Like biomass production and morphology, gas exchange-related traits in P. australis are highly plastic. Within a phylogeographic region, the prevailing climatic conditions have the strongest effects on gas exchange rates (Lessmann et al., 2001; Hansen et al., 2007; Mozdzer et al., 2016a). Although the climate of the area of origin strongly affects physiological responses, there are also phylogeographic differences in potential responses to environmental change. For example, the NAint Delta lineage was less plastic in its ability to modify gas exchange parameters compared to the highly plastic NAint M lineage when grown across 14° of latitude (Mozdzer et al., 2016a). Furthermore, tropical and subtropical populations of P. australis have a higher photosynthetic capacity and photosynthetic pigment concentration than populations in the temperate zone (Nguyen et al., 2013). Similarly, NAint P. australis has a higher photosynthetic capacity and pigment concentrations than NAnat (Mozdzer and Zieman, 2010; Guo et al., 2014). Nguyen et al. (2013) proposed the existence of a diversified C₃ pathway within P. australis that is modified to maintain high enzymatic efficiencies in tropical and Mediterranean climates but can be down-regulated to accommodate the lower temperature and irradiance of temperate regions.

Despite the typical C₃-photosynthetic features displayed by P. australis, C₄-like strategies have also been observed. A prominent sheath layer that is especially pronounced in young P. australis leaves surrounds the vascular bundles in the mesophyll, resembling the foliar Kranz anatomy of C₄ plants (Henriques and Webb, 1989). However, due to the lack of chloroplasts in this layer, there is no functional correlation with true C₄ plants (Henriques and Webb, 1989). Doubts about the photosynthetic pathway of P. australis have also emerged due to relatively high PEPcase activities, higher activities of the decarboxylating NADP-dependent malic enzyme (NADP-ME), and a possible C₃–C₄ intermediate pathway associated with ecotypes from arid or salt-affected habitats (Rintamaki and Aro, 1985; Zheng et al., 2000; Zhu et al., 2012). Most of the known C₄ species occur in the Poaceae, in which C₄-evolution has occurred independently several times and, thus, genes are present in P. australis that can rapidly develop C₄ functions including the gene coding for NADP-ME (Christin et al., 2009).

Nevertheless, P. australis has, in most studies, been shown to possess characteristics typical of C₃ plants, including a high Rubisco/PEPcarboxylase ratio, high photorespiration rates, and a high CO₂ compensation point (Antonielli et al., 2002; Hansen et al., 2007; Eller and Brix, 2012). The photosynthetic pathway of P. australis therefore remains unresolved, as the range of the abovementioned studies suggests that the photosynthetic pathway may vary within the species. The distinct bundle sheath cells in P. australis leaves also raise the possibility of C₂ photosynthesis, which is the evolutionary bridge between C₃ and C₄ photosynthesis (Sage, 2016); however, evidence of this possibility has yet to be found.

By far the greatest number of scientific studies on P. australis have been concerned with the species’ tremendous potential for nutrient removal, which makes it an ideal candidate species for wastewater treatment in constructed wetlands (e.g., Brix and Schierup, 1989; Brix, 1997; Bragato et al., 2006; Vymazal, 2013; Hernández-Crespo et al., 2016). Genetically determined differences in nutrient uptake and assimilation capacity result in distinct reed ecotypes with differences in productivity (Tho et al., 2016). Some ecotypes sustain high nutrient assimilation rates and high allocation to aboveground biomass, while others have high nutrient translocation rates to rhizomes for storage and thus high belowground biomass allocation (Kühl et al., 1997; Tripathee and Schäfer, 2014). Reed genotypes with dissimilar nutrient demands and productivity can thus grow at similar nutrient levels in naturally adjacent stands. Such distinct ecophysiological strategies confer greater population plasticity and performance to a genetically diverse stand compared to a monoclonal stand (Rolletschek et al., 1999). Pronounced differences in the nitrate uptake kinetics of distinct reed genotypes are possibly caused by distinct transcript abundances of nitrate transporter genes, and a likely reason for the genotypic differences in nutrient acquisition strategies (Araki et al., 2005). In general, P. australis is well-adapted for growth in nutrient-rich habitats (Mozdzer et al., 2010; Caplan et al., 2015) but can also acclimate to low nutrient availability by increasing the affinity for ammonium uptake (Romero et al., 1999; Tylova-Munzarova et al., 2005; Mozdzer and Megonigal, 2012).

