Effects of plant species traits on ecosystem processes: experiments in the Patagonian steppe.
ABSTRACT Several experiments have shown that aboveground net primary productivity increases with plant species richness. The main mechanism proposed to explain this relationship is niche complementarity, which is determined by differences in plant traits that affect resource use. We combined field and laboratory experiments using the most abundant species of the Patagonian steppe to identify which are the traits that determine niche complementarity in this ecosystem. We estimated traits that affect carbon, water, microclimate, and nitrogen dynamics. The most important traits distinguishing among species, from the standpoint of their effects on ecosystem functioning, were potential soil nitrification, rooting depth, and soil thermal amplitude. Additionally, we explored the relationship between trait diversity and aboveground net primary production (ANPP) using a manipulative field experiment. ANPP and the fraction of ANPP accounted for by trait diversity increased with number of traits. The effect of trait diversity decreased as the number of traits increased. Here, the use of traits gave us a mechanistic understanding of niche complementarity in the Patagonian steppe.
- [Show abstract] [Hide abstract]
ABSTRACT: Fungal plant pathogens are common in natural communities where they affect plant physiology, plant survival and biomass production. Conversely, pathogen transmission and infection may be regulated by plant-community characteristics such as plant-species diversity and functional composition that favor pathogen diversity through increases in host diversity while simultaneously reducing pathogen infection via increased variability in host density and spatial heterogeneity. Therefore, a comprehensive understanding of multi-host-multi-pathogen interactions is of high significance in the context of biodiversity-ecosystem functioning. We investigated the relationship between plant diversity and aboveground obligate parasitic fungal pathogen ("pathogens" hereafter) diversity and infection in grasslands of a long-term, large-scale biodiversity experiment with varying plant-species (1-60 species) and plant functional group diversity (1-4 groups). To estimate pathogen infection of the plant communities, we visually assessed pathogen-group presence (i.e., rusts, powdery mildews, downy mildews, smuts, and leaf-spot diseases) and overall infection levels (combining incidence and severity of each pathogen group) in 82 experimental plots on all aboveground organs of all plant species per plot during four surveys in 2006. Pathogen diversity, assessed as the cumulative number of pathogen groups on all plant species per plot, increased log-linearly with plant-species diversity. However, pathogen incidence and severity and hence overall infection decreased with increasing plant-species diversity. In addition, co-infection of plant individuals by two or more pathogen groups was less likely with increasing plant-community diversity. We conclude that plant-community diversity promotes pathogen-community diversity while at the same time reducing pathogen infection-levels of plant individuals.Ecology 01/2014; · 5.18 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Resource partitioning, facilitation, and sampling effect are the three mechanisms behind the biodiversity effect, which is depicted usually as the effect of plant-species richness on aboveground net primary production. These mechanisms operate simultaneously but their relative importance and interactions are difficult to unravel experimentally. Thus, niche differentiation and facilitation have been lumped together and separated from the sampling effect. Here, we propose three hypotheses about interactions among the three mechanisms and test them using a simulation model. The model simulated water movement through soil and vegetation, and net primary production mimicking the Patagonian steppe. Using the model, we created grass and shrub monocultures and mixtures, controlled root overlap and grass water-use efficiency (WUE) to simulate gradients of biodiversity, resource partitioning and facilitation. The presence of shrubs facilitated grass growth by increasing its WUE and in turn increased the sampling effect, whereas root overlap (resource partitioning) had, on average, no effect on sampling effect. Interestingly, resource partitioning and facilitation interacted so the effect of facilitation on sampling effect decreased as resource partitioning increased. Sampling effect was enhanced by the difference between the two functional groups in their efficiency in using resources. Morphological and physiological differences make one group outperform the other; once these differences were established further differences did not enhance the sampling effect. In addition, grass WUE and root overlap positively influence the biodiversity effect but showed no interactions.Oecologia 09/2013; · 3.01 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Ecologists recognize that plants capture nitrogen in many chemical forms that include amino acids. Access to multiple nitrogen types in plant communities has been argued to enhance plant performance, access to nitrogen and alter ecological interactions in ways that may promote species coexistence. However, data supporting these arguments have been limited. While it is known that plants uptake amino acids from soil, long term studies that link amino acid uptake to measures of plant performance and potential reproductive effort are not typically performed. Here, a series of experiments that link uptake of nitrate, glutamine or asparagine with lifetime reproductive effort in Arabidopsis thaliana are reported. Nitrogen was offered either singly or in mixture and at a variety of combinations. Traits related to reproductive output were measured, as was the preference for each type of nitrogen. When plants were supplied with a single nitrogen type at concentrations from 0.1-0.9 mM, the ranking of nitrogen types was nitrate > glutamine > asparagine in terms of the relative performance of plants. When plants were supplied with two types of nitrogen in mixture at ratios between 0.1:0.9-0.9:0.1 mM, again plants performed best when nitrate was present, and poorly when amino acids were mixed. Additionally, stable isotopes revealed that plants preferentially captured nitrogen types matching the hierarchy of nitrate > glutamine > asparagine. Comparing between the two experiments revealed that mixed nitrogen nutrition was a net cost to the plants. Plant performance on mixed nitrogen was less than half the performance on equal amounts of any single nitrogen type. We asked: why did A. thaliana capture amino acids when doing so resulted in a net cost? We argue that available data cannot yet answer this question, but hypothesize that access to lower quality forms of nitrogen may become important when plants compete.BMC Ecology 07/2013; 13(1):28.
