Journal of Vegetation Science 28 (2017) 149–159

Plant community assembly in Mediterranean grasslands: understanding the interplay between grazing and spatio-temporal water availability ~a Peco Cristina Rota, Pablo Manzano, Carlos P. Carmona, Juan E. Malo & Begon

Keywords Annual plants; Diversity partitioning; Environmental filtering; Functional diversity; Functional redundancy; Functional structure; Plant functional traits; Productivity; Species richness; Stocking density Abbreviations CMT = community mean trait value; FD = functional diversity, measured as Rao’s quadratic entropy; SLA = specific leaf area; TD = taxonomic diversity, measured as species richness. Nomenclature Flora Europaea (Tutin et al. 1964–1980) Received 5 April 2016 Accepted 29 August 2016 Co-ordinating Editor: Sandor Bartha

Rota, C. (corresponding author, [email protected])1, Manzano, P. ([email protected])1, Carmona, C.P. ([email protected])2, Malo, J.E. ([email protected])1, Peco, B. ([email protected])1 1 Terrestrial Ecology Group, Department of Ecology, Autonomous University of Madrid, Madrid 28049, Spain; 2 Department of Botany, Faculty of Science, University of South Bohemia,  e  Bud Cesk ejovice 37005, Czech Republic

Abstract Questions: How does grazing affect taxonomic diversity and functional structure of Mediterranean grassland communities? How do spatial and inter-annual variations in water availability, as a proxy for productivity, modulate grazing effects? Are shifts in taxonomic diversity systematically mirrored by analogous changes in functional diversity along these gradients? Location: Mediterranean grasslands in central Spain. Methods: We surveyed grassland plant communities in 3 yrs with contrasting mean annual rainfall (total n = 441 plots). Grazing gradients were quantified by periodic visual observation. DEM and annual rainfall data were used to quantify water availability. We examined the effects of grazing and spatio-temporal water availability on taxonomic diversity (TD; species richness), functional diversity (FD; Rao’s Q) and community mean trait values (CMT) for three key plant traits (specific leaf area –SLA–, height and seed mass). Functional redundancy was discussed through the relationship between TD and FD trends. Results: The results for TD, FD and CMT showed that environmental filtering determined the differences between plots with different grazing and water availability conditions. In contrast to seed mass FD, FD of vegetative traits (height and SLA) was highly decoupled from TD as a result of both spatial and interannual variations in water availability. Grazing reduced TD and functional redundancy only in the wettest year (i.e. in the absence of drought filtering), but selected for species with grazing tolerance and grazing avoidance strategies (reflected by high SLA CMT and low height CMT, respectively) in all water availability conditions.

Conclusions: Our results highlight the importance of grazing and both spatial and temporal variation in water availability as drivers of the assembly of Mediterranean grassland communities. The complex and decoupled responses that we found confirm that future studies should combine the use of different analytical methods to elucidate the multiple facets of community change, along with an optimal characterization of livestock pressure and its potential interactions with habitat productivity.

Introduction Semi-natural grasslands are among the most speciesdiverse habitats throughout Europe, hence constituting a central issue in nature conservation (Vandewalle et al. 2014). However, changes in livestock management (i.e. grazing intensification or abandonment; Peco et al. 2012), along with climate change (Cantarel et al. 2013),

which in Mediterranean environments is associated with reductions in rainfall, are already threatening their plant diversity patterns. In fact, livestock grazing is one of the main drivers of change in composition, structure and functioning of grassland plant communities (Sternberg et al. 2000). Characterizing these effects represents a challenge that requires integrating the multiple facets of community change, as well as an adequate characterization of

Journal of Vegetation Science Doi: 10.1111/jvs.12470 © 2016 International Association for Vegetation Science

149

Community assembly in Mediterranean grasslands

C. Rota et al.

livestock pressure and its potential interactions with habitat productivity. Traditionally, grazing effects on grassland communities have been studied using measures of taxonomic diversity (TD), such as species richness (Osem et al. 2002; Dorrough et al. 2006). Nevertheless, ecologists are increasingly focusing on the functional aspects of communities, since they provide a more general and mechanistic basis for characterizing patterns of response across environmental gradients (McGill et al. 2006). Environmental filtering limits the range of trait values within communities by excluding species lacking traits for persistence under a certain set of environmental constraints (Dıaz et al. 1998), such as those imposed by grazing. To analyse these dynamics it is useful to describe the functional structure of communities considering two complementary measures (Ricotta & Moretti 2011): first, an indicator of community functional composition, describing average trait values within a given species assemblage (e.g. community mean trait value, CMT; Roscher et al. 2013); second, an indicator of community functional diversity (FD), describing the dispersion of trait values (e.g. Rao’s quadratic entropy; Rao 1982). Thus, whereas functional composition can be appropriately employed to summarize changes in mean trait values within communities as a result of environmental selection for certain trait values, FD can be used to interpret patterns of trait convergence and divergence (Ricotta & Moretti 2011). Despite the potential of functional metrics, using FD as a surrogate for TD (or vice versa) can lead to erroneous conclusions (Mayfield et al. 2010). For instance, TD and FD do not always co-vary along grazing gradients (de Bello et al. 2006; Sasaki et al. 2009). Hence, an additional criterion to consider is the actual relationship between TD and FD. This relationship depends on the degree of functional redundancy within the assemblage, i.e. the number of species that exhibit similar functional traits and, consequently, perform equivalent ecological roles. Functionally redundant species can thus be removed from a given community without impacting ecosystem processes (i.e. decreases in TD are not mirrored by comparable decreases in FD; Rosenfeld 2002). Therefore, studying the relationship between TD and FD is crucial for gaining a mechanistic understanding of community assembly and ecosystem functioning along environmental gradients. Lastly, although the effects of grazing on TD and FD have generally been approached within communities, local assemblages are the result of processes operating at different scales (de Bello et al. 2009). For example, environmental conditions that prevail in a certain year can not only affect within-communities TD and FD, but also their extent at the regional scale (Carmona et al. 2012). Hence, partitioning the regional TD and FD (c) into their within-