Like almost all plants that can grow vigorously in habitats where soil saturation and flooding are common (Vartapetian and Jackson, 1997), P. australis aerates flooded tissues by transporting oxygen through a well-developed network of internal airspaces, or aerenchyma (Armstrong and Armstrong, 1991; Jackson and Armstrong, 1999). These internal airspaces are continuous from the leaf sheaths and culms, through the rhizomes, and into the root cortex, where aerenchyma are particularly well-developed through lysigeny (Armstrong et al., 1996a; White and Ganf, 2002). Rhizomes are segmented internally and have secondary aeration channels in the internode cortex, such that airflow is maintained even if rhizome cavities become damaged and filled with water (Soukup et al., 2000). More efficient root aeration also allows for greater respiration rates and, thus, sustained nutrient uptake capacity and root development, even in hypoxic soils (Nakamura et al., 2013).

Phragmites australis is also one of the few wetland species that does not rely solely on simple diffusion for gas transport; it supplements its aeration with convective gas flow (Brix, 1989; Brix et al., 1992, 1996; Armstrong et al., 1996a). Convection is induced by humidity gradients generated in lacunae (i.e., sub-stomatal cavities) in leaf sheaths of live culms (Armstrong et al., 1996a,c). The pressure that builds up in lacunae pushes air down through live culms and rhizomes; air is vented out of the plant through damaged or dead culms (Brix, 1989; Armstrong et al., 1996c; Afreen et al., 2007).

Little attention has been paid to potential intraspecific differences in gas transport among P. australis lineages. Tulbure et al. (2012) showed that the ventilation efficiency of the invasive NAint M lineage in North America was 300 times higher than that of native P. australis subsp. americanus, when differences in stem densities between lineages were accounted for. Since gas flux is a physically determined process and is strongly affected by internal anatomy (Rolletschek et al., 1999), different gas flow behavior can be expected in plants with genotype-specific morphological characteristics. Moreover, gas flow characteristics of wetland plants affect not only oxygen transport but also plant-mediated methane emission (Brix et al., 2001), and lineage-specific differences in factors controlling gas flow are known to affect methane fluxes (Armstrong et al., 1996b; Kim et al., 1998). For example, NAint M roots more deeply than other lineages and, through changes in soil organic matter dynamics, can lead to increased rates of CO₂ losses to the atmosphere (Bernal et al., 2017). Differences in gas flow capacity and rhizosphere oxygenation among lineages are therefore very likely and deserve greater attention.

Without considerable greenhouse gas reductions, the global rise in mean surface temperature of Earth is very likely to exceed 1.5–4°C by the end of the 21st century, with the greatest increases in the Northern Hemisphere (IPCC, 2014). Heatwaves and extreme precipitation events are expected to occur more frequently and with longer durations in many regions, but occasional cold temperature extremes can also be expected (IPCC, 2014).

Phragmites australis exhibits lineage-specific responses to temperature regimes in terms of morphology, growth, and to a certain extent, photosynthetic traits (Table 1; Clevering et al., 2001; Lessmann et al., 2001; Eller and Brix, 2012; Eller et al., 2013; Mozdzer et al., 2016a). Rates of P. australis growth (especially shoot height and length), as well as rates of transpiration and photosynthesis, are generally greater at lower latitudes due to warmer temperature regimes and longer day lengths (Haslam, 1975; Lissner et al., 1999a,b; Zemlin et al., 2000; Lessmann et al., 2001; Karunaratne et al., 2003; Mozdzer et al., 2016a). Reciprocal transplant experiments in common gardens have shown that, for lineages originating from lower latitudes, higher temperatures are needed to initiate growth and, after being transplanted to higher latitudes, panicles either emerge late or do not flower at all (Brix, 1999b; Clevering et al., 2001; Karunaratne et al., 2003; Lambertini et al., 2012c). Adaptation to the climate in the region of origin significantly affects plant species’ performance and plasticity (Franks et al., 2014; Molina-Montenegro et al., 2016; Allen et al., 2017; Bhattarai et al., 2017a,b; Münzbergová et al., 2017). Hence, P. australis belonging to the MED lineage can be a model for Mediterranean, subtropical, and even tropical plant species, while populations of the EU lineage can be a model for temperate species found at higher latitudes (Figure 3).

