Biological Invasions 4: 7–24, 2002. © 2002 Kluwer Academic Publishers. Printed in the Netherlands.
Original article
Grazing, environmental heterogeneity, and alien plant invasions in temperate Pampa grasslands Enrique J. Chaneton∗ , Susana B. Perelman, Marina Omacini & Rolando J.C. Le´on IFEVA – CONICET, Departamento de Recursos Naturales, Facultad de, Agronom´ıa, Universidad de Buenos Aires, Av. San Mart´ın 4453, CPA1417DSE Buenos Aires, Argentina; ∗ Author for correspondence (e-mail:
[email protected]; fax: +54-11-4514 8730) Received 12 October 2001; accepted in revised form 30 May 2002
Key words: disturbance, diversity, environmental fluctuation, exotics, Flooding Pampa, functional groups, herbivory, invasibility, soil fertility gradient, spatial heterogeneity Abstract Temperate humid grasslands are known to be particularly vulnerable to invasion by alien plant species when grazed by domestic livestock. The Flooding Pampa grasslands in eastern Argentina represent a well-documented case of a regional flora that has been extensively modified by anthropogenic disturbances and massive invasions over recent centuries. Here, we synthesise evidence from region-wide vegetation surveys and long-term exclosure experiments in the Flooding Pampa to examine the response of exotic and native plant richness to environmental heterogeneity, and to evaluate grazing effects on species composition and diversity at landscape and local community scales. Total plant richness showed a unimodal distribution along a composite stress/fertility gradient ranging several plant community types. On average, more exotic species occurred in intermediate fertility habitats that also contained the highest richness of resident native plants. Exotic plant richness was thus positively correlated with native species richness across a broad range of flood-prone grasslands. The notion that native plant diversity decreases invasibility was supported only for a limited range of species-rich communities in habitats where soil salinity stress and flooding were unimportant. We found that grazing promoted exotic plant invasions and generally enhanced community richness, whereas it reduced the compositional and functional heterogeneity of vegetation at the landscape scale. Hence, grazing effects on plant heterogeneity were scale-dependent. In addition, our results show that environmental fluctuations and physical disturbances such as large floods in the pampas may constrain, rather than encourage, exotic species in grazed grasslands. Introduction A common feature of many terrestrial ecosystems worldwide is the pervasive occurrence of alien plant species originally from diverse, often distant biogeographic regions. Such biotic exchanges constitute one of the most visible aspects of global changes produced by human activity over recent historical times (Vitousek et al. 1997; Mooney and Hobbs 2000). Temperate grasslands have been particularly susceptible to exotic plant invasions as a result of disturbances associated with cultivation and domestic grazing (Mack
1989; Hobbs and Huenneke 1992). Present-day natural grasslands typically represent modified versions of the original ones, and may sometimes harbour a relatively large exotic flora (e.g. Harrison 1999; Stohlgren et al. 1999a). The self-sustained persistence of exotic plants in their new range, or naturalisation (Richardson et al. 2000), poses serious concerns for the conservation of recipient ecosystems. Moreover, ongoing anthropogenic alterations of environmental drivers and biotic controls (Tilman and Lehman 2001) greatly complicate predictions on the dynamics of invaders in response to management (Zavaleta et al. 2001). Where grasslands
8 are managed for livestock, a conflict deriving from the risk of invasion may arise between preserving native plant diversity through controlled grazing (Collins et al. 1998) and maintaining ecosystem production (Smith and Knapp 1999). Thus, there is an urgent need to understand how exotic versus native species diversity changes as a function of grazing disturbance and habitat heterogeneity in remnant natural grasslands. In principle, there is no reason to believe that exotic plants in their new range will not follow the same basic ‘rules’ determining the spatial distribution and assembly of native species in local communities (Crawley 1987; Lodge 1993; Levine and D’Antonio 1999; Grime 2001). Exotic species diversity can be expected to reflect factors such as the size of species pools, propagule pressure, habitat gradients, biotic interactions, and disturbance (Tilman and Pacala 1993; Huston 1994, 1999; Levine 2000; Shea and Chesson 2002). Theory has traditionally emphasised resource-use patterns and competition intensity as determinants of community invasibility (see Levine and D’Antonio 1999; Wardle 2001). This has led to the notion that species-rich communities may be less invasible (Elton 1958; MacArthur 1972), because they use resources more efficiently and offer fewer ‘niche opportunities’ (Shea and Chesson 2002) than depauperate communities. It follows that the number of established exotic species within a local assemblage, a measure of invasion, should correlate negatively with native plant richness (Lodge 1993; Tilman 1997; Levine 2000). However, exotic species must confront environmental hazards that often limit resident native species as well (Huston 1994; Shea and Chesson 2002). Hence, habitats favouring native plant diversity may also contain a richer exotic flora, thus appearing more invasible than habitats supporting species-poor communities (Levine and D’Antonio 1999; Lonsdale 1999; Stohlgren et al. 1999a; Wardle 2001). Several current models propose that plant species richness depicts a unimodal response along gradients of soil fertility and/or abiotic stress (Tilman and Pacala 1993; Huston 1994; Grime 2001). These habitat gradients are usually surrogates for resource supply (Tilman and Pacala 1993) and potential productivity (Grace 1999), with plant diversity being constrained by physical stress on low-productivity sites and by competitive exclusion on fertile, highly productive sites. Intermediate levels of environmental stress or productivity are thought to enhance plant richness by reducing the potential for competitively superior
