Plant hydraulics and photosynthesis of 34 woody species from different successional stages of subtropical forests

Authors

  • SHI-DAN ZHU,

    1. Key Laboratory of Vegetation Restoration and Management of Degraded Ecosystems, South China Botanical Garden, Chinese Academy of Sciences, Xingke Road 723, Tianhe District, Guangzhou 510650, China
    Search for more papers by this author
    • These two authors contributed equally to this work.

  • JUAN-JUAN SONG,

    1. Key Laboratory of Vegetation Restoration and Management of Degraded Ecosystems, South China Botanical Garden, Chinese Academy of Sciences, Xingke Road 723, Tianhe District, Guangzhou 510650, China
    Search for more papers by this author
    • These two authors contributed equally to this work.

  • RONG-HUA LI,

    1. Key Laboratory of Vegetation Restoration and Management of Degraded Ecosystems, South China Botanical Garden, Chinese Academy of Sciences, Xingke Road 723, Tianhe District, Guangzhou 510650, China
    2. University of Chinese Academy of Sciences, 19A Yuquan Road, Beijing 100049, China
    Search for more papers by this author
  • QING YE

    Corresponding author
    1. Key Laboratory of Vegetation Restoration and Management of Degraded Ecosystems, South China Botanical Garden, Chinese Academy of Sciences, Xingke Road 723, Tianhe District, Guangzhou 510650, China
    Search for more papers by this author

Q. Ye. Fax: +86 020 3725 2615; e-mail: qye@scbg.ac.cn

ABSTRACT

It is important to understand the ecophysiological characters of plants when exploring mechanisms underlying species substitution in the process of plant succession. In the present study, we selected 34 woody species from different stages of secondary succession in subtropical forests of southern China, and measured their hydraulic conductivity, gas exchange rates, leaf nutrients and drought-tolerance traits such as xylem resistance to cavitation, turgor loss point and carbon isotope ratio. Principal component analysis revealed that early-, mid- and late-successional species were significantly separated along axis 1, which was strongly associated with hydraulic-photosynthetic coordination. In contrast to species distributed in late-successional forest, early-successional species had the highest hydraulic conductivity, net photosynthetic rates, photosynthetic nitrogen and phosphorus use efficiencies, but had the lowest photosynthetic water-use efficiency. However, changes of the measured drought-tolerance traits of the 34 species along the succession did not demonstrate a clear trend – no significant correlations between these traits and plant successional stages were found. Moreover, the trade-off between hydraulic efficiency and safety was not identified. Taken together, our results suggested that hydraulic efficiency and photosynthetic function, rather than drought tolerance, play an important role in species distributions along plant succession in subtropical forests.

INTRODUCTION

From heavily disturbed open sites to dense forests, the physical environments (e.g. light density, air humidity and temperature, soil nutrient and water availability) vary distinctly. Thus, plants from different stages of secondary succession may exhibit contrasting ecophysiological traits, which have long been an important research topic in plant ecophysiology (Bazzaz & Pickett 1980; Reich, Ellsworth & Uhl 1995; Navas et al. 2010). Understanding the ecophysiology of plant succession not only helps us explain how species replacement occurs, but also provides guidelines for forest management and rebuilding (McGill et al. 2006).

Plant functional traits are key characteristics used to classify plant groups adapted to different environmental factors (Garnier et al. 2004; Targetti et al. 2012). A number of studies have suggested that functional traits underlying the growth survival trade-off of plants might be instructive to predictions of species abundance and distribution along the succession (Coley, Bryant & Chapin 1985; Poorter 2009). For instance, early-successional plants are usually fast-growing and light-demanding species, having features like higher specific leaf area (SLA), faster leaf turnover rate, greater photosynthetic capacity and growth rate compared with late-successional species. This is because growth rather than survival is more important for plants to establish in open, resource-rich habitats (Riddoch et al. 1991; Navas et al. 2003; Poorter & Bongers 2006). Recently, McCulloh et al. (2011) found that early- and late-successional tropical tree species differed in plant hydraulic architecture, with early-successional species having lower wood density and greater hydraulic conductance. Plant hydraulics is closely related to carbon economy, a well-accepted example being the hydraulic-photosynthetic coordination of plants (Brodribb & Feild 2000; Brodribb, Holbrook & Gutierrez 2002; Santiago et al. 2004; Zhang & Cao 2009). Reflecting a balance between carbon gain and water loss, such coordination may be an important determinant of species distribution at different stages of plant succession, and is widely used to identify the adaptive strategy of different plant groups, such as in comparative studies between gymnosperms and angiosperms (Feild & Balun 2008), sun and shade trees (Campanello, Gatti & Goldstein 2008), or trees and lianas (Zhu & Cao 2009).

