Chapter 13
Large herbivores transform savanna ecosystems most fundamentally by selectively consuming plants, thereby altering competitive relations among plant species and the structural balance between trees and grasses. Their impacts on vegetation affect the spread of fires and what proportion of the nutrients contained in the vegetation are recycled via soil organisms rather than through herbivore guts. Through locally concentrating their grazing, browsing, breakage and trampling, they contribute to the spatial heterogeneity that is an inherent feature of Africa’s savannas. Elephants have an especially great impact through their capacity to topple or uproot trees and kill even large trees through bark removal. Mega-grazers in the form of white rhinos and hippos transform grasslands by promoting grazing lawns in place of taller grasslands through their broad-lipped cropping actions. Smaller browsers suppress the establishment of tree seedlings by selectively consuming green leaves. Grazers of all sizes affect the ability of grass plants to re-establish their foliage above ground at the start of the wet season and following droughts or other disturbances. By reducing the fuel load, they affect the intensity of fires and hence the effect that fires have in restricting the expansion of woody plants or bush encroachment. Over what time frame can ecosystems be interpreted as stable, given the frequency of disturbances from variable rainfall, fire, and the shifting impacts of large herbivores?1
Ecosystems with few large herbivores would look very different from those retaining a full complement of these herbivores. This chapter considers how large herbivores alter the structure, composition and functioning of African savannas.
How Much Grass Do Grazers Consume?
The impacts that grazers have on the grass layer depend on how much of the grass biomass produced they consume annually. The proportion left unconsumed gets either incinerated by fire or deposited as litter for decomposition by soil organisms. This partitioning has important consequences for nutrient cycling. Fires result in nitrogen losses while concentrating inorganic mineral nutrients in ash. Digestion via a warm herbivore gut accelerates the conversion of part of the structural carbon into methane and carbon dioxide gases, thereby concentrating organic and mineral nutrients for more rapid recycling than takes place in the soil.
The total consumption of vegetation by large herbivores depends on the aggregate biomass of these consumers and their relative food requirements. Because of metabolic scaling (see below), larger animals require less food per unit of their biomass than do smaller ones. Aggregate biomass levels depend on the rainfall, governing the annual amount of vegetation produced, and on the soil fertility, governing the proportion of these plants that is potentially edible and digestible.
The highest aggregate biomass density of herbivores in Africa, by a considerable margin, was recorded within protected areas in Uganda, before local warfare intruded (Figure 13.1). The major contribution came from the huge numbers of elephants and hippos that were present then. The Hluhluwe-iMfolozi Park in South Africa, with white rhinos substituting for hippos, pips the Serengeti–Mara ecosystem for second place. A broader region of southern Africa – combining Kruger NP in South Africa, Hwange NP in Zimbabwe, Chobe NP in Botswana, and Ruaha NP in southern Tanzania – supports a herbivore biomass totalling over half of that in Serengeti, largely dominated by elephants. Western Africa, representing moist but dystrophic ecosystems, with Malawi in the east added, exhibits much lower herbivore biomass despite high rainfall. However, elephants were probably more abundant there in the past than they are today. Both Hluhluwe-iMfolozi and Serengeti still have growing elephant populations, which will raise their aggregate herbivore biomass levels closer towards the Uganda situation. More broadly across Africa, areas underlain by volcanic deposits or clay-rich sediments generating relatively fertile soils support aggregate herbivore biomass levels 5–10 times greater than found in savannas with similar rainfall underlain by infertile sandy soils.2,3
Figure 13.1
(A) Aggregate herbivore biomass and (B) projected combined consumption of vegetation in representative parks retaining largely intact assemblages of large mammals. Uganda is represented by Murchison Falls and Queen Elizabeth National Parks before their elephant and hippo numbers got decimated; HiP is the Hluhluwe-iMfolozi Park in south-eastern South Africa; Eastern Africa plateau grasslands are represented by the Serengeti–Mara ecosystem; southern Africa by Kruger, Chobe, Hwange, and Ruaha national parks combined; and western Africa by a combination of Comoe and Bouba Njida parks in Cameroon and Kasungu National Park in Malawi, all three predominantly miombo woodlands.
To accommodate the metabolic scaling of food requirements, species-specific biomass contributions need to be transformed into their metabolic mass equivalents. As was outlined in Chapter 10, mass-specific metabolic rates decline with increasing body size according to body mass raised to the power minus one quarter (M–0.25). For example, a wildebeest typically consumes grass amounting to around 2.5 percent of its body mass per day, or approaching 10 times its own weight over the course of a year. A white rhino eating 1.5 percent of its body mass per day will consume five times its body mass over a year. As a consequence, Serengeti is closer to Uganda and ahead of Hluhluwe-iMfolozi in terms of amount of vegetation consumed, because a greater proportion of its herbivore biomass is in the form of medium–large grazers rather than megaherbivores. The contribution of browsers, excluding elephants, to aggregate consumption is much smaller than that of grazers in all of these ecosystems.
To relate vegetation consumed to amount produced, an estimate is needed of the total biomass of vegetation generated annually above ground, focusing specifically on grass. Left ungrazed, the standing biomass of grass by the end of the growing season typically amounts to between a half and one gram of dry matter per millimetre of rainfall per square metre.4,5 However, the amount of grass actually produced could be over twice the peak biomass, because of ongoing turnover of leaves and stems.6,7 Root growth and turnover below ground can equal or exceed the production of leaves and stems above ground.8 For Serengeti, the projected annual biomass produced by grasses is about 100 times greater than the total biomass of large herbivores.
