Chapter 6
How might we classify the variability in structure and composition inherent within savannas? Reviewing savanna vegetation formations world-wide, Cole1 recognised the following structural divisions: (1) open parkland with widely scattered trees, (2) denser woodland with fairly tall and more closely spaced trees, (3) low tree and shrub savanna with a fairly dense woody canopy, (4) grassland lacking trees, and (5) thicket with little grass (Figure 6.1). The rangeland classification developed for eastern African savannas distinguishes woodland, bushland, wooded grassland, bushed grassland, dwarf shrubland and pure grassland, based on tree cover and height.2,3 The structural classification of vegetation developed by UNESCO applies the following definitions: (1) forest – a continuous stand of trees taller than 10 m with interlocking crowns; (2) woodland – an open stand of trees at least 8 m tall with woody cover >40 percent and a field layer of grasses; (3) bushland – an open stand of trees or shrubs 3–8 m tall with woody cover >40 percent; (4) thicket – a closed stand of trees or shrubs usually 3–8 m tall; (5) shrubland – an open or closed stand of shrubs up to 2 m tall; (6) wooded grassland – tree cover however tall of 10–40 percent; and (7) grassland – woody cover <10 percent. Almost all of these divisions focus attention on the woody plant cover rather than features of the grass layer. None are readily mapped at continental scale because patches of forest, varying tree cover and open grassland are commonly interspersed within local landscapes. Variation in grass structure and composition exists at even finer scales within these structural formations defined by tree cover.
Figure 6.1
Structural savanna formations in Africa. (A) Broad-leaved miombo (Brachystegia) woodland with fairly tall trees, Malawi; (B) bushwillow (Combretum)–clusterleaf (Terminalia) woodland of medium height, Kruger NP; (C) open parkland of paperbark thorn (Acacia sieberiana) and sausage trees (Kigelia africana) in Kidepo NP, Uganda; (D) low tree savanna mainly of fine-leaved acacias, Samburu, Kenya; (E) open grassland on shallow soils, Serengeti Plains, Tanzania; (F) subtropical thicket, Addo NP, South Africa, distinct from savanna because of the lack of much grass.
Within southern Africa’s savanna biome, six ‘bioregions’ were distinguished.4 These were named (1) Central Bushveld, located in the north; (2) Lowveld, situated below the escarpment in the east; (3) Sub-Escarpment Savanna, present in parts of KwaZulu-Natal in the south-east; (4) Mopane Woodland, located in lower-lying regions of the north; (5) Eastern Kalahari Bushveld; and (6) Kalahari Duneveld. Note that the Afrikaans word ‘veld’ recognises the presence of a field layer, i.e. grass. Furthermore, the Kalahari region is regarded as arid savanna, not desert, because its MAR exceeds 250 mm, which is adequate to support occasional fires in wet years. Its aridity for animals, and people, is due especially to the lack of surface water because of the deep sand cover. Subtropical thicket is interpreted as a distinct biome and not as a form of savanna, due to its rather patchy grass cover. The grassland biome, separated from savanna, encompasses (1) Mesic Highveld, in the east; (2) Dry Highveld, in the west; (3) Montane Grassland, below the Drakensberg escarpment; and (4) Sub-Escarpment Grassland, elsewhere in KwaZulu-Natal and the Eastern Cape, including coastal regions. Temperate grassland including heath-like shrubs prevails at elevations exceeding 2800 m on top of the Drakensberg/Maloti escarpment in Lesotho, and similar grassy vegetation is represented on hillcrests within the Nama-Karoo biome.
For eastern Africa, the broadest mapped vegetation zone was ‘steppe’, represented by a mixture of grass, ‘brush’ and thicket (Figure 6.2). It grades into ‘savanna grassland’ towards the west where the rainfall is higher and into ‘woodland savanna’ in the south-west and south. In western Africa, a broad zonation in the woody vegetation cover reflects the rainfall gradient diminishing inland from the coast. The Guinean zone takes the form of a forest–savanna mosaic in the south, replaced by broad-leaved savanna woodlands in the Sudanian zone and by an open cover of low trees, mainly acacias, in the Sahelian region, bordering the Sahara Desert (Figure 6.3). Sudanian savanna formations extend westward into Uganda and southern Ethiopia, mostly as open woodland with broad-leaved trees along with tall thatch grasses.
Figure 6.2
Vegetation mosaic in eastern Africa.
Units labelled deciduous forest, woodland, savanna grassland and steppe represent forms of savanna (http://exploringafrica.matrix.msu.edu/wp-content/uploads/2016/05/Vegetation-of-East-Africa-Map.png).
Figure 6.3
Western African savanna formations.
Guinean, Sudanian and Sahelian savanna types form bands across the diminishing rainfall gradient from the coastal forest to the Sahara desert (eros.usgs.gov/westafrica/sites/default/files/inline-images/160823_Climate_zones.jpg)
Climatic Controls: Total Annual Rainfall Versus Dry Season Duration
Climatically, tropical savanna formations span an extreme range in MAR from 1750 mm bordering the transition into rainforest down to as little as 200 mm in the driest parts of the Kalahari and where the Sahel grades into the Sahara Desert.5 The defining feature is the duration and intensity of the dry season. Savanna prevails where the dry season lasts five months or longer spanning the cooler months of the year.5,6 In south-central Africa, dry season rainfall diminishes to as little as 5 percent of the MAR and the dry season extends over seven months (see Table 1.1). In eastern Africa near the equator, the dry season is less intensely dry and as much as 25 percent of the annual rainfall total may be received during five dry season months.
The tree canopy cover tends to increase with rainfall, but there is much variation among savanna sites in the actual woody plant cover exhibited locally (Figure 6.4).6,7 Below a rainfall threshold of around 650 mm, the maximum tree cover recorded diminishes with decreasing MAR. Above this threshold, the local tree canopy can potentially reach 80 percent, but generally is maintained below 40 percent by fire and other influences.8 Once the MAR exceeds 1000 mm, two alternative states can prevail: either (a) a densely wooded savanna, or (b) open grassland interspersed with forest patches.9 The latter tends to prevail at altitudes above 1800 m where temperatures get cooler: for example, below the Drakensberg escarpment, on the Nyika Plateau in Malawi, in the eastern highlands of Zimbabwe, and along the Eastern Arc Mountains in Tanzania (Figure 6.5). The broad-leaved deciduous woodlands known as miombo prevail through south-central Africa where hot conditions occur from August through November before the rainy season begins. Sudanian savanna formations replace it north of the equator under similar climatic regimes. Once forest trees become established, the dense shade they cast restricts the herbaceous cover and hence fire penetration.10 If the forest canopy gets opened, by people or elephants, the evergreen forest trees have little capacity to resist being burnt and may get eliminated.
Figure 6.4
Woody vegetation cover in relation to mean annual precipitation (MAP) recorded at various sites within savanna regions of Africa. Diminishing rainfall restricts the tree cover below 650 mm, while above this threshold the tree cover is widely variable.
