Articles
A Framework for Exploring the Determinants of Savanna and Grassland Distribution ANTHONY J. MILLS, KEVIN H. ROGERS, MARC STALMANS, AND ED T. F. WITKOWSKI
An understanding of the factors governing grass–tree coexistence in savannas and exclusion of trees in grasslands remains elusive. We contend that progress in understanding these factors is impeded by a reliance on a falsification approach and an excessive concern over type I errors (false positives), which results in premature rejection of hypotheses, inadequate attention to scale, and a miring rather than galvanizing of ecological discussions. An additional hindrance to progress may be that investigations tend to focus on processes within either savannas or grasslands, while ignoring the boundary between the two. We propose a new scientific framework for identifying determinants of savanna and grassland distribution, which advocates (a) the recognition of ecosystems and biomes as complex adaptive systems, (b) a scientific methodology based on adaptive inference, and (c) explicit consideration of patch boundaries at various scales. Analysis of processes operating at dynamic savanna– grassland boundaries should permit better separation of ultimate from proximate factors controlling grass–tree interactions within the individual biomes. The proposed savanna–grassland framework has potential for application in other areas of ecology facing similar problems. Keywords: boundary, scale, ecotone, nutrient limitation, savanna problem
T
he coexistence of two very different life-forms within savannas, namely grasses and trees, has perplexed ecologists for decades and has been labeled the “savanna problem” (Sarmiento 1984). Because the lack of trees in many grasslands is equally perplexing, we suggest that there is actually a “savanna–grassland problem”that applies to a large proportion of the global terrestrial landscape. To date, much of the research on the savanna problem has taken a classical, Popperian approach, whereby data (usually derived from small-scale experiments) are presented to refute or falsify theories about processes that must have operated over long time frames and over a range of spatial scales. In our view, this heavy reliance on falsification is impeding progress toward an understanding of the ultimate factors governing savanna and grassland distribution. In this article we depart from such a falsification approach and instead plot a fundamentally new course for tackling this problem: a course based on adaptive inference (Holling and Allen 2002), consideration of scale, and recognition of ecosystems and biomes as complex adaptive systems (Levin 1998). Using the analogy of a horse race, the classic hypotheticodeductive approach, in which ecologists quickly select and pursue a candidate hypothesis using a process that avoids type I errors (false positives), is analogous to shooting most of the horses at the starting gate before the race gets under way. When the chosen horses fail to stand the test of time, the jockeys go back to the start and either try to revive some of www.biosciencemag.org
the prematurely culled horses or else begin the race over again with different horses, but with the same rules that eliminated most of them before the race. We propose a fundamentally different approach whereby all horses (even those that appear lame on first appearance) are nurtured and coaxed to their full capacity. With this approach, there is not necessarily a single winner, but rather a recognition that ecological understanding will arise from the results of races run at different scales at different times. These races may take place in different contexts, such as macroecology, microbiology, or soil science. The difficulty of advocating this new approach is that ecologists mired within the falsification paradigm will be tempted to shoot it down before it has seen the light of day. In all likelihood, it will be up to a different generation of ecologists to take up the challenge and develop this new paradigm further. Many hypotheses have been put forward to explain the savanna–grassland problem (box 1). Only one, however— namely, fire (Bond et al. 2003, Bond and Keeley 2005, Bond Anthony J. Mills (e-mail:
[email protected]) is a soil scientist and ecologist in the Department of Soil Science, Stellenbosch University, Matieland, 7602, South Africa, and the South African National Biodiversity Institute, Claremont, 7735, South Africa. Kevin H. Rogers and Marc Stalmans are ecologists, and Ed T. F. Witkowski is a plant ecologist, in the School of Animal, Plant and Environmental Sciences, University of the Witwatersrand, PO Wits, 2050, Johannesburg, South Africa. © 2006 American Institute of Biological Sciences.
July 2006 / Vol. 56 No. 7 • BioScience 579
Articles Box 1. Hypotheses explaining the lack of trees in South African grasslands.
Humans. According to one early hypothesis, grasslands were anthropogenically derived and maintained by fire over the last 1000 years (Acocks 1953). This view has been discounted because of the great diversity and large degree of endemism within these grasslands (McKenzie 1989), and because palynological data show the presence of this biome since the late Pleistocene (>10,000 years before the present; Meadows and Linder 1993), well before cultivation (Feely 1987). Waterlogging. Another hypothesis suggests that regolith structures (e.g., shallow, impermeable layers or illuvial B horizons), which cause waterlogging in the wet season and yet reduce water storage capacity in the dry season, exclude trees from grasslands (Tinley 1982). Although grasses are often dominant on poorly drained soils within savannas, this contention is incompletely supported, as grasslands do occur extensively on well-drained, loose, deep, sandy soils (Werger and Coetzee 1978). Frost. A third hypothesis proposes that frost excludes woody elements across large parts of the grassland biome (Acocks 1953). Although frost limits the pool of tree species available for colonization of grassland, it does not seem to be an adequate explanation for their exclusion, because there are many exceptions, such as Acacia karroo, Rhus lancea, Leucosidea sericea, and other woody plants that tolerate frost. Fire. A more widely accepted view is that frequent and intense fires (promoted by specific climatic conditions) exclude trees from the grassland biome. High productivity (as a result of relatively high rainfall), seasonal growth, and frost curing result in high fuel loads, which, together with a high incidence of lightning, sustain frequent and intense fires. Fire frequency is likely to increase with rainfall (Van Wilgen et al. 2004), but O’Connor and Bredenkamp (1997) noted that invoking fire as the main agent excluding trees from grasslands is “possibly inadequate given the occurrence of satellite grasslands not dependent on fire and some savannas well adapted to high fire frequencies.” Coetzee and colleagues (1994) also noted that the fire regimes of savannas and grasslands north and south of Pretoria, respectively, do not differ. Nutrient availability. The water content of topsoils (0 to 10 centimeters) during the growing season is greater in grasslands than in adjacent, drier savannas. One hypothesis is that this promotes continuous mineralization of soil organic matter and release of nutrients, which renders grasses more competitive than woody species, irrespective of fire regimes (Mills 2003). Dominance of grasses over woody species, even in the absence of fire, is not without precedent (Knoop and Walker 1985). Peters (2002) modeled grass–shrub interactions from field observations in semiarid New Mexico and predicted that increased summer precipitation would result in dominance of the grass Bouteloua eriopoda over the shrub Larrea tridentata. The grass was more competitive on physiological grounds, and its dominance was not related to fire. Similarly, Palmer and colleagues (1999) note that in the South African Karoo, dominance of shrubs over grasses increases as the coefficient of variation of mean annual rainfall increases. et al. 2005)—appears to be considered seriously at present. Our purpose is not to dismiss this hypothesis, but rather to point out potential contradictions and flaws in the assumptions used to support it, and to examine why other hypotheses have failed to gain support. Our overall aim is to restart the race with a new set of conventions and philosophies, call the favorite back to the start, resuscitate the injured, and finally encourage the birth of some new foals.
