Legacy effects of top–down disturbances on woody plant species composition in semi‐arid systems

Savanna vegetation is controlled by bottom‐up (e.g. soil and rainfall) and top–down (e.g. fire and herbivory) factors, all of which have an effect on biodiversity. Little is known about the relative contribution of these factors to biodiversity, particularly the long‐term effects of top–down disturbance on patterns of woody plant composition. The aim of this study was to identify if various degrees of disturbance regimes create distinct woody species community assemblages. Data were collected over 1820 plots across Kruger National Park, South Africa. Woody species were identified and categorized into one of three height classes: shrub (0.75–2.5 m), brush (2.5–5.5 m), and tree (>5.5 m). Species richness and composition were calculated for each site and height class. A combination of long‐term fire and elephant density data were used to delineate areas with varying degrees of top–down disturbance (i.e. low, medium and high). Using these degrees of disturbance, species composition was identified and community assemblages constructed according to each disturbance regime. Our results suggest that areas with similar disturbance regimes have similar species composition. Shrub composition was mainly responsive to the number of fires between the years 1941–1990, while tree composition was more responsive to elephant disturbance. A few dominant species were found equally under all degrees of disturbance at all height classes, while others were more regularly found under specific disturbance regimes at particular height classes. This study highlights that while species richness does not appear to be influenced by long‐term, top–down disturbance regimes, species community composition may be responsive to these disturbances. Most species and structural classes persisted across all disturbance regimes, but the long‐term effects of top–down disturbances can influence compositional and structural biodiversity. This information provides context for management policies related to artificial water provision, elephants and fire.     

Woody cover in African savannas: the role of resources, fire and herbivory

Aim To determine the functional relationships between, and the relative importance of, different driver variables (mean annual precipitation, soil properties, fire and herbivory) in regulating woody plant cover across broad environmental gradients in African savannas. Location Savanna grasslands of East, West and Southern Africa. Methods The dependence of woody cover on mean annual precipitation (MAP), soil properties (texture, nitrogen mineralization potential and total phosphorus), fire regimes, and herbivory (grazer, browser + mixed feeder, and elephant biomass) was determined for 161 savanna sites across Africa using stochastic gradient boosting, a refinement of the regression tree analysis technique. Results All variables were significant predictors of woody cover, collectively explaining 71% of the variance in our data set. However, their relative importance as regulators of woody cover varied. MAP was the most important predictor, followed by fire return periods, soil characteristics and herbivory regimes. Woody cover showed a strong positive dependence on MAP between 200 and 700 mm, but no dependence on MAP above this threshold when the effects of other predictors were accounted for. Fires served to reduce woody cover below rainfall‐determined levels. Woody cover showed a complex, non‐linear relationship with total soil phosphorus, and was negatively correlated with clay content. There was a strong negative dependence of woody cover on soil nitrogen (N) availability, suggesting that increased N‐deposition may cause shifts in savannas towards more grassy states. Elephants, mixed feeders and browsers had negative effects on woody cover. Grazers, on the other hand, depressed woody cover at low biomass, but favoured woody vegetation when their biomass exceeded a certain threshold. Main conclusions Our results indicate complex and contrasting relationships between woody cover, rainfall, soil properties and disturbance regimes in savannas, and suggest that future environmental changes such as altered precipitation regimes, N‐enrichment and elevated levels of CO₂ are likely to have opposing, and potentially interacting, influences on the tree–grass balance in savannas.     