Even in Serengeti, the aggregate offtake of grass by large herbivores represents less than 20 percent of the total annual amount of grass produced above ground. This projected offtake is consistent with field measurements in Serengeti, which showed a 30 percent reduction in end-of-season biomass in plots open to grazing compared with fenced exclosures.9 However, the average estimate obscures wide spatial variation in grass consumption as well as variation between years in response to variable rainfall. Much grass remains ungrazed on hillsides and other localities where few ungulates normally go. Spatially, the total amount of vegetation consumed by large herbivores in Serengeti could entail the removal of nearly half of the plant biomass produced over half of the ecosystem and none over the rest. In drought years when rainfall and hence grass growth may be reduced to merely half of the mean, almost all of the grass produced above ground can be consumed over extensive areas (Figures 12.2 and 13.2). Places close to water where grazing ungulates congregate during the dry season routinely become denuded of grass, with hippos adding to the offtake by land-based herbivores. A much smaller fraction of tree foliage produced is consumed by browsers, because most remains out of reach and little of it is retained by trees through the dry season. Nevertheless, woody plants with evergreen or semi-evergreen leaves may have most of their foliage removed as high as herbivores can reach by late in the dry season.
Figure 13.2
Grazing lawns. (A) An extensive short-grass lawn cultivated by white rhinos in Mfolozi GR; (B) lawn grass cover in Mfolozi GR; (C) lawn grassland cultivated by wildebeest and other grazers in Serengeti plains; (D) lawn grassland generated by wildebeest grazing in Kruger NP; (E) lawn generated by hippos being maintained by puku, Luangwa Valley, Zambia; (F) lawn sustained by grazing in Central Kalahari, Botswana.
Spatial variation in grazing impacts was the incentive for my white rhino study in the Hluhluwe-iMfolozi Park. Some areas were being grazed down to bared soil, while elsewhere stands of tall grass were left ungrazed. White rhinos were blamed for locally over grazing the grassland, accentuating soil erosion to the detriment of ecosystem productivity and food for other herbivores. The protected area seemed in danger of becoming a white rhino slum, at the cost of broader biodiversity conservation. My investigations revealed how the nutritious short grasses cultivated by white rhino were preferentially grazed as long as they retained sufficient edible material. Once too little grass remained in these areas, white rhinos shifted their grazing to remaining reserves of taller grass. Instead of being labelled ‘overgrazed’, areas with short grasses are now recognised as productive ‘grazing lawns’, benefiting other grazers favouring short grass as well as white rhinos.10 Grazing lawns cultivated by white rhinos covered 10–30 percent of local landscapes within the Hluhluwe-iMfolozi Park and a similar proportion in parts of Kruger NP.11,12,13
Grazing lawns (Figure 13.2) were first recognised in the context of hippo grazing.14 Short-grass mosaics generated by hippos may extend up to 3 km from water bodies.14,15,16 In western Africa, kob concentrations can maintain local grazing lawns flanking dambos.17 Processes involved in the generation and maintenance of lawn grasslands have been especially well studied in Serengeti, with the south-eastern plains representing one vast, although locally variable, lawn grassland (Figure 13.2C).18 Similar lawn-forming grasslands are cultivated locally by wildebeest herds in Kruger NP on uplands underlain by gabbro bedrock (Figure 13.2D).19 If grazing is excluded, taller grasses soon replace the short grasses.20 Repeated episodes of consumption during the course of the growing season favour short grasses, which are promoted by local concentrations of grazers in places where the latter are more secure from predation.21
The impact of grazing on future grass growth depends on when consumption takes place during the seasonal cycle. Defoliation of tall grasses during the wet season lowers their competitive superiority for light, opening space for shorter grasses to establish. The latter are better adapted to tolerate defoliation during the wet season through having a larger proportion of their leaves close to the soil and thus not readily removed by grazers. However, grass regrowth depends on an adequate supply of soil nutrients, potentially contributed by the dung and urine deposited by the grazing ungulates. If this is missing, less-palatable grasses increase to the detriment of forage quality. Consumption occurring during the dry season has little impact on future grass growth because grasses are dormant then. Although large areas may be reduced to bare soil during the dry season in drought years, grasses can recover rapidly during the following wet season provided the soil remains intact.
Grazing and Fire
Recurrent fires are intrinsic to savannas because of the accumulation of potential fuel in the grass layer during the dry season. However, by grazing down this dry grass and thereby promoting the spread of lawn-forming grasses, large herbivores suppress the extent of fires.10 Fires tend not to spread once the proportion of the landscape covered by short grass exceeds around 40 percent.22 During the early 1960s, about 80 percent of Serengeti grasslands, excluding the short-grass plains, burnt annually.23 Following the expansion of the wildebeest population to 1.3 million animals, the annual proportion of the ecosystem burnt has been halved to around 30 to 40 percent.24 The area burnt is reduced further in drought years because less grass remains ungrazed. Plant material that is digested by herbivores escapes the nitrogen loss that occurs with fire. However, both fire and grazing restrict the build-up of decomposing plant material or humus in the soil, reducing its capacity to retain nutrients against leaching. Nevertheless, nutrients are released more slowly through decomposition of plant detritus in the soil than via dung and urine produced by herbivores.
