Tropical grassy biomes (TGBs) are changing rapidly the world over through a coalescence of high rates of land-use change, global change and altered disturbance regimes that maintain the ecosystem structure and function of these biomes. Our theme issue brings together the latest research examining the characterization, complex ecology, drivers of change, and human use and ecosystem services of TGBs. Recent advances in ecology and evolution have facilitated a new perspective on these biomes. However, there continues to be controversies over their classification and state dynamics that demonstrate critical data and knowledge gaps in our quantitative understanding of these geographically dispersed regions. We highlight an urgent need to improve ecological understanding in order to effectively predict the sensitivity and resilience of TGBs under future scenarios of global change. With human reliance on TGBs increasing and their propensity for change, ecological and evolutionary understanding of these biomes is central to the dual goals of sustaining their ecological integrity and the diverse services these landscapes provide to millions of people.
This article is part of the themed issue ‘Tropical grassy biomes: linking ecology, human use and conservation’.
Historically extensive across the global tropics, tropical grassy biomes (TGBs) are now changing rapidly through high rates of land clearance , increasing land-use intensity [2,3], woody encroachment  and disruption of the disturbance regimes [5,6] that maintain ecosystem function. These biomes were the cradle of human evolution , and in our contemporary world, they support the livelihoods and wellbeing of over one billion people . With the population of Africa alone set to treble by 2050 , the continuing pace of climate change , increasing atmospheric CO2 concentrations  and the increasing agricultural development of TGBs [3,10], there is an urgent need to understand the unique ecology of these systems. TGBs, like tropical forests, are subject to a complex set of pressures as a result of human actions. However, unlike other biomes, the contrasting life forms and physiologies of the dominant C3 woody plant species and the grass species using the C4 photosynthetic pathway sees the future of TGBs linked, in a profound way, to the ever-rising atmospheric CO2 concentration and the global political agenda to reduce these emissions. Furthermore, TGBs have generally few policy and legislative mechanisms in place for their protection (, see also the example provided by do Espírito Santo et al. ).
TGBs contribute 30% of global terrestrial net primary productivity and store 15% of the world's carbon . While TGBs are less carbon dense than forests (by an order of magnitude or more), their productivity is such that large proportions of the carbon gained in a single year are rapidly released back to the environment via fire, herbivory and human use ([14,15] and see analysis of this in Archibald & Hempson ). Indeed, the disequilibrium nature of TGB vegetation dynamics means that these biomes are highly sensitive to annual and decadal changes in environmental controls . The degree to which this dynamism will influence trajectories of vegetation change in grassy biomes into the future is unresolved. However, it is apparent that many intact savannahs are now on a trajectory of increasing woody biomass, although the degree of gain varies regionally, with Australian savannahs most stable over time .
TGBs first arose approximately 10 million years ago and expanded such that by 2 million years ago, tropical savannahs and grasslands were a dominant biome covering the tropics . Today, these biomes cover in excess of 20% of the global land surface. At the last glacial maximum, TGBs extended more widely throughout Asia, Africa and the Americas than today [18–20]. The extent of these vast biomes has shifted with glacial–inter-glacial cycles in response to changing atmospheric CO2 concentrations, rainfall, rainfall seasonality, temperature and fire [19,21]. Given that all of these aspects of our environment are now changing, and generally at rates not previously observed, extensive alterations in the distribution and dynamics of TGBs over the coming century are inevitable and are likely already being observed ([1,4,22], see the analysis provided by Stevens et al. ).
The last decade has seen a revolution in our understanding of the evolution [17,24,25], antiquity [26–28], distribution [29,30] and ecosystem dynamics of TGBs [14,31], as well as their role in the global carbon cycle . Some of these advances have sparked controversies that are now active debates in the literature (e.g. ancient grasslands and afforestation policies [33–35]; the existence or not of alternative vegetation states [29,36,37]). Indeed, insights and theory from savannah ecology have challenged long-standing ecological assumptions of climate determinism in defining the limits of biomes .