Prefire grazing by cattle increases postfire resistance to exotic annual grass (Bromus tectorum) invasion and dominance for decades

相似文献   

Elephant herbivory,frost damage and topkill in Kalahari sand woodland savanna trees

Questions: Which factors best predict the probability of elephant and frost damage in Kalahari sand woodland savanna? What is the association between tree mortality and the disturbance regime? Location: Western Zimbabwe. Methods: Elephant and frost damage, topkill, and whole‐plant mortality were quantified in ten common tree species in a Kalahari sand savanna in Zimbabwe. Individual trees were tagged in 20 plots and monitored over a two‐year period. A model selection approach was used to test the association between the probability of damage and size, prior damage, and neighbourhood effects, and to investigate the effect of damage on mortality. Results: Elephant damage differed strongly among species, and was not influenced by neighbourhood effects or prior disturbance. Frost damage also varied across species, and declined as a function of stem size and neighbourhood tree cover, and ‐ against expectations ‐ prior disturbance. Topkill increased as a function of elephant and frost damage, but was lower in previously damaged than in undamaged trees. Conclusions: Frost and elephant damage are influenced by community composition, and frost damage is also correlated with community structure and prior disturbance. Frost is an important and generally overlooked disturbance agent in southern African woodlands, where it may play a key role in association with other disturbance factors ‐ such as elephant herbivory ‐ that reduce woodland canopy cover.     

A continental-scale analysis of tree cover in African savannas

Aim We present a continental‐scale analysis that explores the processes controlling woody community structure in tropical savannas. We analyse how biotic and abiotic factors interact to promote and modify tree cover, examine alternative ecological hypotheses and quantify disturbance effects using satellite estimates of tree cover. Location African savannas. Methods Tree cover is represented as a resource‐driven potential cover related to rainfall and soil characteristics perturbed by natural and human factors such as fire, cattle grazing, human population and cultivation. Within this framework our approach combines semi‐empirical modelling and information theory to identify the best models. Results Woody community structure across African savannas is best represented by a sigmoidal response of tree cover to mean annual precipitation (MAP), with a dependency on soil texture, which is modified by the separate effects of fire, domestic livestock, human population density and cultivation intensity. This model explains c. 66% of the variance in tree cover and appears consistent across the savanna regions of Africa. Main conclusions The analysis provides a new understanding of the importance and interaction of environmental and disturbance factors that create the broad spatial patterns of tree cover observed in African savannas. Woody cover increases with rainfall, but is modified by disturbances. These ‘perturbation’ effects depend on MAP regimes: in arid savannas (MAP < 400 mm) they are generally small (< 1% decrease in cover), while in semi‐arid and mesic savannas (400–1600 mm), perturbations result in an average 2% (400 mm) to 23% (1600 mm) decrease in cover; fire frequency and human population have more influence than cattle, and cultivation appears, on average, to lead to small increases in woody cover. Wet savannas (1600–2200 mm) are controlled by perturbations that inhibit canopy closure and reduce tree cover by, on average, 24–34%. Full understanding of the processes determining savanna structure requires consideration of resource limitation and disturbance dynamics.     

Fine‐scale habitat heterogeneity influences browsing damage by elephant and giraffe

Effects of large mammalian herbivores on woody vegetation tend to be heterogeneous in space and time, but the factors that drive such heterogeneity are poorly understood. We examined the influence of fine‐scale habitat heterogeneity on the distribution and browsing effects of two of the largest African terrestrial mammals, the elephant and giraffe. We conducted this study within a 120‐ha (500 x 2,400 m) ForestGEO long‐term vegetation monitoring plot located at Mpala Research Center, Kenya. The plot traverses three distinct topographic habitats (“plateau,” “steep slopes,” and “valley”) with contrasting elevation, slope, soil properties, and vegetation composition. To quantify browsing damage, we focused on Acacia mellifera, a palatable tree species that occurs across the three habitat categories. Overall tree density, species richness, and diversity were highest on the steep slopes and lowest on the plateau. Acacia mellifera trees were tallest and had the lowest number of stems per tree on the steep slopes. Both elephant and giraffe avoided the steep slopes, and their activity was higher during the wet season than during the dry season. Browsing damage on Acacia mellifera was lowest on the steep slopes. Elephant browsing damage was highest in the valley, whereas giraffe browsing damage was highest on the plateau. Our findings suggest that fine‐scale habitat heterogeneity is an important factor in predicting the distribution of large herbivores and their effects on vegetation and may interact with other drivers such as edaphic variations to influence local variation in vegetation structure and composition.     