150

communities (a) and among-communities (b) components (Whittaker 1975) can provide additional information about the assembly and co-existence of species (de Bello et al. 2009). A complete determination of community change also requires a detailed characterization of livestock pressure and its interactions with other environmental factors. In fact, one of the main difficulties in the study of grazing effects on plant communities is that they strongly depend on habitat productivity (Osem et al. 2002; Carmona et al. 2012). In this sense, the generalized model of Milchunas et al. (1988) remains as the standard reference for interpreting and predicting grazing impacts on plant communities with respect to its interaction with habitat productivity (Oesterheld & Semmartin 2011). This model hypothesizes that grazing and aridity act as convergent selective pressures, since the traits that confer resistance to grazing and drought are the same. Consequently, grazing effects should be more conspicuous on wet than on dry areas, where they are more likely to be overridden by the underlying impact of aridity (Dıaz et al. 2007; Quiroga et al. 2010). In particular, there are two contrasting strategies by which plants cope with grazing: avoidance and tolerance (Cingolani et al. 2005; Dıaz et al. 2007). Avoidance strategies minimize the probability of being grazed, involving slow-growing tissue with considerable structural defence, and leading to short-sized plants with small and unpalatable leaves. Many of the traits presented by grazing avoiders overlap with those presented by drought avoiders, which resist water stress by preventing tissue dehydration (Kenney et al. 2014). In contrast, grazing-tolerant species rapidly replace leaf area removed by herbivores, involving fast tissue regrowth with very low structural defence, and leading to plants with palatable leaves. Notably, many of the traits presented by grazing-tolerant species overlap with those presented by drought escapers, which resist water stress by having short life cycles that can be completed during the period of sufficient water supply (Kenney et al. 2014). The prevailing strategy of grazing resistance is actually context-dependent, with avoidance favoured over tolerance in less-productive systems and vice versa in more-productive systems (Cingolani et al. 2005; Dıaz et al. 2007; Carmona et al. 2015a). An important but often neglected point to consider is that productivity does not only vary spatially, but also temporally. The latter is especially relevant in Mediterranean grasslands, since the mediterranean climate has large inter-annual rainfall variability, resulting in significant disparities in resource availability, species composition and diversity among years (Carmona et al. 2012, 2015a). Therefore, studies covering only 1 yr may be biased by year-specific responses of plant communities, giving a distorted picture of the overall patterns. Finally, although

Journal of Vegetation Science Doi: 10.1111/jvs.12470 © 2016 International Association for Vegetation Science

Community assembly in Mediterranean grasslands

C. Rota et al.

some studies have already addressed the effects of grazing on plant communities by jointly considering its interaction with spatial (de Bello et al. 2006) or even inter-annual variation in productivity (Osem et al. 2002; Carmona et al. 2012, 2015a), these works lack a quantitative characterization of habitat productivity and/or the degree of grazing by livestock. Exhaustive quantifications on continuous scales of both predictors are needed for a full understanding of their effects on grassland communities. In this study, we assessed the influence of grazing on assembly processes in Mediterranean grassland communities. We addressed the following questions: (1) how does grazing affect within-communities TD (a-TD) and functional structure (CMT and a-FD); (2) how do spatial and inter-annual variations in water availability, as a proxy for productivity, modulate grazing effects on both components of a-diversity; (3) are shifts in a-TD systematically mirrored by analogous changes in a-FD along these gradients; (4) how are TD and FD partitioned across spatial scales? Specifically, our main hypotheses were: (1) because of a stronger effect of habitat filtering, low water availability (both temporal and spatial) should filter out unsuitable species whose traits reflect lower fitness, hence reducing FD and shifting CMT along the water availability gradient; (2) if grazing and aridity act as convergent selective forces, the filtering process promoted by grazing should be stronger in wet conditions than in dry ones; (3) since grazing avoiders should be favoured over grazing-tolerant species in less-productive systems and vice versa in more-productive systems, CMT of traits related to these strategies should differ between dry and wet conditions in highly grazed scenarios.

Methods Study area The study area is located in central Spain, about 20 km north of Madrid (40°360 N, 3°450 W), with average altitude of 700 m (ranging from 690 to 750 m). The climate is continental mediterranean, with cold winters, hot and dry summers and rainfall mostly concentrated in spring and autumn. The mean annual temperature is 14.8 °C (0.5) and the average annual rainfall is 450 mm (109.5). The area is underlain by arkosic sands, and the vegetation consists mainly of dry grasslands dominated by annuals (80% of total species richness), such as Bromus hordeaceus L., Plantago lagopus L., Trifolium campestre Schreb. and Vulpia muralis (Kunth) Nees. Vegetation sampling We selected two estates – A and B; 27.9 and 24.5 ha, respectively – 3 km apart from each other, comprising

pastures for livestock. Both estates were grazed by sheep at an average stocking density of 1.7 sheepha1. We surveyed plant communities from both estates in the spring of three consecutive years with contrasting mean annual rainfall (442.0 mm in 2003; 486.6 mm in 2004; 340.4 mm in 2005), using quadrats (20 cm 9 20 cm) that were randomly positioned each year: 146 in 2003 (84 in A; 62 in B), 144 in 2004 (84 in A; 60 in B) and 151 in 2005 (80 in A; 71 in B), totalling 441 plots. For each one, we determined plant species composition (presence/absence) and species richness (hereafter a-TD). Determination of the soil water availability and grazing gradients We calculated a soil water availability index for each plot using topographic attributes derived from a 5-m resolution DEM. The index was computed by dividing the topographic wetness index (TWI; Wolock & McCabe 1995) of the pixel where each plot was located from its heat load (HL; McCune & Keon 2002), both previously rescaled between zero and one (see detailed procedure in Appendix S1, Section 1). This quotient defines a gradient from the most water-stressed environments (high HL and low TWI) to the most humid ones (low HL and high TWI). We also elaborated a 5-m resolution mean stocking density map of the study area through periodic visual observations (totalling 350 observation events on each estate from February 2003 to January 2005). The mean stocking density of each plot was obtained from this map by averaging the stocking density values of all the observation events for the particular pixel where the plot was located (see detailed procedure in Appendix S1, Section 2). These procedures allowed analyses along fine-scale environmental gradients, instead of simply comparing between discrete conditions (dry vs wet or low vs high stocking density). Quantification of functional structure We characterized the functional structure of each plot for three key plant traits: plant height, SLA and seed weight. Trait values were retrieved from a regional database containing average trait values for several species measured following standard procedures (Cornelissen et al. 2003) in grasslands located no more than 25 km away from the study area (Azc arate et al. 2002; Sanchez et al. 2002; Peco et al. 2005; Carmona et al. 2012, 2015b). Functional structure was quantified only in those plots where functional trait information was available for at least 75% of the species, totalling 430 plots for height and SLA analyses and 390 for seed weight. Data on the three traits were logtransformed prior to analyses to achieve normal distributions.

