SEARCH

SEARCH BY CITATION

Keywords:

  • aridity;
  • biodiversity;
  • biotic homoeostasis;
  • carbon isotopes;
  • functional traits;
  • leaf nitrogen;
  • North East China Transect;
  • scaling relationships

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • The leaf carbon isotope ratio (δ13C) of C3 plants is inversely related to the drawdown of CO2 concentration during photosynthesis, which increases towards drier environments. We aimed to discriminate between the hypothesis of universal scaling, which predicts between-species responses of δ13C to aridity similar to within-species responses, and biotic homoeostasis, which predicts offsets in the δ13C of species occupying adjacent ranges.
  • The Northeast China Transect spans 130–900 mm annual precipitation within a narrow latitude and temperature range. Leaves of 171 species were sampled at 33 sites along the transect (18 at ≥ 5 sites) for dry matter, carbon (C) and nitrogen (N) content, specific leaf area (SLA) and δ13C.
  • The δ13C of species generally followed a common relationship with the climatic moisture index (MI). Offsets between adjacent species were not observed. Trees and forbs diverged slightly at high MI. In C3 plants, δ13C predicted N per unit leaf area (Narea) better than MI. The δ13C of C4 plants was invariant with MI. SLA declined and Narea increased towards low MI in both C3 and C4 plants.
  • The data are consistent with optimal stomatal regulation with respect to atmospheric dryness. They provide evidence for universal scaling of CO2 drawdown with aridity in C3 plants.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Stable carbon isotope ratios (δ13C values) measured on the leaves of C3 plants are linearly and inversely related to the time-averaged drawdown of CO2 concentration during photosynthesis (Farquhar et al., 1982). Many studies of foliar δ13C in C3 plants have reported a trend towards higher (less negative) values at drier sites (e.g. Stewart et al., 1995; Miller et al., 2001; Midgley et al., 2004; Liu et al., 2005; Wittmer et al., 2008). This relationship also applies globally, and dominates relationships of δ13C with all other geographic and climatic variables (Diefendorf et al., 2010). One study showed a similar scaling of the δ13C of respired CO2 with moisture availability both within and between communities dominated by four tree species (Bowling et al., 2002). However, a few studies of leaf δ13C along moisture gradients have yielded results that are harder to interpret (e.g. Schulze et al., 1996, 2006). Schulze et al. (1998, 2006) proposed the hypothesis that within-species responses of δ13C are steeper than the overall community-level response, implying a homeostatic role for beta diversity in maintaining photosynthetic rates. No published evidence unambiguously supports this hypothesis. However, its opposite, that is, universal scaling (with similar responses shown by species and communities), is not strongly supported either, because very few studies have sampled individual species repeatedly along aridity gradients.

Experiments have shown that stomatal conductance (gs) responds to transpiration rate (Mott & Parkhurst, 1991; Pieruschka et al., 2010). This is indistinguishable in the field from a response to vapour pressure deficit (D). D is an atmospheric property shared by the plants in a community, controlled in part by transpiration itself (Monteith, 1995). Conditions where transpiration is restricted because of low soil water availability typically co-occur with high values of D. The ratio of internal to ambient CO2 concentration (ci : ca) is related to gs through the diffusion equation, A = gsca(1 − ci : ca), where gs is the stomatal conductance to CO2 and A is the net assimilation rate. The optimality hypothesis of Cowan & Farquhar (1977) predicts a dependence of the form ci : ca = 1/(1 + ξ√D) or, with a further approximation, ci : ca = 1 − ξ√D (the two equations are close within the commonly encountered ranges of ci : ca and D), where ξ is related to Cowan and Farquhar’s λ parameter which represents an ‘exchange rate’ between carbon and water. This prediction can readily be derived for an assumed linear relationship between A and ci, but it has recently been shown to remain a good approximation under more realistic assumptions (Katul et al., 2010; B. Medlyn, pers. comm.). It is theoretically expected that ξ will increase with soil moisture deficit, leading to a steeper relation of ci : ca to D along soil moisture gradients than would be predicted for constant ξ. This expectation is supported by direct measurements of the gsD relationship in different species and environments (e.g. Palmroth et al., 1999; B. Medlyn, pers. comm.). In principle, ξ could also vary among species and plant functional types. However, very little empirical information is available to indicate whether such differences exist, or whether adjustment in ξ along aridity gradients occurs mainly within species, occurs mainly between species, or is universal (i.e. following the same relationship within and between species).

Although D is highly variable in time, we would expect that δ13C measurements would relate to time-averaged values of D and, most importantly, that these values of D would be similar for all plants at a site (except in dense vegetation where subcanopy species experience reduced D). Thus, leaf δ13C measured on all species in a community should provide evidence for or against a universal scaling of ci : ca with D. Differences among species or plant functional types might be attributed to differences in ξ related, for example, to differences in conducting tissue properties or rooting strategy.