Fire severity mediates climate‐driven shifts in understorey community composition of black spruce stands of interior Alaska

Question: How do pre‐fire conditions (community composition and environmental characteristics) and climate‐driven disturbance characteristics (fire severity) affect post‐fire community composition in black spruce stands? Location: Northern boreal forest, interior Alaska. Methods: We compared plant community composition and environmental stand characteristics in 14 black spruce stands before and after multiple, naturally occurring wildfires. We used a combination of vegetation table sorting, univariate (ANOVA, paired t‐tests), and multivariate (detrended correspondence analysis) statistics to determine the impact of fire severity and site moisture on community composition, dominant species and growth forms. Results: Severe wildfires caused a 50% reduction in number of plant species in our study sites. The largest species loss, and therefore the greatest change in species composition, occurred in severely burned sites. This was due mostly to loss of non‐vascular species (mosses and lichens) and evergreen shrubs. New species recruited most abundantly to severely burned sites, contributing to high species turnover on these sites. As well as the strong effect of fire severity, pre‐fire and post‐fire mineral soil pH had an effect on post‐fire vegetation patterns, suggesting a legacy effect of site acidity. In contrast, pre‐fire site moisture, which was a strong determinant of pre‐fire community composition, showed no relationship with post‐fire community composition. Site moisture was altered by fire, due to changes in permafrost, and therefore post‐fire site moisture overrode pre‐fire site moisture as a strong correlate. Conclusions: In the rapidly warming climate of interior Alaska, changes in fire severity had more effect on post‐fire community composition than did environmental factors (moisture and pH) that govern landscape patterns of unburned vegetation. This suggests that climate change effects on future community composition of black spruce forests may be mediated more strongly by fire severity than by current landscape patterns. Hence, models that represent the effects of climate change on boreal forests could improve their accuracy by including dynamic responses to fire disturbance.     

Elephant (Loxodonta africana) Disturbance to Riparian Woodland: Effects on Tree-Species Richness,Diversity and Functional Redundancy

The substantial increase in elephant populations across many areas in southern Africa over past decades is prompting concerns about the effects on biodiversity. We investigated the outcomes of elephant disturbance on tree-species presence, density, and richness, and on alpha and beta diversity within riparian woodland in Chobe National Park, Botswana. We enumerated all tree species occurring in 32 plots (0.06 ha) along the Chobe riverfront. Plots were stratified by soil type (nutrient-rich alluvium vs. nutrient-poor Kalahari sand covering alluvium) and elephant impact (high vs. low impact on both soil types). We tested four predictions: elephants reduce tree density, richness, and alpha diversity; beta diversity is greater in vegetation subjected to high elephant impact; elephant impact on tree-species composition is greater on nutrient-poor than on nutrient-rich soil; and the loss or decline of abundant tree species on heavily disturbed sites is offset by an increase in abundance of functionally similar species, ones that are minor on lightly disturbed sites. Elephant browsing substantially affected tree-species composition, reducing density, species richness, evenness, and alpha diversity but had no effect on beta diversity. The dominant species on relatively undisturbed areas were partly replaced by functionally similar species on heavily disturbed sites. Soil type influenced species composition on lightly disturbed sites but was less important at higher elephant densities. Our findings are important for areas with extreme dry-season densities of elephants but should not be extrapolated to infer purported effects of elephants on tree diversity at lower densities.     

Land‐use legacies and present fire regimes interact to mediate herbivory by altering the neighboring plant community