Journal of Vegetation Science Doi: 10.1111/jvs.12470 © 2016 International Association for Vegetation Science

151

Community assembly in Mediterranean grasslands

C. Rota et al.

We estimated the community mean trait value (hereafter CMT; Roscher et al. 2013) for each trait and plot as: S P

xi

CMT ¼ i¼1 S

ð1Þ

where xi is the trait value of species i (i = 1, . . ., S). From an ecological viewpoint this metric reflects the most frequent values of a particular trait in a particular community. We also calculated the Rao’s quadratic entropy (Rao 1982) for every trait and plot, to estimate its functional diversity (hereafter a-FD):  2 ! S X S X 1 Rao ¼ dij S i¼1 j¼1

ð2Þ

where i and j represent every pair of co-existing species, and dij is the dissimilarity between their trait values obtained from a Gower dissimilarity matrix (Gower 1971). Afterwards, we applied Jost’s correction to the observed Rao values (Jost 2007; de Bello et al. 2010). Since we assigned equal proportional abundances for all the species present in each plot, the Rao index in this study corresponds to a measure of functional richness rather than one of functional divergence (Mason et al. 2012). It is appropriate to note here that since we focused on fixed trait averages (taken from databases) and used a measure of functional richness for quantifying FD (which does not take into account species abundances), our results fully correspond to the turnover in species identity between plots. This allowed us to examine the precise role of this component in assembly processes along environmental gradients without the confounding effects of species dominance and within-species trait variability (Leps et al. 2011). From here on, the functional metric of each trait appears beside its name (TraitCMT; Traita-FD). Data analyses All analyses were performed using R v 3.0.3 (R Foundation for Statistical Computing, Vienna, AT). Prior to analyses, water availability and stocking density were z-standardized. Stocking density was squared after standardization to reduce the correlation between the linear and quadratic terms (Schielzeth 2010). We performed linear regressions for the analyses of CMT and a-FD of each trait and a GLM with Poisson distribution and log link for the analysis of aTD to account for the count data nature of species richness. Estate, year, water availability, stocking density (as both simple and quadratic terms, to detect non-linear relationships) and year 9 water availability 9 stocking density

152

triple interaction (alongside all the possible pair-wise interactions between these three predictors) were included at the initial stage of modelling. The full models were then reduced via backward selection until only significant variables (P < 0.05) remained in the final models. We used Type III sums of squares to account for unequal replication and interaction terms (Fox & Weisberg 2011). However, to control for estate-specific factors, estate was retained in all models regardless of its statistical significance. Tukey’s post-hoc tests were performed when year was retained in the finals models to determine where differences between groups occurred. As none of the final models displayed spatial autocorrelation in the residuals according to onetailed Mantel’s permutation tests (Mantel 1967), we did not use explicit spatial models. Moreover, using the R function ‘Rao’ (de Bello et al. 2010), we calculated and plotted the partitioning of both TD and FD for each year to examine the effects of inter-annual rainfall variability across scales: within plots (a), among plots (b) and within years (c). Thus, a- and b-diversity summarize the amount of within and between plot diversity, respectively, within a given year, so that: mean a + b = c. Finally, functional redundancy is addressed in the Discussion section through the relationship between a-TD and a-FD trends (i.e. whether or not increases and decreases in a-TD were mirrored by increases and decreases in a-FD, respectively).

Results Within-plot analysis (a-TD, a-FD and CMT) Estate explained a significant proportion of the variability in heighta-FD, SLAa-FD and seed weighta-FD, whereas its contribution to that of a-TD, heightCMT, SLACMT and seed weightCMT was not significant (Tables 1,2). Effect of spatial variation in water availability Variation in soil water availability had an effect on both a-TD and functional structure (Tables 1,2, Fig. 1a). a-TD decreased significantly with water availability (b = 0.036), peaking in water-stressed locations. The decline in a-TD with increasing water availability was Table 1. a-TD model: analysis of deviance table (Type III tests) testing the significance of each predictor; significant results (P < 0.05) are in bold.

Water Availability Year Stocking Density Stocking Density^2 Estate Stocking Density 9 Year

df

Chisq

P

1 2 1 1 1 2

8.061 238.246 2.558 4.677 3.649 8.652

0.005 <0.001 0.110 0.031 0.056 0.013

Journal of Vegetation Science Doi: 10.1111/jvs.12470 © 2016 International Association for Vegetation Science

Community assembly in Mediterranean grasslands

C. Rota et al.

Table 2. Functional structure (CMT and a-FD) models of each trait: analysis of deviance tables (Type III tests) testing the significance of each predictor; significant results (P < 0.05) are in bold. df Plant Height CMT Stocking Density Estate a-FD Water Availability Year Stocking Density Estate SLA CMT Stocking Density Estate a-FD Water Availability Year Estate Water Availability 9 Year Seed Weight CMT Water Availability Estate a-FD Water Availability Year Estate

F

P

1 1

4.682 0.999

0.031 0.318

1 2 1 1

6.586 16.749 3.956 10.570

0.011 <0.001 0.047 0.001

1 1

9.264 0.853

0.002 0.356

1 2 1 2

12.511 2.354 3.953 4.672

<0.001 0.096 0.047 0.010

1), a-TD was significantly higher in the wettest year (2004, mean 19.2 speciesplot1) and lower in the driest one (2005, mean 12.2 speciesplot1) compared to the year with intermediate rainfall (2003, mean 16.7 speciesplot1). However, there was an interaction between stocking density and year, so that the aforementioned differences in a-TD between years were maximized at low levels of stocking density (Fig. 1b). Furthermore, the decrease in a-TD in 2005 was mirrored by the decrease in seed weighta-FD, which was significantly lower in 2005 than in the other 2 yrs (2003 and 2004). Conversely, heighta-FD was significantly higher in 2005, whereas mean SLAa-FD did not vary significantly across the 3 yrs (Appendix S2, Section 1, Figure S2-1). Nonetheless, the abovementioned interaction between water availability and year revealed that while at water-stressed locations no obvious differences were observed between years, at higher levels of soil water availability SLAa-FD decreased with annual rainfall (Fig. 1a). Effect of grazing