In comparison to mature forests, early-successional forests usually have higher soil and air temperature, greater vapour pressure deficit and lower relative air humidity and soil water content (Pineda-Garcia, Paz & Meinzer 2012). Given the differentiation of microhabitats (e.g. spatiotemporal variation in soil water availability) in forests at different successional stages, one would expect that the distribution of species along the succession might partially depend on plant drought resistance/tolerance (Lopez et al. 2005; Canham, Froend & Stock 2009; Pineda-Garcia, Paz & Tinoco-Ojanguren 2011; Markesteijn et al. 2011a). Markesteijn et al. (2011b) found that pioneer tree species and shade-tolerant species showed contrasting hydraulic strategies, that is the former had higher hydraulic conductivity but were more vulnerable to cavitation, while the latter had lower hydraulic conductivity but were more resistant to cavitation. A number of studies have shown a trade-off between plant hydraulic efficiency (hydraulic conductivity) and safety (xylem vulnerability to cavitation; Kolb & Sperry 1999; Pinol & Sala 2000; Hacke et al. 2006). However, contradictory results were found in quite a few studies where such a trade-off is obscure (Tyree, Davis & Cochard 1994; Meinzer et al. 2010), particularly when evaluating the data with phylogenetically independent contrasts (PIC; Maherali et al. 2006; Bhaskar, Valiente-Banuet & Ackerly 2007; Fan et al. 2011). In addition, traits like water potential at turgor loss point (ψtlp) (Kolb & Sperry 1999; Baltzer et al. 2008) and long-term water-use efficiency (δ13C; Farquhar, Ehleringer & Hubick 1989; Bai et al. 2008) are often measured to evaluate the tolerance of plant to drought stress. In general, species distributed in low-water availability sites have low ψtlp (Bartlett, Scoffoni & Sack 2012) and high δ13C (Van de Water, Leavitt & Betancourt 2002).

Most primary forests in southern China have been substantially destroyed during the last century, a major consequence of which has been severe soil erosion (Yu & Peng 1996). Heavily disturbed sites are quickly colonized by shrubs and grasses, change to heliophilous broad-leaved forests (on a time scale of several decades) and then gradually become mesophilous evergreen broad-leaved forests (over several hundred years). This process is identified as typical secondary succession in the subtropical areas of southern China (Peng 1996). Plants established in forests at different successional stages in this area may experience dramatically different environmental conditions (e.g. light, temperature and nutrients). This observation prompted us to look at patterns of plant functional traits (SLA, leaf nutrients, photosynthetic capacity and hydraulic conductivity) across species from different successional forests, with the objective of identifying mechanisms contributing to species substitution along the succession in terms of trait–environment relationships. On the other hand, subtropical forests in southern China are influenced by monsoon climate, resulting in distinct rainfall seasonality. It is known that relative air humidity and soil water content in early-successional forests are lower than in mid- and late forests (Peng 1996). This situation could be exacerbated during the dry season (from October to March), and may become a big challenge to plant survival. Hence, drought tolerance of plants in this area might be a factor determining species distributions along the succession. Therefore, a secondary objective of the present study was to investigate plant traits related to drought tolerance such as xylem resistance to cavitation, turgor loss point and carbon isotope ratio.

MATERIALS AND METHODS

Study site and plant materials

The present study was carried out in Dinghushan Forest Ecosystem Research Station (DFERS; 21°09′21′′–21°11′30′′ N, 112°30′39′′–112°33′41′′ E), Chinese Academy of Sciences, central Guangdong, south China. This region has a typical monsoon climate, with a distinct dry season from October to March. Mean annual total precipitation is c. 1900 mm, of which about 80% occurs in the wet season (April to September). Mean annual temperature is 21.4 °C, with monthly temperature ranging from 12.6 °C (January) to 28.0 °C (July). Being a hilly land, soil water-holding capacity is limited in this region (Peng 1996). There is an entire subtropical secondary succession sere in DFERS:

  • 1Early-successional stage, 200–300 m in altitude. The vegetation is under frequent human disturbances. Light-demanding shrub species such as Alchornea trewioides, Rhodomyrtus tomentosa and Mallotus paniculatus dominate in the field. Canopy height is about 2 m. Soil (0–10 cm) pH 4.29, organic matter (SOM) 31.50 g kg–1, N 1.10 g kg–1, P 0.40 g kg–1. Mean annual soil water storage (SWS; 0–75 cm) is 254.36 mm.
  • 2Mid-successional stage, 220–300 m in altitude. This forest is about 40–50 years old. Canopy height is about 12 m. The most dominant tree species are Castanopsis chinensis and Schima superba. Soil pH 3.95, SOM 44.00 g kg–1, N 1.40 g kg–1, P 0.37 g kg–1, SWS 324.98 mm.
  • 3Late-successional stage, 220–300 m in altitude. This forest is over 400 years old and viewed as climax vegetation in southern China. Canopy height and coverage are 15 m and 95%, respectively. Tree species such as Cryptocarya chinensis, Machilus chinensis and Syzygium levinei dominate in this forest. Soil pH 3.89, SOM 48.69 g kg–1, N 1.76 g kg–1, P 0.29 g kg–1, SWS 381.03 mm.

In the present study, a total of 34 woody species were tested from the three successional forests in DFERS (Table 1). Three to five mature individuals were selected for each species. All the selected species are common and typical in each stage of the succession according to long-term community studies of DFERS (Zhang 2011). Most species are evergreen (31 species). To avoid the impact of seasonal drought on ecophysiological traits of these species, all the experiments were carried out from June to August in the year 2011, except for the measurements of photosynthetic photon flux density (PPFD), which were conducted in June and July of 2012.