Some lineages seem to be more plastic to changes in temperature than others, as they show a large acclimation capacity to both increases and decreases in temperature (Figure 3; Lessmann et al., 2001; Eller and Brix, 2012). This is the case for NAint M in North America, for example, where temperature fluctuations have been shown to enhance its distribution (Guo et al., 2013). The North American invasion is therefore likely to accelerate with climate change. It has previously been suggested that lineages originating in areas with high fluctuating temperatures also have higher plasticity to temperature changes, and may therefore be better adapted to withstand climatic changes (Molina-Montenegro and Naya, 2012). The same has been shown for P. australis lineages; EU genotypes from higher latitudes in temperate areas have generally shown higher plasticity toward differences in growth temperature (Lessmann et al., 2001; Eller and Brix, 2012; Nguyen et al., 2013). It can be assumed that the high plasticity of NAint M derives from its origin in the highly plastic EU populations, emphasizing the potential model role of P. australis lineages from high-latitudes for temperate plant responses to temperature differences (Figure 3).

Other lineages appear to be pre-adapted to predicted future temperature regimes and are therefore likely to extend their range northward (Clevering et al., 2001; Lessmann et al., 2001; Eller et al., 2014a,b; Mozdzer et al., 2016a). It is possible that lineages originating from lower latitudes may expand their distributions northward in the warming world (Figures 2, 3; Guo et al., 2013; Mozdzer et al., 2016a), as frost and cool temperatures limit growth or sexual reproduction at mid and high-latitudes (Mozdzer et al., 2016a). Also, the expansion of the invasive NAint Delta lineage can be attributed, in part, to warmer temperatures in its invasive range than in its native range (Guo et al., 2013), as advancement of a population can be expected if a high phenotypic plasticity to temperatures is inherent (Figure 2). Alternatively, lower-latitudinal lineages may be unable to cope with the rapidity of temperature changes due to a narrow niche-breadth or acclimation-capacity, and may become genetically diminished (Figure 2). Using P. australis as model for global warming, a two-way scenario can be expected as the species responds to temperature increases. On the one hand, species with high phenotypic plasticity, and therefore greater niche breadths, will likely be able to cope with warming and thrive equally well or even extend their range northward. Another species in which this is likey to occur is Nothofagus pumilio (Mathiasen and Premoli, 2016). On the other hand, species with limited plasticity and narrower niche breadths may fail to acclimate, facing local extinction in the worst case (Figure 2). Some herbaceous alpine species occurring at high elevation provide a good example of narrow niche breadth leading to local extinction (Schmid et al., 2017). Indirect changes and interactions of temperature with other abiotic factors, such as increased drought and salinity, may impose additional challenges on P. australis populations in already warm areas, possibly resulting in more favorable growth conditions at higher latitudes (Figure 3; Brix, 1999b; Eller et al., 2014a).

Atmospheric CO₂-equivalents are likely to exceed 720 ppm, and possibly reach 1000 ppm, by the late 21st century if greenhouse gas emissions are not restricted substantially (IPCC, 2014). As a C₃ plant, P. australis will benefit from rising atmospheric CO₂ concentrations, but there is growing evidence that the magnitude of its response may be lineage-specific due to differences in phenotypic and physiological plasticity (Figure 3). For example, in an experiment in which CO₂ was elevated to ∼700 ppm, both the NAint M and NAnat lineage responded positively to CO₂ elevation but the NAint M lineage had considerably greater plasticity in nearly every trait measured (Mozdzer and Megonigal, 2012; Caplan et al., 2014). In contrast, several studies focusing on other lineages of P. australis have found no significant effects of elevated CO₂ on biomass or morphological parameters, though some photosynthetic enhancements have been reported (Scholefield et al., 2004; Milla et al., 2006; Kim and Kang, 2008; Eller et al., 2013). We note that these studies either did not measure below-ground biomass productivity or did not account for respiration rates, which may partly explain the lack of biomass stimulation by elevated CO₂.