Past and present human activities, such as historic agriculture and fire suppression, are widespread and can create depauperate plant communities. Although many studies show that herbivory on focal plants depends on the density of herbivores or the composition of the surrounding plant community, it is unclear whether anthropogenic changes to plant communities alter herbivory. We tested the hypothesis that human activities that alter the plant community lead to subsequent changes in herbivory. At 20 sites distributed across 80 300 hectares, we conducted a field experiment that manipulated insect herbivore access (full exclosures and pseudo‐exclosures) to four focal plant species in longleaf pine woodlands with different land‐use histories (post‐agricultural sites or non‐agricultural sites) and degrees of fire frequency (frequent and infrequent). Plant cover, particularly herbaceous cover, was lower in post‐agricultural and fire suppressed woodlands. Density of the dominant insect herbivore at our site (grasshoppers) was positively related to plant cover. Herbivore access reduced biomass of the palatable forb Solidago odora in frequently burned post‐agricultural sites and in infrequently burned non‐agricultural woodlands and increased mortality of another forb (Pityopsis graminifolia), but did not affect two other less palatable species (Schizachyrium scoparium and Tephrosia virginiana). Herbivory on S. odora exhibited a hump‐shaped response to plant cover, with low herbivory at low and high levels of plant cover. Herbivore density had a weak negative effect on herbivory. These findings suggest that changes in plant cover related to past and present human activities can modify damage rates on focal S. odora plants by altering grasshopper foraging behavior rather than by altering local grasshopper density. The resulting changes in herbivory may have the potential to limit natural recovery or restoration efforts by reducing the establishment or performance of palatable plant species.     

Strong influences of a dominant,ground‐nesting ant on recruitment,and establishment of ant colonies and communities

Many factors drive the organization of communities including environmental factors, dispersal abilities, and competition. In particular, ant communities have high levels of interspecific competition and dominance that may affect community assembly processes. We used a combination of surveys and nest supplementation experiments to examine effects of a dominant ground‐nesting ant (Pheidole synanthropica) on (1) arboreal twig‐nesting, (2) ground‐foraging, and (3) coffee‐foraging ant communities in coffee agroecosystems. We surveyed these communities in high‐ and low‐density areas of P. synanthropica over 2 years. To test for effects on twig ant recruitment, we placed artificial nesting resources on coffee plants in areas with and without P. synanthropica. The first sampling period revealed differences in ant species composition on the ground, in coffee plants, and artificial nests between high‐ and low‐density sites of P. synanthropica. High‐density sites also had significantly lower recruitment of twig ants and had species‐specific effects on twig ant species. Prior to the second survey period, abundance of P. synanthropica declined in the high‐density sites, such that P. synanthropica densities no longer differed. Subsequent sampling revealed no difference in total recruitment of twig ants to artificial nests between treatments. Likewise, surveys of ground and coffee ants no longer showed significant differences in community composition. The results from the first experimental period, followed by survey results after the decline in P. synanthropica densities suggest that dominant ants can drive community assembly via both recruitment and establishment of colonies within the community.     

Disturbance Effects on the Seed Bank of Mexican Cloud Forest Fragments1

The density and floristic composition of the soil seed bank was assessed in six cloud forest fragments with different levels of human disturbance in central Veracruz, Mexico. A total of 8416 seeds germinated in 60 soil samples, at 5‐cm depth, corresponding to 107 species, 85 genera, and 48 families. Significant differences were found among study sites in seed densities with values ranging from 873 to 3632/m². Tree species contributed 20 percent of the total soil seed bank in four sites and herbs accounted for the majority of the species in each site. Among tree species, Trema micrantha displayed the highest seed density, accounting for 84 percent of the germinated seeds. In general, the tree species composition of the soil seed bank did not closely reflect the composition of the tree community. Results suggest that disturbance produced by human activities (trail use, selective cutting of trees, livestock) may influence the size and composition of the soil seed bank in forest fragments. Sites where human activity has been reduced showed the highest proportion of dormant seeds.     

Forage nutritive quality in the Serengeti ecosystem: the roles of fire and herbivory

Fire and herbivory are important determinants of nutrient availability in savanna ecosystems. Fire and herbivory effects on the nutritive quality of savanna vegetation can occur directly, independent of changes in the plant community, or indirectly, via effects on the plant community. Indirect effects can be further subdivided into those occurring because of changes in plant species composition or plant abundance (i.e., quality versus quantity). We studied relationships between fire, herbivory, rainfall, soil fertility, and leaf nitrogen (N), phosphorus (P), and sodium (Na) at 30 sites inside and outside of Serengeti National Park. Using structural equation modeling, we asked whether fire and herbivory influences were largely direct or indirect and how their signs and strengths differed within the context of natural savanna processes. Herbivory was associated with enhanced leaf N and P through changes in plant biomass and community composition. Fire was associated with reduced leaf nutrient concentrations through changes in plant community composition. Additionally, fire had direct positive effects on Na and nonlinear direct effects on P that partially mitigated the indirect negative effects. Key mechanisms by which fire reduced plant nutritive quality were through reductions of Na-rich grasses and increased abundance of Themeda triandra, which had below-average leaf nutrients.     

