• allometry traits;
  • co-occurring stress;
  • drought stress;
  • leaf traits;
  • oxygen stress;
  • root porosity;
  • seed buoyancy;
  • temperate climate


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  1. Our understanding of the generality of plant functional responses to water availability is limited; current field studies use either very rough approximations of water and oxygen availability or only focus on water-stressed ecosystems. Studies that relate species' responses to a surplus of water are limited to controlled experiments.
  2. The aim of this study was to investigate how traits are selected along a gradient of soil moisture, ranging from oxygen-stressed to drought-stressed. We tested 15 traits: eight leaf traits, two root traits, two seed traits and three allometry traits and related their community means to process-based measures of drought stress and oxygen stress for 171 plots in the Netherlands. Because the trait values had been taken from a large database, an independent field survey was carried out to validate the relationships thus derived.
  3. We show that root porosity and seed floating capacity are mostly strongly related, although still moderately, to oxygen and drought stress (R2 = 27% and 42%, respectively). Leaf traits responded weakly to either of the stressors. The field survey yielded similar relationships. Trait combinations were much more closely related to oxygen or drought stress than individual traits, suggesting that there are multiple trait solutions at a given level of water and oxygen stress.
  4. The relatively weak relationships found between traits and water-related stressors contrast with the strong control of other environmental drivers (disturbance, nutrients) on traits and suggest that these strong constraints imposed by other environmental drivers necessitate varied solutions to cope with water availability.


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

A central goal of community ecology is to understand patterns of plant species distribution. Research of the past decade shows that this can successfully be done by studying species' traits that reflect how species are adapted to their environment. These successes were mainly obtained by linking soil nutrients and light availability to suites of traits (Falster & Westoby 2005; Ordoñez et al. 2009), but less so for water-related stressors. Yet, it is particularly important to understand the response of plants to water availability, as climate change is expected to strongly affect precipitation (both intensity and frequency) and evapotranspiration (via changes in temperature and atmospheric CO2 concentration) and hence species distributions.

The importance of water in shaping plant communities is emphasized by several recent global analyses showing the effect of water availability on distributions of plant traits. These studies reveal that, across large water availability gradients, species in dry environments have deep roots (Schenk & Jackson 2002), slightly denser stems (Swenson & Enquist 2007), thick and dense leaves and comparatively more nitrogen per leaf area (Wright et al. 2005) to optimize water use efficiency (Song et al. 2008), all of which are adaptations to increase water uptake and/or to reduce water use. Although these studies provide insight into the role of water in shaping species traits, these studies suffer from two main limitations. Firstly, water availability is estimated only approximately, for example, by mean annual precipitation or potential/actual evapotranspiration or combinations of these, neglecting local hydrology and other factors such as soil water-holding capacity (Schenk & Jackson 2002; Wright et al. 2005; Moles et al. 2009 among others). Secondly, these global studies mainly show the effects of drought on species traits because wetlands and wetland species are absent in these analyses. Regional-level studies tend to apply more accurate measures of soil moisture (such as soil water content), but have mostly been carried out in ecosystems with a (severe) drought period, see, for example, Cornwell & Ackerly (2009), with a notable exception of Ordoñez et al. (2010a,b).

At the other end of the water availability gradient, that is, where there is a surplus of water, plants have to deal with a shortage of oxygen in the soil, which requires a rather different set of adaptations (known as aeration stress or oxygen stress, hereafter called oxygen stress, Voesenek et al. 2006; Niinemets & Valladares 2006). As a consequence, niche segregation takes place because different adaptations are needed for oxygen stress and drought stress (Silvertown et al. 1999). Important adaptations to cope with oxygen stress are the formation of aerenchyma (both in shoots and roots, Jackson & Armstrong 1999), enhanced shoot elongation (Voesenek et al. 2004) and increased anaerobic metabolism. Most studies investigating the effect of oxygen stress on plant traits focus on the physiological mechanisms of adaptation (Engelaar et al. 1993; Blom & Voesenek 1996 among others) and use controlled experiments to test species response to flooding (Visser, Blom & Voesenek 1996; Visser et al. 2000; Mommer, Pons & Visser 2006). In contrast, the number of ecological field studies on trait responses is highly limited (but see van Eck et al. 2007). As a result, the responses of plant traits to field gradients of water availability remain unclear.

The aim of this study is to investigate how plants respond along a gradient of soil water availability ranging from oxygen-stressed (waterlogged) to drought-stressed (groundwater independent) through a suite of commonly measured plant traits. To separate those effects, direct measures of both oxygen and drought stress were used as explanatory variables in this study. The response of species to soil water and soil oxygen availability is tested on ecosystems within the temperate biome because in this climatic zone a broad gradient from drought-stressed to water-stressed can be found.

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

General Approach

We combined species composition data from plots distributed across a broad water availability gradient with a species–trait database to calculate community-level trait means for a variety of traits. We then constructed a (linear) regression model relating these traits (singly and in combination) to site-level measures of drought and oxygen stress. Finally, to independently validate the modelled relationships, we conducted a field survey of community trait means across a water availability gradient within a single dune–valley complex.

Plot Selection

To test the effect of water availability on the functional composition of species assemblages, one needs combined measurements of water availability and species composition across the water gradient. The only data set available that spanned this whole range is from Bartholomeus et al. 2011a; who simulated oxygen and drought stress for the data of Runhaar (1989), De Jong (1997), Jansen et al. (2000), Jansen & Runhaar (2005), Hommel et al. (2007). Altogether, the data set contained information about species composition, soil physical properties and water supply in 171 plots from natural ecosystems throughout the Netherlands. The mean summer temperature in the Netherlands is 16·4 °C, the mean winter temperature is 3·2 °C, the mean annual precipitation is 754 mm (locally varying from 831 to 712 mm) and the precipitation surplus is 191 mm (locally varying from 125 to 250 mm). The precipitation deficit during the growing season (April–August) is 122 mm (Royal Netherlands Meteorological Institute (KNMI); Together, the plots covered a range from dry to wet, nutrient-poor to nutrient-rich and frequently disturbed to undisturbed ecosystems. None of the plots had had experienced any major changes in hydrological conditions in the years before the plots were sampled.