Surface water distribution affects the spatial distribution of grazing. Areas remote from surface water during the dry season escape heavy grazing during the dry season so that most of the grass remains to be burnt periodically. Areas close to few sources of water get heavily grazed and trampled, perhaps locally to bare soil. Hippos add their grazing pressure within a few kilometres of the water bodies where they seek security during the day, leading to rampant soil erosion in the vicinity. White rhinos cultivate grazing lawns further afield, but need to remain within cruising range of water sources. Thus, grazing and fire in combination generate a broad-scale zonation in their relative vegetation impacts.
Browsers and Bush Encroachment
Heavily grazed areas tend to be invaded by woody plants, resulting in the development of quite dense thickets.25,26 Browsers and mixed feeders like impala27 can potentially suppress the establishment of woody seedlings and saplings through selective defoliation. Juvenile woody plants emerging in grazing lawns and bared areas are exposed to browsing by the lack of grass cover.28 Heavy browsing of taller saplings by giraffe can prevent woody saplings from growing beyond the fire trap (Figure 13.3A).29 Black rhinos heavily prune small saplings (Figure 13.3D). In Serengeti, wildebeest concentrations contribute additionally to suppressing small woody plants through the breakage they impose. However, the most extreme impacts on woody seedlings came from little dikdiks living in tiny territories at high local densities.30
Figure 13.3
Browsing impacts on woody plants. (A) Acacia ‘topiaries’ heavily pruned by giraffe from above; (B) kudus defoliating acacias at middle height; (C) impala browsing at lower height; (D) acacia sapling with stems heavily clipped by black rhino.
Because of the mechanisms that savanna trees and shrubs have to withstand the effects of fires (Chapter 8), browsing may be needed in addition to suppress the growth of the invading woody plants.31 Cohorts of establishment by acacia saplings have been related to windows of opportunity following disease outbreaks that temporarily reduced impala numbers.32
However, the effectiveness of browsers in suppressing woody plant expansion depends on which species these herbivores choose to eat and where and when their browsing pressure is greatest. These considerations prompted my study on factors influencing diet selection by browsing ruminants undertaken in the Nylsvley Nature Reserve. In this moist/dystrophic savanna, browsing by kudus and impalas actually had little effect on the most common woody species prevalent there, because their leaves generally contained condensed tannins or other chemical deterrents.33 The spinescent acacias escaped heavy impact by shedding their leaves quite early in the dry season, while browse remained abundant. Moreover, while small browsers like impala, dikdik and other small antelope can have locally severe impacts on woody seedings, they are patchily distributed.34 They may suppress woody plant invasion locally, but there will be vast areas where they are absent and trees are potentially able to expand, if permitted by fire patterns and other influences. Over wider landscapes, the main agents of woody plant mortality are fire and elephants, separately or in combination.
Elephants and Trees
Elephants are the supreme mixed feeders, breaking branches, toppling trees and uprooting grasses35 (Figures 13.4 and 13.5). They can push over trees up to 60 cm in trunk diameter.36 They kill even bigger trees by bark removal, thereby exposing trees to further damage from fires burning into the trunk and from wood-boring beetles. They uproot shrubs and saplings and dig out and consume roots. By persistently breaking leader shoots, they can prevent tree saplings from growing taller, producing browsing hedges, the counterpart of grazing lawns.37,38 The impacts of elephants on the tree canopy cover can exceed that of fires.39,40 In Serengeti, increasing numbers of elephants are reversing the expansion by woody plants that had followed the suppression of fires by wildebeest grazing.23,41,42 In Kruger NP, the expanding elephant population has opened the tree cover on basaltic soils, while granitic soils have shown much less change.43 Annual rates of tree toppling were amplified sixfold over those shown in fenced areas from which elephants and other herbivores had been excluded.39
Figure 13.4
Elephants feeding by (A) breaking branchlets of mopane sapling and (B) uprooting grasses.
Figure 13.5
Tree damage inflicted by elephants. (A) Mopane trees felled in northern Botswana; (B) bark stripped from giraffe thorn tree, northern Botswana; (C) mopane tree showing bark response to cumulative damage; (D) debarked dead marula tree, Hluhluwe-Mfolozi Park; (E) baobab tree hollowed, Luangwa Valley, Zambia; (F) elephant yanking acacia sapling, Maasai Mara NR, Kenya; (G) sapling left with top broken off.
The most extreme transformation of the tree cover by elephants took place in Murchison Falls NP in Uganda.44 A broad-leaved woodland of clusterleaf and bushwillow trees became converted into a tall, mostly treeless grassland with shrinking forest patches within a few decades by elephants hemmed in by surrounding human settlements at a density exceeding 3 elephants per km2. However, most of these elephants were eliminated during civil warfare in the 1980s. When I visited the park in 2018, the broad-leaved woodland south of the Nile River had become re-established, although the fire regime had not changed (Figure 13.6A). In the Maasai Mara NR in Kenya, elephants and fire in combination transformed an acacia savanna into mostly open grassland, maintained by elephants through their consumption of woody seedlings (Figure 13.6B).45,46 The elimination of the trees was ascribed to a succession of very hot fires during years with exceptionally high rainfall. However, I suspect that elephants had actually been the primary agent of woodland removal, toppling trees after lower-level browse had been eliminated by the severe fires. By continuing to break regenerating saplings, elephants are preventing the woodland from regenerating there. In northern Botswana, elephants locally toppled a large proportion of canopy trees in circumstances after either fire or frost had reduced food available in the shrub layer (Figure 13.6D).36 Woodland destruction of this magnitude may take place episodically where accessible food is lacking, leaving only stumps remaining (Figure 13.6C).
