Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice Bioscience . 2026 Feb 28;76(4):385–396. doi: 10.1093/biosci/biag008 Search in PMC Search in PubMed View in NLM Catalog Add to search Spatiotemporal variation in top-kill agents relative to ontogenetic resprouting strategies in African savannas Fezile Mtsetfwa Fezile Mtsetfwa 1 School of Animal Plant & Environmental Sciences, University of the Witwatersrand, 1 Jan Smuts Ave, Braamfontein 2000, Johannesburg, South Africa Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing - original draft Find articles by Fezile Mtsetfwa 1, ✉ , Sally Archibald Sally Archibald 2 School of Animal Plant & Environmental Sciences, University of the Witwatersrand, 1 Jan Smuts Ave, Braamfontein 2000, Johannesburg, South Africa Conceptualization, Funding acquisition, Investigation, Supervision, Writing - review & editing Find articles by Sally Archibald 2 Author information Article notes Copyright and License information 1 School of Animal Plant & Environmental Sciences, University of the Witwatersrand, 1 Jan Smuts Ave, Braamfontein 2000, Johannesburg, South Africa 2 School of Animal Plant & Environmental Sciences, University of the Witwatersrand, 1 Jan Smuts Ave, Braamfontein 2000, Johannesburg, South Africa ✉ Corresponding author. Email: [email protected] Roles Fezile Mtsetfwa : Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing - original draft Sally Archibald : Conceptualization, Funding acquisition, Investigation, Supervision, Writing - review & editing Received 2024 Feb 16; Revised 2025 Nov 20; Accepted 2026 Jan 13; Collection date 2026 Apr. © The Author(s) 2026. Published by Oxford University Press on behalf of the American Institute of Biological Sciences. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License ( https://creativecommons.org/licenses/by-nc/4.0/ ), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact [email protected] PMC Copyright notice PMCID: PMC13069568 PMID: 41971257 Abstract Fire is a key control on woody plant size and abundance in semiarid tropics. Other disturbances, including cyclones, elephants, drought, frost, and tree harvesting also damage woody structures, causing top-kill or full mortality, depending on species’ resprouting abilities and disturbance type. We outline expectations for resprouting adaptations in tropical African woody plant communities by integrating spatiotemporal information on disturbances with ecological impacts. Frost and fire affect smaller trees (less than 4 meters tall), cyclones and elephants affect medium to large trees, and human impacts are diverse. Historically, most of sub-Saharan African savanna trees evolved with frequent fires and high elephant densities, with implications for their resprouting patterns. Cyclone-induced toppling seldom coincides spatially with frost or drought. A considerable part of the region has lost a key top-kill agent, elephants, whereas human impacts have both spread and increased. However, in approximately 21% of the areas that lost elephants, this important ecological agent has not been replaced. Keywords: disturbances, frost, large trees, savanna, top-kill agents Disturbances that restrict aboveground woody biomass are a component of almost all ecosystems—more so in the past, when large megaherbivores globally dominated the herbivore biomass (Gill 2014 ). Tropical ecosystems are exposed to a range of disturbance agents, whose form and intensity vary across environmental gradients, leading to alterations in vegetation structure and species filtering. Tree species adopt different strategies for regeneration, such as resprouting from existing structures or reproducing via seed (Bellingham and Sparrow 2000 ). Resprouting enables (often rapid) recovery from biomass loss, allowing individuals to persist, grow, and thrive between disturbance events—known as the persistence niche (Bond and Midgley 2001 ). Seed-based reproduction, in addition to allowing regeneration after a disturbance, enhances dispersal potential, facilitates colonization, and promotes genetic diversity. Although both resprouting and seed-based reproductive strategies promote persistence, they often trade off (Bond and Midgley 2003 , Clarke et al. 2013 ), and we lack a clear understanding of the disturbance conditions that favor one strategy over the other. Moreover, whether a tree resprouts is not a binary response but varies with environmental conditions, and this variation has yet to be thoroughly explored in savannas where disturbances are common and resprouting prolific (Bond and Midgley 2001 , Pausas et al. 2016 ). Not all plant species resprout equally vigorously, and resprouting also varies with plant size and age (Bond and Midgley 2001 , Pausas et al. 2016 ). Although resprouting has been recognized as a globally important trait in woody species (Pausas et al. 2016 ), its importance has not permeated general ecological theory, and it has been overlooked in the parameterization of vegetation models related to recruitment dynamics (Fricke et al. 2022 ). Loss of aboveground biomass is often mislabeled as mortality , but the term top-kill is more appropriate when woody individuals do not die but regrow from the same stem or a new basal shoot (Hoffmann et al. 2019 , De Antonio et al. 2021 ). Although disturbances have large and immediate effects on aboveground biomass of woody species, their longer-term impact depends on the adaptations of the species involved, which, in turn, should depend on the form, intensity, and frequency of disturbances in their environment over evolutionary time (Pausas and Keeley 2014 , Lamont et al. 2019 ). Fire and herbivory are often conceived as constraining the recruitment of seedlings and saplings to adult sizes (Wakeling et al. 2012 , Jurisch et al. 2013 , Shen et al. 2014 ), but top-kill is widely and sometimes disproportionately experienced by adult-size trees. For example, drought is as likely to top-kill large trees as it does seedlings and saplings in some species (Case et al. 2019 ), megaherbivores can preferentially top-kill medium size trees of between 10 and 40 centimeters (cm) diameter at breast height (DBH; Gadd 2002 ), or 2–4 meters in height Cardoso et al. 2020 ), whereas category 5 cyclones are more likely to snap and uproot large trees (Box 1 ; Ibanez et al. 2019 , Delaporte et al. 2022 ). Humans select medium to large trees, depending on whether they are cutting wood for subsistence or commercial purposes (Luoga et al. 2004 , Neke et al. 2006 ) Box 1. Mechanisms of top-kill and the selectivity of their impact. Cyclones: Maximum sustained wind speeds of at least 70 meters per second have the power to snap and uproot shrubs and large trees (Simpson 1974 , Delaporte et al. 2022 ). Larger trees (more than 20 cm diameter) are more likely to be affected both because they have more exposure to the wind, and because their above to belowground biomass ratios might be greater (Brade 2022 ). Certain species are also more susceptible, presumably because of different rooting patterns (Brade 2022 ). Megaherbivore density: Top-kill by megaherbivores is life stage specific, disproportionately affecting medium-size trees (10–40 cm DBH), and there is strong species selectivity that results in differential damage (Cardoso et al. 2020 , Thompson et al. 2022 ). Trees are often top-killed by snapping (breaking the trunk) or uprooting, but larger trees can have their bark stripped off, making them more vulnerable to effects of other top-kill agents (Gadd 2002 , Asner and Levick 2012 , Cardoso et al. 2020 , Thompson et al. 2022 ). Fire: Top-kill by fire can occur via death of the cambium, where the vascular transport system of the plant is damaged beyond repair (Midgley et al. 2011 , Hood et al. 2018 , Hoffmann et al. 2021 ). Trees with protective thick bark can resist the fire’s damaging effects (Hoffmann and Solbrig 2003 ), but alternative tolerance strategies can involve adaptive underground