A challenge for ecologists Conservation managers rely on ecology to provide answers to problems encountered on the ground, yet answers are often conspicuously absent. When it comes to savanna and grassland conservation, an understanding of the fundamental processes governing tree abundance in a landscape would probably help managers to tackle three critical questions: (1) Why are grasslands largely treeless? (2) Why are tree densities increasing in many savannas? and (3) Why are trees encroaching into many grasslands? To take South Africa as a 580 BioScience • July 2006 / Vol. 56 No. 7
case study, fire and frost are often cited as the causal agents maintaining largely treeless landscapes on the Highveld (Acocks 1953, Bond et al. 2003), yet numerous tree species, both within South Africa (e.g., Acacia karroo, Protea caffra, Protea roupelliae, and Leucosidea sericea) and worldwide, are capable of withstanding frequent, intense fires as well as frost (Rundel 1981). The lack of trees in South African grasslands has also been attributed to accidents of evolutionary history, whereby trees tolerant of both frequent grass fires and cooler climates (such as eucalypts) did not evolve in the region (William J. Bond, Department of Botany, University of Cape Town, Cape Town, South Africa, personal communication, 30 March 2005). In the face of imperfect knowledge of what controls grass–tree interactions, it is difficult to assess the role of chance and of environmental factors in shaping grasslands and savannas. Problems of scale also emerge when trying to answer the questions posed above. Results from grass–tree plot-scale experiments (hundreds to thousands of square meters) within www.biosciencemag.org
Articles a patch of savanna may contradict results of broadscale studies across biomes (hundreds of square kilometers). For example, frost may be more frequent in grasslands than in savannas at a biome scale in South Africa, but in some localities, as a result of cold air drainage, frost is more frequent on lower wooded slopes than on the upper grassy slopes. Gosz and Sharpe (1989) similarly noted that biome delineation is often correlated with large-scale climatic features, whereas finescale ecotones can be determined by site-specific characteristics such as soil discontinuities. The challenge for ecologists is to identify processes operating at different scales and to differentiate the drivers from the modifiers of biome structure. Can factors such as fire or herbivory, for example, change a savanna into a grassland or prevent trees from encroaching into grasslands, or are these factors only “modifiers” (Stott 1991), with the ultimate driver of biome structure being climate? Contingency may also play a role, whereby the vegetation structure of a particular landscape is dependent on a unique set of interacting factors (McNaughton 1983). Subtle abiotic differences and biotic interactions within apparently similar landscapes could thereby result in vastly different vegetation structures. Levin (1998) observed that traditional approaches in ecology are inadequate for broadscale questions, such as what factors determine biome distributions, because of a divide between population and ecosystem scientists. He suggested that a way forward is to recognize ecosystems as complex adaptive systems, in which patterns at higher levels emerge from localized interactions and selection processes operating at lower levels. Such systems are nonlinear, with historical dependency and multiple possible outcomes of dynamics. In addition, the systems have been assembled from parts that have evolved over longer timescales and broader spatial scales than the current system or biome. Levin (1998) proposed several pertinent questions for determining the degree to which system features are determined by environmental conditions or by self-organization. These include, among others, (a) Are patterns of biodiversity distribution and organization uniquely determined by local conditions, or are they historically and spatially contingent? (b) How do ecosystems become assembled over time, particularly with respect to evolutionary processes? and (c) What are the relationships between ecosystem structure and functioning? We propose that to analyze the determinants of grassland and savanna distributions within the context of complex adaptive systems and Levin’s (1998) questions, a new, broadbased and integrative conceptual framework is required. Frameworks serve as scientific maps for new areas of endeavor and show how facts, hypotheses, models, and expectations are linked, thereby indicating the scope to which a generalization or model applies (Pickett et al. 1999). They also encourage interdisciplinary interaction at appropriate scales and help to order phenomena and material, thereby revealing patterns (Rapport 1985). No such framework has been developed for the savanna problem, because most ecologists have persisted in the pursuit of a single cause for the distinction www.biosciencemag.org
between the two biomes. The framework we propose should allow ecologists to separate ultimate from proximate determinants at different scales. It is based on principles of landscape ecology (Gosz and Sharpe 1989) and adaptive inference. The framework emphasizes the potential importance of (a) scale, (b) processes operating across boundaries, (c) confirmatory data, and (d) the development of multiple lines of reasoning.