Figure 13.6
Landscapes transformed by elephants. (A) Woodland that has regenerated in place of grassland following elimination of most elephants from Murchison Falls South NP in Uganda; (B) open grassland generated by elephants and fire in combination in Maasai Mara NR; (C) mopane woodland reduced to stumps in northern Botswana; (D) saplings locally broken by elephants following a fire, northern Botswana; (E) vegetation denuded around Aruba Dam in Tsavo NP, Kenya; (F) dense knobbly bushwillow shrubs replacing acacia woodland, Linyanti, Botswana.
Within Tsavo East NP in Kenya, elephants hemmed in by hunting in surrounding areas uprooted or felled wild myrrh shrubs and umbrella thorn trees and almost entirely eliminated baobabs. The area around a large dam became almost completely denuded of vegetation (Figure 13.6E).47,48 Nevertheless, much recovery has since taken place in the woodland following the reduction of the elephant population by illegal hunting.49 In Hwange NP in Zimbabwe where the regional elephant density has reached around 3 animals per km2, elephants have converted areas of broad-leaved woodland into coppice regrowth. Nevertheless, woodlands dominated by deep-rooted Zimbabwe teak and other large trees not utilised by elephants remain little impacted (personal observations). Furthermore, woody species not eaten by elephants are increasing in compensation. Alongside the Chobe River in northern Botswana, high local elephant densities have converted the formerly dense woodland flanking the river into open shrubland.50,51 Further west along the Linyanti River, knob thorn and giraffe thorn trees have been mostly eliminated from the riparian woodland and replaced by knobbly bushwillow (Combretum mossambicensis), a shrub not utilized by elephants (Figure 13.6F).52
Nevertheless, some trees have bark resistant to removal by elephants, like leadwood. Others have very deep roots, and are thus not readily pushed over, like giraffe thorn. Some have foliage rejected by elephants, like African blackwood (Erythrophleum africanum). Although not eaten by elephants, knobbly bushwillow is readily browsed by ruminants.53 Other bushwillow species are consumed by elephants. Such selective utilisation can lead to compositional shifts in the vegetation without overall opening of the woody canopy cover.
However, elephants are water-dependent and need to drink every 2–4 days. This means that the spatial extent of the vegetation transformation that they induce is restricted to <10 km of rivers, lakes or other long-lasting water bodies.54 In northern Botswana, trees pollarded by elephants occur mostly within 5 km of surface water and elephant impacts diminish further away.55 Despite the enormous numbers of elephants present in northern Botswana, extensive mopane and sandveld woodlands remote from water remain little damaged by elephants.56
The habitat transformation brought about by elephants need not be adverse for other large herbivores. Trees get replaced by lower-growing saplings and shrubs, bringing more food within the height reach of medium-large browsers.57 Opening of the tree cover promotes grasses and hence more food for grazers. Nutrients contained within trees get cycled more rapidly than if locked up until trees eventually die of old age. However, the structural and compositional diversity of the woodland can be reduced,52 and birds dependent on tall trees for habitat and for nesting sites, like vultures, may lose out.
Savanna regions were more open than they are today over much of Africa even quite recently. Many wildlife reserves were proclaimed following the resurgence of tsetse flies and the sleeping sickness that they transmit as a consequence of woody thickening, which necessitated the relocation of human settlements. Bush thickening has been widespread throughout southern African over recent decades, except in places retaining elephants.58 The collapse of large herbivore populations during the rinderpest epizootic during the 1890s, coupled with a reduction in fires lit by people, have both been invoked as the causes of the woody plant expansion elsewhere in Africa. I suspect that the main cause of this vegetation transformation was actually the elimination of elephants by hunting for ivory, rampant during the nineteenth century. The woody vegetation cover in the presence of elephants will inevitably be more open than in places lacking these ecosystem engineers, at least within cruising range of surface water.
A further side benefit of elephants should not be overlooked: the network of trails they develop, many leading to long-lasting water sources. Trails formed by white rhinos greatly facilitated my movements on foot in the Hluhluwe-iMfolozi Park. Elephant trails crossing ridges and valleys have served to guide road construction in many parts of Africa.
The combined roles of elephants and fires in suppressing the tree canopy cover in savannas has generally been regarded as adverse. However, the productive potential of savannas lies mainly in the grass layer, which is promoted by both. Trees are needed somewhere as an essential component of savanna heterogeneity. They persist alongside rivers and drainage lines as well as in areas remote from surface water. Africa’s savanna trees and shrubs have evolved along with elephants as well as fires for many millennia. Mopane trees have an amazing capacity to resist elephant impacts by growing back from the stumps of toppled trees. The overall savanna is potentially more productive, and broadly more diverse, in their presence than where their effects are missing. During earlier times, while humans evolved, mega-browsers and mega-grazers were more abundant and diverse than they remained into the Holocene. Their impacts on savanna structure and composition would have been correspondingly greater, at least locally within reach of water.