Table 1. Characteristics of 34 woody species tested in the present study
SpeciesFamilyCodeGrowth formLeaf habitLayer in forestSampling siteCanopy (sampling) height
Early-successional stage       
 Alchornea trewioides (Benth.) Muell. Arg.EuphorbiaceaeAtShrubDeciduousCanopyOpen site2 m (2 m)
 Clerodendrum fortunatum L.VerbenaceaeCfShrubEvergreenCanopyOpen site2 m (2 m)
 Mallotus paniculatus (Lam.) Muell. Arg.EuphorbiaceaeMpShrubEvergreenCanopyOpen site2 m (2 m)
 Melastoma sanguineum SimsMelastomataceaeMsShrubEvergreenCanopyOpen site2 m (2 m)
 M. candidum D. DonMelastomataceaeMcShrubEvergreenCanopyOpen site2 m (2 m)
 Rhodomyrtus tomentosa (Ait.) Hassk.MyrtaceaeRtShrubEvergreenCanopyOpen site1.5 m (1.5 m)
Mid-successional stage       
 Castanopsis chinensis (Sprengel) HanceFagaceaeCcTreeEvergreenCanopyForest road9–10 m (6–7 m)
 Ca. fissa (Champ. ex Benth) Rehd. et Wils.FagaceaeCafTreeEvergreenCanopyForest road8–9 m (6 m)
 Diospyros morrisiana HanceEbenaceaeDmTreeEvergreenMid-canopyForest road4–5 m (4 m)
 Melicope pteleifolia (Champ. ex Benth.) HartleyRutaceaeMepTreeEvergreenUnderstoryForest road2.5 m (2.5 m)
 Sapium sebiferum (L.) Roxb.EuphorbiaceaeSasTreeDeciduousMid-canopyForest road5–6 m (4 m)
 Schefflera heptaphylla (L.) FrodinAraliaceaeShTreeEvergreenMid-canopyForest road5 m (4 m)
 Schima superba Gardn. et Champ.TheaceaeScsTreeEvergreenCanopyForest road8–10 m (6–7 m)
 Toxicodendron succedaneum (L.) O. KuntzeAnacardiaceaeTsTreeDeciduousMid-canopyForest road4–5 m (4 m)
Late-successional stage       
 Acmena acuminatissima (Blume) Merr. et PerryMyrtaceaeAaTreeEvergreenMid-canopyGap6–7 m (5 m)
 Acronychia pedunculata (L.) Miq.RutaceaeApTreeEvergreenMid-canopyForest road5–6 m (5 m)
 Aidia canthioides (Champ. ex Benth.) Masam.RubiaceaeAcTreeEvergreenUnderstoryForest road3 m (3 m)
 Aporusa dioica (Roxb.) Muell. Arg.EuphorbiaceaeAdTreeEvergreenMid-canopyGap5–6 m (5 m)
 Ardisia quinquegona Bl.MyrsinaceaeAqTreeEvergreenUnderstoryForest road3 m (3 m)
 Blastus cochinchinensis Lour.MelastomataceaeBcTreeEvergreenUnderstoryForest road3 m (3 m)
 Cryptocarya chinensis (Hance) Hemsl.LauraceaeChTreeEvergreenCanopyGap9–11 m (6–7 m)
 Cr. concinna HanceLauraceaeCrcTreeEvergreenCanopyGap9–10 m (6–7 m)
 Diplospora dubia (Lindl.) Masam.RubiaceaeDdTreeEvergreenUnderstoryForest road3 m (3 m)
 Gironniera subaequalis Planch.UlmaceaeGsTreeEvergreenMid-canopyGap7–8 m (5 m)
 Machilus chinensis (Champ. ex Benth.) Hemsl.LauraceaeMacTreeEvergreenCanopyGap8–10 m (6–7 m)
 Memecylon ligustrifolium Champ.MelastomataceaeMlTreeEvergreenMid-canopyForest road5–6 m (5 m)
 Microdesmis caseariifolia Planch.PandaceaeMicTreeEvergreenMid-canopyForest road5–6 m (5 m)
 Mischocarpus pentapetalus (Roxb.) Radlk.SapindaceaeMipTreeEvergreenMid-canopyGap5–6 m (5 m)
 Psychotria rubra (Lour.) Poir.RubiaceaePrTreeEvergreenUnderstoryForest road3 m (3 m)
 Pygeum topengii Merr.RosaceaePtTreeEvergreenMid-canopyGap6–7 m (5 m)
 Sarcosperma laurinum (Bench.) Hook. f.SapotaceaeSalTreeEvergreenMid-canopyForest road6–7 m (5 m)
 Syzygium levinei Merr. et PerryMyrtaceaeSylTreeEvergreenMid-canopyForest road7–8 m (5 m)
 Sy. rehderianum Merr. et PerryMyrtaceaeSrTreeEvergreenMid-canopyForest road6–7 m (5 m)
 Xanthophyllum hainanense HuPolygalaceaeXhTreeEvergreenMid-canopyForest road5–6 m (5 m)

