Habitat-specific responses of leaf traits to soil water conditions in species from a novel alpine swamp meadow community

The physiological and morphological responses of plants to different water conditions in a newly formed alpine swamp meadow community in the eastern Tibetan Plateau were examined. Results showed significant habitat-specific responses and thus suggested that the alpine swamp meadow will be vulnerable to further alteration of water conditions under future climate change.

Zoige wetland, located in eastern part of the Qinghai-Tibet Plateau, is the largest alpine wetland in China and plays a critical role as a water resource for the Yellow River and for conservation of many plant and animal species (Xiang et al., 2009). However, overgrazing and misuse in conjunction with global warming during recent decades have led to severe ecosystem change (Zhang et al., 2011). In particular, there has been a significant trend of aridification in most wetland areas (Qiu et al., 2009;Xiang et al., 2009). With the altered soil water availability, a novel 'alpine swamp meadow' community has emerged, which is composed of species typical of both the alpine wetland and the alpine meadow (Xiang et al., 2009;Ma et al., 2011). Thus, the soil water conditions in the swamp meadow are likely to play a key role in shaping its species composition and structure. Furthermore, the Qinghai-Tibet Plateau has been described as extremely vulnerable to global climate change (Wang and French, 1994;Thompson et al., 2000), which is predicted also to influence the local precipitation conditions. More extreme weather events, including both heavy rain and drought, will be more common under changed climate (Knapp et al., 2008;Smith, 2011); this could have pronounced impacts on the local vegetation, particularly on the novel swamp meadow community.
The species composition of a given plant community is assembled through habitat-filtering processes (Weiher and Keddy, 1999;Cornwell et al., 2006). Keedy (1992) suggested that the habitat selects for a set of species that are somewhat equivalent functionally. As a result of human activity and climate change in recent decades, many novel environments have been created (Mahlstein et al., 2011(Mahlstein et al., , 2013, because of which novel plant communities are emerging, where species composition is mixed from adjacent habitats. The response of species from these communities to environmental factors should be partly derived from and determined by their original habitat (Kudo and Hirao, 2006;Song et al., 2013); for example, their traits may provide a fitness advantage in environments that are similar to the original local environmental conditions (local adaptation; see Kawecki and Ebert, 2004). Alternatively, the plants may be highly plastic and thus show convergent traits; for example, adjusting their traits to accommodate new habitat conditions (Nicotra et al., 2010). To our knowledge, few studies have examined the habitat-specific characteristics or plastic response of species that originated elsewhere and now occur in a newly formed habitat, such as the alpine swamp meadow.
Soil water conditions have long been seen as one of the most important environmental factors for plant establishment, growth and distribution in many vegetation types (Bray, 1997;Chaves et al., 2002;Otsus and Zobel, 2004;García-Sánchez et al., 2007;Flexas et al., 2009;Li et al., 2011). In the alpine area, for instance, studies conducted where periods of drought are frequent during the growing season show that water availability limits plant growth and ecosystem productivity (Scott and Billings, 1964;Walker et al., 1994;Fisk et al., 1998). Leaf morphological traits, such as specific leaf area (SLA) and leaf thickness, play an important role in plant growth, and they are often influenced by environmental factors (Cunningham et al., 1999;Niinements, 2001). Gasexchange properties are the most fundamental of plant physiological processes and are also affected by environmental conditions (Wang and Gao, 2001;Wang and Yuan, 2001;Almeida et al., 2014). For example, photosynthetic capacity strongly affects the growth and dry-matter production of plants (Hirasawa and Hsiao, 1999;Chen et al., 2005;El-Sharkawy, 2011;Catoni and Gratani, 2014) and is greatly affected by soil water conditions (Kimenov et al., 1989;Anderson et al., 1995;Sanhueza et al., 2015). Thus, assessing plant performance in response to water conditions by measuring morphological traits and photosynthetic capacity in different water conditions is a useful approach for understanding the contrasting biology of co-occuring plants in the swamp meadow community.
Using a transplant experimental design, we simulated the two contrasting water conditions that are likely to prevail as consequence of future climate change in the alpine swamp meadow: a mosaic of waterlogged soil and drier soils that are subjected to repeated dry-down. We assessed the responses of leaf morphological (e.g. SLA, leaf thickness) and physiological traits (e.g. gas exchange) to soil water content. The objectives of this study were to determine whether there are habitatspecific differences in trait means and response to water conditions indicative of local adaptation to conditions similar to historic water regimens and to describe the pattern of the response of these species to different water conditions. A priori, we expected a significant but different response of these groups of species to a change in soil water availability reflecting the habitat specificity and/or local adaptation of the species. As a result, we predicted that the plant community of alpine swamp meadow should be unstable and largely influenced by future water regimen, especially precipitation. Understanding the patterns of change in precipitation in future climate change scenarios and reducing other disturbances (such as overgrazing) should be considered in management and conservation decisions for this novel community.

