Understanding plant drought resistance in a Mediterranean coastal sand dune ecosystem: differences between native and exotic invasive species

Abstract

Aims

Mediterranean coastal dunes are habitats of great conservation interest, with a distinctive and rich flora. In the last century, Acacia spp., native from Australia, have been introduced in Portugal, with the objective of stabilizing sand dunes, and since have become dominant in numerous sand dune habitats. This invasion process led to the reduction of native plant species richness, changed soil characteristics and modified habitat’s microclimatic characteristics. The aim of this research was to typify and compare, in Mediterranean sand dune ecosystems, the ecophysiological responses to drought of Helichrysum italicum and Corema album, two native species, and Acacia longifolia, an exotic invasive species. We addressed the following specific objectives: (i) to compare water relations and water use efficiencies, (ii) to evaluate water stress, (iii) to assess water use strategies and water sources used by plants and (iv) to evaluate the morphological adaptations at leaf and phyllode level.

Methods

In order to obtain an integrative view of ecophysiological patterns, water relations and performance measuring methods have been applied: predawn (ψ_PD) and midday (ψ_MD) water potential, chlorophyll a fluorescence, oxygen isotopic composition of xylem, rain and groundwater (δ¹⁸O) and leaf carbon isotopic discrimination (Δ¹³C). The leaf characteristics of the three species, as well as the histochemistry of non-glandular trichome cell walls, were also studied to identify morpho-traits related to drought resistance.

Important Findings

The results support our initial hypothesis: although A. longifolia clearly possesses a degree of resistance to water stress, such ability is provided by a different water strategy, when compared to native species. Natives relied on morphological adaptations to restrict water loss, whereas the invasive species adjusted the water uptake as a way to balance their limited ability of restricting water loss. We corroborate that woody native species (i) have a conservative water-saving strategy and minor seasonal variations relative to invasive species, (ii) use enriched water sources during drought periods, indicating different water sources and root systems comparing with invasive species and (iii) present drought leaf morpho-functional adaptations related with limiting water loss. Comparing the physiological performance of invasive and native species can offer causal explanations for the relative success of alien plant invasions on sand dunes ecosystems.

ecophysiology, water use strategy, leaf morphology, Acacia longifolia, Corema album, Helichrysum italicum

INTRODUCTION

Coastal sand dunes are highly diverse and are important for coastal defence against erosion, recreation and as reservoirs of biodiversity (Pardini et al. 2015; UNEP 2006). They comprise various environments, presenting a collection of habitats that change over time and space and undergo critical periods of stress (Garcia Novo et al. 2004). At a global scale, coastal dunes are regularly subjected to the same types of abiotic stress factors, such as salt spray, episodic overwash, highly permeable substrate, substrate mobility, low field capacity, high irradiation and temperatures, drought and high winds (Garcia Novo et al. 2004; Hesp 1991). Sandy dune vegetation is well adapted to strong environmental gradients and zonation, allowing for a wide variety of functional groups to coexist within a relatively small area (Bionti 2007; Carboni et al. 2010). In Mediterranean regions, apart from these factors, sand dune vegetation is also influenced by strong seasonality, with two distinct periods: rainy, cold winters and dry, hot summers with high irradiance and little or no precipitation. Thus, plants experience intense stress in the summer, caused by drought, large evaporative demand and high irradiance at high temperatures (e.g. Bionti 2007; Carboni et al. 2010; Zunzunegui et al. 2005).

Additionally, coastal sand dune ecosystems are highly susceptible to biological invasions due to frequent disturbances and the existence of open patches free of plant competition, resulting in threats and ecological impacts that occur globally (Carboni et al. 2010; Del Vecchio et al. 2013; Lorenzo et al. 2012; Pardini et al. 2015). These ecosystems are particularly threatened by exotic plants that have been intentionally introduced for dune stabilization (Marchante 2003, 2008, 2015; Peperkorn et al. 2005). Competition for the resources to support the invasive’s growth capacity depends on characteristics of both the invaded area and the invader’s biological traits (Rascher et al. 2010; Huang et al. 2018). Moreover, the invasiveness of exotic species is related with (i) specific traits that allow for success in the new environment and (ii) on the interaction between the invader and the organisms of the invaded plant community (Dietz and Steinlein 2004; Redmond and Stout 2018). Interspecific variation in the responses of native and invasive species may lead to community-level changes in species dominance, composition, diversity and functioning (Pysek et al. 2012; Vilà et al. 2010; Gioria et al. 2018). It has therefore been pointed out that there is an urgent need for more community-level studies under natural field conditions to understand climate change effects on both invasive and native species and their habitats (Badalamenti et al. 2014; Weltzin et al. 2003). The understanding of the mechanisms and processes behind the ability of exotic invasive species to compete with native species is particularly relevant under Mediterranean climatic conditions, where seasonality will demand a strong physiological regulation of water use in response to soil water depletion and water vapour changes under typical drought conditions (Alessio et al. 2004; Nardini et al. 2014; Serrano et al. 2005).