Based on the above, differential responses to elevated CO₂ may result in lineage-specific shifts, increases in competitiveness and distribution changes. However, interactions with other abiotic factors such as salinity and nutrients make it more difficult to predict the effects of elevated CO₂ in natural environments. For example, whilst shoot elongation rates are enhanced by elevated CO₂ and temperature to similar degrees in both the invasive NAint Delta and the invasive NAint M lineages, the NAint Delta lineage outperforms the NAint M lineage when grown at 20‰ soil salinity (Eller et al., 2014a). The stronger growth response of NAint Delta is facilitated, in large part, by intrinsically greater photosynthetic rates (Table 1; Nguyen et al., 2013; Eller et al., 2014a). Overall, the strongest effects on growth and carbon assimilation rates are expected to result from changes in CO₂ that are accompanied by increases in temperature or nutrient enrichment, especially nitrogen (N) (Figures 2, 3; Mozdzer and Megonigal, 2012; Eller et al., 2013, 2014a,b; Caplan et al., 2015), which has been shown for other C₃ species (Ainsworth and Rogers, 2007).

Changes induced by elevated atmospheric CO₂ concentrations may influence ecosystem services in P. australis dominated wetlands. For example, elevated CO₂ increases the methane emission rate of both NAnat and NAint lineages (Mozdzer and Megonigal, 2013), which may offset the net carbon fixation of P. australis wetlands that would otherwise be greenhouse gas sinks (Brix et al., 2001). Moreover, elevated CO₂ induces greater belowground productivity and rooting depths in the NAint M lineage (Mozdzer et al., 2016b), which are likely to increase rates of belowground biomass accumulation and surface elevation gain. Such effects could enhance the ability of P. australis dominated wetlands to keep pace with sea level rise (Rooth et al., 2003; Caplan et al., 2015; Mozdzer et al., 2016b). Responses to elevated CO₂ have predominantly been investigated through short-term studies and have only investigated a few P. australis lineages (including NAnat, NAint M, NAint Delta, EU, and Med; Eller et al., 2013; Mozdzer and Megonigal, 2013; Eller et al., 2014a,b); more research is needed to determine if enhancement of growth and methane emission rates apply to the whole species. Due to its high plasticity to atmospheric CO₂ concentration, the NAint M lineage can be used as a model for studying the responses of invasive C₃ species to elevated CO₂ (Figure 3), as high phenotypic plasticity is a common trait in invasive species (Drenovsky et al., 2012; Gioria and Osborne, 2014; Colautti et al., 2017).

Saltwater intrusion due to global sea level rise is becoming a major issue in both brackish saltmarshes and tidal freshwater wetlands (Beckett et al., 2016). Moreover, regions with high salinity and high evaporation rates are likely to become more saline, while regions of low salinity and high precipitation will become fresher, inducing greater extremes in salinity in wetlands globally (IPCC, 2014). Finally, some climate change models predict an increase in the intensity and frequency of tropical storms and hurricanes (e.g., Bender et al., 2010; Knutson et al., 2010), which may lead to flooding and salt intrusion in near-coastal habitats. Hence, soil salinity regimes are shifting such that salinity tolerance is of increasing importance to biotic communities in coastal ecosystems.

The ecological amplitude of P. australis extends from freshwater to saline tidal wetlands, with plants persisting at salinities as high as 65‰ (recorded in Delaware, eastern United States; Engloner, 2009), with tolerances of 22.5‰ (Lissner and Schierup, 1997) to 35‰ salinity reported for juveniles (Engloner, 2009) and a limit of 30‰ reported for seed germination (Yu et al., 2012). The mechanisms of salt tolerance in P. australis include Na⁺ exclusion or vacuolar compartmentalization, tissue dehydration or compatible osmotic solute accumulation, and increased gene expression of oxidative stress response enzymes (Matoh et al., 1988; Lissner and Schierup, 1997; Lissner et al., 1999a; Pagter et al., 2005; Vasquez et al., 2005; Achenbach and Brix, 2013; Achenbach et al., 2013; Eller et al., 2014b). Certain ecotypes of P. australis are more salt tolerant than others (Table 1), with higher salinities yielding greater germination rates and better developed root systems (Rechav, 1967; Van der Toorn, 1972). Several controlled experimental studies have shown that NAint Delta has higher salt tolerance than NAint M, though both perform better at higher salinities than Med or the NAnat lineages (Achenbach and Brix, 2013; Eller et al., 2014a). Since the NAnat lineage has considerably greater N uptake rates than the invasive NAint M lineage at salt concentrations up to 20‰ (Mozdzer et al., 2010), it is primarily limited to oligohaline and mesohaline wetlands in the mid-Atlantic United States (Vasquez et al., 2005; Packett and Chambers, 2006; Mozdzer et al., 2010). However, in other regions of the United States, like New England and the Midwest, native North American populations are not limited by salinity and occur in brackish river systems as well as in saltmarshes (e.g., Burdick et al., 2001; Kettenring and Whigham, 2009; Kettenring and Mock, 2012; Cronin et al., 2015). These results indicate that salt tolerance is genotype-specific rather than lineage-specific, and highly variable within the species. Plant responses to changes in soil salinity can therefore be elucidated by studying locally adapted genotypes rather than lineages; the genetic composition of a population exposed to changing salinity regimes will be altered according to its pre-adaptation for salt tolerance (Figure 2).