species to monopolise resources and become dominant. Although considerable support exists for this diversity–productivity relationship in herbaceous systems (Huston 1994; Grace 1999), little work has been done to determine how patterns of exotic plant richness fit this empirical model (cf. Harrison 1999; Stohlgren et al. 1999a, b). For instance, by generally increasing the likelihood of invasion, recurrent disturbances such as vertebrate grazing may obscure the response of exotic species diversity to existing habitat gradients (Smith and Knapp 1999; Safford and Harrison 2001). It has been shown that vertebrate grazing may enhance grassland diversity by lowering the intensity of competition, or by increasing microhabitat heterogeneity and the number of colonisation gaps (e.g. McNaughton 1983; Collins et al. 1998; Olff and Ritchie 1998). Yet, grazing may not only increase the richness of subordinate or rare native species, but could also facilitate the establishment of exotic plants (Mack 1989; Hobbs and Huenneke 1992). Such grazinginduced shifts in species composition can be particularly dramatic in humid grasslands (Milchunas and Lauenroth 1993). On the other hand, the magnitude and nature of grazing effects on plant diversity may change across scales (McNaughton 1983; Chaneton and Facelli 1991), as well as over soil fertility gradients (Stohlgren et al. 1999b; Safford and Harrison 2001). These complexities have not always been appreciated in generalisations about the impact of herbivory on grasslands (Olff and Ritchie 1998). It is therefore far from obvious what local consequences the removal of grazers will have on the exotic/native balance in systems with a rich exotic species pool (Zavaleta et al. 2001). In this paper, we analyse patterns in exotic versus native species diversity as influenced by environmental heterogeneity and grazing for a vast region of natural grasslands in the Flooding Pampas of eastern Argentina. The study system represents a welldocumented case of a regional flora that has been extensively modified by the massive invasion of exotic plants induced by human activity during post-settlement times (Darwin 1839; Hudson 1892, 1918; Soriano 1992; Rapoport 1996). Vegetation heterogeneity in the Flooding Pampa is strongly determined by landscape-level patterns in topography and soil chemistry (Perelman et al. 2001), which interact with cattle grazing to control plant community structure (Sala et al. 1986). Here, we synthesise evidence from region-wide vegetation surveys and long-term exclosure studies in the
9 Flooding Pampa to address several specific hypotheses: (1) native and exotic species richness show similar responses along landscape gradients of soil fertility and habitat stress; (2) exotic plant richness in local communities may not be limited by native species richness, and so they are positively correlated across a broad range of habitats and community types; (3) grazing is expected to favour invasions, although its impact on plant diversity will vary with the spatial scale of observation (see Sala et al. 1986; Chaneton and Facelli 1991); and (4) exotic species abundance in local communities fluctuates in response to enviromental variation, being negatively affected during wet years characterised by severe flooding events (Chaneton et al. 1988; Insausti et al. 1999). We begin by describing the contribution of exotic species to the Flooding Pampa flora; we then examine patterns in plant diversity and composition in response to habitat gradients and to livestock grazing at various scales, and we end by assessing the impact of interannual abiotic fluctuations and extreme events on exotic species invading flood-prone environments. The study system The Flooding Pampa covers 90,000 km2 in the province of Buenos Aires, between 35◦ and 38◦ S latitude in eastcentral Argentina (Soriano 1992; Perelman et al. 2001). The regional climate is temperate subhumid with a mean annual rainfall varying from ca. 1000 mm in the north (R´ıo Salado basin) near Buenos Aires to 850 mm in the south (Laprida basin). Mean annual temperatures range north–south between 15.9 ◦ C and 13.8 ◦ C. The landscape is characterised by its extremely flat topography and treeless physiognomy. Soils are finetextured Mollisols and Alfisols varying from slightly acidic (pH = 6.2–6.8) to alkaline (pH = 8.1–8.7) in the top horizon. Soil drainage is generally impeded by the lack of slope and reduced infiltration, except in ridge areas with well-drained sandy soils. The region is widely devoted to cattle husbandry on unfertilised, natural grasslands. Less than 20% of the area has ever been cultivated due to limitations imposed by soil properties and periodic flooding (Soriano 1992). Floods occur almost annually during autumn–spring in lowland areas; extensive flooding events lasting over one month are less frequent but have drastic effects on grassland structure (Chaneton et al. 1988; Alconada et al. 1993) and function (Ginzo et al. 1986; Insausti et al. 1999). The natural history of plant invasions in the pampas has attracted considerable attention ever since
Darwin (1839) first described the widespread dominance of exotic thistles near Buenos Aires. Although livestock was introduced to the pampas by European settlers in the late 1500s, historical accounts (Hudson 1892, 1918) indicate that the most profound, irreversible modification of native grasslands by grazing took place from 1880 onwards (Soriano 1992; Rapoport 1996). As early as 1833, when crossing the central Flooding Pampas, Darwin (1839) asked whether differences in vegetation structure between areas subjected to contrasting levels of grazing were ‘owing to the introduction of new species, to the altered growth of the same, or to a difference in their proportional numbers’. It seems then likely that botanical changes in these pampean grasslands involved both a shift in relative abundance of prostrate over tussock grasses and the expansion of alien plants (see Mack 1989). At present, Flooding Pampa grasslands comprise a fine-grained mosaic of plant community types (Perelman et al. 2001). We focused on the shortgrass-like vegetation that covers most of the region under current grazing conditions, excluding tall tussock grasslands dominated by Paspalum quadrifarium, which are managed through prescribed burning and hence become heavily invaded by exotic thistles and legumes (Laterra 1997). Our terminology for plant invaders follows that recommended by Richardson et al. (2000); the term ‘exotic’ (= alien) refers to plant species occurring in a given area because they were accidentally or intentionally introduced from elsewhere as a result of human activities. In our case, the majority of exotics are herbaceous taxa originating from Eurasia, which presumably entered the R´ıo de la Plata region after the Spanish settlement in southern South America in the sixteenth century (Cabrera 1963–1970; Cabrera and Zardini 1978; S¨oyrinki 1991). Little historical data exists to estimate with any accuracy how long different exotic taxa have been present in the pampas, yet most of the currently naturalised alien plants were likely introduced >100 years ago (Rapoport 1996).