Mediterranean plant communities are often dominated by woody species and present some common functional traits, such as sclerophylly, low growth rate, low water and nutrient availabilities and a significant presence of sprouting species (Mooney 1989). Under Mediterranean climate conditions, plant species acquired morphological characteristics in response to the severe drought periods. At leaf level, xerophytic sclerophyllous vegetation can exhibit features such as increased thickness of the outer epidermal walls and cuticles, presence of epicuticular waxes, high trichome density, sunken stomata or stomata confined to pits and grooves on the abaxial surface of the leaf, increased development of palisade parenchyma in the mesophyll with a corresponding decrease of spongy parenchyma and extensive sclerenchyma development (De Micco and Aronne 2012). Also, they can present temporary specializations such as leaf loss, changes in leaf angle and rolling and folding of leaf blades (Volaire et al. 2009). The morphological characteristics are accompanied by a set of physiological adaptations, which together determine plants’ strategies to resist drought (Chaves et al. 2003; Fotelli et al. 2000; Nardini et al. 2014; Sardans and Peñuelas 2013). In coastal dunes’ harsh conditions, plant resistance, which is the plant’s ability to minimize the negative impact of environmental adverse conditions, relies either on avoidance or tolerance (Levitt 1972; Lo Gullo and Salleo 1988; Puijalon et al. 2011). Avoidance entails traits that enable plants to resist adverse conditions by preventing their deleterious effects (via enhanced water uptake and reduced water loss) whereas tolerance consists in traits that enable plants to endure adverse conditions (via osmotic adjustment, antioxidant capacity and desiccation tolerance) (Levitt 1972; Puijalon et al. 2011). Regardless, the adaptation of specific plant species to periods of drought stress is not limited to the use of a single mechanism, but can use both avoidance and tolerance traits as a means to survival (Nilsen and Orcutt 1996). Some of the most common traits involved in drought resistance include the development of deep and extensive root systems (Baldocchi and Xu 2007; Padilla and Pugnaire 2007; West et al. 2012), foliar sclerophylly, high density of foliar trichomes (Lo Gullo and Salleo 1988; Sardans and Peñuelas 2013), low xylem cavitation vulnerability and leaf protections (Vilagrosa et al. 2010). In Portugal, coastal dunes are endangered by the invasion of non-native Acacia species, which have been introduced deliberately at the beginning of the last century with the objective of stabilizing sand dunes and for ornamental purposes (Marchante et al. 2003, 2008). The so called Australian Acacia species or ‘wattles’ have been planted around the world and particularly in the Mediterranean regions (Alberio and Comparatore 2014; Hui et al. 2011; Richardson and Rejmánek 2011). Such is the case of Acacia longifolia, whose introduction in Portuguese sand dune systems led to the decrease of native plant diversity and richness while increasing total vegetation cover through the formation of dense, monospecific Acacia stands (Marchante 2003, 2008, 2015; Rascher et al. 2011a). Acacia longifolia alters soil properties and nutrient availability (e.g. Marchante et al. 2015; Rascher et al. 2012) changing soil nutrient cycling, with substantial inputs of N-enriched litter which leads to changes in neighbourhood natives since the very early stages of invasion (Hellmann et al. 2011; Rascher et al. 2012). It also has a marked impact on ecosystem water cycling (Rascher et al. 2010, 2011b) due to high resource utilization (Funk 2013; Marchante et al. 2008; Werner et al. 2010), symbiotic N₂-fixation (Morris et al. 2011; Rascher et al. 2012) and a water spender strategy (Peperkorn et al. 2005; Rascher et al. 2011b). The increased water use by Australian acacias in invaded areas (Le Maitre 2004; Le Maitre et al. 2000; Rascher et al. 2010, 2011b) may be a reflection of a larger above-ground biomass when compared with native vegetation (Morris et al. 2011), as well as of increased transpiration rates per leaf area (Máguas et al. 2011). Plant water uptake is dependent on the size, surface area and depth of its roots as well as how these roots are spatially distributed across the soil profile (Schenk and Jackson 2002). Further information about rooting depth and water used by plants is imperative in order to understand the ability of invasive Australian acacias to access deep water (and associated nutrient sources).

In this paper, we aim to compare the response of native and exotic invasive plant species to drought in a sand dune habitat, in order to understand which morpho-functional traits better explain survival, distribution and competitive success. Two drought adapted native shrubs, Helichrysum italicum ssp. picardi and Corema album were compared with Acacia longifolia, an exotic invasive species in Portuguese coastal sand dune ecosystems. We tested the hypothesis that although A. longifolia clearly shows resistance to water stress, contributing to its invasive character, such ability is provided by a different strategy of water relations and response to drought, when compared to native species. With that aim, we addressed the following specific parameters, comparing native and invasive species: (i) water relations in terms of water potential—predawn (ψ_PD) and midday (ψ_MD)—and water use efficiency, (ii) water stress, by means of chlorophyll a fluorescence, (iii) main water used by plants, through isotopic composition (δ¹⁸O) and better understanding of their root system and (iv) morphological adaptations at leaf and phyllode level. Comparing the physiological performance of invasive and native species can offer causal explanations for the relative success of alien plant invasions in sand dunes ecosystems.

MATERIALS AND METHODS

Study site and selected species

The study site is a coastal dune system located at Pinheiro-da-Cruz (38°14′N, 8°46′W), 70 km Southeast of Lisbon, Portugal. In the study area, summers are considered dry with high temperatures, and precipitation mostly concentrated in the winter months. Soils have poor water retention capacity (Rascher et al. 2011a), low organic matter content and low N and P availability (Hellmann et al. 2011). Due to low summer precipitation, and thus reduced groundwater recharge, the water table is deeper in the summer months and most likely with a depth of >10 m (based on regional information, Mendonça et al. 2003; Outras 2009), therefore hardly accessed by plants. The coastal dune system is characterized by natural vegetation that follows the morphological gradient of the dune. Fieldwork was conducted on the secondary dunes, which are mainly constituted by shrubby species such as C. album, Stauracanthus spectabilis, Santolina impressa, H. italicum, Thymus carnosus and Halimium calycinum forming an open community with low cover. Acacia longifolia is also present in our study area, having been introduced in the Portuguese littoral, mainly during the 1970s, to stabilize dune systems.