(from Sankaran et al. (2005) Nature 438: 846–849)
Figure 6.5
Grassland–forest mosaics. (A) Lope NP, Gabon, with forest pockets in lowlands (MAR 1750 mm); (B) Nyika Plateau, Malawi, with forest patches on mid-slopes (MAR 1200 mm); (C) Drakensberg foothills, South Africa, with forest patches on slopes and lowlands (MAR 1050 mm); (D) Maasai-Mara, Kenya, with forest flanking Mara River (MAR 1000 mm); (E) Eastern Highveld grassland, MAR 700 mm; (F) mostly treeless grassland in the Bateke plateau in south-eastern Gabon, underlain by deep sandy soils derived from Kalahari sand (MAR 1500 mm).
A potential division thus exists between moist savannas, with MAR exceeding 650 mm, where broad-leaved woodlands prevail, and dry savannas with MAR less than 650 mm, where rainfall limits the cover of mainly fine-leaved trees. But the woody plant cover is controlled not only by seasonal moisture – the nature of the soil also influences the kinds of trees and grasses that predominate, as I will explain next.
Geological Substrate and Soil Fertility
Soil fertility is governed by the capacity of soil particles to retain the mineral nutrients needed for plant growth. This is determined largely by the clay fraction, as outlined in Chapter 5. The clay component is derived from the geological substrate, with mafic volcanic rocks contributing much clay and felsic granitic rocks forming rather sandy soils. The clay contributed by sedimentary rocks depends on whether they were consolidated from mud, silt or sand. The bedrock contributes also to the mineral nutrient pool, while the humus component helps retain nutrients, especially nitrogen, derived from organic matter, not rocks.
The effective fertility is influenced also by rainfall. Water percolating through the soil leaches nutrients and weathers the clay content to kaolinite with lowered cation-holding capacity. Where rainfall is low, water penetrates less deeply into soils and mineral nutrients get retained near the surface. With high amounts of rainfall, soil fertility depends largely on the organic matter input, whatever the geological substrate. This leads to a ‘double-barrelled’ subdivision between dry and/or ‘eutrophic’ and moist and/or ‘dystrophic’ savanna forms, reflecting the interaction between rainfall and the bedrock geology.11 Clay-rich soils are typically ‘well-nourished’ (‘eu-trophic’) for plants, while sandy soils provide poor nourishment (‘dys-trophic’). South African cattle ranchers recognise the implications for the performance of their livestock, distinguishing ‘sweet bushveld’, capable of supporting cattle year-round, from ‘sour bushveld’, where livestock lose condition during the dry season.12 It is in the middling range of MAR between 500 and 1000 mm, prevalent through much of Africa’s savanna biome, that bedrock geology is most influential. With higher rainfall, nutrients get intensely leached, whatever the soil parent material. With lower rainfall, water does not penetrate far and nutrients get retained (Box 6.1).
Box 6.1What Governs the Distinction in Soil Fertility Between Dry/Eutrophic and Moist/Dystrophic Savannas?
Soil fertility controls how fast plants can potentially grow when water is sufficient. It is governed by nutrients like nitrogen recycled through decaying organic matter, basic elements like potassium, magnesium and calcium held on clay particles and phosphorus contributed from degrading rock material. Fertiliser applied to crops and gardens consists basically of nitrate, phosphate and potassium salts (NPK).
The bedrock geology supplies the mineral elements and phosphate (PO4–), but more importantly provides the clay particles that hold cations, including nitrogen in the form of ammonia (NH4+), against leaching by water percolating through the soil. Once generated, the pool of both mineral and organic nutrients gets recycled mostly through the soil organic matter or humus content via growth and decay of plants, below- as well as above-ground, plus animal contributions in the form of dung, urine and dead bodies. However, the humus content of African savanna soils is generally low,14 because primary productivity is restricted by low rainfall and plant litter is decomposed rapidly by soil microbes in hot environments. Moreover, there are leakages. Much of the nitrogen is volatilised in smoke during fires, and some nitrogen in the form of nitrate (NO3–) is lost through leaching. However, phosphates are highly insoluble and can contribute to maintaining high fertility for a long time in sites enriched by settlements where animal bones accumulate.15
The capacity of clay particles to hold cations like K+ and Ca+ depends on the mineral form of the clay. Clay is constituted by alternating layers of silica and aluminium oxides, which develop negative surface charges attracting cations. These charges are best developed where the layers form a 2:1 lattice, called smectite or illite. After they get weathered to a 1:1 lattice (kaolinite, the substance of talcum powder) this capacity is greatly reduced, and mineral nutrients are held more precariously in the organic matter.
Legumes in the subfamilies Mimosoideae (including the acacias) and Papillionoideae (including beans) can reclaim nitrogen from the atmosphere through the agency of bacteria housed in root nodules. However, this biological nitrogen fixation is energetically costly, and beneficial only if phosphorus is sufficiently available to complement the nitrogen gained.16 Where phosphorus is deficient, in sandy soils derived from granitic rocks or sandstone that typically underlie miombo woodlands, no benefit would be derived from nitrogen fixation.
In African savannas, more nutrients get recycled via clay particles and less through soil organic matter than in temperate latitudes, which is why the bedrock geology has such a great influence on fertility. Where rainfall is high, all soils become reduced to deeply weathered ‘oxisols’ with little capacity to retain nutrients. Accordingly, in tropical forests most recycling take place via the mulch of leaves and other plant litter on the surface before it enters the soil. The link between soil fertility and geology in African savannas is thus via (1) the clay amount and form generated by the bedrock type, which determines the cation exchange capacity (ability to retain nutrients), and (2) the content of mineral nutrients, especially P, in the parent rock material.
The subdivision between dry/eutrophic and moist/dystrophic savanna forms is associated with distinctions in the predominant tree species. Dry/eutrophic savannas are typified by the prevalence of fine-leaved legumes formerly grouped in the genus Acacia within the subfamily Mimosoideae (see Box 6.2). Moist/dystrophic savannas prevalent in south-central Africa are dominated by broad-leaved legumes in the subfamily Caesalpinioideae (recently revised to Detariodeae), especially the genera Brachystegia, Julbernadia and Isoberlinia (Appendix 6.1 lists common and scientific names of species mentioned in the text along with their families). North of the equator in the Sudanian zone, broad-leaved trees in the family Combretaceae (Combretum and Terminalia spp.) tend to predominate and grow also in the understorey of miombo woodlands. Then there is the anomalous mopane (Colophospermum mopane), also a member of the Caesalpinioideae, which attains almost monospecific dominance in low-lying hot and dry areas in southern Africa, especially but not exclusively on clay soils. While the predominant trees are deciduous in both savanna subdivisions, they retain leaves longer into the dry season in moist/dystrophic savannas than in dry/eutrophic savannas.
Box 6.2Status of the Genus Acacia
The fine-leaved legumes in the Mimosoideae that were grouped in the genus Acacia have recently become partitioned between newly erected genera Vachellia and Senegalia for the African representatives, with the generic label Acacia confined to the Australian wattles. Species now subsumed within Vachellia have fairly straight spines developed from stipules located at the base of leaves, while the species placed in Senegalia have recurved prickles located at nodes along stems. The former produce yellow or cream pom-pom flowers, while the latter produce white or cream catkins. Species labelled Vachellia tend to flower throughout the wet season, while those in Senegalia generally flower early, before leaf flush. The former produce either indehiscent pods containing very hard seeds that fall to the ground intact, or dehiscent pods that split to shed their seeds, while the latter produce only dehiscent pods.