We set out to test the hypothesis of universal scaling by analysing leaf samples of the most common vascular plant species at 33 sites distributed along the Northeast China Transect (NECT; Fig. 1). The NECT runs east–west (so there is no confounding with latitude) with little variation in mean annual temperature (0–6°C), but a major variation in annual precipitation from 130 to 900 mm. There is a steady trend of decreasing stature, density and foliage projective cover towards the dry end of the transect, with trees largely confined to the wet end (Ni & Zhang, 2000; Ni & Wang, 2004; Wang & Ni, 2005). Thus, the NECT has ideal properties for an aridity gradient study. In order to examine other correlates of leaf δ13C we measured leaf dry matter, carbon (C) and nitrogen (N) contents and specific leaf area (SLA), allowing us to evaluate ancillary findings in the literature that relate to the potential trade-off between leaf N and stomatal closure. Sampling density was sufficient that many species were sampled at more than one site. Altogether 18 species, distributed fairly evenly along the gradient, were sampled at five or more sites, allowing us to analyse the within-species patterns and to compare these with the multi-species trends.

image

Figure 1.  Location of the Northeast China Transect (NECT; rectangle) in China. Sites are shown as closed black circles. Gradients in the climatic moisture index (MI) are indicated by colour fields.

Download figure to PowerPoint

Materials and Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Environmental data

Mean monthly values of temperature, precipitation and percentage of possible sunshine hours were derived from 1814 meteorological stations across China (740 stations have observations from 1971–2000, and the rest from 1981–1990: China Meteorological Administration, unpublished data). The climate data were interpolated at 1-km resolution using a smoothing spline interpolation (anusplin version 4.36; Hutchinson & Hancock, 2006) and the STRM 1-km digital elevation model (Farr et al., 2007). Mean annual temperature (MAT), mean annual precipitation (MAP), and mean precipitation in June, July and August (PJJA) were calculated for each site. Bioclimatic variables more closely related to the physiological controls on plant growth were then derived as in Prentice et al. (1993) and Gallego-Sala et al. (2010): the mean temperature of the coldest month (MTCO), the daily mean during the growing season (the period with daily-interpolated temperatures > 0°C) of accumulated growing degree days above 0°C (mGDD0, equivalent to growing-season mean temperature), the daily mean during the growing season of photosynthetically active radiation (mPAR0), annual equilibrium evapotranspiration (EET), the climatic moisture index (MI = MAP/EET), annual actual evapotranspiration (AET) and the alpha index (alpha = AET/EET). The last two variables require data on available water holding capacity (AWHC). Texture data (sand, silt and clay fractions) were obtained for soil types, digitized from the 1 : 1 million soil map of China (Shi et al., 2004). AWHC was then calculated as the difference between field capacity and wilting point, estimated as the soil water content at matric potentials of −0.033 and −1.5 MPa, respectively, using the equations of Saxton & Rawls (2006) with a correction (based on linear interpolation of the cumulative log size–frequency plot) to convert from International Society for Soil Sciences (ISSS) to US Department of Agriculture (USDA) definitions of the sand–silt boundary.

Sampling strategy

The sampling sites (Table 1) were selected to represent the aridity gradient along the transect. All sampling took place during August 2006. All sites were occupied by visually homogeneous uncultivated vegetation. Most of the grasslands are grazed, and it is virtually impossible to find undisturbed sites; we chose sites with minimal signs of recent disturbance, and avoided sites with evidence of exceptionally high grazing pressure.

Table 1.   Characteristics of the sites from the Northeast China Transect (NECT) used in this analysis
Site no.Longitude (degree)Latitude (degree)Elevation (m)Vegetation typesMax height (m)No. of species sampledMoisture index (MI) at the site
NECT01118.4842.881024Steppe0.3670.50
NECT02119.0243.64781Steppe0.58130.50
NECT03129.7843.02136Mixed deciduous broad-leaved forest10.5170.77
NECT04130.0842.98114Mixed deciduous broad-leaved forest7.5150.78
NECT05131.1543.30289Mixed conifer–deciduous broad-leaved forest17.5250.79
NECT06131.0043.12244Mixed deciduous broad-leaved forest26.0200.78
NECT07129.6743.39224Mixed conifer–deciduous broad-leaved forest19.0150.82
NECT08128.6443.25601Conifer forest23.5130.96
NECT09127.0343.73390Mixed deciduous broad–leaved forest23.0190.95
NECT10125.6843.81252Mixed conifer–deciduous broad–leaved forest6.0140.79
NECT11123.5144.59146Meadow steppe0.75110.55
NECT12123.2744.43150Meadow steppe0.2070.51
NECT13121.8443.60203Meadow steppe0.53100.44
NECT14121.7744.12202Meadow steppe0.3650.43
NECT15120.5544.39448Steppe0.90110.51
NECT16120.3744.22372Steppe0.53120.45
NECT17119.3843.88601Steppe0.57120.48
NECT18119.1243.76729Steppe0.76120.47
NECT19118.4943.34707Steppe0.4350.47
NECT20117.7643.19889Steppe0.75110.48
NECT21117.2443.221259Steppe1.30130.57
NECT22116.8943.391267Steppe0.4270.55
NECT23116.6843.551261Steppe0.70130.51
NECT24116.6443.691211Steppe0.58100.49
NECT25116.3143.911199Steppe0.90120.46
NECT26115.3243.901196Steppe0.65130.38
NECT27114.6143.941123Desert steppe0.6090.30
NECT28113.8343.831166Desert steppe0.2380.27
NECT29113.3643.801017Desert steppe0.2860.22
NECT30112.5943.72974Desert steppe0.3090.18
NECT31112.1743.63999Desert steppe0.2480.18
NECT32111.9243.661005Desert steppe0.2080.17
NECT33111.8943.651017Desert steppe0.2670.17