The species C. album (L.) D. Don, H. italicum ssp. picardi (Boiss. & Reuter) Franco (afterwards referred as H. italicum) and A. longifolia (Andrews) Willd. were selected for this study, since they have an ubiquitous distribution in the Portuguese coastal dunes, namely in the foredunes where the ecological stresses are stronger. Helichrysum italicum and C. album are dominant and well adapted to the stressful dune conditions, whereas A. longifolia is an exotic invasive in this habitat. The native species H. italicum is a chamaephyte from the Asteraceae family, with lanceolate, revolute and densely pubescent leaves and C. album, an endemic species of the Atlantic coast of the Iberian Peninsula, is a nanophanerophyte plant from the Ericaceae family (Castroviejo et al. 1993). The latter evergreen shrub, which rarely exceeds 1 m in height, is densely branched from the base with whorled ericoid bright leaves that persist for only two growing seasons (Castroviejo et al. 1993). Acacia longifolia, an invasive Australian legume (Fabaceae), is a phanerophyte species with heteroblasty, i.e. distinct juvenile and adult leaf forms, compound leaves on the seedling stage and phyllodes on the adult stage.

Climatic conditions

Climatic data (monthly mean temperature, T_m, and accumulated monthly precipitation, P) were obtained from the nearest meteorological station, Alcácer do Sal (38º23’ N, 8º31’ W) ~20 km from Pinheiro-da-Cruz. During ecophysiological measurements microclimatic conditions such as air temperature (T_air), relative humidity (RH) and photosynthetic active radiation (PAR) were recorded every hour with sensors coupled to a Hobo data logger (H08-007-02, Onset Computer Corporation, USA). Vapour pressure deficit (VPD) was calculated from air temperature and RH values.

Ecophysiological measurements

For each of the studied species, 3–5 randomly isolated individuals were chosen to perform ecophysiological measurements, which were carried out under contrasting environmental conditions: in winter (14 February), spring (13 May) and summer (28 July).

Stem water potential

Water potential (Ψ) of five individuals per species, considering one small terminal shoot (H. italicum and C. album) or phyllode (A. longifolia) per each individual, was measured during predawn before 6 a.m., and hourly from 9 a.m. to 4 p.m., with a Scholander-type pressure chamber (Manofrígido, Portugal) (Turner 1981).

Chlorophyll a fluorescence

Effective quantum yield (Φ_PSII) was measured in the field with a portable pulse-modulated fluorimeter (Mini-Pam Photosynthesis Yield Analyzer, Walz, Effeltrich, Germany). The measurements were performed in three small terminal leafs or phyllodes per plant, with similar positions and orientations in the canopy, during predawn and hourly from 9 a.m. to 4 p.m. Φ_PSII or actual efficiency of energy conversion in PSII in a light adapted state was estimated according to Genty et al. (1989) as ${\Phi }_{\text{PSII}}=({F^{\prime }}_{\text{m}}-F)/{F^{\prime }}_{\text{m}}$⁠, where ${F^{\prime }}_{\text{m}}$ and F are maximal and steady state fluorescence under actinic irradiance, respectively.

Water isotopic analyses (δ¹⁸O)

The composition of stable isotopes in water remains unaltered during plant root absorption. Thus, when the isotopic ratio of the available water sources is known, an analysis of the oxygen isotopic ratio of xylem water provides information about the water sources being used by the plant at the time of study (Chimner and Cooper 2004; Ehleringer and Dawson 1992).

Three lignified and mature stems of three to five individuals per species were collected at midday (0–2 p.m.) in each season and immediately stored in watertight vials. Xylem water from sampled twigs was extracted by cryogenic vacuum distillation (Ehleringer and Dawson 1992; Ellsworth and Williams 2007). To determine available water sources for the roots, samples from rainwater and from the water table were collected at the same dates. Rainwater was collected in pluviometers with added liquid paraffin to prevent evaporation. Groundwater samples were extracted in a well that reached the water table near the study site. All samples were kept frozen at −20°C until analysis to prevent isotopic fractionation. Oxygen stable isotope ratio analyses were performed at the Stable Isotopes and Instrumental Analysis Facility (SIIAF), Centro de Ecologia, Evolução e Alterações Ambientais (CE3C), at the Faculdade de Ciências, Universidade de Lisboa, Portugal, by headspace equilibration, on an Isoprime (Micromass, UK) SIRMS, coupled on continuous flow mode to a Multiflow (Micromass, UK) auto-sampler and sample equilibration system. Delta calculation was performed according to δ = [(R_sample − R_standard)/R_standard] × 1000, where R is the ratio between the heavier isotope and the lighter one. δ¹⁸O values are referred to V-SMOW (Vienna Standard Mean Ocean Water). The working reference materials used were Medium Natural Water (Elemental Microanalysis Ltd, UK; δ¹⁸O_V-SMOW = −10.18 ± 0.2‰) and Zero Natural Water (Elemental Microanalysis Ltd, UK; δ¹⁸O_V-SMOW = 0.56 ± 0.23‰), regularly checked against IAEA-VSMOW and IEAE-GISP (Coleman and Meier-Augenstein 2014). Analytical precision is <0.1‰.