Separating ultimate from proximate factors in the grassland–savanna problem The perplexing absence of trees in South African grasslands (where mean annual rainfall exceeds 700 millimeters [mm] in many parts) is a prime example of the savanna–grassland problem. According to Tainton and Walker (1993), “The interaction of rainfall, temperature (particularly frost), fire and soil type determines the type of vegetation, but it’s not always clear how or why some areas are pure grasslands and others not” (p. 271). O’Connor and Bredenkamp (1997) concluded that the distribution of grasslands is governed by a “subtle interplay of climate, topography, fire and grazing.” The difficulty of isolating the factors and subtle interplays determining vegetation structure is not, however, restricted to grasslands and savannas. Orians and Solbrig (1977) noted three decades ago that “predictive theories about community structure and functioning are nearly absent in ecology” (p. 254), a statement that still largely rings true today. A variety of factors is likely to affect the development of a grassland or savanna (figure 1). Feedback effects between several of the components are evident. Separating proximate from ultimate factors is difficult. A chicken and egg problem often develops. Fire, for example, may be a proximate factor, an inevitable result of the combination of dry grass and lightning. The presence of large herds of herbivores such as black wildebeest (Connochaetes gnou), springbok (Antidorcas marsupialis), and extinct large mammals such as the giant buffalo (Pelorovis antiquus) and giant hartebeest (Megalotragus priscus; Klein 1984) during the evolution of southern African grasslands may also have been a proximate factor, given that grazers are likely to be attracted to flushes of grass after fire. The classification of biomes and vegetation types tends to perpetuate rather than alleviate the chicken and egg problem. This is because various abiotic and biotic characteristics (e.g., frequency of fire, cover of grass, incidence of frost, hydrological status of soils) are used to divide the landscape into defined units, and the abiotic characteristics are often subsequently assumed to be causal. Hypotheses of grass–tree coexistence also suffer from chicken and egg problems. Demographic-bottleneck models, such as the storage effect hypothesis (Higgins et al. 2000), contend that frequent fires, competition from grass, and soil moisture limitations usually prevent the recruitment of tree seedlings in savannas, and that adult trees store the potential for seedling recruitment for the few occasions when abiotic conditions are suitable. Competition-based models, by July 2006 / Vol. 56 No. 7 • BioScience 581
Articles content, soil temperatures, rates of mineralization, and light availability (Stock and Lewis 1986, Blair 1997, Knapp et al. 1998), all of which are likely to influence the competitive ability of grasses and trees. For example, the mean sunlit photosynthetic rates of several trees in a hardwood forest in Wisconsin were stimulated after fire, probably because of an increase in nitrogen availability (Reich et al. 1990). There may also be contingency effects, whereby fire interacts with soil moisture, nutrient availability, light availability, and grass–tree competition in subtle, nonlinear ways. As McNaughton (1983) noted with respect to Figure 1. Hypothetical relationships between factors likely to affect the evolution of a grassland the Serengeti grasslands,“the or savanna. Note that there are feedback effects from several of the center components; even cli- proximate mechanisms regmate is not immune from feedback. There are also numerous feedback effects between the cenulating species abundances ter components (e.g., between fire and herbivory pressure, and between fire and are many weak forces acting mineralization) not indicated in the diagram. probabilistically, so that the cumulative effects are large, but the individual effects are minor, interactive, and uncercontrast, propose that trees and grasses coexist because of their tain” (p. 315). He highlighted grazing as one force dependent differential ability to acquire and partition resources (see on many intersecting probability functions, such as species Walter [1971] for an explicit example of classic niche sepacomposition, phenological stages of grasses, tree canopy denration through separate rooting zones). Both models propose sity, species of grazers, density of grazers, soil properties, frethat one main factor or several interacting factors determine quency of burning, and soil water content. The influence of a particular ecosystem state. Yet the factors are not indepeneach one of these factors is likely to vary through time, and dent of the state, and consequently there is a danger of circular it may prove more fruitful to map and describe such dyreasoning. Grass competition, fire frequency, and soil moisnamic probability functions rather than pursuing static, linture are, for example, highly dependent on the amount of grass ear chains of cause and effect when examining factors that and tree biomass. govern biome distribution. Effects of fire exclusion on vegetation structure have recently Despite the apparently compelling evidence from models been modeled using dynamic global vegetation models and long-term experiments supporting the present role of fire (DGVMs). Results from the Sheffield DGVM suggest that “vast in tree exclusion, an intriguing inconsistency remains: Why areas of humid C4 grasslands and savannas, especially in did fire-tolerant tree floras not evolve to dominate all fireSouth America and Africa, have the climate potential to form prone ecosystems, when trees from families such as Caeforests” (Bond et al. 2005), and that if fire were excluded, most salpiniaceae, Fagaceae, Pinaceae, and Myrtaceae cover vast of the eastern half of South Africa would be “dominated by areas of the globe? Trees show great plasticity with respect to trees instead of grasses”(Bond et al. 2003). Findings from longfire tolerance, with adaptive features such as corky bark (e.g., term fire exclusion experiments in grasslands and savannas Quercus suber), seedlings with protective grassy covering appear to support the model results, in that woody biomass (e.g., Pinus palustris), lignotubers, and epicormic buds (e.g., usually increases with a decrease in fire frequency. The posAcacia and Eucalyptus spp.). Furthermore, convergent evosible mechanisms by which fire excludes trees from grasslands lution shows that plants in similar environments on differand the potential role of grass–tree competition require some ent continents evolve into remarkably similar growth forms discussion. (Orians and Solbrig 1977, Cody and Mooney 1978, Cowling The effects of fire are likely to be considerably more comand Witkowski 1994). The principle of convergent evoluplex than the direct damage caused to living tissues. Fire intion seems, however, to falter when it comes to fire-prone flofluences subsequent (postfire) nutrient availability, soil water 582 BioScience • July 2006 / Vol. 56 No. 7
www.biosciencemag.org