Over the coming decade, we anticipate important in-roads will be made in reconciling the complex ecology and biogeochemical cycling of these geographically dispersed biomes via integration of remote sensing, modelling, ecology and evolution. However, it will be critical to incorporate the role of people in shaping and responding to changing ecosystem dynamics and function across this global region, as in the Anthropocene people will be increasingly important agents of landscape change, directly and indirectly influencing the environmental controls and ecological processes that structure TGBs from global to local scales.
2. This issue
TGBs are expansive and changing rapidly, yet our capacity to predict trajectories of change in these biomes is limited, despite their importance to human livelihoods, biodiversity and biogeochemical cycling. In this issue, we highlight the need for integration among research related to the ecology and dynamics of these biomes: characterization and definition of TGBs, complex ecology, patterns and drivers of change, and human use and ecosystem services. For the first time, analyses are presented on the biogeography and potential distributions of Asian savannahs . Other significant steps forward in our understanding include: methods for characterizing ancient versus derived grassy biomes [40,41], comparative data on the species diversity of TGB regions across the globe , an improved understanding of the complex ecology of herbivory and fire [16,43] and the context-dependent response of vegetation to global change , and finally, tools to examine trade-offs in biodiversity, carbon and agriculture to aid land-use planning and policy .
3. Defining tropical grassy biomes
TGBs include C4 grass-dominated savannahs and grasslands (following the definitions of [30,45]). Definitions of TGBs have historically been varied and fraught with problems. Functionally, TGBs are characterized by a grassy ground layer (generally dominated by grasses using the C4 photosynthetic pathway—with a noted exception in Brazil  and Indochina ) and an overstorey varying from 0% up to 60–80% woody cover . The biota, depending on its biogeographic and environmental settings, is tolerant of any, or all of, fire, grazing and browsing . However, universally, the flora is shade intolerant, at least at the establishment phase, due to the open canopy overstorey . While the biodiversity value of these systems to-date has been typically overshadowed by that of tropical forests , Murphy et al.  illustrate the biodiversity value of TGBs, particularly of vertebrates and range-restricted species, and emphasize variation in diversity among the TGB regions (the South American region generally being the richest).
The disequilibrium nature of tropical grassy biome vegetation dynamics, with varying levels of woody cover, has consistently posed problems for the categorization of these ecosystems . This problem has been compounded by a focus on trees, rather than ground layer composition and function; for example, the Millennium Ecosystem Assessment focuses on drylands and forests [47,48], but does not explicitly consider tropical savannahs. While tropical rainforests have been mapped globally, no accurate global map of the TGBs exists. The most widely used general vegetation map and classification scheme is Olson et al.'s 2001 ecoregions , although this biome classification is problematic because it does not recognize some of the world's major savannahs and grasslands, including those in Asia (e.g. India, Thailand, Burma) and Madagascar. Wide use of such maps for research, policy and conservation has the potential to have adverse impacts on landscape management and the perceived conservation value of these regions (e.g. conversion of TGBs, perceived as degraded land, for agriculture, see ). For the first time, Ratnam et al.  review the scattered literature on the distribution of Asian savannahs and evidence for the antiquity and diversity of TGBs across Asia.
In many regions, including Madagascar, Southeast Asia and South America, grassy biomes have historically been considered either a degraded form of forest of anthropogenic origin created via tree clearing, burning and grazing, or a subclimax or secondary successional stage [28,39]. While true in some locations , in the majority of areas this perspective is misplaced . A wealth of new information including dated phylogenetic analyses demonstrates the antiquity of both tree and grass species (and lineages) specialized to these biomes [25,27,28]. The presence of endemic plant lineages and species, as well as species with unique life histories, functional traits and architectures, including forbs with large underground storage organs, are strong indicators of the antiquity of TGBs [35,41].
The fauna of these regions also contains numerous endemic species specialized to open and grassy environments providing additional evidence for the origin and age of the tropical grassy systems. Fauna include species of granivorous birds (e.g. the Madagascan mannikin, Lonchura nana), a suite of grazing ungulates (e.g. the critically endangered Kouprey, Bos sauveli, from Cambodia, and the chital deer, Axis axis, from India) and a high diversity of small marsupials in Australia . Many of these species are endangered and threated with imminent extinction . Fire is a frequent disturbance in the TGBs and has been part of these systems for millions of years [17,21]; consequently, the plants and animals they contain are generally adapted to its occurrence .