Locally adapted genotypes may, however, not be the only strategy P. australis employs to persist in saline environments. Salt avoidance by below-ground labor division may also play a significant role in acclimation to shifts in salinity regime, as has also been found in the clonal species Schoenoplectus americanus in a brackish tidal wetland (Ikegami et al., 2008). An important component of P. australis’ ability to grow in soils spanning a wide range of salinities is its extensive rhizome and root system. Thus, most of the belowground biomass of lineages occurring in North America has been found in the upper 70 cm of soil (Moore et al., 2012), though depths can exceed 3 m even at coastal sites (Mozdzer et al., 2016b). This morphology may grant the species access to freshwater resources at soil depths less affected by tides. The importance of belowground organs to salt tolerance has also been demonstrated in Asia in landscapes with patchy soil salinity, where the genetic variation of P. australis is closely correlated with habitat heterogeneity (Gao et al., 2012).

Despite being able to survive and grow in saline soil conditions, P. australis has historically been considered a fresh to brackish water species (Raunkiaer, 1893; Haslam, 1973; Matoh et al., 1988). Several studies have identified negative effects of greater salinity levels on various traits including biomass production, culm height, stand density, culm diameter, and rhizome carbohydrate content (Lissner et al., 1999a; Engloner, 2009; Achenbach et al., 2013; Tang et al., 2013; Eller et al., 2014a). Physiologically, P. australis responses to high salinity are associated with decreases in tissue water potential, stomatal conductance and transpiration rates, photosynthetic efficiency of PSII, and nitrogen uptake rates (Chambers et al., 1998; Lissner et al., 1999b; Naumann et al., 2007; Pagter et al., 2009; Mozdzer et al., 2010; Zhang and Deng, 2012). The photosynthetic recovery and re-opening of stomata after short-term exposure to high salinity differs between genotypes of different lineages, demonstrating that sensing and responding to osmotic stress is a genotype-specific feature (Achenbach and Brix, 2014). Also, high-affinity K⁺ transporters isolated from salt tolerant reed plants are more efficient in K⁺ uptake and less permeable to Na⁺ than transporters from salt-sensitive plants, offering an explanation for their difference in salt-sensitivity (Takahashi et al., 2007).

Salinity increases in freshwater wetlands are likely to affect the natural distribution of P. australis genotypes and to alter the competitive dynamics between less and more salt-resistant plants. Spread of P. australis into salt marshes might also be accelerated in El Niño years due to temporary decreases in salinity from heavy rains that open windows for seedling establishment (Minchinton, 2002) and also due to the expansion of patches that maintain access to less saline groundwater in other parts of the stand. Storm surge from tropical storms, cyclones, and hurricanes can flood near-coastal freshwater wetlands and greatly elevate salinity levels. In North America, the spread of invasive NAint M populations is strongly positively correlated with the frequency of these storms and it has been argued that NAint M lineage thrives because it is more salt tolerant than native wetland plants (Burdick and Konisky, 2003; Bhattarai and Cronin, 2014). Unlike climatic adaptations that can be attributed, in part, to the phylogeographic origins of P. australis lineages, salt tolerance cannot simply be ascribed to a specific phylogenetic background, but is rather a consequence of the single pre-adapted genotype (Gao et al., 2012; Achenbach et al., 2013). Locally adapted genotypes of different reed lineages may therefore serve as models for studying responses to changes in soil salinity. Depending on the lineage, however, the outcome of interactions with other global change drivers can be estimated. Overall, future changes in P. australis salt tolerance are very likely, as responses to salinity have been shown to interact with temperature and CO₂, and may confer greater salt resistance on P. australis due to improved osmotic acclimation and higher assimilation rates (Lissner et al., 1999a; Eller et al., 2014a). A lineage with inherently high phenotypic plasticity, such as NAint M, can be expected to benefit more from interactive effects of salinity and elevated CO₂ than NAnat.