Articles ras, with some fire-prone systems being dominated by firetolerant trees and others by fire-tolerant grasses. This is a major inconsistency when invoking fire as a primary determinant of vegetation structure across large parts of the planet. An alternative view that may resolve the issue is that vegetation structure is primarily governed by climate, and fire is proximate and incidental. If the climate in a fire-prone environment favors the dominance of low-growing herbaceous vegetation for physiological reasons, it is conceivable that through evolutionary time a positive feedback could develop, whereby the vegetation benefits from fire because nutrients are returned to topsoils or moribund material is removed, or both. The vegetation is likely to evolve attributes that promote fire, and the competitive ability of the vegetation may become largely dependent on fire. Woody plants invading herbaceous vegetation are likely to be limited by the competitive ability of the herbaceous vegetation (through competition for resources such as light, water, and nutrients), but will also probably not be tolerant of the specific fire regime that coevolved with the herbaceous vegetation. If, however, climate in another fire-prone environment favors the growth of trees, then through evolutionary time, a tree flora may emerge that is tolerant of or promotes fire, and this flora may even come to depend on nutrient-rich ash beds for the germination and successful recruitment of seedlings. This could explain the wide range in vegetation structure across fire-prone ecosystems and the increase in woody vegetation in fire-exclusion experiments. Herbaceous vegetation that coevolved with fire over millions of years is unlikely to be competitive when fire is removed. Removing fire from fire-dependent grasslands is likely to reduce grass vigor and thereby create an artificially competitive advantage for trees. Consequently, it may be premature to conclude from the results of fire-exclusion experiments (Bond et al. 2005) that fire is more important than climate in shaping vegetation structure. Indeed, the wide range of fire regimes evident across southern African vegetation types suggests that, through evolutionary time, plants exerted a strong influence on the fire regime rather than the other way around. Subtropical thicket in the Eastern Cape, South Africa, for example, occurs in a warm, semiarid climate (250 to 650 mm mean annual rainfall) with a large potential for growth of grass, yet can exclude fire because of the relative paucity of grass cover and the predominance of succulent shrubs such as Portulacaria afra (Vlok et al. 2003). Rainforest patches in northern Australian savannas provide another example of vegetation that excludes fire. Bowman and colleagues (2004) showed that rainforest patches in some regions occur on the most clayey soils. They suggested that the rainforest species thrive in these fertile, moist soils and can persist by preventing grass recruitment and excluding fire. Much of our understanding of fire–vegetation interactions has been derived from experimental studies at the plot scale. It is difficult to use experiments (e.g., fire exclusion, watering, nutrient addition) to separate ultimate from proximate causes because of mismatches in temporal scale. Savanna and www.biosciencemag.org
grassland biomes did not develop under the conditions imposed in such experiments. If such conditions had been present through evolutionary time, and yet the ultimate factors were still in place, it is conceivable that similar biomes would have arisen, but with different ecological processes. If fire, for example, had been excluded from the South African Highveld for several million years, a grassland might have developed that relied on microbial decomposition and herbivores, rather than fire, for nutrient cycling. It is conceivable that today’s grasslands are a response to a certain suite of abiotic conditions in soils (e.g., water content, degree of aeration, specific macro- or even micronutrient content) that promote grasses, and thereby enable them to dominate and exclude trees, irrespective of the fire regime. Although experiments can be instructive, it is increasingly acknowledged that without complementary emphasis on large-scale phenomena through time and space, it is difficult to determine which results reflect idiosyncrasies of individual treatments, species, or site conditions, and which reflect the operation of more universal processes (Brown 1995). An alternative explanation for the lack of fire-tolerant trees in South African grasslands is an accident of evolutionary history. It may be that tree genera that are extremely tolerant of fire in other parts of the world (e.g., Eucalyptus, Melaleuca, Banksia, Pinus, Quercus) by chance did not migrate to or evolve in South Africa (Richardson et al. 1992). Ackerley (2004) notes that the presence of a species within an ecosystem does not necessarily reflect a process of natural selection and adaptation, because many species may have migrated into the system from surrounding areas. Nevertheless, convergent evolution (e.g., such as occurs across Cactaceae and Euphorbiaceae) suggests that plant growth forms are flexible, highly responsive to abiotic conditions, and unlikely to be limited by the available gene pool. Furthermore, many tree families present in South African grasslands (in forest patches, along drainage lines, or on rocky slopes) have fire-tolerant tree genera in other parts of the world (e.g., Proteaceae and Myrtaceae). This suggests that genetic limitations for the development of a fire-tolerant tree flora were unlikely, and we consequently propose that the principle of convergent evolution warrants an in-depth analysis in the context of the savanna–grassland problem. Furthermore, we caution that the “accident of evolutionary history” explanation is potentially dangerous when deployed within the falsification paradigm, because in its extreme form, no abiotic explanations of vegetation structure are necessary, given that all observed differences between biomes can be explained away by chance. Unfortunately, this explanation starts from an inherently domineering position (in that it can be invoked to explain any pattern), and it inevitably stultifies or nullifies research focusing on abiotic explanations for ecological differences at large scales (e.g., across continents). It is evident that scientists are no closer than Tainton and Walker (1993) were 13 years ago to solving the question of why some areas are pure grassland and others are not. The principle of convergent evolution suggests that fire-tolerant woodJuly 2006 / Vol. 56 No. 7 • BioScience 583
Articles lands should have developed in many pure grasslands. Smallscale experiments are unlikely to separate ultimate from proximate factors governing vegetation structure, because of mismatches in temporal scale, and research within the individual biomes tends to run up against problems of circular reasoning. Ecologists need to regroup and develop a new strategy to deal with these problems and to tackle this intractable question.