organs to enable resprouting after top-kill (Veldman et al. 2015 , Harrison et al. 2021 ). Many savanna trees tend to escape the fire trap once they have grown above approximately 3 meters in height (Hoffmann and Solbrig 2003 , N’Dri et al. 2022 ), but some thin-barked species remain susceptible even when very large. Frost: Frost damage is experienced when minimum temperatures drop below freezing and exhibits a clear disturbance trap that disproportionately affects smaller trees (Holdo 2006 , Whitecross et al. 2012 , Hoffmann et al. 2019 ). Globally, many trees are resistant to frost damage, but most tropical and subtropical tree species are susceptible to top-kill, and lethal frost temperatures for tropical tree species can be as high as −2 degrees Celsius (Hoffmann et al. 2019 , De Antonio et al. 2021 ). For example, once Colophospermum mopane trees have grown taller than approximately 4 meters, they are usually no longer susceptible to frost (Whitecross et al. 2012 ). Drought: The mechanisms of drought death fall into two main categories: carbon starvation and hydraulic failure due to cavitation and loss of hydraulic capacity (Sankaran 2019 ). Top-kill by drought (Zeppel et al. 2015 , Pausas et al. 2016 ) is skewed toward the smallest and largest size classes because younger trees have less well-developed root and vascular systems, whereas larger trees are susceptible to hydraulic failure because of high water demands and to carbon starvation because of high metabolic demand (Case et al. 2019 , Sankaran 2019 ). Differences in water use strategies make some species more susceptible to drought than others (Case et al. 2019 , González-M et al. 2021 , Oliveira et al. 2021 ). Harvesting: Fuel wood harvesters in communal areas will target stems as small as 1 cm DBH, although the preferred size range is 4–20 cm DBH (Okello et al. 2001 , Neke et al. 2006 ) and is skewed toward the larger size (i.e., between 10 and 20 cm for charcoal and building materials; Okello et al. 2001 , Luoga et al. 2004 ). On the other hand, commercial harvesters in sub-Saharan Africa have preferences for much larger stem sizes, with minimum harvesting diameters of approximately 35–40 cm (Therrell et al. 2007 ). As with elephants, some woody species are preferred and others avoided, meaning that the impacts are not uniform in a tree community. Today, humans are one of the most common agents of top-kill and true mortality for woody plants, and this is considered a major global cause of degradation (Ellis and Ramankutty 2008 , Sandel and Svenning 2013 ). The iron age created the need for both wood for smelting and the tools to chop down trees (House and Bamford 2019 ). This impact has intensified with increasing human populations fueling demand for fuel wood, charcoal, and timber. To understand the ecological consequences, it is important to assess how current levels of human wood harvesting differ from the top-kill agents that trees would have evolved with, and to do this requires linking spatial data sets with information on demographic impacts. Evolutionary selection for resprouting Top-kill and resprouting are common and well described in sub-Saharan fire-prone ecosystems (Bond and Midgley 2001 ), but resprouting has only recently been recognized as a globally important plant trait (Salguero-Gomez et al. 2012 , Pausas et al. 2016 ). It has been proposed that the ability to resprout may come with a carbon cost of maintaining active meristems by the plant and, possibly, the difficulty of maintaining without degradation the DNA for the many decades to centuries that trees can live (Clarke et al. 2013 ). Taken together with the idea that many species can outgrow their top-kill disturbances, this potentially explains the observed decline in resprouting ability as trees age and explains why it is most prolific in younger age classes (Bond and Midgley 2001 , Neke et al. 2006 , Handavu et al. 2011 ). Conversely, plants need to accumulate or maintain stored resources to be able to resprout successfully, so there might also be a lower age or size limit to resprouting (Cardoso et al. 2016 , Archibald et al. 2021 ). The fact that resprouting does not evolve in many woody species and may even be secondarily lost in environments where disturbances are not common suggests that it occurs at some cost to the plant (Bellingham and Sparrow 2000 , Lamont et al. 2019 ). In addition, many woody species can resprout in more than one way (basal, epicormic, or root suckering), particularly toward more arid regions (Clarke et al. 2013 ), with the resprouting mechanism often determined by the severity of the disturbance (Bellingham and Sparrow 2000 ). Therefore, one might expect that the mechanism of resprouting and the size class where plants resprout would vary depending on the disturbances experienced and that organisms would tend to evolve and retain this ability during demographic stages where top-kill is likely. In the present article, we synthesize spatiotemporal information on top-kill agents with ecological information on the demographic stages most affected to set up expectations on the resprouting adaptations found in woody plant communities across tropical Africa. We use available literature to quantify the average size classes of trees that are susceptible to top-kill by fire, drought, frost, cyclones, and megaherbivory (elephants). We map each top-kill agent and the spatial overlap between them to infer what resprouting responses would be most appropriate in different parts of the continent. We then contrast a past and current top-kill regime, dominated by elephants and human impacts respectively to gain insight on what levels of human mediated top-kill are comparable to the top-kill pressures that trees evolved with. Box 2. Characterizing top-kill probabilities. From an individual tree perspective, the ecological impact of a disturbance is a combination of the probability that it will occur, its intensity when it does occur and the resistance or resilience to the disturbance. The impact of abiotic disturbances like fire, frost, drought, and cyclones can be quantified with two metrics: the return time and intensity (i.e., different intensities of events occur at different frequencies). Seasonal (3–8 month) droughts, for example, occur annually across most tropical savannas, but extreme multiannual droughts occur at frequencies of 7–20 years (Khan and Gilani 2021 ). Similarly, although cyclones themselves are common, high intensity conditions with the power to top-kill trees are relatively infrequent on the African continent (in the southwest Pacific, winds up to category 4 occurred once in 20 years and affected at least 22% of the cyclone region; Delaporte et al. 2022 ). On the other hand, about 50% of sub-Saharan savanna areas experience a burn at least once a decade, with fire return intervals ranging from annual to up to 20 years (Archibald et al. 2010 , 2013 , Archibald and Hempson 2016 , Harrison et al. 2021 ). Sites with shorter fire return intervals typically experience less intense fires (Archibald et al. 2013 ). Frost regimes are not well characterized in the tropics and subtropics (Franco and Álvarez-Yépiz 2022 ), but both the intensity and the frequency of frost varies widely and with landscape position (Holdo 2005 , 2006 , Muller et al. 2016 , Hoffmann et al. 2019 , De Antonio et al. 2021 ). Unlike episodic events like droughts, biotic disturbances like elephant and human damage are chronic but stochastic events from a perspective of a tree: The overall rate of harvesting or damage in a landscape will predict the probability that an individual tree will be damaged. Data indicate that elephant damage is correlated with elephant density (Guldemond and Van Aarde 2008 ) but is even more strongly predicted by elephant habitat preference (Cardoso et al. 2020 , Abraham et al. 2021 ), and similarly, woody tree harvesting is more common in places with higher