Ecology, 93(2), 2012, pp. 227–234
? 2012 by the Ecological Society of America
Effects of plant species traits on ecosystem processes:
experiments in the Patagonian steppe
PEDRO FLOMBAUM1,3AND OSVALDO E. SALA2
1Centro de Investigaciones del Mar y la Atmo ´sfera, Departamento de Ciencias de la Atmo ´sfera y los Oce´anos,
and Instituto Franco-Argentino sobre Estudios de Clima y sus Impactos, CONICET/FCEN-UBA/UMI, Pabello´n II Piso 2,
Ciudad Universitaria, C1428EGA Buenos Aires, Argentina
2School of Life Sciences and School of Sustainability, Arizona State University, Tempe, Arizona 85287-4501 USA
increases with plant species richness. The main mechanism proposed to explain this
relationship is niche complementarity, which is determined by differences in plant traits that
affect resource use. We combined field and laboratory experiments using the most abundant
species of the Patagonian steppe to identify which are the traits that determine niche
complementarity in this ecosystem. We estimated traits that affect carbon, water,
microclimate, and nitrogen dynamics. The most important traits distinguishing among
species, from the standpoint of their effects on ecosystem functioning, were potential soil
nitrification, rooting depth, and soil thermal amplitude. Additionally, we explored the
relationship between trait diversity and aboveground net primary production (ANPP) using a
manipulative field experiment. ANPP and the fraction of ANPP accounted for by trait
diversity increased with number of traits. The effect of trait diversity decreased as the number
of traits increased. Here, the use of traits gave us a mechanistic understanding of niche
complementarity in the Patagonian steppe.
Several experiments have shown that aboveground net primary productivity
nitrification; soil thermal amplitude; trait diversity.
aboveground net primary production (ANPP); niche complementarity; rooting depth; soil
Ecologists’ interest in understanding the role of
biodiversity on ecosystem functioning has increased
since they realized that global biodiversity was being
drastically altered by anthropogenic activities and that it
was urgent to address its consequences. The central
hypothesis of the effect of biodiversity on ecosystem
functioning states that decreases in species richness
would result in decreases in rates of ecosystem processes
(Vitousek and Hooper 1993). The first controlled tests of
this hypothesis were based on experimental plant species
richness gradients, which were created in artificial
ecosystems by sowing different numbers and combina-
tions of species (Tilman et al. 1997a, Hector et al. 1999).
These experiments yielded consistent patterns that as
number of species increased the rate of net primary
production increased up to a point when ecosystem
response saturated and that niche complementarity
accounted for most observed effects of species richness
on ecosystem functioning (Loreau and Hector 2001).
Moreover, an experiment in a natural ecosystem in
Patagonia showed a biodiversity effect that was signif-
icantly stronger than previously reported from experi-
ments with synthetic communities and identified niche
complementarity as the dominant mechanism account-
ing for the increase in production with species richness
(Flombaum and Sala 2008). The effect of niche
complementarity on ecosystem functioning results from
larger amounts of resources used in ecosystems contain-
ing species with different ecological niches (Tilman et al.
1997b). The ecological niche of a species depends on the
collection of its traits, and these are shaped by
evolutionary processes that maximize species fitness
(Chase and Leibold 2003). Examples of species differ-
entiation in traits occur as a result of long-term species
interactions with each other (Brown and Wilson 1956),
or as a result of local adaptation that occur in coexisting
populations of different species exposed to local
environmental conditions (Kawecki and Ebert 2004).
