Effects of herbivory on competition intensity in two arctic-alpine tundra communities with different productivity

The effects of long-term (11 yr) exclusion of vertebrate herbivores on competition intensity and plant community structure were studied using manipulative field experiments in two arctic-alpine plant communities with contrasting productivity: an unproductive snowbed and a considerably more productive tall herb meadow. In the snowbed, the exclusion of herbivorous mammals resulted in a significant increase in the biomasses of vascular plants and cryptogams, whereas no corresponding response was observed on the tall herb meadow. The intensity of competition, measured with a neighbour removal experiment, did not differ significantly between three of the four habitat×treatment combinations – snowbed exclosures, meadow exclosures and open meadow plots – but was significantly lower on open snowbed plots. Our results thus suggest that the low competition intensity in the unproductive snowbed is caused by herbivorous mammals, which tend to depress plant biomass in relatively unproductive habitats. When herbivorous mammals have been excluded for a sufficiently long time to allow the build-up of plant biomass even in unproductive habitats, between-habitat differences in competition intensity disappear.     

Local and regional processes in low-productive mountain plant communities: the roles of seed and microsite limitation in relation to grazing

It is becoming widely accepted that plant community structure is determined not only by local scale factors, but that regional factors may play considerable role. The research studying the associated processes in different environments with different species assemblages is still limited. We conducted a two-year seed sowing experiment to test whether a plant community in a low-productive mountain snowbed is limited by seed or microsite availability and how these variables depend on natural grazing. In a factorial design, half of the plots received a mixture of seeds of fourteen species naturally occurring at the study site and above ground biomass was removed from half of the plots. These treatments were applied to plots with long term grazer exclosures and to plots accessible to grazers. Both sowing and biomass removal increased the number of seedlings, the species richness of seedlings and total species richness. The number of seedlings was higher in open plots than in exclosures in the second year. Both seedling richness and total species richness were higher in open plots. Seedling recruitment was negatively related to the amount of above ground biomass and positively to the initial species richness. These results suggest that even fairly low-productive environments can be both seed and microsite limited and that these depend on grazing pressure. Natural grazing by mammal herbivores (e.g. lemmings and reindeer) favours species colonization and seedling emergence. Low-productive mountain snowbeds are prone to colonization from the local species pool and even high species richness may not constrain ingression of new species.     

The vegetation of the Øvre Heimdalen valley

The catchment area of Øvre Heimdalsvatn reaches from the upper half of the subalpine birch forest zone to the lowermost part of the high-alpine subzone within the alpine zone. The units dominating the area include mountain birch forest, tall herb meadow, mire vegetation, bilberry heath, snowbed communities, chionophobous heath vegetation, and areas without a continuous vegetation layer. It is shown that the potential natural vegetation corresponds closely to the actual vegetation distribution.     

From protégé to nurse plant: establishment of thorny shrubs in grazed temperate woodlands

Question: Thorny shrubs play keystone roles in grazed ecosystems by defending non‐protected plants against herbivores, but their establishment in grazed ecosystems is poorly understood. Which factors control establishment of recruits of thorny nurse shrubs in grazed temperate woodlands? Location: Ancient grazed temperate woodlands (52°32′N, 6°36′E), The Netherlands. Methods: We surveyed biotic and abiotic factors for saplings of thorny nurse shrubs in plots with and without saplings. To disentangle these factors, we performed a transplantation experiment over two growing seasons with nurse shrub saplings (Prunus spinosa and Crataegus monogyna) planted in two dominant vegetation types – tall unpalatable swards and short grazed lawns – half of them protected from herbivory via exclosures. Results: Plots with shrub saplings had taller surrounding vegetation, higher soil pH and higher soil moisture than plots without saplings. These plots predominantly contained unpalatable sward species, while plots without saplings mainly contained palatable lawn species. After transplantation, sapling survival was higher in exclosures than in the open, and higher in sward exclosures than in lawn exclosures. Sapling growth was higher in swards than in lawns, higher inside than outside exclosures, and higher for Prunus than Crataegus, while browsing on saplings was higher in lawns. Conclusion: Unpalatable swards form essential establishment niches for thorny shrubs in grazed temperate woodlands: they protect against herbivores before thorns fully develop in saplings, and sapling growth is better due to improved micro‐environmental conditions. Once established and thorny, shrub saplings grow out of the protective range of the swards and in turn facilitate tree seedlings, which are essential for long‐term persistence of grazed temperate woodlands. This study shows that nurse plants may start as protégés before becoming facilitators for other plants in a later life stage. This may be common for nurse plants in various ecosystems. We argue that improved understanding of establishment of nurse plants and their constraining factors is crucial for effective conservation and restoration in various ecosystems.     