Species composition and vegetation structure were surveyed at each site. In each of the eight forested sites at the eastern (wet) end of the transect, we surveyed one 10 × 10 m2 quadrat for the tree layer, five for the shrub layer (5 × 5 m2 each) and five for the herb layer (2 × 2 m2 each). In the dry steppe and desert steppe vegetation in the middle to the west of the transect, we investigated five quadrats in each of the 25 sites. The quadrat size was 1 × 1 m2 for grasses and 4 × 4 m2 for shrubs. We created a checklist of vascular species present at each site, and sampled the common species.

Foliage sampling and analysis

Sunlit leaves were obtained with long-handled twig shears. At least 10 g of leaves was collected for each species, except for a small number of species with very small leaves at the dry end of the transect. The samples were subdivided for the measurement of SLA and leaf dry matter content (LDMC) (sample sizes for replicate measurements varied between 0.02 and 2.3 g), C content, N content and δ13C. The measurements used here are the average of three replicates, except in the case of the δ13C where one measurement was made per individual. Leaves were scanned with a laser scanner and the leaf areas were measured using Photoshop on the scanned images. Leaf weight was measured in the field; dry weight was obtained after air-drying for several days and then oven-drying at 75°C for 48 h. Leaf C content was measured by the potassium dichromate volumetry method. N content was measured by the micro-Kjeldahl method. δ13C was measured using a Finnigan MAT DELTAplusXP Isotope Ratio Mass Spectrometer (Finnigan Corporation, San Jose, CA).

Statistical analysis

Principal components analysis (PCA) of climate variables was performed using the spss statistical package (SPSS, Chicago, IL, USA). Redundancy analysis (RDA; ter Braak & Prentice, 1988) of traits (as predictands) with respect to climate variables (as predictors) was performed using the program canoco (ter Braak & Smilauer, 2002). The traits entered in the analysis were loge LDMC, loge C, loge N, loge SLA and δ13C. The use of loge-transformed variables allows the directions of various ratios to be inferred from a vector plot (for example, as Narea = N/SLA, the vector for loge Narea is the resultant of the vector for loge N and the vector pointing in the opposite direction to the one for loge SLA). We used ordinary least-squares regression throughout. Tests for homogeneity of regressions were performed with the smatr package (Falster et al., 2006).

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Variation in climate variables and leaf traits

PCA of climate variables across sampling sites confirmed that there was a single dominant climatic gradient (Table 2) accounting for 68% of the variance. From the wet to the dry end, this gradient was characterized by decreasing MAP, MI, PJJA and alpha, and increasing mPAR0 (attributable to an increasing fraction of hours without clouds). However, a second axis with a nontrivial eigenvalue, accounting for 24% of the variance, was also present, indicating the existence of temperature (MAT and mGGD0) variations independent of the main moisture gradient.

Table 2.   Principal components analysis (PCA) of climate variables, showing strong correlations between all of the moisture-related indices (axis 1), and similarly strong correlations between the temperature-related indices (axis 2)
 Component 1Component 2
  1. MAP, mean annual precipitation; MAT, mean annual temperature; mGDD0, daily mean during the growing season of accumulated growing degree days above 0°C; MI, moisture index; mPAR0, the daily mean during the growing season of photosynthetically active radiation; PJJA, mean precipitation in June, July and August.