1 1

7.365 0.163

0.007 0.687

1 2 1

14.874 10.143 38.337

<0.001 <0.001 <0.001

mirrored by the decrease in seed weighta-FD (b = 0.010), but not by a change in heighta-FD or SLAa-FD. While heighta-FD increased with increasing water availability (b = 0.007), an interaction between this predictor and year revealed that the variation of SLAa-FD with soil water availability depended on annual water availability (b2003 = 0.012, significantly different from zero, P < 0.01; b2004 = 0.004, not significantly different from zero, P < 0.26; b2005 = 0.021, significantly different from zero, P < 0.01). Hence, the positive effect of increasing water availability on SLAa-FD became noticeably stronger with decreasing annual water availability (Fig. 1a). Concerning CMT analysis (Table 2), seed weightCMT decreased with water availability (b = 0.053), so that communities in water-stressed locations tended to have heavier-seeded species. In contrast, neither heightCMT nor SLACMT showed a significant relationship with soil water availability. Effect of inter-annual variation in rainfall Rainfall fluctuations between years also had an effect on both a-TD and a-FD of the communities under study, but did not affect CMT of any trait (Tables 1,2). According to Tukey’s post-hoc tests (Appendix S2, Section 1, Figure S2-

Stocking density was retained in the final models of a-TD, heighta-FD, heightCMT and SLACMT (Tables 1,2, Fig. 1b). The interaction between stocking density and year revealed a non-linear decrease of a-TD along the grazing gradient in the year with higher mean annual rainfall (2004), whereas in the other years (2003 and 2005) the curve reached a slight maximum at intermediate levels of stocking density (Fig. 1b). In contrast, heighta-FD decreased linearly with increasing stocking density across the 3 yrs (b = 0.005), while SLAa-FD and seed weighta-FD remained constant. With respect to CMT analysis (Table 2), heightCMT consistently exhibited a significant linear decrease with increasing stocking density (b = 0.026), while SLACMT showed the opposite trend (b = 0.016). On the contrary, stocking density did not explain a significant proportion of the variability in seed weightCMT. Diversity partitioning The overall partitioning of TD revealed that the betweenplot component (b-TD) was always larger than the withinplot one (a-TD), regardless of year. In contrast, most of the observed variability in FD was consistently recorded at the within-plot scale (a-FD; Appendix S2, Section 1, Figure S2-1).

Discussion The present study supports the expectation that grazing and both spatial and inter-annual variation in water

Journal of Vegetation Science Doi: 10.1111/jvs.12470 © 2016 International Association for Vegetation Science

153

Community assembly in Mediterranean grasslands

C. Rota et al.

FD

TD Height

(a)

Seed weight

Soil water availability effect within plots 35

intermediate year wet year dry year

30 25 20 15 10 5 –1

0

1

2

3

4

1.40

1.40

α − FD (Rao)

α − TD (Richness)

SLA

1.35

1.30

1.35

1.30

1.25

1.30

1.25

1.20

1.25

1.20 1.15 1.10

1.10

1.05

1.05

5

1.20

1.15

–1

0

1

2

3

4

1.15 1.10 –1

5

0

1

2

3

4

5

–1

0

1

2

3

4

5

water availability

Grazing effect within plots 35

1.40

30

1.35

1.30

1.35

1.30

1.25

1.30

1.20

1.25

α − FD (Rao)

α − TD (Richness)

(b)

25 20 15

1.40

1.25 1.20

10

1.10

1.10

5

1.05

1.05

0

1

2

3

4

1.20

1.15

1.15

0

1

2

3

4

1.15 1.10 0

1

2

3

4

0

1

2

3

4

stocking density

Fig. 1. Scatterplots of a-TD (first column) and a-FD of the three different functional traits: height (second column), SLA (third column) and seed weight (fourth column) against standardized soil water availability (a) and stocking density (b). Lines represent the best-fit functions obtained for the whole data set, according to the multiple regression models in Tables 1 and 2, while keeping stocking density (a) or soil water availability (b) fixed to its mean value (i.e. to 0, since predictors were z-standardized prior to analyses). The Pearson correlation coefficient between soil water availability and stocking density was 0.068. Since year was retained in all final models, a separate curve was plotted for each year (regardless of the significance of water availability (a) or stocking density (b)) to show inter-annual differences: Intermediate year, 2003, 442.0 mm; wet year, 2004, 486.6 mm; dry year, 2005, 340.4 mm.

availability jointly shape local species assemblages in Mediterranean grasslands through complex interactions (e.g. Carmona et al. 2012, 2015a). The results for a-TD, aFD and CMT showed that environmental filtering determined the differences between plots with different grazing and water availability conditions. However, these responses were trait-dependent, with patterns of convergence and divergence of vegetative traits (height and SLA) consistently decoupled from those of seed weight (Lavorel et al. 2011). Moreover, the differences between TD and FD found within plots also extended across scales. Thus, while there was a high species turnover between plots, the functional diversity turnover was remarkably low (see Appendix S2, Section 2 for a more detailed discussion on diversity partitioning results). The precise effects of each of the three predictors on the different components of diversity are discussed in detail in the following sections. A summary of the main results of within-plot analyses (aTD, a-FD and CMT) can be found in Table 3. Effect of spatial variation in water availability As expected, soil water availability acted as a selective filter determining the trait values that were viable in local

154

communities (Dıaz et al. 1998). In particular, low soil water availability reduced the range of vegetative trait values (height and SLA). Although it is important to note here that the slope of the relationship between SLAa-FD and soil water availability was steeper in the driest year, the implications of this interaction will be discussed in more detail in the next section. The reduction in a-FD of height and SLA in stressful conditions was not accompanied by an analogous decrease in a-TD (Table 3a). On the contrary, aTD was higher in water-stressed locations, suggesting a relaxation of competitive exclusion processes that might be widespread in humid sites (Grime 2001). When considered together, the patterns of a-TD and a-FD of the vegetative traits suggest an increase in the functional redundancy of species from humid to water-stressed locations (Table 3a). Thus, assemblages in dry plots consisted of several species with similar – convergent – vegetative traits. Conversely, seed weighta-FD mirrored a-TD patterns, with larger ranges of seed weight in water-stressed locations (Table 3a). Accordingly, species in dry plots tended to have different – divergent – seed weight values. This contradiction between the assembly processes of vegetative and reproductive traits suggests that convergence constitutes a predominant pattern for vegetative traits, while