Estimates of Water Availability: Oxygen Stress and Drought Stress

Although water is essential for plant growth, a shortage will limit photosynthesis and plant transpiration, while a surplus limits diffusion of oxygen to roots and hence limits respiration and the metabolic activity of plants. Groundwater levels alone do not fully represent the moisture conditions, and consequently, oxygen and drought stress, in the root zone. Therefore, to test the effects of each component, individual measures of oxygen and drought stressors were applied. We calculated process-based oxygen and drought stress for terrestrial vegetation plots from a variety of natural habitats. We used the reductions in respiration and transpiration because of low oxygen and water availability, respectively, to characterize these stresses. As a result, respiration reduction includes effects of both extreme rainfall events and high temperatures, known to affect vegetation composition (Sojka, Joseph & Stolzy 1972; Drew 1983). Transpiration reduction accounts for the effects of both prolonged dry periods and high atmospheric demand for plant transpiration, that is, factors that determine drought stress of plants (Porporato, Daly & Rodriguez-Iturbe 2004).

For an unbiased comparison of oxygen and drought stress between sites, and following Dyer (2009), we simulated the daily respiration and transpiration reduction for a hypothetical reference vegetation instead of the actual vegetation. By doing so, we obtained stress measures that reflect the soil moisture and oxygen status, independent of the actual vegetation. The use of a reference vegetation improves the applicability of models in which stress measures are implemented (Dyer 2009). By adaptation, the natural vegetation reduces stress. Hence, poor relations between the actual stress and vegetation characteristics are expected, although quantifying actual stress would be useful when aiming at calculating fluxes of, for example, water and carbon. Our reference vegetation is defined as a temperate natural grassland not adapted to oxygen and drought stress, that is, a grassland as defined by Van Dam (2000) and Bartholomeus et al. (2008b).

Oxygen stress, defined as the reduction in root respiration relative to respiration under optimal soil aeration, is a function of soil texture, soil organic matter content, gas-filled porosity and soil temperature (Bartholomeus et al. 2008b, 2011b). In the Netherlands, with its rather shallow groundwater tables, these variables greatly depend on the groundwater table depth, soil type and soil temperature. Daily groundwater levels were calculated from fortnightly series of groundwater levels measured with piezometers located in or closely to the plots ( These observed groundwater levels were first interpolated to daily values and then extrapolated to a 30-year period (1971–2000) with the help of the impulse–response (impulse = precipitation and reference evapotranspiration and response = groundwater levels; climate data were taken from the KNMI) software Menyanthes (Von Asmuth, Bierkens & Maas 2002). As is shown by Bartholomeus et al. (2008a), a 30-year time frame of daily values is needed to obtain a robust estimate of the moisture conditions of a particular plot. These daily values were subsequently used in a detailed soil–water–atmosphere–plant model (SWAP, Van Dam et al. 2008) together with soil texture, air temperature, rainfall and evapotranspiration data to calculate gas-filled porosity and soil temperature for different layers in the soil profile. These variables were used in an additional model to calculate actual and potential root respiration (Bartholomeus et al. 2008b). A difference between potential and actual root respiration under the influence of an oxygen deficiency in the root zone indicates the potential oxygen stress experienced by a plant. The yearly maximum respiration reduction in a 10-day period averaged across 30 years was taken as a measure of oxygen stress (OS–kg O2 m−2 10 day−1). For a detailed description and justification of this procedure, see Bartholomeus et al. (2011a,b)

Drought stress was calculated as the reduction in potential plant transpiration relative to conditions where they transpire at potential rate (sufficient supply of water to meet the demands), as determined by global radiation, air humidity, wind speed, air temperature and atmospheric CO2 concentration. The transpiration is reduced when water becomes limited. This reduction in transpiration was calculated for the same reference vegetation as used for the OS calculations. Analogous to the OS measure, drought stress was defined as the yearly maximum reduction in a 10-day period averaged across 30 years (DS–m 10 day−1). For a detailed description and justification of this procedure, see Bartholomeus et al. (2011a).

Our measures of oxygen and drought stress have been shown to be strongly related to the fractions of xerophytes and hygrophytes in a vegetation plot (Bartholomeus et al. 2011a) and have been used with success in a vegetation model to predict vegetation distribution (Douma et al. (2012c).

Trait Selection

A vascular plant species–trait database was compiled for species in the Netherlands, taken from Douma et al. (2012a) and BioBase (2003), LEDA-traitbase (Kleyer et al. 2008,, Hodgson et al. 2011 and (P.M. van Bodegom, T.D. Colmer, B.K. Sorrell, E.J.W. Visser, P. Adam, W. Armstrong et al. unpublished data). Compiled trait information of wetland species around the globe and included approximately 200 studies and 1200 species (see We selected 15 plant traits which (i) together reflect how species disperse, establish and persist, (ii) can reasonably be assumed to be under selective pressure of water-related stressors, based on ecological theory and previous regional and global analyses (see 'Introduction') and (iii) are commonly measured and available for many species. These 15 traits included eight leaf traits: specific leaf area (SLA in mm2 mg−1), leaf nitrogen content (LNC mg g−1), leaf nitrogen per area (Narea, mg mm−2), leaf density (LD, mg mm−3), leaf volume (LV, mm3), leaf thickness (LT, mm), leaf phosphorus content (LPC in mg g−1) and leaf size (LS, mm2); two root traits: rooting depth (RD in m) and root porosity (RP in%); two seed traits: seed mass of the germinule (SM_g in mg; seed without fruits or detachable appendages) and seed buoyancy (BYC; floating capacity defined as percentage seeds floating after 1 week; see details of this measure in Knevel et al. 2005); and three allometry traits: seedling maximum relative growth rate (RGR in day−1), maximum canopy height (maxCH in m) and specific stem density (SSD in g cm−3). The number of species for which trait data were available is shown in Table 1. Leaf traits based on leaf size or leaf thickness (LS, LT, SLA, Narea, LV and LD) of species with cylindrical leaves were omitted from the analysis. Leaf area measurements of species are mostly estimated by the projected area (Cornelissen et al. 2003), but these estimates may be problematic for species with cylindrical leaves as it does not reflect to the total surface area of that leaf, which is important when related to water availability. Four traits were log10-transformed prior to analysis because their geometrical mean is more closely related to their ecosystem functioning: maxCH, LS, RD and SM_g (Leps et al. 2006). We coupled the species–trait database to the plot–species database to construct a plot–trait database. Plot mean trait values indicate the average response of species to environmental drivers, given that species filtering takes place at the community level (Ackerly & Cornwell 2007). Therefore, plot mean traits were regressed against water and oxygen stress in separate single regressions and their combined effect, including interactions, in a multiple regression. Only those plots that had trait information for at least five species were selected. To obtain reliable parameter estimates, a weighted regression was applied (Quinn & Keough 2002), weighing each plot by the fraction of species that had a trait value.

