SEARCH

SEARCH BY CITATION

Keywords:

  • climate effects;
  • growth;
  • leaf nitrogen;
  • leaf size;
  • mortality;
  • plant development and life-history traits;
  • seed mass;
  • specific leaf area;
  • tree height;
  • wood density

Summary

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

1. Surprisingly little is known about the relationship between functional traits and demographic rates of tree species under field conditions, particularly for non-tropical species.

2. We studied the interspecific relationship between key functional traits (wood density (WD), maximum tree height, specific leaf area, nitrogen (N) content of leaves, leaf size and seed mass), demographic rates (relative growth rate (RGR) and mortality rate (MR)) and climatic niche for the 44 most abundant tree species in Spain.

3. Demographic data were derived from the Spanish Forest Inventory, a repeated-measures scheme including c. 90 000 permanent plots spread over a territory of c. 500 000 km2. Functional traits data came primarily from a more detailed forest inventory carried out in Catalonia, NE Spain.

4. Our study region covers a wide range of climatic conditions and, not surprisingly, the studied species differed markedly in their climatic niche. Despite that fact, our results showed that the variability in demographic rates across species was much more related to differences in functional traits than to differences in the average climate among species.

5. Maximum tree height and, particularly, WD, emerged as key functional traits, and were the best predictors of demographic rates in our study. These two variables also mediated the marginally significant relationship between RGR and MR, suggestive of a weak trade-off between growth and survival.

6. The main aspects of our results were not altered by the explicit incorporation of phylogenetic effects, suggesting that the observed relationships are not due to divergences between a few major clades.

7.Synthesis. Our study gives support to the notion that variation in functional traits across species allows them to perform largely independently of climatic conditions along environmental gradients.


Introduction

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

One of the key goals of ecology is to understand the factors that determine fitness and its components (survival, growth, reproductive effort). Fitness is related both to environmental conditions and to phenotypic attributes that determine the performance of individual species in specific environments (Ackerly 2003). Much effort has been devoted to understanding the effect of the environment on plant performance (e.g.Woodward 1987) and the relationships and trade-offs between key functional traits at the inter-species level (Wright et al. 2004; Chave et al. 2009). However, little is known about the relationships between species-specific functional traits and demographic rates, such as growth or mortality, for plants growing in the field, despite the fact that the very definition of functional trait generally implies an impact in fitness (Violle et al. 2007). This lack of knowledge is particularly patent for trees, as their long life span requires monitoring relatively large populations for long periods of time. Additionally, the few studies looking at the association between functional traits and demography in trees are limited to tropical species (e.g. Sterck, Poorter & Schieving 2006; Poorter et al. 2008). Because of the diverse nature of tropical communities, these studies tend to be limited to a single community or a small number of them, and are thus not well suited to assess the interaction between climate and species-to-species differences in functional traits in their effect on plant performance.

Within a species, plant growth and survival are determined to a large extent by climate (Harper 1977). We know much less, however, about the variation of demographic rates across species and the way they are affected by climate. As species themselves change along climatic gradients and tend to do so predictably according to some functional attributes (e.g.McGill et al. 2006), the variability of these attributes across species is expected to buffer somewhat the effects of climate on plant performance. Out of the vast number of functional traits of potential relevance to plants, some have been identified as particularly good predictors of demographic rates, at least in tropical forests (cf. Poorter et al. 2008), and are the basis of this study.

Maximum plant height, or potential canopy height, is a complex trait with a fundamental role in light interception (King et al. 2005; King, Davies & Nur Sapardi 2006a). Within a site, taller plants intercept more radiation on average and thus potentially realize higher growth rates, a relationship that has been reported for tropical trees (Poorter et al. 2008). However, height also incurs many costs in terms of construction and maintenance of support and transport structures (Westoby et al. 2002; Mencuccini 2003), and increased risks of buckling and dysfunctions in the water transport system (Koch et al. 2004; Sperry, Meinzer & McCulloh 2008). Wood density (WD) is also regarded as a key functional trait (Chave et al. 2009). On the one hand, it represents the plant’s investment in biomass per unit of wood volume and thus plants with denser wood are expected to realize lower growth rates than plants with lighter wood (King et al. 2005, 2006b; Poorter et al. 2008). On the other hand, WD is related to the mechanical properties of wood (Poorter 2008; Chave et al. 2009) and to the hydraulic properties of the xylem (Hacke et al. 2001; Jacobsen et al. 2007; Zanne et al. 2010). Finally, seed size is also an integrative trait showing large variation across species (Westoby et al. 2002). Small-seeded species produce more seeds per unit of energy invested, and thus tend to have higher dispersal success (Dalling & Hubbell 2002; Moles & Westoby 2006). However, the seedlings of large-seeded species are better able to cope with many of the stresses encountered during establishment and usually have higher survival through early seedling establishment (Moles & Westoby 2004, 2006). In the recent study by Poorter et al. (2008) in tropical forests, seed size was the best predictor of adult mortality rate (MR).

Leaf functional traits have also been shown to be good predictors of plant performance in tropical forests (Poorter & Bongers 2006; Sterck, Poorter & Schieving 2006), which is not surprising considering that most of the biochemical processes that shape plant function take place in leaves. Leaf traits also illustrate the important fact that functional traits do not tend to vary in isolation. A close coordination among leaf functional traits has been found in studies across species (Reich, Walters & Ellsworth 1997; Wright et al. 2004), from which specific leaf area (SLA), representing the amount of light-capturing area per unit of biomass invested in leaf tissue, has emerged as a key trait. Plants with high SLA tend to have high nitrogen concentrations per unit of mass, high photosynthetic assimilation and respiration rates, and lower leaf life span (Wright et al. 2004, 2005). The area of individual leaves is believed to represent another, apparently independent, dimension of trait variation (Wright et al. 2007). Species with larger leaves tend to have less frequent branching, bear larger fruits, and have thicker boundary layers. As a result, they overheat more easily than species with smaller leaves, leading to higher respiration and transpiration costs (Givnish 1978). Additionally, leaf size is intimately linked to xylem hydraulics (e.g.Zwieniecki, Boyce & Holbrook 2004), which determines yet another mechanism by which this trait can influence plant performance.

The coordination among functional traits is not limited to leaf traits. Recent studies have shown, for instance, that WD and leaf size tend to be related, probably through the hydraulics of the water transport system (Ackerly 2004; Pickup, Westoby & Basden 2005; Wright et al. 2007). Likewise, taller species tend to have larger seeds, a pattern that might be related to the time needed to reach reproductive maturity (cf. Moles et al. 2004; Wright et al. 2007). More generally, these relationships among traits are likely to reflect broader ecological strategies positioning plants along the disturbance spectrum, from fast-growing species specialized to ephemeral environments to longer-lived, stress-tolerant species (Grime 2001). This distinction gives rise to the trade-off between growth and survival or, more precisely, between a species’ ability to grow quickly under optimum conditions vs. its ability to avoid mortality when conditions are more stressful. This trade-off, particularly in relation to the light environment, is frequently observed in tropical forests (Wright et al. 2003; Poorter et al. 2008).