human population density and close to towns and roads (Bailis et al. 2015 ). Plant adaptations can alter the severity of all events through traits that confer resistance or avoidance to damage (e.g., bud protection from frost, thick bark protecting trees from fire, and increased wood density from being broken by elephants or cyclones; De Antonio et al. 2021 , Scalon et al. 2021 , Dantas and Pausas 2022 ). In the case of elephants and humans, plant traits can also alter the probability of events by determining whether the plant will be selected in the first place (Archibald et al. 2019 ). Variation in the probability of top-kill across tree life stages The risk of top-kill varies over a tree’s lifespan, both because of variations in its ability to withstand damage and because of variations in the intensity and severity of the disturbance at different life stages. We identified estimates of minimum size, target size, and escape size for each disturbance agent from the literature and used this to establish a mean functional form of the probability of top-kill relative to stem diameter for each disturbance (figure 1 , table 1 ). Cyclone damage increases with size to plateau at a high level for plants larger than approximately 10 cm DBH. This lower damage threshold is similar for elephants, but they also tend to avoid extremely large (more than 40 cm DBH) trees. Fire damage, in contrast, tends to drop off after trees exceed 5 cm DBH, although interactions with termites and elephants can increase fire mortality for very large trees (Pringle et al. 2015 , Woolley et al. 2018 ). Frost, like fire, is most destructive to smaller trees, and drought damage can affect all plants but is at a minimum for intermediate-size trees. Human wood harvesting covers a very large range of size classes but seldom affects trees smaller than 4 cm DBH (figure 1 ). Figure 1. Open in a new tab Summary from information in the literature of how the probability of top-kill by six dominant top-kill agents changes along the life history of a woody savanna plant. The arrows indicate reported cutoff sizes for targeted tree size (as diameter at breast height [DBH] in centimeters [cm]). Table 1. Top-kill target and escape heights extracted from a systematic literature review for savanna trees in Africa. Minimum target size Escape size Top-kill agent Type Occurrence Species susceptibility Height (in meters) Stem DBH (in cm) Height (in meters) Stem DBH (in cm) Fire Episodic Flame heights and energy released control intensity—higher closer to the ground Differential damage due to varying fire-resistant traits such as thick bark 0 0 More than 3 5 Megaherbivore Chronic (selective) Occurrence depends on density Differential damage due to elephant preference 3 6 9 37 Cyclone Episodic Intense events rare Differential damage due to rooting patterns 4.1 10 None, damage increases exponentially with height/stem diameter Drought Episodic Quantifying drought usually relative to prevailing water availability Differential damage due to different physiological strategies 0 to 5 Frost Episodic Frost most extreme closest to the ground usually Very little known about different frost tolerance—generally frost is thought to be lethal to small tropical trees exposed in the “frost trap” 0 0 More than 4 more than 10 Human: subsistence Chronic (selective) Differential damage due to human preference, but selectivity goes down when demand increases. 2.33–8.75 4–35 8.75 35 Human: commercial Chronic (selective) Differential damage due to human preference 8.75–9.50 35 None, timber becomes more valuable, unless large trees become hollow Open in a new tab Note: A list of references for the literature reviewed on each top-kill agent is provided in supplemental table S1 . Where values were initially reported as diameter at breast height (DBH), in centimeters (cm), we have converted them to height (in meters) and rounded to the nearest meter, following Moncrieff and colleagues ( 2014 ). Mapping disturbances and top-kill Fire, megaherbivores, drought, and human wood harvesting have expansive ranges but are heterogenous in their intensities across the landscapes they affect (figure 2 ; see box 1 and the methods in the supplemental material for details). Reconstructed historical megaherbivore density was low in both arid and humid environments (Hempson et al. 2015 ). In environments between approximately 600 and approximately 1200 millimeters of annual rainfall, however, elephant density could often have reached three individuals per square kilometer (Hempson et al. 2015 ), but woody tree species damage would have varied with preferred habitat, distance to water (Guldemond and Van Aarde 2008 ), and dominant plant traits in the community (e.g., wood density, water storage; Dantas and Pausas 2022 ). Fires become more frequent at higher levels of rainfall, reaching a maximum in mesic savannas and dropping off again in (wetter) closed-canopy forest vegetation. In contrast, the probability of extreme drought (measured with Foley’s drought index) correlates negatively with mean annual precipitation and arid places are more likely to be exposed to sustained periods of below-average rainfall, which can last for several years (Khan and Gilani 2021 ). Frost is largely related to topography with high-altitude plateaus and escarpments being exposed to up to 100 days of frost a year in some places (figure 2 ), but frost pockets can also occur in valleys where cold-air pooling occurs (Finckh et al. 2021 ). Likewise, cyclones are correlated with altitude and longitude and are focused on the low-lying areas to the east of southern Africa (Ramsay 2017 ). Wood fuel harvesting occurs in most parts of Africa but is patchily distributed with concentrations close to major urban areas. Higher population landscapes in East Africa tend to have higher harvesting rates (Bailis et al. 2015 ). Figure 2. Open in a new tab Regional spatial distribution of large tree top-kill agents in savannas, showing: cyclone frequency (cyclones per year), megaherbivore density (in kilograms per square kilometer), the number of frost days, fire frequency (the number of burns in a 22-year period), drought intensity (Foley’s drought index), and human impact (from the Global Human Footprint Index as a percentage). The scale is reversed for drought because lower values suggest higher intensities A clustering analysis (see the methods in the supplemental material ) reveals six distinct clusters that differ in their exposure to woody species top-kill (figure 3 ). The areas associated with cluster 1 have moderate fires and high elephant densities, occasionally affected by moderate drought conditions. This cluster occupies approximately 33.7% of sub-Saharan Africa and covers the largest spatial extent. The second largest cluster (27.6%), unlike cluster 1, has frequent fires with areas experiencing a burn every 1–2 years and high elephant densities but experiences minimal drought. Cluster 3 experiences the full drought spectrum, including intense droughts with variable fire return intervals, but can also be prone to frost. Cluster 4 is explicitly characterized by rare fire occurrences (i.e., one fire every 5 to 20 years or more), but drought, and megaherbivore density can still cause woody species top-kill. It covers approximately 13.7% of sub-Saharan Africa, compared with 4.43% for cluster 5, which is prone to cyclones. The cyclone cluster occurs together with frequent fires along the eastern coast of Southern Africa, specifically Mozambique and throughout Madagascar. Finally, the smallest cluster (cluster 6), which covers only 3.0% of sub-Saharan Africa and is concentrated in the southern hemisphere, was prone to frost but also high elephant densities and low to moderate rates of fire (approximately every 4–20 years). Figure 3. Open in a new tab Spatial clustering of top-kill agents based on hierarchical clustering (a), (b) and a principal component analysis (c). The x -axis on panel (b) shows cyclone frequency (cyclones per year), megaherbivore density (in kilograms per square kilometer), the number of frost days, fire frequency (the number of burns in a 22-year period), drought intensity (Foley’s drought index) and human impact (the Global Human Footprint Index) Identifying high top-kill risk regions We further determined a risk of woody species’ exposure to top-kill by identifying pixels where the top-kill agents; cyclone frequency (cyclones per year), megaherbivore density (in kilograms per square kilometer), the number of frost days, fire frequency (the number of burns in a 22-year period), drought intensity (Foley’s drought index) or human impact (from the Global Human Footprint Index as a percentage) occurred at higher-than-average measures. A sum of top-kill agents in a pixel defined the risk of exposure to multiple agents (Archibald and Hempson 2016 ). We compared top-kill exposure in the past, when human impacts were low and elephants were dominant and unconstrained (Hempson et al. 2015 ), to a current scenario where human impacts are dominant, elephants extirpated in most areas, reduced in numbers and constrained in protected areas (Chase et al. 2016 , Venter et al. 2016 ). Examining the past scenario, we found regions exposed from 1 out of 5 up to 4 out of 5 top-kill agents (figure 4a ). Low exposure (i.e., exposure to one or two top-kill agents) occurred in areas experiencing long fire return intervals (more than 100 years) or prone to extreme droughts. These low-exposure areas dominate the continent, occurring throughout most of Central Africa (too wet) and southeastern Africa (too dry). High exposure areas (i.e., where woody species are exposed to three or more top-kill agents) are dominant across the remainder of sub-Saharan Africa and the east coast through Mozambique. However, high-exposure areas also form a discontinuous belt across the top of the sub-Saharan caused by high estimated past elephant densities. High-exposure areas had high rates of fire return (at least 3 years), high elephant density (at least 3000 kilograms per square kilometer), drought index up to −0.98, and either annual frost day frequency up to 2.5 days or an average cyclone frequency of at least 0.006 cyclones per year. There were no pixels that experienced both cyclones and frost. Figure 4. Open in a new tab Top-kill risk exposure across the African landscape represented by the cumulative number of top-kill agents likely to occur in a 0.5° × 0.5° pixel grid at higher than mean values: (a) historical exposure, expected to have driven resprouting adaptations; (b) current exposure, where elephant top-kill is reduced and human impacts have intensified; (c) exposure quantified for three size classes of trees (less than 5 cm, targeted by fire, frost, and drought; 5–10 cm, targeted by frost, drought, and now also human impact; and greater than 10 cm, drought, cyclones, and elephants, and now humans). Moving from megaherbivore to human dominated landscapes When we add the human top-kill agent and map current, in contrast to reconstructed past elephant densities, we observe changes in the spatial patterns. Woody tree species in approximately 18% of sub-Saharan Africa are now exposed to one additional agent because of the addition of human impacts in the current scenario (i.e., these areas are currently exposed to both elephant and human impacts). Interestingly, approximately 21% of the region is currently exposed to one less top-kill agent. These are areas previously exposed to high levels of elephant impact that no longer contain elephants but still have lower than average levels of human impact. Across nearly two-thirds (61%) of sub-Saharan Africa, there was no change to the degree of exposure, implying that high levels of human impact now occur in places where elephant impacts have been reduced. Notably, in Madagascar, which never had elephants, large areas now experience exposure to more disturbance agents than previously. Ontogenetic risk of exposure to top-kill We evaluated the risk of being exposed to top-kill for different size classes of woody trees based only on agents known to target different size classes: less than 5 cm DBH (fire, frost and drought), 5–10 cm DBH (frost, droughts, and human impacts; i.e., harvesting in the current scenario), and more than 10 cm DBH (drought, cyclones, and elephants and also including human impacts in the current scenario; table 1 ). Overall, all size classes of trees are exposed to at least one agent in most of the sub-Saharan Africa area. Comparing historical and current scenarios, trees less than 5 cm DBH are still affected by the same top-kill agents (fire, frost, and drought) and over the same spatial range. Medium-size trees have been the most vulnerable to temporal change because of increased harvesting, whereas the exposure to top-kill for larger trees greater than 10 cm DBH has been reduced in many areas as elephants are extirpated in most of their historical ranges. In some areas, potential reductions in top-kill by elephants has been compensated for by increased wood-fuel harvesting impacts. Expected tree resprouting responses Resprouting is common in tropical ecosystems (Pausas et al. 2016 ), but so far, it has been difficult to disentangle the relative influence of different top-kill agents on resprouting responses in an evolutionary context. Resprouting has been repeatedly lost and regained within evolutionary lineages when the disturbance selecting for resprouting changed (Pausas and Keeley 2014 , Lamont et al. 2019 ). For example, the question of whether drought or fire is the primary selective force for basal resprouting has been the subject of debate (Axelrod 1980 , López-Soria and Castell 1992 , Archibald et al. 2018 ), although there is now more conclusive evidence for the role of fire in shaping this plant trait (Archibald et al. 2018 , He and Lamont 2018 ). In the present article, we explored how aligned various top-kill agents are—both in their spatiotemporal occurrence and in the size classes of trees most affected—to help set up expectations for the types of resprouting responses we expect to find in savannas evolving under different combinations of disturbances. Implications of spatial associations between top-kill agents We provide evidence of spatial aggregation in top-kill agents that affect woody trees in African savannas (figure 2 ). Our observations suggest that fire and elephants had similar spatial patterns that can further be separated into high- or low-drought areas. Although cyclones are concentrated in a small section of southeast Africa, these regions also have high levels of fire impact. In contrast, frost is largely concentrated in regions without high elephant densities, but it can also co-occur with drought. We suggest that these unique aggregations create hotspots where combinations of top-kill agents are dominant, potentially imposing aligned or conflicting selection pressures on recruitment strategies in those regions. Fire top-kills smaller size classes less than 5 cm DBH (Hoffmann and Solbrig 2003 ) across many woody species, whereas megaherbivores preferentially top-kill medium to large size classes of trees of selected species (Holdo 2005 , Cardoso et al. 2020 , Thompson et al. 2022 ). Therefore, at high levels of rainfall, where both fire and elephants are common but drought impacts are rare, we expect prolific resprouting of most trees at early stages to escape the fire trap, but only species that are preferred and selected by elephants will benefit from retaining this ability as adults. In regions where extreme droughts can impose top-kill events on all size classes of trees (Case et al. 2019 , Sankaran 2019 ), we would expect the dominant tree species to maintain their resprouting potential throughout their lives but also that the resprouting response should start at a very young age—within a