It is known that light levels may significantly influence the hydraulic properties (Cochard, Lemoine & Dreyer 1999; Raimondo et al. 2009) and photosynthetic characteristics (Thomas & Bazzaz 1999) of plants. Hence, to compare these ecophysiological traits of different successional species, it is critically important that samples are taken from plants growing in habitats with similar light conditions (Markesteijn et al. 2011b). The three successional forests in DFERS are located at a similar altitude of around 260 m above sea level. Early-successional species are grown in open sites exposed to full sun. Therefore, samples of mid- and late-successional forests species were taken from individuals growing in large gaps or along forest roads, such that plants could receive full or substantial amounts of overhead light (Fan et al. 2011; Markesteijn et al. 2011a,b). To further ensure this, four plant sampling sites (used for hydraulic and photosynthetic measurements) from each of the three forests were selected to measure PPFD. Using three sets of quantum sensors (Li-1400, Li-Cor, Lincoln, NE, USA), light levels in the three forests were determined simultaneously. Sensors were positioned at similar heights to branches sampled for hydraulic and photosynthetic measurements (Supporting Information Table S1). The results showed that plant samples taken from different successional forests experienced similar light levels, including the period of 09:00–11:00 h, when photosynthetic measurements were carried out (Supporting Information Fig. S1).

Branch samplings and photosynthetic measurements for plants in early-successional communities were relatively easy because plants were typically less than 2 m high. It was more difficult, however, to sample terminal branches from the canopy of some plant species in mid- and late-successional forests; for trees that were between 7 and 10 m high (e.g. Gironniera subaequalis, Sc. superba and Sy. levinei etc.), we were able to sample their terminal or lateral sun-exposed branches using long-reach pruners (4 m in length). For taller trees such as Ca. chinensis, Cr. chinensis and M. chinensis, we managed to obtain their sun-exposed branches with the aid of long-reach pruners by standing on ladders (3 m high). For photosynthetic measurements on tall plant species from the mid- and late-successional forests, sun-exposed branches (at similar heights as those taken for hydraulic measurements) were cut off with the aid of long-reach pruners. The cut ends of the detached branches (at least 20 mm in diameter) were immediately immersed in water and re-cut, prior to subsequent leaf gas exchange measurements (Thomas & Bazzaz 1999; Wang et al. 2008; Wyka et al. 2012).

Hydraulic conductivity, sapwood density and xylem vulnerability curve

Ten terminal branches (8–10 mm in diameter) were sampled early in the morning from three to five mature individuals for each species, sealed in black plastic bags with moist towel and immediately transported to the laboratory. All the stem samples were re-cut underwater and the cut ends were trimmed with a razor blade. The length of stem segments for measurements of stem hydraulic conductivity was 20–25 cm.

Stem hydraulic conductivity was measured using the method described by Sperry, Donnelly & Tyree (1988). In an air-conditioned laboratory (26 °C), branch segments were flushed at a pressure of 0.1 MPa for 20 min to remove air embolisms. The flush fluid was 20 mmol filtered (0.2 µm) KCl solutions. Hydraulic conductivity per unit pressure gradient (kh) is equal to the ratio between water flux through an excised stem segment and the pressure gradient causing the flow. Sapwood specific conductivity (kS; kg m–1 s–1 MPa–1) is equal to kh divided by sapwood cross-section area. Leaf-specific hydraulic conductivity (kL; kg m–1 s–1 MPa–1) was calculated as the ratio of kh to the leaf area, which is a measure of the hydraulic sufficiency of the stem to supply water to distal leaves. Leaf area/sapwood area ratio (AL/AS; m2 cm–2) was calculated as the ratio of leaf area attached per unit sapwood cross-section area.

Sapwood density (WD) was measured from the same branches used for hydraulic conductivity measurements. After removing the bark or pith, the fresh sapwood was immersed in distilled water overnight allowing for saturation of the sample. Volume of sapwood was determined by water displacement, and its dry mass was determined after oven drying at 70 °C for 72 h. WD was the ratio of dry mass to fresh volume.

The air injection method was used to establish xylem vulnerability curves (Cochard, Cruiziat & Tyree 1992). Segments were placed inside a pressure chamber (PMS, Corvallis, OR, USA) with both ends protruding. Proximal ends were connected to the measuring equipment, and maximum hydraulic conductivity was measured. We raised the pressure inside the chamber to 0.5 MPa and maintained it for 5–10 min, before reducing it to a basal value of 0.01 MPa. We then waited for 10–20 min to allow the system to equilibrate, before repeating the conductivity measurement. We repeated this process, raising the injection pressure by 0.5 or 1 MPa each time until more than 80% of kS was lost. The residual pressure inside the chamber was maintained to make sure that no refilling could occur during measurements. We fitted a vulnerability curve by following the equation (Pammenter & Vander Willigen 1998):

image

Where PLC is percentage loss of hydraulic conductivity, P is the applied pressure, b is the pressure causing 50% loss of hydraulic conductivity (-P50).