Termites and Decomposition
The other major bioengineers in savannas altering structure and function physically, besides elephants, are termites. Their activities are undertaken largely underground and thus out of sight. They are the primary agents of the decomposition of plant biomass not eaten by herbivores and not burnt by fires. The termites that feed mainly on woody material are members of the genus Macrotermes. Those dependent mostly on grasses are represented by three genera: Odontotermes, Trinervitermes and Hodotermes. Macrotermes spp. build large earth mounds, Trinervitermes spp. small domed mounds, and Odontotermes spp. low spread-out mounds, while Hodotermes colonies are hidden entirely underground (Figure 13.7).59 Cubitermes spp. construct small domes and digest the soil organic matter that they consume via the agency of cellulose-degrading microbes in their gut. Other mound-building termites cultivate symbiotic fungi within chambers inside their mounds, able to degrade the lignocellulose component of woody plant parts. The termites feed mostly on these fungi. Harvester termites contribute to much grass removal, including some green leaves, especially when drought conditions take hold (Figure 13.7E,F). Large Macrotermes mounds with basal diameters up to 10 m achieve densities of 2–10 mounds per ha (Figure 13.7C), while the much smaller mounds of other species can reach densities of 1000 or more per ha.60,61 Mounds are present more frequently on crests than lower down on slopes.
Figure 13.7
Termite activities. (A) Macrotermes mound in grassland, western extension of Serengeti NP; (B) Macrotermes mound in woodland, Kruger NP; (C) concentration of Macrotermes mounds, Kruger NP; (D) low Odontotermes mound on western edge of plains, Serengeti NP; (E) Hodotermes gathering grass remains into their underground nest; (F) grass cut by harvesting Hodotermes.
Nutrients contained in plant litter get concentrated in the interiors of termite mounds. Termites play a further role in soil engineering by lifting clay from deeper levels to constitute much of the mound, along with mineral nutrients. Concurrently, stone pebbles gravitate downwards to form stone lines in subsoils.
Low Odontotermes mounds support the short grasses favoured by white rhinos and other grazers. The raised eminences of Macrotermes mounds provide refuges for woody plant species sensitive to fire.62,63 Trees growing on mounds attract browsing by black rhinos and elephants because of their relatively palatable foliage compared with the surrounding woodland.64,65
The abundance of termites in African savannas is testified by the many mammals that feed on them, some nearly exclusively. These include aardvarks (Orycteropus afer), pangolins (Manis spp.) and aardwolves (Proteles cristata), while numerous birds are attracted to the winged alates, which fly off in vast swarms following rain. In nutrient-poor savannas, where most of the grass is too fibrous to be digested by ungulates, the total biomass of termites underground can greatly exceed the combined weight of the mammalian herbivores.66,67 This situation gets reversed in nutrient-rich savannas, at least locally.
However, the presence of termites of different kinds can vary quite widely spatially. Mound-building termites need clay particles to build these structures, restricting their presence to regions where there is adequate clay in the soil or subsoil. There is also turnover among mounds, as testified by dispersal of the alates seeking sites to establish colonies. The consequences of such dynamics for herbivores, fire and nutrient recycling have yet to be investigated.
Nutrient Cycling
I have emphasised the recycling of the nutrients contained in plant biomass by large herbivores, specifically nitrogen, phosphorus, potassium and calcium. However, there is another mineral nutrient critically important for herbivores that is made available very differently. It is sodium. Most plants actively exclude sodium in their tissues because it is so avidly sought by herbivores.68 The sodium accumulators include many of the short grasses prevalent in grazing lawns, such as pan dropseed (Figure 6A.4G).69,70 These grasses benefit indirectly through attracting grazing because they are better able to tolerate defoliation than the taller grasses which are able to shade them out. The bare soil exposed by the reduced grass cover promotes evaporation, which concentrates sodium.71 Trampling may contribute further by soil compaction, restricting water infiltration. Grazing ungulates tend to be most abundant in drier savannas where sodium is not leached. Notably, white rhinos were absent historically from cool Highveld grasslands. Unpublished studies revealed that grazing lawns promoted by ruminants in the the Suikerbosrand Nature Reserve south of Johannesburg did not contain the sodium-accumulating grasses prevalent elsewhere. In nutrient-deficient miombo woodlands, sodium tends to accumulate downslope on the margins of grassy dambos, attracting grazers such as roan and sable antelope to these regions. Salt licks formed in other localities may also become sources of attraction for herbivores (Figure 13.8). Termite mounds locally concentrate sodium in soils and hence in plant parts. Elephants tended to favour water bodies containing higher sodium levels in regions underlain by nutrient-poor Kalahari sand.72
Figure 13.8
Sable antelope herd at a salt lick in northern Botswana.
Overview
Large herbivores have fundamental impacts on savanna structure and composition, in interaction with fire. Large grazers cultivate lawn-forming grasses in place of taller tufted grasses, suppressing fires. Browsers of varying sizes help suppress the expansion of woody plants, particularly in the absence of frequent fires. Elephants are major agents of the mortality of trees and shrubs, capable of opening or even removing the tree canopy over large areas, especially in interaction with fire. Concentrations of grazers near surface water can exclude fires through their denudation of the grass cover. Areas remote from surface water burn more frequently and intensely because of the absence of water-dependent grazers. All of these features contribute to the spatial heterogeneity that is an inherent feature of Africa’s savannas. Grass consumption by grazers plus tree felling by elephants accelerates rates of nutrient recycling. This accentuates landscape heterogeneity in interaction with soil moisture, bedrock geology and local fire intensity. In these various ways, herbivores contribute to the spatial heterogeneity that is integral to African savannas. The outcome is the green, brown and black world mosaic that characterises savanna landscapes,73 observable from above via satellite, aircraft or drone imagery. The patch mosaic would appear somewhat different in the absence of large herbivores.