In this study, we explore the interspecific relationships between climate, key functional traits and demographic rates (growth and survival) for 44 tree species characteristic of temperate and Mediterranean ecosystems, corresponding to the most widely distributed forest species in Spain. To our knowledge, this is the first time these relationships are studied in adult trees outside the Tropics. The climate in Spain is very variable spatially (Ninyerola, Pons & Roure 2005; Fig. 1) and includes harsh environments that pose unfavourable conditions for trees to live in, ranging from very low temperatures at mountain tops (cold timberline) to extremely dry Mediterranean conditions in the SW of the country (dry timberline). The fact that our analysis is not limited to one particular community or a set of a few communities, but to all forest communities within the study region, allows us to work with species averages across a relatively large area and explore to what extent the differences in demographic rates across species are more related to climatic differences among their ranges or to their specific functional traits. As climate is considered to be the main determinant of the abundance and distribution of plant species (e.g.Woodward 1987), we expected to find a strong effect of climate on demographic rates. We asked three main questions: (i) are the differences in demographic rates across species more related to climatic differences among their ranges or to their functional traits? (ii) are the relationships between functional traits and demographic rates in the studied tree species similar to those found for tropical species? and (iii) is there a negative relationship between growth rate and survival, indicative of a trade-off, and is it mediated by the relationships of these variables with the studied functional traits?

image

Figure 1.  Location of the c. 90 000 sampling plots of the Spanish National Forest Inventory (IFN) represented on a map of mean annual precipitation (MAP) of the Iberian Peninsula. Source of MAP data: Digital Climatic Atlas of the Iberian Peninsula (Ninyerola, Pons & Roure 2005).

Download figure to PowerPoint

Materials and methods

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

Study area

The area covered by this study includes all the territories of Spain identified as forests by the Spanish National Forest Inventory (Dirección General de Conservación de la Naturaleza 2006) (Fig. 1). The overall climate is temperate according to Köppen’s classification. In the North-West the climate is relatively wet and cold, with a dominant Atlantic influence, whereas most of the remaining area falls under different types of Mediterranean climate, with the exception of some mountain ranges (e.g. Pyrenees). There is also a continental gradient inland towards the centre of the Iberian Peninsula, with colder winters and very hot summers. Climate extremes cover a wide range of values, with average annual temperatures ranging between < 5 °C at high altitudes in mountain ranges (particularly in N Spain) to c. 19 °C in some parts of the Guadalquivir valley (S Spain), and total annual rainfall varying between > 2000 mm in some locations in Galicia (NW Spain) and < 250 mm in Almería (SE Spain) (Fig. 1). Vegetation types across the climatic-topographic gradients include sclerophyllous, evergreen shrublands and forests, deciduous forests, and coniferous forests (Rivas Martínez 1987), all with a long history of human management (cf.Blondel & Aronson 1999).

Demographic rates

The data on tree growth and mortality were obtained from the second and third Spanish National Forest Inventories (IFN2 and IFN3, respectively; Dirección General de Conservación de la Naturaleza. 2006). Those inventories sampled the whole forested area of Spain (c. 10.7 × 106 ha) at a density of 1 plot km−2 following a regular design, for a total of c. 90 000 plots (Fig. 1). The exact position of each plot was recorded to allow for cross-referencing with climatic data bases and previous surveys. The IFN surveys include exhaustive information on the composition of canopy and understorey woody species, as well as on forest production and structure. Plots were circular with variable radius, so that the size of the plots depended on the diameter of the measured trees. Four plot radii were used: within 5 m of the centre of the plots all trees with diameter at breast height (d.b.h.) ≥ 7.5 cm were identified and measured, between 5 and 10 m around the centre of the plots only trees with d.b.h. ≥ 12.5 cm were considered, whereas 10–15 m around the centre of the plots only trees with d.b.h. ≥ 22.5 cm were included, and at 15–25 m around the centre only very large trees (d.b.h. ≥ 42.5 cm) were measured.

A large fraction of the plots inventoried during the second IFN (IFN2, period 1986–96) were revisited and measured during the third IFN (IFN3, 1997–2007), c. 11 years later (54 300 plots including at least one tree with d.b.h. > 7.5 cm were resampled). Because all trees were individually identified, it was possible to calculate tree growth at the individual level as well as to follow the fate of individual trees. Only plots that could be positively relocated in the field were included in our analyses. As forest management could introduce a significant bias in our estimates of demographic rates, we constrained our analyses to plots for which no evidence of management (cutting or thinning) was recorded in the plot or the surrounding area during the IFN3 survey (= 40 373 plots). For the same reason, crops and plantation species that are primarily grown for timber production in Spain (such as Eucalyptus spp., Pinus radiata, Populus x canadensis and Pseudotsuga menziesii) were excluded from the analyses.

The next step was to calculate species averages of growth and MR (see below). For large tree species (typical d.b.h. > 15 cm and typical height > 10 m), only individuals with d.b.h. > 12.5 cm were considered in the calculations to exclude juvenile individuals with potentially distinct demographic rates (see below). All individuals with d.b.h. > 7.5 cm were included in the case of small trees. The IFN3 includes information on c. 100 tree species. In order to obtain representative estimates of demographic rates, we included only species for which at least 100 of the individual trees measured during the IFN2 were measured again in the IFN3 survey. This lowered our sample size to 44 tree species (Fig. 2).

image

Figure 2.  Phylogenetic tree of the 44 tree species included in the study (see text for details). The names of major clades are indicated on the corresponding nodes.

Download figure to PowerPoint

The relative growth rate (RGR) was used as a measure of growth. Relative growth rate was calculated for each individual tree as (ln[DBHIFN3]−ln[DBHIFN2])/time, where DBHIFN2 and DBHIFN3 are the diameters at breast height measured at the second and third Spanish Forest Inventories, respectively, and time is the time interval between measurements. Relative growth rate values were then averaged by species (= 100 – 69 037 individuals, depending on the species). The contribution of each individual to the average was weighted according to its d.b.h. class to account for the fact that plot size was not constant (see above). The mean RGR calculated in this way represents the average RGR that would have been obtained had individual trees been sampled randomly, regardless of their size.

Annual MR were calculated as (ln[NIFN2]−ln[NIFN3])/time, where NIFN2 is the number of trees of a given species that were recorded at the IFN2, NIFN3 the number of those that were still alive at the IFN3, and time is the time interval between surveys. Mortality rate values were calculated separately for each d.b.h. class and averaged taking into account the unequal sampling effort across tree sizes as above. The number of individuals included in the MR calculation ranged between 107 (Populus alba) and 68 441 (Pinus sylvestris). We evaluated the potential bias in MR for species with relatively low sample size by estimating mortality in the commonest species (P. sylvestris) by randomly resampling a number of individuals equal to the minimum sample size for any species (= 107). The outcome of these tests showed that the resampled estimates for those simulations (= 100) were on average within 5% of the actual value, and that 62% of the individual trials were within 25% of the actual value.