Pressure–volume relations

Terminal branches were harvested from three to five individuals for each species. Branch ends were re-cut underwater and rehydrated until leaf water potential exceeded −0.05 MPa, after which leaves were detached for P-V curve determination. Leaf weight and water potential were measured periodically during slow desiccation of the sample in the laboratory. After all balanced pressure-weight measurements, leaves were oven-dried for 72 h at 70 °C to determine the dry weight. Leaf water potential at turgor loss point (ψtlp) was determined by a pressure–volume relationship analysis program developed by Schulte & Hinckley (1985).

SLA, nutrients and carbon isotope ratio

For SLA, 20 fully expanded sun-exposed leaves were collected from each individual. Total areas were measured with a leaf area meter (Li-3000A; Li-Cor). The leaves were then oven dried for 72 h at 70 °C to determine their dry weight. SLA was calculated as leaf area per dry mass. The dry leaves were then ground and homogenized for subsequent analyses.

Leaf chemical analyses were conducted in the Public Laboratory of South China Botanical Garden. Total nitrogen concentrations (N) were determined by Kjeldhal analysis. Total phosphorus concentrations (P) were determined using atomic absorption spectrum photometry. Leaf carbon isotope discrimination ratio (δ13C) can be used to estimate long-term water-use efficiency of leaves in natural vegetation (Farquhar et al. 1989). The δ13C was measured with an elemental analyser (Flash EA 1112, Thermo Electron Corporation, Waltham, MA, USA) interfaced to an isotope ratio mass spectrometer (Thermo Finnigan MAT DELTAplusXP, Thermo Electron Corporation, MA, USA) at the Institute of Desertification Studies, Chinese Academy of Forestry (Beijing, China). δ13C was calculated as:

image

Where Rsample and Rstandard were the ratios of 13C/12C in the sample and in the Pee Dee Belemnite standard, respectively.

Leaf gas exchange rate

Maximum net CO2 assimilation rate (A) and stomatal conductance (gs) were measured between 9:00 and 11:00 h with a portable photosynthesis system (Li-6400, Li-Cor). The PPFD was set between 1000 and 1500 µmol m–2 s–1, which corresponds to photosynthetic saturation for all species. Three to five individuals were selected for each species and five to six sun-exposed leaves were selected from each individual for photosynthetic measurements. Instantaneous water-use efficiency (WUEi) was calculated as A/gs. Photosynthetic nitrogen and phosphorus use efficiencies (PNUE and PPUE) were calculated as A/N and A/P, respectively.

Statistical analyses

Differences in hydraulic and photosynthetic traits among different successional groups were tested with one-way anova by using SPSS version 13.0 software package (SPSS, Chicago, IL, USA). Multivariate associations of the 16 leaf and stem traits (Supporting Information Table S3) were analysed with a principal component analysis (PCA), using log10-transformed values of the means of the traits. Mean factor loading values of different successional species on the first two PC axes were also compared to examine whether these groups were significantly separated along PC axes.

Relationships among hydraulic conductivity and WD, xylem vulnerability and photosynthetic function were analysed using traditional regression analysis, and the correlations were also calculated using PIC to evaluate whether the observed traits associations were the result of repeated evolutionary divergences. Assuming equal branch lengths for all analyses, we generated a phylogenetic tree using Phylomatic program (Webb & Donoghue 2005) based on the Angiosperm Phylogeny Group III classification of angiosperms. All 34 species tested in the present study were kept and 33 evolutionary contrasts were generated in the final phylogenetic tree (Supporting Information Fig. S2). Evolutionary associations were assessed with ‘analysis of traits’, a module of Phylocom version 4.1 (Webb, Ackerly & Kembel 2008).

RESULTS

Plants from late-successional stages had significantly lower kS compared to those from early- and mid-successional forests in which there were virtually no differences in plant kS (Fig. 1a). Plant kL decreased significantly in the order of early–mid–late-successional stages (Fig. 1b). This may be due to the fact that, for a given branch sampled for the experiments, plants from early-successional stages had a smaller value of AL/As than those in mid-successional stages, and the highest AL/AS was found for plants in late-successional forest (Supporting Information Table S2).

Figure 1.

Mean values (±SE) of sapwood-specific hydraulic conductivity (kS) and leaf-specific hydraulic conductivity (kL) for 34 woody species in three successional communities. White, grey and black bars represent species from early-, mid- and late-successional stages, respectively. Species abbreviations follow Table 1. Different letters at the top of each column denote significant differences among communities (P < 0.05, anova).

For gas exchange characteristics, both Aarea and gs of the tested 34 species showed a significant decreasing pattern in the order of early- to late-successional stages, indicating a decrease of photosynthetic capacity of the plants along the succession (Fig. 2).

Figure 2.

Mean values (±SE) of maximum CO2 assimilation rate (Aarea) and stomatal conductance (gs) for 34 woody species in three successional communities. White, grey and black bars represent species from early-, mid- and late-successional stages, respectively. Species abbreviations follow Table 1. Different letters denote significant differences among communities (P < 0.05, anova).