More research is needed to determine which genetic factors underlie the high salt tolerance of genotypes within P. australis lineages, and how and why these factors arise in these genotypes. Due to their increased likelihood of including salt-resistant genotypes, populations and stands with high genetic variability will probably have the strongest prospects of adapting to changes in salinity. Moreover, individual stands of P. australis will likely face genetic impoverishment following the extinction of salt-sensitive genotypes under shifting soil salinity regimes (Figure 2).

Climate projections indicate that greater variability in precipitation will cause more frequent extremes in precipitation and discharge in many areas. This will increase the frequency and magnitude of inland and coastal floods, which will be compounded by larger storm surges and rising sea levels (IPCC, 2014).

Although P. australis seedlings are extremely vulnerable to flooding (Chambers et al., 2003; Mauchamp and Methy, 2004; Baldwin et al., 2010; Kettenring et al., 2015), once established, seedlings and adult plants are highly tolerant of inundation. Specifically, the survival, physiology, and growth of P. australis are less affected by submersion than are many other wetland plants (Gries et al., 1990; Brix et al., 1992; Armstrong et al., 1996b). Moreover, susceptibility to flooding decreases with ontogeny in the species (Bart and Hartman, 2003; Chambers et al., 2003; Whyte et al., 2008; Tulbure and Johnston, 2010). P. australis is also tolerant of greater amplitude fluctuations (±45 cm) in water level than other species, provided that its elevation is close to the mean water level (White et al., 2007). However, high water during extreme flooding years caused reed belts to decline along lakes in southern Germany and Austria; stands rejuvenated only in low-water years (Ostendorp, 1999; Ostendorp et al., 2003).

We found no studies that directly assessed genotypic differences in flooding tolerance. However, the number of genotypes represented in a lakeshore stand in Hungary decreased with water depth (Engloner and Major, 2011), which the authors attributed to genotype-specific flooding tolerance. Another study found that seasonal profiles of amino acids and carbohydrates differed by genotype and flooding regime in a German fen (Koppitz, 2004; Koppitz et al., 2004), though genotype by flooding regime interactions were not reported. These findings indicate that responses to flooding are a consequence of pre-adapted genotypes rather than adaptation at the lineage scale. Like soil salinity, plant responses to fluctuating water levels can best be studied in locally adapted genotypes, and the population response to be expected will be an alteration of its genetic composition (Figure 2).

Responses to flooding depend on the interaction of other drivers of global change in a way that is similar to soil salinity. For example, belowground, deeper water induces a shallower rhizome depth distribution (Weisner and Strand, 1996; Vretare et al., 2001; White and Ganf, 2002; Mozdzer et al., 2016b), but this rooting depth will deepen with rising CO₂ concentrations (Mozdzer et al., 2016b). Vretare et al. (2001) suggested that P. australis increases allocation to stem height when growing in deeper water and simultaneously decreases stem density and belowground allocation. While stem density appears to be consistently lower in deeper water (Yamasaki and Tange, 1981; Vretare et al., 2001; Bodensteiner and Gabriel, 2003), stem height has been reported to both increase and decrease in response to deeper flooding (e.g., Hellings and Gallagher, 1992; Coops et al., 1996; Vretare et al., 2001; Engloner, 2004). Genotypic differences in the P. australis plants studied may contribute to the variation in outcomes from these studies, though this has not been assessed in the majority of cases. Facing more frequent flooding regimes with global change (IPCC, 2014), natural selection of flooding-resistant genotypes can be anticipated such that flooded populations may become genetically impoverished. Previous studies have often focused on the growth and morphological acclimation of P. australis to flooding, but there is a need to investigate the physiological consequences of inundation more thoroughly. Recent advances in research on photosynthesis in submerged shoots showed that elevated CO₂ can alleviate flooding stress (Winkel et al., 2014). Although not investigated in P. australis, this capability would be especially relevant to determining seedling responses to inundation.