Methods Regional floristic surveys We based our analysis on three independent data sets of different spatial extent and resolution. The first one corresponded to a region-wide survey of grassland
10 vegetation comprising 749 censuses from four phytosociological inventories carried out at four contrasting latitudes within the study region (35◦ 20 S–37◦ 20 S; Perelman et al. 2001). The spatial location of floristic censuses in each latitudinal survey was guided by detailed physiographic analyses using aerial photographs (scale = 1 : 20,000) to help identify major geomorphological patterns and landscape elements. Sampling locations in the field were selected to encompass most of the existing variation in land forms and associated soil types, as revealed by the aerial analysis and by soil maps of comparable scale. All sampled stands occurred on grassland sites, which were managed for cattle grazing by private landowners and had not been cultivated for the prior eight years. Each census contained a full list of vascular plant species in a 0.25-ha stand and a cover estimate per species; sampling was conducted in early summer when most flowering plants could be easily identified. We compiled species richness data from these floristic surveys to quantify the contribution of exotics to the regional species pool. To characterise the presence of different life-forms in the exotic flora, all herbaceous alien species occurring in our censuses were classified into grasses, forbs, annuals and perennials. We also separated exotic species according to their phenological pattern between cool-season and warm-season plants (Sala et al. 1981; Rusch and Oesterheld 1997). Differences among the four latitudinal surveys in the proportional richness of various life-forms of exotic plants were evaluated using χ 2 tests of homogeneity. Landscape-level patterns of diversity and invasibility Recently, Perelman et al. (2001) analysed this floristic database to produce a synthesis of regional- and landscape-level compositional patterns for Flooding Pampa grasslands. These authors classified vegetation stands using cluster analysis (presence–absence data) into several community types, which ranged from prairies on well-drained soils typical of elevated topographic positions, to various meadows and steppes on either acidic or alkaline/saline soils occurring in the flood-prone lowlands. These plant community types form intricate landscape mosaics closely associated with differences in soil chemistry and topographic height. Such vegetation-habitat mosaics are repeated at different latitudes across the entire region, so that latitudinal factors play a minor role in determining
the floristic heterogeneity of these grasslands (see Perelman et al. 2001). We used this extensive floristic database to examine landscape-level patterns in native and exotic species diversity along a generalised habitat stress/fertility gradient. This composite gradient was represented by the amount of soil organic carbon (%), which in our system is positively correlated with soil nitrogen content (Berasategui and Barberis 1982; Collantes et al. 1988; Batista 1991; Chaneton and Lavado 1996), and thus provides a crude measure of resource supply. The term habitat ‘stress’ is used in the sense of reduced soil C content (or fertility), a condition that in these grasslands corresponds with highly saline soils supporting low-productivity, halophyte steppes (Ginzo et al. 1986; Alconada et al. 1993). We focused on soil fertility because it bears on current models proposed to explain plant diversity–productivity relations (Tilman and Pacala 1993; Huston 1994; Grace 1999), and it has been perceived as an important factor for exotic species distribution in other grasslands (Stohlgren et al. 1999a, b; Grime 2001). The use of other abiotic variables (topographic height, flooding) as a basis for integrating diversity patterns in the study system would have produced misleading vegetation–habitat relations, by putting together stands from community types (lowland acidic meadows and halophyte steppes) that become similarly flooded, but are also known to greatly differ in soil chemistry, plant cover, and primary productivity (Sala et al. 1981; Ginzo et al. 1982, 1986; Collantes et al. 1988; Perelman et al. 2001). Soil carbon data were obtained from published studies having an adequate description of the local plant community (Berasategui and Barberis 1982; Collantes et al. 1988; Lavado and Taboada 1988; Lavado et al. 1992; Chaneton and Lavado 1996; Alvarez and Lavado 1998). Soil organic C ranged from 0.8% to 6.5% across grassland habitats varying from halophyte steppes (low fertility) to acidic meadows (high fertility), through mesic prairies in intermediate fertility soils (2.6–3.5% C). These habitats range from ca. 150 to 600 g m−2 yr−1 in total aboveground plant productivity (Sala et al. 1981; Ginzo et al. 1982, 1986; Hidalgo and Cauh´ep´e 1991). The plant richness–soil fertility relationship could be determined only at the level of plant community types, because the available soil data had a lower spatial resolution than our vegetation database. Hence, we assigned the same soil carbon value to the mean richness of all stands that had been classified into the same community type and shared the
11 same topographic position and soil type (Perelman et al. 2001). Since the vegetation gradient was described by a rank order of community types corresponding to different soil fertility levels, differences in their mean species richness were evaluated using one-way ANOVA followed by Tukey tests (P < 0.05) for unequal sample sizes (n = 47–102 stands/community). The relationship between exotic and native species richness was evaluated through simple regression analysis including all stands sampled as part of the regional survey (n = 749). This analysis comprised the whole range of grassland habitats within the study region. In addition, we separately tested for correlations between the local richness of native and exotic plants in grasslands occurring on well-drained soils (n = 432) and for grasslands in regularly flooded environments (n = 317). The latter analysis examined invasibility patterns for contrasting habitat ranges characterised by different overall levels of plant richness (see Results).
Grazing effects at landscape and local-community scales To examine grazing impacts at landscape and local scales, we used a second data set obtained as part of a long-term exclosure study in flood-prone acidic grasslands at the centre of the R´ıo Salado basin, ca. 200 km south of Buenos Aires (36◦ 30 S, 58◦ 30 W). We monitored plant community dynamics in grazed and ungrazed (exclosure) plots established on two sites located 6 km apart, at slightly different topographic positions (relative height difference ≈0.25 cm). These sites were chosen to represent two common plant community types, hereafter referred to as ‘lowland’ and ‘upland’ communities (Sala et al. 1986; Chaneton and Lavado 1996). Soil carbon content varied from 3.1% to 4.7% and total soil N from 0.28% to 0.42% in the upland and lowland grasslands, respectively. Both sites were grazed by domestic cattle on a year-round basis (stocking rate = 0.55 cow ha−1 yr−1 ), being located in the same ranch (Estancia ‘Las Chilcas’, Pila County). The area is flooded almost every year during winter–spring; depth and duration of floods are typically greater in the lowland site. In each grassland site, one exclosure was established on a formerly homogeneous stand, so that vegetation inside and outside the fenced area could not be initially distinguished with regard to floristic composition, canopy structure, and soil substrate (R.J.C Le´on, pers.