Plants in early-successional stages showed the highest P50 value on average (Fig. 3a). No significant difference was found for P50 when comparing species from mid- and late-successional forests. As for water potential at turgor loss point (ψtlp), plants in mid-successional stages had the lowest value on average, while species from early- and late-successional forests had higher ψtlp but were similar to each other on average (Fig. 3b).

Figure 3.

Mean values (±SE) of xylem tension at 50% loss of hydraulic conductivity (P50) and water potential at turgor loss point (ψtlp) for 34 woody species in three successional communities (P50 for each species is shown without standard error). White, grey and black bars represent species from early-, mid- and late-successional stages, respectively. Different letters denote significant differences among communities (P < 0.05, anova).

Cross-species analysis showed that kS was negatively related to WD, but correlation between kS and P50 was not significant (Fig. 4a,b). Phylogenetic analysis results were similar to those of the cross-species analysis (Fig. 4c,d).

Figure 4.

Correlations between sapwood-specific hydraulic conductivity (kS) and xylem tension at 50% loss of hydraulic conductivity (P50), and sapwood density (WD). (a,b) traditional cross-species analysis; (c,d) phylogenetically independent contrasts (PIC). White, grey and black circles indicate species from early-, mid- and late-successional stages, respectively.

k L was positively related to Aarea, gs, PNUE and PPUE, and negatively related to WUEi (Fig. 5a–e). Correlations based on PIC (Fig. 5f–j) were consistent with the above cross-species analysis, although the correlation coefficients (rp) were somewhat lower in strength than rc.

Figure 5.

Correlations among leaf-specific hydraulic conductivity (kL) and photosynthetic traits. (a–e) traditional cross-species analysis; (f–j) phylogenetically independent contrasts (PIC). White, grey and black circles indicate species from early-, mid- and late-successional stages, respectively. Aarea, maximum CO2 assimilation rate per unit area; gs, stomatal conductance; WUEi, instantaneous water-use efficiency; PNUE, photosynthetic nitrogen use efficiency; PPUE, photosynthetic phosphorus use efficiency.

PCA was employed to evaluate how traits of hydraulics and photosynthesis were associated for subtropical woody species at different successional stages (Fig. 6). PCA axis 1 showed strong positive loadings for plant hydraulic traits (kS and kL) and photosynthetic traits (Aarea, gs, Amass, PNUE and PPUE), and had negative loadings for WD and WUEi. Leaf nutrient contents (N and P) and SLA were positive loads of PCA axis 2. Neither PCA axis 1 nor axis 2 was significantly associated with traits related to drought-tolerance (P50, ψtlp and δ13C; Fig. 6a, Supporting Information Table S5). Plant species from different successional stages were well separated along the first PCA axis (Fig. 6b, Supporting Information Table S2). In contrast to species in late-successional forest, early-successional species showed the highest hydraulic conductivity (kS and kL), Aarea, PNUE and PPUE, but had the lowest WUEi.

Figure 6.

Principal component analysis for (a) 16 traits and (b) 34 woody species for the first two axes. White, grey and black circles indicate species from early-, mid- and late-successional stages, respectively. kS, sapwood-specific hydraulic conductivity; kL, leaf-specific hydraulic conductivity; AL/AS, leaf area/sapwood area ratio; WD, sapwood density (WD); P50, xylem tension at 50% loss of kS; ψtlp, water potential at turgor loss point; SLA, specific leaf area; N, leaf nitrogen; P, leaf phosphorus; Amass, maximum CO2 assimilation rate per unit mass; δ13C, leaf carbon isotope discrimination ratio.

DISCUSSION

Changes in plant hydraulic conductivity, photosynthetic capacity and SLA along the forest succession

When compared to species in late-successional forest, plants distributed at early stages of succession had the highest kS, kL, Aarea, Amass and gs (Figs 1 & 2). In the subtropical areas of southern China, long-term human disturbances ruined primary forests and produced large-area open sites, resulting in the establishment of fast-growing and light-demanding shrub species such as A. trewioides and R. tomentosa. These species showed fairly high Aarea and gs at mean values of 15.1 µmol m–2 s–1 and 0.56 mol m–2 s–1, respectively (Fig. 2). Higher gas exchange rates were accompanied by more efficient water transport (e.g. higher kS and kL) as shown in Fig. 1 and Supporting Information Table S2. The early-successional species had relatively low WD, indicating that a large fraction of sapwood might be occupied by wide vessels (Zanne et al. 2010), resulting in high kS (McCulloh et al. 2011). AL/AS was found to be smaller for early-successional species, suggesting less leaf area was supported by sapwood at a given area, which in turn promoted higher kL of these species. In mid-successional forest, pioneer tree species (e.g. Ca. chinensis and Sc. superba) formed a closed canopy; species dominant in early-successional forest were filtered out, providing suitable conditions for shade-tolerance species such as Cr. chinensis, M. chinensis and Sy. levinei to prosper. Late-successional species might invest more biomass in survival rather than growth in shaded environments with limited resources (Coley et al. 1985; Poorter 2009), thus bearing lower Aarea, gs, Amass, kS and kL (Supporting Information Table S2).