One potential caveat of our approach is that demographic rates are likely to be affected by tree size, and tree size distributions vary substantially across species. The effect of size could be particularly important for the comparisons of growth across species (Poorter et al. 2008). In order to gauge the impact that the differences in tree size across species could have in our analyses we studied in some detail the relationship between d.b.h. and RGR both within and among species. In all cases, there was a negative relationship between d.b.h. and RGR within species (significant, < 0.05, in 38 of 44 cases), whereas the overall relationship between d.b.h. and RGR among species was positive (R2 = 0.26, < 0.05). This result shows that the relationship between size and growth we found across species cannot be due to differences in their size distributions, as species with larger size tended to have higher, not lower, RGR. We also used the within-species relationship between d.b.h. and RGR for each species to estimate growth of 20 cm d.b.h. trees (cf.Poorter et al. 2008). These values were highly correlated with the average RGR used in our analyses (Pearson’s = 0.94, < 0.001) with a slope of 0.93, indicating that size-dependent growth did not bias our results.

Functional traits

The information on functional traits was obtained mostly from the Catalan Ecological and Forest Inventory (IEFC, Burriel et al. 2000-2004; http://www.creaf.uab.es/iefc/). This inventory was carried out between 1988 and 1998, and included a total of 10 664 plots randomly distributed throughout the forested area of Catalonia, NE Spain (1.2 × 106 ha of forest, accounting for 11.2% of Spanish forests). In each of these 10 m radius plots all trees with d.b.h. > 5 cm were labelled and measured. In a random subsample of 20% of the plots, an additional sampling was carried out on one or two representative individuals of each 5 cm diameter class of the dominant tree species in the plot. From each of these individuals, branches of different sizes and order were sampled and weighted in order to get allometric relationships, and leaf and wood samples were taken to the laboratory to measure nutrient composition, SLA and WD.

Six functional traits are considered in this study (cf.Westoby et al. 2002; Westoby & Wright 2006; Wright et al. 2007): maximum tree height (Hmax, m), WD (g cm−3), SLA (mm2 mg−1), nitrogen content of leaves (Nmass, % mass), leaf area (LA, cm2) and seed mass (SM, mg). Of those, Hmax was taken entirely from the IFN data set; WD, SLA and Nmass were taken mostly from the IEFC data set, completed with published values for species for which the number of individuals sampled in the IEFC was < 3. Leaf size was taken from Bolós et al. (1990) and seed volume was taken entirely from published sources (see Appendix S1 in Supporting Information for a complete list of methods, and Table S1 for data sources by variable and species). Only values measured in field populations using methodologies comparable to those used in the IEFC were included in our data set. Literature values were taken primarily from measurements taken in Spanish populations. In the very few cases (= 3) in which those were not available we used values measured elsewhere.

Using average values for species functional traits, without explicitly accounting for the spatial variability of those traits within a species, assumes that functional traits are more variable between than within species. We tested this assumption using the IEFC data set for WD, Nmass and SLA. A variance components analysis revealed that in all cases variation was greater across species (69%, 86% and 82% of total variation was explained by between-species differences for WD, Nmass and SLA, respectively), supporting our approach. Additionally, we used maximum tree height data, the only trait for which we had information at both the Catalan and the Spanish scale, to compare the estimates at these two spatial scales. The results confirm that the values are indeed similar (Hmax [Catalonia] = 0.94 Hmax [Spain] −1.3, R2 = 0.85, = 44).

Climatic data

Climatic data for each IFN plot was obtained from the Digital Climatic Atlas of the Iberian Peninsula (Ninyerola, Pons & Roure 2005; http://opengis.uab.es/wms/iberia/index.htm). The average values of mean annual temperature (MAT) and mean annual precipitation (MAP) for the period 1951–99 were used to characterize the climate of each plot. A mean climate was obtained for each species by averaging each of the two previous variables for all the plots where the species was present. We also characterized the climatic amplitude of each species by conducting a principal components analysis (PCA) of the standard deviations of the main climate variables across each species’ range and using the scores on the first axis of the PCA (accounting for 56% of variance) as a measure of amplitude. The variables included in the PCA were the standard deviations of MAT, MAP, minimum annual temperature, maximum annual temperature, annual potential evapotranspiration and the ratio of summer (June–August) precipitation to potential evapotranspiration. Average MAP, average MAT and climatic amplitude were not significantly correlated to each other across species (> 0.05), and were thus used as indicators of approximately orthogonal climatic axes.

Analyses

All variables were checked for normality and transformed by applying logarithms or square roots whenever required (cf.Table 1). Simple linear regression analysis was used to quantify the influence of functional traits and climate on demographic rates and Pearson correlation coefficients were used to test for association between pairs of functional traits. Because species descend hierarchically from common ancestors, they generally cannot be considered as independent data points in statistical analyses. We used phylogenetically independent contrasts (PICs; Felsenstein 1985) to asses to what extent correlations between species traits correspond to an evolutionary association of those traits or largely reflect trait differences among clades. In PIC analyses each evolutionary divergence contributes a data point. In regression analyses, we explicitly took into account phylogenetic effects by using phylogenetic generalized least squares (PGLS, Freckleton, Harvey & Pagel 2002). In this case, PGLS is more robust and generic than PICs, as it properly represents the actual phylogenetic load that each variable incorporates instead of assuming that the load is maximal and equal to the phylogenetic distance. We fitted each PGLS model assuming a Brownian or an Ornstein–Uhlenbeck (OU) model of character evolution (Martins & Hansen 1997). As the results for these two models of character evolution were similar, and the OU model tended to fit the data slightly better, we only report the results obtained using the OU model. The proportion of the variation of each variable explained by phylogenetic distance was calculated using phylogenetic autoregression as in Paradis (2006).

Table 1.   Relationship between functional traits and demographic rates modelled using partial least squares, both with and without including phylogenetic effects (see text). Significant relationships are highlighted in bold. An asterisk after the corresponding Akaike Information Criterion (AIC) value indicates that the fit of the model including phylogenetic effects is significant better
ModelWithout phylogenetic effectsIncluding phylogenetic effects
AICCoefficientPAICCoefficientP
  1. RGR, relative growth rate; MR, mortality rate; Hmax, maximum tree height; WD, wood density; SLA, specific leaf area; Nmass, nitrogen content of leaves; LA, leaf area; SM, seed mass; MAT, mean annual temperature; MAP, mean annual precipitation; Amplitude, climate amplitude of the species’ range (see text).

log(RGR)–Hmax12.940.027< 0.00112.010.026< 0.001
log(RGR)–WD12.30−1.495< 0.0018.55*−1.525< 0.001
log(RGR)–log(SLA)31.880.0450.60824.59*0.1340.279
log(RGR)–sqrt(Nmass)24.470.4880.0065.66*1.011< 0.001
log(RGR)–sqrt(LA)34.540.0250.12421.25*0.0530.007
log(RGR)–log(SM)33.04−0.0240.12827.83*−0.0220.270
log(RGR)–MAT35.630.0180.45026.46*0.0280.167
log(RGR)–MAP44.17< 0.0010.19635.50*< 0.0010.136
log(RGR)–Amplitude35.68−0.0160.54428.11*< 0.0010.967
log(MR)–Hmax128.910.0010.941123.46*−0.0010.946
log(MR)–WD109.63−3.5270.001108.66−3.2200.008
log(MR)–log(SLA)119.040.1760.486111.66*0.2190.522
log(MR)–sqrt(Nmass)121.470.4370.418116.07*0.3160.654
log(MR)–sqrt(LA)126.920.0100.832121.22*0.0140.815
log(MR)–log(SM)117.38−0.0690.126116.88−0.0160.767
log(MR)–MAT126.180.0160.822120.83*0.0320.620
log(MR)–MAP132.650.0010.074127.23*0.0010.072
log(MR)–Amplitude124.10−0.1050.162120.44*−0.0460.519