To date, it has been difficult to distinguish ancient, old-growth grasslands and savannahs from secondary systems given superficial similarities in structure. Here, Veldman  and Zaloumis & Bond , examining the Neotropics and South Africa respectively, differentiate ancient and secondary systems, noting differences in species composition with the former particularly rich in forbs, many with well-developed underground storage organs that facilitate survival in seasonally dry climates with frequent fire. The challenge is to test the generality of these compositional characteristics across TGB regions.
The ecology of TGBs is complex by virtue of the numerous environmental controls, acting across different scales of influence, both directly and indirectly to structure these ecosystems (figure 1). The last decade has seen a shift from a long-standing view of deterministic relationships among vegetation, climate and soils, focused on niche separation between trees and grasses for water use, to one that integrates niche separation (e.g. phenological, water use) with the controls of fire and mammal herbivory structuring vegetation via the restriction of woody plant growth . Archibald & Hempson  explore trade-offs in fire and mammalian herbivory across the African continent where realms of influence can change through space and time. Complementing this research is that of Anderson et al. , who examine spatial associations of African mammalian herbivores relative to body size and influences on ecosystem function relative to species composition. Both of these studies raise important questions about the function or dysfunction of TGBs in the context of changing disturbance regimes.
Integration of bottom up (e.g. climate and soils) and top down (e.g. fire and mammalian herbivory) controls in structuring TGB vegetation has significantly improved our process understanding of the dynamics and limits of these systems [14,53]. But, it has also highlighted the degree to which contemporary dynamics of TGBs, from local to continental scales, are a function of historical contingencies [44,54]. In assessment of regional patterns and dynamics of vegetation, emergent patterns of woody cover can appear almost stochastic, due to the array of structural states possible for a given set of environmental conditions . At the heart of the current disagreement around alternate vegetation states prevalent in the tropical savannah and forest literature (all of state shifts between savannah and forest, grassland and savannah) may be a lack of recognition of both the role of contingency in influencing contemporary dynamics of TGBs, and that the relative role of environmental controls in structuring vegetation varies across savannah systems and productivity gradients: i.e. some savannahs probably exist due to soil barriers to woody plant growth, while others exist because of controls, such as prevalent fire, that also act to limit plant recruitment and growth. That is, the similarities in structure among TGBs (open canopied vegetation with a predominantly C4 grassy ground layer) have led to an assumption in the literature that the processes regulating vegetation structure across these varied and geographically dispersed ecosystems are directly equivalent. Finally, the presence of numerous, well-documented, feedbacks structuring TGBs where the species composition can influence the strength and direction of effects (figure 1; for example, tree cover–fire; fire–grazing; grazing–browsing), combined with the importance of historical contingencies means that multiple states influencing both the limits and structure of TGBs are highly likely. Oliveras & Malhi  examine the shades of green in our understanding of the processes structuring the limits of TGBs highlighting how biotic and abiotic processes operate at different scales and that the nature of vegetation dynamics is context-dependent.
Savannah vegetation dynamics have been shown to vary as a function of plant traits that aggregate from the individual to ecosystem level [14,56]. However, current model simulations generally represent TGBs as functionally identical, in contrast to ecological knowledge (although in this issue, see Moncrieff et al. ). TGBs constitute a geographically dispersed set of regions, with the flora and fauna of each representing unique evolutionary and environmental histories . The relative importance of environmental controls in structuring these systems varies across these geographical regions, and relative to the environmental niche of each region . For example, the high rainfall Australian savannahs dominated by tall, fast growing, narrow-canopied evergreen Eucalyptus species are less sensitive to fire than the wide-canopied deciduous Brachystegia and Julbernardia species that dominate a savannah region equivalent in area across southern Africa [57,58]. Thus, for a given set of environmental conditions, similar fire frequencies and intensities could produce different vegetation structures, and the difference in sensitivity to fire of these floras is highly likely underpinned by the functional traits of the plant species themselves . It is increasingly appreciated that the functional biogeography of TGBs has critical implications for our capacity to determine the sensitivity and resilience of TGB regions to global change (e.g. ), and yet our quantitative understanding of the functional biogeography of TGBs remains limited. This information is needed as our capacity to predict future change will rely on a quantitative representation of the aggregation of the traits that characterize these floras in influencing ecosystem dynamics and responding to environmental variation.
