in water-limited environments, but their comparative study at a common site has been limited. The primary aim of this study was to determine whether selective pressures on species originating in drought-prone environments constrain hydraulic traits among related species grown under common conditions.

Methods Leaf tissue water relations, xylem anatomy, stomatal behaviour and vulnerability to drought-induced embolism were measured on six Eucalyptus species growing in a common garden to determine whether these traits were related to current species climate range and to understand linkages between the traits.

Key Results Hydraulically weighted xylem vessel diameter, leaf turgor loss point, the water potential at stomatal closure and vulnerability to drought-induced embolism were significantly (P < 0·05) correlated with climate parameters from the species range. There was a co-ordination between stem and leaf parameters with the water potential at turgor loss, 12 % loss of conductivity and the point of stomatal closure significantly correlated.

Conclusions The correlation of hydraulic, stomatal and anatomical traits with climate variables from the species’ original ranges suggests that these traits are genetically constrained. The conservative nature of xylem traits in Eucalyptus trees has important implications for the limits of species responses to changing environmental conditions and thus for species survival and distribution into the future, and yields new information for physiological models.

INTRODUCTION

Sensitivity to drought has been identified as an important factor shaping plant species distributions and community composition (BassiriRad et al., 1999; Engelbrecht et al., 2007; Kursar et al., 2009). The ability of plant species to grow and survive under different moisture regimes and to withstand drought is dependent on a number of physiological, anatomical and morphological traits including stomatal regulation of water loss, osmotic adjustment of leaf turgor loss point, rooting depth and xylem structure, as well as whole-plant adjustments such as root and leaf mass fraction, leaf area ratio and root length (Anderson et al., 1996; Villar-Salvador et al., 1997; Schulze et al., 2006; Baltzer et al., 2008; Bourne et al., 2015; Pfautsch et al., 2016). Xylem cavitation resistance has emerged as a key variable in relation to plant drought resistance and the distribution of plants across gradients of water availability (Kursar et al., 2009; Choat et al., 2012; Schuldt et al., 2015; Anderegg et al., 2016). Low precipitation and high evaporative demand will drive high potential evapotranspiration and a decline in plant water potential and increase in xylem tension. Although declining water potentials can be delayed by stomatal closure and leaf shedding, prolonged drought will eventually result in extensive xylem cavitation, a sharp reduction in hydraulic conductivity, canopy die back and ultimately death of the whole plant (Choat, 2013; Nardini et al., 2013).

Stomatal function is linked with hydraulic characteristics because (1) the plant hydraulic system controls the rate at which water can be transported to the canopy (Sperry et al., 2002; Brodribb and Holbrook, 2003) and (2) stomatal regulation must maintain xylem water potentials within a range that limits cavitation to maintain hydraulic integrity (Delzon and Cochard, 2014). Plants with high stomatal conductance (g_s) and photosynthetic rates therefore generally require high hydraulic conductivity (Brodribb and Feild, 2000; Santiago et al., 2004; Choat et al., 2011). Hydraulic conductivity is proportional to the sum of all conduit diameters to the fourth power (Tyree and Ewers, 1991), thus a small increase in vessel diameter results in large increases in hydraulic conductivity (K_h). However, xylem traits that promote high K_h may also increase the risk of cavitation and hydraulic dysfunction within the plant (Loepfe et al., 2007; Sperry et al., 2008). Subsequently, plants face trade-offs between hydraulic safety (e.g. high cavitation resistance) and hydraulic efficiency (high K_h) in order to survive and maintain productivity within the climate of their native distribution (Sperry et al., 2008; Pittermann et al., 2010; Johnson et al., 2011).

In the context of plant hydraulics, efficiency is defined as the amount of water transpired for a given pressure drop along the hydraulic pathway (Tyree et al., 1994); thus, a more efficient hydraulic system requires a smaller pressure drop to transport a given amount of water, and is usually promoted by wider, longer xylem vessels. Conversely, hydraulic safety is promoted by narrower, shorter, more frequent xylem vessels and thicker, less porous pit membranes in the hydraulic system (Sperry et al., 2006, 2008; Loepfe et al., 2007; Jansen et al., 2009). These adaptations enhance drought tolerance (Baas, 1986; Choat et al., 2005), the degree to which a plant has adapted to maintain cell water status under dry or arid conditions (Oliver et al., 2010), by limiting the spread of embolism between xylem vessels via air-seeding (Loepfe et al., 2007; Choat et al., 2008). However, these adaptations also compromise hydraulic efficiency, resulting in lower maximum K_h and therefore the rate at which water can be delivered to the transpiring leaf surface for a given pressure gradient (Sperry et al., 2008; Pittermann et al., 2010). This trade-off between safety and efficiency (Nardini and Salleo, 2000; Sperry et al., 2008; Johnson et al., 2011) forms the basis for the co-ordination of stem hydraulic traits and consequently their relationship to stomatal traits previously reported (Brodribb et al., 2003; Meinzer et al., 2009). Studies examining the co-ordination between stomatal conductance and stem hydraulics have reported that the point of stomatal closure occurs prior to the onset of cavitation and is related to stem vulnerability and the water potential at turgor loss point (Ψ_TLP) in different species (Brodribb and Holbrook, 2003; Brodribb et al., 2003; Martorell et al., 2014). However, uncertainty remains about how these safety vs. efficiency trade-offs are related to species climatic distribution, particularly when species are grown in a common climate. Additionally, some species are reported to maintain low hydraulic safety and efficiency (Gleason et al., 2016), presenting a challenge for understanding xylem evolution. Understanding this relationship is an important precursor to predicting how species survival and productivity can be affected under current and future climatic conditions.