Influences of chronic and current season grazing by collared pikas on above-ground biomass and species richness in subarctic alpine meadows

We studied an alpine herbivory gradient established by collared pikas, a small central place foraging lagomorph, to examine the effects of multiple grazing levels on above-ground live biomass (AGB) and species richness (SR) in alpine tundra. The effects of within-season (four sampling periods), multi-season (across three summers) and longer-term dynamics (inferred from spatial location of vegetation with respect to pika haypiles) were examined. Along the grazing gradient, we found support for and against hypotheses that propose biphasic, increasing, or decreasing responses to herbivory, both in terms of AGB and SR. Our results suggest that plant–herbivore predictability is still weak. To further examine the impact of herbivory, we experimentally removed pikas using mesh exclosures placed at increasing distance from the edge of talus occupied by pikas. AGB after the second consecutive year of herbivore exclusion increased by 125% compared to control plots in highly grazed areas adjacent to talus (<1 m). In more lightly grazed sites at distances 1–6 m from talus, AGB increased by more than 40% after pikas were removed. No differences were observed in the ungrazed sites >6 m from talus. AGB was highest in meadow patches previously grazed by pikas compared to those with little grazing history, but this response was only observed after two seasons following release from herbivory. Grazed sites at distances of 1–6 m had the highest SR. These results indicate that multi-year measurements of growth are particularly relevant in ecosystems dominated by long-lived perennials in regions where productivity is low. Infrequent herbivore vacancies may provide local short-term release from pika grazing, thereby contributing to the persistence of productive, highly palatable vegetation.     

Herbivore-induced expansion of generalist species as a driver of homogenization in post-disturbance plant communities

We used a controlled experiment to investigate how disturbance scale (canopy gap area) and herbivory influence post-disturbance plant community dynamics. Twenty canopy gaps were installed in a temperate hemlock-hardwood forest during the winter of 2002–2003: seven small gaps (50–150 m²), seven medium gaps (151–250 m²), and six large gaps (251–450 m²). Within each gap, we established 4–12 sample plots (depending on gap size); 1–3 of which were enclosed with wire mesh white-tailed deer (Odocoileus virginianus) exclosures in 2005. Gaps were revisited and intensively sampled in 2007. After five growing seasons, ground-layer plant communities in non-exclosed plots were more similar compositionally than exclosure plots. Non-exclosed plots in small and medium gaps were more similar to non-exclosed plots in large gaps than they were to exclosed plots in their respective gap size class. Shade-tolerant forbs and trees were less common outside exclosures, while generalist species associated with higher understory light levels and exotics were more prevalent outside exclosures. Our results suggest that even in forests with relatively low deer densities (6.5–9.3 deer km⁻²), white-tailed deer herbivory may influence the developmental trajectory of post-disturbance plant communities and be a mechanism for decreasing β-diversity along environmental gradients.     