Across TGBs, vegetation composition, woody cover and grass biomass are considered key determinants of ecosystem function. However, quantitative links between structure and function, ultimately, remain poor and there is no consensus of these relationships among TGB regions . Despite, the antagonistic dynamics between tree and grass dominance being central to savannah ecology, we retain a limited predictive capacity of vegetation structure. It could be argued that our current lack of knowledge about the physio-ecological responses of TGBs to global change is hindered by both the functional differences among the TGB regions and our weak quantitative understanding of the processes that structure vegetation due to the complexity of interactions and scales of feedbacks in operation (figure 1). To aid the management of global change impacts for both people and biodiversity, we need to determine the relative sensitivities of savannah vegetation types to key environmental controls—CO2, water availability and disturbance dynamics—and identify structural thresholds where critical ecosystem functions change.
5. Drivers of change
Tropical vegetation is changing at broad spatial scales but there is a limited understanding of current trends. On one hand, rates of land-use change are increasing [3,11], and on the other, woody encroachment is widespread across savannahs especially in Brazil and South Africa [23,59]. The extent to which drivers that enhance tree growth (e.g. increasing atmospheric CO2 concentrations ([CO2]a), reduced disturbance, improved plant water-use efficiencies), prevail over drivers of enhanced tree mortality (e.g. reduced rainfall, increasing intensity of El Niño, increased temperature, increased harvesting) is unknown, but this is the key to the future management and integrity of the biome.
Rising [CO2]a has long been hypothesized to be a key driver in the re-organization of tropical vegetation, specifically in savannahs where the contrasting life forms and physiologies of the dominant C3 trees and C4 grasses are expected to respond differently . While modelling of the proposed mechanisms underpinning shifts in the competitive interactions between C3 trees and C4 grasses is improving (i.e. increased plant water-use efficiencies of C3 plant species, specifically woody plant species; carbon allocation and storage patterns that vary between life forms; reduced photorespiration in C3 grasses), demonstrating the potential for regional shifts in biome extent and woody biomass , there is a major gap in the experimental evidence of the responses of tropical plant species to altered CO2 concentrations, especially with regards to interactions with other environmental controls . In particular, dominant woody taxa in each savannah region have different life histories, allocation strategies and architectures [14,56,62]. Increasingly, functional traits are recognized as phylogenetically conserved , and differential responses to CO2 would likely be expected relative to both ecological and environmental settings. Looking across a rainfall gradient and land-use types in South Africa, Stevens et al.  report large increases in woody cover in just a few decades providing support for a global driver , while also noting the interaction with megaherbivores (elephants). Woody encroachment may provide carbon benefits, but will undoubtedly come at a biodiversity cost .
TGBs are characterized by seasonally dry and hot climates . While climates across this swath of the world are changing, particularly in terms of the frequency and intensity of El Niño drought events, disagreement among model predictions contributes to the lack of certainty for climate change predictions across tropical regions . Novel climates, in combination with rising CO2, will generate novel interactions among organisms, where small shifts in the season and timing of rainfall may have large consequences for the phenological cycles of plants and animals, and stark consequences for crop production . By contrast, small changes in total rainfall may be of limited consequence, where increasing atmospheric CO2 concentrations will drive improvements in plant water-use efficiencies . Temperature is assumed important in determining plant distributions and function primarily based on assumptions from the Northern Hemisphere. Yet, the importance of temperature in the dynamics of TGBs is poorly understood. The small body of research suggests if there is sufficient water, a warming climate may enhance plant success through improved germination [67,68] and sapling growth rates , and an extended growing season .
