The heart of vegetation science is an understanding of spatial variation and temporal change in vegetation. This variation and change in turn provides a template for ecosystem functioning. Spatial variation in African savanna vegetation at a continental level has been related to climate and geology (Bell 1982), and at a regional level to topo-edaphic variation and land forms (see Mucina & Rutherford 2006) that can be further modified by fire and large mammalian herbivores (Frost et al. 1986). Biotic influences such as termites can exert pronounced local-scale effects (Sileshi et al. 2010; Jouquet et al. 2011).
Van der Plas et al. (2013) sought to examine whether Macrotermes mounds had an expected influence on woody vegetation composition and structure, and consequently on foraging by mammalian browsers in a mesic African savanna. These savannas are synonymous with the wildlife that they support; at least two dedicated journals address the ecology of large African mammals. By comparison, a recent review of the ecological effects of termites lists 133 references (Jouquet et al. 2011) – about one volume of the Journal of Vegetation Science. However, the biomass of termites within African savanna is in the order of 70–110 kg·ha−1 (Wood & Sands 1978), whereas that of ungulates is 10–80 kg·ha−1 (Bell 1982). Termites' ubiquitous influence is demonstrated by mound densities of up to six per hectare, foraging that covers most of an area, and the removal of up to 15 000 kg·ha−1 of litter to a colony per annum (Dangerfield et al. 1998). Any redress of this disparity in research effort is to be welcomed.
That termite mounds affect vegetation is well known (Jouquet et al. 2011), but their importance for foraging by animals as diverse as elephants (Holdo & McDowell 2004) and ants (Brody et al. 2010) has been recognized only recently (Table 1). Mound-building using subsoil results in clays being concentrated near the soil surface, greater exchange capacity, and thus improved concentrations of N, P and bases in foliage. Browsers continually face the challenge of finding sufficient food of adequate quality during the dry season when most woody plants have shed their leaves, while evergreen species are commonly less palatable (Owen-Smith & Cooper 1987). At this time, herbivores need to seek out key forage resources from ‘hotspots’ within their landscape.
Table 1. Environmental context of studies reporting that termite mounds create a foraging hotspot for browsers in African savannas
|Location||Mean annual rainfall (mm)||Parent geology/Soils; vegetation||Termite species||Browsers selecting for mounds||Source|
|Botswana: Okavango||520||Kalahari sands plus alluvium|| Macrotermes michaelseni ||Impala||McCarthy et al. (1998)|
|Kenya: Laikipea||550||‘Black cotton’ vertisol soils||Odontotermes sp.||Eland, Steenbuck, Grant's gazelle||Brody et al. (2010)|
|South Africa: Hluhluwe-iMfolozi park||720–950||Shale, Sandstone, Dolerite, Basalt|| Macrotermes natalensis ||None||Van der Plas (2013)|
|South Africa: Kruger NP||700||Granite-derived sandy soils; Broad-leaved woodland||Macrotermes sp.||Elephant, Giraffe, Kudu, Impala||Levick et al. (2010a,b)|
|Uganda: Lake mburo NP||800||Soils: Ferrasols, Histosols, Vertisols, Leptosols|| Macrotermes herus ||Impala||Mobaek et al. (2005)|
|Zimbabwe: Hwange||650||Kalahari Sands; Baikiaea plurijuga Woodland|| Macrotermes michaelseni ||Elephant||Holdo & McDowell (2004)|
|Zimbabwe: Iwaba||697||Granite-Derived sandy soils; Miombo||Macrotermes sp.||Elephant, Black Rhinoceros||Loveridge & Moe (2004)|
Against this background, Van der Plas et al. undertook what was intended to be a confirmatory study that Macrotermes mounds in the Hluhluwe-iMfolozi Park, South Africa, function as islands of high nutrient and water availability and low fire frequency that support distinct communities of trees, which are attractive to browsers. A novelty was that interpretation of vegetation differences between mounds and adjacent savanna was based on plant traits. They selected 22 traits that reflected the environmental demands of irradiance, availability of water and nutrients, and browsing pressure. Mounds were well dispersed across the landscape, thus allowing for an effective sampling design of a mound paired with adjacent matrix. Feeding preferences of browsing herbivores for woody species were quantified.
As expected, mounds and the matrix differed in both species composition and plant traits. There were more evergreen and broad-leaved, but fewer legume, tree species on mounds, whereas the matrix had more deciduous, fine-leaved and thorny species. Dominant species on mounds were less mechanically defended, less nutritious, had larger leaves and grew more rapidly than the species dominating matrix areas. Traits of mound species suggest they are adapted for light competition (high growth rates, large leaves) but not for drought and fires. In essence, Acacia species dominated within the matrix but were scarce to absent on mounds; these species determined the difference in traits of species growing on mounds vs. adjacent savanna. In contrast, the pattern of usage by browsers was not as expected. Browsers behaved as browsers should, they sought out plants with high foliage concentrations of N and P, but these were the legume tree species that were rare on mounds, despite their clay-enriched soils.
This exception to previous observation highlights the importance of environmental context for comparative studies. The potential for mounds to enhance nutrient availability depends on the texture of the surrounding substrate. Clay enrichment of surface soils by termites occurs far less when the surface soil is already clay-rich (Fig. 2b in Sileshi et al. 2010). Browsing herbivores potentially experience higher nutritional limitations in dystrophic systems (Bell 1982), in which a hotspot such as a termite mound should play a greater role. Empirical studies to date, which have shown termite mounds serve as foraging hotspots, have mainly been conducted in relatively dystrophic systems on sandy soils (Table 1). As Van der Plas et al. comment, most soils within the Hluhluwe-iMfolozi Park are relatively fertile as a consequence of being derived from siltstones, mudstones and base-rich volcanic rocks (Whateley & Porter 1983). Another exception is the effect of mounds on the palatability of Acacia drepanolobium on volcanic-derived vertic soils of the Laikipea plateau (Brody et al. 2010), but this vegetation is in effect a monoculture of this species.
These findings suggest two avenues of research deserving of greater attention. If the influence of termite mounds depends on environmental context, then questions are begged about the distribution and importance of mounds across large environmental gradients. Mounds may exhibit specific habitat relationships within a local landscape (Dangerfield et al. 1998), but their distribution and abundance throughout African savannas is not well understood. The largest spatial scale analysed to date, using LIDAR, is 50 000 ha of granitic landscape within the Kruger National Park in South Africa (Levick et al. 2010a,b). Macrotermes mounds occurred at all landscape positions within the semi-arid regions, but were restricted to the crests and upper slopes in the mesic areas. Similar insights are wanting for an extended rainfall gradient and for heavy textured soils. A second avenue is increased understanding of the population dynamics of these species and their mounds. Does the 20-yr colony longevity for Macrotermes bellicosus in a West African savanna (Collins 1981) apply to the other 165 species of Macrotermes across Africa (Sileshi et al. 2010)? Are mounds a shifting mosaic across a landscape patch so that soils of an entire area might eventually be altered, or is it commonplace that recolonization of mounds occurs continuously (Moore & Picker 1991) so that ‘superhot spots’ are maintained indefinitely?
The effect of these ecosystem engineers on the vegetation of African savannas deserves more attention, but focused effort on a single agent needs to be balanced with holistic studies that address all purported agents affecting vegetation of a single locality.