Niche complementarity has been related to resource use
but could also occur via other dimensions of the
ecological niche; here we will refer to resource-use niche
complementarity. Other factors that may account for
the observed effects of species richness on ecosystem
functioning are positive interactions and sampling effect
(Huston 1997, Tilman et al. 1997b).
Although niche complementary is one of the major
mechanisms accounting for the effect of biodiversity on
ecosystem functioning, it is still unclear how niche
complementarity occurs, what characteristics of species
result in niche complementarity, and ultimately, what
species traits affect the functioning of ecosystems. The
Manuscript received 24 April 2011; revised 27 September
2011; accepted 29 September 2011. Corresponding Editor: M. J.
link between niche complementarity and ecosystem
functioning has been established through analytical
tools (Loreau and Hector 2001, Fargione et al. 2007,
Flombaum and Sala 2008) and, just recently, using
direct estimates of niche complementarity (Cardinale
2011). Here, we specifically addressed the following
questions. (1) How does niche complementarity in the
Patagonian steppe occur? (2) What are the species traits
that account for most of the observed niche comple-
mentarity? And finally, since species affect ecosystem
processes through their traits, (3) how does trait
diversity affect ecosystem functioning?
Here, we report on a study of plant species traits and
their effect on ecosystem functioning. We used the
Patagonian steppe as a model ecosystem, because its low
natural diversity makes it relatively simple to study, and
because the relationship between plant species richness
and biodiversity has been shown to be driven by species
niche complementarity (Flombaum and Sala 2008).
MATERIALS AND METHODS
Our study site is located in the Patagonian steppe,
Southern South America (see Plate 1). We worked in
INTA Rı´o Mayo, Argentina (458410S 708160W), where
the climate is semiarid with 170 mm of annual
precipitation, which occurs mostly during fall and
winter. Average monthly temperatures range from 28C
to 148C for winter and summer, respectively. Soils are
coarse textured with gravel and stones. Vegetation
corresponds to the Occidental floristic district charac-
terized by two life forms, grasses and shrubs, and
relatively low species richness. Three grass species (Stipa
speciosa Trin et Ruprecht, Stipa humilis Cav., and Poa
ligularis Nees ap. Steud) and three shrub species
(Mulinum spinosum (Cav.) Pers., Adesmia volckmanni
(Philippi), and Senecio filaginoides DC) account for 97%
of aboveground net primary productivity (Flombaum
and Sala 2008).
Plant species traits that affect ecosystem functioning
For all the dominant plant species of the Patagonian
steppe, we evaluated relative growth rate, potential soil
respiration, rooting depth, plant phenology, soil thermal
amplitude, potential soil net nitrification and net
ammonification, and nitrogen-form relative preference,
which are plant traits that influence carbon, water,
microclimate, and nitrogen dynamics in an ecosystem
(see Appendix A for a detailed description of methods).
We assessed species relative growth rate to investigate
how species could influence net primary production, and
potential soil respiration as a pathway by which species
can influence soil carbon release to the atmosphere.
Carbon fixation and release are processes related to the
energy flow and, in arid ecosystems, these processes are
closely related to water availability (Yahdjian et al.
2006). We considered plant species influence on water
flow from the soil to the atmosphere by addressing
rooting depth, which gives an indication of a plant
capacity to draw down soil moisture with depth (Eviner
and Chapin 2003), and plant annual phenological cycles,
which have important effects on soil seasonal water
demand and carbon sequestration (Eviner and Chapin
2003). Plant species effects on soil microclimate can have
a large effect on belowground biological activity. We
evaluated the thermal amplitude below different species,
which can alter the biological activity rate (Smith et al.
2003). We also evaluated plant species effects on
nitrogen mineralization, the main pathway by which
organic nitrogen becomes available to plants and soil
microorganisms, and ammonium and nitrate relative
absorption by plants because of their differential
distribution in the soil profile. The importance of traits
in affecting the relationship between trait diversity and
ecosystem functioning depends on the distribution of
traits among species. If all species in a community have a
similar level for a trait, that trait will explain little of the
effect of biodiversity. On the contrary, traits that have
different levels for different species should have a large
contribution to the effect of biodiversity on ecosystem
functioning. To compare trait distribution among
species, we made traits relative to minimum and
maximum levels and therefore they ranged between 0
and 1. We divided this scale into six levels, and used H0
(modified Shannon index) to examine the frequency of
species occurring in each level.
We used a plant species diversity gradient experimen-
tally created in Patagonia (Flombaum and Sala 2008) to
define various trait diversity gradients. The gradient of
plant richness had 1, 2, 4, and 6 species with all 6, 15, 15,
and 1 possible assemblages (37 in total) using the same six
species mentioned in Materials and methods: Study site.