Study species and area
Twelve common species (from eight families) were selected from the alpine swamp meadow plant community (Table 1); six are typical wetland species, whereas the others are usually abundant in alpine meadows (Wu et al., 1994). All of these species are perennial herbaceous plants, in which only the below-ground structures survive through winter. In total, these species account for >50% of total above-ground biomass (55.43 ± 4.32%) and coverage (58.46 ± 3.81%) of the swamp meadow community (H. Li, A. B. Nicotra, D. Xu and G. Du, unpublished work).
The research was conducted at the Field Station of Alpine Meadow and Wetland Ecosystems of Lanzhou University (Maqu branch), in the eastern Qinghai-Tibet Plateau. The site is situated at latitude 33°39′N, longitude 101°53′E, and the average elevation is 3660 m. This site has a typical plateau continental climate with short, cool summers and long, cold winters (Rui et al., 2012). The annual average temperature is 2.2°C and annual precipitation is 672 mm, falling primarily as rain between the months of June and August (Chu et al., 2008, Niu et al., 2014. The growing season is short, with July the season of peak plant growth. Leaves become senescent from mid-August.

Experimental design
In mid-May 2012, 12 species were transplanted from the field into individual plastic pots containing soil from the swamp meadow site. The diameter of the pots was 20 cm at the top and 18 cm at the bottom, and the height was 14 cm. Forty plants per species, each of which had only one or two small leaves emerged (enough for successful identification of the species) were planted, one per pot. Individuals were selected at a consistent size and, presumably, age for each species. The plastic pots were filled with soil from the same site [soil was carefully sieved and mixed before use, with total nitrogen 0.81 ± 0.03% and total phosphorus 0.87 ± 0.09 mg g −1 , which were measured by taking soil cores and analysing with an elemental analyser (FIAstar 5000 Flow Injection Analyzer, FOSS, Denmark)] and arranged in a plastic-covered metal shelter to prevent the pots from receiving natural rainfall. The transmittance of the plastic roof was 85%, measured by a PAR light meter (Decagon Devices, Pullman, WA, USA). A plastic sheet below prevented the plants from accessing ground moisture and kept the irrigation water from seeping into ground soil.
Two weeks after the plants been transplanted, the four least healthy plants per species were excluded and the remaining 36 plants per species were randomly assigned to two water regimens, either the waterlogged treatment (W) or the repeated dry-down treatment (D), to simulate the potential water conditions as consequence of climate change. The 18 pots per treatment were randomly divided into one of three sub-plots containing six pots per species. Pots of a given sub-plot were placed on plastic sheets, which were laid on the ground with the edges fixed a little higher than ground level. During the experimental time, the waterlogged treatment was irrigated with standing water to about 5 cm deep around the pots at all times. The repeated dry-down treatment for a given sub-plot was adjacent to the waterlogged treatment, and plants were given 300 ml water (∼80% of soil field capability) and rewatered every 4 days, at which point the soil water content was about 40-50% of field capability. Plants were observed to ensure that the treatment did not cause wilting during the experimental period (from June to August).