Table 1. Traits used for the analyses, the number of vascular plant species involved (in total 386) and its sources
TypesTraitScale and unitsTotal number of speciesSources
  1. 1) Douma et al. 2012a; 2) van Bodegom et al. (unpublished data), 3) Hodgson et al. 2011, 4) Kleyer et al. 2008; LEDA-traitbase (

Leaf traitsLeaf nitrogen content (LNC)Continuous (mg g−1)2081,2
Leaf nitrogen per area (Narea)Continuous (g m−2)1771,2
Leaf phosphorus content (LPC)Continuous (mg g−1)1981,2
Specific leaf area (SLA)Continuous (mm2 mg−1)3251,2
Leaf size (LS)Continuous (mm2)2111
Leaf thickness (LT)Continuous (mm)3283
Leaf density (LD)Continuous (mg mm−3)1971,2,3
Leaf volume (LV)Continuous (mm3)1081,2,3
Structural traitsRelative growth rate (RGR)Continuous (day−1)1301
Stem specific density (SSD)Continuous (g cm−3)901
Max. canopy height (maxCH)Continuous (m)3801
Root traitsRooting depth (RD)Continuous (m)2291, 2
Root porosity (RP)Continuous (%)932
Reproduction traitsSeed mass of the germinule (SM_g)Continuous (mg)3591
Seed buoyancy (BYC, proportion of seeds floating after 1 week)Continuous (%)914

Validation of Trait–Water Relationships by a Field Survey

To test the generality of the relationships between OS, DS and trait values and to evaluate potential biases because of the use of trait values from databases (not accounting for trait plasticity or missing trait information for particular species), a field survey was carried out at 10 sites along a water gradient from wet to dry. To avoid other confounding factors such as soil type and soil nutrient status, the validation was carried out within one ecosystem. Dune slack valleys are very suitable for this validation as the water-table varies by almost 2 m (from + 0·3 m + s.s. to −1·7 m + s.s.) within a distance of 25–50 m. They are nutrient poor, and the parent material (sand) does not change along this gradient. In total, two dune–valleys were sampled with five plots per valley that were equally distributed over the water gradient. Within each plot (approximately 1–4 m2), the five most dominant species were selected, assuming these species represent the local conditions best. In total, 48 species samples from 23 different species and 10 different sites were taken.

When available, five individuals of each dominant species were selected for each site. For these species, a selected set of traits was measured: specific leaf area (SLA; m2 kg−1), leaf nitrogen concentration (LNC; mg g−1), leaf size (LS, mm2; including petiole), leaf thickness (mm) and stem specific density (SSD; mg mm−3) were determined at peak biomass (from mid-July until the end of August) and measured following standardized protocols from Cornelissen et al. (2003). In addition, root porosity (RP;%) was determined. For this, non-woody roots attached to the stem of the selected individual were dug up and removed. Samples, including adherent soil particles, were packed in a moist paper bag and inserted in a sealed plastic bag. Samples were transported in moist paper and stored under cooled conditions. After collection, the roots were washed and root porosity was determined on 1–3 g of root material using the pycnometer procedure, as described by Burdick (1989).

Groundwater levels were measured directly adjacent to the plots at three occasions during the growing season. Two nearby located piezometers were used, after accounting for the differences in soil surface height, as reference for the long term groundwater dynamics in that plot. Drought stress and oxygen stress were calculated following the same procedure as described in the section ‘Estimates of water availability: Oxygen stress and drought stress’. Nutrient availability, as approximated by C : N ratio, was homogeneous across the plots (results not shown).

We tested to what extent the slopes and intercepts of the regressions derived from the database and the in situ trait validation differed. This was tested as follows: a dummy variable (0 and 1 for the two data sets, respectively) was included in the regression: Y = a*X + c*group + d*group*X. If the slope of the subset is significantly different from the full set, then the parameter d will be significantly different from zero. Likewise, significant differences in the intercept are shown by c being significantly different from zero. Also here, plots were weighted by the fraction of species that had a trait value. The plots from the field survey were assigned maximal weights (1) because in situ measurements were used. The dummy regression was only performed for oxygen stress as drought stress did not vary sufficiently among the plots for a meaningful analysis. Note that neither the sites, nor the in situ trait measurements were used for the database predictions and therefore the validation is strictly independent of the database predictions.

Finally, to test whether oxygen stress and drought stress lead to the selection of a suite of traits, a multiple regression was run on the database observations with oxygen stress or drought stress as dependent and the traits as predictor. A backward stepwise regression was run to obtain a minimal model with high predictive power (of the models with lowest AIC, the model with highest adjusted R2 was selected). As LNC and LPC were highly correlated (r = 0·87), LPC was omitted in this analysis.


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

Root porosity showed the strongest correlation of all traits to oxygen stress (R2 = 0·24), followed by seed buoyancy (R2 = 0·22, Fig. 1a, Table 2; for full details see Appendix S1, Tables 1 and 2 in the Supporting Information). The response of both traits to drought stress was even stronger (R2 = 0·25 and 0·45, respectively, Fig. 1b, Table 2); root porosity and seed buoyancy increased with increasing oxygen stress and decreased with increasing drought stress. Note that the results for seed buoyancy need to be interpreted with care; see Appendix S2, Supporting information. In addition, rooting depth, relative growth rate and stem specific density consistently showed a significant negative response to oxygen and a positive response to drought stress. Leaf traits showed a much weaker response to both of the stressors. Leaf nitrogen, both on a mass and area basis, showed a negative response to oxygen stress and positive to drought stress (Table 2 and Fig. 1a,b). In contrast, SLA did not respond significantly to oxygen stress and had a weak response to drought stress (marginally significant, P = 0·04). Leaf size was consistently lower at dry sites compared with wet sites, although this relationship was rather weak for oxygen stress and not significant for drought stress. Leaf thickness was positively related to drought stress (a quadratic relationship between leaf thickness and oxygen stress did not perform better than a linear one). Leaf density did not respond significantly to either stressor. Leaf volume showed a weak positive relationship with oxygen stress. Maximum canopy height was not related to either oxygen or drought stress. Seed mass showed a negative, although weak, response to oxygen stress (and the reverse relationship with drought stress).