We generated a species richness gradient by removing
plant species, and we equalized total plant cover on the
plots by removing variable portions of individuals that
were not targeted for removal (Flombaum and Sala
2008). In this way, our removal disturbance was the same
among all plots. This approach resulted in experimental
plots having the same vegetation cover along the
gradient, but differing in species number and assemblages
(Flombaum and Sala 2008).
We redefined the plant species richness gradient into
various trait-richness gradients. Here, it is important to
distinguish between traits, which are characters, and
levels, which refer to specific values of traits. For
example, rooting depth is a trait, and 40 cm deep is
the level of S. speciosa for rooting depth. Species were
considered to have different levels if their means were
different. For example, the trait plant phenology had
species in three levels corresponding to the means 5, 7,
and 12 months. We used single traits and also
combinations of traits to define various trait gradients.
We evaluated 255 trait richness gradients that resulted
from combining one to eight traits. For example, we
PEDRO FLOMBAUM AND OSVALDO E. SALA228Ecology, Vol. 93, No. 2
evaluated a richness gradient of plant phenology,
another richness gradient of plant phenology combined
with potential soil respiration, another for all traits
combined, and so on.
With the list of species assemblages (from the species
diversity gradient), and each species assigned to a level
(based on its mean), we redefined the species richness
gradient into trait richness gradients. For each species
assemblage, we counted the number of levels for a trait.
For example, the two-species assemblage of S. speciosa
and P. ligularis counted as one level of plant phenology
(since both species had 12 months of green leaves); while
the two-species assemblage of S. speciosa and M.
spinosum counted as two levels of plant phenology
(since species had 12 and 7 months of green leaves;
Appendix B). In trait richness gradients that combined
more than two traits, we counted the total number of
levels per assemblage; for example, in the gradient that
combined plant phenology and rooting depth, the two-
species assemblage of S. speciosa and P. ligularis
counted as three levels (Appendix B). In all trait richness
gradients, one-species plots (monocultures) had the
lowest level of trait richness, and the mixture with
species richness of six (the highest species richness) had
the highest combination of trait levels; combinations
with two and four species could be closer to monocul-
tures or to the six species mixture.
In each experimental plot, we estimated aboveground
net primary production (ANPP) with four, 5 m long,
equally spaced, parallel lines. We recorded vegetation
cover on the lines and used a nondestructive method
that correlated cover and biomass to estimate ANPP
(Flombaum and Sala 2007). ANPP was estimated as the
increase in biomass from early spring (September 2002)
To compare differences in species traits, we conducted
one-way analyses of variance (ANOVA) and a posteriori
Fisher’s LSD contrast, with the exception of phenology
and rooting depth, for which we considered that two
species differed in rooting depth if the difference was
more than 30 cm and that species differed in plant
phenology if the difference in the extent of the period
with green leaves was more than three months. Also, we
performed a priori contrasts to compare between grasses
and shrubs. Using the species average for each trait, we
performed trait correlations and a principal components
analysis (PCA). Last, to assess the effect of trait
diversity on ANPP we estimated the portion of variance
explained (R2) by trait diversity, and estimated the best
fit between R2and the number of traits considered in the
Plant species traits that affect ecosystem functioning
We found many differences between life forms and
among species for the traits analyzed. Regarding traits
that influence carbon dynamics in the ecosystem, grasses
and shrubs had similar values for relative growth rate
and differed in potential soil respiration (Fig. 1). Shrubs
had 1.6 times larger potential soil respiration rates than
grasses (Fig. 1), which could reflect higher organic
matter accumulated underneath shrub canopies. The
study of species effects revealed that relative growth rate
was similar for all species except for A. volckmanni,
which was higher (Fig. 1). S. filaginoides showed the
highest potential soil respiration while P. ligularis and S.
speciosa had the lowest, which were similar to bare soil
(5.6 6 0.6 lg C?[g dry soil]?1?d?1). M. spinosum, A.
volckmanni, and S. humilis fell in the middle (Fig. 1).
Differences in potential soil respiration reflect differenc-
es in litter quality, nutrient availability, and/or differ-
ences in microbial species composition.