Yet, across TGBs, rates of land-use and cover change appear to exceed the effects of climate change [1,11]. With increasing global scarcity of lands for agriculture and increasing food demands , land-use intensity is only likely to increase. There are a multitude of land-use types across TGBs, many of which are context-dependent, from shifting cultivation and grazing lands to commercial agriculture . However, all affect the continuity of ecosystems and, some land-use types more so than others . Increasing land-use intensity and fragmentation disrupts disturbance regimes and vegetation dynamics , potentially amplifying a global change-induced trend of encroachment by further reducing tree mortality. To date, 50% of the Brazilian cerrado has been transformed for agriculture, a rate of land-use change roughly double that of the Amazon forest [73,74]. Land has historically been perceived as being of marginal agricultural value across TGBs [10,58]. However, technical innovations in managing highly weathered tropical red soils and the breeding of suitable crop varieties have transformed agriculture in Brazil . This tropical agricultural revolution has been proposed as a viable development model for wetter African savannahs [3,10]. Do Espírito Santo et al.  document land abandonment and encroachment in secondary savannahs of the Brazilian cerrado , where this development policy has been to the detriment of the integrity of the system and where, globally, rates of encroachment are highest .
Finally, people also directly influence disturbance regimes at broad scales [6,72]. Active suppression of fire in the savannah regions of Asia and Brazil in particular has facilitated woody or weed encroachment [59,75]. The increasing extent of roads and fences act as fire breaks and also prohibit large-scale animal movements . These changes to the major savannah processes of fire and herbivory, combined with the impacts and effects of the poaching crisis of Africa and Asia, are likely to have profound consequences for ecosystem function [16,43].
6. Human use and value
Quantifying and understanding the value of TGBs to humans is challenging because in many regions, particularly in Asia, data on the value of TGBs to human livelihoods are limited. TGBs are sometimes described as ‘unused’ or ‘degraded’, although these systems provide fundamental resources and ecosystem services supporting the livelihoods of the millions of people living in these regions. Furthermore, the people of these regions are among the world's poorest and most vulnerable , and global change will inevitably affect ecosystem services and resource availability. Perhaps more so than any other TGB region, direct use by local communities is greatest in Africa and Asia , given the urbanization of Latin America and the sparse population densities of Australia. Ryan et al.  highlight the diversity and number of ecosystem services (supporting, regulating and cultural services) provided by TGBs (specifically in relation to southern Africa) that, to-date, have typically been either overlooked or considered at small spatial scales. Critical resources provided include food (wild fruits, tubers, nuts, edible insects, bushmeat), non-timber forest products for sale (e.g. honey, beeswax, insects), fuel (fire wood and charcoal), construction materials (e.g. thatching grass, timber), water, nutrient cycling and medicinal plants .
Across Africa, Asia and even South America, fuelwood harvesting is a significant activity [12,78]. Although the fuelwood crisis predicted in the 1970s has not materialized, projections surrounding levels of fuelwood sustainability are varied . In some regions (e.g. South Africa), fuelwood harvesting is considered sustainable due to regeneration after coppicing , possibly facilitated by CO2 fertilization (e.g. ). Elsewhere, wood demand is anticipated to increase due to increasing populations and a switch to charcoal, which has an increasing international market . The implications of changing wood demands are raised by Woollen et al. : wood for construction material is now traded off against wood for charcoal. How changes in wood resources in the context of global change will influence the integrity and functioning of TGBs needs urgent attention.
Water availability will change with growing populations, altered landscapes and global change, and in the seasonally dry climates of these regions water availability is critical to both ecosystem dynamics and human use of landscapes, yet an understanding of the nature of this change and the implications is restricted to a few regions . Honda & Durigan  provide, for the first time, estimates of rainfall partitioning in the cerrado and demonstrate how fire suppression, indirectly via woody encroachment, can reduce rain interception. Given many TGB regions are experiencing woody encroachment, this is a reminder of the functional role of disturbance in the provision of water resources. Across TGBs, afforestation (often of exotic species) is common , in part, to meet fuelwood needs; improved examination of the trade-offs between fuelwood and woody encroachment with water are desperately required. Further, much needed are examinations of the biophysical and biogeochemical consequences of woody encroachment and land cover change, such as albedo and nutrient cycling.