While there are these valid distinctions between these subgroups, the ecological unity among the African acacias has thereby been obscured. It would have been more meaningful ecologically had the distinctions that exist among the African acacias been recognised only at subgeneric level. Furthermore, the new generic assignments were erected irregularly by reassigning the type species, and so remain contentious. In this book, I will retain the original assignment of Africa’s fine-leaved thorn trees to the genus Acacia. Common names together with scientific names and other features of all plant species mentioned are listed in the Appendix at the end of this book.
The rainfall influence alone was explored along a transect consistently on infertile Kalahari sand stretching from the Northern Cape of South Africa through Botswana into Zambia.13 There was a transition from acacia trees to bushwillow or mopane once MAR exceeded 400 mm, and a variety of broad-leaved leguminous trees became predominant above 600 mm MAR. The prevalent grass types also changed, from aristoid (three-awn or needle grasses, Aristida spp.) at the dry end to panicoid (buffalo grass Panicum spp. and finger grass Digitaria spp.) in the middle, and andropogonoid species (thatch grasses Hyparrhenia spp.) in the wettest north.
The soil effect is strikingly apparent in South Africa’s Kruger NP. The eastern region underlain by basalt supports knob thorn (Acacia nigrescens) parkland, under MAR regimes between 550 and 650 mm, while the western region on granitic gneiss carries a broad-leaved bushwillow woodland under similar rainfall. Where gabbro or dolerite sills intrude within the granite, acacias once again predominate. In the Serengeti ecosystem in Tanzania where soils are widely volcanically enriched, acacia savanna extends into regions with MAR exceeding 1000 mm in the north. Red grass (Themeda triandra) features in the herbaceous layer on clay-rich soils, while thatch grasses and various love grasses are prevalent on soils with a higher sand content.
Broad-leaved savanna woodlands known as miombo are typically associated with sandy soils derived from basement granite, sandstone or Kalahari sand, mostly in regions with MAR exceeding 800 mm. Where soils are highly infertile, broad-leaved savanna formations occur even under MAR regimes as low as 600 mm. Examples include woodlands dominated by wild seringa (Burkea africana) on sandstone-derived soils in the Nylsvley Nature Reserve and elsewhere in the ‘sour’ bushveld of South Africa; and places with Zimbabwe teak (Baikiaea plurijuga) on deep Kalahari sand in northern Botswana and adjacent parts of Zimbabwe where MAR is similarly low.17 Sudanian savannas are associated with regions where soils became intensely weathered down to iron concretions.18 This is derived from when they were located close to the climatic equator while Africa was drifting northward after the breakup of Gondwana.
Phosphorus seems to be the key soil nutrient governing the prevalence of fine-leaved acacias. Nitrogen availability becomes most limiting where the available phosphorus is adequate. Legumes in the Mimosoideae overcome the nitrogen limitation via a symbiotic relationship with bacteria housed in nodules on their roots.19 These bacteria are capable of fixing atmospheric nitrogen, subsidised energetically by carbohydrates from the host plant. The phosphorus influence was evident in the Nylsvley Nature Reserve where the prevalent broad-leaved woodland gave way to patches dominated by acacia trees on sites of former human occupation.20 In these places, the phosphorus content in the sandy soil had been enriched by wood-burning and bone accumulations. The grassland composition also changed from dominance by tall and highly fibrous broom love-grass (Eragrostis pallens) to a prevalence of love grasses (other Eragrostis spp.) and blue buffalo grass (Cenchrus ciliaris) in the acacia patches.
Thickets of low-growing trees, commonly evergreen, occur patchily within savanna regions, either on deeper soils that retain moisture longer into the dry season or on rocky hills. Only in South Africa are thickets sufficiently extensive to be regarded as forming a distinct biome, occurring in a region of the Eastern Cape where rainfall is low but distributed year-round and deep sandy soils retain moisture. Once formed, thickets resist fires because of the sparse grass cover. These climatically generated thickets are compositionally distinct from thickets that have developed due to fire exclusion or lack of grass cover brought about by heavy grazing.21
Soil Water: Topo-hydrology
The amount of moisture effectively available for plant growth depends not simply on the rainfall, but also on how rainwater is redistributed within landscapes. The deluge during thunderstorms can exceed the infiltration capacity of the soil, so that a substantial fraction of the water gravitates down slope, eroding and leaching in the process.22 How much seeps in or runs off depends on the soil texture as well as the soil depth. Uplands tend to be driest, while lowlands accumulate water and soil, especially finer clay particles. The resultant topo-sequence linking vegetation formations is known as a catena. In general, tree cover tends to be least and grasses shortest on uplands where soils are shallowest, while woodland or forest flanks river channels23 (Figure 6.6A). However, this pattern gets reversed in broad-leaved savannas, especially on sandy soils. Tree cover becomes densest on crests, while zones of treeless grassland, known locally as dambos, mbugas or vleis, occupy the broad valleys where soils are seasonally waterlogged (Figure 6.6B–D). Seep zones may also promote open grassland on lower slopes where water gets forced to the surface over bedrock or a hardpan. Sandier soils allow greater infiltration and retain moisture for longer than compacted clay soils, while the latter dry more intensely. Consequently, sandy soils tend to support a denser tree cover than clay-rich soils. Soil depth may be restricted by hardpan layers in the subsoil.
Figure 6.6
Topographic sequences of vegetation formations within local landscapes. (A) Open grassy upland grading into shrub cover near drainage lines, Serengeti NP, Tanzania; (B) grassy seep zone with tall grassland in mid-slope and clusterleaf woodland on crests, south-west Kruger NP; (C) grassy dambo intersecting miombo woodland, Zambia; (D) wide dambo grassland flanked by broad-leaved woodland, north-west Serengeti NP.
The open short-grass plain in south-eastern Serengeti is edaphically controlled by a calcrete hardpan developed from volcanic ash and tuff deposited by nearby volcanoes.24 With increasing rainfall westward, soils get deeper and the vegetation shifts from treeless grassland into acacia savanna with a taller grass cover. Zones of open grassland more generally are associated with reduced rates of water infiltration into soils.22,25 Tree cover decreases with increasing intensity of rainfall on clay soils.26 Various forms of hardpan can restrict the rooting depths of trees elsewhere in Africa. On ancient land surfaces in South Africa’s Highveld region, as well as further north in Africa, erosion over many millions of years has formed ferricrete (or plinthite), silcrete or calcrete hardpans at shallow depths, inhibiting tree establishment.27,28 Shrubs occur on rocky outcrops where soil moisture is concentrated between the rocks and fires are deflected.