Journal of Vegetation Science Doi: 10.1111/jvs.12470 © 2016 International Association for Vegetation Science

Community assembly in Mediterranean grasslands

C. Rota et al.

Table 3. Main results of within-plot analysis summarized as the effects of the three analysed stresses, (a) low soil water availability, (b) low annual water availability and (c) high stocking density, and their interactions on a-TD, a-FD, CMT and functional redundancy. Functional redundancy is examined through the relationship between a-TD and a-FD trends. Arrows indicate positive (↑) and negative (↓) effects, and dashes () indicate non-significant effects. HG, high stocking density; LG, low stocking density; DS, dry soils; WS, wet soils; DY, dry year (2005); IY, intermediate year (2003); WY, wet year (2004). Stress

a-TD

SLA

Seeds

Height

SLA

Seeds

Height

SLA

Seeds

WY  IY ↓ DY ↓ WS ↑

























HG–WS ↓ HG–DS  LG–WS ↓ LG–DS ↓ WY ↓ IY  DY 

HG ↑





(b) Low Annual Water Availability

HG 



LG ↓ WY ↓ IY  DY 

Redundancy

Height (a) Low Soil Water Availability

(c) High Stocking Density

CMT

a-FD

DS  ↓





divergence is more recurrent for reproductive traits (Grime 2006; Carmona et al. 2015a). Further, the high diversity of seed weight in dry plots emphasizes the importance of regeneration niche differentiation within communities as a mechanism promoting species co-existence in diverse plant communities (Grubb 1977; Lavorel et al. 2011). Relevant niche differences enabling species co-existence in plant communities may hence only be evident during early life-history stages. This is likely to be especially relevant in Mediterranean grasslands due to the predominance of annual species, implying that establishment stages play a crucial role in the composition and structure of these systems (Espigares & Peco 1995; Carmona et al. 2015a). Contrary to our expectations, although a-FD of vegetative traits declined from wet to dry conditions, CMT of these traits remained constant along the entire gradient (Table 3a). This suggests the existence, in wet soil conditions, of rare species with extreme vegetative trait values, such as increased height, which are absent in dry soil conditions, where these traits confer low fitness. These compositional differences entailed large reductions in a-FD, associated with a reduction in the range of vegetative trait values present in these plots, but did not affect CMT to the same extent. Particular examples of this process may be Festuca ampla Hack. or Vicia lutea L., large species (41.86 cm and 38.89 cm, respectively) that were completely excluded in dry soil conditions. This effect might be further emphasized by the use of species presence/absence data in this study, which over-represents the effect of less abundant species. In contrast, both seed weighta-FD and seed weightCMT increased in arid locations (Table 3a) due to the appearance of species with heavier seeds. Echium plantagineum L., a heavy-seeded species (2.49 mg) that was only found in dry soil conditions, may be a particular example of this process. This result is consistent with the







WY  IY ↑ DY ↑

LG  WY ↓ IY  DY 

notion that heavy-seeded species have higher survival rates during early life-history stages, particularly in poor environments (Metz et al. 2010; Arellano & Peco 2012). Effect of inter-annual variation in rainfall Temporal fluctuations in water availability altered both aTD and functional structure, although the largest effects were mainly observed in the driest year (2005). The mean annual rainfall of the study area is 450 mm (109.5), so that 2003 (with 442.0 mm) and 2004 (with 486.6 mm) had annual rainfalls close to the average value, while 2005 (with 340.4 mm) experienced a sharp decrease in water availability. This suggests that the constraint imposed by drought in 2005 could have entailed significant restrictions to the range of trait values present in communities. However, vegetative and reproductive traits were again decoupled, showing opposite patterns. First, the decrease in a-TD in the driest year was not accompanied by an analogous decrease in heighta-FD or SLAa-FD, indicating a marked reduction in the functional redundancy of species in vegetative traits (Table 3b). This probably reflects a divergence in strategies of drought stress resistance. Whereas decreased soil productivity filtered out some unsuitable species, further decreases in water availability in the driest year reached sufficiently dry conditions to allow some specific drought-resistant species to establish in the communities. Due to the promotion of species occupying both extremes of the vegetative trait axes (e.g. very small winter annual species, which complete their cycle in early spring before summer drought, or very large herbaceous species with deep root systems that increase their water extraction capability), the range of vegetative trait values present in these plots was conspicuously expanded but without substantially shifting community-aggregated means. Particular examples of this process may be Euphorbia exigua L. or

Journal of Vegetation Science Doi: 10.1111/jvs.12470 © 2016 International Association for Vegetation Science

155

Community assembly in Mediterranean grasslands

C. Rota et al.

Teesdalia coronopifolia (J.P.Bergeret) Thell., which have notably low height (1.5 cm and 0.74 cm, respectively) and were only found in the driest year. However, in contrast to plant height patterns, strategies related to SLA only diverged in 2005 in high water availability scenarios (Table 3b). In relatively fertile sites, dry years might select species with high SLA, which complete their cycle in early spring before summer drought (drought escapers), as well as species with low SLA, which reduce water loss from evapotranspiration (drought avoiders). Thus, our results suggest that responses of drought stress resistance related to SLA patterns arose only under sufficient resource availability. Finally, the decrease in seed weighta-FD in the driest year may simply be a side effect of the decrease in a-TD, since these parallel trends were not accompanied by any change in CMT values (Table 3b). Reproductive traits are relatively stable from year to year despite considerable variation in rainfall conditions, whereas long-term differences in water availability, such as those produced by different topographic positions, are more likely to have a larger impact on seed weight patterns (Carmona et al. 2015a). However, a previous analysis of 16-yr permanent plots (1980–1995) along a topographic gradient in the same study area showed that large-seeded species were more abundant in dry years (Peco et al. 2009). This discrepancy could be due to the presence of extremely dry years in the latter study, with annual precipitation as low as 273 mm.