obs.). The exclosures were constructed in 1972 (lowland: 2 ha) and 1983 (upland: 4 ha), and were annually sampled from 1985 (lowland) and 1989 (upland) through 2000, in January, at the time of peak plant richness. In each year, plant-basal cover was measured within two 0.1-ha permanent plots marked inside each exclosure and on the adjacent grazed area, using four 5 m long linear transects placed at random in each plot. The distance between samples in grazed and ungrazed plots ranged from 100 to 250 m; samples within a given plot were at least 20 m apart. To determine how grazing may influence vegetation heterogeneity at the landscape level, we assessed the magnitude of floristic differences between upland and lowland grasslands subjected to either grazed or ungrazed conditions. The working hypothesis was that grazing decreases floristic differences between communities occurring at different topographic positions (Sala et al. 1986). A 4-year (1989–1992) subset of plant cover data from the upland and lowland sites (grazed vs exclosure) was subjected to non-metric multidimensional scaling (NMDS), using the PC-ORD package (McCune and Mefford 1999). This iterative ordination technique summarises compositional differences among a set of samples with the only constraint being that the relative distances between samples in an ordination space of specified dimensions (e.g. two or three) should reflect the ranking of pairwise sample differences in a distance matrix (Digby and Kempton 1987). NMDS has the advantage of not assuming a particular model for species responses along environmental gradients. This non-parametric method best served our purpose of comparing grassland plots that were known to differ in grazing regime and habitat conditions. The initial sample configuration used to feed NMDS was obtained through detrended correspondence analysis (Digby and Kempton 1987). We used Sorensen’s quantitative dissimilarity index (Magurran 1988) to calculate pairwise floristic distances among samples (excluding species present in 0.25). On average, exotics accounted for 15% and 3.5% of total plant richness in grazed versus ungrazed areas, respectively. In contrast, grazing did not have a consistent impact on native forbs richness (F1,1 = 21.9, P = 0.13), as its mean (positive) effect size on this group varied between lowland and upland plots (grazing × block: F1,22 = 3.85, P = 0.06) and with the study year (F11,22 = 2.14, P = 0.06). There was a slightly positive impact of grazing on native grass richness (Figure 7; F1,1 = 81.4, P = 0.07), but the observed effect was again influenced by the sampling year (F11,22 = 3.28, P = 0.01). Scale dependence of diversity in grazed versus ungrazed grassland Species–area curves for the lowland grassland suggested that grazing differentially affected species richness over a range of spatial scales within the same habitat (Figure 8). The log (species)–log (area) relationship was much steeper for the ungrazed exclosure (z = 0.331, SE = 0.006; intercept = 0.53, SE = 0.02) than for the adjacent grazed grassland (z = 0.193, SE = 0.004; intercept = 0.99, SE = 0.01). The predicted enhancement of plant richness under continuous grazing was only evident at scales 0.40).
Discussion Our figures for exotic species in the Flooding Pampa flora match those of S¨oyrinki (1991), who reported 23% exotics out of a total of 1730 plant species in the area around Buenos Aires (after Cabrera and Zardini 1978), and those for the whole province of Buenos Aires where exotics have been estimated to contribute up to 25% of all 2106 vascular plants (Rapoport 1996). The proportion of exotics in Flooding Pampa grasslands falls near the upper boundary of the range (8–24%) reported by Stohlgren et al. (1999a) for various North American central plain grasslands. Interestingly, annual cool-season forbs predominate among the exotic species found in the regional surveys. Such a biased representation of life-forms within the established exotic flora could reflect the action of major environmental constraints favouring invaders with certain functional traits. Alternatively, it might represent a proportional ‘sample’ of successful invaders from the pool of exotics that has reached the pampas (see Crawley 1987; Shea and Chesson 2002). Landscape patterns of diversity and invasibility The observed patterns of variation in native and exotic species richness have various important implications. First, they conformed to the well-known unimodal relationship between plant richness and habitat stress/fertility gradients found across plant communities in other systems (Tilman and Pacala 1993; Huston 1999; Grime 2001). Moreover, this pattern applied to native and exotic species alike, suggesting that equivalent processes govern the coexistence and spatial distribution of diversity for both sets of plant species. All sampled grasslands were grazed by domestic herbivores. It is then intriguing that the unimodal pattern held, despite the potentially confounding
influence of grazing (see Olff and Ritchie 1998; Safford and Harrison 2001). However, in the Flooding Pampa, grazing appears to similarly increase species richness in different grasslands and habitat types (Sala et al. 1986; Chaneton and Facelli 1991; Alconada et al. 1993; Rusch and Oesterheld 1997). Second, factors limiting native richness at both ends of the soil fertility gradient appear to be strong enough to reduce the diversity of exotics occurring in those habitats, although mechanisms operating on each extreme of this gradient are likely to be different (Huston 1994, 1999; Grace 1999; Grime 2001; Wardle 2001). The low fertility end of the gradient (Figure 2) comprises halophyte steppes in which plant cover is usually less than 40% (Ginzo et al. 1986; Alconada et al. 1993). In those sites, stressful conditions produced by low nutrient supply and high soil salinity would determine that only a reduced number of species were capable of becoming established. Towards the more elevated, mesic prairie habitats, increased soil resources coupled with a reduction in salinity stress (Perelman et al. 2001) would result in better conditions for more native species, and also for a larger number of exotics. On the fertile end of the gradient, floodprone acidic meadows have nearly 100% cover and thus establishment opportunities are probably reduced (see Huston 1994; Wardle 2001). The dominant native grasses in these habitats are well adapted to flooding (Rubio et al. 1995) and show enhanced biomass productivity during prolonged floods, excluding subordinate forbs (Ginzo et al. 1986; Alconada et al. 1993; Insausti et al. 1999). Thus, in high fertility habitats, flooding may indirectly limit invasion by increasing the competitive dominance of flood-tolerant graminoids (Chaneton et al. 1988; Chaneton and Facelli 1991; Insausti et al. 1999). In addition, only a few exotic species could be able to tolerate direct physical stress from frequent soil waterlogging. Third, our results imply that native plant diversity may not reduce invasibility across a range of grassland types, including both stressful habitats in saline/infertile soils and high fertility habitats in acidic soils (Figures 2 and 3a). If conditions favouring species coexistence, e.g. moderate fertility and grazing, act in preventing competitive exclusion, they could also decrease community resistance to exotic invasions (Huston 1994, 1999; Wardle 2001). This apparent discrepancy between classical theory and results from correlative studies like ours (see Levine and D’Antonio 1999) has various interpretations. Naeem et al. (2000)
19 pointed out that covarying extrinsic factors obscure the effect of competition and diversity on the realised success of invaders in broad-scale observational studies. In contrast, Wardle (2001) argued for the universal role of resource availability and competition in determining plant diversity and invasibility along stress/fertility gradients. However, our finding that the proportional richness of exotics decreased with increased native richness only across well-drained grassland sites suggests that local species intractions and the community position along habitat gradients are both important controls on invasibility (Levine 2000; Shea and Chesson 2002). We hypothesise that an upper boundary may exist to plant diversity in the more species-rich grasslands under current resource supply rates, grazing regime, and propagule pressure. Only manipulative experiments replicated across relevant habitat gradients may help distinguish the factors underlying these invasibility patterns.