Quite extensive grasslands are found on floodplains and former lake beds where clay seals restrict water penetration, as shown in the Savuti and Mababe depressions in northern Botswana.29 Besides the Okavango Delta, extensive floodplains border sections of the Kafue River and occur in the Bangweulu Swamp in Zambia, below Gorongoza Mountain on the edge of the rift in Mozambique and in the Katavi region of western Tanzania. These sump grasslands represent hotspots of nutrient enrichment within predominantly nutrient-deficient woodlands. Termite mounds raised within the floodplain allow trees to escape waterlogging in local clumps. Under considerably drier conditions in the southern Kalahari, pan depressions and infilled valleys have likewise remained treeless due to clay seals. Tree establishment may also be inhibited on very deep sands, as seems to be the case on the Bateke Plateau in Congo Brazzaville and adjoining Gabon underlain by ancient deposits of Kalahari sand (Figure 6.5F). The influential factor in such localities may be the difficulty for woody seedlings to access water at depth.30
Combined influences from the redistribution of rainwater, bedrock geology and hydrology generate a complex mosaic of vegetation patterns, most extremely between treeless grassland and closed forest. The composition of the grass layer also responds to soil depth and texture, with shortest grasses prevailing on uplands and tallest grass in lowlands and wetlands. These effects are modified additionally by the recurrent fires promoted by the seasonal dryness.
Fire Regimes: Recurrent Incineration
Recurrent fires are an intrinsic feature of savannas due to the seasonal drying of the grass layer. The impact of these fires on the vegetation cover depends on the frequency of the burns, the intensity of the fires and the stage within the seasonal cycle when the fires occur.31 Currently, the fire return interval is typically 2–3 years across a range in MAR from 700 to 1200 mm. In regions with lower rainfall than this, there is less fuel to support the spread of fires and burns occur mostly during unusually wet years, perhaps at quite long intervals.
However, intervals between fires tend to vary quite widely locally. Some places may escape being burnt for many years, due to local soil moisture or grass removal by herbivores. Within Serengeti NP, some localities got burnt twice annually, while others escaped being burnt for 10 years or longer, mostly on the heavily grazed plains.32,33 Regular annual fires promote an open tree canopy, without necessarily affecting the density of woody plants, because small trees and shrubs can remain hidden in the grass layer. With protection from fire, the increased tree cover can shift the grass layer towards shade-loving grasses less supportive of fires, reinforcing the woody thickening trend.34 Variable intervals between burns may enable juvenile trees to grow beyond the flame zone. With long intervals, moribund grass accumulated over several years can promote fires hot enough to kill tree saplings several metres tall plus some canopy trees. Fires occurring in drier savannas, although less intense, can still be quite destructive to woody plants that lack adequate protective features.
Higher rainfall promotes greater grass growth and hence hotter fires (Figure 6.7). Weather conditions at the time of the fire are also influential.35 Extremely hot ‘firestorms’ are generated by 30:30:30 conditions: air temperature >30°C, relative humidity <30 percent and wind speed >30 km/h.36 Head fires fanned by wind are hot but brief, and temperatures at the soil surface may not rise much. Back burns into the wind can be more destructive to grass tufts as well as tree saplings, because flames remain longer. Intense ‘firestorms’ can kill even established trees by burning through their protective bark, especially if some bark had been removed by elephants, porcupines or other animals.
Figure 6.7
Savanna fires. (A) Hot fire burning tall savanna grassland, Kruger NP (photo: E. Le Roux); (B) gentle fire burning grass layer in miombo woodland, Luangwa Valley; (C) fire burning into forest margin, Gabon; (D) herbivores grazing green-up following burning, Serengeti NP.
Fires can get ignited by lightning when the first thunderstorms occur while the grass still remains dry. If the onset of the rains is delayed, burns may extend into the wet season months, and thus be especially detrimental to trees that had flushed new leaves. However, humans have been major agents of fire ignition for at least 300,000 years in Africa, perhaps even much earlier during the Pleistocene.37 Hunter-gatherers lighting fires to clear impeding grass and attract animals to the post-burn flush of green grass would have set them earlier in the dry season than those caused by lightning. After humans acquired livestock, early dry season burns may have been used to improve the nutritive value of the grass cover for these domestic herbivores during the dry season. In moist miombo woodlands in Zambia, early season burns promote an open understorey coupled with pockets of resistant evergreen thicket, while annual late burns cause the death of some canopy trees, replaced by coppice regeneration.38,39 Wetland grasslands ranging from dambos to floodplains maintained by waterlogging can support fires after floodwaters have retreated and grasses have dried out.
Regular burns help maintain a sharp boundary between grassland and forest patches in grassland–forest mosaics.40 Shrubs bordering forest margins may form partial barriers to fire penetration into the forest. Recurrent annual or semi-annual fires contribute to the mosaic of open grassland and stunted woodland interspersed with forest patches in regions of montane grassland.41,42
Recurrent fires thus help maintain an open woody canopy within the savanna biome, especially in moister regions. The suppression of fires by European settlers, combined with heavy stocking with domestic grazers, has contributed to the widespread expansion in woody vegetation cover within savanna regions of southern Africa43 and in parts of eastern Africa44 over recent decades. A further factor may have been the post-industrial rise in atmospheric carbon dioxide levels, favouring woody plants over grasses.45,46 However, the susceptibility of the grass cover to being burnt is modified by the activities of large herbivores, both grazers and browsers. These interactions will be considered within an ecosystem context in Part III of the book.
River Corridors
River corridors contribute importantly to vegetation patterns within savannas. Regions bordering river channels, from headwater gullies to broad river basins, have more fertile soils and retain soil moisture longer into the dry season than the surrounding savanna matrix, favouring woodland or forest development. Fires get suppressed not only because of wetter soils and greater shading, but also because heavy grazing and trampling by herbivores concentrating near water reduces the fuel load.47 Tall trees characteristic of river margins include sycamore figs (Ficus sycomorus), ana trees (Faidherbia albida), jackalberry (Diospyros mespilliformis), leadwood (Combretum imberbe) and mahogany (Trichilia emetica) (see Appendix 6.1). Fever trees (Acacia xanthophloea) may form groves in regions subject to seasonal flooding. Stands of reeds (Phragmites spp.) develop in sandy river beds.
Periodic floods can remove even quite large trees bordering the river channel and redistribute the areas covered by water, sand, reeds or forest.48 Hence the current state of the vegetation alongside or within the river channel depends largely on the time elapsed since the last big flood. Pools in rivers support hippos and attract concentrations of grazers and browsers to feed on the locally greener vegetation as well as to access water for drinking, increasing the spatial heterogeneity within the riparian zone.
Open Grasslands: Why No Trees?
Regions with predominantly open grassland are mapped as a distinct biome in South Africa, but not generally elsewhere in the continent. Various factors have been proposed to explain the lack of trees, but no consensus has been reached on their relative importance. Temperatures sufficiently cool to generate winter frosts have been invoked, but grasslands without trees are not restricted to highlands where frosts are frequent. Highveld climates do not prevent shrubs from growing on rocky outcrops, where they gain some protection from fire. Grassland and forest are interspersed in a mosaic bordering the rainforest biome and even in coastal regions, where frosts never occur. Fires burn intensely hot through open grasslands, especially when fanned by winds, but the presence of a few trees with fire-resistant bark does little to reduce the intensity, which depends mainly on the grass fuel load. Soil depth on eroding uplands and elsewhere where hardpan layers have formed may block deeper penetration by tree roots27; but open grasslands exist on deep sands in the Bateke Plateau in Congo-Brazzaville and on the Mozambican coastal plain. The lack of trees may even be a legacy of the last glacial maximum when atmospheric levels of CO2 dropped too low to support much woody plant growth, but there should have been adequate time since then for trees to spread back.