The selective pressures that prevail within a particular climate constrain species adaptations and subsequently which characteristics are maintained in future generations (Ramirez-Valiente et al., 2010). While species from arid regions may benefit from maintaining a low stomatal conductance, increasing water use efficiency by minimizing transpiration, they also benefit from maintaining a hydraulic system that is resistant to cavitation and offers greater redundancy, i.e. narrow, short and frequent xylem vessels. Alternatively, a species originating from humid areas where water is readily available may benefit from anatomical characteristics oriented towards higher hydraulic conductance, assimilation rates and productivity (Fichot et al., 2009). Species oriented towards hydraulic efficiency typically exhibit stomatal closure at higher leaf and stem water potentials, allowing them high transport rates when water is available but minimizing the risk of cavitation when water deficits develop (Nolf et al., 2015). In this way, while trait values favouring hydraulic safety may be selected for in species from more arid regions (e.g. Schuldt et al., 2015; Pfautsch et al., 2016), hydraulic efficiency and enhanced productivity through access to light, water and nutrient resources are more likely to drive selection in species from more humid regions.

Previous studies have reported correlations between climate and stomatal behaviour (Héroult et al., 2013; Bourne et al., 2015), hydraulic characteristics (Brodribb and Hill, 1999; Maherali et al., 2004; Mitchell et al., 2008; Beikircher and Mayr, 2009; Choat et al., 2012), vessel anatomy (Villar-Salvador et al., 1997; Noshiro and Baas, 2000; Fisher et al., 2007; Medeiros and Pockman, 2014; Schreiber et al., 2015; Pfautsch et al., 2016) and leaf tissue water relations (White et al., 2003; Bartlett et al., 2012). However, many of these studies compare different species along environmental gradients, which does not allow for determination of climatic effects on hydraulic, stomatal and leaf physiological traits of individual species. Common garden studies, which enable environmental conditions to be kept constant, can provide information on the influence of genotype on hydraulic traits. Despite the importance of these traits in determining drought tolerance, survival and water use under different climatic conditions, relatively few studies have utilized common gardens to examine genotypic constraints on hydraulic traits. Previous common garden studies have provided valuable insights into the effects of species climate range on hydraulic characteristics (Vander Willigen et al., 2000), leaf water relations (Baltzer et al., 2008) and xylem anatomy (Schreiber et al., 2015). However, to date, there are no studies that address the influence of species climate range on stomatal, hydraulic and anatomical traits in trees species constrained to a single genus.

Using a common garden plantation of Eucalyptus species, we ask whether species climate range, specifically the mean annual precipitation (MAP) and vapour pressure deficit (D), influences the leaf water relations, xylem anatomy, and hydraulic and stomatal characteristics of a species. We expect that selective pressures from the species native climate constrain the genotypes maintained within a species and thus that the water use strategies employed by different species will be linked to the climate of species range. Our specific hypotheses for this study are: (1) xylem anatomy, leaf tissue water relations, and stomatal and hydraulic traits will correlate with precipitation and D of the native distribution of species; and (2) there is a co-ordination of stem hydraulics with leaf tissue water relations and the point of stomatal closure across species from different climatic ranges. Comparing multiple species from the same genus within a common garden, where confounding environmental and phylogenetic factors are minimized, allows extrapolation of genetic constraints on hydraulic, anatomical, stomatal and leaf tissue water relations traits.