Component matrix
 MAT0.2900.921
 MAP0.9850.051
 PJJA0.9470.081
 mGDD0−0.3360.906
 MI0.993−0.067
 mPAR0−0.861−0.078
 Alpha0.993−0.092
Eigenvalues
 % of variance68.224.2
 Cumulative %92.5

RDA of traits of C3 plants vs the same set of climate variables again showed a single dominant axis reflecting the moisture gradient (Table 3, Fig. 2). The C content of dry weight was almost invariant. Along the gradient from wet to dry, δ13C, LDMC and Narea increased while SLA decreased. However, in contrast to the PCA of climate variables alone, the eigenvalue of the second axis was vanishingly small and contributed nothing to the information explained by the climate variables – in other words, although there was some independent variation of climate along a temperature axis, there was no significant associated variation in leaf traits. The sampled variation along the gradient can thus be taken to reflect control of leaf traits by aridity (with a possible contribution from the associated variation in photosynthetically active radiation), with minimal confounding by temperature.

Table 3.   Redundancy analysis of C3 species traits in relation to climate variables
 Axis
1234
Eigenvalue0.7010.0030.0010
Trait–environment correlation0.8720.3460.1760.133
Cumulative % variance of trait data70.170.470.570.5
Cumulative % variance of trait–environment relations99.499.899.9100
image

Figure 2.  Redundancy analysis plot showing vectors (representing correlations with the first two axes) for traits and climate variables: δ13C, carbon isotope ratio; loge LDMC, loge leaf dry matter content; loge C, see Q21; loge N, see Q21; PJJA, mean precipitation in June, July and August; alpha; MI, moisture index; MAP, mean annual precipitation; loge SLA, loge specific leaf area; MAT, mean annual temperature; mGDD0, daily mean during the growing season of accumulated growing degree days above 0°C; mPAR0, the daily mean during the growing season of photosynthetically active radiation.

Download figure to PowerPoint

Variation in δ13C along the moisture gradient

C3 plants showed a systematic trend of increasing δ13C with dryness (Fig. 3), regardless of whether moisture availability was plotted as MAP, MI or alpha. By a small margin, the best R2 (0.712) was provided by MI, which was used in all subsequent analyses. In a multiple regression of C3 plant δ13C with MI, mGDD0 and mPAR0 as predictors (Supporting Information Table S1) only MI had a significant regression coefficient, indicating that temperature and light made no significant additional contributions to explaining the variation in δ13C.

image

Figure 3.  Relationships between the carbon isotope ratio (δ13C) and three measures of water availability: mean annual precipitation (MAP), moisture index (MI), and alpha. Species are grouped by photosynthetic pathway (C3 species, closed circles and solid line; C4 species, open circles and dashed line). ns, not significant. Asterisks indicate the significance levels.

Download figure to PowerPoint

We used linear regressions throughout for simplicity of statistical analysis. However, all three plots showed a tendency for δ13C to flatten off at the dry end of the transect. C4 plants showed characteristically higher values of δ13C, and a slight (nonsignificant) trend of decreasing δ13C with dryness.

Responses of δ13C for different life forms

The responses of δ13C to MI for different C3 life forms (Fig. 4) showed individually significant relationships in all cases but one (lianas/climbers). (None of the C4 life forms had a significant slope; see Table S2.) There were also small but significant differences (Table S2) between the fitted linear regressions for individual life forms and the regression for all C3 plants (Figs 1, 4). Woody C3 plants collectively, and trees, had shallower slopes. Nonwoody C3 plants collectively, and perennial forbs, had steeper slopes, but perennial grasses had a shallower slope. A comparison of average δ13C values for each life form within three MI ‘bins’ corresponding to the dry, middle and wet parts of the transect (Table 4) suggested that these differences arose from slight deviations from linearity in the response of different life forms to MI. At the dry end of the transect, there were only very small differences in δ13C between life forms. In the middle part of the transect, perennial grasses showed significantly higher δ13C values than shrubs and forbs. At the wet end of the transect, there were systematic and significant differences between life forms, with highest δ13C values shown by trees, intermediate values by shrubs, and low values by forbs. Subcanopy and especially field-layer species in forests experience lower D than canopy trees, so this divergence is consistent with a common response to atmospheric dryness.

image

Figure 4.  Relationship between carbon isotope ratio (δ13C) and moisture index (MI) for woody and nonwoody C3 species, and for C3 life forms (linear fit, solid line). The linear fit for all C3 species is shown as a dashed line on all of the panels. ns, not significant. Asterisks indicate the significance levels.

Download figure to PowerPoint

Table 4.   Multiple comparisons of carbon isotope ratio (δ13C) values for C3 plant life forms grouped into three bins (moisture index (MI) < 0.35, 0.35–0.7 and > 0.7)
GroupMultiple comparisons when MI ≤ 0.35Multiple comparisons when 0.35 < MI ≤ 0.7Multiple comparisons when MI > 0.7
  1. Values are mean ± standard deviation.

  2. Shaded cells indicate that no comparison is possible because the life form is not represented in this part of the transect. Groups sharing a letter do not differ significantly.