Rooting depth and plant phenology, traits that
influenced water dynamics in the ecosystem, were
different between life forms, with more variability
among species for shrubs than for grasses (Fig. 1). A
gradient existed in rooting depth from grasses to shrubs,
with a more than fourfold difference between the
shallowest (S. humilis) and the deepest (S. filaginoides)
rooted species (Fig. 1). All grasses had perennial plant
phenology with multiple leaf cohorts coexisting in the
same tussock and continuous leaf production; shrub
species S. filaginoides was semi-deciduous with a very
short overlap between successive year leaves, and the
other shrubs were deciduous with marked differences in
the leaf time span (Fig. 1).
Plant traits that modify microclimatic conditions
strongly differed among life forms. Soil thermal
amplitude was higher for grasses than for shrubs (Fig.
1); a difference that was consistent among species with
the exception of A. volckmanni, which presented an
intermediate value (Fig. 1). M. spinosum and S.
filaginoides have closed canopies and project a deep
shadow to the soil underneath, contrasting with
scattered shadows of A. volckmanni and grasses. The
thermal amplitude of soil underneath grasses did not
differ from bare soil (10.4 6 1.28C, t test, P . 0.05).
Life forms did not differ in traits that affect ecosystem
nitrogen dynamics, although species characteristics were
dissimilar in many cases (Fig. 1). Soil potential net
nitrification was very high underneath A. volckmanni,
which is the only legume species studied (Fig. 1). Soil
potential net ammonification differed in sign among
species; soils under S. humilis showed net positive
mineralization while the rest showed NH4-N immobili-
zation. Species differences in soil inorganic nitrogen
mineralization could result from differences in litter
quality accumulated underneath the canopy, species
composition of the community of microorganisms, and
the amount of soil organic matter accumulated under
the canopy of each species. Plant preference for N form
was marginally different, but was probably because of
the effect of A. volckmanni that absorbed larger
quantities of NO3-N than NH4-N (Fig. 1).
February 2012 229SPECIES TRAITS AND ECOSYSTEM PROCESSES
FIG. 1. Trait values related to carbon (green), water (blue), microclimate (violet), and nitrogen dynamics (red) for life forms
(open bars) and plant species (solid bars) in the Patagonian steppe. Traits are: RGR, relative growth rate; PSR, potential soil
respiration; RD, rooting depth; PP, plant phenology, measured as months with green leaves per year; STA, soil thermal amplitude;
PSN, potential soil nitrification; PSA, potential soil ammonification; and NRP, nitrogen form relative preference. Life forms and
species abbreviations are: Shr, shrubs; Gra, grasses; MS, M. spinosum; AV, A. volckmanni; SF, S. filaginoides; PL, P. ligularis; SS, S.
speciosa; SH, S. humilis. Values are meansþSE. Error terms are not available for RD of species because data came from previously
published papers where error was calculated for horizontal rather than vertical rooting distribution. Error terms are not available
for PP because all sampled individuals within a species were at the same phenological stage at each sampling date. For species,
different letters above bars represent significant differences (P , 0.05) according to Fisher’s LSD contrast. For life forms,
significance was determined by ANOVA with a priori contrasts.
* P , 0.05; *** P , 0.001; ? P , 0.06; ns, P . 0.06.
PEDRO FLOMBAUM AND OSVALDO E. SALA 230Ecology, Vol. 93, No. 2
PCA analysis separated species according to their
traits and levels (Fig. 2). The first two axes accounted for
77.1% of total variance, and separated A. volckmanni as
the most distinctive species, from the other two shrubs,
M. spinosum and S. filaginoides, and more distantly to
the three grasses P. ligularis, S. speciosa, and S. humilis.
The first axis had high loads for relative growth rates,
plant phenology, and relative preference of inorganic
nitrogen form (Table 1A), three traits that identified A.
volckmanni (Fig. 1). The second axis separated the two
shrubs, M. spinosum and S. filaginoides, from the rest of
the species; A. volckmanni, which had intermediate
values between grasses and shrubs for the traits on the
second axis, is closer to the grasses than to the other two
shrubs. The second axis had high loads for potential soil
respiration, species effect on soil thermal amplitude, and
species effect on potential soil nitrification (Table 1A).
The third axis separated S. humilis from the rest of the
species because of its high potential ammonification
values (eigenvector ¼ 0.84). Traits were not correlated
among each other (Appendix C). The highest correla-
tion (?0.896) was between relative growth rate and
nitrogen-form preference because of the distinctive effect
of one single species, A. volckmanni. However, no
correlation was observed if A. volckmanni was removed
suggesting that traits, in general, were not associated.
We found a positive relationship between trait
diversity and aboveground net primary production.