Resource availability and plant diversity explain patterns of invasion of an exotic grass

Aims In this study, we examine two common invasion biology hypotheses—biotic resistance and fluctuating resource availability—to explain the patterns of invasion of an invasive grass, Microstegium vimineum.Methods We used 13-year-old deer exclosures in Great Smoky Mountains National Park, USA, to examine how chronic disturbance by deer browsing affects available resources, plant diversity, and invasion in an understory plant community. Using two replicate 1 m 2 plots in each deer browsed and unbrowsed area, we recorded each plant species present, the abundance per species, and the fractional percent cover of vegetation by the cover classes: herbaceous, woody, and graminoid. For each sample plot, we also estimated overstory canopy cover, soil moisture, total soil carbon and nitrogen, and soil pH as a measure of abiotic differences between plots.Important findings We found that plant community composition between chronically browsed and unbrowsed plots differed markedly. Plant diversity was 40% lower in browsed than in unbrowsed plots. At our sites, diversity explained 48% and woody plant cover 35% of the variation in M. vimineum abundance. In addition, we found 3.3 times less M. vimineum in the unbrowsed plots due to higher woody plant cover and plant diversity than in the browsed plots. A parsimonious explanation of these results indicate that disturbances such as herbivory may elicit multiple conditions, namely releasing available resources such as open space, light, and decreasing plant diversity, which may facilitate the proliferation of an invasive species. Finally, by testing two different hypotheses, this study addresses more recent calls to incorporate multiple hypotheses into research attempting to explain plant invasion.     

Functional identity versus species richness: herbivory resistance in plant communities

The resistance of a plant community against herbivore attack may depend on plant species richness, with monocultures often much more severely affected than mixtures of plant species. Here, we used a plant–herbivore system to study the effects of selective herbivory on consumption resistance and recovery after herbivory in 81 experimental grassland plots. Communities were established from seed in 2002 and contained 1, 2, 4, 8, 16 or 60 plant species of 1, 2, 3 or 4 functional groups. In 2004, pairs of enclosure cages (1 m tall, 0.5 m diameter) were set up on all 81 plots. One randomly selected cage of each pair was stocked with 10 male and 10 female nymphs of the meadow grasshopper, Chorthippus parallelus. The grasshoppers fed for 2 months, and the vegetation was monitored over 1 year. Consumption resistance and recovery of vegetation were calculated as proportional changes in vegetation biomass. Overall, grasshopper herbivory averaged 6.8%. Herbivory resistance and recovery were influenced by plant functional group identity, but independent of plant species richness and number of functional groups. However, herbivory induced shifts in vegetation composition that depended on plant species richness. Grasshopper herbivory led to increases in herb cover at the expense of grasses. Herb cover increased more strongly in species-rich mixtures. We conclude that selective herbivory changes the functional composition of plant communities and that compositional changes due to selective herbivory depend on plant species richness.     

The species richness–productivity relationship in the herb layer of European deciduous forests

Aim In contrast to non‐forest vegetation, the species richness–productivity (SR‐P) relationship in forests still remains insufficiently explored. Several studies have focused on the diversity of the tree layer, but the species richness of temperate deciduous forests is mainly determined by their species‐rich herb layer. The factors controlling herb‐layer productivity may differ from those affecting tree layers or open herbaceous vegetation, and thus the SR‐P relationship and its underlying processes may differ. However, the few relevant studies have reported controversial results. Here we explore the SR‐P relationship in the forest herb layer across different areas from oceanic to continental Europe, and put the effect of habitat productivity on species richness into context with other key factors, namely soil pH and light availability. Location North‐western Germany, Czech Republic, Slovakia and southern Urals (Russia). Methods We measured herb‐layer species richness and biomass, soil pH and tree‐layer cover in 156 vegetation plots of 100 m² in deciduous forests. We analysed the SR‐P relationship and the relative importance of environmental variables using regression models for particular areas and separate forest types. Results We found a consistent monotonic increase in the herb‐layer species richness with productivity across all study areas and all forest types. Soil pH and light availability also affected species richness, but their relative importance differed among areas. Main conclusions We suggest that the monotonically increasing SR‐P relationship in the forest herb layer results from the fact that herb‐layer productivity is limited by canopy shading; competition within the herb layer is therefore not strong enough to exclude many species. This differs fundamentally from open herbaceous vegetation, which is not subject to such productivity limits and consequently exhibits a unimodal SR‐P relationship. We present a conceptual model that might explain the differences in the SR‐P relationship between the forest herb layer and open herbaceous vegetation.     