C3 tree  −29.47 ± 1.34f
C3 shrub−24.98 ± 0.87a−27.09 ± 1.02c−30.47 ± 1.40gi
C3 perennial forb−24.80 ± 0.46ab−26.91 ± 1.41ce−31.36 ± 1.27h
C3 annual forb −27.23 ± 1.35ce−31.63 ± 3.63ghi
C3 perennial grasses−24.39 ± 0.87ab−25.75 ± 0.84d 
C3 sedges/rushes  −30.23 ± 1.61fgi
C3 lianas/climbers  −30.76 ± 1.29gh
C3 geophytes−24.02 ± 0.74b−26.23 ± 1.17de 

Responses of other leaf traits

There was a weak tendency for LDMC of C3 species to increase with dryness, as has been observed previously (Niinemets, 2001; Lavorel et al., 2007), and a very weak tendency towards narrower C : N ratios (more N per unit C), although this relationship was steeper for C4 species (Fig. 5). Similarly weak relationships were found for these traits with δ13C in C3 plants. By contrast, the data showed a strong increase in Narea and decrease in SLA with dryness among both C3 plants and C4 plants, and even stronger relationships of Narea and SLA to δ13C in C3 plants (Fig. 5). These relationships are highly conserved among life forms and even between C4 and C3 plants (Figs 6, 7). The putative relationship of SLA and Narea to ci : ca cannot be tested for C4 plants, however, because of the more complex interpretation of δ13C in C4 plants (Farquhar, 1983). In multiple regressions of loge Narea with MI and δ13C as predictors (Table S1), C3 plants showed a significant effect of δ13C on Narea, additional to that of MI.

image

Figure 5.  Relationships between leaf dry matter content (LDMC), loge (C : N), loge specific leaf area (SLA) and loge N per unit leaf area (Narea) and carbon isotope ratio (δ13C) for C3 species (a-d) and C4 species (e-h), and moisture index (MI) for C3 species (i-l) and C4 species (m-p). C3 species, closed circles and solid line; C4 species, open circles and dashed line. ns, not significant. Asterisks indicate the significance levels.

Download figure to PowerPoint

image

Figure 6.  Relationship between (loge Narea) and moisture index (MI) for C3 and C4 woody and nonwoody species, and for C3 life forms. ns, not significant. Asterisks indicate the significance levels.

Download figure to PowerPoint

image

Figure 7.  Relationship between loge N per unit leaf area (Narea) and carbon isotope ratio (δ13C) for C3 woody and nonwoody species, and for C3 life forms (linear fit, solid line). The linear fit for all C3 species is shown as a dashed line on all of the panels. ns, not significant. Asterisks indicate the significance levels.

Download figure to PowerPoint

Responses of δ13C for different species

Fig. 8 illustrates the responses of δ13C to MI found within species of C3 plants, for the 15 C3 species that were sampled at five or more sites. The individual species’ lines generally did not depart from the overall trend. There was no systematic pattern of offsets and, in particular, there was no indication that species of drier environments compensate by having lower δ13C at a given MI, as hypothesized by Schulze et al. (1998, 2006). Instead, species tended to be replaced (in the direction of increasing dryness) by species that had higher δ13C, but the same δ13C at a given MI. The only species with an individually significant regression that differed significantly from the general line was the geophyte of dry environments Allium racemosum, whose slope against MI was shallower than the norm, consistent with the general flattening of the multi-species δ13C–MI relationship at low MI. Species at the wet end of the transect also showed a (nonsignificant) tendency to diverge in their δ13C–MI relationship; for example, the tree species Quercus mongolica showed higher δ13C values than the shrub species Lespedeza bicolor. We also examined the responses of δ13C to MI for three C4 species (Cleistogenes squarrosa, Pennisetum flaccidum and Salsola collina): none showed significant slopes (Table S2).

image

Figure 8.  Relationship between carbon isotope ratio (δ13C) and moisture index (MI) for all C3 species that were sampled at five or more sites. AD, Asparagus dauricus; AF, Artemisia frigida; AM, Allium mongolicum; AR, Allium ramosum; CH, Corylus heterophylla; CM, Caragana microphylla; CS, Caragana stenophylla; LB, Lespedeza bicolour; LD, Lespedeza davurica; LC, Leymus chinensis; MR, Melitoides ruthenica; PC, Potentilla chinensis; QM, Quercus mongolica; SK, Stipa krylovii; UD, Ulmus davidiana var. japonica.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

These findings, obtained under optimum field sampling conditions on a long precipitation gradient (130–900 mm) with minimal confounding by temperature or latitude, provide evidence for close to universal scaling (i.e. the same response both within and between C3 species and life forms) of δ13C. C3 species that successively replaced one another as the environment became drier fell along a single overall trend of δ13C in relation to MI. They did not have lower δ13C at the same MI, as one would expect according to the biotic homoeostasis hypothesis proposed by Schulze et al. (1998, 2006). Thus, in common with Wittmer et al. (2008), our results show that species- and community-level responses to changing aridity are strikingly similar. Given the lack of any compelling evidence presented to date in favour of the biotic homoeostasis hypothesis, and its clear rejection by our results, we suggest that it should be abandoned.