largely independent of a-TD due to high functional redundancy within communities (Carmona et al. 2012). According to this interpretation, functionally redundant species are removed as stocking density increases, leading to the observed decrease of a-TD across the grazing gradient with almost no impact on a-FD (Table 3c). Furthermore, we found that grazing favoured both tolerant species with palatable leaves, i.e. high SLA, as well as grazing avoiders of small stature, regardless of spatial and temporal water availability conditions (Table 3c). Since traits associated with grazing tolerance entail a less efficient conservation of resources than those associated with grazing avoidance (Cingolani et al. 2005), we expected tolerant species to be selected by grazing only under conditions of high resource availability. This lack of evidence for a shift from tolerance to avoidance strategies in less productive sites, consistent with the results of Lavorel et al. (2011), may reflect two particularities of our systems: (1) a regional species pool mainly composed of fast-growing annuals and lacking tough grasses or spiny shrubs typically promoted by grazing in other Mediterranean landscapes (Osem et al. 2004); (2) a water availability gradient not reaching sufficiently dry conditions to entail the non-viability of grazing-tolerant species. Finally, our results suggest that stocking density has no effect on seed weightCMT. Results reported by Lavorel et al. (2011) again support our observation, although it is widely recognized that seed weight responses to grassland management may be inconsistent across sites (Pakeman 2004) and years (Carmona et al. 2015a).

Effect of grazing Our results highlight the role of water availability as a modulator of grazing effects. In particular, rainfall fluctuations modified the effect of stocking density on a-TD, which decreased with grazing in the wettest year across the entire grazing gradient. In contrast, in drier years the decline of a-TD with grazing was less pronounced, and only detectable under high stocking densities (Table 3c). This is consistent with the generalized model of Milchunas (Milchunas et al. 1988), suggesting that grazing mostly affects plant communities in high productivity scenarios. Reduced water availability at low productivity scenarios imposes strong constraints on plants, so that grazing effects are more likely to be overridden under these conditions (Kikvidze et al. 2011; Carmona et al. 2012). However, contrary to our expectations, only heighta-FD was affected by stocking density (Table 3c). This reinforces the notion of disturbance as a driver of convergence in trait patterns (Mason et al. 2011; Laliberte et al. 2013). However, the low significance of the effect of stocking density on heighta-FD (Table 2), along with the non-significant SLAaFD and seed weighta-FD results, suggests that a-FD may be

156

Concluding remarks The main objective of the present study was to investigate whether and how grazing, together with both spatial and temporal variation in water availability, influences assembly processes in Mediterranean grassland communities. While confirming the importance of all three predictors and their interactions in shaping local species assemblages in Mediterranean grasslands, our results point out the relevance of both functional redundancy and functional differentiation in species co-existence (Fukami et al. 2005). Notably, patterns in vegetative traits (height and SLA) were repeatedly decoupled from those of seed weight, highlighting the importance of single-trait analyses to assess mechanisms of community response to environmental and land-use change (Lavorel et al. 2011). Our findings also revealed that TD and FD do not always covary along water availability or grazing gradients, but may even show opposite responses. This effectively means that different biodiversity components are frequently independent, and, consequently, TD should not be used as surrogate for FD, or vice versa (de Bello et al. 2006; Mayfield

Journal of Vegetation Science Doi: 10.1111/jvs.12470 © 2016 International Association for Vegetation Science

Community assembly in Mediterranean grasslands

C. Rota et al.

et al. 2010). Interestingly, our findings highlight the functional stability of Mediterranean grasslands in the face of human disturbances, thanks to the high levels of functional redundancy of these communities, which provides insurance against grazing. Nonetheless, increased stocking density fostered species turnover towards more grazingresistant species, either by avoidance or tolerance mechanisms, suggesting that changes in grazing management should be considered with caution, especially in a context of reduced water availability due to climate change. The complex and decoupled responses that we found confirm that future studies should combine the use of different analytical methods to gather the multiple facets of community change (Mouillot et al. 2013; Carmona et al. 2016), with an optimal characterization of livestock pressure and its potential interactions with habitat productivity. We believe that future research on this topic should aim at disentangling the specific role of turnover in species identity and dominant species and on the role of intraspecific trait variability.

Acknowledgements Financial support was provided by the Spanish MINECO (Project CGL2014-53789-R) and the Madrid Regional Government (Project REMEDINAL-3). CR was supported by a FPU fellowship from the Spanish MECD (AP20122849). CPC was supported by a Marie Curie Intra-European Fellowship within the 7th European Community Framework Programme (TANDEM; project id. 626392). We thank Francisco M. Azcarate for fieldwork assistance and valuable advice, and Carly Golodets and Takehiro Sasaki for helpful comments on an earlier version of this paper.

References Arellano, G. & Peco, B. 2012. Testing the role of seed size in annual legume seedling performance under experimental autumn moisture conditions. Journal of Vegetation Science 23: 690–697. Azc arate, F.M., S anchez, A.M., Arqueros, L. & Peco, B. 2002. Abundance and habitat segregation in Mediterranean grassland species: the importance of seed weight. Journal of Vegetation Science 13: 159–166. Cantarel, A.A.M., Bloor, J.M.G. & Soussana, J.F. 2013. Four years of simulated climate change reduces above-ground productivity and alters functional diversity in a grassland ecosystem. Journal of Vegetation Science 24: 113–126. Carmona, C.P., Azc arate, F.M., de Bello, F., Ollero, H.S., Leps, J. & Peco, B. 2012. Taxonomical and functional diversity turnover in Mediterranean grasslands: interactions between grazing, habitat type and rainfall. Journal of Applied Ecology 49: 1084–1093.