few weeks of germination, where even a brief period without rain can top-kill a seedling (Cardoso et al. 2020 ) and where even small herbivores can cause top-kill (Archibald et al. 2021 ). Cyclones and elephants target similar size class trees, so although Madagascar evolved without elephants, the resprouting strategies found there might be quite similar to those found in the mainland—with potentially a higher proportion of adult resprouters because of the less selective nature of cyclone damage. From a tree perspective, frost and fire damage aboveground biomass in similar ways (Botha et al. 2020 ), affecting similar size classes (Holdo 2005 , Bond and Van Wilgen 2012 ), and the appropriate resprouting responses are also likely to be comparable: vigorous resprouting as a seedling or sapling but, once the sapling has escaped the frost or fire trap, resprouting would no longer be selected for. The combination of these disturbances might also increase their individual impacts: with fire removing the protective layer of grass and exposing seedlings to frost and frost maintaining plants in size classes where they are exposed to the more intense damage caused by fires (Holdo 2005 , Hoffmann et al. 2019 ). Both frost and fire have been proposed as the main driver of the geoxyle growth form—an extreme resprouting response where woody species retreat belowground, preventing stem exposure to disturbance agents, and only resprout leaves and flowers between disturbance events (Maurin et al. 2014 , Finckh et al. 2016 , Courtenay et al. 2023 ). Importantly, frost-prone ecosystems are also found in combination with elephants and droughts (figure 3 ), and the area exposed to frost might have been much larger as recently as 20,000 years ago during the last glacial maximum (Partridge et al. 1999 ). Moving from megaherbivore- to human-dominated landscapes Humans have been modifying African savanna vegetation for millennia through harvesting, fire initiation and suppression, increasing domestic herbivore densities, and cropland conversion. Early records for the use of fire date back perhaps 1.5 million years (Stevens et al. 2022 ), but human impacts have recently intensified (Venter et al. 2016 ). Some areas that had a high density of elephants in the past currently have high levels of human impact, highlighting the potential for humans to replace an important ecological process that is now missing from the landscape. To fully evaluate this, it would be necessary to have better insights into the levels of tree harvesting that are ecologically representative of past elephant damage. Elephants and humans may often select similar size classes of trees (figure 1 ), but their preferences differ: Softer-wooded, more palatable species are more often top-killed by elephants (Teren et al. 2018 , Ferry et al. 2021 ), whereas humans tend to select the harder-wooded species for timber and charcoal (Pelletier et al. 2019 ). In addition, human impacts are concentrated around settlements (Pelletier et al. 2019 ), and elephant impacts are concentrated close to water bodies and rivers (Owen-Smith et al. 2019 ). Moreover, harvesting and removing trees, instead of leaving the toppled wood in the ecosystem, will result in changed biogeochemistry and carbon cycling. We know that elephants reduce plant densities of preferred tree species in savannas (Coetsee et al. 2023 ) and can also push some species to local extirpation (Young et al. 2021 , O’Connor et al. 2024 ). There is also evidence that human impacts cause degradation and conversion of savanna woodlands (Fisher et al. 2012 ), often exceeding the rates of natural top-kill agents like elephants. Nevertheless, across large portions of the African landscape, designing sustainable tree-harvesting regimes that aim to replicate the impacts of the missing keystone elephant is certainly worth careful investigation (Okello et al. 2001 ). There are substantial portions of Africa where elephants have been lost but human impacts appear to remain low, and land management and conservation in this region should account for this when exploring rewilding options (Young et al. 2021 , Fløjgaard et al. 2022 ). However, there is an equivalently large area where human impacts are additional to existing top-kill processes, and these might be particularly vulnerable, especially if the species did not evolve with elephants or cyclones and therefore might not retain resprouting capacity as large trees. Mesoherbivores as top-kill agents Smaller herbivores such as duiker and impala can top-kill very small seedlings and might select for early resprouting in areas with high browsing pressure (Archibald et al. 2021 ). Limitations to the size of stem that can be bitten (Wilson and Kerley 2003 ) means that mesoherbivores generally do not top-kill larger stems, and they were not included in this analysis. However, through chronic browsing of small seedlings, they can trap trees in smaller size classes—exposed to other top-kill agents for years (LaMalfa et al. 2021 ). Past and future fire regimes One key piece of information missing from this analysis is the extent and frequency of fire in the past. Little is known and much surmised about past natural and anthropogenic fire regimes, and our assumption in this article is that the spatial pattern is similar (driven by environmental factors driving fuel accumulation and drying). This seems reasonable but might underpredict the past fires experienced in areas that have been substantially transformed by agriculture and where fire is suppressed (Archibald et al. 2010 , Jones et al. 2022 ). Moreover, intensities were probably higher (more late dry season fires) and return times lower (fewer ignition events) in the past, and higher intensity fires tend to top-kill even larger trees (Ryan and Williams 2011 ). We expect fire regimes in Africa to respond to global change: Model ensembles predict burned areas will decrease in Africa as water limitations from high temperatures reduce grassy fuel loads below that which can easily spread a fire (Moritz et al. 2012 ). Indeed, where fire is suppressed, increase in woody cover occurs because of the lack of top-kill agents (Case and Staver 2017 ), and the loss of elephants also results in woody encroachment (Stevens et al. 2016 ). However, it is also likely that extreme fires will be more common in high-biomass forest or thicket environments, because increasing global temperatures increase the probability of firestorms and canopy fires (Cardoso et al. 2023 ). Therefore, fire-induced top-kill or mortality could become important in tree communities previously rarely exposed to fire, and the ecological consequence will depend on the resprouting responses of the species involved. Conclusions African savannas continue to be some of the most heavily used ecosystems on the globe, providing important ecosystem services to some of the world’s poorest and most vulnerable populations (Lehmann and Parr 2016 ). The fact that these systems evolved with a range of top-kill agents and that many tree species resprout vigorously after damage likely contributes to their resilience. At least one study that has quantified tree harvesting (Bailis et al. 2015 ) suggested that about 70% of this is sustainable (i.e., less than or equal to tree growth rates and not reducing the aboveground biomass); including ecological information on the rates of top-kill under different elephant densities would provide a new lens for defining and discussing sustainable harvesting. These results also help to explain the global nature of resprouting adaptations (Pausas and Keeley 2017 ), which could be evidence of the past disturbance regimes and could help guide insights into how best to rewild and manage both tropical and nontropical ecosystems (Churski et al. 2017 ). This analysis extends our understanding of the resprouting response of larger trees: From the data presented in this article, we expect