However, much of savanna Africa is not thronged with abundant herbivores. The broad-leaved woodlands typifying moist/dystrophic savanna support much lower densities of big herbivores, except locally in floodplains or around dambos. Soil fertility based on mineral nutrients plays a major role in limiting the regional abundance of large herbivores. Sodium availability may serve as a major restriction on herbivore abundance in moister savannas where soils are leached of this mineral.
Elsewhere in the world, the role of herbivores in restricting the tree cover is less apparent, because most were eliminated by human hunters during the late Pleistocene. Prior to these extinctions, mammoths, bisons and horses contributed to maintaining the grassy ‘mammoth steppe’ that stretched from Siberia through Alaska.74 Megaherbivores opened localities within deciduous forests, enabling fires to penetrate and promote meadows.75 Grazing by mammoths would have enhanced the compositional diversity of tall-grass prairies. Gomphotheres and other megaherbivores formerly inhabiting South America’s cerrado savannas must have had a similarly great impact on the tree cover there. Nevertheless, the tree-damaging effects of the elephants occupying tropical Asian forests seem quite minor.76 In India, spotted deer attain local densities matching those of mixed-feeding impalas in Africa, but abundant grazers are lacking.77 Forest elephants in Africa similarly have little impact on tree saplings, although their grazing and browsing along forest margins may restrict fire penetration by depressing fuel loads.78
Before turning attention to human evolution, the somewhat different assemblages of herbivores and carnivores that coexisted with emerging human lineages further back in time need to be recognised. These paleo-faunas will be the subject of the next chapter.
SUGGESTED FURTHER READING
Owen-Smith, N. (1988) Megaherbivores. The Influence of Very Large Body Size on Ecology. Cambridge University Press, Cambridge.
REFERENCES
1.Briske, DD, et al. (2020) Strategies for global rangeland stewardship: assessment through the lens of the equilibrium–non‐equilibrium debate. Journal of Applied Ecology 57:1056–1067.
2.Bell, RHV. (1982) The effect of soil nutrient availability on community structure in African ecosystems. In Huntley, BJ; Walker, BH (eds) Ecology of Tropical Savannas. Springer, Berlin, pp. 193–216.
3.Fritz, H; Duncan, P. (1994) On the carrying capacity for large ungulates of African savanna ecosystems. Proceedings of the Royal Society of London Series B: Biological Sciences 256:77–82.
4.Rutherford, MC. (1980) Annual plant production–precipitation relations in arid and semi-arid regions. South African Journal of Science 76:53–57.
5.Deshmukh, IK. (1984) A common relationship between precipitation and grassland peak biomass for east and southern Africa. African Journal of Ecology 22:181–186.
6.Deshmukh, I. (1986) Primary production of a grassland in Nairobi National Park, Kenya. Journal of Applied Ecology 23:115–123.
7.Cox, GW; Waithaka, JM. (1989) Estimating aboveground net production and grazing harvest by wildlife on tropical grassland range. Oikos 54:60–66.
8.Strugnell, RG; Pigott, CD. (1978) Biomass, shoot-production and grazing of two grasslands in the Rwenzori National Park, Uganda. The Journal of Ecology 66:73–96.
9.Anderson, TM, et al. (2007) Rainfall and soils modify plant community response to grazing in Serengeti National Park. Ecology 88:1191–1201.
10.Waldram, MS, et al. (2008) Ecological engineering by a mega-grazer: white rhino impacts on a South African savanna. Ecosystems 11:101–112.
11.Arsenault, R; Owen-Smith, N. (2011) Competition and coexistence among short-grass grazers in the Hluhluwe-iMfolozi Park, South Africa. Canadian Journal of Zoology 89:900–907.
12.Cromsigt, JPGM; te Beest, M. (2014) Restoration of a megaherbivore: landscape‐level impacts of white rhinoceros in Kruger National Park, South Africa. Journal of Ecology 102:566–575.
13.Cromsigt, J, et al. (2017) The functional ecology of grazing lawns – how grazers, termites, people, and fire shape HiP’s savanna grassland mosaic. In Cromsigt, JPGM, et al. (eds) Conserving Africa’s Mega-diversity in the Anthropocene: The Hluhluwe-iMfolozi Park Story. Cambridge University Press, Cambridge, pp. 135–160.
14.Olivier, RCD; Laurie, WA. (1974) Habitat utilization by hippopotamus in the Mara River. African Journal of Ecology 12:249–271.
15.Lock, JM. (1972) The effects of hippopotamus grazing on grasslands. The Journal of Ecology 60:445–467.
16.Kanga, EM, et al. (2013) Hippopotamus and livestock grazing: influences on riparian vegetation and facilitation of other herbivores in the Mara Region of Kenya. Landscape and Ecological Engineering 9:47–58.
17.Verweij, R, et al. (2006) Grazing lawns contribute to the subsistence of mesoherbivores on dystrophic savannas. Oikos 114:108–116.