MATERIALS AND METHODS

Study site

The study was conducted in a plantation of Eucalyptus species at Richmond, New South Wales, Australia (33º33′S, 150º44′E) near the Hawkesbury River on alluvial floodplain at an elevation of 25 m. The local climate is sub-humid temperate, characterized by hot summers and the low possibility of frost during winter (approx. 13 times annually). Mean annual temperature is 17 ºC, with a mean monthly maximum and minimum temperature of 30·1 ºC and 3·5 ºC generally achieved in January and July, respectively. Annual rainfall is 801 mm, with a mean maximum and minimum rainfall of 129 mm and 32 mm per month, respectively, distributed fairly evenly but with the majority of rain falling in summer months. The plantations were established in September 2007 in a paddock with occasional weed control, as described previously in Héroult et al. (2013) and Bourne et al. (2015). The plantation consisted of two replicate, monoculture plots of each of five Eucalyptus species, namely Eucalyptus cladocalyx, E. crebra, E. dunnii, E. grandis and E. tereticornis, where each plot contained 35 individual trees planted at a density of 1000 trees ha^–1 with a spacing of 2·6 m along the row and 3·85 m between rows. An additional monoculture plot of 160 trees of E. saligna was planted alongside these plantations with the same spacing at the same time. Seedlings were fertilized at planting in September 2007, watered by hand for the first 4 months of establishment, and replaced as required within the first year, with survival of all individuals at the time of measurement. Measurements were conducted on 4- to 5-year-old trees of each species with average tree heights of 8·8, 8·3, 13·1, 15·5, 11·1 and 12·8 m for E. cladocalyx, E. crebra, E. dunnii, E. grandis, E. saligna and E. tereticornis, respectively. The driest part of each species range, being the lowest mean annual precipitation observed in the species range, was used as an indication of their ability to tolerate dry conditions, and in accordance with the Koppen–Geiger climate classification system we classified E. cladocalyx and E. crebra as sub-humid climate zone species, E. tereticornis as a sub-humid/humid intermediate species, and E. dunnii, E. grandis and E. saligna as humid zone species. We used climate maps in Peel et al. (2007) overlain with species distributions from the Atlas of Living Australia (ALA; www.ala.org.au) for these classifications, consistent with Bourne et al. (2015).

Leaf tissue water relations

We conducted measurements of pressure volume (PV) isotherms to measure leaf tissue water relations using techniques described by Tyree and Hammel (1972). A mature, sunlit leaf from six different trees of each of six species, namely E. cladocalyx, E. crebra, E. dunnii, E. grandis, E. saligna and E. tereticornis, was collected from the upper third of the tree crown. These measurements provided leaf tissue water relation traits as described by Koide et al. (2000) and Sack et al. (2003) and obtained using a spreadsheet available from PrometheusWiki (Sack et al., 2011).

Anatomical characteristics

Estimates of vessel diameter and density were obtained for E. cladocalyx, E. crebra, E. dunnii, E. grandis, E. saligna and E. tereticornis trees. Sections of 20 μm thickness were taken from one separate branch segment from the upper third of the tree crown from each of five different trees using a sliding microtome. Photographs of these sections, stained with 0·05 % (w/v) Toluidine blue, were taken using a digital camera (ProgRes^® C14, JenOptik, Hallam, Australia) with the Lumix Progress C4 software as seen at × 40 (vessel counts) and ×100 (vessel diameter) magnification under a microscope (Olympus BX60, Center Valley, PA, USA). We analysed these images and measured vessel diameter of all vessels >300 μm² throughout the sapwood of each section using ImageJ processing and analysis software (Abramoff et al., 2004). Hydraulically weighted vessel diameter (D_h), consistent with that measured by Sperry and Saliendra (1994), was calculated for interspecific comparisons of vessel anatomy and correlations with climate.

Stem hydraulics

Five to six branches from separate trees of E. cladocalyx, E. crebra, E. dunnii, E. grandis, E. saligna and E. tereticornis, with up to six 4–9 cm long sub-samples per branch, were harvested from the upper third of the tree crown using a boom lift (Snorkel MHP13/35 Trailer Mounted Lift, Snorkel Ltd, Meadowbrook, Qld, Australia) and measured for stem hydraulic characteristics. Xylem vulnerability curves were constructed using the bench dehydration method (Sperry et al., 1988) where branches were progressively dried down at 25 ºC for between one and 12 cumulative hours dependent on species selected, branch size and the desired water potential. Regardless of segment lengths considered, bench dehydration provides more accurate vulnerability estimates for long-vesseled species, prone to the ‘open vessel’ artefact, than either air injection or centrifuge methods (Choat et al., 2010, 2016; Cochard et al., 2013). As Eucalyptus species have long vessels (>1 m length), bench dehydration was used here.

Xylem tension was relaxed prior to measurements in order to avoid possible excision artefacts (Wheeler et al., 2013; Torres-Ruiz et al., 2015). Preliminary testing of xylem tension relaxation in these species revealed that rehydration for up to 2 h followed by progressive cutting back underwater from upstream and downstream ends achieved the best results. Accordingly, stems were harvested close to dawn and rehydrated for 2 h in a walk-in cold room set to 4 ºC. The first cut was made to the upstream end of the branch underwater, at least 30 cm from the desired segment, and left to rehydrate for 10 min. Successive cuts were made at 2 min intervals, allowing rehydration between each cut, alternating between upstream and downstream ends of the segment. Water potentials on excised leaves were measured prior to cutting the branch using a Scholander type pressure chamber (Model 1505D Pressure Chamber Instrument, PMS Instrument Company, Albany, OR, USA). Initial hydraulic conductivity (K_initial) and final hydraulic conductivity (K_final) were measured using a 0·003 MPa pressure head and logged using a digital liquid flow meter (Liqui-Flow L10, Bronkhorst High-Tech BV, Ruurlo, Gelderland, The Netherlands) and flow analysis programs FlowDDE (V. 4.69) and FlowPlot (Versions 4·69 and 3·34, respectively, Bronkhorst FlowWare, http://downloads.bronhorst.com). K_final was measured after flushing the stem segment with 2 mmol KCl solution at a pressure of 175 kPa for 20 min or until bubbles were no longer emitted from the segment.