A phylogenetic tree for the study species was constructed using Phylomatic (Webb & Donogue 2005) based on the maximally resolved tree (Davies et al. 2004). As this tree was not fully resolved, genus- and species-level polytomies were resolved by obtaining additional phylogenetic information from (i) the user-supplied data repository in Phylomatic for the Pinaceae (Rydin 2002), Rosaceae (Potter 2007) and Betulaceae (Chen 1999) and (ii) from searches of the primary literature for the following clades: the genera Acer (Li, Yue & Shoup 2006), Pinus (Wang, Tank & Sang 2000; Gernandt et al. 2005) and Populus (Hamzeh & Dayanandan 2004). For the genus Quercus, alternative phylogenies were available (Kremer & Petit 1993; Manos, Doyle & Nixon 1999; Bellarosa et al. 2005), differing in the exact position of Quercus suber, which resulted in two slightly different phylogenetic trees. All analyses were conducted using the two trees and, as the results were similar in all cases, only the analyses based on one of the trees (Fig. 2) are presented. As the phylogenetic tree was a composite from multiple sources and we lacked data on branch lengths, all branch lengths were set equal to 1. The r packages ape and nlme were used to carry out the phylogenetic analyses.

Path analysis (structural equation modelling with no latent variables) was employed to compare alternative conceptual models of the way demographic rates were affected by functional traits and climate characteristics. We started with the ‘saturated’ model including all possible directional relationships between functional traits and demographic rates and between climate and demographic rates, plus all possible covariations among functional traits. This model was simplified by stepwise removing the least significant path until the fit of the model did not increase further. An alternative path analysis was constructed using PICs to check for the consistency of the reported relationships. Seed mass and climatic amplitude were not included in the path models based on their poor correlation with the other studied variables. All variables were standardized prior to fitting the path models. The Akaike Information Criterion (AIC) was used as model selection criterion for both PGLS and path analysis. The package amos (SPSS Inc., Chicago, IL, USA) was used to carry out the path analyses. Significance for all statistical tests was accepted at α = 0.05.

Results

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

Mortality vs. growth rates

There was a marginal positive relationship between RGR and MR across species (R2 = 0.07, = 0.088, = 44) (Fig. 3). However, when phylogenetic effects were taken into account using PGLS, the model improved significantly (AIC = 119.6 without phylogenetic effects vs. AIC = 116.2 when those effects were included, = 0.02), and the relationship between RGR and MR disappeared (= 0.314). We did not find any consistent difference between angiosperms and gymnosperms in either RGR or MR (P > 0.68).

image

Figure 3.  Relationship between mortality rate (MR) and relative diameter growth rate (RGR) for the 44 studied tree species. The depicted line corresponds to the regression between the two variables, which was marginally significant. Note that both axes are logarithmic.

Download figure to PowerPoint

Pairwise relationships between demographic rates, functional traits and climate

Relative growth rate was positively related to Hmax (R2 = 0.46, < 0.001) and Nmass (R2 = 0.17, < 0.006) and negatively related to WD (R2 = 0.35, < 0.001). Relative growth rate was unrelated to all other variables, including the climatic ones (> 0.1 in all cases) (Fig. 4). The relationships between RGR and Hmax, Nmass and WD remained significant when the phylogenetic effects were included in the models using PGLS (Table 1). Interestingly, a significant, positive relationship between RGR and LA appeared in the model accounting for phylogenetic effects (Table 1). Mortality rate, on the other hand, was only related to WD, with MR values declining with increasing WD (R2 = 0.23, = 0.001) (Fig. 5). Again, this relationship was conserved when phylogenetic effects were taken into account (Table 1). In all cases, except for the relationship between RGR and LA pointed out above, the results were qualitatively identical between phylogenetic and non-phylogenetic analyses. However, including phylogenetic effects improved model fit significantly in 13 out of 16 cases (Table 1), highlighting the fact that the data showed a significant level of phylogenetic autocorrelation. On average, the percentage of variance explained by phylogenetic distance in the variables of our data set was 46%, ranging between 24% (for Hmax) and 68% (for SLA).

image

Figure 4.  Relationship between relative growth rate (RGR), functional traits and climate. Only significant regressions are depicted. Note that some axes are logarithmic. See Table 1 for abbreviations.

Download figure to PowerPoint

image

Figure 5.  Relationship between mortality rate, functional traits and climate. Significant regressions are depicted as continuous lines; a dotted line indicates a marginally significant relationship (0.05 < < 0.1). Note that some axes are logarithmic. See Table 1 for abbreviations.

Download figure to PowerPoint

Interestingly, the traits that were more tightly related to demographic rates, namely WD and Hmax, were the ones showing poorer relationships with climatic variables (Table 2). Only SM was significantly related to MAT (R= 0.23, = 0.28), whereas SLA, Nmass and LA were related to MAP (R2 = 0.29–0.31, < 0.001 in all cases). The relationship between SM and MAT disappeared when phylogenetic effects were considered, whereas significant relationships appeared between MAP and Hmax and between MAP and WD. Climatic amplitude was unrelated to any of the measured functional traits. Overall model fit was always better for phylogenetic than for non-phylogenetic models (Table 2).

Table 2.   Relationship between functional traits and climatic variables modelled using partial least squares, both with and without including phylogenetic effects (see text). Significant relationships are highlighted in bold. An asterisk after the corresponding Akaike Information Criterion (AIC) value indicates that the fit of the model including phylogenetic effects is significant better
ModelWithout phylogenetic effectsIncluding phylogenetic effects
AICCoefficientPAICCoefficientP
  1. MAT, mean annual temperature; MAP, mean annual precipitation; Amplitude, climate amplitude of the species’ range; Hmax, maximum tree height; WD, wood density; SLA, specific leaf area; Nmass, nitrogen content of leaves; LA, leaf area; SM, seed mass.