Carbon isotope discrimination (Δ¹³C)

Samples consisted in 8–10 leaves or phyllodes from 3 to 5 individuals selected. Leaves were oven-dried at 60°C for 48 h and milled to fine powder for carbon isotopic analysis. Carbon stable isotope ratio analyses were performed at SIIAF. δ¹³C_VPDB was determined by continuous flow isotope mass spectrometry (Preston and Owens 1983), on a Sercon Hydra 20–22 (Sercon, UK) stable isotope ratio mass spectrometer, coupled to a EuroEA (EuroVector, Italy) elemental analyzer for online sample preparation by Dumas-combustion. Delta calculation was performed according to δ = [(R_sample − R_standard)/R_standard] × 1000, where R is the ratio between the heavier isotope and the lighter one. δ¹³C values are referred to PDB (Pee Dee Belemnite). The reference materials used were Sorghum Flour Standard OAS and Wheat Flour Standard OAS (Elemental Microanalysis, UK) with, respectively, δ¹³C_VPDB (Sorghum Flour OAS) = −13.68 ± 0.19‰ and δ¹³C_VPDB (Wheat Flour OAS) = −27.21 ± 0.13‰ (Coleman and Meier-Augenstein 2014). Precision of the isotope ratio analysis, calculated using values from six to nine replicates of secondary isotopic reference material (wheat flour, δ¹³C_VPDB = −27.27 ± 0.19‰) interspersed among samples in every batch analysis, was ≤0.2‰. Carbon isotope discrimination (Δ¹³C) was calculated accordingly to Farquhar et al. (1982) for C3 plants.

Leaf structure and histochemistry

Leaf morpho-functional adaptations of the three species were studied focusing on traits related to the environmental conditions of sand dunes habitats, such as water availability, and high temperatures and radiation. Mature leaves from H. italicum and C. album and mature phyllodes from A. longifolia were collected from three specimens of each species and prepared for scanning electron microscope and light microscope observations.

Scanning electron microscopy

For scanning electron microscopy, samples of ~0.5 cm² were excised from the middle region of the leaf blade and phyllode (n = 6 for each species) and fixed with 2.5% glutaraldehyde in 0.1 M sodium phosphate buffer, at pH 7.2 for 4 h at 4°C. The material was then washed in the fixative buffer, dehydrated in a graded acetone series, critical-point dried with CO₂ and coated with gold. Observations were carried out on a JEOL T220 scanning electron microscope (JEOL Ltd., Tokyo, Japan) at an accelerating voltage of 15 kV. Images were recorded in a Kodak T-Max 100 professional black-and-white negative film using a MAMIYA camera.

Light microscopy

For the histochemical characterization of total lipids and pectins, transverse hand-cut sections from the middle region of the leaf blade and phyllode (n = 12 for each species) were stained with Sudan IV and ruthenium redss, respectively. Observations were made with a Leica DM-2500 microscope (Leica Microsystems CMS GmbH, Wetzlar, Germany), and images were recorded digitally using a Leica DFC-420 camera (Leica Microsystems Ltd., Heerbrugg, Switzerland).

Statistical analysis

To test differences between species in each season (for δ¹⁸O and Δ¹³C variables), a non-parametric analysis of variance (Kruskal–Wallis test) was carried out since the assumption of normality was not found. For correlations between VPD and Ψ and between Ψ and Φ_PSII, the Spearman correlation coefficient (ρ) was calculated using mean values of each sampling hour per species (n = 17). Data analysis was performed using Statistica 9.0 (StatSoft, Tulsa, OK, USA) and R version 3.2.3.

RESULTS

During the study period, climatic conditions followed a typical-Mediterranean climatic pattern with two distinct periods: rainy cold winter and dry hot summer with high irradiance and low or no precipitation (Fig. 1). The average annual precipitation was 405 mm indicating a very dry year with a monthly mean temperature of 16°C. Higher precipitation occurred during autumn and winter, until May. A pronounced drought from June to August was observed with a total of 18 mm of rain during the 3-month span. The lowest temperatures were observed in December and February (9 and 10°C, respectively) and the highest in the summer months of June, July and August (mean temperature of 22.5, 23.5 and 22.5°C, respectively) (Fig. 1).

Figure 1:

climatic conditions in the study site. Bars represent total monthly rainfall and points represent average monthly air temperature. Arrows indicate the period of ecophysiological measurements.

Accordingly, the microclimatic variables presented a typical pattern for these seasons, with VPD, air temperature (T_air) and PAR increasing from morning to afternoon with higher values during summer (Fig. 2A). In general, A. longifolia showed more negative water potentials (Ψ) than native species, especially during the dry season (Fig. 2B). In winter, the three species showed a similar Ψ pattern during the day and the Ψ values found were not significantly different. In spring and summer, with higher PAR and concomitantly higher VPD (30 mbar), A. longifolia showed significantly more negative Ψ from 9 a.m. to the end of afternoon, with values decreasing more than twice those presented by both native species (Fig. 2B). In spring, C. album and H. italicum showed Ψ values that were similar and significantly higher than the exotic species during the day; however, in summer, C. album Ψ values also dropped significantly, when compared to H. italicum, reaching values of −1.87 MPa (Fig. 2B). Still, no significant differences were observed at predawn between the exotic and the native species (Fig. 2B). Among the three studied species, only H. italicum presented similar Ψ values in all seasons (varying from −0.2 to −0.93 MPa).