The results have valuable implications for large-scale modelling. They support the simplifying assumption that different C3 species and life forms react through stomatal adjustment in a similar way in response to the aridity gradient. This is consistent with the theoretical expectation that the plants are responding to a common atmospheric property (D). However, the magnitude of the variation in δ13C along the gradient was too large to be plausibly accounted for by the response of ci : ca to D with constant ξ (B. Medlyn, pers. comm.). The results therefore imply a systematic adjustment of the ξ parameter along the gradient. They further indicate that (within the sampled range of MAP) this adjustment is universal, following the same pattern within as between species. Differences among life forms were slight, and could reasonably be neglected for modelling purposes.

The fact that the relationships of Narea and SLA in C3 plants were stronger with δ13C than MI suggests that there is a systematic relationship that applies to a part of the variation within sites at the same MI – in other words, that there is a mechanistic link among the leaf traits such that low ci : ca ratios tend to be accompanied by low SLA and high Narea, and that this may account at least in part for the observed relationships of SLA and Narea to MI. The observed relationships to δ13C in C3 plants are qualitatively consistent with the hypothesis (e.g. Wright et al., 2003) that high photosynthetic capacity coupled with high CO2 drawdown represents an adaptation to drought. Low ci : ca ratios, adaptive under water-stressed conditions, imply that leaves require a high Narea in order to achieve the high carboxylation capacity required to effectively utilize the available light when ci is low. This mechanism may be further strengthened because PAR increases with decreasing cloudiness, so more light is available to be utilized towards the dry end of the gradient. Low SLA can be viewed as a requirement for leaves to be capable of high potential carboxylation rates.

However, the magnitude of the increase in Narea (by a factor of 4.5) along the gradient was considerably larger than would be indicated by the hypothesis that carboxylation capacity at the leaf level adapts to the environment in such a way as to achieve maximal net photosynthesis, that is, around the point of co-limitation by Rubisco and PAR (Farquhar et al., 1982; Haxeltine & Prentice, 1996). This hypothesis would predict an increase by a factor of only 1.6 in the carboxylation capacity of C3 plant leaves as a result of the combination of increasing mPAR0 (by a factor of 1.25 along the transect) and decreasing ci : ca. Thus, the magnitude of the changes in Narea and SLA along the transect requires an additional functional explanation. Furthermore, any comprehensive explanation for these observations must apply to C4 as well as C3 plants: the response of Narea and SLA to MI in C4 plants contrasts sharply with their lack of response in δ13C. The low SLA and high Narea of plants towards the dry end of the gradient may be functionally related to the greater resistance to dehydration conferred by stiffer leaves (Niinemets, 2001; Harrison et al., 2010), or to the requirement for leaves to be small or narrow to avoid overheating under conditions of high radiation load and low transpiration rates (Midgley et al., 2004; Harrison et al., 2010).

The plant species sampled were almost all angiosperms. There is some evidence that the ci : ca ratio (Lloyd & Farquhar, 1994) and δ13C (Marshall & Waring, 1984) of gymnosperms react more steeply to increasing D, although one study of the δ13C of respired CO2 from temperate angiosperm- and gymnosperm-dominated forests in the same climate indicated no difference in the response to D (Mortazavi et al., 2005). The possible difference could be tested by sampling along a comparable moisture gradient in a temperate winter rainfall regime where conifers dominate. The gymnosperms sampled on the NECT (Pinus koraensis, Pinus tabulaeformis and Larix olgensis) did not have atypical δ13C values for their position along the transect. However, no gymnosperm species were encountered in the middle or dry sections of the transect.