We estimated this relationship with 255 different trait-
diversity gradients and 228 presented a positive and
significant relationship (P , 0.05, Fig. 3). The average
coefficient of determination (R2) increased with the
number of traits, which were combined to create trait-
diversity gradients (P .. 0.001) from 0.18 to 0.26 for
one to eight combined traits (Fig. 3).
How does niche complementarity occur?
The life forms, shrubs and grasses, complement each
other because they differed in water absorption patterns
(Sala et al. 1989), soil thermal amplitude, and potential
soil mineralization. Shrubs have deep roots and absorb
that affect carbon, water, nitrogen, and microclimate in the
Patagonian steppe. Symbols represent species scores on the first
two axes of principal components analysis. The first axis
accounted for 48.2% of the variability and separated A.
volckmanni (AV), from grasses (PL, P. ligularis; SS, S. speciosa;
and SH, S. humilis) and the other two shrubs (MS, M.
spinosum; and SF, S. filaginoides). The second axis accounted
for 28.9% of the variability and separated MS and SF from the
rest of the species. Trait loads are reported in Table 1A.
Plant species ordination on the basis of species traits
TABLE 1.(A) Trait loads for PCA reported in Fig. 2 and (B) diversity of trait values calculated with Shannon Index.
A) Trait loads for PCA reported in Fig. 2
PCA axis 1
PCA axis 2
B) Trait values calculated with Shannon index
Shannon index (H’)188.8.131.520.871.011.560.870.45
Notes: Traits are RGR, relative growth rate; PSR, potential soil respiration; RD, rooting depth; PP, plant phenology, measured
as months with green leaves per year; STA, soil thermal amplitude; PSN, potential soil nitrification; PSA, potential soil
ammonification; and NRP, nitrogen-form relative preference. For the Shannon index, we made values for species traits relative
using a 0–1 scale and applied the Shannon modified index (H0) to explore which traits would have the largest contribution to niche
relationship between aboveground net primary production
(ANPP) and trait diversity in the Patagonian steppe. Each
dot (255 in total) represents the R2of a regression analysis
between ANPP and plant trait diversity. The x-axis represents
the number of traits combined to generate each trait-diversity
gradient. The discontinuous line represents a fitted function (R2
¼0.04 lnxþ0.19) to the number of traits included (P , 0.001).
Coefficient of determination (R2) obtained from the
February 2012 231SPECIES TRAITS AND ECOSYSTEM PROCESSES
water from deeper layers whereas grasses with their
shallow roots take up water from upper soil layers.
Water resources in these two layers have different
seasonal dynamics and consequently shrubs and grasses
are linked to two distinct resources. Deep soil layers are
refilled in the winter, and this resource is utilized slowly
until the middle of the growing season. Water availabil-
ity in upper layers is a transient resource that could be
available any time of the year. Upper layers get wet after
a rainfall event and dry out fast due to the high soil
evaporation characteristic of deserts. Shrubs with closed
and tall canopies reduced soil thermal amplitude
whereas grasses did not affect soil temperature. Also,
shrub canopies concentrate production and functioned
as a trap for organic matter, which resulted in 3.4 times
more inorganic nitrogen (Lopez et al. 2003) and 6.7
times more carbon (Gonzalez-Polo and Austin 2009).
The difference in soil organic matter between microsites
located underneath shrubs and in bare soil can partially
explain the observed 60% higher potential soil respira-
tion below shrubs than below grasses.
Species in Patagonia show high degree of niche
complementarity as a result of traits affecting carbon,
water, microclimate, and nitrogen dynamics of the
ecosystem. Grass species had similar effects on water
dynamics (similar rooting depth and the same phenol-
ogy) but surprisingly they differed in carbon- and
nitrogen-related traits (potential soil respiration and
ammonification, respectively). Within shrubs, the most
distinctive species was A. volckmanni, a legume, with a
unique effect on nitrogen dynamics (with a high
potential soil nitrification and a high preference for
nitrate). A. volckmanni also had different traits regarding
carbon (high relative growth rate) and water (shortest
period with green leaves), and its scattered canopy that
resulted in a grass-like microclimate effect. The other
shrub species, M. spinosum and S. filaginoides, differed
in several traits but were less contrasted than A.
volckmanni. Shrub species were much more heteroge-
neous as a group than grass species, probably because
shrubs species are located in more taxonomically distant
families (M. spinosum: Apiaceae, S. filaginoides: Aster-
aceae, A. volckmanni: Fabaceae) than grasses (Poaceae).