Figure 2:

microclimatic conditions during measurements: (A) VPD, temperature (T_air) and PAR; (B) water potential (Ψ) and (C) effective quantum yield of Helichrysum italicum (□), Corema album (▲) and Acacia longifolia (●), measured throughout the day in winter (February), spring (May) and summer (July). Values are means ± standard deviation.

Both exotic and native species showed the highest efficiency of PSII in winter and in predawn, with values of effective quantum yield of photosystem II (Φ_PSII) between 0.75 and 0.64 (Fig. 2C). No significant differences were detected in this season between exotic and native species (Fig. 2C). The highest daily decline of Φ_PSII was observed in spring and summer after sunrise in all the three species studied. Helichrysum italicum reached significantly lower values of Φ_PSII in spring when compared to the other species after midday (Fig. 2C). In summer, from 11 a.m. to 4 p.m., the Φ_PSII of the native species H. italicum had no substantial variation, contrasting with A. longifolia that showed a pronounced decline of Φ_PSII (reaching values of 0.11) (Fig.2C). The Φ_PSII values of C. album were more or less stable after sunrise, but increased at midday in spring and summer, reaching values of 0.64 and 0.39, respectively (Fig. 2C).

In order to assess the studied species’ main water sources, oxygen isotopic ratio (δ¹⁸O) of precipitation, groundwater and xylem water was compared. Seasonal δ¹⁸O of precipitation was −1.9 ± 0.2‰ in winter and −3.0 ± 0.1‰ in spring. No precipitation occurred in July. δ¹⁸O of groundwater remained stable across seasons: −4.4 ± 0.1‰ in winter, −4.3 ± 0.1‰ in spring and −4.6 ± 0.1‰ in summer (Fig. 3). The comparison of the isotopic ratio of water sources and xylem water indicates that both native and exotic species were using different water sources, depending on the season. In winter, the native species showed δ¹⁸O values matching the rain signal, contrasting with A. longifolia that was using water with a more depleted signal, from deeper soil layers (closer to the groundwater isotopic ratio) (Fig. 3). In spring, a similar oxygen isotopic ratio on all the three species indicates that they were using mainly the same water source (precipitation) (Fig. 3). In summer, H. italicum presented a significantly different δ¹⁸O xylem water signal (close to 0‰), which reveals that this species was using topsoil or dew water during this period (Fig. 3). Despite the relative increase of δ¹⁸O in summer, A. longifolia was still using more depleted water, indicating that it was tapping deeper soil layers when compared to shrubby native species, where water probably reflects the isotopic composition of winter precipitation (Fig. 3). None of the studied species was using groundwater as the main source in summer. The differences in water sources between the species were more evident between H. italicum and A. longifolia than between C. album and A. longifolia (Fig. 3).

Figure 3:

oxygen isotopic composition (δ₁₈O) of rainwater, groundwater and xylem water of Helichrysum italicum, Corema album and Acacia longifolia, collected in winter, spring and summer. Horizontal lines indicate isotopic signal of rain and groundwater (see details in inner legend), and standard boxplots represent xylem water isotopic signal for each species. Different letters represent significant differences between species within a season (KW P < 0.05).

A significant negative correlation between VPD and Ψ was found for A. longifolia (rho = −0.96, P < 0.001) and C. album (rho = −0.91, P < 0.001) (Fig. 4A). On the contrary, no significant correlation between these two variables was verified for H. italicum (P = 0.184). For the same conditions of increasing VPD (from 4 to 48 mbar), the native species H. italicum kept the Ψ values between −0.19 and −0.96 MPa, contrasting with A. longifolia that decreased linearly the Ψ values from −0.49 MPa to −3.68 MPa (Fig. 4A). Φ_PSII was positively and significantly correlated with Ψ for A. longifolia and H. italicum (rho = 0.75, P < 0.01 and rho = 0.54, P = 0.03, respectively) (Fig. 4B). Helichrysum italicum showed a sharper decrease in Φ_PSII in a smaller range of decreasing Ψ (Fig. 4B).

Figure 4:

correlation between (A) VPD and water potential (Ψ) and (B) water potential (Ψ) and effective quantum yield (Φ PSII) for Acacia longifolia, Corema album and Helichrysum italicum considering all sampling dates. rho and regression lowess line are present for significant correlations (P < 0.05).

The relationship between the leaf carbon isotope discrimination (Δ¹³C) and the difference between midday and predawn leaf water potential (Ψ_MD − Ψ_PD) showed distinct patterns depending on season and species (Fig. 5). The three species showed an increasing disparity from winter to summer, presenting a higher similarity in Δ¹³C and Ψ_MD − Ψ_PD in winter and a higher discrepancy in summer (Fig. 5). The native species, H. italicum maintained Δ¹³C and Ψ_MD − Ψ_PD across seasons, showing significant differences: (i) in Δ¹³C, with C. album in all seasons and with A. longifolia in spring and (ii) in Ψ_MD − Ψ_PD, with C. album only in summer and with A. longifolia in spring and summer (Fig. 5). The exotic species, A. longifolia, decreased its Δ¹³C from 20.4 to 17.8‰ from winter to spring and increased its Δ¹³C in summer_. The lowest Ψ_MD − Ψ_PD value was observed in A. longifolia in summer (significantly lower compared with the other species) (Fig. 5). The highest values of Δ¹³C were observed in H. italicum (21.2‰ in winter and 21.2‰ in spring) and the lowest were found in A. longifolia in the spring (17.8‰) (Fig. 5).