Toward a new framework that encompasses complex adaptive systems, scale, boundaries, and adaptive inference Frameworks facilitate linkages between different paradigms and prepare common ground for scientists using different conceptual approaches. In this light, we discuss below several different approaches that we advocate using in tandem for homing in on the factors ultimately determining biome distributions. We suggest that new insights are likely to arise from the mere process of combining methodologies and ideas. In the words of Jacob (1977), “Novelties come from previously unseen association of old materials. To create is to recombine” (p. 1163). Complex adaptive systems and scale. The identification of ultimate and proximate factors governing vegetation structure requires consideration of complexity, chance, evolution, and scale. Levin (1998) brings this diverse range of issues together under the banner of “complex adaptive systems” and thereby provides a useful starting point for investigating the
savanna–grassland problem. Tackling the different issues will require research on several fronts. To date, most grassland and savanna research has taken place within the biomes. We suggest that a new focus on scale and processes operating at grassland–savanna boundaries will generate additional insights into factors influencing the individual biome structures and distributions. A scaled approach was recently used by Gillson (2004) in an East African savanna. Using the concept of hierarchical patch dynamics (Wu and Loucks 1996), Gillson suggested that different ecological processes—namely, soil type, disturbance by fire or herbivory, and climate—determine tree abundance over hundreds of years at micro, local, and landscape scales, respectively. A similarly scaled perspective on boundaries (Gosz and Sharpe 1989), from individual plant patches to the biome boundary (figure 2), could be very useful in separating ultimate and proximate causation in grassland and savanna differentiation. Using boundaries to separate proximate from ultimate causation. Boundary research has principally aimed at documenting and understanding boundaries. We suggest, however, that boundaries can be used for understanding ecological processes shaping the communities or biomes on either side of the boundary. This is a major conceptual departure both from current boundary research and from “within-biome” research. The potential relevance of boundaries to withinbiome research is highlighted by Cadenasso and colleagues (2003), who note that because the patches that the boundary separates are distinguished from each other by some defin-
Figure 2. A hierarchical patch-dynamic approach to the study of the grassland–savanna boundary. Factors governing nitrogen mineralization and boundary patchiness highlight the effect of scale on ecological processes. Abbreviations: cm, centimeter; km, kilometer. 584 BioScience • July 2006 / Vol. 56 No. 7
www.biosciencemag.org
Articles ing characteristic, the gradient in that characteristic is steeper in the boundary than in either of the neighboring patches. This suggests that ultimate factors governing grassland and savanna distribution are likely to show sharp changes over the boundaries between the vegetation types and are likely to control shifts in boundaries over time. By contrast, we suggest that proximate factors within grasslands or savannas are likely to be more loosely tied, and may or may not show changes at boundaries (figure 3). Scrutiny of processes occurring across boundaries can therefore potentially identify proximate factors and by a process of exclusion develop a shortlist of potential ultimate factors (figures 3, 4). Boundaries that are shifting through time are likely to be particularly powerful in this regard (figure 4). Potential ultimate factors driving a boundary shift (e.g., soil moisture in surface soils) would, for example, be expected to be closely tied to the present boundary line and might even be slightly ahead of it (figure 4). In contrast, potential proximate soil properties (e.g., soil organic matter, pH, base status and redox status at depth) may lag behind the boundary shift. It is likely that processes and differences across boundaries, as yet not envisaged, would be uncovered with a systematic examination, especially if conducted at different scales. Although ecological properties such as soil type can change across boundaries (Cole 1992), it cannot be assumed that abrupt changes in grassland–savanna boundaries will necessarily be reflected in abiotic variables. This is because ecological
boundaries may also be the product of nonlinear behavior, whereby gradual changes in environmental variables elicit dramatic changes in population and community variables when thresholds are reached (Fagan et al. 2003). Boundaries may also have properties that are unique to the boundary (Fagan et al. 2003), thereby complicating the isolation of ultimate factors. As an example, the woody species Maesa lanceolata occurs mostly at the boundary between Afromontane forests and grasslands but is largely absent from either of the two formations. We note that there are likely to be multiple causes for changes in vegetation structure across boundaries and consequently advocate that research is conducted on numerous fronts using numerous hypotheses. Adopting an adaptive inference approach with multiple lines of reasoning. Grassland and savanna research has focused predominantly on experimentation, falsification, and avoidance of type I error, and has tended to generate inconclusive bivariate empirical studies and factorial experiments (Weiher et al. 2004). The main drawback of this use of Popperian philosophy is that, because it does not deal well with contingency and multiple causality, it results in type II errors (false negatives), shutting down promising avenues of research. We suggest that adaptive inference (Holling and Allen 2002) is a more suitable approach for trying to separate the proximate and ultimate factors shaping vegetation structure and for analyzing systems with complex (nonlinear) dynamics. Adaptive inference draws on both experimental and correlative data.
Figure 3. Changes in ecological processes or factors across a grassland–savanna boundary. Abbreviations: cm, centimeter. www.biosciencemag.org
July 2006 / Vol. 56 No. 7 • BioScience 585
Articles
Figure 4. Factors that may lag, keep abreast of, or move ahead of dynamic grassland–savanna boundaries and thereby provide clues for separating proximate from ultimate processes governing biome distributions. It specifically tries to avoid type II errors in the early phases of advancing understanding by keeping numerous hypotheses alive and actively seeking confirmatory data for each hypothesis. Research at this early stage would focus on building cases for each hypothesis, rather than trying to refute them. Given that processes are likely to vary considerably at different scales of analysis and experimentation, the rejection of any hypothesis in the early stages of investigation is probably unwarranted. The investigation of numerous lines of reasoning at different scales cannot be achieved by an individual scientist, and therefore requires a collective and focused effort by the appropriate ecological community (Pickett et al. 1994). Adaptive inference has been shown to be effective (Levin 1998, Holling and Allen 2002) because, as understanding improves and some lines of reasoning are slowly winnowed out, the most parsimonious explanations are identified. Ecologists can then set quantitatively more specific, more precise, and narrower tests of components of the lines of reasoning. This in turn encourages a switch to more deductive tests (e.g., manipulative or controlled experiments, null hypothesis testing) that avoid type I errors.