Phragmites australis is well-adapted for life in flooded environments but is tolerant of the full range of wetland hydrological conditions, including drought (Pagter et al., 2005). Wetland hydrology can be highly variable, with relatively dry conditions being common or even extreme in times of drought (Mitsch and Gosselink, 2007). With climate change, drought is predicted to develop more quickly and increase in intensity in many regions of the world (IPCC, 2014; Trenberth et al., 2014). Phragmites australis deals with drought through both short-term tolerance mechanisms (i.e., by making physiological or biochemical adjustments) and longer-term avoidance strategies that affect morphological and developmental traits (Morgan, 1984; Chaves et al., 2002; Pagter et al., 2005; Touchette et al., 2007). Following extreme low-water conditions, reed stands employ a “guerilla strategy” to efficiently and quickly occupy new wet habitats; they produce tillers across the uninhabited littoral zone as well as “legehalme,” which are rapidly elongating, horizontal shoots from which new culms emerge at the nodes (Ostendorp and Dienst, 2012).

Under dry soil conditions (in situ), P. australis substantially decreases leaf osmotic potential and accumulates more soluble sugars, amino acids, protein metabolites, proline, and nutrient elements than under moist conditions (Elhaak et al., 1993). When subjected to mild water stress, P. australis reduces total leaf area and biomass, but severe water stress induces changes in osmolality, leaf proline concentration, leaf chlorophyll a content, stomatal conductance, and photosynthetic rates (Pagter et al., 2005). Similarly, terrestrial dryland ecotypes of P. australis from northwest China increase their capacity for osmotic adjustment, significantly decrease stomatal conductance, reduce net photosynthetic rate, and their cover and height declines (Cui et al., 2010). Compared to wetland ecotypes, they also exhibit greater water use efficiency, increased activity of C₄ photosynthetic enzymes, protective down-regulation of photosynthetic enzyme activities, and greater antioxidant enzyme activity (contributing to oxidative stress protection; Wang et al., 1998; Zhu et al., 2001, 2003a,b; Gong et al., 2011; Xiang et al., 2012). With the onset of complete drought (in controlled experimental studies), P. australis showed signs of drought in leaf xylem pressure potentials by the second day, stomatal conductance and photosynthesis by days four to eight, and leaf rolling and wilting by day five (Saltmarsh et al., 2006; Naumann et al., 2007). Field-based phenological studies of P. australis in Great Britain showed that years with spring drought can induce later emergence and flowering, as well as shorter culms, compared to years with normally flooding patterns. Also, years with fall drought may lead to earlier senescence compared to years when the stand is flooded. Nonetheless, P. australis rhizomes can penetrate up to 2 m into the soil to access deeper groundwater (Haslam, 1970).

As with salinity tolerance, it is very likely that some P. australis genotypes are more drought tolerant than others, even within lineages (Figure 3). However, drought events are more tightly coupled to climate than are high salinity periods, with the implication that drought-resistance could be phylogeographically determined in the species. It remains to be determined if phylogenetically determined drought tolerant lineages are also salt tolerant. For example, it seems that the North American native lineage is more sensitive to drought in some regions, such as the southwestern United States, where it is often associated with small streams and springs, which are sensitive to small changes in water availability (Meyerson et al., 2010a; Kettenring and Mock, 2012; Kettenring et al., 2012). In contrast, short-term drought that leads to temporary drawdowns may benefit colonization of the NAint M lineage by fostering seedling recruitment (Alvarez et al., 2005; Tulbure et al., 2007; Whyte et al., 2008; Kettenring et al., 2015, 2016). These studies serve as examples for using P. australis as model to study, whether physiologically similar responses to different global change factors result from similar adaptations or are independent of the plants’ phylogeographic origin.

Increases in nutrients from atmospheric deposition, agriculture, and development are a well-known component of global change (Galloway et al., 2004). The ability of P. australis to efficiently take up nutrients, especially nitrogen (N; i.e., NO₃^-, NH₄⁺, and dissolved organic nitrogen), suggests that increased eutrophication from human activities will have a positive impact on the spread of the species, particularly its invasive lineages. Wetland eutrophication is expected to increase in areas such as in agricultural and densely populated urban and suburban areas where nutrient loads continue to increase. However, the distribution of eutrophication under global change is likely to be spatially heterogeneous across regions. In some places, increased water resources due to glacier melting or increased precipitation may dilute N concentrations, whereas, in other places, evaporation and decreased precipitation could exacerbate the effects of pollutants and nutrients (IPCC, 2007, 2014).