Figure 1. Relationships between traits and oxygen stress (a) or drought stress (b) from database predictions (grey dots) and independently measured field data as black dots. Significant relationships between the drivers are shown with the estimated regression line for both the traits derived from the database (solid line) as the fieldwork (dashed line). See Table 1 for abbreviations of traits and main text for definition of oxygen and drought stress.

Download figure to PowerPoint

Table 2. Results of regression analysis of traits compiled in our database with oxygen stress and drought stress showing the standardized regression coefficient of the relationship (correlation, the change in SD units of a trait by a change of one SD in the predictor), the explained variance and its significance (See Table 1 for abbreviations of traits and main text for definition of oxygen and drought stress)
TraitOxygen stressDrought stressd.f.
Standardized slope R 2 P Standardized slope R 2 P

The combination of the single effects of oxygen stress and water stress were significant for only two traits: relative growth rate and root porosity. (Appendix S1, Table 3, Supporting information). In contrast, the interaction between oxygen stress and drought stress was significant for many more traits: LNC, LPC, SM_g, RD and LT. This shows that although water and oxygen stress are highly correlated (−0·75), their curvilinear relationship (Appendix S3, Fig. 1, Supporting information) means that the degree of response of species to drought stress depends on the level of oxygen stress, and vice versa. The results suggest that at a given level of one stressor, the response of species from sites with high co-occurring stress, the traits are modified in the direction of the response of the other stressor. For example, a high level of oxygen stress and a co-occurring high level of drought stress lead to higher leaf nitrogen and phosphorus content, heavier seeds, thicker leaves and deeper roots than when examined for oxygen stress alone.

The regression analysis of the in situ trait validation measurements had a much smaller power (because of lower number of observations) and was significant only for the two traits most strongly related to oxygen stress in the database analysis (i.e. root porosity and stem specific density, Table 3). However, in these cases the R2-values of the regression were much higher compared with the R2-values of the database analysis, showing a tight relationship when accounting for potentially confounding factors.

Table 3. Results of regression analysis (standardized regression coefficient, explained variance and significance respectively) of traits that have been measured in the field and oxygen and drought stress (See Table 1 for abbreviations of traits, Table 2 for explanation of standardized regression coefficient and main text for definition of oxygen and drought stress)
TraitOxygen stressd.f.
Standardized slope R 2 P

Comparing the slopes of the regression equations for both datasets showed that the relationships derived from the validation set were not significantly different from the database derived relationships, except for Narea and leaf density (both slope and intercept) and SLA and leaf volume (intercept only), see Appendix S1, Table 4, Supporting information for details. All traits that had been more strongly coupled to oxygen and drought stress (LNC, RP, SSD) showed no significant differences.

A multiple regression to test how traits are collectively selected by oxygen or drought stress showed that a combination of traits can explain the impacts of water and oxygen availability well. The best model (AIC = −647·94, d.f. = 87), consisted of seven traits SSD, LNC, LD, RGR, LS, RP and BYC and explained up to 67% of the variation in oxygen stress (Table 4), compared with 24% for the single trait with the highest predictive power RP. In contrast to what would be expected from the single regressions, not RP but SSD, BYC and LNC were the most important in explaining the variation in oxygen stress. A model without BYC resulted in similar standardized coefficients for the other traits (Appendix S2, Table 1, Supporting information). The variation in drought stress (64%, Table 5) was best explained by BYC, RP, Narea, RD, SLA, SSD, LD and LT and was again much higher than the predictive power of the best single trait (43%, AIC = −1101·79, d.f. = 102). Here, BYC and RP were the most important in explaining drought stress as they were in the single regressions. A model without BYC resulted in similar standardized coefficients for the other traits (Appendix S2, Table 2, Supporting information).

Table 4. Best multiple trait model to predict oxygen stress (see Table 1 for abbreviations of traits). Stepwise backward selection. R2 adj = 0·67, AIC = −647·94, d.f. = 87. Standardized regression coefficients and standardized standard errors are shown (see Table 2 for full explanation of standardized regression coefficient)
TraitStandardized slopeSt. Std. Error P
Table 5. Best multiple trait model to predict drought stress (see Table 1 for abbreviations of traits). Stepwise backward selection. R2 adj = 0·64, AIC = −705·97, d.f. = 86. Standardized regression coefficients and standardized standard errors are shown (see Table 2 for full explanation)
TraitStandardized slopeSt. Std. Error P

As the predictors were measured at different scales, for all regressions standardized regression coefficients were calculated to assess the relative effects (effect sizes) of the different predictors on the response variable (Quinn & Keough 2002). All analyses were performed in R (R Development Core Team 2009).


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

To our knowledge, this is the first field study that simultaneously quantifies the role of oxygen and drought stress on trait selection in species assemblages along a gradient of dry to wet ecosystems. Both the database analysis and the field survey validation consistently showed that across a range of dry to wet ecosystems within the temperate biome the functional characteristics of species assemblages can partly be explained by the combined action of water and oxygen stress. Root porosity and seed buoyancy appeared to be particularly affected by differences in water availability, while leaf traits responded only weakly to either stress measure.