Stomatal conductance and water relations measurements

Measurements of g_s and stem water potential (Ψ_stem) were made on progressively dehydrated E. cladocalyx, E. crebra, E. dunnii, E. grandis, E. saligna and E. tereticornis branches over 13 sunny days at times when trees were well hydrated post-rainfall events and evaporative demand was not sufficiently high to cause rapid dry-down of branches. Branches were collected from the upper third of five trees of each species using a boom lift, cut underwater and then transferred to water-filled buckets. Branches were kept in the sun and left to transpire for 30 min prior to measurements, while leaves on secondary branches adjacent to g_s measurement leaves were bagged and foiled, ready for measurements of Ψ_stem. Four open-flow portable photosynthesis systems with 6400-01 leaf chambers (Li6400, Li-Cor, Inc., Lincoln, NE, USA) and a Scholander-type pressure chamber were used to measure g_s and Ψ_stem, respectively. Cuvette light levels and temperatures were adjusted to ambient conditions, as determined from a nearby weather station (Héroult et al., 2013; Bourne et al., 2015) and the portable photosynthesis systems. Chamber humidity was maintained as close as possible to ambient conditions at the beginning of measurements (approx. 70 % humidity) by removing a small amount of water vapour commensurate with that expected to be added by leaves during measurements. For each measurement, the CO₂ mole fraction within the chamber and at the leaf surface (C_a) was set at ambient levels (390 μmol mol^–1 CO₂) by mixing CO₂ from an internal source (CO₂ cartridge) with outside air. Stomatal ratios determined from microscopic observations of stomatal peels were used to correct stomatal ratios for each species. Three to eight fully expanded, healthy leaves on each branch were placed inside the cuvette and then logged five times for each g_s observation once the leaf reached equilibrium (coefficient of variation for A_net and g_s < 1 %). Eucalyptus tereticornis and E. saligna only had overwintering leaves, while all other species had fully expanded, hardened leaves, yet stomatal conductance values were consistent with those conducted on fully expanded, hardened leaves for E. tereticornis from an earlier measurement campaign. Measurement leaves and adjacent foiled and bagged leaves were then measured for Ψ_leaf and Ψ_stem, respectively. Branches were then dried to more negative water potentials by removing them from the bucket, and g_s and Ψ_stem were measured regularly until all available leaves were exhausted or g_s reached zero.

Statistical analysis

Species climate range and distributional co-ordinates of the different species were mined from the ALA (http://spatial.ala.org.au/; accessed 12 December 2014) and synthesized and plotted using R (R Development Core Team, 2014). A drawback of this approach is that it may include planted trees as well as those in their native distribution. The ALA species distributional data were trimmed to include only data between the first and 99th percentile as a small number of locations, such as those from botanical gardens and herbariums, were outside the natural distributional range of species. We conducted pairwise Pearson correlations of traits from species leaf tissue water relations, stomatal behaviour and hydraulic measurements against specific climate parameters of the species range using JMP (v. 5.0.1, SAS, Cary, NC, USA). A large number of climate parameters averaged across the species range are inter-related, and we focused analysis on a few of these as key climate indicators (mean annual precipitation and maximum monthly vapour pressure deficit), with a wider set of climate parameters considered elsewhere (see Supplementary Data Table S1). Maximum monthly vapour pressure deficit (D_max) was used in analyses in the form of 1/D_max, consistent with stomatal models based on hydraulic theory and the negative exponential relationship between stomatal conductance and D (Tardieu and Davies, 1993; Leuning, 1995; Oren et al., 1999). The Ψ_stem at stomatal closure was assessed at two points, the beginning of stomatal closure (g_s50; when g_s reached 50 % of g_smax) and complete stomatal closure (g_s5; when g_s reached 5 % of g_smax) by considering the point at which g_s declined below a certain point of the maximum measured g_s using all g_s measurements. These parameters were used as indicators of stomatal behaviour as complete stomatal closure (g_s = 0) was not measured. Sigmaplot (v. 11.0, Systat Software, San Jose, CA, USA) was used for fitting g_s vs. Ψ_stem curves and obtaining 95 % confidence intervals (CIs) using sigmoidal three parameter (E. cladocalyx and E. crebra), sigmoidal four parameter (E. saligna and E. tereticornis), Chapman (E. dunnii) and quadratic (E. grandis) functions based on those that best represented the data. A Weibull function as re-parameterized by Ogle et al. (2009) was used for fitting vulnerability curves, from which 50 % loss of conductivity (P₅₀) values and 95 % CIs were calculated in R (R Development Core Team, 2014) using the fitplc package (Duursma and Choat, 2017).