Our sampling stopped at a MAP of 130 mm, so we did not sample vegetation in hyper-arid environments. There is an indication in the overall data (Fig. 3) and in the data for one species (Allium racemosum; Fig. 8) that the general response of leaf δ13C to MI levelled off at MI values below 0.3. The values of δ13C reached at this low level of MI may be approaching minimum sustainable values of ci : ca. Taking δ13C in atmospheric CO2 as −8.3‰ (Mauna Loa, July 2006: http://scrippsco2.ucsd.edu/data/atmospheric_co2.html) and applying standard approximations Δ ≈ δair −δleaf ≈+ (− a) ci : ca, = 4.4 and = 27, a leaf δ13C of −24‰ would correspond to a ci : ca ratio of 0.5. Lower average ci : ca ratios may not allow sufficient carbon assimilation for plant survival. Vegetation in the driest environments may therefore depend strictly on access to groundwater to allow increased transpiration rates. This could be an explanation for the extremely variable results obtained by Schulze et al. (2006) for Eucalyptus spp. at low MAP. Dependence of vegetation on aquifers was also invoked by Schulze et al. (1996) to explain the atypical constancy of δ13C values of C3 foliage along a precipitation gradient eastward from the Andes mountains. A different modelling approach may be required to describe vegetation subsisting under hyperarid conditions, and perhaps also in situations where groundwater is the dominant water source. The indication in our data that the δ13C values of C3 plant leaves at the dry end of the NECT are approaching a maximum could be tested by a westward extension of the transect.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank Tingting Yao, Shengjun Ji, Juan Wang, Xun Tian, Honsou Eshara and Lucy Harrison-Prentice for field assistance; Angela Gallego-Sala for providing the program to estimate bioclimate variables; the Chinese Academy of Sciences for funding; and Remko Duursma, Josh Fisher, Belinda Medlyn, Ian Wright and Xu Liang for discussions. The fieldwork was carried out while I.C.P. and S.P.H. were Guest Professors at the Chinese Academy of Sciences, Institute of Botany, Beijing. This work was supported by a key project of the National Natural Science Foundation of China (grant no 30590383) and the Chinese Academy of Sciences.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • Bowling D, McDowell N, Bond B, Law B, Ehleringer J. 2002. 13C content of ecosystem respiration is linked to precipitation and vapor pressure deficit. Oecologia 131: 113124.
  • ter Braak CJF, Prentice IC. 1988. A theory of gradient analysis. Advances in Ecological Research 18: 271317.
  • ter Braak CJF, Smilauer P. 2002. CANOCO reference manual and user’s guide to Canoco for Windows: software for canonical community ordination (version 4.53). Ithaca, NY, USA: Microcomputer power.
  • Cowan IR, Farquhar GD. 1977. Stomatal function in relation to leaf metabolism and environment. Society Experimental Biology Symposium 31: 471505.
  • Diefendorf AF, Mueller KE, Wing SL, Koch PL, Freeman KH. 2010. Global patterns in leaf 13C discrimination and implications for studies of past and future climate. Proceedings of the National Academy of Sciences, USA 107: 57385743.
  • Falster DS, Warton DI, Wright IJ. 2006. SMATR: Standardised major axis tests and routines, ver 2.0 [WWW document]. URL http://www.bio.mq.edu.au/ecology/SMATR/. Last accessed: 1 August 2010.
  • Farquhar GD. 1983. On the nature of carbon isotope discrimination in C4 species. Australian Journal of Plant Physiology 10: 205226.
  • Farquhar GD, O’Leary MH, Berry JA. 1982. On the relationship between carbon isotope discrimination and the intercellular carbon dioxide concentration in leaves. Australian Journal of Plant Physiology 9: 121137.
  • Farr TG, Rosen PA, Caro E, Crippen R, Duren R, Hensley S, Kobrick M, Paller M, Rodriguez E, Roth L et al. 2007. The shuttle radar topography mission. Reviews of Geophysics 45: RG2004. doi:10.1029/2005RG000183.
  • Gallego-Sala AV, Clark JM, House JI, Orr HG, Prentice IC, Smith P, Farewell T, Chapman SJ. 2010. Bioclimatic envelope model of climate change impacts on blanket peatland distribution in Great Britain. Climate Research, doi:10.3354/cr00911.
  • Harrison SP, Prentice IC, Barboni D, Kohfeld KE, Ni J, Sutra J-P. 2010. Ecophysiological and bioclimatic foundations for a global plant functional classification. Journal of Vegetation Science 21: 300317.
  • Haxeltine A, Prentice IC. 1996. A general model for the light use efficiency of primary production. Functional Ecology 10: 551561.
  • Hutchinson M, Hancock PA. 2006. Spatial interpolation of large climate data sets using bivariate thin plate smoothing splines. Environmental Modelling and Software 21: 16841694.
  • Katul G, Manzoni S, Palmroth S, Oren R. 2010. A stomatal optimization theory to describe the effects of atmospheric CO2 on leaf photosynthesis and transpiration. Annals of Botany 105: 431442.