New foci for grassland–savanna research McNaughton (1983) noted that the composition of vegetation is probably governed not by a single overpowering force but rather by constellations of weak forces. Following this 586 BioScience • July 2006 / Vol. 56 No. 7
approach, the outcome of grass–tree competition may be a function of numerous interacting factors, such as fire, soil water content, availability of energy (sunlight), nutrients, frost, and microbial activity. Although this constellation of forces may operate in complex ways and at different scales (see, e.g., factors affecting nitrogen mineralization shown in figure 2), the art of investigating such complexity may lie in the identification of those processes and agents that play the greatest role in shaping the system (i.e., the ultimate factors). To this end, we suggest several new research foci in the paragraphs below. First, the role of fire in the exclusion of trees from grasslands warrants critical analysis. Data on fire intensity and frequency gradients across grassland–savanna boundaries are likely to be instructive, as are data on the relationships between fire intensity and survival of trees at different life history stages at the boundary (figure 5). Interactions between fire and grass–tree competition are largely unexplored. For example, the increased nutrient availability after fire as a result of incorporation of ash into topsoils (Stock and Lewis 1986) and increased rates of mineralization (Knapp et al. 1998) could increase grass vigor and enable the dominance of grasses over tree seedlings, whether fire damages the tree seedlings or not. Experiments that shield tree seedlings from fire, yet allow adjacent grass plants to burn, are a possible way of teasing out such interactions. Furthermore, considerably more information on the effects of herbaceous vegetation on woody www.biosciencemag.org
Articles plant growth, recruitment, seed production, and seedling establishment is required to separate the effects of competition on grassland and savanna structure from those of disturbance (House et al. 2003). Second, determining the effects of climate on nutrient availability in topsoils (Hooper and Johnson 1999), and the implications of these effects for grass–tree competition, could be revealing. Savanna models have tended to focus on water rather than nutrients as the main resource limiting growth (Sankaran et al. 2004). The importance of nutrient availability with respect to grass–tree interactions and vegetation structure is intuitive yet largely unexamined (Walker and Langridge 1997, House et al. 2003). Recent experimental work shows that nitrogen fertilization in semiarid savannas can result in an increase in grass vigor and reduced tree seedling survival (David Ward, School of Biological and Conservation Sciences, University of KwaZulu-Natal, Scottsville, South Africa, personal communication, 20 December 2005). This suggests that grasses Figure 5. A dynamic grassland–savanna boundary near Barbermay dominate in landscapes where the supply of nutrients ton, Mpumalanga, South Africa. Fire frequency is likely to be from mineralization is sustained above a certain thresh- similar across the boundary, although differences in grass species old during periods of plant growth. Mineralization is con- flammability may influence fire intensity at a microscale and trolled by numerous factors operating at different scales thereby affect tree seedling establishment. Photograph: Marc (figure 2). At the macro or biome scale, the desiccation of Stalmans. topsoils, which occurs frequently during the growing season in South African savannas, may reduce grass com(Mills and Fey 2004), tends to have more soil organic matter petitiveness not only through reduced water availability but than the deeper layers within the conventionally defined A also through a reduced nutrient supply from mineralization. horizon, and consequently mineralization is often disproIf the dominance of grasses over trees is largely dependent on portionately greater in this surface layer than below it (Woods a sustained supply of nutrients, this desiccation, or “switch1989, Purnomo et al. 2000). Seasonal and daily fluctuations ing off the nutrient pump,” may be a key factor enabling in the water content of the pedoderm (such as occur under trees to get a foothold in savanna systems. conditions of mist and dew) are likely to influence the availThird, the jury is still out as to whether fire and herbivory ability of nutrients, and therefore to affect grass–tree comare modifiers (Stott 1991, Sankaran et al. 2004) or primary depetition. A pedoderm that seldom dries out during the terminants (Bell 1982, Higgins et al. 2000, Bond et al. 2005) growing season, for example, may be necessary for grasses to of savanna structure. Even extremely intense fire regimes or dominate tree seedlings. Ellery and colleagues (1991) noted browsing pressure have, to our knowledge, never converted that mean temperature during the growing season is genera savanna into a treeless grassland (Laws 1970). This suggests ally lower in South African grasslands than in savannas. Does that these factors are modifiers rather than drivers of the this climatic pattern influence biome structure through a system. But is the same true at the boundary? It is conceivconstellation of interacting factors, such as the water content able that manipulation of fire or herbivory is more likely to of the pedoderm, rates of soil organic matter mineralizaexclude trees at grassland–savanna boundaries than within the tion, and rates of photosynthesis in grass leaves? Our undersavanna system itself. The vulnerability of trees to herbivores standing of grass–tree interactions is likely to improve if and fire may, for example, be a function of distance from the research paradigms are shifted to a much finer vertical resoboundary (Palmer et al. 2003), and may correlate with spelution when examining the uppermost part of soil profiles. cific environmental factors such as the nutrient supply from We hope that housing these different research foci within topsoils or the distribution of water in soil profiles. the proposed framework of a scaled exploration at the boundFinally, the vertical distribution of water in the soil profile ary, with an emphasis on adaptive inference, will provide has been invoked as a key determinant of grass–tree ratios in new vigor and direction for researchers tackling the vexing savannas, with tree biomass increasing as the amount of subsavanna–grassland problem. soil water increases (Walter 1971, Walker and Noy-Meir The challenge concluded 1982). The focus of these classic equilibrium models of saIn this article we have explored the many theories of grass–tree vannas has been on soil water content in topsoils relative to interaction that have been invoked to explain the differences subsoils. Effects of water content on vegetation structure in between grasslands and savannas. We have also introduced grasslands and savannas may, however, also play out at another some new ideas, and different ways of thinking about old ideas, scale. The top few centimeters of mineral soil, or pedoderm www.biosciencemag.org
July 2006 / Vol. 56 No. 7 • BioScience 587
Articles into the debate. The list of ideas and theories can be summarized as follows: One system is favored over another as a result of (a) the present suite of abiotic conditions; (b) abiotic history; (c) chance-driven evolutionary history; (d) a particular regime of fire, frost, herbivory, nutrients, or water; (e) contingent interactions between some or all of these regimes; and (f) complex adaptive processes involving some or all of the above. Each of these lines of reasoning has merit as a working hypothesis, model, or theoretical construct, but our current understanding does not provide a useful basis for accepting one over the other, or for choosing one or a combination as the most profitable avenue of scientific exploration. There are three main reasons for the limitations in current understanding of the savanna–grassland problem: First, previous studies of the problem have been conducted within savanna or grassland systems, making it almost impossible to separate ultimate drivers from proximate modifiers in the exploration of causation. Second, the predominance of a falsification paradigm behind the design of previous studies has led to premature debates over which single factor is the dominant cause of the differences between grasslands and savannas. Third, the lack of attention to spatial and temporal scale, particularly evident in experimental and falsification approaches to science, has lead to confusion between cause and effect. The challenge ahead is therefore to acknowledge the complexity of the practical, analytical, and paradigmatic problems faced when exploring ecosystem-level questions and to adopt a fresh approach. We recommend that a community of scientists from different disciplines explore and winnow the working models by (a) adopting an adaptive inference paradigm of scientific exploration; (b) applying analytical approaches that allow type I and type II errors at appropriate junctures of the scientific process; and (c) focusing scientific effort at the grassland–savanna boundary, where the consequences of scale, history, contingency, and adaptive feedback are most likely to be unraveled. Such a scaled, adaptive, and interdisciplinary approach will provide a challenge for future generations of ecologists and collaborating scientists.
Acknowledgments The authors extend their grateful thanks to the Mellon Foundation for funding the research.