RESULTS

Climate range is related to leaf tissue water relations, xylem anatomy, and stomatal and hydraulic characteristics

All except one of the Eucalyptus species in our study were distributed along the eastern coastline of Australia, with some species, such as E. crebra, distributed into more arid, inland regions of Australia (Fig. 1). The exception to this was E. cladocalyx, which was distributed along the south-eastern coastline of Australia and was also distributed into more arid, inland regions (Fig. 1). Eucalyptus dunnii, E. cladocalyx, E. grandis and E. saligna had narrower distributional ranges compared with the other species, and E. dunnii had the most restricted distribution of all species (Fig. 1), though the native range of E. cladocalyx is also very narrow (not shown) existing in few native pockets in the Flinders Ranges, Eyre Peninsula and on Kangaroo Island. The D and MAP of climate range differed across species, with sub-humid species, E. cladocalyx and E. crebra, having a lower MAP and higher D compared with the humid species, E. saligna, E. dunnii and E. grandis, where E. tereticornis was intermediate between the two (Fig. 1).

Fig. 1.

Distribution of six Eucalyptus species and climatic range for two climate variables related to water relations. Species distribution data were extracted from Atlas of Living Australia (ALA; http://www.ala.org.au; accessed 5 February 2015). In maps, black points represent the species distribution according to ALA. Boxplots indicate the maximum mean monthly D (left) and MAP (right) from the species range. The natural distribution for E. cladocalyx is actually more restricted than depicted here, but there is currently no reliable way to obtain this information.

Species climate range was related to leaf tissue water relations and xylem anatomy of the six Eucalyptus species studied (Fig. 2; Tables 1 and 2; Supplementary Data Fig. S1; Table S1). Water potential at turgor loss (Ψ_TLP), volume index growth rate (VIGR), osmotic potential at full turgor (Π_o), g_s50, D_h and P₅₀ were correlated with MAP (R² = 0·76, 0·84, 0·66, 0·93, 0·75 and 0·53, respectively; Fig. 2A–F;Table S1) such that species from more humid regions with a higher mean annual precipitation of species range had higher Ψ_TLP, VIGR and D_h and higher (less negative) P₅₀ than species from sub-humid regions. These correlations were qualitatively similar for the inverse of maximal D (Table S1). Mean annual aridity index and monthly maximum water deficit were also significantly correlated with leaf tissue water relations, and anatomical, hydraulic and stomatal characteristics (Table S1). Furthermore, a co-ordination and correlation between these stem and leaf characteristics was also observed (Fig. 3; Table S1) where Ψ_TLP was significantly positively correlated with VIGR (R² = 0·73, P = 0·03) and negatively correlated with g_s50 and 12 % loss of conductivity (P₁₂; R² = 0·89 and 0·64, respectively; Fig. 3; Table S1).

Fig. 2.

Correlations of leaf tissue water relations, xylem anatomy, growth and hydraulic characteristics with mean annual precipitation (MAP) of the species range for six climatically contrasting Eucalyptus species. Correlations and their significance are presented in Table S1. Dashed lines are shown as a guide. Error bars are the s.e.m. for each species or climate parameter, and are bidirectional.

Table 1.

Leaf tissue water relations, xylem anatomy and climate characteristics of six Eucalyptus species grown together in a common garden, with the s.e.m. across seven trees of each species

Species	Class	RWCTLP (%)	ε (MPa)	VDens (mm–2)	VArea (μm2)	VDiam (μm)	Dh (μm)	MAP (mm)	AI	Dmax (kPa)	WDmax (mm)
E. cladocalyx	Sub-humid	84·1 ± 0·7	17·4 ± 0·6	113·01 ± 10·78	917·5 ± 58·0	33·1 ± 0·9	36·0 ± 1·3	537·49 ± 3·16	0·59 ± 0·004	0·78 ± 0·007	120·3 ± 1·0
E. crebra	Sub-humid	89·2 ± 1·5	26·4 ± 3·9	96·79 ± 5·89	1001·0 ± 130·3	34·3 ± 2·2	38·4 ± 2·0	743·37 ± 2·21	0·48 ± 0·001	1·00 ± 0·003	124·9 ± 0·4
E. tereticornis	Intermediate	90·0 ± 0·3	22·1 ± 0·9	78·36 ± 6·01	1195·1 ± 97·9	37·6 ± 1·6	41·2 ± 1·8	971·09 ± 3·13	0·70 ± 0·002	0·67 ± 0·002	87·6 ± 0·4
E. saligna	Humid	89·6 ± 0·6	22·6 ± 0·8	64·98 ± 3·85	1151·0 ± 96·0	31·4 ± 1·8	40·8 ± 1·7	1190·93 ± 3·98	0·93 ± 0·003	0·57 ± 0·001	51·2 ± 0·6
E. dunnii	Humid	87·6 ± 1·3	15·3 ± 1·8	95·23 ± 3·94	1154·2 ± 83·7	36·1 ± 1·3	39·6 ± 3·5	1109·58 ± 11·22	0·83 ± 0·009	0·57 ± 0·003	33·6 ± 1·2
E. grandis	Humid	88·9 ± 0·5	20·0 ± 0·8	97·52 ± 6·85	1171·3 ± 184·0	36·5 ± 2·4	41·4 ± 3·5	1476·37 ± 7·52	1·05 ± 0·006	0·53 ± 0·002	50·7 ± 0·7