Consistent but Weak Response of Individual Traits to Water and Oxygen Stress

Individual traits showed only a moderately strong response to water and oxygen stress: none of the traits exceeded an explained variance of 45%. Traits that were best explained by water and oxygen stress were traits that are not part of the commonly accepted plant strategy schemes (Westoby 1998; Weiher et al. 1999; Diaz et al. 2004), which mainly reflect adaptations to light, nutrients and disturbance. This shows that coping with water availability requires a specific set of adaptations both during the persistence and dispersal phase (sensu Weiher et al. 1999). Root porosity (one of the traits most strongly related to oxygen stress) is an important adaptation to cope with waterlogged conditions as it enables oxygen transport to the roots. In dry sites, porous roots may be disadvantageous as porous roots are more vulnerable to external compressive soil forces (Striker et al. 2007). Seed buoyancy (adj. R2 42% from the multiple regression) is a specific adaptation to disperse seeds via water to new sites and was found to be higher in wet environments than in dry environments [in concert with findings of van den Broek, Diggelen & Bobbink (2005)]. Paradoxically, species that rely on buoyant seeds as a dispersal mechanism are as a consequence also limited to those locations that are inundated more or less frequently. This may explain the strong relationship between water stress with seed buoyancy and oxygen stress with seed buoyancy, although inundation frequency would be have been a more direct measure for this trait. The increase in buoyancy along the water gradient may also explain the decrease in seed mass with increasing oxygen stress, because seed mass was negatively correlated with seed buoyancy (Pearson correlation = −0·50). The relationships derived from the database were not significantly different from the response found in the field survey (except for Narea and LD). The congruence between the results of the database and the field survey gives confidence in the reliability of database predictions and suggest that they could be safely used to interpret water availability trait relationships.

In general, leaf traits responded weakly to both measures of water supply, which contrasts with the relatively strong relation between the approximate measures of water availability (through precipitation and evapotranspiration) and traits found in global analyses. Still, supporting earlier findings of Wright et al. (2005) on a global scale, our analyses also indicate that species growing under dry conditions generally have a higher leaf nitrogen content (both on an area and mass basis), most likely to optimize their water use efficiency and reduce transpiration (Wright, Reich & Westoby 2003). Previous regional studies (Ackerly 2004; chaparral vegetation and Ordoñez et al. 2010b; temperate vegetation) did not find a relationship between LNC and water availability, probably because their water gradient was narrower. SLA seems to have a more varied response to differences in water availability. In our data set, we did not find a significant relationship, and neither did Ackerly (2004) for chaparral species, while Wright et al. (2005) and Ordoñez et al. (2010b) did find a positive relationship between water availability and SLA. The differential response of SLA to water availability is perhaps because SLA is additionally dependent on other factors, such as soil nutrients (Ordoñez et al. 2009), disturbance (Douma et al. 2012b) and growth form (Ordoñez et al. 2010a). Despite the absence of a relationship between SLA, leaf density and water availability, species at dry sites had smaller but thicker leaves and therefore less voluminous leaves than at sites with higher water availability, but again this effect was only moderate.

The average trait response of plant communities to oxygen and drought stress displayed an inverse relationship, for example, an increase in oxygen stress leads to a decrease in LNC, and an increase in drought stress leads to an increase in LNC. This is probably because both stressors are naturally correlated; sites with a high level of oxygen stress are not likely to suffer from drought stress in the same growing season. Interestingly, the effect of oxygen stress on trait responses is dampened by co-occurring high drought stress. In those cases, plants generally have higher leaf nitrogen and phosphorus contents, heavier seeds, thicker leaves and surprisingly deeper roots than without co-occurring stress. This effect becomes stronger upon rising levels of co-occurring stressors. Consequently, species at sites with co-occurring stresses lack full adaptation to either of the stresses, which may hamper species performance relative to their performance on sites without co-occurring stress (Silvertown et al. 1999; Niinemets & Valladares 2006; Bartholomeus et al. 2011a).

Are We Measuring the Wrong Traits?

This study tested, as a first step, the relationships of commonly measured traits to water and oxygen stress. A surprising result was that most commonly measured plant traits do not appear to be strongly constrained by water and oxygen stress. There may be two explanations for the relatively low explained variance of individual traits by oxygen and drought stress. Firstly, as water availability varies strongly in space and time, species have to adapt on both short [e.g. closing of the stomata (seconds–hours) and long time-scales (e.g. avoiding unfavourable periods (months–years)]. Plants with different life-history strategies may respond differently to variation in soil moisture within a site (Blom & Voesenek 1996; Grubb 1998). As a result, the community mean trait may be only weakly affected by either of the stressors (van Ommen Kloeke et al. 2011). Secondly, we may have measured the wrong traits. Other traits, which are currently not available in databases and which were not measured in the field because of time constraints, are probably more directly related to optimizing water supply and transpiration and oxygen stress and respiration. For example, leaf size/twig size ratio and its whole plant analogue leaf area/sapwood area ratio (Westoby et al. 2002; Martinez-Vilalta et al. 2009) may be a better indicator of transpirational demand and water supply. In addition to traits that are related to root oxygen supply, such as root porosity, traits related to oxygen demand in the root zone may also be strongly constrained by oxygen stress. For example, root respiration is largely determined by root nitrogen (Reich et al. 2008) and lowering root nitrogen may provide an alternative way to deal with reduced levels of oxygen in the soil.

Alternatively, traits relevant for coping with unfavourable periods probably would have been strongly related to water and oxygen stress. For example, Raunkaier (Raunkiaer 1934) showed that the position of the growth bud is a trait demonstrating an adaptation to unfavourable conditions, and as such, it could be an important trait to classify both species and communities.. This view is supported by regressing the percentage of helophytes in a community (flowering parts above water and growth buds submerged) against oxygen stress and therophytes (annuals, surviving the critical period via seeds) against drought stress. This analysis revealed that 51% and 20% of the variation in the helophytes and therophytes could be explained by the level of oxygen stress or drought stress. (for details see Appendix S4, Supporting information).

Are There Multiple Trait Solutions to Cope with Water Shortage or Surplus?