Hmax–MAT305.24−0.4930.405293.16*−0.5340.266
WD–MAT−44.690.0170.064−63.87*0.0080.262
log(SLA)–MAT82.93−0.0110.80343.27*−0.0210.377
sqrt(Nmass)–MAT20.850.0120.550−15.98*0.0040.729
sqrt(LA)–MAT223.320.1520.495199.25*−0.0470.762
log(SM)–MAT214.080.5230.028187.11*0.1870.217
Hmax–MAP312.300.0080.071298.97*0.0100.026
WD–MAP−32.40> −0.0010.362−59.87*> −0.0010.010
log(SLA)–MAP77.780.001< 0.00146.82*0.0010.010
sqrt(Nmass)–MAP16.510.001< 0.001−10.51*< 0.0010.046
sqrt(LA)–MAP218.750.006< 0.001202.05*0.0040.010
log(SM)–MAP228.50−0.0010.782197.98*> −0.0010.883
Hmax–Amplitude305.79−0.0210.975293.60*0.4080.456
WD–Amplitude−41.38−0.0010.906−63.81*−0.0080.313
log(SLA)–Amplitude81.86−0.0440.38743.11*0.0210.495
sqrt(Nmass)–Amplitude20.08−0.0210.327−17.80*0.0170.187
sqrt(LA)–Amplitude221.84−0.3180.183198.57*0.1280.459
log(SM)–Amplitude217.610.2870.277188.02*0.1080.516

As expected, the studied functional traits did not vary independently of each other. Leaf traits, for instance, were highly intercorrelated, with positive relationships between SLA, LA and Nmass (Fig. 6a). Additionally, WD showed a negative relationship with Hmax and a positive relationship with SM, whereas Hmax was positively related to Nmass and LA (Fig. 6a). This structure of the relationships was not fundamentally altered when PICs were used instead of raw values, although the significance of some individual relationships changed (Fig. 6b). In particular, the relationship between the PICs of WD and SM was no longer significant (= 0.28, = 0.07), the PICs of Hmax and SLA became positively correlated, and the PICs of WD were negatively associated with the PICs of all leaf-level traits.

image

Figure 6.  Pairwise correlations between functional traits for (a) raw values and (b) phylogenetic independent contrasts (PICs). Dashed lines indicate negative relationships. Some variables were transformed to achieve normality (cf.Table 1). See Table 1 for abbreviations.

Download figure to PowerPoint

Determinants of demographic rates

The results of the path analysis aided the interpretation of the previously reported relationships. The final path model provided a reasonable fit to the data (χ2 = 4.03, = 42, d.f. = 7, = 0.78, AIC = 46.03; note that because the goal of path analysis is to develop a model that fits the data, a non-significant chi-squared value is desired) (Fig. 7a). The model accounted for a large proportion of the variability in RGR (67%), whereas the variance explained was lower for MR (38%). The significant paths and the sign of the path coefficients were generally consistent with the regression and correlation analyses reported above: RGR was negatively related to WD and positively related to Hmax and Nmass, whereas MR declined with increasing WD and increased with increasing MAP (marginally significant in the PGLS analysis). Some additional relationships were detected by the path model, mostly related to indirect effects of climate on demographic rates through modifications of WD, Hmax and Nmass (Fig. 7a). A significant negative relationship appeared between Hmax and MR, although this relationship is complicated by the indirect effect through the covariation between Hmax and WD (Fig. 7a). Finally, the path model also identified the significant correlations between WD and Hmax and between Hmax and Nmass reported before. All the significant relationships reported above were preserved when the path model was constructed using PICs instead of raw trait values (not shown).

image

Figure 7.  (a) Path model relating demographic rates, functional traits and climate for the studied species. (b) The same model as in (a) but setting all directional effects of functional traits to zero to illustrate the fact that when only the direct effects of climate were taken into account the model explained a low proportion of the variability in demographic rates. If a directional path was added between MAP and RGR the model improved slightly and accounted for 11% of the variability in RGR, still much lower than the value corresponding to model (a). Arrows indicate the proposed links between variables (single headed: directional paths, double-headed: covariances). Dashed lines indicate negative relationships, and thick lines correspond to stronger relationships (absolute value of standardized path coefficients > 0.45, < 0.001). The values close to each line indicate standardized path coefficients (shown only for directional relationships for clarity). The number in brackets over a given endogenous variable in the path diagram corresponds to the R2 value indicating the percentage of the variance in that variable that is accounted for by the model. Some variables were transformed to achieve normality (cf.Table 1). See Table 1 for abbreviations.

Download figure to PowerPoint

In order to quantify the relative importance of functional traits vs. climatic variables in accounting for the intraspecific variability in demographic rates, we recalculated the previous path model by setting to zero the direct effects of either group of variables on RGR and MR. The results were conclusive: if climatic effects were not considered the path model still accounted for 57% of the variability in RGR and 24% of the variability in MR. In contrast, if the direct effect of functional traits was removed, the resulting model accounted for virtually none of the variability in RGR and only 10% of the variability in MR (Fig. 7b).

Discussion

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

The ranges of values we obtained for the studied functional traits are similar to those reported for tropical species (cf.Poorter et al. 2008 for Hmax, WD and SLA). Similarly, the values of SLA and Nmass reported here are around the middle range of the values included in the global leaf economics spectrum (Wright et al. 2004). In the case of demographic rates, our MR and RGR values partially overlapped the ranges reported for tropical forests, but there were some noticeable differences, as our MR values tended to be lower, and our RGR values higher, on average, than those reported in any of the tropical communities measured by Poorter et al. (2008).

Coordination among functional traits

In terms of the coordination among functional traits, our results are generally consistent with the relationships reported previously across species. In particular, leaf traits were highly inter-correlated, with strong positive relationships between SLA and Nmass (cf.Wright et al. 2004). The positive relationship between SLA and LA found in this study is consistent with some previous reports looking at the relationship across species from different sites (e.g. Niinemets 2001), although this relationship does not seem to hold universally (e.g. Wright et al. 2007) and the relationship between the two traits may actually be negative within a given ecological community (Ackerly et al. 2002). The positive association between SLA and LA across species from a wide range of climates including dry sites is likely to be mediated by the convergence of species with low SLA and low LA in low rainfall areas (Givnish 1987; Ackerly et al. 2002), as suggested by our own results (Table 2), and does not necessarily imply a direct link between the two variables.

We found evidence of coordination at the whole-plant level, as shown by significant correlations between leaf traits and other plant functional traits. These relationships became particularly clear when phylogenetic effects were taken into account (Fig. 6b). The positive association between tree height and SLA, LA and Nmass contrasts with the suggestion by Wright et al. (2007) that plant height may be considered essentially orthogonal to other plant strategy dimensions. On the other hand, the relationships between WD and the other functional traits measured in this study are consistent with a whole-plant strategy by which species with larger leaves and more photosynthetic capacity have higher volumetric growth and lower WD (Chave et al. 2009). This strategy may be mediated by plant hydraulics and, in particular, by the negative relationship between WD and hydraulic efficiency (Wright et al. 2007), although the evidence for this relationship across species is not conclusive (Chave et al. 2009; Zanne et al. 2010). In contrast to the close coordination among functional traits described in the previous paragraphs, seed size (SM) was essentially independent of all the other studied traits, consistent with previous results showing that it represents a separate dimension of plant trait variation (Wright et al. 2007).

Functional traits as the main determinants of growth and mortality across species

One of the main findings of this work is the fact that the studied functional traits explained a large proportion of the variation in demographic rates across species. While this is consistent with previous reports (King, Davies & Nur Sapardi 2006a; King et al. 2006b; Sterck, Poorter & Schieving 2006; Osunkoya et al. 2007; Poorter et al. 2008), it should be noted that our study is unique in at least two aspects. First, it is the first time, to our knowledge, that these relationships are explored outside tropical forests. Secondly, and partially because of the inherent differences in species richness between tropical and temperate forests, our study was not limited to one or a small set of communities. Instead, we computed species-specific averages of functional traits and demographic rates by combining data from forests within a large (c. 500 000 km2) study region, including a substantial proportion of each of the species’ range. Although this approach has some evident limitations, as it mixes sources of variability at different levels (e.g. the spatial variability within a species’ range was not explicitly accounted for), the nature of the sampling design was such that it allowed the computation of species-specific values without incurring any obvious bias (cf. Materials and methods). In this context, it is particularly remarkable that our path models, including only a coarse representation of functional traits and average climate across the species’ ranges in Spain, accounted for 67% of the variation in growth and 38% of the variation in mortality across species.