Sawdust Addition Reduces the Productivity of Nitrogen‐Enriched Mountain Grasslands

Anthropogenic nutrient enrichment of mountain grasslands has boosted grasses and fast‐growing unpalatable plants at the expense of slow‐growing species, resulting in a significant loss in biodiversity. A potential tool to reduce nutrient availability and aboveground productivity without destroying the perennial vegetation is carbon (C) addition. However, little is known about its suitability under severe climatic conditions. Here, we report the results of a 3‐year field study assessing the effects of sawdust addition on soil nutrients, aboveground productivity, and vegetational composition of 10 grazed and ungrazed mountain grasslands. Of particular interest was the effect of C addition on grasses and on the tall unpalatable weed Veratrum album. After 3 years, soil pH, ammonium, and plant‐available phosphorus were not altered by sawdust application, and nitrate concentrations were marginally higher in treatment plots. However, the biomass of grasses and forbs (without V. album) was 20–25% lower in sawdust‐amended plots, whereas the biomass of V. album was marginally higher. Sawdust addition reduced the cover of grasses but did not affect evenness, vegetation diversity, or plant species richness, although species richness generally increased with decreasing biomass at our sites. Our results suggest that sawdust addition is a potent tool to reduce within a relatively short time the aboveground productivity and grass cover in both grazed and ungrazed mountain grasslands as long as they are not dominated by tall unpalatable weeds. The technique has the advantage that it preserves the topsoil and the perennial soil seed bank.     

Long-term effects of grazing on subalpine and alpine grasslands in the Central Alps,Austria

The effects of grazing exclusion on species diversity and functional diversity were analyzed along an elevation gradient from the subalpine (1960 m a.s.l.) to the lower and upper alpine zone (2275 m–2650 m a.s.l.) in the Austrian Central Alps for 15 years. Nine sites were chosen, including grasslands at different elevations, a bog and a glacier foreland site at the lower alpine zone and a snowbed at the upper alpine zone. Data were acquired by frequency counts in 1 m² permanent plots inside each fenced area (three plots per site) and outside in grazed areas (three plots). Diversity indices and functional diversity were analyzed by means of generalized linear mixed models (GLMMs). Exclosure, duration of exclusion and exclosure*years (interaction effect) were defined as predictor variables. Multivariate ordination techniques were used to (i) determine species responses to grazing exclusion (pRDA, partial redundancy analysis) and (ii) to create a distance matrix representing the changes between exclosures and control plots per year (NMDS, non-metric multidimensional scaling). At the subalpine grassland, first differences between exclosure and control plots occurred already only after three years, at the upper alpine zone after four and five years. Contrary to our expectation, dwarf shrubs did not increase within the exclosures of the subalpine grassland. Instead, mainly the tall forb Geranium sylvaticum increased. Species richness significantly decreased at the exclosures of the subalpine zone, the snowbed and at one upper alpine grassland sites. The communities of the glacier foreland and the bog were hardly affected by grazing exclusion.We conclude that plant species and communities react individually depending on elevation and grazing animals. Grazing exclusion studies at high elevations should definitely be carried out in the long-term.     

Mesoherbivores affect grasshopper communities in a megaherbivore-dominated South African savannah

African savannahs are among the few places on earth where diverse communities of mega- and meso-sized ungulate grazers dominate ecosystem functioning. Less conspicuous, but even more diverse, are the communities of herbivorous insects such as grasshoppers, which share the same food. Various studies investigated the community assembly of these groups separately, but it is poorly known how ungulate communities shape grasshopper communities. Here, we investigated how ungulate species of different body size alter grasshopper communities in a South African savannah. White rhino is the most abundant vertebrate herbivore in our study site. Other common mesoherbivores include buffalo, zebra and impala. We hypothesized that white rhinos would have greater impact than mesoherbivores on grasshopper communities. Using 10-year-old exclosures, at eight sites we compared the effects of ungulates on grasshopper communities in three nested treatments: (i) unfenced plots (‘control plots’) with all vertebrate herbivores present, (ii) plots with a low cable fence, excluding white rhino (‘megaherbivore exclosures’), and (iii) plots with tall fences, excluding all herbivores larger than rodents (‘complete ungulate exclosures’). In each plot, we collected data of vegetation structure, grass and grasshopper community composition. Complete ungulate exclosures contained 30 % taller vegetation than megaherbivore exclosures and they were dominated by different grass and grasshopper species. Grasshoppers in complete ungulate exclosures were on average 3.5 mm longer than grasshoppers in megaherbivore exclosures, possibly due to changes in plant communities or vegetation structure. We conclude that surprisingly, in this megaherbivore hotspot, mesoherbivores, instead of megaherbivores, most strongly affect grasshopper communities.     