Gas exchange, leaf morphology and growth measurements
Both leaf gas exchange and morphological traits were measured for one fully expanded, newly developed, healthy leaf on each of six plants (randomly selected from 18 pots) for each species in each treatment in mid-July, 6 weeks after water treatment was imposed. These leaves had all developed and expanded after the treatment was imposed (personal observation). Leaf gas exchange was conducted between 10.00 and 12.30 h during a sunny, clear day (third day after watering for dry-down treatment) in mid-July with a GSF-3000 portable photosynthesis system (GFS-3000; Heinz Walz, Germany). The measurements were taken at 1800 µmol m −2 s −1 , which equates to the average light intensity of sunlight during the measurement (measured by the ambient light sensor of GFS-3000), thus ensuring that light was not limiting. The CO 2 concentration was ∼340 ppm (reflecting the measured ambient concentration on site), and relative humidity was between 50 and 65%. The instantaneous water use efficiency (WUEi) was calculated as photosynthetic rate (A max )/transpiration rate (E).
For leaf morphology, one fully expanded, newly developed, healthy leaf from each of the six randomly selected pots was first weighed, then scanned (EPSON 1670; Seiko Epson Corporation, Japan), then dried at 80°C for 48 h and reweighed. The area of each leaf was calculated using ImageJ (Rasband, 2005). The specific leaf area (SLA = area/dry weight, in square metres per gram), leaf thickness (measured with digital callipers to the nearest millimetre), leaf area (LA, in square centimetres) and relative leaf water content [RLWC; (fresh weight-dry weight)/fresh weight, expressed as a percentage] were calculated for each leaf.
At the end of the experiment (mid-August), the total number of leaves (including those that were not fully expanded) generated during the water treatment were counted for each plant as a proxy for growth.

Statistical analysis
The effect of water conditions and species' original habitat (wetland or meadow) on leaf morphology, gas-exchange traits and growth were analysed using two-way ANOVA (type III sum of squares, with treatment and group as main effects, and species as random factors, see Supplementary material Fig. S1 for the responses of individual species to the water conditions). The differences between treatments or between groups for all measured traits were analysed using ANOVA at P = 0.05. Data that violated the ANOVA assumptions of normality of variance were log-transformed. All statistical analyses were performed in SPSS 16.0 for Windows (SPSS, Chicago, IL, USA).

Results
The growth (indicated by leaf production) of the two groups showed strong differences in response to water treatment (Table 2A and Fig. 1). There were significantly more leaves generated in the waterlogged than the dry-down treatment for the wetland group (P < 0.001), whereas more leaves were generated in the dry-down than the waterlogged conditions for the meadow group ( Fig. 1; not significant, P = 0.107).
There were significant differences between groups for leaf morphological traits, but neither group showed significant effects of water treatment (Table 2B and Fig. 2). The meadow group had significantly lower SLA, LA and RLWC but greater leaf thickness compared with the wetland group ( Fig. 2; P < 0.05).
In contrast, both water treatment and the interaction between water treatment and species groups significantly affected gas-exchange traits (Table 2C and Fig. 3). The wetland group had significantly higher A max when grown in waterlogged than in dry-down conditions (P < 0.05), whereas the rankings were reversed for the meadow group ( Fig. 3A; P < 0.05). Likewise, the stomatal conductance (g s ) of wetland group was significantly higher in waterlogged than dry-down conditions ( Fig. 3B; P < 0.001), whereas the meadow group showed no significant effect of growth water conditions on g s (P = 0.535). In dry-down conditions, the A max and g s of meadow group species was significantly higher than that of the wetland group (P < 0.05), whereas the A max of the wetland group was significantly higher than that of the meadow group in waterlogged conditions (P = 0.005). There was no significant difference between groups for g s in waterlogged conditions (P = 0.362). The WUEi of both groups was higher in dry-down conditions, and there was not a significant difference between species groups ( Fig. 3C; P > 0.05).