Overview
The savanna biome is defined by the presence of a grass layer sufficiently dense to support recurrent fires over most of the landscape. The local tree cover is climatically allied with a dry season period when plant growth ceases. Geological influences on soil fertility, modified by rainfall, underlie a functional subdivision between fine-leaved dry/eutrophic savannas, characterised by thorny acacias, and broad-leaved moist/dystrophic savannas where other leguminous trees predominate. Treeless grasslands are found on cooler uplands and elsewhere where shallow soils restrict rooting depth, although not restricted to such conditions. A forest–grassland mosaic prevails adjoining the rainforest zone. The spatiotemporal redistribution of soil moisture contributes to spatial heterogeneity in the woody plant canopy cover. Variable fire return intervals further modify the tree and grass cover.
The tropical savanna biome is less extensive in other continents. South American savanna formations, locally known as ‘cerrado’, prevail in a fairly large block south of the equator on the Brazilian highlands and in a smaller area just north of the equator in Venezuela, known as the llanos, separated by the forested Amazon basin. Australia has an arc of savanna in its north, running from Western Australia through Northern Territory into Queensland. In parts of India and south-east Asia, vegetation formations mapped as ‘tropical scrub forest’ and ‘tropical deciduous forest’ have been allied with savanna forms because of their substantial grass cover.49 Savanna vegetation extends into regions with higher rainfall in other continents than in Africa: up to 2500 mm in South America and 2000 mm in Australia, compared with 1750 mm in Africa.5 Fire incidence has less influence on the woody plant cover in both of these southern continents than in Africa.50
The dry/eutrophic savanna subdivision seems to be exclusive to Africa. South American cerrado found on Brazil’s Parana plateau is associated with a range in MAR from 1100 to 1900 mm.51,52,53 The tree cover there gives way to open grassland towards higher elevations, disrupted by shrubby trees on rocky hills, but the underlying bedrock, whether sedimentary sandstone or mafic basalt/diabase, has little effect on vegetation structure or composition. Soils are generally highly weathered, degrading the clay content to kaolinite with little cation-holding capacity.54 Trees retaining evergreen foliage are intermingled among those shedding their leaves during the dry season. Grasses are generally low in nutritional value and include a substantial component of species with the C3 photosynthetic pathway (see Chapter 7). The grassy llanos occupies a low-lying plain in Venezuela and adjoining parts of Colombia where MAR ranges from 1200 mm up to 2750 mm. Over much of its extent the open aspect is maintained by seasonal flooding, despite the 5–6 months dry season. Smaller blocks of grassland–forest mosaic emerge from the Amazon rainforest near the border between Brazil, Venezuela and Guyana. The closest approach to African savannas is the mix of dry woodland, thicket and grassland called ‘chaco’ found in Paraguay and adjoining countries. In north-eastern Brazil, a dry spiny thicket or low forest with stem-succulents prominent called ‘caatinga’ replaces cerrado where MAR lies between 450 and 1000 mm with a less distinctive dry season.55 The Pantanal in south-western Brazil is a vast wetland grassland, although some areas do burn during the dry season. Well south of the tropics, savanna grades into temperate ‘pampas’ grassland and steppe in Argentina.
Australia exhibits a lower tree cover relative to MAR than Africa, perhaps due to tree mortality during prolonged extreme droughts.56 Australian savanna formations are dominated mostly by evergreen or semi-evergreen gum trees (Eucalyptus spp.).57 Although wattle trees (various Acacia spp.) tend to predominate where soils are less infertile, there is no clear division between moist savanna and dry savanna. Trees assigned to the genus Acacia there lack thorns and many have phyllodes (expanded stalks) functioning as leaves. Treeless grassland occurs in the Barkly Tableland under MAR as low as 350 mm on vertisols derived from limestone, which exclude trees by swelling and shrinking in response to moisture variation. Other areas of open grassland occur locally on basalt substrates. The form of red grass (Themeda sp.) growing in Australia lacks the ability of the African species to support grazing and African grasses have been introduced to promote the livestock industry.
Within India and parts of south-east Asia, vegetation formations resembling the moist savannas of Africa are present in uplands of the Western Ghats on basalt and more locally elsewhere.49 Dipterocarp woodlands with a grassy understorey, which does burn periodically, occur amid closed-canopy forest under MAR levels approaching 2000 mm in continental south-east Asia.1 Dry areas with thorn trees prominent prevail in parts of north-west India, but lack sufficient grass cover to be called savanna. Dense human occupation plus a policy of actively suppressing fires has perhaps obscured savanna affinities with Africa in these regions.
Thus, tropical savannas outside Africa are generally rather wet and mostly extremely dystrophic. Within Australia recurrent fires burn the woody canopy as well as the grass layer due to the flammability of the prevalent eucalypt trees. Fires seem less influential in tropical Asia, perhaps because they are generally suppressed, while dry seasons there are less dry than within Africa and Australia.
To establish how grasses exclude trees, we need to explore the mechanisms of competition between these plant forms, taking place below-ground as well as above-ground. This is the topic to be addressed in the next chapter. Before pitting trees against grasses, you need to be reminded that individual species of each of these plant types differ in their ecologies. Profiles of some of the prominent species, plus pictures, are presented in Appendix 6.1.
SUGGESTED FURTHER READING
Bond, WJ. (2019) Open Ecosystems. Ecology and Evolution Beyond the Forest Edge. Oxford University Press, Oxford.
Cole, M. (1986) The Savannas. Biogeography and Geobotany. Academic Press, New York.
Huntley, BJ; Walker, BH. (1982) Ecology of Tropical Savannas. Springer, Berlin.
Lehmann, CER, et al. (2011) Deciphering the distribution of the savanna biome. New Phytologist 191:1970209.
McClenahan, TR; Young, TP. (1996) East African Ecosystems and their Conservation. Oxford University Press, Oxford.
REFERENCES
1.Cole, MM. (1986) The Savannas. Biogeography and Geobotany. Academic Press, New York.
2.Pratt, DJ; Gwynne, M. (1977) Rangeland Management and Ecology in East Africa. Hodder and Stoughton, London.
3.Grunblatt, J, et al. (1989) A hierarchical approach to vegetation classification in Kenya. African Journal of Ecology 27:45–51.
4.Mucina, L; Rutherford, M. (2006) The Vegetation of South Africa, Lesotho and Swaziland. South African National Biodiversity Institute, Pretoria.
5.Lehmann, CER, et al. (2011) Deciphering the distribution of the savanna biome. New Phytologist 191:197–209.
6.Good, SP; Caylor, KK. (2011) Climatological determinants of woody cover in Africa. Proceedings of the National Academy of Sciences of the United States of America 108:4902–4907.
7.Sankaran, M, et al. (2005) Determinants of woody cover in African savannas. Nature 438:846–849.
8.Sankaran, M, et al. (2008) Woody cover in African savannas: the role of resources, fire and herbivory. Global Ecology and Biogeography 17:236–245.
9.Staver, AC, et al. (2011) History matters: tree establishment variability and species turnover in an African savanna. Ecosphere 2:1–12.