Effects of grazing exclusion on species composition in high-altitude grasslands of the Central Alps

Our study examines the effects of grazing exclusion on low-productive subalpine and alpine grasslands of the Central Alps (UNESCO Biosphere Park Gurgler Kamm, Obergurgl, Austria). A long-term exclusion experiment was established in 2000 in the subalpine, the lower, and the upper alpine zone. With exception of the subalpine zone, domestic herbivores have been grazing during the whole growing season. In grazed and exclosure plots species frequencies were recorded for 7 years. We analysed exclosure effects on species number, community composition, life forms, and functional groups.Species richness did not decrease significantly within the exclosures, but changes in species composition occurred in each zone, although some were transitory in nature. The dynamic trends of the plots were significantly explained by the ‘treatment×year’ effect along the whole altitudinal gradient, but the effects decreased considerably with altitude. In the subalpine and upper alpine exclosures, stress-tolerators, species of low or no nutritive value, and mosses showed a decreasing trend, whereas tall grasses (subalpine exclosures), competitors, and species with high or medium nutritive values (lower alpine exclosures) tended to increase.Overall, our 7-year study revealed that several functional groups reacted to grazing, according to our main expectations. We suggest that these effects will intensify in the long term.     

Effects of local extinction of the plains vizcacha (Lagostomus maximus) on vegetation patterns in semi-arid scrub

We studied spatial and temporal effects of local extinction of the plains vizcacha (Lagostomus maximus) on plant communities following widespread, natural extinctions of vizcachas in semi-arid scrub of Argentina. Spatial patterns in vegetation were examined along transects extending outward from active and extinct vizcacha burrow systems. Responses of vegetation to removal of vizcachas were assessed experimentally with exclosures and by documenting vegetation dynamics for 6 years following extinctions. Transect data demonstrated clear spatial patterns in plant cover, particularly an increase in perennial grasses, outward from active vizcacha burrows. These patterns were consistent with predictions based on foraging theory and studies that document grasses as the preferred food of vizcachas. Removal of vizcachas, experimentally and with extinctions, resulted in an immediate increase in perennial and annual forbs indicating that intense herbivory can depress forb cover, as well as grasses. After a 1-year lag following cessation of herbivory, cover of grasses increased. Forbs declined as grasses increased. The long-term effect of extinction of vizcachas was a conversion of colony sites from open patches dominated by forbs to dense bunch grass characteristic of the matrix. Major changes in vegetation occurred within 2–3 years after extinction, resulting in a large pulse of landscape change. However, some species of grasses were uncommon until 5–6 years after the vizcacha extinction. With extinction and colonization, vizcachas generate a dynamic mosaic of patches on the landscape and create temporal, as well as spatial, heterogeneity in semi-arid scrub.     

Restoring riparian meadows currently dominated by Artemisa using alternative state concepts ‐ above‐ground vegetation response