Early-successional species also showed significantly higher photosynthetic nutrient use efficiency. Phosphorus limitation is a common phenomenon in subtropical forests in southern China (Han et al. 2005). Generally, leaf N:P >16 suggests that plants suffer P limitation (Koerselman & Meuleman 1996). In this study, we found species from all the three successional forests had mean values of leaf N:P >24 (Supporting Information Table S2). Early-successional species that could quickly colonize open sites after disturbances might be partially explained by their high PNUE and PPUE (Supporting Information Table S2). Similar to results found in Amazonian forests by Reich et al. (1995), plants from early-, mid- and late-successional stages had similar leaf total N, thus higher photosynthetic rates of the early-successional species indicated a larger fraction of organic nitrogen was allocated to photosynthetic apparatus in those fast-growing species (Poorter & Evans 1998).

SLA has been regarded as a key leaf trait involved in the ‘leaf economics spectrum’, and it is closely related to other important functional traits of hydraulic conductivity (Mitchell et al. 2008; Bucci et al. 2009; Fan et al. 2011), photosynthetic function and growth rate (Poorter & Bongers 2006). A number of studies have shown that early-successional species had higher SLA than that of late-successional species in different forest types, for example, in the Mediterranean region (Garnier et al. 2004; Navas et al. 2010) or in a lowland tropical area of French Guiana (Rijkers, Pons & Bongers 2000). In the present study, we did not find clear patterns of SLA among plant groups from the three subtropical successional forests. It turned out that early and late species had higher mean values of SLA than species in mid-successional forest (Supporting Information Table S2). The lack of a clear pattern in SLA might be due to the large variations of SLA across sites at different successional stages and/or among species within the same site (Bassow & Bazzaz 1997; Jacobsen et al. 2008), leading to the lack of significant correlations between SLA and plant functional traits (e.g. kS, kL, WD, P50, Aarea, gs and PNUE) measured in the present study (Supporting Information Table S4).

Traits related to drought tolerance of plants in different successional groups

Unlike hydraulic conductivity and photosynthetic properties, xylem resistance to cavitation (P50) did not show a clear pattern along the succession (Fig. 3). Early-successional species had the highest mean value of P50 among the three successional groups, but no significant difference was found between mid- and late-successional species. The results indicating a lack of a clear pattern in P50 for different successional species are consistent with findings from a recent article by Pineda-Garcia et al. (2012). These authors explored mechanisms of drought resistance in early- and late-successional species from a tropical dry forest, and concluded that it may be misleading to evaluate plant performance in response to drought stress solely based on xylem resistance to cavitation. There might be other mechanisms contributing to drought resistance, including high stem water storage capacity and variations in the leaf-shedding strategy of plants.

Large interspecific variations of P50 were found for species from mid- and late-successional stages. For example, within the late-successional community, P50 ranged from the most vulnerable value (−0.8 MPa) for Acmena acuminatissima to the most resistant value (−4.4 MPa) for Blastus cochinchinensis. Similar results were found by Lopez et al. (2005) in a tropical rain forest where shade-tolerant woody species showed a wide range of P50 (from −1.9 MPa to −4.1 MPa). Since xylem vulnerability was positively correlated with leaf water potential (Lopez et al. 2005), it is interesting that pre-dawn leaf water potential in the dry season (December) for A. acuminatissima and B. cochinchinensis was found to be −0.1 MPa and −0.7 MPa, respectively (Zhu et al., unpublished data). Because pre-dawn leaf water potential is a proxy of soil water potential at root zone, assuming plants had no transpiration and water refill at night (Bucci et al. 2005), it might be plausible that in the dry season, A. acuminatissima could be capable of accessing soil water at depth, but shallow-rooted species such as B. cochinchinensis might suffer water deficit; high resistance to cavitation should therefore be helpful for the avoidance of hydraulic dysfunction (Pockman & Sperry 2000).

Turgor loss point (ψtlp) for leaves or terminal shoots provides a promising index to estimate leaf-level desiccation tolerance of plants from different rainfall sites (Lenz, Wright & Westoby 2006). This leaf trait was particularly used to describe drought adaptations of sandy plants (Dong & Zhang 2001) or seedlings (Baltzer et al. 2008). Our results did not show a clear trend with changes in leaf ψtlp along the succession. Mid-successional species had lower ψtlp on average compared to plants from early- and late-successional forests (Fig. 3b). Leaves of early-successional species and some shade-tolerant tree species were less resistant to desiccation as they showed greater values of leaf ψtlp. This might be due to the relatively higher SLA of these plants (Supporting Information Table S4), a similar result as found by Bucci et al. (2004), as leaves with lower SLA are usually thicker and denser and have more dry matter content (Vendramini et al. 2002), enabling plants to maintain living tissues under more negative water potential (low ψtlp; Baltzer et al. 2008). Values of leaf δ13C showed no significant difference among different successional communities (Supporting Information Table S2). It is known that leaf δ13C is influenced by environmental factors such as altitude (Hultine & Marshall 2000), light density (Holtum & Winter 2005) and water availability (Knight, Livingston & Van Kessel 1994). Hence, the combined effects of environmental factors on isotopic composition of leaves from different successional species might diminish the potential difference of leaf δ13C in wet seasons.