Discussion
Using a controlled experiment, we tested the responses of plant species in a newly formed 'novel' swamp meadow 4 Conservation Physiology • Volume 3 2015 Research article A, the growth traits; B, leaf morphological traits; C, leaf physiological traits. Degree of freedom for treatment and group = 1 and the denominator d.f. = 120. Abbreviations and terms: A max , net photosynthesis rate; g s , the stomatal conductance for CO 2 ; LA, leaf area; leaf thickness, the thickness of a single leaf in the middle; number of leaves, the total number emerged after the treatment applied, is a proxy for growth; RLWC, relative leaf water content; SLA, specific leaf area; and WUEi, instantaneous water use efficiency. *Indicates significant effect of treatment or their interaction (P-value < 0.05). community to different soil water treatments; the significant but distinctive responses of the two groups of species to changes in soil water availability followed our predictions. Given that the swamp meadow community was formed only recently, following the drying of wetland (Xiang et al., 2009;Ma et al., 2011), the species that emerged in this site could either reflect species from each of the original communities that were highly plastic and thus would show convergent traits in the novel community or, alternatively, could have maintained their original habitat-specific characteristics (local adaptation) and thus show stronger group differences than treatment effects. Song et al. (2013) found that the responses of germination and seedling establishment to soil water conditions are habitat specific in alpine plants on the Tibetan Plateau. Similar signals of habitat specificity are also evident in our study. We found that overall, wetland species performed best (as indicated by leaf growth) in flooded conditions and meadow species best in periodic drought conditions. We also found that gas-exchange traits reflected significant effects of both water treatment and habitat of origin. In contrast, we did not find a significant effect of water conditions on morphological traits for either group of species, although the species groups differed markedly in these traits. This difference in the pattern of response could reflect that ours was a relatively modest water stress (compared with others; see Kramer, 1969) that might not have been strong enough to affect morphological traits (e.g. Nicolás et al., 2005), that we did not impose the stress for long enough to elicit a morphological shift or, alternatively, the result may suggest that morphology is a 'more expensive' shift to achieve and that in this system the species have evolved high levels of plasticity in physiological traits that underlie canalization in morphological ones. Given that significant effects of water treatment on growth were observed, we find the final option most plausible.
Our finding that the wetland group performed better in waterlogged conditions, whereas the meadow group performed better in dry-down conditions is consistent with the habitat specificity, potentially reflecting local adaptation of these groups. A recent study that examined the responses of dominant grass species in a mesic tallgrass prairie to drought has reported contrasting strategies that different species employed to respond to water shortage (Hoover et al., 2014). The results from our study also indicate the sensitivity of these alpine species to water availability, especially for the wetland species (Fenner and Freeman, 2011). In addition, our results highlight the important role of soil water availability in controlling this novel plant community and the high vulnerability of the current community structure to future changes in precipitation regimen, which is expected to become increasingly variable according to climate change projections for the Tibetan Plateau (Liu et al., 2006;You et al., 2008). On the one hand, the fluctuating water conditions may effectively create additional niches that species can occupy, thus promoting coexistence and biodiversity, but on the other hand, if cycles of water availability occur on longer time scales they could result in widespread mortality in either wetland or meadow species. Resulting areas of bare ground could lead to exacerbated water loss, erosion and further degradation of water-holding capacity of the swamp meadow.
Species from both wetland and meadow exhibited higher WUEi in dry-down than waterlogged conditions. The underlying mechanism, however, reflected differences in the physiology and morphology of these two groups of species. In the waterlogged conditions, the wetland species had significantly higher rates of carbon uptake compared with the dry-down conditions, whereas in the meadow species the pattern was reversed. The higher WUEi in the dry-down treatment for wetland species was associated with a steep decrease of g s , whereas for meadow species increased WUEi resulted from increased A max , seemingly somewhat uncoupled from g s (see Farquhar and Sharkey, 1982). Likewise, the combination of higher SLA and LA and lower leaf thickness in the wetland group is typical of species found in mesic environments (Niinemets, 2001;Nicotra et al., 2007). These results suggest a degree of local adaptation (Kawecki and Ebert, 2004).
Theory suggests that the greater the difference in functional traits between species, the more likely there will be a strong difference in response to limiting factors (Adler et al., 2013). Here, we demonstrated significant differences in leaf traits and growth response between the wetland and meadow in response to water availability. The results suggest that the community composition of the alpine swamp meadow in the eastern Qinghai-Tibet Plateau is likely to remain highly sensitive to changes in water conditions and could be substantially influenced by future water supply; thus, it is unstable. At this point, we cannot make a firm prediction about which way it will go; therefore, further research is needed in order to inform management and conservation decisions adequately. Wetland aridification is accelerating in the Tibetan Plateau as well as on a global scale (Dugan, 1992;Xiang et al., 2009). Our study makes a substantive contribution to understanding of species and community dynamics of the novel swamp meadow in this time of significant global climate change (Liu et al., 2006;Knapp et al., 2008;You et al., 2008;Smith, 2011). Our results may assist with conservation, management and plant community restoration actions; for example, management to reduce disturbances that may affect soil water availability (e.g. overgrazing, drainage and peat exploitation) could improve stability of the community.

Supplementary material
Supplementary material is available at Conservation Physiology online.