The diverse phylogenetic origins characteristic of the
shrub group would increase the tendency of organisms
to be dissimilar (McKitrick 1993). Our findings of
complementarity among species from this water-limited
ecosystem are similar to the results from a nitrogen-
limited ecosystem where niche complementarity among
species was based on rooting depth, plant phenology,
and nitrogen-form relative preference (McKane et al.
2002, Kahmen et al. 2006).
What are the species traits that account for most
of the observed niche complementarity?
We expected that traits with levels more uniformly
spread among species would have the largest effect on
niche complementarity and that, on the contrary, traits
for which most of the species present the same value
would have the smallest contribution to niche comple-
mentarity. We made values for species traits relative
using a 0–1 scale and used the modified Shannon index
(H’) to explore which traits would have the largest
contribution to niche complementarity. Traits associat-
ed with water had the largest contribution to niche
complementarity and those associated with nitrogen
smallest (Fig. 4, Table 1B). Species had homogeneously
spread rooting depth and potential soil nitrification
different species. Traits that were more uniformly spread across species should have a larger effect on complementarity. In contrast,
traits with little variability should contribute less to niche complementarity. Here, we made each trait relative to its minimum (Min)
and maximum (Max) values. Table 1B provides a quantification of species distribution for each trait.
Relative values of traits for all species analyzed. Each row represents a different trait, and each symbol represents a
PEDRO FLOMBAUM AND OSVALDO E. SALA232Ecology, Vol. 93, No. 2
characteristics and had lumped characteristics for
nitrogen-form relative preference, potential soil ammo-
nification, and plant phenology (Fig. 4, Table 1B). There
was a clear gradient for species rooting depth, a key trait
in arid ecosystems since it is related to resource uptake
in space and time (Golluscio and Sala 1993, Schwinning
and Ehleringer 2001). Similarly, species traits were
evenly distributed for potential soil nitrification; how-
ever the impact of this trait in the fitness of plant species
is less direct.
How does trait diversity affect ecosystem functioning?
Trait diversity increased aboveground net primary
production consistently with the hypothesis that relates
biodiversity and ecosystem functioning and with niche
complementarity as the main causal mechanism (Vitou-
sek and Hooper 1993, Tilman et al. 1997b). Comple-
mentarity is a product of species interactions and
therefore combinations of traits, rather than traits of
individual species in isolation. We observed an asymp-
totic increase in the fraction of ANPP accounted for by
the number of traits along trait-diversity gradients (Fig.
3), which could be partly explained by differences in trait
contribution to niche complementarity (Fig. 4), trait
effects on ANPP, and correlation among traits (Fig. 2).
For example, rooting depth had large effects on niche
complementarity and ANPP; relative preference for
nitrate or ammonia, on the contrary, had small effects.
Our results provide evidence on how niche comple-
mentarity occurs, identifies which traits had the largest
contribution to niche complementarity, and shows a
direct link between trait-based niche complementarity
and aboveground net primary production in the
We thank M. L Yahdjian, L. Vivanco, D. H. Morse, M.
Bertness, S. Porder, J. B. H. Martiny, and two anonymous
reviewers for their suggestions during the experiment and
manuscript preparation, and J. Vrsalovic, M. and L. Covalschi,
M. Gonzalez Polo, C. Gonzalez-Fisher, L. Gherardi, and G. A.
Gil for their field and lab assistance. We express our gratitude
to the Institute for Agricultural Plant Physiology and Ecology
(IFEVA) and to the Agricultural Research Service of Argentina
(INTA) of Rı´o Mayo for providing us with support and
facilities. This work was funded by the Argentina National
Council for Research PIP 112-200801-01788 (CONICET),
PRH-PICT 1-1 0106 (Agencia), University of Buenos Aires,
and Arizona State University.
Brown, W. L., and E. O. Wilson. 1956. Character displacement.
Systematic Zoology 5:49–64.
Cardinale, B. J. 2011. Biodiversity improves water quality
through niche partitioning. Nature 472:86–89.
Chase, J. M., and M. A. Leibold. 2003. Ecological niches:
linking classical and contemporary approaches. University of
Chicago Press, Chicago, Illinois, USA.
Eviner, V. T., and F. S. Chapin. 2003. Functional matrix: a
conceptual framework for predicting multiple plant effects on
ecosystem processes. Annual Reviews of Ecology, Evolution
and Systematics 34:455–485.
Fargione, J., D. Tilman, R. Dybzinski, J. H. R. Lambers, C.
Clark, W. S. Harpole, J. M. H. Knops, P. B. Reich, and M.