Traits correlations and PCA

A number of studies showed that WD, kS and P50 were closely related to each other. Species with higher WD had lower kS but were more resistant to xylem cavitation (Hacke et al. 2001; Bucci et al. 2004; Jacobsen et al. 2007; Markesteijn et al. 2011b). In the present study, an evolutionary correlation between kS and WD was found (Fig. 4), species with dense wood often having xylem largely occupied by non-vessel components such as fibres, thus contributing to low kS (few vessel components; Zanne et al. 2010). However, we did not find a significant relationship between WD and P50 (Supporting Information Table S4), and trade-off between hydraulic efficiency and safety was obscured according to both traditional cross-species correlations and PIC analysis (Fig. 4). For example, the mid-successional species Sc. superba had moderate WD and kS (0.57 g cm–3 and 2.82 kg m–1 s–1 MPa–1, respectively), but was highly resistant to drought-induced cavitation (P50 = −4.3 MPa). Several late-successional species (e.g. Memecylon ligustrifolium, Randia canthioides and Xanthophyllum hainanense) had quite high WD (0.86, 0.79 and 0.77 g cm–3, respectively) and low kS (0.55, 0.45 and 0.16 kg m–1 s–1 MPa–1, respectively), but showed relatively higher P50 values (−1.86, −1.46 and −1.50 MPa, respectively). Since kS is proportional to vessel diameter to the fourth power according to the Hagen–Poiseuille equation, the lack in safety efficiency trade-off suggests that P50 might be relatively less dependent on xylem vessel size (Meinzer et al. 2010) and better explained by the inter-vessel pit properties (Choat et al. 2003; Wheeler et al. 2005; Lens et al. 2011). Other hydraulic processes such as stomatal regulation (Salleo et al. 2000; Domec et al. 2006; Maherali et al. 2006), refilling of embolized vessels (Bucci et al. 2003), stem wood storage (Barnard et al. 2011; Pineda-Garcia et al. 2011) and regulation of AL/AS (Martinez-Vilalta et al. 2009) could also contribute to drought tolerance of plants, and lead to the decoupling of the safety efficiency trade-off. Therefore, integrated studies considering variation in hydraulic traits from cell to whole plant levels are needed for a better understanding of hydraulic safety and efficiency (Sperry, Meinzer & McCulloh 2008; Meinzer et al. 2010; McCulloh et al. 2012).

Significant correlations between leaf-specific hydraulic conductivity and plant photosynthetic traits (Aarea, gs, PNUE, PPUE and WUEi) were found in the present study (Fig. 4), which was consistent with a number of other studies in different forest types (Brodribb & Feild 2000; Brodribb et al. 2002; Santiago et al. 2004; Zhang & Cao 2009). Similar to the results found by Hao et al. (2011), phylogenetic correlations between kL and photosynthetic traits were significant. PCA further strengthened the notion that plant hydraulic conductivity and associated photosynthetic traits (PC axis 1), rather than SLA and nutrient contents (PC axis 2), might be key factors underlying species replacement along subtropical forest succession (Fig. 6). Axis 1 also revealed several trade-offs widely found in other studies, for example, the trade-off between photosynthetic nutrient and water-use efficiency (McDowell 2002; Santiago et al. 2004), and the trade-off between water transport and leaf water-use efficiency (Sobrado 2003; Zhang & Cao 2009). Along PC axis 1, early-successional species exhibited typical ‘fast-growing’ characteristics such as higher kS, kL, Aarea, gs, PNUE and PPUE, but used water more ‘luxuriously’ (lower WUEi). By contrast, late-successional species had lower water transport and photosynthetic nutrient use efficiency, but used water more ‘conservatively’.

CONCLUSION

Our results highlighted that hydraulic-photosynthetic coordination was the key factor explaining plant distributions along subtropical forest succession. Compared to plants in late-successional forest, early-successional species had higher hydraulic conductivity as well as photosynthetic capacity. Although in the research site, plants may experience seasonal water deficit for 5 months (from October to March) and water availability is variable across different successional communities, changes in drought tolerance traits measured in the present study did not show a clear pattern, suggesting a limited role of these traits in determining species distributions along the succession. Given the profound impact of seasonal drought on the growth and development of plants, possible hydraulic adjustments to seasonal drought stresses taken by plants at different successional stages need to be further investigated.

ACKNOWLEDGMENTS

The authors are grateful to Edward Morrison for reading and discussing the manuscript. We thank Li-Dong Wang, Ding-Shen Mo, Juan Song and Hui Liu for their assistance in fieldwork and data analysis. DFERS provided the data relating to soil chemistry and plant community investigations. Acknowledgments are extended to the anonymous reviewers who made helpful suggestions and comments on the manuscript. This work was funded by the National Natural Science Foundation of China (grants no. 31100293 and 31070231), and the Chinese Academy of Sciences through its Hundred Talent Program and Knowledge Innovation Project (KSCX2-EW-J-28).

Ancillary