Soil phosphorus and disturbance influence liana communities in a subtropical montane forest

Questions: What are the effects of soil, topography, treefall gaps, tree species composition, and tree density on liana species composition and total liana abundance? Location: A 6‐ha permanent plot in a subtropical montane forest in northwest Argentina. Methods: Multiple regressions were used to quantify associations of liana species composition and total liana abundance with edaphic, disturbance and tree community variables. Gradients in liana and tree species composition were quantified using principal components analysis (PCA). Results: Liana species composition was correlated most strongly with soil phosphorus concentration (R²=0.55). Total liana aanased with phosphorus and the density of recent treefall gaps (R²=0.60). Conclusions: In our study area, liana composition and abundance are most strongly correlated with features of the physical environment, rather than host tree characteristics. Our results support the hypothesis that recent increases in liana abundance in mature tropical forests may be related to increased rates of gap formation.     

Short‐term Effects of Single Species Browsing Release by Different‐sized Herbivores on Sand Forest Vegetation Community,South Africa

Manipulations of herbivores in protected areas may have profound effects on ecosystems. We examine short‐term effects on tree species assemblages and resource utilization by a mesoherbivore and small‐size herbivores (ungulates <20 kg) in Sand Forest, after browsing release from a megaherbivore (elephant), or both a mega‐ and mesoherbivore (nyala), respectively. Effects were experimentally separated using replicated exclosures where all trees were counted, identified to species and browsing events recorded. Tree species assemblages were impacted by both elephant and nyala, and by each herbivore species individually. Tree turnover rates were higher where both herbivore species were present than in their combined absence. Diet was segregated among elephant, nyala and small‐size herbivores. Both resource specificity and browsing pressure by nyala increased in absence of elephant; small‐size herbivores increased resource specificity in absence of elephant, and increased browsing pressure in absence of both elephant and nyala. This implies interference competition with competitive release. The indirect effect of the manipulation of herbivore populations, through the removal of one or two herbivore species, caused a shift in tree species composition and diet of smaller‐size herbivores. These indirect effects, especially on tree species composition, can become critical as they affect vegetation dynamics, biodiversity and ecosystem processes. Therefore, in order to conserve habitats and biodiversity across all trophic levels, conservation managers should consider the effects of: (1) the full herbivore assemblage present; and (2) any effects of altering the relative and absolute abundance of different herbivore species on other herbivore species and vegetation.     

Resilience and restoration of tropical and subtropical grasslands,savannas, and grassy woodlands

Despite growing recognition of the conservation values of grassy biomes, our understanding of how to maintain and restore biodiverse tropical grasslands (including savannas and open‐canopy grassy woodlands) remains limited. To incorporate grasslands into large‐scale restoration efforts, we synthesised existing ecological knowledge of tropical grassland resilience and approaches to plant community restoration. Tropical grassland plant communities are resilient to, and often dependent on, the endogenous disturbances with which they evolved – frequent fires and native megafaunal herbivory. In stark contrast, tropical grasslands are extremely vulnerable to human‐caused exogenous disturbances, particularly those that alter soils and destroy belowground biomass (e.g. tillage agriculture, surface mining); tropical grassland restoration after severe soil disturbances is expensive and rarely achieves management targets. Where grasslands have been degraded by altered disturbance regimes (e.g. fire exclusion), exotic plant invasions, or afforestation, restoration efforts can recreate vegetation structure (i.e. historical tree density and herbaceous ground cover), but species‐diverse plant communities, including endemic species, are slow to recover. Complicating plant‐community restoration efforts, many tropical grassland species, particularly those that invest in underground storage organs, are difficult to propagate and re‐establish. To guide restoration decisions, we draw on the old‐growth grassland concept, the novel ecosystem concept, and theory regarding tree cover along resource gradients in savannas to propose a conceptual framework that classifies tropical grasslands into three broad ecosystem states. These states are: (1) old‐growth grasslands (i.e. ancient, biodiverse grassy ecosystems), where management should focus on the maintenance of disturbance regimes; (2) hybrid grasslands, where restoration should emphasise a return towards the old‐growth state; and (3) novel ecosystems, where the magnitude of environmental change (i.e. a shift to an alternative ecosystem state) or the socioecological context preclude a return to historical conditions.     