VDens, VArea and VDiam are vessel density, vessel area and vessel diameter as an arithmetic mean, respectively. AI and WD_max are representative of the species range, as obtained using Atlas of Living Australia (http://www.ala.org.au; accessed 12 December 2014; climate parameters are for 2007–2010).

Table 2.

A comparison of leaf tissue water relations, hydraulic parameters and growth rates of six Eucalyptus species in a common garden

Species	ΨTLP (MPa)	P50 (–MPa)	P12 (–MPa)	P88 (–MPa)	gs50 (–MPa)	gs5 (–MPa)	VIGR (m3 ha–1 year–1)
E. cladocalyx	–3·13 ± 0·08	5·27 [6·23, 4·57]	2·74 [3·97, 1·22]	8·13 [2·30, 1·32]	3·51 [5·07, 1·94]	6·91	27 ± 5
E. crebra	–2·99 ± 0·11	4·90 [5·16, 4·49]	3·26 [3·79, 2·69]	6·41 [7·14, 5·95]	3·30 [4·13, 2·53]	6·35	52 ± 11
E. tereticornis	–2·47 ± 0·07	3·89 [4·30, 3·51]	2·18 [2·75, 1·60]	5·70 [6·99, 4·87]	2·43 [3·30, 1·68]	3·98 [4·74 ,3·27]	60 ± 4
E. saligna	–2·50 ± 0·04	3·40 [3·78, 3·08]	1·81 [3·60, 0]	5·16 [6·72, 4·34]	2·34 [3·23, 1·73]	4·57	133
E. dunnii	–2·27 ± 0·09	4·93 [5·46, 4·49]	2·18 [2·76, 1·56]	8·44 [11·12, 7·12]	2·42 [2·86, 1·84]	3·55 [4·74, 2·37]	156 ± 31
E. grandis	–2·34 ± 0·04	3·83 [4·16, 3·54]	2·14 [2·62, 1·55]	5·63 [7·16, 4·90]	1·95 [2·33, 1·50]	2·28 [2·61, 1·89]	165 ± 13

Errors presented are standard errors, and numbers in square brackets are [lower, upper] 95 % confidence intervals, respectively.

No standard error is given for E. saligna VIGR as there was only one replicate plot.

In cases where upper and lower confidence intervals could not be estimated accurately, they are not given.

g_s5 and g_s50 are the Ψ_stem at complete stomatal closure (5 % of g_smax) and the beginning of stomatal closure (50 % of g_smax), respectively.

Fig. 3.

Correlations of Ψ_TLP with the point of stomatal closure (g_s50), P₅₀ and VIGR for six climatically contrasting Eucalyptus species. Correlations and their significance are presented in Table S1 where P = (A) 0·03, (B) 0·004 and (C) 0·05, respectively. Dashed lines are shown as a guide. Error bars are the s.e.m. for each species or climate parameter, and are bidirectional.

Stomatal and hydraulic responses to Ψ_stem

Stomatal conductance declined as Ψ_stem became progressively more negative in all species (Fig. 4). Species from humid regions tended to have higher maximal g_s, lower minimum g_s and a steeper decline in g_s for a given drop in Ψ_stem compared with species from sub-humid regions (Fig. 4), though E. saligna had comparatively low g_s for a humid species (Fig. 4D). Further, sub-humid species tended to maintain some level of g_s at Ψ_stem as low as –6 MPa (Table 2; Fig. 4A, B), whereas humid species had a faster decline in g_s, completely closing stomata (g_s5) around or before –4 MPa (Table 2; Fig. 4C–F). Stomatal closure (g_s50) started at a Ψ_stem of –3·5 and –3·3 MPa in the sub-humid species E. cladocalyx and E. crebra (Table 2; Fig. 4A, B), whereas in the intermediate and humid species E. tereticornis, E. saligna, E. dunnii and E. grandis, g_s50 was at less negative Ψ_stem of –2·43, –2·34, –2·42 and –1·95 MPa, respectively (Table 2; Fig. 4C–F). Leaf level transpiration (E_L) followed the same pattern as g_s for all species (data not shown).

Fig. 4.