Figure 5:

leaf carbon isotope discrimination (Δ₁₃C) and the difference between midday and predawn leaf water potential (Ψ_MD − Ψ_PD) for Helichrysum italicum, Corema album and Acacia longifolia in winter, spring and summer. See detailed species legend in the figure. Different capital letters represent significant Ψ_MD − Ψ_PD differences between species within a season, and different small letters represent significant Δ₁₃C differences between species within a season.

The leaves of H. italicum and C. album are small, linear and revolute, their lamina rolled from dorsal to ventral surface. In the first species, the mid-vein protrudes above the rolled leaf margins forming two chambers, while in the latter species, only the adaxial surface of the leaf is visible and the abaxial surface outlines a single cavity formed by the rolled leaf edges (Fig. 6A and C). Unlike the leaves of H. italicum that are velvety and nearly grey-whitish due to a very dense lanate indumentum that partially obscures the epidermis, the leaves of C. album are green and glossy (Fig. 6A and C). Both species show uniseriate epidermis, with thick, cutinized walled cells that stain intensely with Sudan IV (Fig. 7A and C), although the cuticles and the outer walls of the adaxial epidermal cells of C. album are thicker than those present at the abaxial surface (Figs 6C and 7C) and on both leaf surfaces of H. italicum (Figs 6A and 7A). On the two species, the abaxial surface of leaves is covered by uniseriate non-glandular trichomes and capitate glandular trichomes, with multicellular secretory heads and short stalks (Fig. 6B and D). The non-glandular trichomes of Helichrysum are smooth, very long and flexuous, do not react with Sudan IV and stain rose with ruthenium red (Fig. 7A and B). By contrast, non-glandular trichomes of Corema are erect, unicellular with thick walls that stain red with Sudan IV and give a negative reaction with ruthenium red (Fig. 7C and D). On both species, the stomata, confined to the grooves (stomatal crypts), are slightly raised above the abaxial surface and the mesophyll is characterized by two layers of palisade parenchyma and a spongy tissue with large intercellular spaces (Figs 6A–D and 7C).

Figure 6:

scanning electron micrographs of mature leaves and phyllodes from the three studied species. Cross sections of the revolute leaves of (A, B) Helichrysum italicum and (C, D) Corema album and (E, F) Acacia longifolia phyllodes. General views showing on (A) the abaxial surface the two grooves defined between the midrib and the leaf edges and (C) the single groove formed by the rolling of the leaf margins. Note in (A) the dense indumentum of flexuous non-glandular trichomes that covers the leaf adaxial surface and in (C) the very thick outer walls of the adaxial epidermal cells. (B, D) Details of the leaves spongy parenchyma and of the abaxial grooves. Slightly raised stomata (arrows), non-glandular trichomes and capitate glandular trichomes are well seen. (E) Cross section of a fully expanded phyllode. Note the compact isobilateral palisade parenchyma. (F) Stomata on the adaxial epidermis, where wax deposits are clearly seen. Scale bars: A, C, E = 100 μm; B, D, F = 50 μm.

Figure 7:

light micrographs of fresh hand-cut sections. (A–D) Cross sections of Helichrysum italicum and Corema album mature leaves. (A) The cuticle of the epidermal cells is Sudan positive, but the walls of the non-glandular trichomes do not react with this lisocrome. (B) Ruthenium red stained pink on the walls of non-glandular trichomes. (C) The epidermal cell walls and non-glandular trichome walls stained intensely red with Sudan IV but (D) do not show any reaction with ruthenium red. Cross sections of fully expanded phyllodes from Acacia longifolia staining with (E) Sudan IV and (F) ruthenium red. Note the isobilateral palisade parenchyma surrounding the central achlorophyllous parenchyma. Scale bars: 50 μm.

The ‘lamina’ of A. longifolia phyllodes is flat, with thick epidermal cell walls heavily cutinized (Figs 6E and 7E) and with epicuticular waxes (Fig. 6F). The stomata, numerous on both sides, are level with the epidermis or even somewhat elevated and protected by projecting ridges of cuticle (Fig. 6E and F). The photosynthetic tissue consists of a compact, two-layered palisade parenchyma on both adaxial and abaxial surfaces, surrounding five to six central layers of large cells without chloroplasts; a typical spongy parenchyma is absent. Vascular bundles with lignified fibbers caps occur in pairs, one on each side of the phyllode, while a single one is present along each margin (Figs. 6A and 7E and F).

DISCUSSION

On this oligotrophic, Mediterranean sand dune ecosystem, our results suggest that although both invasive and native species share a common resistance to drought, there is a clear diversity on the water use strategy. In broad terms, the native plants resist to drought by decreasing water loss while the exotic does so by maintaining water uptake. Both woody native species showed a conservative water use strategy relative to the invasive species, reducing water loss through stomatal control and with the concurrence of several morphological adaptations to withstand the harsh conditions.