Our study region covers a wide range of climatic conditions, including very harsh climates. Not surprisingly, the studied species differed markedly in their climatic niche. For instance, average rainfall is 523 ± 155 mm (mean ± SD) in Pinus halepensis plots, while it is 1487 ± 287 mm for Quercus robur, a species that only grows under relatively humid conditions. Similarly, for temperature, the average values range between 6.1 ± 1.4 °C for Pinus uncinata and 16.1 ± 1.2 °C for Quercus canariensis. It is well known that climate gradients affect demographic rates of individual species and also macroscopic properties of ecosystems such as productivity (Huston & Wolverton 2009). It is thus noteworthy that most of the explanatory power of our path models came not from differences in the average climate across species ranges but from differences in species-specific functional traits. This result implies that functional traits are tuned to the specific environment of each species so that species changes along climate gradients, together with the corresponding changes in functional traits, are able to buffer the effects of climate on plant performance. Obviously, large scale climatic gradients are superimposed on microclimatic effects associated with topography, aspect, wind and soils, which together with disturbance history actually determine vegetation composition and structure (cf.Bonan 2002). This finer-scale variation should be also taken into account if we are to fully understand the variation in demographic rates along climatic gradients. In any case, our results are consistent with the view that functional traits and their variation along abiotic environmental gradients play a key role in community ecology (cf.McGill et al. 2006).

With regards to the effects of functional traits on demographic rates, our results are generally consistent with the studies carried out in tropical forests (cf.Poorter et al. 2008). Specifically, we confirmed the importance of tree height and particularly WD in influencing growth and MR also in temperate and Mediterranean forests. In our case, WD, Hmax and Nmass were the best predictors of RGR, in agreement with previous studies (King et al. 2005, 2006b; Sterck, Poorter & Schieving 2006; Poorter et al. 2008). On the other hand, WD was by far the best predictor of MR in our data set. This is only partially consistent with previous reports (Osunkoya et al. 2007; Poorter et al. 2008), as Poorter et al. (2008) in their analysis combining data from different tropical forests found MR to be better correlated with seed size (seed volume in their case) than with WD. Large seeds have been associated with shade tolerance and a suit of characters (Moles & Westoby 2006) that may confer inherently lower growth and MR (cf.Poorter et al. 2008). However, the relationship between shade tolerance and seed size has been disputed, at least in the tropics (Metcalfe, Grubb & Turner 1998; Valladares & Niinemets 2008). The low association of seed size with demographic rates found in our study suggests that the relationship between seed size and (adult) mortality reported for tropical forests may not be applicable to other ecosystem types, and that the higher survival of large-seeded species during seedling establishment (Moles & Westoby 2004) does not necessarily have carry-over effects on tree performance at the adult stage.

One important aspect that emerges from recent work on plant functional traits is the central importance of WD as an integrator of wood properties (Chave et al. 2009). Our study expands the evidence gathered so far, coming almost exclusively from the tropics, to temperate and Mediterraean ecosystems. The reasons why WD is such a good predictor of demographic rates are only partially understood. Dense wood tends to be more resistant to stress, both in terms of mechanical damage (e.g.Curran et al. 2008) and drought (Hacke et al. 2001), which might explain why trees with denser woods tend to show lower MR. Hydraulic aspects related to drought resistance are likely to be particularly relevant in our study system (McDowell et al. 2008), as most of Spain is characterized by a typical Mediterranean climate with dry to very dry summers.

The reason why growth rates decline with increasing WD are less clear, particularly if RGR is used as a measure of growth. Because WD is likely to vary little over short periods of time, RGR in terms of biomass and in terms of volume (or tree diameter) should scale roughly isometrically, and a relationship between RGR and WD is not necessarily to be expected. Some studies have suggested a relationship between WD and specific hydraulic conductivity, which is believed to limit transpiration and, ultimately, growth (Chave et al. 2009). However, a recent meta-analysis clearly shows that WD is almost entirely disconnected from hydraulic conductivity across angiosperm species (Zanne et al. 2010). In a recent study across 42 rain forest tree species, Poorter et al. (2010) found growth rate to be more tightly linked to WD than to xylem anatomical properties or hydraulic conductivity, to the point that the latter relationships actually disappeared when phylogenetic independent contrasts were used instead of raw trait values.

Trade-off between growth and survival

Our results suggest a loose, but negative, relationship between growth rate and survival across the studied species (Fig. 3), mediated by key functional traits such as WD and maximum tree height. However, this relationship disappeared when phylogenetic effects were taken into account. An interspecific demographic trade-off between a species’ ability to grow quickly vs. its ability to avoid mortality is commonly found within tropical communities (Wright et al. 2003; Poorter et al. 2008). However, it is much less clear whether such a trade-off is to be expected in temperate or Mediterranean forests; in particular when different communities are compared. The latter is because variable access to resources can confound the relationships among traits (cf.van Noordwijk & de Jong 1986) and it is likely to result in different associations between growth and mortality along gradients of resource availability (Russo et al. 2008).

Conclusion

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

We found that the variation in demographic rates across 44 temperate and Mediterranean tree species was more related to differences in functional traits than to climatic differences among the species’ ranges, suggesting that variation in functional traits across species allows them to perform largely independently of climatic conditions along environmental gradients. Plant growth and survival depend on a suit of coordinated traits (Lambers & Poorter 1992; Reich et al. 2003). Some of those traits, such as WD and maximum tree height, seem to summarize many other plant features, and are thus emerging as the best single predictors of demographic rates across species. The mechanisms behind the key role of these two traits are not yet completely understood, but they are probably best seen from the point of view of whole-plant ecological strategies and the trade-off between investment in durable structures (i.e. high WD) and the possibility to realize high volumetric growth rates (i.e. high maximum tree height). Finally, the fact that the main aspects of our results were not altered by the explicit incorporation of phylogenetic effects suggests that the observed relationships have emerged repeatedly throughout evolutionary history and are not due to divergences between a few major clades.