biogeographical differences in the occurrence of adult resprouting response, driven by spatial variation in the top-kill agents that affect adult trees. However very little is known about resprouting in larger size classes and under what conditions top-kill of large trees actually results in full mortality. Better quantification of this would improve our understanding of the evolutionary pressures on trees in tropical ecosystems and would also enable harvesting guidelines that prevent mortality and increase ecosystem productivity (see Okello et al. 2001 ). To advance our understanding, we encourage future research using empirical data to test hypotheses about expected ontogenetic resprouting responses of tree species and individuals in these disturbance-prone regions, where top-kill probabilities vary. This will not only allow us to move beyond theoretical assumptions but to understand how tree species adapt and survive under different disturbance regimes. This information is critical for predicting ecological outcomes, guiding conservation efforts, and informing land-management practices tailored to local environmental stresses. Supplementary Material biag008_Supplemental_File biag008_supplemental_file.docx (159.9KB, docx) Acknowledgments The authors thank Gareth Hempson and Thomas Ibanez for assisting with spatial data layers and Caroline Lehmann, Truman Young, William Bond, and two anonymous reviewers for helpful discussions on the text and ideas. Authors declare no conflict of interests. Author Biography Fezile Mtsetfwa is a postdoctoral researcher, in the School of Animal, Plant, and Environmental Sciences at the University of Witwatersrand, South Africa. Sally Archibald is a professor and co–principal investigator of the Future Ecosystems for Africa Program in the School of Animal, Plant, and Environmental Sciences at the University of Witwatersrand, South Africa. Contributor Information Fezile Mtsetfwa, School of Animal Plant & Environmental Sciences, University of the Witwatersrand, 1 Jan Smuts Ave, Braamfontein 2000, Johannesburg, South Africa. Sally Archibald, School of Animal Plant & Environmental Sciences, University of the Witwatersrand, 1 Jan Smuts Ave, Braamfontein 2000, Johannesburg, South Africa. Data availability The data underlying this article were provided by multiple researchers working on the different disturbances by permission or available in Dryad Digital Repository at https://doi.org/10.5061/dryad.052q5 . Private data will be shared on request to the corresponding author with permission of third party researches. Author contributions Fezile Mtsetfwa (Conceptualization, Data curation, Formal Analysis, Investigation, Methodology, Writing – original draft), Sally Archibald (Conceptualization, Funding acquisition, Supervision, Writing – review & editing). Funding This research was funded by the Oppenheimer Generations Research and Conservation, Future Ecosystems for Africa Program and the National Research Foundation (NRF Grant Number: PSTD230504101921). References cited Abraham JO, Goldberg ER, Botha J, Staver AC. 2021. Heterogeneity in African savanna elephant distributions and their impacts on trees in Kruger National Park, South Africa. Ecology and Evolution. 11: 5624–5634. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Archibald S, Hempson GP. 2016. Competing consumers: Contrasting the patterns and impacts of fire and mammalian herbivory in Africa. Philosophical Transactions of the Royal Society B. 371: 20150309. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Archibald S, Twine W, Mthabini C, Stevens N. 2021. Browsing is a strong filter for savanna tree seedlings in their first growing season. Journal of Ecology. 109: 3685–3698. [ Google Scholar ] Archibald S. 2018. Biological and geophysical feedbacks with fire in the Earth system. Environmental Research Letters. 13: 033003. [ Google Scholar ] Archibald3 S, Scholes RJ, Roy DP, Roberts G, Boschetti L. 2010. Southern African fire regimes as revealed by remote sensing. International Journal of Wildland Fire. 19: 861–878. [ Google Scholar ] Archibald4 S, Lehmann CER, Gómez-Dans JL, Bradstock RA. 2013. Defining pyromes and global syndromes of fire regimes. Proceedings of the National Academy of Sciences. 110: 6442–6447. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Archibald5 S, Hempson GP, Lehmann C. 2019. A unified framework for plant life-history strategies shaped by fire and herbivory. New Phytologist. 224: 1490–1503. [ DOI ] [ PubMed ] [ Google Scholar ] Asner GP, Levick SR. 2012. Landscape-scale effects of herbivores on treefall in African savannas. Ecology Letters. 15: 1211–1217. [ DOI ] [ PubMed ] [ Google Scholar ] Axelrod DI. 1980. History of the Maritime Closed-Cone Pines, Alta and Baja California. University of California Press. [ Google Scholar ] Bailis R, Drigo R, Ghilardi A, Masera O. 2015. The carbon footprint of traditional woodfuels. Nature Climate Change. 5: 266–272. [ Google Scholar ] Bellingham PJ, Sparrow AD. 2000. Resprouting as a life history strategy in woody plant communities. Oikos. 89: 409–416. [ Google Scholar ] Bond WJ, Midgley JJ. 2001. Ecology of sprouting in woody plants: The persistence niche. Trends in Ecology and Evolution. 16: 45–51. [ DOI ] [ PubMed ] [ Google Scholar ] Bond WJ, Midgley JJ. 2003. The Evolutionary Ecology of Sprouting in Woody Plants. International Journal of Plant Sciences. 164: S103–S114. [ Google Scholar ] Bond WJ, Van Wilgen BW. 2012. Fire and Plants. Springer. [ Google Scholar ] Botha M, Archibald S, Greve M. 2020. What drives grassland-forest boundaries? Assessing fire and frost effects on tree seedling survival and architecture. Ecology and Evolution. 10: 10719–10734. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Brade TK. 2022. Constraints on Tree Growth, Impacts of Tropical Cyclones and Outcomes of Community Management in the Miombo Woodlands. PhD dissertation. University of Edinburgh, Edinburgh, Scotland, United Kingdom. [ Google Scholar ] Cardoso AW, Beckett H, Bond WJ. 2023. How forests survive alongside flammable open ecosystems: Conservation implications for Africa. Frontiers in Conservation Science. 4: 1150516. [ Google Scholar ] Cardoso AW, Malhi Y, Oliveras I, Lehmann D, Ndong JE, Dimoto E, Bush E, Jeffery K, Labriere N, Lewis SL. 2020. The role of forest elephants in shaping tropical forest–savanna coexistence. Ecosystems. 23: 602–616. [ Google Scholar ] Cardoso AW, Medina-Vega JA, Malhi Y, Adu-Bredu S, Ametsitsi GKD, Djagbletey G, van Langevelde F, Veenendaal E, Oliveras I. 2016. Winners and losers: Tropical forest tree seedling survival across a West African forest–savanna transition. Ecology and Evolution. 6: 3417–3429. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Case MF, Staver AC. 2017. Fire prevents woody encroachment only at higher-than-historical frequencies in a South African savanna. Journal of Applied Ecology. 54: 955–962. [ Google Scholar ] Case MF, Wigley-Coetsee C, Nzima N, Scogings PF, Staver AC. 2019. Severe drought limits trees in a semi-arid savanna. Ecology. 100: e02842. [ DOI ] [ PubMed ] [ Google Scholar ] Chase MJ. 2016. Continent-wide survey reveals massive decline in African savannah elephants. PeerJ. 4: e2354. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Churski M, Bubnicki JW, Jędrzejewska B, Kuijper DPJ, Cromsigt JPGM. 2017. Brown world forests: Increased ungulate browsing keeps temperate trees in recruitment bottlenecks in resource hotspots. New Phytologist. 214: 158–168. [ DOI ] [ PubMed ] [ Google Scholar ] Clarke PJ, Lawes MJ, Midgley JJ, Lamont BB, Ojeda F, Burrows GE, Enright NJ, Knox KJE. 2013. Resprouting as a key functional trait: How buds, protection and resources drive persistence after fire. New Phytologist. 197: 19–35. [ DOI ] [ PubMed ] [ Google Scholar ] Coetsee C, Botha J, Case MF, Manganyi A, Siebert F. 2023. The hard lives of trees in African savanna: Even without elephants. Austral Ecology. 