The high water potential (Ψ) measured at predawn suggests that the three species avoided the drought and were not overall limited by water availability, even during summer. However, A. longifolia showed a limited ability to restrict water loss as well as indications of possible water limitation during drier periods of the day. Contrastingly to native species, the decrease of midday Ψ in A. longifolia under drier conditions points to an increase in water uptake, supporting the hypothesis that they display a water spending strategy, as already described in other studies (Máguas et al. 2011; Peperkorn et al. 2005). Diurnal Ψ is mainly determined by transpiration rate and the rate at which water is supplied to the leaves from soil via the conduction system. High transpiration rates in combination with lower conductivity rates resulted in low leaf Ψ. Leaf Ψ showed reciprocal responses to diurnal changes in transpiration rates declining to minima at or just after midday and recovering during the afternoon. Additionally, there was a seasonal maintenance of stomatal conductance from winter to summer and a strong response of water potential to variation in VPD in A. longifolia, which corroborates its limited ability to prevent water loss, as well as adjustments of water uptake to preserve the soil–plant–atmosphere continuum under drier conditions. This process could be facilitated within plant cells by osmotic adjustments, a mechanism that helps plants to acclimatize to the dry conditions. Thus, the water balance could be achieved by dropping Ψ, linked to increased water absorption, and, as a water spender, this species could potentially have an associated high transpiration rate. However, as soon as the absorption rate becomes insufficient to keep up with water loss, the water spenders will reduce the photosynthetic activity. The low Φ_PSII of A. longifolia in drier conditions seems to support this idea.

Native species H. italicum and C. album showed a seasonal maintenance of Ψ and a response to drought by lowering the Φ_PSII in drier conditions, indicating a water-saving strategy. The natives’ drought avoidance by restricting water loss would imply a higher stomatal control, which was not reflected in the values of Δ¹³C. As pointed out by Werner and Máguas (2010), leaf phenology governs seasonal responsiveness of Δ¹³C to drought, which over expresses its applicability as an indicator of water use efficiency, particularly in Mediterranean evergreen species with short growing periods. In fact, native species restrict biomass production to periods of higher water availability, reflecting the physiological conditions at this particular time, rather than the drought stressful period. However, Δ¹³C remains a good functional tracer and indicator of interspecific variations (Werner and Máguas 2010). Corema album had lower Δ¹³C than H. italicum throughout the seasons and a higher Φ_PSII compared to H. italicum in spring. These results showed a general lower photosynthetic capacity of H. italicum and a better stomatal control of C. album, which could be related with leaf morphological characteristics: C. album with leaf rolling forming a closed chamber, contributing to reutilization of CO₂ and a higher photosynthetic capacity, while H. italicum with a higher leaf pubescence leading to a reduction in light absorbance and to a reduction on photosynthetic capacity.

The natives’ common conservative water use strategy was supported by morpho-functional traits indicative of low-resource environments and where plant productivity is severely limited (e.g. Carboni et al. 2010; Cramer et al. 2014; Pardini et al. 2015). In fact, the two native species presented leaf adaptations for protection against water deficit and high solar radiation in the summer, such as the dense covering indumentum of H. italicum leaves and the thick and cutinized upper epidermal cell walls in C. album. Also, in both species the leaves are rolled dorsiventrally to create deep stomatal grooves that are covered by glandular and non-glandular trichomes. These features could be interpreted as additional adaptations to control the plant water content, regulating the leaf temperature and the boundary layer of the leaf surface (De Micco and Aronne 2012; Rotondi et al. 2003). In general, leaves with a high density of trichomes are water repellent, but in some species, where trichomes’ walls are smooth and non-waxy, they can entrap water droplets (Koch 2008). In H. italicum, the walls of non-glandular trichomes showed no cutin or waxes and are essentially constituted by pectins, highly hydrophilic compounds, well known by their ability to form gels with water. The high frequency of dew condensation on leaves, in Mediterranean habitat (Pitacco et al. 1992), together with the wall characteristics of these non-glandular trichomes, leads us to consider that the lanate, felt-like indumentum of H. italicum, formed by air filled, non-glandular, trichomes, not only reflects the visible light making the surfaces appear white but also may retain and probably adsorb dew water. According to other authors, leaf hairiness allows to retain the fog humidity and keep leaf surface wet, and thus a significant amount of water may enter through the cuticle which appears relatively permeable when humid (e.g. Grammatikopoulos and Manetas 1994; Grammatikopoulos et al. 1995; Kyparissis and Manetas 1993; Limm et al. 2009). Supporting this hypothesis, the summer xylem water δ¹⁸O observed probably reflects dew as an important water source for this species, but further work is needed to confirm the main water source of this species. On the contrary, the invasive A. longifolia presented no leaf pubescence. However, the exotic species presented phyllodes and not true leaves, which, although by different mechanisms, can be favourable under drought conditions, since they are better adapted to water stress (e.g. Hansen and Steig 1993; Walters and Bartholomew 1984). Their sclerophyllous nature is a specific adaptation to drought and the isobilateral structure of the mesophyll with a significant large number of palisade layers may account for a better use of light. In addition, the amphistomatous phyllodes, with high stomatal density, may be an effective feature for CO₂ capture and could contributed, together with the rich palisade parenchyma, to a high photosynthetic capacity. The presence of an achlorophyllous central parenchyma with several layers of large cells surrounded by the palisade parenchyma may be of particular advantage as it is thought to be a water storage tissue (Boughton 1990). Accordingly to Pasquet-Kok et al. (2010), in Acacia koa the phyllodes were related with a higher water storage capacity, rather than an ability to develop and maintain a high physiological performance, thus enabling a longer survival after stomatal closure during the drought period (Sack et al. 2003). Indeed, the water storage capacity evidenced by Australian acacias is an important trait to drought tolerance of this genus, and may well be one of the most successful leaf traits of this invasive species. Moreover, Sommerville et al. (2012) showed that the large morphological diversity of phyllodes across the genus Acacia might explain a correspondent diversity in hydraulic response to rapid exploitation of water resources in environments with unpredictable rainfall. Such could be the case of A. longifolia, since previous reports in a sand dune pine forest, showed that after single rain events, this invasive species presented a higher ability of water use when compared with Pinus pinaster (Rascher et al. 2011b). In winter and spring, no differences were detected among the three species regarding the efficiency of energy conversion determined in the absence of any excess of light. This is in accordance with previous studies involving a large number of ecologically diverse species not subjected to severe stresses (Björkman and Demmig 1987). However, during the day, the responses to an increase of PAR and a decrease of Ψ were remarkably different among species. In H. italicum, the Φ_PSII dropped precipitously with a relatively small decrease of Ψ, from −0.5 MPa to −1.0MPa. In A. longifolia, this Φ_PSII decline was more gradual and the lowest values were obtained at very low values of Ψ (around −3.5 MPa). This implies that A. longifolia’s water potential decreased substantially in response to a possible high water loss and that the plant presented a physiological adjustment complementary to the higher water uptake demand. In case of an external water stress, it is crucial for A. longifolia that water can be extracted from soil rapidly enough to compensate for water loss. The use of water from deeper soil layers, where water is more available during the day due to less evaporative demand, could be a beneficial adjustment in this process. The δ¹⁸O values of A. longifolia xylem water pointed to a different water source compared with the native species, being this species relying mainly in deeper soil water during summer. The xylem isotopic analyses showed a similar isotopic ratio in H. italicum and C. album, which indicates that both species were using the same ¹⁸O enriched water source (most probably water from topsoil layers or dew) during summer. Contrastingly, A. longifolia was able to explore deeper soil water resources that may be mainly constituted by the more abundant winter precipitation. The groundwater isotopic composition, typically less oscillating, is scarcely affected by local and punctual precipitation but rather reflects the general average precipitation and the overall annual fresh water inputs, generally more ¹⁸O depleted. In summer, this water source presented a low contribution to the xylem water of all species, since the water table is more distant during the dry season and not easily reachable by roots. The difference in water sources used by native and exotic species indicates that the studied species have different root systems, H. italicum being the one with shallower roots. More superficial root systems, as present in the native species, can be an advantage to capture air humidity and dew formed on the soil during the night or dawn. In these environments, where rainfall is scarce especially in the dry season, soil water input from night condensation, fog and dew formations (Kidron 2000; Lo Gullo et al. 1986) or soil water vapour adsorption (Agam and Berliner 2006; Kosmas et al. 1998) could be an important water source for plants.