There is a large difference between the rather low explained variance of the single regressions and the high explained variance of the two multiple regressions that predicted oxygen and drought stress by a combination of traits (67% and 64%, respectively). A combination of several traits [BYC and SSD combined with one or two leaf traits (and RP for oxygen stress)] contributed significantly to explaining oxygen and drought stress. This shows that at a given level of oxygen or drought stress, multiple functional solutions seem to prevail across species assemblages. These differences may be caused by differences in other environmental drivers, for example, nutrients or disturbance that also co-determine trait composition of species assemblages. Compared with water, disturbance and nutrient availability have a relatively strong constraining effect on traits. For example, Douma et al. (2012c) have shown for the same temperate climate that the explained variance of SSD by time since disturbance and LPC by soil C/P ratio is 78% and 44%, respectively. In the case of SSD, this implies that, assuming that disturbance and water availability can vary independently, only 22% of the variance remains to be explained by water. Therefore, we hypothesize that because of this strong coordination of traits by disturbance and nutrients (Westoby et al. 2002; Wright et al. 2004), multiple solutions for these traits are needed to cope with a level of oxygen and drought stress. In conclusion, the extent to which multiple solutions are indeed a common way to deal with water availability, or whether these results were because of the type of traits selected during this study, remains to be seen. Future studies would benefit from estimating the relative effects of different environmental drivers on traits to further test this hypothesis.