References cited Ackerley DD. 2004. Adaptation, niche conservatism, and convergence: Comparative studies of leaf evolution in the Californian chaparral. American Naturalist 163: 654–671. Acocks JPH. 1953. Veld types of South Africa. Memoirs of the Botanical Survey of South Africa 28: 1–192. Bell RHV. 1982. The effect of soil nutrient availability on community structure in African ecosystems. Pages 193–216 in Huntley BJ, Walker BH, eds. Ecology of Tropical Savannas. Berlin: Springer-Verlag. Blair JM. 1997. Fire, N availability, and plant response in grasslands: A test of the transient maxima hypothesis. Ecology 78: 2359–2368.
588 BioScience • July 2006 / Vol. 56 No. 7
Bond WJ, Keeley JE. 2005. Fire as a global “herbivore”: The ecology and evolution of flammable ecosystems. Trends in Ecology and Evolution 20: 385–394. Bond WJ, Midgley GF, Woodward FI. 2003. What controls South African vegetation—climate or fire? South African Journal of Botany 2003: 79–91. Bond WJ, Woodward FI, Midgley GF. 2005. The global distribution of ecosystems in a world without fire. New Phytologist 165: 525–538. Bowman DMJS, Cook GD, Zoppi U. 2004. Holocene boundary dynamics of a northern Australian monsoon rainforest patch inferred from isotopic analysis of carbon (14C and ␦13C) and nitrogen (␦15N) in soil organic matter. Austral Ecology 29: 605–612. Brown JH. 1995. Macroecology. Chicago: University of Chicago Press. Cadenasso ML, Pickett STA, Weathers KC, Bell SS, Benning TL, Carreiro MM, Dawson TE. 2003. An interdisciplinary and synthetic approach to ecological boundaries. BioScience 53: 717–722. Cody ML, Mooney HA. 1978. Convergence versus non-convergence in Mediterranean-climate ecosystems. Annual Review of Ecology and Systematics 9: 265–321. Coetzee JP, Bredenkamp GJ, van Rooyen N. 1994. An overview of the physical environment and vegetation units of the Ba and Ib land types of the Pretoria–Witbank–Heidelberg area. South African Journal of Botany 60: 49–61. Cole MM. 1992. Influence of physical factors on the nature and dynamics of forest–savanna boundaries. Pages 63–75 in Furley PA, Proctor J, Ratter JA, eds. Nature and Dynamics of Forest–Savanna Boundaries. London: Chapman and Hall. Cowling RM, Witkowski ETF. 1994. Convergence and non-convergence of plant traits in climatically and edaphically matched sites in Mediterranean Australia and South Africa. Australian Journal of Ecology 19: 220–232. Ellery WN, Scholes RJ, Mentis MT. 1991. An initial approach to predicting the sensitivity of the South African grassland biome to climate change. South African Journal of Science 87: 499–503. Fagan WF, Fortin M-J, Soykan C. 2003. Integrating edge detection and dynamic modeling in quantitative analyses of ecological boundaries. BioScience 53: 730–738. Feely JM. 1987. The Early Farmers of Transkei, Southern Africa, before A.D. 1870. Oxford (United Kingdom): British Archaeological Reports. Cambridge Monographs in African Archaeology 24. British Archaeological Reports International Series 378. Gillson L. 2004. Evidence of hierarchical patch dynamics in an East African savanna? Landscape Ecology 19: 883–894. Gosz JR, Sharpe PJH. 1989. Broad-scale concepts for interactions of climate, topography, and biota at biome transitions. Landscape Ecology 3: 229–243. Higgins SI, Bond W, Trollope WSW. 2000. Fire, resprouting and variability: A recipe for grass–tree coexistence in savanna. Journal of Ecology 88: 213–229. Holling CS, Allen CR. 2002. Adaptive inference for distinguishing credible from incredible patterns in nature. Ecosystems 5: 319–328. Hooper DU, Johnson L. 1999. Nitrogen limitation in dryland ecosystems: Responses to geographical and temporal variation in precipitation. Biogeochemistry 46: 247–293. House JI, Archer S, Breshears DD, Scholes RJ. 2003. Conundrums in mixed woody–herbaceous plant systems. Journal of Biogeography 30: 1763–1777. Jacob F. 1977. Evolution and tinkering. Science 196: 1161–1166. Klein RG. 1984. The large mammals of southern Africa: Late Pliocene to recent. Pages 107–148 in Klein RG, ed. Southern African Prehistory and Paleoenvironments. Rotterdam (The Netherlands): A. A. Balkema. Knapp AK, Conrad SL, Blair JM. 1998. Determinants of soil CO2 flux from a sub-humid grassland: Effect of fire and fire history. Ecological Applications 8: 760–770. Knoop WT, Walker BH. 1985. Interactions of woody and herbaceous vegetation in a southern African savanna. Journal of Ecology 73: 235–253. Laws RM. 1970. Elephants as agents of habitat and landscape change in East Africa. Oikos 21: 1–15. Levin SA. 1998. Ecosystems and the biosphere as complex adaptive systems. Ecosystems 1: 431–436.