48: 532–551. [ Google Scholar ] Courtenay AP, Moonlight PW, Toby Pennington R, Lehmann CE. 2023. Underground trees inhabit varied environmental extremes across the Afrotropics. Annals of Botany. 133: 757–772. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Dantas VL, Pausas JG. 2022. The legacy of the extinct Neotropical megafauna on plants and biomes. Nature Communications. 13: 129. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] De Antonio AC, Hoffmann WA, Rossatto DR. 2021. The role of morpho-physiological traits in frost tolerance of neotropical savanna trees. Trees. 35: 1687–1696. [ Google Scholar ] Delaporte B, Ibanez T, Despinoy M, Mangeas M, Menkes C. 2022. Tropical cyclone impact and forest resilience in the Southwestern Pacific. Remote Sensing. 14: 1245. [ Google Scholar ] Ellis R. 2008. Putting people in the map: Anthropogenic biomes of the world. Frontiers in Ecology and the Environment. 6: 439–447. [ Google Scholar ] Ferry N, Dray S, Fritz H, Ipavec A, Wigley BJ, Charles-Dominique T, Bourgarel M, Sebele L, Valeix M. 2021. Long-term high densities of African elephants clear the understorey and promote a new stable savanna woodland community. Journal of Vegetation Science. 32: e13101. [ Google Scholar ] Finckh M, Revermann R, Aidar MPM. 2016. Climate refugees going underground: A response to Maurin et al. (2014). New Phytologist. 209: 904–909. [ DOI ] [ PubMed ] [ Google Scholar ] Finckh M, Wendefeuer J, Meller P. 2021. Frost-driven lower treelines in Angola and their implications for tropical forest–grassland mosaics. Journal of Vegetation Science. 32: e13084. [ Google Scholar ] Fisher JT, Witkowski ETF, Erasmus BFN, Van Aardt J, Asner GP, Wessels KJ, Mathieu R. 2012. Human-modified landscapes: Patterns of fine-scale woody vegetation structure in communal savannah rangelands. Environmental Conservation. 39: 72–82. [ Google Scholar ] Fløjgaard C, Pedersen PBM, Sandom CJ, Svenning J-C, Ejrnæs R. 2022. Exploring a natural baseline for large-herbivore biomass in ecological restoration. Journal of Applied Ecology. 59: 18–24. [ Google Scholar ] Franco AC, Álvarez-Yépiz JC. 2022. Adaptive responses of tropical savanna trees to frost. Trees. 36: 1–5. [ Google Scholar ] Fricke EC, Ordonez A, Rogers HS, Svenning JC. 2022. The effects of defaunation on plants’ capacity to track climate change. Science. 375: 210–214. -+. [ DOI ] [ PubMed ] [ Google Scholar ] Gadd ME. 2002. The impact of elephants on the marula tree Sclerocarya birrea . African Journal of Ecology. 40: 328–336. [ Google Scholar ] Gill JL. 2014. Ecological impacts of the late Quaternary megaherbivore extinctions. New Phytologist. 201: 1163–1169. [ DOI ] [ PubMed ] [ Google Scholar ] González-M R, et al. 2021. Diverging functional strategies but high sensitivity to an extreme drought in tropical dry forests. Ecology Letters. 24: 451–463. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Guldemond R, Van Aarde R. 2008. A meta-analysis of the impact of African elephants on savanna vegetation. Journal of Wildlife Management. 72: 892–899. [ Google Scholar ] Handavu F, Syampungani S, Chisanga E. 2011. The influence of stump diameter and height on coppicing ability of selected key Miombo woodland tree species of Zambia: A guide for harvesting for charcoal production. Journal of Ecology and the Natural Environment. 3: 461–468. [ Google Scholar ] Harrison SP. 2021. Understanding and modelling wildfire regimes: An ecological perspective. Environmental Research Letters. 16: 125008. [ Google Scholar ] He T, Lamont BB. 2018. Baptism by fire: The pivotal role of ancient conflagrations in evolution of the Earth’s flora. National Science Review. 5: 237–254. [ Google Scholar ] Hempson GP, Archibald S, Bond WJ. 2015. A continent-wide assessment of the form and intensity of large mammal herbivory in Africa. Science. 350: 1056–1061. [ DOI ] [ PubMed ] [ Google Scholar ] Hoffmann WA, Flake SW, Abreu RCR, Pilon NAL, Rossatto DR, Durigan G. 2019. Rare frost events reinforce tropical savanna–forest boundaries. Journal of Ecology. 107: 468–477. [ Google Scholar ] Hoffmann WA, Rodrigues AC, Uncles N, Rossi L. 2021. Hydraulic segmentation does not protect stems from acute water loss during fire. Tree Physiology. 41: 1785–1793. [ DOI ] [ PubMed ] [ Google Scholar ] Hoffmann WA, Solbrig OT. 2003. The role of topkill in the differential response of savanna woody species to fire. Forest ecology and management. 180: 273–286. [ Google Scholar ] Holdo RM. 2005. Stem mortality following fire in Kalahari sand vegetation: Effects of frost, prior damage, and tree neighbourhoods. Plant Ecology. 180: 77–86. [ Google Scholar ] Holdo RM. 2006. Elephant herbivory, frost damage and topkill in Kalahari sand woodland savanna trees. Journal of Vegetation Science. 17: 509–518. [ Google Scholar ] Hood SM, Varner JM, Mantgem Pv, Cansler CA. 2018. Fire and tree death: Understanding and improving modeling of fire-induced tree mortality. Environmental Research Letters. 13: 113004. [ Google Scholar ] House A, Bamford MK. 2019. Investigating the utilisation of woody plant species at an Early Iron Age site in KwaZulu-Natal, South Africa, by means of identifying archaeological charcoal. Archaeological and Anthropological Sciences. 11: 6737–6750. [ Google Scholar ] Ibanez T, Keppel G, Menkes C, Gillespie TW, Lengaigne M, Mangeas M, Rivas-Torres G, Birnbaum P. 2019. Globally consistent impact of tropical cyclones on the structure of tropical and subtropical forests. Journal of Ecology. 107: 279–292. [ Google Scholar ] Jones MM, Fletcher R Jr, Kruger L, Monadjem A, Simelane P, McCleery R. 2022. Drought limits large trees in African savannas with or without elephants. Austral Ecology. 47: 1668–1684. [ Google Scholar ] Jurisch K, Hahn K, Wittig R, Bernhardt-Römermann M. 2013. Land-use impact on the growth and survival of seedlings and saplings in W est A frican savannas. Journal of Vegetation Science. 24: 101–112. [ Google Scholar ] Khan R, Gilani H. 2021. Global drought monitoring with big geospatial datasets using Google Earth Engine. Environmental Science and Pollution Research. 28: 17244–17264. [ DOI ] [ PubMed ] [ Google Scholar ] LaMalfa EM, Riginos C, Veblen KE. 2021. Browsing wildlife and heavy grazing indirectly facilitate sapling recruitment in an East African savanna. Ecological Applications. 31: e02399. [ DOI ] [ PubMed ] [ Google Scholar ] Lamont BB, He T, Yan Z. 2019. Evolutionary history of fire-stimulated resprouting, flowering, seed release and germination. Biological Reviews. 94: 903–928. [ DOI ] [ PubMed ] [ Google Scholar ] Lehmann CER, Parr CL. 2016. Tropical grassy biomes: Linking ecology, human use and conservation. Philosophical Transactions of the Royal Society B. 371: 20160329. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] López-Soria L, Castell C. 1992. Comparative genet survival after fire in woody Mediterranean species. Oecologia. 91: 493–499. [ DOI ] [ PubMed ] [ Google Scholar ] Luoga EJ, Witkowski ETF, Balkwill K. 2004. Regeneration by coppicing (resprouting) of miombo (African savanna) trees in relation to land use. Forest Ecology and Management. 189: 23–35. [ Google Scholar ] Maurin O, Davies TJ, Burrows JE, Daru BH, Yessoufou K, Muasya AM, van der Bank M, Bond WJ. 2014. Savanna fire and the origins of the ‘underground forests’ of Africa. New Phytologist. 204: 201–214. [ DOI ] [ PubMed ] [ Google Scholar ] Midgley JJ, Kruger LM, Skelton R. 2011. How do fires kill plants? The hydraulic death hypothesis and Cape Proteaceae “fire-resisters.”. South African Journal of Botany. 77: 381–386. [ Google Scholar ] Moncrieff GR, Lehmann CE, Schnitzler J, Gambiza J, Hiernaux P, Ryan CM, Shackleton CM, Williams RJ, Higgins SI. 2014. Contrasting architecture of key A frican and A ustralian savanna tree taxa drives intercontinental structural divergence. Global ecology and biogeography. 23: 1235–1244. [ Google Scholar ] Moritz MA, Parisien M-A, Batllori E, Krawchuk MA, Van Dorn J, Ganz DJ, Hayhoe K. 2012. Climate change and disruptions to global fire activity. Ecosphere. 