  • Lavorel S, Díaz S, Cornelissen JHC, Garnier E, Harrison SP, McIntyre S, Pausas JG, Perez-Harguindeguy N, Roumet C, Urcelay C et al. 2007. Plant functional types: are we getting any closer to the Holy Grail? In: CanadellJG, PatakiD, PitelkaL, eds. Terrestrial ecosystems in a changing world. The IGBP Series. Berlin, Heidelberg, Germany: Springer-Verlag, 149160.
  • Liu W, Xiahong F, Youfeng N, Qingle Z, Yunning C, Zhisheng AN. 2005. δ13C variation of C3 and C4 plants across an Asian monsoon rainfall gradient in arid northwestern China. Global Change Biology 11: 10941100.
  • Lloyd JJ, Farquhar GD. 1994. 13C discrimination during CO2 assimilation by the terrestrial biosphere. Oecologia 99: 201215.
  • Marshall JD, Waring RH. 1984. Conifers and broadleaf species: stomatal sensitivity differs in western Oregon. Canadian Journal of Forest Research 14: 905908.
  • Midgley GF, Aranibar JN, Mantlana KB, Macko S. 2004. Photosynthetic and gas exchange characteristics of dominant woody plants on a moisture gradient in an African savanna. Global Change Biology 10: 309317.
  • Miller JM, Williams RJ, Farquhar GD. 2001. Carbon isotope discrimination by a sequence of Eucalyptus species along a subcontinental rainfall gradient in Australia. Functional Ecology 15: 222232.
  • Monteith JL. 1995. Accomodation by transpiring vegetation and the convective boundary layer. Journal of Hydrology 166: 251263.
  • Mortazavi B, Chanton JP, Prater JL, Oishi C, Oren R, Katul GG. 2005. Temporal variability in 13C of respired CO2 in a pine and a hardwood forest subject to similar climatic conditions. Oecologia 142: 5769.
  • Mott KA, Parkhurst DF. 1991. Stomatal responses to humidity in air and helox. Plant, Cell & Environment 14: 509515.
  • Ni J, Wang GH. 2004. Northeast China Transect (NECT): ten-year synthesis and future challenges. Acta Botanica Sinica 46: 379391.
  • Ni J, Zhang X-S. 2000. Climate variability, ecological gradient and the Northeast China Transect (NECT). Journal of Arid Environments 46: 313325.
  • Niinemets N. 2001. Global-scale climatic controls of leaf dry mass per area, density, and thickness in trees and shrubs. Ecology 82: 453469.
  • Palmroth S, Berninger F, Nikinmaa E, Lloyd J, Pulkkinen P, Hari P. 1999. Structural adaptation rather than water conservation was observed in Scots pine over a range of wet to dry climates. Oecologia 121: 302309.
  • Pieruschka R, Huber G, Berry JA. 2010. Control of transpiration by radiation. Proceedings of the National Academy of Sciences, USA 107: 1337213377.
  • Prentice IC, Sykes MT, Cramer W. 1993. A simulation model for the transient effects of climate change on forest landscapes. Ecological Modelling 65: 5170.
  • Saxton KE, Rawls WJ. 2006. Soil water characteristic estimates by texture and organic matter for hydrologic solutions. Journal of the Soil Science Society of America 70: 15691578.
  • Schulze E-D, Mooney HA, Sala OE, Jobbagy E, Buchmann N, Bauer G, Canadell J, Jackson RB, Loreti J, Oesterheld M. 1996. Rooting depth, water availability, and vegetation cover along an aridity gradient in Patagonia. Oecologia 108: 503511.
  • Schulze E-D, Turner NC, Nicolle D, Schumacher J. 2006. Leaf and wood carbon isotope ratios, specific leaf areas and wood growth of Eucalyptus species across a rainfall gradient in Australia. Tree Physiology 26: 479492.
  • Schulze ED, Williams RJ, Farquhar GD, Schulze W, Langridge J, Miller JM, Walker BH. 1998. Carbon and nitrogen isotope discrimination and nitrogen nutrition of trees along a rainfall gradient in northern Australia. Australian Journal of Plant Physiology 25: 413425.
  • Shi XZ, Yu DS, Warner ED, Pan XZ, Petersen GW, Gong ZG, Weindorf DC. 2004. Soil database of 1:1,000,000 digital soil survey and reference system of the Chinese genetic soil classification system. Soil Survey Horizons 45: 129136.
  • Stewart GR, Turnbull MH, Schmidt S, Erskine PD. 1995. 13C natural abundance in plant communities along a rainfall gradient: a biological integrator of water availability. Australian Journal of Plant Physiology 22: 5155.
  • Wang GH, Ni J. 2005. Responses of plant functional types to an environmental gradient on the Northeast China Transect. Ecological Research 20: 563572.
  • Wittmer MHOM, Auerswald K, Tungalag R, Bai YF, Schäufele R, Schnyder H. 2008. Carbon isotope discrimination of C3 vegetation in Central Asian grassland as related to long-term and short-term precipitation patterns. Biogeosciences 5: 913924.
  • Wright IJ, Reich PB, Westoby M. 2003. Least-cost input mixtures of water and nitrogen for photosynthesis. American Naturalist 161: 98111.

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Table S1 Results from multiple regression analyses

Table S2 Simple regression analyses: relationship, slope, intercept, R2, P-value for slope, and P-value for the change from the slope of all C3 species or all C4 species (as appropriate)

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office.

FilenameFormatSizeDescription
NPH_3579_sm_TableS1-S2.doc172KSupporting info item