Abstract. Livestock overgrazing and stream incision in the western USA often result in encroachment and dominance of Artemisia tridentata ssp. tridentata (Big sagebrush) in riparian areas that formerly supported meadows. To define the alternative states and thresholds for these ecosystems, we conducted a restoration experiment that included sites with high, intermediate or low water tables. We used a paired‐plot approach in which one plot on each site was burned and seeded with native grasses and forbs typical of naturally occurring dry meadow and Artemisia/Leymus cinereus ecological types, while adjacent unburned plots served as controls. Sites with high and intermediate water tables had greater initial abundances of perennial grasses typical of dry meadows, such as Leymus triticoides and Poa secunda ssp. juncifolia, and these species increased after the burn. In contrast, sites with low water tables were dominated by annual forbs such as Chenopo‐dium album and Descurainia pinnata after the burn. Biomass increased progressively from 1997 to 1999 on burned plots, while controls showed little change. Burning effects were microsite specific, with former Artemisia microsites exhibiting lower biomass than interspaces initially, but similar or higher biomass by the third year. Establishment of seeded species was low and species composition was determined largely by pre‐burn vegetation. Artemisia dominated sites with high water tables appear to represent an alternative state of the dry meadow ecological type, while sites with low water table sites have crossed an abiotic threshold governed by water tables and represent a new ecological type. Burning is an effective tool for restoring relatively high water table sites, but low water table sites will require burning and seeding with species adapted to more xeric conditions.     

Effects of reindeer on boreal forest floor vegetation: Does grazing cause vegetation state transitions?

Intensive reindeer grazing has been hypothesized to drive vegetation shifts in the arctic tundra from a low-productive lichen dominated state to a more productive moss dominated state. Although the more productive state can potentially host more herbivores, it may still be less suitable as winter grazing grounds for reindeer, if lichens, the most preferred winter forage, are less abundant. Therefore, such a shift towards mosses may have severe consequences for reindeer husbandry if ground-growing lichens have difficulties to recover. We tested if reindeer cause this type of vegetation state shifts in boreal forest floor vegetation, by comparing plant species composition and major soil processes inside and outside of more than 40-year-old exclosures. Lichen biomass was more than twice as high inside exclosures than in grazed controls and almost 5 times higher than in heavily grazed patches. Contrary to our predictions, net N mineralization and plant production were higher in the exclosures than in the grazed controls. The lack of response of phytometer plants in a common garden bioassay indicated that changed soil moisture may drive effects of reindeer on plant productivity in these dry Pine forest ecosystems.     

Plant species richness–productivity relationships in a low-productive boreal region

Local species richness–productivity (SR–P) relationship is usually reported as unimodal if long productivity gradients are sampled. However, it tends to be monotonically increasing in low-productive environments due to the decreasing part of the SR–P curve being truncated. Previous work indicated that this can hold true for forest herb layers, because of an upper bound on productivity caused mainly by canopy shading. Here, we ask whether the same pattern exists in a region with an upper bound on productivity caused by a harsh climate. We sampled herbaceous vegetation of boreal forests and grasslands in a low-productive region of central Yakutia (NE Siberia) with dry and winter-cool continental climate. We collected data on species composition, herb-layer productivity (aboveground herbaceous biomass), soil chemistry and light availability. We applied regression models to discriminate between monotonically increasing, decreasing and unimodal responses of herb-layer species richness to measured variables and analysed trends in the species-pool size and beta diversity along the productivity gradient. Our expectation of the monotonically increasing SR–P relationship was confirmed for neither forest herb layers nor grasslands. In the forest herb layers, no relationship was detected. In grasslands, the relationship was unimodal with species richness decline starting at much lower productivity levels than in more productive temperate grasslands. Potential causes for this decline are either limitation of local species richness by the species pool, which contains few species adapted to more productive habitats, or competitive exclusion, which can become an important control of species richness under lower levels of productivity than is the case in temperate grasslands.     

Influence of livestock grazing on meadow pipit foraging behaviour in upland grassland