10.Charles-Dominique, T, et al. (2018) Steal the light: shade vs fire adapted vegetation in forest–savanna mosaics. New Phytologist 218:1419–1429.
11.Huntley, BJ. (1982) Southern African savannas. In Huntley, BJ; Walker, B (eds) Ecology of Tropical Savannas. Springer, Berlin, pp. 101–119.
12.Ellery, FN. (1995) The distribution of sweetveld and sourveld in South Africa’s grassland biome in relation to environmental factors. African Journal of Range and Forage Science 12:38–45.
13.Scholes, R, et al. (2002) Trends in savanna structure and composition along an aridity gradient in the Kalahari. Journal of Vegetation Science 13:419–428.
14.Crowther, TW, et al. (2019) The global soil community and its influence on biogeochemistry. Science 365:eaav0550.
15.Augustine, DJ. (2003) Long-term, livestock-mediated redistribution of nitrogen and phosphorus in an East African savanna. Journal of Applied Ecology 40:137–149.
16.Hogberg, P. (1986) Nitrogen-fixation and nutrient relations in savanna woodland trees (Tanzania). Journal of Applied Ecology 23:675–688.
17.Childes, SL; Walker, BH. (1987) Ecology and dynamics of the woody vegetation on the Kalahari sands in Hwange National Park, Zimbabwe. Vegetatio 72:111–128.
18.Spinage, CA. (1988) First steps in the ecology of the Bamingui‐Bangoran National Park, Central African Republic. African Journal of Ecology 26:73–88.
19.February, EC, et al. (2019) Physiological traits of savanna woody species: adaptations to resource availability. In Scogings, PF; Sankaran, M (eds) Savanna Woody Plants and Large Herbivores. Wiley, Oxford, pp. 309–329.
20.Blackmore, AC, et al. (1990) The origin and extent of nutrient-enriched patches within a nutrient-poor savanna in South-Africa. Journal of Biogeography 17:463–470.
21.Bond, WJ, et al.(2017) Demographic bottlenecks and savanna tree abundance. In Cromsigt, JPG, et al. (eds) Conserving Africa’s Mega-Diversity in the Anthropocene. Cambridge University Press, Cambridge, pp. 161–188.
22.Jager, T. (1982) Soils of the Serengeti woodlands, Tanzania. PhD thesis, Wageningen University, Wageningen.
23.Colgan, MS, et al. (2012) Topo-edaphic controls over woody plant biomass in South African savannas. Biogeosciences 9:1809–1821.
24.de Wit, HA. (1978) Soils and grassland types of the Serengeti Plains (Tanzania). PhD thesis, Wageningen University, Wageningen.
25.Holdo, RM, et al. (2020) Spatial transitions in tree cover are associated with soil hydrology, but not with grass biomass, fire frequency, or herbivore biomass in Serengeti savannahs. Journal of Ecology 108:586–597.
26.Case, MF; Staver, AC. (2018) Soil texture mediates tree responses to rainfall intensity in African savannas. New Phytologist 219:1363–1372.
27.Tinley, K. (1982) The influence of soil moisture balance on ecosystem patterns in southern Africa. In Huntley, BJ; Walker, BH (eds) Ecology of Tropical Savannas. Springer, Berlin, pp. 175–192.
28.O’Connor, TG; Bredenkamp, GJ. (1997) Grassland. In Cowling, RM, et al. (eds) Vegetation of Southern Africa. Cambridge University Press, Cambridge, pp. 215–257.
29.Sianga, K; Fynn, R. (2017) The vegetation and wildlife habitats of the Savuti–Mababe–Linyanti ecosystem, northern Botswana. Koedoe 59:1–16.
30.Knoop, WT; Walker, BH. (1985) Interactions of woody and herbaceous vegetation in a southern African savanna. The Journal of Ecology 73:235–253.
31.Archibald, S, et al. (2013) Defining pyromes and global syndromes of fire regimes. Proceedings of the National Academy of Sciences of the United States of America 110:6442–6447.
32.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.
33.Probert, JR, et al. (2019) Anthropogenic modifications to fire regimes in the wider Serengeti–Mara ecosystem. Global Change Biology 25:3406–3423.
34.Higgins, SI, et al. (2007) Effects of four decades of fire manipulation on woody vegetation structure in savanna. Ecology 88:1119–1125.
35.Govender, N, et al. (2006) The effect of fire season, fire frequency, rainfall and management on fire intensity in savanna vegetation in South Africa. Journal of Applied Ecology 43:748–758.
36.Browne, C; Bond, W. (2011) Firestorms in savanna and forest ecosytems: curse or cure? Veld & Flora 97:62–63.
37.Archibald, S, et al. (2012) Evolution of human-driven fire regimes in Africa. Proceedings of the National Academy of Sciences of the United States of America 109:847–852.
38.Trapnell, CG. (1959) Ecological results of woodland and burning experiments in Northern Rhodesia. The Journal of Ecology 47:129–168.
39.Smith, P; Trapnell, C. (2002) Chipya in Zambia: a review. Kirkia 18:16–34.
40.Titshali, LW, et al. (2000) Effect of long-term exclusion of fire and herbivory on the soils and vegetation of sour grassland. African Journal of Range and Forage Science 17:70–80.
41.Oliveras, I; Malhi, Y. (2016) Many shades of green: the dynamic tropical forest–savannah transition zones. Philosophical Transactions of the Royal Society B – Biological Sciences 371.
42.Walters, G. (2012) Customary fire regimes and vegetation structure in Gabon’s Bateke Plateaux. Human Ecology 40:943–955.
43.Stevens, N, et al. (2017) Savanna woody encroachment is widespread across three continents. Global Change Biology 23:235–244.
44.Sinclair, ARE, et al. (2007) Long‐term ecosystem dynamics in the Serengeti: lessons for conservation. Conservation Biology 21:580–590.
45.Bond, WJ; Midgley, GF. (2000) A proposed CO2‐controlled mechanism of woody plant invasion in grasslands and savannas. Global Change Biology 6:865–869.
46.Buitenwerf, R, et al. (2012) Increased tree densities in South African savannas: >50 years of data suggests CO2 as a driver. Global Change Biology 18:675–684.
47.Smit, IP; Archibald, S. (2019) Herbivore culling influences spatio‐temporal patterns of fire in a semiarid savanna. Journal of Applied Ecology 56:711–721.
48.Rountree, M, et al. (2000) Landscape state change in the semi-arid Sabie River, Kruger National Park, in response to flood and drought. South African Geographical Journal 82:173–181.
49.Ratnam, J, et al. (2016) Savannahs of Asia: antiquity, biogeography, and an uncertain future. Philosophical Transactions of the Royal Society B – Biological Sciences 371.
50.Lehmann, CER, et al. (2014) Savanna vegetation–fire–climate relationships differ among continents. Science 343:548–552.
51.Eiten, G. (1982) Brazilian ‘savannas’. In Huntley, BJ; Walker, BH (eds), Ecology of Tropical Savannas. Springer, Berlin, pp. 25–47.
52.Sarmiento, G. (1984) The Ecology of Neotropical Savannas. Harvard University Press, Cambridge, MA.
53.Borghetti, F. (2020) South American savannas. In Scogings, PF; Sankaran, M (eds) Savanna Woody Plants and Large Herbivores. Wiley, Oxford.