The response of g_s to declining Ψ_stem of six climatically contrasting Eucalyptus species. The black dashed line is a spline to highlight the relationship only. Vertical, grey lines represent Ψ_TLP.

Vulnerability to drought-induced embolism varied between the six species (Fig. 5). Species from sub-humid regions tended to have a shallower increase in percentage loss of conductivity (PLC) for a given drop in Ψ_stem (Fig. 5A, B) compared with humid zone and intermediate species (Fig. 5C–F). The sub-humid species, E. cladocalyx and E. crebra, maintained a PLC of < 50 % until Ψ_stem –5·3 and –4·9 MPa, respectively (Table 2; Fig. 4A, B) whereas the intermediate and humid species, E. tereticornis, E. saligna, E. dunnii and E. grandis, reached P₅₀ at –3·9, –3·4, –4·9 and –3·8 MPa, respectively (Table 2; Fig. 4C–F). For all but E. dunnii, sub-humid species P₅₀ was significantly more negative than humid species P₅₀, based on the 95 % CI (Fig. 4). Water potential at turgor loss (Ψ_TLP) was more negative in sub-humid than humid species, negatively correlated with g_s50 and P₁₂ (R² = 0·89 and 0·64, respectively; Fig. 3B, C) and in all species was similar to, and within the confidence limits of, the Ψ_stem at P₁₂ and g_s50 (Table 2; Fig. 4). The difference between the lower 95 % CI of P₅₀ and upper 95 % CI of g_s50 for E. cladocalyx, E. crebra, E. tereticornis, E. saligna, E. dunnii and E. grandis was –0·50, 0·36, 0·21, –0·15, 1·63 and 1·21, respectively (Table 2). Accordingly, species from humid regions tended to have no overlap between the Ψ_stem at P₅₀ and g_s50, starting to close stomata on average 0·72 MPa before P₅₀ was reached, whereas in sub-humid species g_s50 tended to be within the 95 % CIs of P₅₀ (Table 2).

Fig. 5.

The response of PLC to declining Ψ_stem of six climatically contrasting Eucalyptus species. The dashed line represents the relationship for all PLC vs. Ψ_stem data. Grouped vertical lines indicate P₅₀ (dark grey centre line) and the 95 % CI (light grey adjacent lines) around the mean as obtained through R (R Development Core Team, 2014) using the fitplc package (Duursma and Choat 2017). The separate grey vertical line indicates the Ψ_TLP, as in Table 2 and Fig. 4.

DISCUSSION

Climate range related to leaf tissue water relations, xylem anatomy, and stomatal and hydraulic characteristics

We found evidence supporting our first hypothesis, that xylem anatomy, leaf tissue water relations, and stomatal and hydraulic characteristics would correlate with the MAP and D of a species range when species are grown under common environmental conditions. We observed that species from more arid regions, indicated by a lower 1/D_max and MAP, had lower Ψ_TLP, Π₀, VIGR and D_h and lower (more negative) P₅₀ and g_s50 than species from more humid regions (Fig. 2). At the leaf level, a more negative Π₀ increases the cell’s capacity to maintain turgor at lower leaf water status (White et al., 2003; Bartlett et al., 2012); more negative Π₀ and Ψ_TLP also allow stomata to remain open at lower water potentials. Consequently, these characteristics facilitate stomatal closure at lower water potentials (more negative g_s50) and may confer a competitive advantage in sub-humid environments by enabling continued photosynthesis under water-limited conditions. Our results suggest that a hydraulic system oriented towards high cavitation resistance is favoured in regions where drought is frequent; however, these adaptations constrain the efficiency of water transport, subsequently limiting growth and disadvantaging the plant in areas of high water availability. Consistent with this, Pfautsch et al. (2016) reported that Eucalyptus species growing in arid environments have narrower xylem vessels than those growing in wet environments. The available data therefore strongly support the hypothesis that hydraulic traits of Eucalyptus species are adapted to the water availability of their native environment, consistent with Mitchell and O’Grady (2015) who reported strong selective pressures of aridity for shaping Ψ_TLP across 174 Australian species.

The correlations between hydraulic traits and species climate range provide evidence that these traits are genetically constrained in Eucalyptus. We observed that sub-humid species with distributions extending further inland (E. cladocalyx and E. crebra) had a higher drought tolerance than humid species with coastal distributions (E. tereticornis, E. saligna, E. dunnii and E. grandis), when considering drought tolerance to be a function of leaf tissue water relations (Fig. 2; Table 2), xylem anatomy (Fig 2; Fig. S1), stomatal behaviour (Fig. 4) and vulnerability to cavitation (Fig. 5). A study examining variation in xylem anatomical traits of Eucalyptus species growing under different rainfall regimes reported low intraspecific variation between populations, an observation consistent with low plasticity in these traits (Pfautsch et al., 2016). Similarly, previous work has suggested that phenotypic plasticity in cavitation resistance is limited (Martinez-Vilalta et al., 2009; Lamy et al., 2014), although there is some evidence of plasticity in this trait from other studies (Lopez et al., 2013; Schuldt et al., 2015). Consistent with the results of the present study, other common garden experiments have shown that differences in cavitation resistance are maintained when water availability is held constant (Sparks and Black, 1999; Vander Willigen et al., 2000), providing further evidence that this trait is under strong genetic control.