Changes in grazing management are believed to be responsible for declines in populations of birds breeding in grassland over the last decades. The relationships between grazing management regimes, vegetation structure and composition and the availability of invertebrate food resources to passerine birds remain poorly understood. In this study, we investigated the foraging site selection of meadow pipits (Anthus pratensis L.) breeding in high intensity sheep-grazed plots or low intensity mixed (i.e. sheep and cattle)-grazed plots. We sampled above-ground invertebrates, measured vegetation height and density and conducted a vegetation survey in areas where meadow pipits were observed to forage and areas that were randomly selected. Birds foraged in areas with a lower vegetation height and density and in areas containing a lower proportion of the dominant, tussock-forming grass species Molinia caerulea. They did not forage in areas with a total higher invertebrate biomass but at areas with preferred vegetation characteristics invertebrate biomass tended to be higher in foraging sites than random sites. The foraging distance of meadow pipits was higher in the intensively grazed plots. Our findings support the hypothesis that resource-independent factors such as food accessibility and forager mobility may determine patch selection and are of more importance as selection criteria than food abundance per se. Food accessibility seems to become an even more important selection criterion under high grazing intensity, where prey abundance and size decrease. In our upland grazing system, a low intensity, mixed grazing regime seems to provide a more suitable combination of sward height, plant diversity, structural heterogeneity and food supply for meadow pipit foraging activity compared to a more intensive grazing regime dominated by sheep.     

The interactive effects of grazing ungulates and aboveground production on grassland diversity

The variable and nonlinear relationships between plant species richness (SR) and aboveground production (NAP) among terrestrial ecosystems indicate that the energetic capacity of ecosystems interacts with other environmental factors to control diversity. One contributing factor determining plant diversity is herbivory; but few studies have effectively examined the interaction of herbivores and NAP on SR. The objective of this study was to investigate how NAP and herds of native migrating ungulates determine plant SR in grasslands of Yellowstone National Park. Plant SR at peak aboveground biomass was compared inside and outside ungulate exclosures at two spatial scales, 1.0 m² (“local”) and 100 m² (“community”), in ten variable grasslands. NAP also was determined inside and outside exclosures. The relationship between SR and NAP was unimodal for grazed and ungrazed grassland at both spatial scales. Grazers increased local SR, independent of NAP. In contrast, herbivore effects on community SR ranged from no effect among low-productive grassland to an increasingly positive influence as NAP increased. In addition, ungulates reduced beta diversity (the contribution to community SR attributed to variability among local patches) at dry, low-productive and wet, high-productive sites. These results suggest that the size of the pool of species available to colonize grassland is an important factor controlling the response of grassland SR to herbivory, particularly from low- to intermediate-productive grassland.     

An experimental study of the impacts of understorey forest vegetation and herbivory by red deer and rodents on seedling establishment and species composition in Waitutu Forest, New Zealand

Introduced mammalian herbivores are changing the structure and composition of New Zealand’s forest ecosystems and may modify forest succession after natural disturbances. We studied how introduced ungulates (red deer and feral pigs) and rodents (rats and house mice) affected the rate of recovery (i.e. the engineering resilience) of the forest understorey following artificial disturbance. We imposed disturbances by clearing understorey vegetation dominated by Blechnum ferns in forests on relatively fertile alluvium and elevated infertile marine terraces, and recorded recovery of vegetation (seedling establishment, species composition, cover and volume) in herbivore exclosures and controls. Seedlings quickly established on cleared plots: after 2 years, numbers of woody seedlings and ground cover of vascular plants relative to initial values were similar on cleared and uncleared treatments. Volume of plant biomass <2 m remained low on cleared subplots. Ungulates significantly reduced the re-establishment of woody seedlings ≥ 10 cm tall: only one seedling reached this height outside exclosures, compared with 29 seedlings inside. The number of seedlings <10 cm tall, expressed relative to numbers present pre-clearing, was not significantly affected by ungulates. The species composition of regenerating vegetation was more similar (Jaccard index) to pre-clearing understorey vegetation inside ungulate exclosures than outside. No consistent effect of rodents (primarily house mice) on seedling establishment or species composition was detected after 2 years, and rodent exclosures did not significantly affect survival of seedlings (Griselinia littoralis and Aristotelia serrata) planted as an index of rodent herbivory pressure. No significant differences in vegetation recovery were apparent between forest types. Rapid seedling recruitment in the absence of understorey vegetation and the presence of herbivores provided evidence that understorey vegetation competes with seedlings for light. Ungulate effects were consistent with other experiments that showed herbivores reduced the rate and altered the trajectory of vegetation regrowth after disturbance.