www.biosciencemag.org
Articles McKenzie B. 1989. Medium-term changes of vegetation pattern in Transkei. South African Forestry Journal 150: 1–6. McNaughton SJ. 1983. Serengeti grassland ecology: The role of composite environmental factors and contingency in community organization. Ecological Monographs 53: 291–320. Meadows ME, Linder HP. 1993. A palaeoecological perspective on the origin of Afromontane grasslands. Journal of Biogeography 20: 345–355. Mills AJ. 2003. Reciprocal relationships between vegetation structure and soil properties in selected biomes of South Africa. PhD dissertation. University of Stellenbosch, Bellville, South Africa. Mills AJ, Fey MV. 2004. Frequent fires intensify soil crusting: Physicochemical feedback in the pedoderm of long-term burn experiments in South Africa. Geoderma 121: 45–64. O’Connor TG, Bredenkamp GJ. 1997. Grassland. Pages 215–257 in Cowling RM, Richardson DM, Pierce SM, eds. Vegetation of Southern Africa. Cambridge (United Kingdom): Cambridge University Press. Orians GH, Solbrig OT. 1977. Convergent Evolution in Warm Deserts. Stroudsburg (PA): Dowden, Hutchinson and Ross. Palmer AR, Novellie PA, Lloyd JW. 1999. Community patterns and dynamics. Pages 208–223 in Dean WRJ, Milton SJ, eds. The Karoo: Ecological Patterns and Processes. Cambridge (United Kingdom): Cambridge University Press. Palmer SCF, Hester AJ, Elston DA, Gordon IJ, Hartley SE. 2003. The perils of having tasty neighbors: Grazing impacts of large herbivores at vegetation boundaries. Ecology 84: 2877–2890. Peters DPC. 2002. Plant species dominance at a grassland–shrubland ecotone: An individual-based gap dynamics model of herbaceous and woody species. Ecological Modelling 152: 5–32. Pickett STA, Kolasa J, Jones CG. 1994. Ecological Understanding: The Nature of Theory and the Theory of Nature. San Diego: Academic Press. Pickett STA, Burch WR, Grove JM. 1999. Interdisciplinary research: Maintaining the constructive impulse in a culture of criticism. Ecosystems 2: 302–307. Purnomo E, Black AS, Smith CJ, Conyers MK. 2000. The distribution of net nitrogen mineralisation within surface soil, 1: Field studies under a wheat crop. Australian Journal of Soil Research 38: 643–652. Rapport A. 1985. Thinking about home environments: A conceptual framework. Pages 255–286 in Altman I, Werner CM, eds. Home Environments. New York: Plenum Press. Reich RB, Abrams MD, Ellsworth DS, Kruger EL, Tabone TJ. 1990. Fire affects ecophysiology and community dynamics of central Wisconsin oak forest regeneration. Ecology 71: 2179–2190. Richardson DM, MacDonald IAW, Holmes PM, Cowling RM. 1992. Plant and animal invasions. Pages 271–308 in Cowling RM, ed. The Ecology of Fynbos: Nutrients, Fire and Diversity. Cape Town (South Africa): Oxford University Press.
www.biosciencemag.org
Rundel PW. 1981. Fire as an ecological factor. Pages 501–538 in Lange LO, Nobel PS, Osmond CB, Ziegler H, eds. Physiological Plant Ecology I: Responses to the Physical Environment. Berlin: Springer-Verlag. Sankaran M, Ratnam J, Hanan NP. 2004. Tree–grass coexistence in savannas revisited—insights from an examination of assumptions and mechanisms invoked in existing models. Ecology Letters 7: 1–11. Sarmiento G. 1984. The ecology of neotropical savannas. Cambridge (MA): Harvard University Press. Stock WD, Lewis OAM. 1986. Soil nitrogen and the role of fire as a mineralizing agent in a South African coastal fynbos ecosystem. Journal of Ecology 74: 317–328. Stott P. 1991. Recent trends in the ecology and management of the world’s savanna formations. Progress in Physical Geography 15: 18–28. Tainton NM, Walker BH. 1993. Grasslands of southern Africa. Pages 265–290 in Coupland RT, ed. Natural Grasslands: Eastern Hemisphere and Résumé. Amsterdam: Elsevier. Tinley KL. 1982. The influence of soil moisture balance of ecosystem patterns in southern Africa. Pages 175–192 in Huntley BJ, Walker BH, eds. Ecology of Tropical Savannas. Berlin: Springer-Verlag. Van Wilgen BW, Govender N, Biggs HC, Ntsala D, Funda XN. 2004. Response of savanna fire regimes to changing fire-management policies in a large African national park. Conservation Biology 18: 1533–1540. Vlok JHJ, Euston-Brown DIW, Cowling RM. 2003. Acocks’ Valley Bushveld 50 years on: New perspective on the delimination, characterisation and origin of subtropical thicket vegetation. South African Journal of Botany 69: 27–51. Walker BH, Langridge JL. 1997. Predicting savanna vegetation structure on the basis of plant available moisture (PAM) and plant available nutrients (PAN): A case study from Australia. Journal of Biogeography 24: 813–825. Walker BH, Noy-Meir I. 1982. Aspects of stability and resilience in savanna ecosystems. Pages 556–590 in Huntley BJ, Walker BH, eds. Ecology of Tropical Savannas. Berlin: Springer-Verlag. Walter H. 1971. Ecology of Tropical and Subtropical Vegetation. Edinburgh (Scotland): Oliver and Boyd. Weiher E, Forbes S, Schauwecker T, Grace JB. 2004. Multivariate control of plant species richness and community biomass in blackland prairie. Oikos 106: 151–157. Werger MJA, Coetzee BJ. 1978. The Sudano–Zambezian region. Pages 301–462 in Werger MJA, ed. Biogeography and Ecology of Southern Africa. The Hague: Junk. Woods LE. 1989. Active organic matter distribution in the surface 15 cm of undisturbed and cultivated soil. Biology and Fertility of Soils 8: 271–278. Wu J, Loucks OL. 1996. From balance of nature to hierarchical patch dynamics: A paradigm shift in ecology. Quarterly Review of Biology 70: 439–466.
July 2006 / Vol. 56 No. 7 • BioScience 589
BACK ISSUES OF BIOSCIENCE NOW AVAILABLE ONLINE WITH JSTOR FOR INSTITUTIONAL AND INDIVIDUAL SUBSCRIBERS!
Back issues of BioScience are now available online for institutional and individual subscribers through JSTOR, the not-for-profit, online digital archive. Users can browse, search, view, download, and print the fulltext PDF files of BioScience from its original inception as the AIBS Bulletin in 1951 up to the most recent five years. AIBS maintains a five-year moving wall between BioScience content archived with JSTOR and more recent BioScience content. Institutions that participate in JSTOR’s new Biological Sciences Collection can access back issues of BioScience at www.jstor.org. The Biological Sciences Collection will include at least 100 titles when it is completed at the end of 2007, including the 29 journals in JSTOR’s existing Ecology and Botany Collection. The journals in this collection offer greater depth in fields such as biodiversity, conservation, paleontology, and plant sciences, in addition to introducing new areas such as cell biology and zoology. Individual members of AIBS can access BioScience back issues through JSTOR at www.aibs.org/bioscience, where both individuals and institutions can also access the most recent five years of BioScience online. AIBS and BioScience are proud to collaborate with JSTOR to preserve and make widely available the historic literature of our field.