Grazing and vegetation heterogeneity across habitat gradients One major consequence of invasions is the reduction of compositional differences at relatively large spatial scales, a process termed ‘biotic homogenisation’ (see Vitousek et al. 1997; Lockwood and McKinney 2001). In the Flooding Pampa, this phenomenon is shown on a regional scale by the fact that exotic species are less narrowly associated with particular habitats and community types than native species (Perelman et al. 2001). However, it is difficult to assess the role grazing has had in biotic homogenisation at such large scales since no unmanaged relicts (or natural reserves) of native grassland are available for comparison. Therefore, we turned to exclosure studies conducted along a topographic gradient. Our observations in lowland and upland grasslands (Figures 4 and 5) supported the hypothesis that grazing decreases vegetation heterogeneity across landscape habitat gradients in this ecosystem. Sala et al. (1986) previously showed that seasonal differences between communities at contrasting topographic positions increased after a few years of cattle exclusion. We found that floristic differences between ungrazed grasslands persisted after >10 years of protection from grazing, and that the similarity between grazed lowland and upland areas was unaffected by interannual climatic variation. This botanical homogenisation may not be driven by changes in soil conditions, as grazing seems
to exaggerate rather than reduce differences in soil chemistry along topographic gradients (Chaneton and Lavado 1996). Instead, colonisation/extinction processes comprising the increased abundance of exotic forbs, and a decline of different grass functional groups in each habitat (Figure 5) apparently accounted for reduced grassland heterogeneity under grazing. Our results show that grass species responses to herbivore removal were strongly habitat-specific, and thus could be interpreted in terms of differential adaptation of grass functional types to prevailing flooding regimes (Sala et al. 1986). While cool-season grasses show a productivity peak in the spring, when floods are most common, warm-season grasses concentrate their growth in summer–autumn. Warm-season grasses appear better adapted to the highly fluctuating water conditions of this ecosystem (Soriano 1992), and would therefore become dominant in frequently flooded habitats. Furthermore, the floristic changes induced by grazing removal suggest that grazed grasslands are relatively less heterogeneous from a functional viewpoint, since the observed species turnover (Figure 5) would cause increased concentration of the primary production during winter–spring in both grassland habitats. Grazing, plant invasions, and local community structure Understanding the processes that influence exotic species colonisation and persistence requires a focus on community dynamics (Tilman 1997; Grime 2001). Local processes affecting species coexistence such as competition and gap disturbance may in turn help explain patterns of invasion and diversity at landscapeto-regional scales (Huston 1994, 1999). Where grasslands have been extensively modified by livestock, opportunities to directly evaluate how herbivory may influence plant invasions are rare. Our approach was to reintroduce cattle to an experimentally protected grassland dominated by a diverse mix of native grasses, and to compare the dynamics of this newly grazed community with that of a nearby exclosure (see also McNaughton 1983). A remarkable replacement of plant growth forms was observed in the newly grazed area. Different growth habits would reflect the potential for different plant species to avoid or tolerate herbivory, and to capture resources under sustained grazing (Mack and Thompson 1982; Noy-Meir et al. 1989). Hence,
20 a decline in tussock grasses was followed by a compensatory increase in prostrate grasses, which horizontally expanded through clonal growth upon disturbance of the original canopy (McNaughton 1983; Noy-Meir et al. 1989). The transient colonisation by L. multiflorum was consistent with the idea that altered light conditions in grazed canopies may trigger seedling emergence from this species (Deregibus et al. 1994) and other opportunistic species (see Oesterheld and Sala 1990; Insausti et al. 1995). Although grazing-induced changes in community structure were relatively rapid and, overall, re-created a mixed exotic/native plant assemblage, our results support the observation that seed banks of exotic species may become locally depleted in long-term ungrazed sites (Oesterheld and Sala 1990). Re-invasion by exotic rosette-forming forbs would require seed dispersal from nearby grazed areas (Facelli 1988). We suggest that low immigration rates coupled with competition from prostrate grasses may limit the colonisation of disturbed grassland by exotic forbs. These dynamics agree with those reported by Facelli (1988) in the only other study in which grazing effects were examined on a nearby protected grassland. Several studies in the Flooding Pampa have shown that in the absence of grazing, tussock grasses dominate low-growing exotic and native forbs (Sala et al. 1986; Facelli et al. 1989; Rusch and Oesterheld 1997). It has therefore been suggested that domestic herbivores increase plant community richness (e.g. Sala et al. 1986; Chaneton and Facelli 1991; Rusch and Oesterheld 1997). However, all previous studies comparing grazed and exclosure areas in this system relied on just 1–2 years of observation. Our results indicate that grazing effects on species richness can be strongly influenced by temporal dynamics presumably associated with climatic fluctuation (Figure 10b), as well as by site-specific responses from different plant communities (Figure 7). It is noteworthy that these sources of variation have rarely been considered when addressing the impact of herbivory on grassland diversity (but see McNaughton 1983; Collins et al. 1998; Stohlgren et al. 1999b). Specifically, we found that the local incidence of native forb species in grazed grasslands was very much dependent on such spatiotemporal variability, and hence their richness was not consistently affected by grazing. Meanwhile, grazing was a chief factor controlling the local richness of exotic species, even within a highly fluctuating environmental context. These exotics are typically cool-season forbs that