People and institutional structures can strongly influence and affect TGBs—their biodiversity, functioning and services . For example, in many areas, declines in mammal species have been reported due to bushmeat hunting and poaching, savannahs have become degraded through overgrazing, land abandonment and afforested monocultures [81–83]. The economic importance of the wildlife across TGBs, particularly in African and some Asian regions (and to a lesser extent Australia), is a unique economic advantage that contributes to the tourism industries of these regions. Major ecological changes driven by the global changes set out above, combined with human pressures, could threaten these economic benefits. Human usage needs to be reconciled with ecological values in these areas, including biodiversity; human activity has already reduced the richest botanical savannah region, the Brazilian cerrado, to a series of, arguably, dysfunctional fragments. Finally, the climate change mitigation agenda could represent a threat to TGBs, as there is increasing talk of the need for ‘negative emissions’ to meet the emissions targets set out at the Paris climate conference (COP21) in 2015 where inappropriate application of these targets could lead to afforestation of TGBs .
The pace and scale of change in TGBs is astonishing and will affect human livelihoods, biodiversity, carbon and biogeochemical cycling. Yet, our capacity to predict the direction and extent of change, as well as the consequences, is currently limited. Research will improve our ecological understanding of TGBs, and it is clear that our capacity to effectively predict the sensitivity and resilience of each TGB region is dependent upon understanding the cascading and interacting effects of ecological, socioeconomic and global drivers across contrasting contexts. This is no easy task even from a purely ecological perspective as these systems are complex, ecosystem dynamics are context-dependent and uncertainty surrounds the influence of global change drivers.
To better conserve and manage TGBs for the future, we must look beyond a simplified view considering only the tree layer, to a perspective that embraces the grassy ground layer and the unique functions associated with this. Only with this broader perspective will we be in a position to consider the range of trajectories and possible states that are likely across the different regions and how the influence of key drivers may vary. To apply the most appropriate conservation and management efforts in the right place, field studies are needed to characterize and determine the antiquity and value of TGBs more broadly. Additionally, field studies will help us understand how the multiple pathways for structural (and compositional) change link to the functioning of these biomes.
Experiments manipulating global change drivers (e.g. water availability, temperature and CO2) will help unravel the complexities of savannah process and dynamics so we are in a stronger position to understand how different TGB regions may respond to future change. Ecologists need to work with land managers, policy agencies and governments to ensure that processes critical to the healthy functioning of TGBs (i.e. fire, herbivory) are maintained. This will mean revisiting carbon mitigation initiatives, taking a more nuanced approach to applying REDD+ in TGBs, and ultimately recognizing TGBs are wholly different to forests in terms of ecosystem function. Many of the world's poorest live in TGBs; it is therefore essential that the dual goals of sustaining the ecological integrity of these biomes and supporting the livelihoods of the people who live in these landscapes are not viewed as competing demands.
C.L.P. and C.E.R.L. contributed equally to the development of this issue.
We declare we have no competing interests.
We received no funding for this study.
We thank William Bond, Nicola Stevens and Sally Archibald for helpful discussions and for comments on this manuscript. Helpful comments from two reviewers improved this manuscript. A huge thank you to Helen Eaton for her constant support and enduring patience with us during this process. Thank you to all the authors for contributing such interesting research to this issue. Finally, we are grateful to the reviewers who so ably (and rapidly!) aided the compilation of this issue.
One contribution of 15 to a theme issue ‘Tropical grassy biomes: linking ecology, human use and conservation’.
- Accepted July 1, 2016.
- © 2016 The Author(s)
Published by the Royal Society. All rights reserved.
Guest editor biographies
Caroline E. R. Lehmann is a lecturer in biogeography in the School of GeoSciences at the University of Edinburgh. Caroline's research examines the distributions and dynamics of tropical biomes, and tropical plant species in an ever changing world. Caroline's current focus is understanding how variation in plant functional traits can aggregate to impact the dynamics and distributions of biomes, particularly in an evolutionary and biogeographic context, and using this information to aid development of useful predictions of tropical savannas in the Anthropocene.
Catherine L. Parr is a Senior Lecturer in the School of Environmental Sciences, at the University of Liverpool, and adjunct faculty at the University of the Witwatersrand, South Africa. She is a community ecologist with a particular focus on ants and termites. Catherine's current research interests include diversity change along gradients and with disturbance, the role of dominant social insects in ecosystem structure and functioning, and conservation of savanna biodiversity. She works mostly in savannas in the tropics, but also in montane grasslands and rainforests.