3: 1. [ Google Scholar ] Muller K, O’Connor TG, Henschel JR. 2016. Impact of a severe frost event in 2014 on woody vegetation within the Nama-Karoo and semi-arid savanna biomes of South Africa. Journal of Arid Environments. 133: 112–121. [ Google Scholar ] N’Dri AB, Kpangba KP, Werner PA, Koffi KF, Bakayoko A. 2022. The response of sub-adult savanna trees to six successive annual fires: An experimental field study on the role of fire season. Journal of Applied Ecology. 59: 1347–1361. [ Google Scholar ] Neke KS, Owen-Smith N, Witkowski ETF. 2006. Comparative resprouting response of Savanna woody plant species following harvesting: The value of persistence. Forest Ecology and Management. 232: 114–123. [ Google Scholar ] O’Connor TG, Pallett NCM, Clegg BW, Shimbani J. 2024. Local extirpation of woody species in Colophospermum mopane woodland under chronic utilisation by elephants. African Journal of Ecology. 62: e13296. [ Google Scholar ] Okello BD, O’connor TG, Young TP. 2001. Growth, biomass estimates, and charcoal production of Acacia drepanolobium in Laikipia, Kenya. Forest Ecology and Management. 142: 143–153. [ Google Scholar ] Oliveira RS, Eller CB, Barros F de V, Hirota M, Brum M, Bittencourt P. 2021. Linking plant hydraulics and the fast–slow continuum to understand resilience to drought in tropical ecosystems. New Phytologist. 230: 904–923. [ DOI ] [ PubMed ] [ Google Scholar ] Owen-Smith N, Page B, Teren G, Druce DJ. 2019. Megabrowser impacts on woody vegetation in savannas. Pages 585–611. in Scogings PF, Sankaran M, eds. Savanna Woody Plants and Large Herbivores. Wiley. [ Google Scholar ] Partridge TC, Scott L, Hamilton JE. 1999. Synthetic reconstructions of southern African environments during the last glacial maximum (21–18kyr) and the Holocene Altithermal (8–6kyr). Quaternary International. 57-58: 207–214. [ Google Scholar ] Pausas JG, Keeley JE. 2014. Evolutionary ecology of resprouting and seeding in fire-prone ecosystems. New Phytologist. 204: 55–65. [ DOI ] [ PubMed ] [ Google Scholar ] Pausas JG, Keeley JE. 2017. Epicormic Resprouting in Fire-Prone Ecosystems. Trends in Plant Science. 22: 1008–1015. [ DOI ] [ PubMed ] [ Google Scholar ] Pausas JG, Pratt RB, Keeley JE, Jacobsen AL, Ramirez AR, Vilagrosa A, Paula S, Kaneakua-Pia IN, Davis SD. 2016. Towards understanding resprouting at the global scale. New Phytologist. 209: 945–954. [ DOI ] [ PubMed ] [ Google Scholar ] Pelletier J, Chidumayo E, Trainor A, Siampale A, Mbindo K. 2019. Distribution of tree species with high economic and livelihood value for Zambia. Forest Ecology and Management. 441: 280–292. [ Google Scholar ] Pringle RM, Kimuyu DM, Sensenig RL, Palmer TM, Riginos C, Veblen KE, Young TP. 2015. Synergistic effects of fire and elephants on arboreal animals in an African savanna. Journal of Animal Ecology. 84: 1637–1645. [ DOI ] [ PubMed ] [ Google Scholar ] Ramsay H. 2017. The Global Climatology of Tropical Cyclones. Oxford Research Encyclopedia of Natural Hazard Science. Oxford University Press. [ Google Scholar ] Ryan CM, Williams M. 2011. How does fire intensity and frequency affect miombo woodland tree populations and biomass?. Ecological Applications. 21: 48–60. [ DOI ] [ PubMed ] [ Google Scholar ] Salguero-Gomez R, Siewert W, Casper BB, Tielbörger K. 2012. A demographic approach to study effects of climate change in desert plants. Philosophical Transactions of the Royal Society B. 367: 3100–3114. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Sandel B, Svenning J-C. 2013. Human impacts drive a global topographic signature in tree cover. Nature Communications. 4: 2474. [ DOI ] [ PubMed ] [ Google Scholar ] Sankaran M. 2019. Droughts and the ecological future of tropical savanna vegetation. Journal of Ecology. 107: 1531–1549. [ Google Scholar ] Scalon MC. 2021. Fire and drought: Shifts in bark investment across a broad geographical scale for Neotropical savanna trees. Basic and Applied Ecology. 56: 110–121. [ Google Scholar ] Shen W, Zhang L, Liu X, Luo T. 2014. Seed-based treeline seedlings are vulnerable to freezing events in the early growing season under a warmer climate: Evidence from a reciprocal transplant experiment in the Sergyemla Mountains, southeast. Agricultural and Forest Meteorology. 187: 83–92. [ Google Scholar ] Simpson RH. 1974. The hurricane disaster: Potential scale. Weatherwise. 27: 169–186. [ Google Scholar ] Stevens N, Bond W, Feurdean A, Lehmann CER. 2022. Grassy ecosystems in the Anthropocene. Annual Review of Environment and Resources. 47: 261–289. [ Google Scholar ] Stevens N, Erasmus BFN, Archibald S, Bond WJ. 2016. Woody encroachment over 70 years in South African savannahs: Overgrazing, global change or extinction aftershock?. Philosophical Transactions of the Royal Society B. 371: 20150437. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Teren G, Owen-Smith N, Erasmus BFN. 2018. Elephant-mediated compositional changes in riparian canopy trees over more than two decades in northern Botswana. Journal of Vegetation Science. 29: 585–595. [ Google Scholar ] Therrell MD, Stahle DW, Mukelabai MM, Shugart HH. 2007. Age, and radial growth dynamics of Pterocarpus angolensis in southern Africa. Forest Ecology and Management. 244: 24–31. [ Google Scholar ] Thompson KE, Ford A, Esteban G, Poupard A, Zoon K, Pettorelli N. 2022. Impacts of African savannah elephants ( Loxodonta africana ) on tall trees and their recovery within a small, fenced reserve in South Africa. African Journal of Ecology. 60: 357–366. [ Google Scholar ] Veldman JW. 2015. Toward an old-growth concept for grasslands, savannas, and woodlands. Frontiers in Ecology and the Environment. 13: 154–162. [ Google Scholar ] Venter O, Sanderson EW, Magrach A, Allan JR, Beher J, Jones KR, Possingham HP, Laurance WF, Wood P, Fekete BM. 2016. Sixteen years of change in the global terrestrial human footprint and implications for biodiversity conservation. Nature Communications. 7: 12558. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Wakeling JL, Cramer MD, Bond WJ. 2012. The savanna-grassland "treeline": Why don’t savanna trees occur in upland grasslands?. Journal of Ecology. 100: 381–391. [ Google Scholar ] Whitecross MA, Archibald S, Witkowski ETF. 2012. Do freeze events create a demographic bottleneck for Colophospermum mopane ?. South African Journal of Botany. 83: 9–18. [ Google Scholar ] Wilson SL, Kerley GIH. 2003. Bite diameter selection by thicket browsers: The effect of body size and plant morphology on forage intake and quality. Forest Ecology and Management. 181: 51–65. [ Google Scholar ] Woolley L-A, Murphy BP, Radford IJ, Westaway J, Woinarski JCZ. 2018. Cyclones, fire, and termites: The drivers of tree hollow abundance in northern Australia’s mesic tropical savanna. Forest Ecology and Management. 419–420: 146–159. [ Google Scholar ] Young TP, Kimuyu DM, Odadi WO, Wells HB, Wolf AA. 2021. Naïve plant communities and individuals may initially suffer in the face of reintroduced megafauna: An experimental exploration of rewilding from an African savanna rangeland. PLOS ONE. 16: e0248855. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Zeppel MJB. 2015. Drought and resprouting plants. New Phytologist. 206: 583–589. [ DOI ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials biag008_Supplemental_File biag008_supplemental_file.docx (159.9KB, docx) Data Availability Statement The data underlying this article were provided by multiple researchers working on the different disturbances by permission or available in Dryad Digital Repository at https://doi.org/10.5061/dryad.052q5 . Private data will be shared on request to the corresponding author with permission of third party researches. Articles from Bioscience are provided here courtesy of Oxford University Press ACTIONS View on publisher site PDF (2.0 MB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top