colonise gaps created by cattle grazing (Oesterheld and Sala 1990; Deregibus et al. 1994) or prolonged flooding (Insausti et al. 1995). Grazing effects on plant diversity may also vary across spatial scales (Olff and Ritchie 1998; Stohlgren et al. 1999b), much like both species richness and heterogeneity are scale-dependent phenomena (Crawley and Harral 2001). In some systems, grazing increases spatial heterogeneity at within-community scales (McNaughton 1983; Knapp et al. 1999). However, Olff and Ritchie (1998) pointed out that herbivory may reduce plant diversity on large scales, while still enhancing it on smaller scales, through selection of a limited pool of grazing-tolerant species. Taken together, our species–area relationships and quadratscale diversities (Figure 9) suggest that herbivory may decrease community heterogeneity via two related processes. Firstly, grazing prevents dominance by a few tussock grasses that competitively reduce small-scale diversity when not grazed (Facelli et al. 1989); secondly, it allows a diverse group of subordinate species to coexist at small scales (Chaneton and Facelli 1991; Rusch and Oesterheld 1997). Whether observed differences in the steepness of species–area curves between grazed and ungrazed vegetation reflect differences in microhabitat heterogeneity (see Crawley and Harral 2001) is not known. The fine-grained packing of species in grazed areas might, in part, reflect plant density effects, as they contain greater densities of small tussocks compared with ungrazed areas (Sala et al. 1986; Facelli et al. 1989). Fragmentation of the coarse-grained grass matrix by grazing may be an important process mediating the increase in small-scale plant diversity. Environmental fluctuation, flooding, and community dynamics Invading species face various hazards generated by changing habitat conditions and extreme events. This environmental variability may itself constitute a barrier to the spread of alien plants intolerant to physical stresses characteristic of the focal system (Shea and Chesson 2002). Most areas in the Flooding Pampa are affected by periodic floods of varied depth and duration that alternate with seasonal droughts (Lavado and Taboada 1988). Both observational studies (Chaneton et al. 1988; Chaneton and Facelli 1991; Alconada et al. 1993) and mesocosm experiments (Insausti et al. 1999) showed that discrete flooding events produce profound
21 changes in plant community composition and diversity. Large floods reduce the cover of forb species including widespread invaders and native weedy plants, whereas they increase the biomass of native graminoids (Insausti et al. 1999) adapted to soil waterlogging (Rubio et al. 1995). Our long-term data series shows that changing environmental conditions strongly altered the performance of exotic plants in grasslands subjected to slightly different flooding regimes (Figure 10). Exotic species were negatively affected during wet years with large floods, in what effectively represented a temporary reversal of compositional changes induced by continuous grazing (see also Chaneton et al. 1988; Insausti et al. 1999). Interestingly, exotic species apparently lagged behind interannual fluctuations in the hydric environment. At present, we may only conjecture that such a delayed response to environmental variation could either reflect direct (demographically driven) or indirect (competition-driven) effects of previousyear conditions on the population dynamics of exotic species (cf. Silvertown et al. 1994). Insausti et al. (1999) proposed two mechanisms for the negative effect of flooding on exotic forbs that are broadly compatible with these hypotheses. On the one hand, flood-intolerant species may be directly killed by physiological stress and should hence recover slowly from in situ propagule sources and/or immigration (Insausti et al. 1995). On the other hand, flooding may increase the amount of competitive interference on next-year exotic recruitment by enhancing the cover of flood-tolerant grasses (Chaneton et al. 1988). Likewise, we suggest that severe droughts might encourage recolonisation of exotics through decreased competition from grasses and by creating gaps for recruitment (Silvertown et al. 1994).
Conclusions Both physical and biotic factors may constrain plant invasion, though controversy persists over which prevail once propagule pressure of invaders is factored out (Tilman 1997; Lonsdale 1999; Levine and D’Antonio 1999; Naeem et al. 2000; Wardle 2001). Our regionwide survey of grazed grasslands showed that exotic species diversity changed across stress/fertility gradients following the same ‘hump-backed’ distribution as that observed for native plant diversity. In general, more exotics occurred in intermediate fertility habitats
also containing the higher richness of indigenous plants (cf. Stohlgren et al. 1999a). This pattern is opposite to that predicted by classical invasion models relying on resource-use patterns and competition (Levine and D’Antonio 1999; Naeem et al. 2000; Wardle 2001). We did find, however, correlative evidence consistent with the hypothesis that native plant diversity may decrease invasibility, but only for a limited range of species-rich grasslands in which salinity and flooding stress were unimportant, and where moderate resource levels could presumably reduce competition from resident plants (Huston 1999; Wardle 2001). Our results suggest that not all disturbance agents will encourage the increased presence of exotic species, which contradicts a basic tenet of invasion theory (Elton 1954; Crawley 1987; Lodge 1993). Physical disturbances that played a major role in the ecosystem’s history, such as fire in North American tallgrass prairie (Collins et al. 1998; Smith and Knapp 1999) and flooding in the pampa grasslands studied here (Chaneton et al. 1988; Insausti et al. 1999), may constrain the local abundance and potential spread of exotic plants as they reinforce native grass dominance and cause selective mortality of invasive species. Grazing has been used as a management tool for preserving diversity in grasslands where conditions lead to native grass dominance (Collins et al. 1998; Knapp et al. 1999; Grime 2001). However, in systems where exotic invasions represent a threat to grazed vegetation, it may be more appropriate to focus on the identity of colonisers rather than on just total diversity (Smith and Knapp 1999; Safford and Harrison 2001). We suggest that generalisations on the beneficial role of grazing in maintaining diversity can be risky for at least two reasons. First, grazing enhances plant richness at within-community scales, but it may also produce biotic homogenisation at larger scales. In the Flooding Pampa, local extinction–colonisation dynamics driven by grazing exert a strong bottom-up influence on grassland heterogeneity at the landscape scale. Second, the switch towards a mixed native/exotic plant assemblage not only reduces the functional heterogeneity of vegetation at various spatiotemporal scales (Perelman et al. 2001), but may alter ecosystem-level processes such as primary productivity (Rusch and Oesterheld 1997) and disturbance resistance (Chaneton et al. 1988). The extant diversity of naturalised plant invaders warrants a stronger focus on the impact of different functional groups of exotics on native plant diversity and ecosystem functioning.
22 Acknowledgements We thank Roxana Arag´on and Diego V´azquez for inviting us to contribute to this special feature. We are especially grateful to W. Batista, S. Burkart, H. Trebino, R. Lavado, and M. Taboada for their contribution to the work presented here. A. Grimoldi, N. Maz´ıa, R. Arag´on, D.V´azquez, and three anonymous reviewers provided insightful comments that helped improve the content and organisation of the manuscript. Our research in the pampas has been supported by the University of Buenos Aires, Consejo Nacional de Investigaciones Cient´ıficas y T´ecnicas (CONICET) and Agencia Nacional de Promoci´on Cient´ıfica y T´ecnol´ogica (FONCYT) of Argentina.