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

This study was carried out in the framework ‘Kennis voor Klimaat’ theme 3. We thank Flip Witte for his contribution in acquiring the plot database, Han Runhaar and Staatsbosbeheer for providing their valuable data and PWN Hubert Kivit for permission to do fieldwork in the Kennemerduinen. Leaf thickness data were provided by Peter Wilson and other members of the Unit of Comparative Plant Ecology and Department of Archaeology at Sheffield University (see Hodgson et al. 2011). We thank the editor and an anonymous reviewer for their constructive comments. Also thanks to James Weedon for proofreading the manuscript.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • Ackerly, D. (2004) Functional strategies of chaparral shrubs in relation to seasonal water deficit and disturbance. Ecological Monographs, 74, 2544.
  • Ackerly, D.D. & Cornwell, W.K. (2007) A trait-based approach to community assembly: partitioning of species trait values into within- and among-community components. Ecology Letters, 10, 135145.
  • Bartholomeus, R.P., Witte, J.P.M., van Bodegom, P.M. & Aerts, R. (2008a) The need of data harmonization to derive robust empirical relationships between soil conditions and vegetation. Journal of Vegetation Science, 19, 799808.
  • Bartholomeus, R.P., Witte, J.P.M., Van Bodegom, P.M., Van Dam, J.C. & Aerts, R. (2008b) Critical soil conditions for oxygen stress to plant roots: substituting the Feddes-function by a process-based model. Journal of Hydrology, 360, 147165.
  • Bartholomeus, R.P., Witte, J.-P.M., van Bodegom, P.M., van Dam, J.C. & Aerts, R. (2011a) Climate change threatens endangered plant species by stronger and interacting water-related stresses. Journal of Geophysical Research, 116, G04023.
  • Bartholomeus, R.P., Witte, J.-P.M., van Bodegom, P.M., van Dam, J.C., de Becker, P. & Aerts, R. (2011b) Process-based proxy of oxygen stress surpasses indirect ones in predicting vegetation characteristics. Ecohydrology, DOI: 10.1002/eco.261.
  • Blom, C. & Voesenek, L. (1996) Flooding: the survival strategies of plants. Trends in Ecology & Evolution, 11, 290295.
  • van den Broek, T., van Diggelen, R. & Bobbink, R. (2005) Variation in seed buoyancy of species in wetland ecosystems with different flooding dynamics. Journal of Vegetation Science, 16, 579586.
  • Burdick, D.M. (1989) Root aerenchyma development in spartina patens in response to flooding. American Journal of Botany, 76, 777780.
  • Cornelissen, J.H.C., Lavorel, S., Garnier, E., Diaz, S., Buchmann, N., Gurvich, D.E., Reich, P.B., ter Steege, H., Morgan, H.D., van der Heijden, M.G.A., Pausas, J.G. & Poorter, H. (2003) A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Australian Journal of Botany, 51, 335380.
  • Cornwell, W.K. & Ackerly, D.D. (2009) Community assembly and shifts in plant trait distributions across an environmental gradient in coastal California. Ecological Monographs, 79, 109126.
  • De Jong, A. (1997) IJking van de vochttoestand van de bodem met de vochtindicatie van de vegetatie, volgens het ecotopensysteem [Calibration of the moisture conditions of the soil with the moisture indication of the vegetation according to the ecotope system]. Landbouw Universiteit Wageningen, Wageningen.
  • Diaz, S., Hodgson, J.G., Thompson, K., Cabido, M., Cornelissen, J.H.C., Jalili, A., Montserrat-Marti, G., Grime, J.P., Zarrinkamar, F., Asri, Y., Band, S.R., Basconcelo, S., Castro-Diez, P., Funes, G., Hamzehee, B., Khoshnevi, M., Perez-Harguindeguy, N., Perez-Rontome, M.C., Shirvany, F.A., Vendramini, F., Yazdani, S., Abbas-Azimi, R., Bogaard, A., Boustani, S., Charles, M., Dehghan, M., de Torres-Espuny, L., Falczuk, V., Guerrero-Campo, J., Hynd, A., Jones, G., Kowsary, E., Kazemi-Saeed, F., Maestro-Martinez, M., Romo-Diez, A., Shaw, S., Siavash, B., Villar-Salvador, P. & Zak, M.R. (2004) The plant traits that drive ecosystems: evidence from three continents. Journal of Vegetation Science, 15, 295304.
  • Douma, J.C., Aerts, R., J.P.M, W., Bekker, R.M., Kunzmann, D., Metselaar, K. & Van Bodegom, P.M. (2012a) A combination of functionally different plant traits provides a means to quantitatively predict a broad range of species assemblages in NW Europe. Ecograpy, 35, 364373.
  • Douma, J.C., Shipley, B., Witte, J.P.M., Aerts, R. & van Bodegom, P.M. (2012b) Disturbance and resource availability act differently on the same suite of plant traits: revisiting assembly hypotheses. Ecology, 93, 825835.
  • Douma, J.C., Witte, J.P.M., Aerts, R., Bartholomeus, R.P., Ordoñez, J.C., Olde Venterink, H., Wassen, M.J. & van Bodegom, P.M. (2012c) Towards a functional basis for predicting vegetation patterns; incorporating plant traits in habitat distribution models. Ecography, 35, 294305.
  • Drew, M.C. (1983) Plant injury and adaptation to oxygen deficiency in the root environment: a review. Plant and Soil, 75, 179199.
  • Dyer, J. (2009) Assessing topographic patterns in moisture use and stress using a water balance approach. Landscape Ecology, 24, 391403.
  • Engelaar, W.M.H.G., Vanbruggen, M.W., Vandenhoek, W.P.M., Huyser, M.A.H. & Blom, C.W.P.M. (1993) Root porosities and radial oxygen losses of Rumex and Plantago species as influenced by soil pore diameter and soil aeration. New Phytologist, 125, 565574.
  • Falster, D.S. & Westoby, M. (2005) Alternative height strategies among 45 dicot rain forest species from tropical Queensland, Australia. Journal of Ecology, 93, 521535.
  • Grubb, P.J. (1998) A reassessment of the strategies of plants which cope with shortages of resources. Perspectives in Plant Ecology, Evolution and Systematics, 1, 331.
  • Hodgson, J.G., Montserrat-Martí, G., Charles, M., Jones, G., Wilson, P., Shipley, B., Sharafi, M., Cerabolini, B.E.L., Cornelissen, J.H.C., Band, S.R., Bogard, A., Castro-Díez, P., Guerrero-Campo, J., Palmer, C., Pérez-Rontomé, M.C., Carter, G., Hynd, A., Romo-Díez, A., de Torres Espuny, L. & Royo Pla, F. (2011) Is leaf dry matter content a better predictor of soil fertility than specific leaf area? Annals of Botany, 108, 13371345.
  • Hommel, P.W.F.M., Arts, G.H.P., Beets, C.P., Schipper, P.C., Smolders, A.J.P. & De Waal, R.W. (2007) Naar een landelijke referentiesysteem voor standplaatsbeschrijvingen van aquatische en grondwaterafhankelijke, terrestrische plantengemeenschappen [To a national reference system for site descriptions of aquatic and groundwater dependent terrestrial vegetation types]. Stratiotes, 35, 1121.
  • Jackson, M.B. & Armstrong, W. (1999) Formation of aerenchyma and the processes of plant ventilation in relation to soil flooding and submergence. Plant Biology, 1, 274287.
  • Jansen, P.C. & Runhaar, J. (2005) Toetsing van het verband tussen het aandeel xerofyten en de droogtestress onder verschillende omstandigheden [Testing of the relationship between the covering of xerophytes and drought stress under different conditions]. Alterra, Wageningen.
  • Jansen, P.C., Runhaar, J., Witte, J.P.M. & van Dam, J.C. (2000) Vochtindicatie van grasvegetaties in relatie tot de vochttoestand van de bodem [Moisture indication of grass vegetations in relation to the moisture conditions of the soil]. Alterra, Wageningen.
  • Kleyer, M., Bekker, R.M., Knevel, I.C., Bakker, J.P., Thompson, K., Sonnenschein, M., Poschlod, P., van Groenendael, J.M., Klimes, L., Klimesova, J., Klotz, S., Rusch, G.M., Hermy, M., Adriaens, D., Boedeltje, G., Bossuyt, B., Dannemann, A., Endels, P., Gotzenberger, L., Hodgson, J.G., Jackel, A.K., Kuhn, I., Kunzmann, D., Ozinga, W.A., Romermann, C., Stadler, M., Schlegelmilch, J., Steendam, H.J., Tackenberg, O., Wilmann, B., Cornelissen, J.H.C., Eriksson, O., Garnier, E. & Peco, B. (2008) The LEDA Traitbase: a database of life-history traits of the Northwest European flora. Journal of Ecology, 96, 12661274.