54.Medina, E; Silva, JF. (1990) Savannas of northern South America: a steady state regulated by water–fire interactions on a background of low nutrient availability. Journal of Biogeography 17:403–413.
55.Bucher, EH. (1982) Chaco and Caatinga – South American arid savannas, woodlands and thickets. In Huntley, BJ; Walker, BH (eds) Ecology of Tropical Savannas. Springer, Berlin, pp. 48–79.
56.Fensham, RJ, et al. (2005) Rainfall, land use and woody vegetation cover change in semi-arid Australian savanna. Journal of Ecology 93:596–606.
57.Williams, RJ, et al. (1997) Leaf phenology of woody species in a North Australian tropical savanna. Ecology 78:2542–2558.
Appendix 6.1Some Tree and Grass Species Typical of African Savannas
Plant species matter ecologically, not just the broad types labelled trees and grasses. Let me introduce you to some of their notable characteristics, in text and in pictures.
All of the acacias are thorny as well as possessing finely divided leaves (Figure 6A.1). Umbrella thorn (Acacia tortilis) is the most widely distributed, found throughout Africa and broadly tolerant of different soil substrates. Other acacias have distinct soil associations. Knob thorn (Acacia nigrescens) is typical of basaltic or doleritic clay soils, but is also present in bottomlands of granitic landscapes where clay accumulates. Giraffe thorn (Acacia erioloba) is characteristic of arid savannas on Kalahari sand and flanks dry river beds, accessing soil moisture deep below. Whistling or ant-gall thorn (Acacia drepanolobium) occurs in shrubby form on heavy clays, while black monkey thorn (Acacia burkei) favours quite sandy soils and common hook-thorn (Acacia caffra) grows on rocky hillsides. White thorn (Acacia polyacantha) is usually found on alluvial soils near rivers. Paperbark thorn (Acacia sieberiana) grows both in alluvial lowlands and in moister savannas elsewhere. Sweet thorn (Acacia karroo) is restricted to South Africa, growing either as a spreading tree or as a pole-like or cage-like shrub. Black thorn (Acacia mellifera) grows as a multi-stemmed shrub forming thickets in dry regions of southern and eastern Africa. Most acacias shed their leaves quite early in the dry season, but splendid (or stinkbark) thorn (Acacia robusta) retains leaves until the end of the dry season, while ana trees (Faidherbia albida) carry leaves throughout the dry season and shed them at the start of the wet season. Both of the latter species commonly flank rivers. Acacia species also differ in their adaptations to withstand fires.1 Thus, each of the acacia species that I know has distinctive features of its ecology – there are no neutrally equivalent niches.
Figure 6A.1
Some representative acacias. (A) Umbrella thorn widespread through Africa, Luangwa, Zambia; (B) giraffe thorn, typically found in arid savanna, central Kalahari, Botswana; (C) paperbark thorn, especially resistant to fires, Kidepo, Uganda; (D) splendid thorn, typically in river margins, Luangwa Valley, Zambia; (E) knob thorn, favouring clay soils, Hluhluwe-iMfolozi Park, South Africa; (F) fever trees, typical of swampy soils, Kruger NP, South Africa.
Broad-leaved savanna woodlands feature numerous species placed within the legume subfamily Caesalpinioideae (recently revised to Detariodeae).2 Three species of Brachystegia and one species of Julbernadia represent miombo woodland at its southern limit in Zimbabwe. Zebrawood or msasa (Brachystegia spiciformis) is ubiquitous on comparatively shallow but well-drained sands from Zimbabwe to the Congo (Figure 6A.2). Prince of Wales feathers or mfuti (Brachystegia boehmi) is also widely distributed, favouring shallow infertile soils. Mountain acacia (Brachystegia tamarindoides) is associated with rocky hills. Munondo (Julbernadia globiflora) is associated with relatively dry conditions at low altitudes, often on rocky slopes with thin soils. Representatives of other genera within this subfamily include wild seringa (Burkea africana) and Zimbabwe teak (Baikiaea plurijuga), which are locally common on sandy soils in drier regions. Mutondo (Isoberlinea angolensis) is characteristic of wetter forms of miombo woodland from Zambia into western Africa. In this heartland, 12 or more species in the miombo group may be intermingled without obvious differences in their ecology. All of these broad-leaved legumes evidently contain chemical deterrents against being munched by browsing ungulates. Miombo woodland trees characteristically produce new leaves several weeks before the early rains, coloured red or yellow by anthocyanin precursors of tannins.
Figure 6A.2
Some miombo woodland and allied trees. (A) Msasa, the most widespread species, Zambia; (B) Brachystegia spp., early leaf flush, Zambia; (C) munondo, found on shallower soils, Zimbabwe; (D) Zimbabwe teak, typical of deep Kalahari sands, Zimbabwe; (E) wild seringa, found in drier sandy soils than typical miombo, Magaliesberg in South Africa; (F) tall mopane woodland, predominant in hot dry lowlands of south-central Africa, Luangwa Valley, Zambia.
Then there is, once again, the enigmatic mopane (Colophospermum mopane; Figure 6A.3). Despite being a broad-leaved member of the Caesalpinioideae, it occurs in hot and dry low-lying regions in an east–west swathe from northern South Africa through Botswana into Namibia and Angola, but no further north than the Luangwa Valley in Zambia, achieving monospecific dominance over much of this range. Mopane can grow either as a woodland of tall trees or a low tree shrubland, dependent on soil depth governed by calcrete hardpans. The grass cover within mopane woodlands is commonly sparse with annuals predominating, reducing the effects of fires. Mopane woodlands in south-central Africa currently support most of Africa’s elephants.
Figure 6A.3
Some widely distributed broad-leaved trees. (A) Widely spread marula tree, Kruger NP; (B) widely distributed sausage tree, Luangwa Valley, Zambia; (C) semi-evergreen jackalberry tree on termite mound, Northern Botswana; (D) clusterleaf tree (Terminalia trichopoda), Serengeti, Tanzania; (E) enormous baobab succulent, Luangwa Valley, Zambia; (F) huge sycamore fig tree, Ndumo GR, South Africa.
Bushwillows in the Combretaceae are commonly found in the understorey of miombo woodlands as well as predominating on sandy soils in regions too dry to support the broad-leaved legumes. Typical genera are Combretum (bushwillows) and Terminalia (clusterleafs), some growing as low trees and others as quite tall trees. A giant among them is leadwood (Combretum imberbe), growing in locations with deep soil water. Its wood is especially heavy and thus resistant to decay, and specimens can live for over a thousand years.
Other broad-leaved trees are not restricted to either savanna form. Marula (Sclerocarya birrea) trees (Figure 6A.3) are widely distributed in both dry/eutrophic and moist/dystrophic savanna and dominate the tree biomass in some regions. Baobab trees (Adansonia digitata) have enormous succulent trunks and are typical of drier regions. Species commonly represented in riverine woodlands, but not restricted to them, include jackalberry (Diospiros mespiliformis) and sausage tree (Kigelia africana). Species of wild myrrh (Commiphora spp.) are a common constituent of dry savannas in eastern Africa, while the evergreen shepherd’s tree (Boscia albitrunca) is revered for shade as well as forage in arid savanna regions.