References Alconada M, Ansin OE, Lavado RS, Deregibus VA, Rubio G and Gutierrez Boem FH (1993) Effect of retention of run-off water and grazing on soil and on vegetation of a temperate humid grassland. Agricultural Water Management 23: 233–246 Alvarez R and Lavado RS (1998) Climate, organic matter and clay content relationships in the Pampa and Chaco soils, Argentina. Geoderma 83: 127–141 Batista WB (1991) Correspondencia entre comunidades vegetales y factores ed´aficos en el pastizal de la Pampa Deprimida. MSc thesis, University of Buenos Aires, Buenos Aires Berasategui LA and Barberis LA (1982) Los suelos de la regi´on de Castelli-Pila, Depresi´on del Salado (Provincia de Buenos Aires). Revista de la Facultad de Agronom´ıa 3: 13–25 Cabrera AL (1963–1970) Flora de la Provincia de Buenos Aires. Colecci´on Cient´ıfica, Instituto Nacional de Tecnolog´ıa Agropecuaria, Buenos Aires Cabrera AL and Zardini EM (1978) Manual de la flora de los alrededores de Buenos Aires. Acme, Buenos Aires Chaneton EJ and Facelli JM (1991) Disturbance effects on plant community diversity: spatial scales and dominance hierarchies. Vegetatio 93: 143–155 Chaneton EJ and Lavado RS (1996) Soil nutrients and salinity after long-term grazing exclusion in a Flooding Pampa grassland. Journal of Range Management 49: 182–187 Chaneton EJ, Facelli JM and Le´on RJC (1988) Floristic changes induced by flooding on grazed and ungrazed lowland grasslands in Argentina. Journal of Range Management 41: 495–499 Collantes MB, Kade M, Miaczynski C and Santanatoglia O (1988) Distribuci´on de las especies en funci´on de factores ed´aficos en un pastizal natural de la Depresi´on del R´ıo Salado (Prov. de Buenos Aires, Argentina). Studia Oecologica 5: 77–93 Collins SL, Knapp AK, Briggs JM, Blair JM and Steinauer EM (1998) Modulation of diversity by grazing and mowing in native tallgrass prairie. Science 280: 745–747 Crawley MJ (1987) What makes a community invasible? In: Gray AJ, Crawley MJ and Edwards PJ (eds) Colonization, Succession and Stability, pp 429–453. Blackwell Scientific, Oxford
Crawley MJ and Harral JE (2001) Scale dependence in plant biodiversity. Science 291: 864–868 Darwin C (1893) Journal of Researches into the Geology and Natural History of the Various Countries Visited by the H.M.S. Beagle. Henry Colburn, London Deregibus VA, Casal JJ, Jacobo EJ, Gibson D, Kauffman M and Rodriguez AM (1994) Evidence that heavy grazing may promote the germination of Lolium multiflorum seeds via phytochromemediated perception of high red/far-red ratios. Functional Ecology 8: 536–542 Digby PGN and Kempton RA (1987) Multivariate Analysis of Ecological Communities. Chapman & Hall, London Elton CS (1958) The Ecology of Invasions by Plants and Animals. Chapman & Hall, London Facelli JM (1988) Response to grazing after nine years of cattle exclusion in a Flooding Pampa grassland, Argentina. Vegetatio 78: 21–25 Facelli JM, Le´on RJC and Deregibus VA (1989) Community structure in grazed and ungrazed grassland sites in the flooding Pampa, Argentina. American Midland Naturalist 121: 125–133 Forbes SP, Schauwecker T and Weiher E (2001) Rarefaction does not eliminate the species richness-biomass relationship in calcareous blackland prairies. Journal of Vegetation Science 12: 525–532 Ginzo HD, Collantes MB and Caso OH (1982) Fertilization of a native grassland in the ‘Depresi´on del R´ıo Salado’, Province of Buenos Aires: herbage dry matter accumulation and botanical composition. Journal of Range Management 35: 35–39 Ginzo HD, Collantes MB and Caso OH (1986) Fertilization of a halophytic natural grassland in Argentina: herbage dry matter, botanical composition, and mineral content. Turrialba 36: 453–459 Grace JB (1999) The factors controlling species density in herbaceous plant communities: an assessment. Perspectives in Plant Ecology, Evolution and Systematics 2: 1–28 Grime JP (2001) Plant Strategies, Vegetation Processes, and Ecosystem Properties, 2nd edn. John Wiley & Sons, Chichester Harrison S (1999) Native and alien species diversity at the local and regional scales in a grazed California grassland. Oecologia 121: 99–106 Hidalgo LG and Cauh´ep´e MA (1991) Producci´on de forraje de las comunidades de la Depresi´on del Salado. Revista CREA 149: 58–62 Hobbs RJ and Huenneke LF (1992) Disturbance, diversity, and invasion – implications for conservation. Conservation Biology 6: 324–337 Hudson WH (1892) The Naturalist in La Plata. Chapman & Hall, London Hudson WH (1918) Far Away and Long Ago. JM Dent & Sons, London Huston MA (1994) Biological Diversity: The Coexistence of Species on Changing Landscapes. Cambridge University Press, Cambridge Huston MA (1999) Local processes and regional patterns: appropriate scales for understanding variation in the density of plants and animals. Oikos 86: 393–401 Insausti P, Soriano A and Sanchez RA (1995) Effect of flood-induced factors on seed germination of Ambrosia tenuifolia. Oecologia 103: 127–132
