Search

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. [ABSTRACT FROM AUTHOR]

Novel fire regimes are an important cause and consequence of global environmental change that involve interactions among biotic, climatic, and human components of ecosystems. Plant flammability is key to these interactions, yet few studies directly measure flammability or consider how multiple species with different flammabilities interact to produce novel fire regimes. Deserts of the southwestern United States are an ideal system for exploring how novel fire regimes can emerge when fire-promoting species invade ecosystems comprised of species that did not evolve with fire. In these deserts, exotic annual grasses provide fuel continuity across landscapes that did not historically burn. These fires often ignite a keystone desert shrub, the fire-intolerant creosote bush, Larrea tridentata ( DC.) Coville. Ignition of Larrea is likely catalyzed by fuels produced by native plants that grow beneath the shrubs. We hypothesize that invasive and native species exhibit distinct flammability characteristics that in combination determine spatial patterns of fire spread and intensity. We measured flammability metrics of Larrea, two invasive grasses, Schismus arabicus and Bromus madritensis, and two native plants, the sub-shrub Ambrosia dumosa and the annual herb Amsinckia menziesii. Results of laboratory experiments show that the grasses carry fire quickly (1.32 cm/s), but burn for short duration (0.5 min) at low temperatures. In contrast, native plants spread fire slowly (0.12 cm/s), but burn up to eight times longer (4 min) and produced hotter fires. Additional experiments on the ignition requirements of Larrea suggest that native plants burn with sufficient temperature and duration to ignite dead Larrea branches (time to ignition, 2 min; temperature at ignition 692°C). Once burning, these dead branches ignite living branches in the upper portions of the shrub. Our study provides support for a conceptual model in which exotic grasses are 'spreaders' of fire and native plants growing beneath shrubs are 'igniters' of dead Larrea branches. Once burning, flames produced by dead branches engulf the entire shrub, resulting in locally intense fires without historical precedent in this system. We suggest that fire models and conservation-focused management could be improved by incorporating the distinct flammability characteristics and spatial distributions of spreaders, igniters, and keystone shrubs. [ABSTRACT FROM AUTHOR]

Question Human-altered disturbance regimes and agricultural land uses are broadly associated with reduced plant species diversity in terrestrial ecosystems. In this study, we seek to understand how fire frequency and agricultural land-use history influence savanna understorey plant diversity through complex relationships (i.e. indirect effects) among multiple biophysical variables. Location Fort Bragg, NC, US, Savannah River Site, SC, US and Fort Stewart, GA, US. Methods We use structural equation modelling ( SEM) to evaluate the relationships among six groups of predictor variables and their influence on local-scale species richness in pine savannas at 256 sites from three locations in the southeastern USA. In the model, fire frequency and agricultural history are hypothesized to control richness through a combination of direct effects, and indirect effects mediated by resource availability, tree abundance, understorey plant abundance and the O horizon (litter and duff depth). Results Frequent fires promote richness by limiting tree abundance, which increases understorey abundance and reduces the O horizon. Frequent fires also limit the O horizon independent of tree abundance. Of the total positive effect of fire on richness, 70% is attributable to reductions in the O horizon and 30% to reduced tree abundance. Agricultural history has a negative effect on richness through a positive correlation with tree abundance, which decreases understorey abundance and increases the O horizon. Agricultural history has a modest negative effect on richness by reducing resource availability as well as a strong direct negative effect (38% of the total effect) that is unrelated to other modelled variables. Conclusions Through a multivariate framework and large-scale data set, this study unites and tests our understanding of the factors that control plant species diversity in a fire-dependent ecosystem. We show that the effects of fire frequency and agricultural history on richness are largely mediated through other ecosystem attributes, including vegetation structure (i.e. tree and understorey abundance), resource availability and the O horizon. Persistent, negative effects of agricultural history demonstrate the slow rates of savanna plant community recovery on post-agricultural land and highlight the conservation value of frequently burned savanna remnants. [ABSTRACT FROM AUTHOR]

Although savanna trees and forest trees are thought to represent distinct functional groups with different effects on ecosystem processes, few empirical studies have examined these effects. In particular, it remains unclear if savanna and forest trees differ in their ability to coexist with understory plants, which comprise the majority of plant diversity in most savannas. We used structural equation modeling (SEM) and data from 157 sites across three locations in the southeastern United States to understand the effects of broadleaf savanna trees, broadleaf forest trees, and pine trees on savanna understory plant communities. After accounting for underlying gradients in fire frequency and soil moisture, abundances (i.e., basal area and stem density) of forest trees and pines, but not savanna trees, were negatively correlated with the cover and density (i.e., local-scale species richness) of C4 graminoid species, a defining savanna understory functional group that is linked to ecosystem flammability. In analyses of the full understory community, abundances of trees from all functional groups were negatively correlated with species density and cover. For both the C4 and full communities, fire frequency promoted understory plants directly, and indirectly by limiting forest tree abundance. There was little indirect influence of fire on the understory mediated through savanna trees and pines, which are more fire tolerant than forest trees. We conclude that tree functional identity is an important factor that influences overstory tree relationships with savanna understory plant communities. In particular, distinct relationships between trees and C4 graminoids have implications for grass-tree coexistence and vegetation-fire feedbacks that maintain savanna environments and their associated understory plant diversity. [ABSTRACT FROM AUTHOR]