Acknowledgements

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

We thank Mireia Burnat for technical assistance doing literature searches, and the IEFC and IFN staff for collecting the data and processing all the samples. This study was supported by the Spanish Ministry of Science and Innovation via competitive grants CGL2007-60120 and CSD2008-0004.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. References
  10. Supporting Information
  • Ackerly, D.D. (2003) Community assembly, niche conservatism, and adaptive evolution in changing environments. International Journal of Plant Sciences, 164, 165184.
  • Ackerly, D.D. (2004) Functional traits of chaparral shrubs in relation to seasonal water deficit and disturbance. Ecological Monographs, 74, 2544.
  • Ackerly, D.D., Knight, C.A., Weiss, S.B., Barton, K. & Starmer, K.P. (2002) Leaf size, specific leaf area and microhabitat distribution of woody plants in a California chaparral: contrasting patterns in species level and community level analyses. Oecologia, 130, 449457.
  • Bellarosa, R., Simeone, R.C., Papini, A. & Schirone, B. (2005) Utility of ITS sequence data for phylogenetic reconstruction of Italian Quercus spp. Molecular Phylogenetics and Evolution, 34, 355370.
  • Blondel, J. & Aronson, J. (1999) Biology and Wildlife of the Mediterranean Region. Oxford University Press, Oxford, UK.
  • Bolós, O., Vigo, J., Masalles, R.M. & Ninot, J.M. (1990) Flora Manual dels Països Catalans. Pòrtic, Barcelona.
  • Bonan, G.B. (2002) Forests and climate change: forcings, feedbacks, and the climate benefits of forests. Science, 320, 14441449.
  • Burriel, J.A., Gracia, C., Ibàñez, J.J., Mata, T. & Vayreda, J. (2000-2004) Inventari Ecològic i Forestal de Catalunya, 10 volumes. CREAF, Bellaterra, Spain.
  • Chave, J., Coomes, D., Jansen, S., Lewis, S.L., Swenson, N.G. & Zanne, A.E. (2009) Towards a worldwide wood economics spectrum. Ecology Letters, 12, 351366.
  • Chen, Z. (1999) Phylogeny and evolution of the Betulaceae as inferred from DNA sequences, morphology, and paleobotany. Plant Systematics and Evolution, 86, 11681181.
  • Curran, T.J., Gersbach, L.N., Edwards, W. & Krockenberger, A.K. (2008) Wood density predicts plant damage and vegetative recovery rates caused by cyclone disturbance in tropical rainforest tree species of north Queensland, Australia. Austral Ecology, 33, 442450.
  • Dalling, J.W. & Hubbell, S.P. (2002) Seed size, growth rate and gap microsite conditions as determinants of recruitment success for pioneer species. Journal of Ecology, 90, 557568.
  • Davies, T.J., Barraclough, T.G., Chase, M.W., Soltis, P.S., Soltis, D.E. & Savolainen, V. (2004) Darwin’s abominable mystery: insights from a supertree of the angiosperms. Proceedings of the National Academy of Sciences, USA, 101, 19041909.
  • Dirección General de Conservación de la Naturaleza. (2006) Tercer Inventario Forestal Nacional 1997–2006, Barcelona. Ministerio de Medio Ambiente, Madrid.
  • Felsenstein, J. (1985) Phylogenies and the comparative method. American Naturalist, 125, 115.
  • Freckleton, R.P., Harvey, P.H. & Pagel, M. (2002) Phylogenetic analysis and comparative data: a test and review of evidence. American Naturalist, 160, 712726.
  • Gernandt, D.S., López, G.G., García, S.O. & Liston, A. (2005) Phylogeny and classification of Pinus. Taxon, 54, 2942.
  • Givnish, T.J. (1978) Ecological aspects of plant morphology: leaf form in relation to environment. Theoretical Plant Morphology (ed. R.Sattler), pp. 83142. Leiden University Press, Leiden, The Netherlands.
  • Givnish, T.J. (1987) Comparative studies of leaf form: assessing the relative roles of selective pressures and phylogenetic constraints. New Phytologist, 106, S131S160.
  • Grime, J.P. (2001) Plant Strategies, Vegetation Processes, and Ecosystem Properties. John Wiley and Sons, Chichester, UK.
  • Hacke, U.G., Sperry, J.S., Pockman, W.T., Davis, S.D. & McCulloh, K.A. (2001) Trends in wood density and structure are linked to prevention of xylem implosion by negative pressure. Oecologia, 126, 457461.
  • Hamzeh, M. & Dayanandan, S. (2004) Phylogeny of Populus (Salicaceae) based on nucleotide sequences of chloroplast TRNT-TRNF region and nuclear rDNA. American Journal of Botany, 91, 13981408.
  • Harper, J.L. (1977) Population Biology of Plants. Academic Press, London.
  • Huston, M.A. & Wolverton, S. (2009) The global distribution of net primary production: resolving the paradox. Ecological Monographs, 79, 343377.
  • Jacobsen, A.L., Agenbag, L., Esler, K.J., Pratt, R.B., Ewers, F.W. & Davis, S.D. (2007) Xylem density, biomechanics and anatomical traits correlate with water stress in 17 evergreen shrub species of the Mediterranean-type climate region of South Africa. Journal of Ecology, 95, 171183.
  • King, D.A., Davies, S.J., Nur Supardi, M.N. & Tan, S. (2005) Tree growth is related to light interception and wood density in two mixed dipterocarp forests of Malaysia. Functional Ecology, 19, 445453.
  • King, D.A., Davies, S.J. & Nur Sapardi, M.N. (2006a) Growth and mortality are related to adult tree size in a Malaysian mixed dipterocarp forest. Forest Ecology and Management, 223, 152158.
  • King, D.A., Davies, S.J., Tan, S. & Supardi, M.N.N. (2006b) The role of wood density and stem support costs in the growth and mortality of tropical trees. Journal of Ecology, 94, 670680.
  • Koch, G.W., Sillett, S.C., Jennings, G.M. & Davis, S.D. (2004) The limits to tree height. Nature, 428, 851854.
  • Kremer, A. & Petit, R.J. (1993) Gene diversity in natural populations of oak species. Annals of forest Science, 50(s1), s186s202.
  • Lambers, H. & Poorter, H. (1992) Inherent variation in growth-rate between higher-plants – a search for physiological causes and ecological consequences. Advances in Ecological Research, 23, 187261.
  • Li, J., Yue, J. & Shoup, S. (2006) Phylogenetics of Acer (Aceroideae, Sapindaceae) based on nucleotide sequences of two chloroplast non-coding regions. Harvard Papers in Botany, 11, 101115.
  • Manos, P.S., Doyle, J.J. & Nixon, K.C. (1999) Phylogeny, biogeography, and processes of molecular differentiation in Quercus Subgenus Quercus (Fagaceae). Molecular Phylogenetics and Evolution, 12, 333349.
  • Martins, E.P. & Hansen, T.F. (1997) Phylogenies and the comparative method: a general approach to incorporating phylogenetic information into the analysis of interspecific data. American Naturalist, 149, 646667.
  • McDowell, N., Pockman, W.T., Allen, C.D., Breshears, D.D., Cobb, N., Kolb, T., Plaut, J., Sperry, J., West, A., Williams, D.G. & Yepez, E.A. (2008) Mechanisms of plant survival and mortality during drought: why do some plants survive while others succumb to drought? New Phytologist, 178, 719739.
  • McGill, B.J., Enquist, B.J., Weiher, E. & Westoby, M. (2006) Rebuilding community ecology from functional traits. Trends in Ecology and Evolution, 21, 178185.