  • Knevel, I.C., Bekker, R.M., Kunzmann, D., Stadler, M. & Thompson, K. (2005) The LEDA Traitbase Collecting and Measuring Standards of Life-History Traits of the Northwest European Flora. Scholma Druk. B.V., Bedum, The Netherlands.
  • Leps, J., de Bello, F., Lavorel, S. & Berman, S. (2006) Quantifying and interpreting functional diversity of natural communities: practical considerations matter. Preslia, 78, 481501.
  • Martinez-Vilalta, J., Cochard, H., Mencuccini, M., Sterck, F., Herrero, A., Korhonen, J.F.J., Llorens, P., Nikinmaa, E., Nole, A., Poyatos, R., Ripullone, F., Sass-Klaassen, U. & Zweifel, R. (2009) Hydraulic adjustment of Scots pine across Europe. New Phytologist, 184, 353364.
  • Moles, A.T., Warton, D.I., Warman, L., Swenson, N.G., Laffan, S.W., Zanne, A.E., Pitman, A., Hemmings, F.A. & Leishman, M.R. (2009) Global patterns in plant height. Journal of Ecology, 97, 923932.
  • Mommer, L., Pons, T.L. & Visser, E.J.W. (2006) Photosynthetic consequences of phenotypic plasticity in response to submergence: Rumex palustris as a case study. Journal of Experimental Botany, 57, 283290.
  • Niinemets, U. & Valladares, F. (2006) Tolerance to shade, drought, and waterlogging of temperate Northern Hemisphere trees and shrubs. Ecological Monographs, 76, 521547.
  • van Ommen Kloeke, A.E.E., Douma, J.C., Ordoñez, J.C., Reich, P.B. & van Bodegom, P.M. (2011) Global quantification of contrasting leaf life span strategies for deciduous and evergreen species in response to environmental conditions. Global Ecology and Biogeography, 21, 224235.
  • Ordoñez, J.C., van Bodegom, P.M., Witte, J.P.M., Wright, I.J., Reich, P.B. & Aerts, R. (2009) A global study of relationships between leaf traits, climate and soil measures of nutrient fertility. Global Ecology and Biogeography, 18, 137149.
  • Ordoñez, J.C., van Bodegom, P.M., Witte, J.P.M., Bartholomeus, R.P., van Hal, J.R. & Aerts, R. (2010a) Leaf habit and stem investments regulate different leaf economy traits at given nutrient supply. Ecology, 91, 32183228.
  • Ordoñez, J.C., van Bodegom, P.M., Witte, J.P.M., Bartholomeus, R.P., van Hal, J.R. & Aerts, R. (2010b) Plant strategies in relation to resource supply in mesic to wet environments: does theory mirror nature? American Naturalist, 175, 225239.
  • Porporato, A., Daly, E. & Rodriguez-Iturbe, I. (2004) Soil water balance and ecosystem response to climate change. American Naturalist, 164, 625632.
  • Quinn, G.P. & Keough, M.J. (2002) Experimental Design and Data Analysis for Biologists, 6th edn. Cambridge University Press, Cambridge.
  • R Development Core Team (2009) R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL
  • Raunkiaer, C. (1934) The life forms of plants and statistical plant geography, being the collected papers of C. Raunkiær. History of Ecology Series (ed. F. N. Egerton), pp. 632. Oxford University Press, Oxford.
  • Reich, P.B., Tjoelker, M.G., Pregitzer, K.S., Wright, I.J., Oleksyn, J. & Machado, J.L. (2008) Scaling of respiration to nitrogen in leaves, stems and roots of higher land plants. Ecology Letters, 11, 793801.
  • Runhaar, J. (1989) Toetsing van het ecotopensysteem [Testing of the ecotope system]. Rijksuniversiteit Leiden, Leiden.
  • Schenk, H.J. & Jackson, R.B. (2002) Rooting depths, lateral root spreads and below-ground/above-ground allometries of plants in water-limited ecosystems. Journal of Ecology, 90, 480494.
  • Silvertown, J., Dodd, M.E., Gowing, D.J.G. & Mountford, J.O. (1999) Hydrologically defined niches reveal a basis for species richness in plant communities. Nature, 400, 6163.
  • Sojka, R.E., Joseph, H.A. & Stolzy, L.H. (1972) Wheat response to short-term heat stress and to soil oxygen stress at three stages of growth. Agronomy Journal, 64, 450452.
  • Song, M.H., Duan, D.Y., Chen, H., Hu, Q.W., Zhang, F., Xu, X.L., Tian, Y.Q., Ouyang, H. & Peng, C.H. (2008) Leaf delta(13)C reflects ecosystem patterns and responses of alpine plants to the environments on the Tibetan Plateau. Ecography, 31, 499508.
  • Striker, G.G., Insausti, P., Grimoldi, A.A. & Vega, A.S. (2007) Trade-off between root porosity and mechanical strength in species with different types of aerenchyma. Plant Cell and Environment, 30, 580589.
  • Swenson, N.G. & Enquist, B.J. (2007) Ecological and evolutionary determinants of a key plant functional trait: wood density and its community-wide variation across latitude and elevation. American Journal of Botany, 94, 451459.
  • Van Dam, J.C. (2000) Field-Scale Water Flow and Solute Transport. Wageningen University, Wageningen.
  • Van Dam, J., Groenendijk, P., Hendriks, R. & Kroes, J. (2008) Advances of modeling water flow in variably saturated soils with SWAP. Vadose Zone Journal, 64, 0653.
  • van Eck, W.H.J.M., van de Steeg, H.M., Blom, C.W.P.M. & de kroon, H. (2004) Is tolerance to summer flooding correlated with distribution patterns in river floodplains? A comparative study of 20 terrestrial grassland species. Oikos, 107, 393405.
  • Visser, E.J.W., Blom, C. & Voesenek, L. (1996) Flooding-induced adventitious rooting in Rumex: morphology and development in an ecological perspective. Acta Botanica Neerlandica, 45, 1728.
  • Visser, E.J.W., Bogemann, G.M., Van de Steeg, H.M., Pierik, R. & Blom, C. (2000) Flooding tolerance of Carex species in relation to field distribution and aerenchyma formation. New Phytologist, 148, 93103.
  • Voesenek, L., Rijnders, J., Peeters, A.J.M., Van de Steeg, H.M.V. & De Kroon, H. (2004) Plant hormones regulate fast shoot elongation under water: from genes to communities. Ecology, 85, 1627.
  • Voesenek, L., Colmer, T.D., Pierik, R., Millenaar, F.F. & Peeters, A.J.M. (2006) How plants cope with complete submergence. New Phytologist, 170, 213226.
  • Von Asmuth, J.R., Bierkens, M.F.P. & Maas, C. (2002) Transer function noise modeling in continuous time using predefined impulse response functions. Water Resources Research, 38, 23012312.
  • Weiher, E., van der Werf, A., Thompson, K., Roderick, M., Garnier, E. & Eriksson, O. (1999) Challenging Theophrastus: a common core list of plant traits for functional ecology. Journal of Vegetation Science, 10, 609620.
  • Westoby, M. (1998) A leaf-height-seed (LHS) plant ecology strategy scheme. Plant and Soil, 199, 213227.
  • Westoby, M., Falster, D.S., Moles, A.T., 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.
  • Wright, I.J., Reich, P.B. & Westoby, M. (2003) Least-cost input mixtures of water and nitrogen for photosynthesis. American Naturalist, 161, 98111.
  • Wright, I.J., Reich, P.B., Westoby, M., Ackerly, D.D., Baruch, Z., Bongers, F., Cavender-Bares, J., Chapin, T., Cornelissen, J.H.C., Diemer, M., Flexas, J., Garnier, E., Groom, P.K., Gulias, J., Hikosaka, K., Lamont, B.B., Lee, T., Lee, W., Lusk, C., Midgley, J.J., Navas, M.L., Niinemets, U., Oleksyn, J., Osada, N., Poorter, H., Poot, P., Prior, L., Pyankov, V.I., Roumet, C., Thomas, S.C., Tjoelker, M.G., Veneklaas, E.J. & Villar, R. (2004) The worldwide leaf economics spectrum. Nature, 428, 821827.
  • Wright, I.J., Reich, P.B., Cornelissen, J.H.C., Falster, D.S., Groom, P.K., Hikosaka, K., Lee, W., Lusk, C.H., Niinemets, U., Oleksyn, J., Osada, N., Poorter, H., Warton, D.I. & Westoby, M. (2005) Modulation of leaf economic traits and trait relationships by climate. Global Ecology and Biogeography, 14, 411421.

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

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.

fec2054-sup-0002-AppendixS1-S4.docxWord document223K

Appendix S1. Details of regression analyses of traits and oxygen and drought stress.

Appendix S2. Interpretation of seed buoyancy impacts.

Appendix S3. Relationship between oxygen stress and drought stress.

Appendix S4. Relationship between oxygen and drought stress and the percentage helophytes and therophytes in a community.

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.