23 Insausti P, Chaneton EJ and Soriano A (1999) Flooding reverted grazing effects on plant community structure in mesocosms of lowland grassland. Oikos 84: 266–278 Knapp AK, Blair JM, Briggs JM, Collins SL, Hartnett DC, Johnson LC and Towne EG (1999) The keystone role of bison in North American tallgrass prairie. Bioscience 49: 39–50 Laterra P (1997) Post-burn recovery in the flooding pampa: impact of an invasive legume. Journal of Range Management 50: 274–277 Lavado RS and Taboada MA (1988) Water, salt and sodium dynamics in a natraquoll in Argentina. Catena 15: 577–594 Lavado RS, Rubio G and Alconada M (1992) Grazing management and soil salinization in two pampean natraqualfs. Turrialba 42: 500–508 Levine JM (2000) Species diversity and biological invasions: relating local process to community pattern. Science 288: 852–854 Levine JM and D’Antonio CM (1999) Elton revisited: a review of evidence linking diversity and invasibility. Oikos 87: 15–26 Lockwood JL and McKinney ML (eds) (2001) Biotic Homogenization. Kluwer Academic Publishers, New York Lodge DM (1993) Biological invasions: lessons for ecology. Trends in Ecology & Evolution 8: 133–137 Lonsdale WM (1999) Global patterns of plant invasions and the concept of invasibility. Ecology 80: 1522–1536 MacArthur RH (1972) Geographical Ecology. Haper & Row, New York Mack RN (1989) Temperate grasslands vulnerable to plant invasions: characteristics and consequences. In: Drake JA, Mooney HA, di Castri F, Groves RH, Kruger FJ, Rejm´anek M and Williamson M (eds) Biological Invasions: A Global Perspective, pp 155–179. John Wiley & Sons, Chichester, UK Mack RN and Thompson JN (1982) Evolution in steppe with few large, hooved mammals. American Naturalist 119: 757–773 Magurran AE (1988) Ecological Diversity and Its Measurement. Princeton University Press, Princeton, New Jersey McCune, B and Mefford, MJ (1999) Multivariate Analysis of Ecological Data. Version 4.0. MJM Software, Gleneden Beach, Oregon McNaughton SJ (1983) Serengeti grassland ecology: the role of composite environmental factors and contingency in community organization. Ecological Monographs 53: 291–320 Milchunas DG and Lauenroth WK (1993) Quantitative effects of grazing on vegetation and soils over a global range of environments. Ecological Monographs 63: 327–366 Minchin PR (1987) An evaluation of the relative robustness of techniques for ecological ordination. Vegetatio 69: 89–107 Mooney HA and Hobbs RJ (eds) (2000) Invasive Species in a Changing World. Island Press, Washington, DC Naeem S, Knops JMH, Tilman D, Howe KM, Kennedy T and Gale S (2000) Plant diversity increases resistance to invasion in the absence of covarying extrinsic factors. Oikos 91: 97–108 Noy-Meir I, Gutman M and Kaplan Y (1989) Responses of Mediterranean grassland plants to grazing and protection. Journal of Ecology 77: 290–310 Oesterheld M and Sala OE (1990) Effects of grazing on seedling establishment: the role of seed and safe-site availability. Journal of Vegetation Science 1: 353–358
Olff H and Ritchie ME (1998) Effects of herbivores on grassland plant diversity. Trends in Ecology and Evolution 13: 261–265 Palmer MW and White PS (1994) Scale dependence and the species– area relationship. American Naturalist 144: 717–740 Perelman SB, Le´on RJC and Oesterheld M (2001) Cross-scale vegetation patterns of Flooding Pampa grasslands. Journal of Ecology 89: 562–577 Rapoport EH (1996) The flora of Buenos Aires: low richness or mass extinction? International Journal of Ecology and Environmental Sciences 22: 217–242 Richardson DM, Pysek P, Rejm´anek M, Barbour MG, Panetta FD and West CJ (2000) Naturalization and invasion: concepts and definitions. Diversity and Distributions 6: 93–107 Rubio G, Casasola G and Lavado RS (1995) Adaptations and biomass production of two grasses in response to waterlogging and soil nutrient enrichment. Oecologia 102: 102–105 Rusch GM and Oesterheld M (1997) Relationship between productivity, and species and functional group diversity in grazed and non-grazed pampas grassland. Oikos 78: 519–526 Safford HD and Harrison SP (2001) Grazing and substrate interact to affect native vs exotic diversity in roadside grasslands. Ecological Applications 11: 1112–1122 Sala OE, Deregibus VA, Schlichter T and Alippe H (1981) Productivity dynamics of a native temperate grassland in Argentina. Journal of Range Management 34: 48–51 Sala OE, Oesterheld M, Le´on RJC and Soriano A (1986) Grazing effects upon plant community structure in subhumid grasslands of Argentina. Vegetatio 67: 27–32 Shea K and Chesson P (2002) Community ecology theory as a framework for biological invasions. Trends in Ecology and Evolution 17: 170–176 Silvertown J, Dodd ME, McConway K, Potts J and Crawley MJ (1994) Rainfall, biomass variation, and community composition in the Park Grass Experiment. Ecology 75: 2430–2437 Smith MD and Knapp AK (1999) Exotic plant species in a C4dominated grassland: invasibility, disturbance, and community structure. Oecologia 120: 605–612 S¨oyrinki N (1991) On the alien flora of the province of Buenos Aires, Argentina. Annales Botanici Fennici 28: 59–79 Soriano A (1992) R´ıo de la Plata grasslands. In: Coupland RT (ed) Natural Grasslands: Introduction and Western Hemisphere. Ecosystems of the World – 8A, pp 367–407. Elsevier, Amsterdam Stohlgren TJ, Binkley D, Chong GW, Kalkhan MA, Schell LD, Bull KA, Otsuki Y, Newman G, Bashkin M and Son Y (1999a) Exotic plant species invade hot spots of native plant diversity. Ecological Monographs 69: 25–46 Stohlgren TJ, Schell LD and vanden Heuvel B (1999b) How grazing and soil quality affect native and exotic plant diversity in Rocky Mountain grasslands. Ecological Applications 9: 45–64 Tilman D (1997) Community invasibility, recruitment limitation, and grassland biodiversity. Ecology 78: 81–92 Tilman D and Pacala S (1993) The maintenance of species richness in plant communities. In: Ricklefs RR and Schluter D (eds) Species Diversity in Ecological Communities, pp 13–25. University of Chicago Press, Chicago
24 Tilman D and Lehman C (2001) Human-caused environmental change: impacts on plant diversity and evolution. Proceedings of the National Academy of Sciences (USA) 98: 5433–5440 Vitousek PM, Mooney HA, Lubchenco J and Melillo JM (1997) Human domination of Earth’s ecosystems. Science 277: 494–499
Wardle DA (2001) Experimental demonstration that plant diversity reduces invasibility – evidence of a biological mechanism or a consequence of sampling effect? Oikos 95: 161–170 Zavaleta ES, Hobbs RJ and Mooney HA (2001) Viewing invasive species removal in a whole-ecosystem context. Trends in Ecology and Evolution 16: 454–459