Carmona, C.P., Mason, N.W.H., Azc arate, F.M. & Peco, B. 2015a. Inter-annual fluctuations in rainfall shift the functional structure of Mediterranean grasslands across gradients of productivity and disturbance. Journal of Vegetation Science 26: 538–551. Carmona, C.P., Rota, C., Azc arate, F.M. & Peco, B. 2015b. More for less: sampling strategies of plant functional traits across local environmental gradients. Functional Ecology 29: 579– 588. Carmona, C.P., De Bello, F., Mason, N.W.H. & Leps, J. 2016. Traits without borders: integrating functional diversity across scales. Trends in Ecology & Evolution 31: 382–394. Cingolani, A.M., Posse, G. & Collantes, M.B. 2005. Plant functional traits, herbivore selectivity and response to sheep grazing in Patagonian steppe grasslands. Journal of Applied Ecology 42: 50–59. Cornelissen, J.H.C., Lavorel, S., Garnier, E., Dıaz, S., Buchmann, N., Gurvich, D.E., Reich, P.B., Ter Steege, H., Morgan, H.D., (. . .) & Poorter, H. 2003. A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Australian Journal of Botany 51: 335–380. de Bello, F., Leps, J. & Sebasti a, M.T. 2006. Variations in species and functional plant diversity along climatic and grazing gradients. Ecography 29: 801–810. de Bello, F., Thuiller, W., Leps, J., Choler, P., Clement, J.C., Macek, P., Sebastia, M.T. & Lavorel, S. 2009. Partitioning of functional diversity reveals the scale and extent of trait convergence and divergence. Journal of Vegetation Science 20: 475–486. de Bello, F., Lavergne, S., Meynard, C.N., Leps, J. & Thuiller, W. 2010. The partitioning of diversity: showing Theseus a way out of the labyrinth. Journal of Vegetation Science 21: 992–1000. Dıaz, S., Cabido, M. & Casanoves, F. 1998. Plant functional traits and environmental filters at a regional scale. Journal of Vegetation Science 9: 113–122. Dıaz, S., Lavorel, S., McIntyre, S., Falczuk, V., Casanoves, F., Milchunas, D.G., Skarpe, C., Rusch, G., Sternberg, M., (. . .) & Campbell, B.D. 2007. Plant trait responses to grazing – A global synthesis. Global Change Biology 13: 313–341. Dorrough, J., Moxham, C., Turner, V. & Sutter, G. 2006. Soil phosphorus and tree cover modify the effects of livestock grazing on plant species richness in Australian grassy woodland. Biological Conservation 130: 394–405. Espigares, T. & Peco, B. 1995. Mediterranean annual pasture dynamics: impact of autumn drought. Journal of Ecology 83: 135–142. Fox, J. & Weisberg, S. 2011. An R Companion to Applied Regression, 2nd ed. Sage, Thousand Oaks, CA, US. Fukami, T., Martijn Bezemer, T., Mortimer, S.R. & Putten, W.H. 2005. Species divergence and trait convergence in experimental plant community assembly. Ecology Letters 8: 1283– 1290. Gower, J.C. 1971. A general coefficient of similarity and some of its properties. Biometrics 27: 857–874.

Journal of Vegetation Science Doi: 10.1111/jvs.12470 © 2016 International Association for Vegetation Science

157

Community assembly in Mediterranean grasslands

C. Rota et al.

Grime, J.P. 2001. Plant Strategies, Vegetation Processes, and Ecosystem Properties. John Wiley and Sons, Chichester, UK. Grime, J.P. 2006. Trait convergence and trait divergence in herbaceous plant communities: mechanisms and consequences. Journal of Vegetation Science 17: 255–260. Grubb, P.J. 1977. The maintenance of species-richness in plant communities: the importance of the regeneration niche. Biological Reviews 52: 107–145. Jost, L. 2007. Partitioning diversity into independent alpha and beta components. Ecology 88: 2427–2439. Kenney, A.M., McKay, J.K., Richards, J.H. & Juenger, T.E. 2014. Direct and indirect selection on flowering time, water-use efficiency (WUE, d13C), and WUE plasticity to drought in Arabidopsis thaliana. Ecology and Evolution 4: 4505–4521. Kikvidze, Z., Suzuki, M. & Brooker, R. 2011. Importance versus intensity of ecological effects: why context matters. Trends in Ecology & Evolution 26: 383–388. Lalibert e, E., Norton, D.A. & Scott, D. 2013. Contrasting effects of productivity and disturbance on plant functional diversity at local and metacommunity scales. Journal of Vegetation Science 24: 834–842. Lavorel, S., de Bello, F., Grigulis, K., Leps, J., Garnier, E., Castro, H., Dolezal, J., Godolets, C., Qu etier, F. & Th ebault, A. 2011. Response of herbaceous vegetation functional diversity to land use change across five sites in Europe and Israel. Israel Journal of Ecology & Evolution 57: 53–72.  Leps, J., de Bello, F., Smilauer, P. & Dolezal, J. 2011. Community trait response to environment: disentangling species turnover vs intraspecific trait variability effects. Ecography 34: 856–863. Mantel, N. 1967. The detection of disease clustering and a generalized regression approach. Cancer Research 27: 209–220. Mason, N.W.H., de Bello, F., Dolezal, J. & Leps, J. 2011. Niche overlap reveals the effects of competition, disturbance and contrasting assembly processes in experimental grassland communities. Journal of Ecology 99: 788–796. Mason, N.W.H., Richardson, S.J., Peltzer, D.A., de Bello, F., Wardle, D.A. & Allen, R.B. 2012. Changes in coexistence mechanisms along a long-term soil chronosequence revealed by functional trait diversity. Journal of Ecology 100: 678–689. Mayfield, M.M., Bonser, S.P., Morgan, J.W., Aubin, I., McNamara, S. & Vesk, P.A. 2010. What does species richness tell us about functional trait diversity? Predictions and evidence for responses of species and functional trait diversity to landuse change. Global Ecology and Biogeography 19: 423–431. McCune, B. & Keon, D. 2002. Equations for potential annual direct incident radiation and heat load. Journal of Vegetation Science 13: 603–606. McGill, B.J., Enquist, B.J., Weiher, E. & Westoby, M. 2006. Rebuilding community ecology from functional traits. Trends in Ecology & Evolution 21: 178–185. Metz, J., Liancourt, P., Kigel, J., Harel, D., Sternberg, M. & Tielb€ orger, K. 2010. Plant survival in relation to seed size along environmental gradients: a long-term study from