Decomposing biodiversity data using the Latent Dirichlet Allocation model,a probabilistic multivariate statistical method

We propose a novel multivariate method to analyse biodiversity data based on the Latent Dirichlet Allocation (LDA) model. LDA, a probabilistic model, reduces assemblages to sets of distinct component communities. It produces easily interpretable results, can represent abrupt and gradual changes in composition, accommodates missing data and allows for coherent estimates of uncertainty. We illustrate our method using tree data for the eastern United States and from a tropical successional chronosequence. The model is able to detect pervasive declines in the oak community in Minnesota and Indiana, potentially due to fire suppression, increased growing season precipitation and herbivory. The chronosequence analysis is able to delineate clear successional trends in species composition, while also revealing that site‐specific factors significantly impact these successional trajectories. The proposed method provides a means to decompose and track the dynamics of species assemblages along temporal and spatial gradients, including effects of global change and forest disturbances.     

The relationship between soil seed bank,above‐ground vegetation and disturbance intensity on old‐field successional permanent plots

Questions: How does disturbance and successional age influence richness, size and composition of the soil seed bank? What is the potential contribution of the soil seed bank to the plant community composition on sites differing in their successional age or disturbance intensity? Location: Experimental Botanical Garden of Göttingen University, central Germany. Methods: Above‐ground vegetation and soil seed bank were studied on formerly arable fields in a 36‐year‐old permanent plot study with five disturbance intensities, ranging from yearly ploughing via mowing to long‐term uninterrupted succession. We compared species compositions, seed densities and functional features of the seed bank and above‐ground vegetation by using several methods in parallel. Results: The seed bank was mainly composed of early successional species typical of strongly disturbed habitats. The difference between seed bank composition and above‐ground vegetation decreased with increasing disturbance intensity. The species of greatest quantitative importance in the seed bank was the non‐native forb Solidago canadensis. Conclusions: The ability of a plant community to regenerate from the soil seed bank dramatically decreases with increasing time since abandonment (successional age) and with decreasing disturbance intensity. The present study underlines that plant species typical of grasslands and woodlands are limited by dispersal capacity, owing to low capacity for accumulation of seeds in the soil and the fact that most species do not build up persistent seed banks. Rare and target species were almost absent from the seed bank and will, after local elimination, depend on reintroduction for continuation of their presence.     

Initial response of understorey vegetation to fire severity and salvage‐logging in the southern boreal forest of Québec

Abstract. In this study we compared the effects of fire on understorey vegetation in the Québec southern boreal forest with effects of salvage‐logging (clear‐cutting after fire). All 61 400‐m² sampling sites were controlled for overstorey composition (Deciduous, Mixed and Coniferous) and disturbance type, which consisted of three fire impact severity (FIS) classes (Light, Moderate and Extreme) and two harvesting techniques (Stem‐only and Whole‐tree Harvesting). Percent‐cover data of vegetation and post‐disturbance environmental characteristics were recorded in the field during the first two years after fire as well as soil texture. Ordination of fire alone demonstrated that, on Coniferous sites, fire initiates a succession whereby the understorey Coniferous sites approaches that of Deciduous‐Mixed sites, due to the release of the understorey from Sphagnum spp. dominance, this pattern being a function of FIS. On Deciduous‐Mixed stands, increased FIS resulted in a transition from herb to shrub dominance. Ordination of all five disturbance types showed that the impact of salvage‐logging on understorey composition was within the range of fire, but marginalized to the extreme end of the FIS spectrum. Variance partitioning demonstrated that overstorey and soil texture were the most important explanatory variables of fire alone, while disturbance type explained the largest independent fraction of understorey variation when salvage‐logging was introduced. Salvage‐logging also results in significant reductions in understorey abundance, richness and diversity, while indicator species analysis suggests that it favours mesoxerophytic to xeric species. Results are interpreted in light of shade‐tolerance dynamics, forest floor disturbance and soil moisture regimes. Implications for sustainable forest management are discussed.     