Loreau. 2007. From selection to complementarity: shifts in
the causes of biodiversity–productivity relationships in a
long-term biodiversity experiment. Proceedings of the Royal
Society B 274:871–876.
Flombaum, P., and O. E. Sala. 2007. A non-destructive and
rapid method to estimate biomass and aboveground net
credit: O. E. Sala.
Patagonian steppe, broad horizons, infinite landscapes, and rare rain showers as depicted here by the rainbow. Photo
February 2012 233 SPECIES TRAITS AND ECOSYSTEM PROCESSES
primary production in arid environments. Journal of Arid
Flombaum, P., and O. E. Sala. 2008. Higher effect of plant
species diversity on productivity in natural than artificial
ecosystems. Proceedings of the National Academy of
Sciences USA 105:6087–6090.
Golluscio, R. A., and O. E. Sala. 1993. Plant functional types
and ecological strategies in Patagonian forbs. Journal of
Vegetation Science 4:839–846.
Gonzalez-Polo, M., and A. T. Austin. 2009. Spatial heteroge-
neity provides organic matter refuges for soil microbial
activity in the Patagonian steppe, Argentina. Soil Biology
and Biochemistry 41:1348–1351.
Hector, A., et al. 1999. Plant diversity and productivity
experiments in European grasslands. Science 286:1123–1127.
Huston, M. A. 1997. Hidden treatments in ecological experi-
ments: re-evaluating the ecosystem function of biodiversity.
Kahmen, A., C. Renker, S. B. Unsicker, and N. Buchmann.
2006. Niche complementarity for nitrogen: an explanation
for the biodiversity and ecosystem functioning relationship?
Kawecki, T. J., and D. Ebert. 2004. Conceptual issues in local
adaptation. Ecology Letters 7:1225–1241.
Lopez, N. I., A. T. Austin, O. E. Sala, and B. S. Mendez. 2003.
Controls on nitrification in a water-limited ecosystem:
experimental inhibition of ammonia-oxidising bacteria in
the Patagonian steppe. Soil Biology and Biochemistry
Loreau, M., and A. Hector. 2001. Partitioning selection and
complementarity in biodiversity experiments. Nature 412:72–
McKane, R. B., L. C. Johnson, G. R. Shaver, K. J.
Nadelhoffer, E. B. Rastetter, B. Fry, A. E. Giblin, K.
Kielland, B. L. Kwaitkowski, J. A. Laundre, and G. Murray.
2002. Resource-based niche provide a basis for plant species
diversity and dominance in Artic tundra. Nature 415:68–71.
McKitrick, M. C. 1993. Phylogenetic constraint in evolutionary
theory: has it any explanatory power? Annual Review of
Ecology and Systematics 24:307–330.
Sala, O. E., R. A. Golluscio, W. K. Lauenroth, and A. Soriano.
1989. Resource partitioning between shrubs and grasses in
the Patagonian steppe. Oecologia 81:501–505.
Schwinning, S., and J. R. Ehleringer. 2001. Water use trade-offs
and optimal adaptations to pulse-driven arid ecosystems.
Journal of Ecology 89:464–480.
Smith, K. A., T. Ball, F. Conen, K. E. Dobbie, J. Massheder,
and A. Rey. 2003. Exchange of greenhouse gases between soil
and atmosphere: interactions of soil physical factors and
biological processes. European Journal of Soil Science
Tilman, D., J. Knops, D. Wedin, P. Reich, M. Ritchie, and E.
Siemann. 1997a. The influence of functional diversity and
composition on ecosystem processes. Science 277:1300–1302.
Tilman, D., C. Lehman, and K. Thomson. 1997b. Plant
diversity and ecosystem productivity: theoretical consider-
ations. Proceedings of the National Academy of Sciences
Vitousek, P. M., and D. U. Hooper. 1993. Biological diversity
and terrestrial ecosystem biogeochemistry. Pages 3–14 in
E. D. Schulze and H. A. Mooney, editors. Biodiversity and
ecosystem function. Springer-Verlag, Berlin, Germany.
Yahdjian, L. M., O. E. Sala, and A. T. Austin. 2006.
Differential controls of water input on litter decomposition
and nitrogen dynamics in the Patagonian steppe. Ecosystems
Plant species traits that affect ecosystem functioning (Ecological Archives E093-021-A1).
Example of construction of a trait richness gradient (Ecological Archives E093-021-A2).
Correlation coefficient among traits (Ecological Archives E093-021-A3).
PEDRO FLOMBAUM AND OSVALDO E. SALA 234Ecology, Vol. 93, No. 2