The exotic invasive species was able to respond to variations in water availability during the growth season by changing its water use efficiency, while displaying higher photosynthetic capacity during the growth period when compared to the native H. italicum, together with increased variation between spring and summer Δ¹³C. Such performance is indicative of a higher plasticity and favours quick responses to environmental conditions (as seen also in Máguas et al. 2011; Peperkorn et al. 2005). In fact, the contrasting physiological pattern of the native and invasive exotic species highlights their different adaptations, the native species being, under changing water availability, generally more stable and the exotic species more dynamic. The results obtained indicate that the invasive A. longifolia occurrence can relate to a high physiological plasticity and to anatomical traits of phyllodes, which allow a higher performance and potentially a higher productivity under favourable periods of available water. Accordingly, some authors refer that non-native invasive plants are often assumed to generally have higher rates of photosynthesis, higher rates of leaf-level water use (Cavaleri and Sack 2010; Leishman et al. 2007) and faster growth rates (Peperkorn et al. 2005; Pysek and Richardson 2007; van Kleunen et al. 2010). However, a recent study from Funk et al. (2016) brought new insights to this discussion, showing a highly diverse range of drought-avoiding traits that favour invasive species’ success in low-resource habitats. According to this study and others (Funk and Vitousek 2007; Matzek 2012; Oliveira et al. 2014), invasive species may also succeed in low-resource habitats by both resource acquisition and conservation, rather than efficiency. One important aspect that was not included in this study and very seldom is elsewhere, is the comparison between natives and invasives during the full vegetative growth period, which could reveal a longer growth period and higher resources efficiency by the invasive species. Such characteristics would then be attributable to a higher competitiveness, as successful competitors will often use limiting resources more efficiently for longer periods or more quickly than co-occurring natives (Rejmánek et al. 2013; Gioria et al. 2018).

In conclusion, the current study shows different leaf traits and ecophysiological responses to drought of native and exotic invasive species. Although the species resist to drought and showed some degree of functional similarity, native and invasive species presented different water strategies: natives relied on morphological adaptations to restrict water loss, whereas the invasive species adjusted the water uptake as a way to balance their limited ability of restricting water loss. These contrasting water strategies were accompanied by the use of different water sources during drought periods, indicating different root systems, H. italicum most likely being the one with shallower roots. Additionally, the seasonal physiological plasticity and the high performance during the growth period of the invasive species A. longifolia can be seen as a major competitive advantage over native species, although further work is needed to fully understand the species successful invasion.

FUNDING

FCT, Portuguese Foundation for Science and Technology (POCTI/BSE/34689/1999); European Union Framework Programme (FP7-PEOPLE-2010-IRSES–269206, INSPECTED.NET project) PhD grant from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES; to C.A.).

ACKNOWLEDGEMENTS

We thank Christiane Werner for stimulating discussions. We acknowledge Rodrigo Maia for valuable comments that greatly improved the manuscript.

Conflict of interest statement. None declared.