18.McNaughton, SJ. (1983) Serengeti grassland ecology – the role of composite environmental-factors and contingency in community organization. Ecological Monographs 53:291–320.
19.Yoganand, K; Owen‐Smith, N. (2014) Restricted habitat use by an African savanna herbivore through the seasonal cycle: key resources concept expanded. Ecography 37:969–982.
20.McNaughton, SJ. (1985) Ecology of a grazing ecosystem: the Serengeti. Ecological Monographs 55:259–294.
21.Anderson, TM, et al. (2010) Landscape‐scale analyses suggest both nutrient and antipredator advantages to Serengeti herbivore hotspots. Ecology 91:1519–1529.
22.Archibald, S, et al. (2017) Interactions between fire and ecosystem processes. In Cromsigt, JPGM, et al. (eds) Conserving Africa’s Mega-Diversity in the Anthropocene: The Hluhluwe-iMfolozi Park Story. Cambridge University Press, Cambridge, pp. 233–261.
23.Sinclair, A, et al. (2008) Historical and future changes to the Serengeti ecosystem. In Sinclair, ARE, et al. (eds) Serengeti III: Human Impacts on Ecosystem Dynamics. University of Chicago Press, Chicago, pp. 7–46.
24.Eby, S, et al. (2015) Fire in the Serengeti ecosystem: history, drivers, and consequences. In Sinclair, ARE, et al. (eds) Serengeti IV: Sustaining Biodiversity in a Coupled Human–Natural System. University of Chicago Press, Chicago, pp. 73–103.
25.Roques, KG, et al. (2001) Dynamics of shrub encroachment in an African savanna: relative influences of fire, herbivory, rainfall and density dependence. Journal of Applied Ecology 38:268–280.
26.O’Connor, TG, et al. (2014) Bush encroachment in southern Africa: changes and causes. African Journal of Range & Forage Science 31:67–88.
27.Moe, SR, et al. (2009) What controls woodland regeneration after elephants have killed the big trees? Journal of Applied Ecology 46:223–230.
28.Voysey, MD, et al. (2021) The role of browsers in maintaining the openness of savanna grazing lawns. Journal of Ecology 109:913–926.
29.Pellew, RA. (1984) The feeding ecology of a selective browser, the giraffe (Giraffa camelopardalis tippelskirchi). Journal of Zoology 202:57–81.
30.Augustine, DJ; McNaughton, SJ. (2004) Regulation of shrub dynamics by native browsing ungulates on East African rangeland. Journal of Applied Ecology 41:45–58.
31.Staver, AC; Bond, WJ. (2014) Is there a ‘browse trap’? Dynamics of herbivore impacts on trees and grasses in an African savanna. Journal of Ecology 102:595–602.
32.Prins, HHT; van der Jeugd, HP. (1993) Herbivore population crashes and woodland structure in East Africa. Journal of Ecology 81:305–314.
33.Cooper, SM; Owen-Smith, N. (1985) Condensed tannins deter feeding by browsing ruminants in a South African savanna. Oecologia 67:142–146.
34.du Toit, JT; Owen-Smith, N. (1989) Body size, population metabolism, and habitat specialization among large African herbivores. The American Naturalist 133:736–740.
35.Owen‐Smith, N, et al. (2019) Megabrowser impacts on woody vegetation in savannas. In Scogings, PF; Sankaran, M (eds) Savanna Woody Plants and Large Herbivores. Wiley, Oxford, pp. 585–611.
36.Chafota, J; Owen-Smith, N. (2009) Episodic severe damage to canopy trees by elephants: interactions with fire, frost and rain. Journal of Tropical Ecology 25:341–345.
37.Cromsigt, JPGM; Kuijper, DPJ. (2011) Revisiting the browsing lawn concept: evolutionary interactions or pruning herbivores? Perspectives in Plant Ecology, Evolution and Systematics 13:207–215.
38.du Toit, JT; Olff, H. (2014) Generalities in grazing and browsing ecology: using across-guild comparisons to control contingencies. Oecologia 174:1075–1083.
39.Asner, GP; Levick, SR. (2012) Landscape‐scale effects of herbivores on treefall in African savannas. Ecology Letters 15:1211–1217.
40.Pellegrini, AFA, et al. (2017) Woody plant biomass and carbon exchange depend on elephant–fire interactions across a productivity gradient in African savanna. Journal of Ecology 105:111–121.
41.Sinclair, ARE, et al. (2008) Historical and future changes to the Serengeti ecosystem. In Sinclair, ARE, et al. (eds) Serengeti III: Human Impacts on Ecosystem Dynamics. University of Chicago Press, Chicago, pp. 7–46.
42.Morrison, TA, et al. (2016) Elephant damage, not fire or rainfall, explains mortality of overstorey trees in Serengeti. Journal of Ecology 104:409–418.
43.Trollope, WSW, et al. (1998) Long-term changes in the woody vegetation of the Kruger National Park, with special reference to the effects of elephants and fire. Koedoe 41:103–112.
44.Laws, RM, et al. (1975) Elephants and Their Habitats. Clarendon Press, Oxford.
45.Dublin, HT, et al. (1990) Elephants and fire as causes of multiple stable states in the Serengeti–Mara woodlands. The Journal of Animal Ecology 59:1147–1164.