Co-ordination of stem hydraulics with stomatal behaviour

In order to test our second hypothesis, that there is a co-ordination of stem hydraulic traits with leaf tissue water relations and the point of stomatal closure across species, we compared these characteristics across six climatically contrasting Eucalyptus species. We observed tight correlations of Ψ_TLP with the point of stomatal closure (g_s50) and P₁₂ (Fig. 3), and a tight correlation between g_s50 and P₁₂ (R² = 0·68, P = 0·04; Table S1) across all species, indicating a co-ordination of stem hydraulics with leaf tissue water relations and the point of stomatal closure. A correlation between Ψ_TLP and g_s50 has previously been reported (Brodribb et al., 2003) and can be expected as stomata will close when guard cells lose turgor (Brodribb et al., 2003; Martorell et al., 2014). The results presented here suggest that Eucalyptus species close stomata prior to the occurrence of significant cavitation (P₅₀), thus protecting the hydraulic system of the stem from dysfunction during periods of drought. We observed only one species, E. cladocalyx, for which Ψ_stem at g_s50 and P₅₀ were statistically similar (overlap in 95 % CIs; Table 2), while species from humid regions tended to have greater margins, with an average of 0·73 MPa difference in the 95 % CIs of these parameters (Table 2; Figs 4 and 5). This is consistent with recent work showing that stem and leaf traits are co-ordinated in woody species such that stomatal closure occurs before critical thresholds for cavitation are crossed in the stem (Delzon and Cochard, 2014; Nolf et al., 2015; Bartlett et al., 2016).

Co-ordination of leaf and stem hydraulics with stomatal behaviour allows the plant to work within narrow safety margins (Choat et al., 2012), ensuring minimal cavitation (safety) while still enhancing transpiration (efficiency) and subsequently productivity, thus benefiting the plant regardless of the environmental conditions it encounters. Consistent with this, we observed a tight relationship between volume growth rates (VIGR) and physiological traits such as leaf water relations (Ψ_TLP and Π₀) and the point of stomatal closure (g_s50, g_s12 and g_s5). This confirms that variation in the hydraulic strategies of these species is translated into differences in productivity and growth at the whole-plant level. Overall, our results support the existence of a fast–slow spectrum of functional traits in Eucalyptus from different climate ranges (Reich, 2014), with species from more arid climates occupying the slow (safe) end of this trait spectrum.

Conclusions

Species climate range was tightly correlated to leaf tissue water relations, xylem anatomy, the Ψ_stem at stomatal closure (g_s50) and hydraulic characteristics of six different Eucalyptus species. Additionally, we observed co-ordination between stem hydraulics, leaf tissue water relations and the point of stomatal closure, which has been hypothesized but infrequently tested, and a difference in water use strategy linked to species climatic range. We observed that species from more humid regions exhibited characteristics of an efficient hydraulic system, preventing hydraulic dysfunction through a conservative approach to stomatal closure. In contrast, sub-humid species with ranges extending into more arid regions maintained a safe hydraulic system, limiting g_s and cavitation with narrower xylem vessels, and maintaining leaf tissue characteristics tolerant of water limitation. By conducting this study on species from a single genus and measuring all species in a common garden, we eliminated much of the variation that could be attributed to different climatic or environmental characteristics. Despite minimizing this variation, we observed a significant correlation of climate with leaf tissue water relations, xylem anatomy, stomatal behaviour and hydraulic characteristics between Eucalyptus species from contrasting climates. Additionally, we observed a tighter co-ordination of stem hydraulics with leaf tissue water relations and the point of stomatal closure in humid compared with sub-humid species, indicating a difference between species in the water use strategy employed. These observations indicate that the species climatic range has a large influence on shaping species responses to the environment, supporting the theory that genetic factors influence the expression of leaf tissue water relations, xylem anatomy, stem hydraulics and stomatal behaviour. In these species, it is likely that genetics control plant structural traits in particular, since they pertain to xylem anatomy and thus are not easily modified when growing in different climates. These findings hold important implications for species survival and adaptations to changes in climate, and may assist plantation and land management decisions to ensure survival of native and planted ecosystems under future climate scenarios.

SUPPLEMENTARY DATA

Supplementary data are available online at https://academic.oup.com/aob and consist of the following. Figure S1: transverse sections of branch xylem tissue imaged with bright field microscopy at × 40 magnification for E. cladocalyx, E. crebra, E. tereticornis, E. saligna, E. dunnii and E. grandis. Table S1: pairwise correlation matrix for hydraulic and anatomical parameters with variables related to the species climate range. Table S2: list of abbreviations.