African savannas support an impressive diversity of grass species forming fine-scale spatial mosaics.3,4 Tribal divisions are associated with ecological distinctions. Grasses in the tribe Andropogoneae tend to grow tall and stemmy and thereby promote hot fires (Figure 6A.4).5,6 Typical representatives are thatch grasses falling within the genera Hyparrhenia and Hyperthelia. They prevail under relatively high rainfall conditions and resist being heavily grazed on account of their high fibre contents. Red grass, sole representative of the genus Themeda, is widely dominant on clay-rich soils. It retains adequate nutritional value for livestock into the dry season, at least in drier areas, making it characteristic of sweetveld. It is promoted by frequent fires. The tribe Paniceae includes grasses highly palatable to grazers, but less supportive of fires, notably species in the genera Panicum (Guinea grass and others), Digitaria (finger grasses), Urochloa (bushveld signal grasses) and Setaria (bristle grasses). Guinea grass (Panicum maximum) is commonly prominent under tree canopies where soils are enriched in nitrogen from leaf litter deposited by the trees. The tribe Chlorideae includes lawn-forming grasses in the genera Cynodon (couch grasses) and Sporobolus (dropseed grasses), as well as numerous species of Eragrostis (love grasses). Pan dropseed (Sporobolus ioclados) attracts grazing by accumulating sodium in its leaves. Couch grass (Cynodon dactylon) is widely cultivated as lawn cover as well as for animal fodder, but can become toxic through a high cyanide content. Three-awn or needle grasses in the tribe Aristideae, including the genera Aristida and Stipagrostis, are most commonly found in arid savannas or even semi-deserts. They have narrow or rolled leaves. Then there are the wetland grasses tolerant of seasonal inundation, including the genera Leersia and Oryza within the tribe Oryzeae (wild rices), associated with tall sedges (Cyperus papyrus), rushes (Typha spp.) and reeds (Phragmites spp.). Genera with aromatic oils inhibiting their consumption by grazers include stinking grasses (Bothriochloa spp.) and lemon grasses (Cymbopogon spp.). All grasses contain silica bodies in their leaves, believed to inhibit grazing by increasing tooth wear of ungulate herbivores. Various grass genera and even species are shared between eastern and southern Africa7,8 and some African grasses have become distributed worldwide.9
Figure 6A.4
Grassland types. (A) Red grass grassland, Serengeti NP, Tanzania; (B) red grass grassland, Kruger NP, South Africa; (C) thatch grass (Hyparrhenia rufa) grassland, Kidepo, Uganda; (D) tall thatch grass, Lope NP, Gabon; (E) Guinea grass, Kruger NP, South Africa; (F) finger grass grassland, Botswana; (G) pan dropseed grass, Kalahari, Botswana; (H) mixed short grass lawn with flowering Ammocharis lily, Mfolozi GR; (I) finger grass (Digitaria macroblephara) showing connecting stolons (runners), Serengeti NP, Tanzania.
Other plant forms found in savannas show various adaptations to cope with recurrent fires as well as herbivory. Shrubs, i.e. multi-stemmed woody plants not growing much taller than 3 m, retain sufficient underground reserves in roots and tubers to regenerate stems and leaves lost during fires.10 Geoxyles growing with their woody stems mostly underground and merely their branch tips bearing leaves protruding are a feature of sandy soils (Figure 6A.5).11 They evade the recurrent loss of woody growth to fires, but need to have their exposed leaves defended against large herbivores by deterrent chemicals. Most genera with geoxyles also have species that grow into tall trees, including Brachystegia, Erythrina, Lannea, Parinari, Dichapetalum and Combretum.
Figure 6A.5
Underground trees or geoxyles. (A) Pygmaeothamnus zeyheri, Magaliesberg; (B) Lannea edulis, Magaliesberg; (C) Parinari capensis, Magaliesberg; (D) highly toxic gifblaar Dichapetalum cymosum, Nylsvley, South Africa; (E,F) unidentified geoxyles, Gabon.
Flowering herbs apart from grasses and sedges commonly get lumped as forbs. These include dwarf shrubs remaining under 0.5 m in height along with herbaceous shoots growing from woody roots, tubers or bulbs below ground. Creepers and woody vines ascend from the herbaceous layer using the support provided by trees and grasses. Forbs and shrublets in the pea family (e.g. Indigofera spp.) make an important contribution to soil fertility through nitrogen fixation in their root nodules, although this capacity seems to diminish as plants mature.12 Forbs of various forms typically contribute most of the taxonomic diversity in the herbaceous layer of savannas. Although forming only a minor component of the vegetation, plants with underground storage organs contributed importantly to the survival of hominins through the dry season.
Then there are plants with succulent stems or leaves, commonly associated with but not exclusive to arid environments, such as aloes and certain euphorbias. Some of these succulents attain tree stature, notably the candelabra euphorbia (Euphorbia ingens), which is distributed widely from Uganda into southern Africa. Baobab trees are effectively enormous succulents with their swollen stems storing moisture in the arid regions they occupy.
REFERENCES
1.Staver, AC, et al. (2012) Top‐down determinants of niche structure and adaptation among African acacias. Ecology Letters 15:673–679.
2.Smith, P; Allen, Q. (2004) Field Guide to the Trees and Shrubs of the Miombo Woodlands. Royal Botanic Gardens Kew, Richmond, pp. 132–133. Includes a picture.
3.McNaughton, SJ. (1983) Serengeti grassland ecology – the role of composite environmental factors and contingency in community organization. Ecological Monographs 53:291–320.
4.Augustine, DJ. (2003) Spatial heterogeneity in the herbaceous layer of a semi-arid savanna ecosystem. Plant Ecology 167:319–332.
5.Hempson, GP, et al. (2019) Alternate grassy ecosystem states are determined by palatability–flammability trade-offs. Trends in Ecology & Evolution 34:286–290.
6.Archibald, S, et al. (2019) A unified framework for plant life-history strategies shaped by fire and herbivory. New Phytologist 224:1490–1503.
7.Fynn, R, et al. (2011) Trait–environment relations for dominant grasses in South African mesic grassland support a general leaf economic model. Journal of Vegetation Science 22:528–540.
8.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 JPG, et al. (eds) Conserving Africa’s Mega-diversity in the Anthropocene: the Hluhluwe-iMfolozi Park Story. Cambridge University Press, Cambridge, pp. 135–160.
9.Linder, HP, et al. (2018) Global grass (Poaceae) success underpinned by traits facilitating colonization, persistence and habitat transformation. Biological Reviews 93:1125–1144.
10.Wigley, BJ, et al. (2009) Sapling survival in a frequently burnt savanna: mobilisation of carbon reserves in Acacia karroo. Plant Ecology 203:1.
11.Maurin, O, et al. (2014) Savanna fire and the origins of the ‘underground forests’ of Africa. New Phytologist 204:201–214.
12.February, EC, et al. (2019) Physiological traits of savanna woody species: adaptations to resource availability. In Scogings, PF; Sankaran, M (eds) Savanna Woody Plants and Large Herbivores. John Wiley & Sons, Oxford, pp. 309–329.