158

semi-arid and Mediterranean annual plant communities. Journal of Ecology 98: 697–704. Milchunas, D.G., Sala, O.E. & Lauenroth, W.K. 1988. A generalized model of the effects of grazing by large herbivores on grassland community structure. The American Naturalist 132: 87–106. Mouillot, D., Graham, N.A.J., Vill eger, S., Mason, N.W.H. & Bellwood, D.R. 2013. A functional approach reveals community responses to disturbances. Trends in Ecology & Evolution 28: 167–177. Oesterheld, M. & Semmartin, M. 2011. Impact of grazing on species composition: adding complexity to a generalized model. Austral Ecology 36: 881–890. Osem, Y., Perevolotsky, A. & Kigel, J. 2002. Grazing effect on diversity of annual plant communities in a semi-arid rangeland: interactions with small scale spatial and temporal variation in primary productivity. Journal of Ecology 90: 936–946. Osem, Y., Perevolotsky, A. & Kigel, J. 2004. Site productivity and plant size explain the response of annual species to grazing exclusion in a Mediterranean semi-arid rangeland. Journal of Ecology 92: 297–309. Pakeman, R.J. 2004. Consistency of plant species and trait responses to grazing along a productivity gradient: a multisite analysis. Journal of Ecology 92: 893–905. Peco, B., De Pablos, I., Traba, J. & Levassor, C. 2005. The effect of grazing abandonment on species composition and functional traits: the case of dehesa grasslands. Basic and Applied Ecology 6: 175–183. Peco, B., Rico, L. & Azc arate, F.M. 2009. Seed size and response to rainfall patterns in annual grasslands: 16 years of permanent plot data. Journal of Vegetation Science 20: 8–16. Peco, B., Carmona, C.P., de Pablos, I. & Azc arate, F.M. 2012. Effects of grazing abandonment on functional and taxonomic diversity of Mediterranean grasslands. Agriculture, Ecosystems and Environment 152: 27–32. Quiroga, R.E., Golluscio, R.A., Blanco, L.J. & Fern andez, R.J. 2010. Aridity and grazing as convergent selective forces: an experiment with an Arid Chaco bunchgrass. Ecological Applications 20: 1876–1889. Rao, C.R. 1982. Diversity and dissimilarity coefficients: a unified approach. Theoretical Population Biology 21: 24–43. Ricotta, C. & Moretti, M. 2011. CWM and Rao’s quadratic diversity: a unified framework for functional ecology. Oecologia 167: 181–188. Roscher, C., Schumacher, J., Lipowsky, A., Gubsch, M., Weigelt, A., Pompe, S., Kolle, O., Buchmann, N., Schmid, B. & Schulze, E.D. 2013. A functional trait-based approach to understand community assembly and diversity–productivity relationships over 7 years in experimental grasslands. Perspectives in Plant Ecology, Evolution and Systematics 15: 139–149. Rosenfeld, J.S. 2002. Functional redundancy in ecology and conservation. Oikos 98: 156–162. Sanchez, A.M., Azcarate, F.M., Arqueros, L. & Peco, B. 2002. Volumen y dimensiones como predictores del peso de

Journal of Vegetation Science Doi: 10.1111/jvs.12470 © 2016 International Association for Vegetation Science

Community assembly in Mediterranean grasslands

C. Rota et al.

semilla de especies herb aceas del centro de la Penınsula Ib erica. Anales Jardın Bot anico de Madrid 59: 249–262. Sasaki, T., Okubo, S., Okayasu, T., Jamsran, U., Ohkuro, T. & Takeuchi, K. 2009. Two-phase functional redundancy in plant communities along a grazing gradient in Mongolian rangelands. Ecology 90: 2598–2608. Schielzeth, H. 2010. Simple means to improve the interpretability of regression coefficients. Methods in Ecology and Evolution 1: 103–113. Sternberg, M., Gutman, M., Perevolotsky, A., Ungar, E.D. & Kigel, J. 2000. Vegetation response to grazing management in a Mediterranean herbaceous community: a functional group approach. Journal of Applied Ecology 37: 224–237. Tutin, T.G., Heywood, V.H., Burges, N.A., Valentine, D.H., Walters, S.M. & Webb, D.A. 1964–1980. Flora Europaea. Cambridge University Press, Cambridge, UK. Vandewalle, M., Purschke, O., de Bello, F., Reitalu, T., Prentice, H.C., Lavorel, S., Johansson, L.J. & Sykes, M.T. 2014. Functional responses of plant communities to management,

landscape and historical factors in semi-natural grasslands. Journal of Vegetation Science 25: 750–759. Whittaker, R.J. 1975. Communities and ecosystems. Macmillan, New York, NY, US. Wolock, D.M. & McCabe, G.J. 1995. Comparison of single and multiple flow direction algorithms for computing topographic parameters in TOPMODEL. Water Resources Research 31: 1315–1324.

Supporting Information Additional Supporting Information may be found in the online version of this article: Appendix S1. Determination of the soil water availability and grazing gradients. Appendix S2. Diversity partitioning.

Journal of Vegetation Science Doi: 10.1111/jvs.12470 © 2016 International Association for Vegetation Science

159

Standard PDF - Wiley Online Library

This article is protected by copyright. All rights reserved. Received Date : 05-Apr-2016. Revised Date : 03-Aug-2016. Accepted Date : 29-Aug-2016. Article type ...

911KB Sizes 12 Downloads 136 Views

Recommend Documents

Standard terminology for the nose and philtrum - Wiley Online Library
16 Jan 2009 - 1Clinical and Molecular Genetics Unit, Institute of Child Health, Great Ormond Street Hospital for Children, UCL, London, UK. 2Department of Pediatrics, Academic Medical Center, UVA, Amsterdam, The Netherlands. 3Department of Medical Ge

PDF(3102K) - Wiley Online Library
Rutgers University. 1. Perceptual Knowledge. Imagine yourself sitting on your front porch, sipping your morning coffee and admiring the scene before you.

TARGETED ADVERTISING - Wiley Online Library
the characteristics of subscribers and raises advertisers' willingness to ... IN THIS PAPER I INVESTIGATE WHETHER MEDIA TARGETING can raise the value of.

Strategies for online communities - Wiley Online Library
Nov 10, 2008 - This study examines the participation of firms in online communities as a means to enhance demand for their products. We begin with theoretical arguments and then develop a simulation model to illustrate how demand evolves as a functio

XIIntention and the Self - Wiley Online Library
May 9, 2011 - The former result is a potential basis for a Butlerian circularity objection to. Lockean theories of personal identity. The latter result undercuts a prom- inent Lockean reply to 'the thinking animal' objection which has recently suppla

Understanding dynamic capabilities - Wiley Online Library
Defining ordinary or 'zero-level' capabilities as those that permit a firm to ... reliance on dynamic capability, by means here termed 'ad hoc problem solving.