46.Dublin, HT. (1991) Dynamics of the Serengeti–Mara woodlands: an historical perspective. Forest & Conservation History 35:169–178.
47.Agnew, ADQ. (1968) Observations on the changing vegetation of Tsavo National Park (East). African Journal of Ecology 6:75–80.
48.Leuthold, W. (1977) Changes in tree populations of Tsavo East National Park, Kenya. African Journal of Ecology 15:61–69.
49.Leuthold, W. (1996) Recovery of woody vegetation in Tsavo National Park, Kenya, 1970–94. African Journal of Ecology 34:101–112.
50.Mosugelo, DK, et al. (2002) Vegetation changes during a 36‐year period in northern Chobe National Park, Botswana. African Journal of Ecology 40:232–240.
51.Skarpe, C, et al. (2014) Plant–herbivore interactions. In Skarpe, C, et al. (eds) Elephants and Savanna Woodland Ecosystems. Wiley, Oxford, pp. 189–206.
52.Teren, G, et al. (2018) Elephant‐mediated compositional changes in riparian canopy trees over more than two decades in northern Botswana. Journal of Vegetation Science 29:585–595.
53.Makhabu, SW. (2005) Resource partitioning within a browsing guild in a key habitat, the Chobe Riverfront, Botswana. Journal of Tropical Ecology 21:641–649.
54.Chamaillé‐Jammes, S, et al. (2007) Managing heterogeneity in elephant distribution: interactions between elephant population density and surface‐water availability. Journal of Applied Ecology 44:625–633.
55.Sianga, K, et al. (2017) Spatial refuges buffer landscapes against homogenisation and degradation by large herbivore populations and facilitate vegetation heterogeneity. Koedoe 59:a1434.
56.Ben-Shahar, R. (1998) Changes in structure of savanna woodlands in northern Botswana following the impacts of elephants and fire. Plant Ecology 136:189–194.
57.Rutina, LP, et al. (2005) Elephant Loxodonta africana driven woodland conversion to shrubland improves dry-season browse availability for impalas Aepyceros melampus. Wildlife Biology 11:207–213.
58.Stevens, N, et al. (2017) Savanna woody encroachment is widespread across three continents. Global Change Biology 23:235–244.
59.Gosling, CM, et al. (2012) Effects of erosion from mounds of different termite genera on distinct functional grassland types in an African savannah. Ecosystems 15:128–139.
60.Goudie, AS. (1988) The geomorphological role of earthworms and termites in the tropics. In Viles, H (ed.) Biogeomorphology. Blackwell, Oxford, pp. 43–82.
61.Davies, AB, et al. (2014) Spatial variability and abiotic determinants of termite mounds throughout a savanna catchment. Ecography 37:852–862.
62.Mobæk, R, et al. (2005) Termitaria are focal feeding sites for large ungulates in Lake Mburo National Park, Uganda. Journal of Zoology 267:97–102.
63.Davies, AB, et al. (2016) Termite mounds differ in their importance for herbivores across savanna types, seasons and spatial scales. Oikos 125:726–734.
64.Holdo, RM; McDowell, LR. (2004) Termite mounds as nutrient-rich food patches for elephants. Biotropica 36:231–239.
65.Loveridge, JP; Moe, SR. (2004) Termitaria as browsing hotspots for African megaherbivores in miombo woodland. Journal of Tropical Ecology 20:337–343.
66.Ferrar, P. (1982) Termites of a South African savanna. IV. Subterranean populations, mass determinations and biomass estimations. Oecologia 52:147–151.
67.Deshmukh, I. (1989) How important are termites in the production ecology of African savannas? Sociobiology 15:155–168.
68.Borer, ET, et al. (2019) More salt, please: global patterns, responses and impacts of foliar sodium in grasslands. Ecology Letters 22:1136–1144.
69.Seagle, SW; McNaughton, SJ. (1992) Spatial variation in forage nutrient concentrations and the distribution of Serengeti grazing ungulates. Landscape Ecology 7:229–241.
70.Griffith, DM, et al. (2017) Ungulate grazing drives higher ramet turnover in sodium‐adapted Serengeti grasses. Journal of Vegetation Science 28:815–823.
71.Stock, WD, et al. (2010) Herbivore and nutrient control of lawn and bunch grass distributions in a southern African savanna. Plant Ecology 206:15–27.
72.Weir, JS. (1972) Spatial distribution of elephants in an African National Park in relation to environmental sodium. Oikos 23:1–13.
73.Bond, WJ. (2005) Large parts of the world are brown or black: a different view on the ‘Green World’ hypothesis. Journal of Vegetation Science 16:261–266.
74.Zimov, SA, et al. (2012) Mammoth steppe: a high-productivity phenomenon. Quaternary Science Reviews 57:26–45.
75.Owen-Smith, N. (1987) Pleistocene extinctions: the pivotal role of megaherbivores. Paleobiology 13:351–362.
76.Mueller-Dombois, D. (1972) Crown distortion and elephant distribution in the woody vegetations of Ruhuna National Park, Ceylon. Ecology 53:208–226.
77.Karanth, KU; Sunquist, ME. (1992) Population structure, density and biomass of large herbivores in the tropical forests of Nagarahole, India. Journal of Tropical Ecology 8:21–35.
78.Cardoso, AW, et al. (2020) The role of forest elephants in shaping tropical forest–savanna coexistence. Ecosystems 23:602–616.