Although P. australis grown under controlled experimental conditions generally responds positively to nutrient addition, e.g., displaying increased biomass production and a greater shoot density (Szczepanska and Szczepanski, 1976; Romero et al., 1999; Tho et al., 2016), eutrophication was a key factor responsible for reed die-back in Europe during the 1990s (van der Putten, 1997). However, the detrimental effects were predominantly indirect and caused by anoxic sediments, phytotoxin production from algal blooms or increased litter production, callus development and blockage of gas transport pathways in rhizomes and roots, and exacerbated by other human-induced impairments of natural reed habitats (Armstrong et al., 1996c; Brix, 1999b). In general, the high aeration capacity of the species allows for high root respiration rates throughout its large root systems, which, in turn, can facilitate high nutrient uptake rates (Nakamura et al., 2013). Phragmites australis lineages with inherently high biomass productivity and high belowground:aboveground ratios are therefore well-adapted for growth under increasingly eutrophic and anaerobic conditions, and appropriate models for investigating nutrient availability responses of highly productive and ruderal species (Figure 3). Increased nutrient availability is also likely to increase P. australis inflorescence and floret production (Kettenring et al., 2011) as well as seedling success given that seedlings will grow more rapidly beyond a vulnerable size (Saltonstall and Stevenson, 2007; Kettenring et al., 2015). Nutrient addition can also alter phenology, inducing culms to grow more rapidly early and late in the year, increasing their heights and annual carbon gains (Caplan et al., 2015). Relative to other wetland species, N affinity is very high for P. australis, but is usually its limiting nutrient (Chambers et al., 1998; Clevering, 1998; Romero et al., 1999; Saltonstall and Stevenson, 2007; Mozdzer et al., 2010). In contrast to phosphate, nitrate availability has been shown to result in altered aboveground:belowground biomass ratio of P. australis by favoring aboveground productivity with increasing N addition (Ulrich and Burton, 1985).

In North America, both native and introduced P. australis lineages have the capacity to rapidly take up and assimilate nutrients including inorganic N (Meyerson et al., 2000; Windham and Meyerson, 2003; Hazelton et al., 2010; Mozdzer and Zieman, 2010) and organic N (Mozdzer et al., 2010). However, most studies indicate that, in response to increased N availability, the NAint M lineage is competitively superior to many other wetland species (Chambers et al., 1998; Meyerson et al., 2000; Windham and Meyerson, 2003; Hazelton et al., 2010; Mozdzer et al., 2010, 2013) as well as to the native NAnat lineage (Saltonstall and Stevenson, 2007; Holdredge et al., 2010; Mozdzer et al., 2010, 2013; Mozdzer and Megonigal, 2012). This may be due to its ability to substantially increase carbon assimilation in response to greater N availability (Caplan et al., 2015). Nevertheless, NAint M is also able to regulate its N metabolism to outperform NAnat under low-N conditions (Mozdzer and Megonigal, 2012). This greater plasticity and ability to use available N in both eutrophic and oligotrophic ecosystems can enhance this lineage’s invasiveness by conferring traits such as shifts in phenology as well as increased height growth, leaf area, specific leaf area, leaf area ratio, root mass fraction, and foraging distance (Meadows, 2006; Holdredge et al., 2010; Mozdzer and Zieman, 2010; Mozdzer et al., 2010, 2013; Mozdzer and Megonigal, 2012).

In contrast to NAint M in North America, reeds from the East Asian/Australian group have lower plasticity, N uptake capacity and assimilation rates than co-occurring Spartina alterniflora, a C₄ grass that displaces P. australis on the east coast of China (Zhao et al., 2010). Reed lineages adapted to nutrient-poor sites, which preferably translocate nutrients to storage organs rather than the assimilating tissue, may be outcompeted by stronger competitors for nutrients when in eutrophied settings. They may also respond by increasing productivity and culm height, which, due to their inherently lower tissue N allocation, may lead to poor culm stability and mechanical impairment (Kühl et al., 1997). Lineages that are capable of utilizing nutrients at higher concentrations, especially by allocating more biomass and N to their aboveground organs, may gain competitive advantages that contribute to invasive behavior in eutrophied habitats (Tho et al., 2016). The impacts of eutrophication on plants that are adapted to low vs. high nutrient availability can be studied by using NAnat and NAint M. Lineages EU and MED are also appropriate model systems to investigate eutrophication, although to a lesser extent than the North American lineages (Figure 3).