Depth distribution and composition of seed‐banks in alien‐invaded and uninvaded fynbos vegetation

South African fynbos vegetation is threatened on a large scale by invasive woody plants. A major task facing nature conservation managers is to restore invaded areas. The aim of this study was to determine the restoration potential of fynbos following dense invasion by the Australian tree Acacia saligna. The impacts of dense invasion on seed‐bank composition and depth distribution were investigated to determine which fynbos guilds and species have the most persistent seed‐banks. Soil samples were excavated at three different depths for invaded and uninvaded vegetation at two sand plain and mountain fynbos sites. Seed‐banks were determined using the seedling emergence approach. Invasion caused a significant reduction in seed‐bank density and richness at all sites. There was a significant, but smaller, reduction in seed‐bank density and richness with soil depth at three sites. Seed‐bank composition and guild structure changed following invasion. Low persistence of long‐lived obligate seeders in sand plain fynbos seed‐banks indicates that this vegetation type will be difficult to restore from the seed‐bank alone following alien clearance. The dominance of short‐lived species, especially graminoids, forbs and ephemeral geophytes, suggests that regenerating vegetation will develop into a herbland rather than a shrubland. It is recommended that seed collecting and sowing form part of the restoration plan for densely invaded sand plain sites. As seed density remained higher towards the soil surface following invasion, there is no general advantage in applying a mechanical soil disturbance treatment. However, if the shallow soil seed‐bank becomes depleted, for example following a hot fire through dense alien slash, a soil disturbance treatment should be given to exhume the deeper viable seed‐bank and promote recruitment.     

Reproductive potential and seed fate of Sclerocarya birrea subsp. caffra (marula) in the low altitude savannas of South Africa

Even though Sclerocarya birrea subsp. caffra (marula) is a well-studied, keystone tree species with high ecological, commercial and cultural value, significant gaps in our understanding of its reproductive biology exist, particularly the factors limiting fruit and seed production, seed fate and the persistence of the seed bank. Therefore, a detailed quantitative assessment of these factors was conducted at five sites in the low altitude savannas of South Africa. Sites varied with respect to fire regime, large mammals, geology and rainfall. Most sites showed male-biased secondary sex ratios and the minimum fruiting stem diameter ranged between 7.1 and 15.7 cm across sites. Sites with higher levels of disturbance (fire and large browsers) had trees producing fruit at larger minimum stem diameters than sites with lower levels of disturbance. Fruit production was highly variable between individuals, within and between sites, and from year to year. Variability in fruit production across years at one site was greater than the variability across sites in one year, indicating that drivers such as weather, insect herbivory, fire and predator numbers, which vary annually, play a greater role than more constant drivers such as mammalian herbivory, soil types or long-term rainfall. No significant relationship was found between environmental variables (rainfall and temperature) and annual fruit production, indicating a trade-off between vegetative growth and reproduction between years. Since marula fruits are large and heavy, the species relies primarily on mammalian dispersal agents such as the African elephant, which have also been shown to increase the germination rate. However, rodents also appear to play a significant role in seed dispersal. Seed predation rates tend to be highest in areas of low disturbance (no fire and no large browsers). While marula has a small persistent seed bank, recruitment appears to be reliant on the current season's fruit crop. This study provides a detailed quantitative assessment of important reproductive and seed fate vital rates for future population modelling.