  • Mencuccini, M. (2003) The ecological significance of long-distance water transport: short-term regulation, long-term acclimation and the hydraulic costs of stature across plant life forms. Plant, Cell and Environment, 26, 163182.
  • Metcalfe, D.J., Grubb, P.J. & Turner, I.M. (1998) The ecology of very small-seeded shade-tolerant trees and shrubs in lowland rain forest in Singapore. Plant Ecology, 134, 131149.
  • Moles, A.T. & Westoby, M. (2004) Seedling survival and seed size: a synthesis of the literature. Journal of Ecology, 92, 372383.
  • Moles, A.T. & Westoby, M. (2006) Seed size and plant strategy across the whole life cycle. Oikos, 113, 91105.
  • Moles, A.T., Falster, D.S., Leishman, M.R. & Westoby, M. (2004) Small-seeded species produce more seeds per square metre of canopy per year, but not per individual per lifetime. Journal of Ecology, 92, 384396.
  • Niinemets, U. (2001) Global-scale climatic controls of leaf dry mass per area, density, and thickness in trees and shrubs. Ecology, 82, 453469.
  • Ninyerola, M., Pons, X. & Roure, J.M. (2005) Atlas Climático Digital de la Península Ibérica. Metodología y aplicaciones en bioclimatología y geobotánica. Universidad Autónoma de Barcelona, Bellaterra, Spain.
  • Van Noordwijk, A.J. & De Jong, G. (1986) Acquisition and allocation of resources: their influence on variation in life history tactics. American Naturalist, 128, 137142.
  • Osunkoya, O.O., Sheng, T.K., Mahmud, N.A. & Damit, N. (2007) Variation in wood density, wood water content, stem growth and mortality among twenty-seven tree species in a tropical rainforest on Borneo Island. Austral Ecology, 32, 191201.
  • Paradis, E. (2006) Analysis of Phylogenetics and Evolution with R. Springer, Berlin.
  • Pickup, M., Westoby, M. & Basden, A. (2005) Dry mass costs of deploying leaf area in relation to leaf size. Functional Ecology, 19, 8897.
  • Poorter, L. (2008) The relationships of wood-, gas-, and water fractions of tree stems to performance and life history variation in tropical trees. Annals of Botany, 102, 367375.
  • Poorter, L. & Bongers, F. (2006) Leaf traits are good predictors of plant performance across 53 rain forest species. Ecology, 87, 17331743.
  • Poorter, L., Paz, H., Wright, S.J., Ackerly, D.D., Condit, R., Ibarra-Manríquez, G. et al. (2008) Are functional traits good predictors of demographic rates? Evidence from 5 neotropical forests. Ecology, 89, 19081920.
  • Poorter, L., McDonald, I., Alarcón, A., Fichtler, E., Licona, J.-C., Peña-Claros, M., Sterck, F., Villegas, Z. & Sass-Klaassen, U. (2010) The importance of wood traits and hydraulic conductance for the performance and life history strategies of 42 rainforest tree species. New Phytologist, 185, 481492.
  • Potter, D. (2007) Phylogeny and classification of Rosaceae. Plant Systematics and Evolution, 266, 543.
  • Reich, P.B., Walters, M.B. & Ellsworth, D.S. (1997) From tropics to tundra: global convergence in plant functioning. Proceedings of the National Academy of Sciences, USA, 94, 1373013734.
  • Reich, P.B., Wright, I.J., Cavender-Bares, J., Craine, J.M., Oleksyn, J., Westoby, M. & Walters, M.B. (2003) The evolution of plant functional variation: traits, spectra, and strategies. International Journal of Plant Sciences, 164, 143164.
  • Rivas Martínez, S. (1987) Memoria del mapa de series de vegetación de España. Organismo Autónomo de Parques, Madrid.
  • Russo, S.E., Brown, P., Tan, S. & Davies, S.J. (2008) Interspecific demographic trade-offs and soil-related habitat associations of tree species along resource gradients. Journal of Ecology, 96, 192203.
  • Rydin, C. (2002) Seed plant relationships and the systematic potion of gnetales based on nuclear and chloroplast DNA: conflicting data, rooting problems, and the monophyly of conifers. International Journal of Plant Science, 163, 197214.
  • Sperry, J.S., Meinzer, F.C. & McCulloh, K.A. (2008) Safety and efficiency conflicts in hydraulic architecture: scaling from tissues to trees. Plant, Cell and Environment, 31, 632645.
  • Sterck, F.J., Poorter, L. & Schieving, F. (2006) Leaf traits determine the growth–survival trade-off across rain forest tree species. American Naturalist, 167, 758765.
  • Valladares, F. & Niinemets, U. (2008) Shade tolerance, a key plant feature of complex nature and consequences. Annual Review of Ecology, Evolution and Systematics, 39, 237257.
  • Violle, C., Navas, M.-L., Vile, D., Kazakou, E., Fortunel, C., Hummel, I. & Garnier, E. (2007) Let the concept of trait be functional!. Oikos, 116, 882892.
  • Wang, X.-Q., Tank, D.C. & Sang, T. (2000) Phylogeny and divergence times in Pinaceae: evidence from three genomes. Molecular Biology Evolution, 17, 773781.
  • Webb, C.O. & Donogue, M.J. (2005) Phylomatic: tree assembly for applied phylogenetics. Molecular Ecology Notes, 5, 181183.
  • Westoby, M. & Wright, I.J. (2006) Land-plant ecology on the basis of functional traits. Trends in Ecology and Evolution, 21, 261268.
  • Westoby, M., Falster, D.S., Moles, A., Vesk, P.A. & Wright, I.J. (2002) Plant ecological strategies: some leading dimensions of variation between species. Annual Review of Ecology and Systematics, 33, 125159.
  • Woodward, F.I. (1987) Climate and Plant Distribution. Cambridge University Press, Cambridge, UK.
  • Wright, S.J., Muller-Landau, H.C., Condit, R. & Hubbell, S.P. (2003) Gap-dependent recruitment, realized vital rates, and size distributions of tropical trees. Ecology, 84, 31743185.
  • Wright, I.J., Reich, P.B., Westoby, M., Ackerly, D.D., Baruch, Z., Bongers, F. et al. (2004) The worldwide leaf economics spectrum. Nature, 428, 621827.
  • Wright, I.J., Reich, P.B., Cornelissen, J.H.C., Falster, D.S., Garnier, E., Hikosaka, K. et al. (2005) Assessing the generality of global leaf trait relationships. New Phytologist, 166, 485496.
  • Wright, I.J., Ackerly, D.D., Bongers, F., Harms, K.E., Ibarra-Manriquez, G., Martinez-Ramos, M. et al. (2007) Relationships among ecologically important dimensions of plant trait variation in 7 Neotropical forests. Annals of Botany, 99, 10031015.
  • Zanne, A.E., Westoby, M., Falster, D.S., Ackerly, D.D., Loarie, S.R., Arnold, S.A.J. & Coomes, D.A. (2010) A new understanding of angiosperm wood structure. American Journal of Botany, 97, 207215.
  • Zwieniecki, M.A., Boyce, C.A. & Holbrook, N.M. (2004) Functional design space of single-veined leaves: role of tissue hydraulic properties in constraining leaf size and shape. Annals of Botany, 94, 507513.

Supporting Information

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

Appendix S1. Methods used to measure functional traits as reported in this study.

Table S1. Data sources of the studied functional traits.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

FilenameFormatSizeDescription
JEC_1718_sm_AppS1-tableS1.doc135KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.