By
J.I. House & D.O. Hall
Division of Life Sciences, King's College London

Savanna structure: tree-grass mix
Plant function
Tree-grass interaction

The relative abundance of woody and herbaceous species is highly dependent on environmental conditions and seasonal and interannual variations. Plant available moisture (a function of climate, soil type and topography) seems to be the key determinant of the tree-grass balance, but structure, function and species composition are also altered by available nutrients, fire and herbivory, and these are all discussed later (Walker & Noy-Meir, 1982; Menaut et al, 1985; Solbrig, 1990; Scholes & Hall, 1996; Solbrig et al, 1996a; Scholes & Archer, 1997; Scholes et al, 1997).

Within this naturally dynamic ecosystem, small environmental/management changes may be buffered in the short-term, but in the long term can lead to shifts in species and even functional types. Larger changes can lead to rapid transformations in vegetation structure which may go beyond certain thresholds into a new structural domain from which it is hard to return (Menaut et al, 1985). For example, fire exclusion and heavy grazing promote increases in tree cover, grasses are suppressed, and there is no fire thereby leading to dense thickets resistant to fire. Conversely, beyond a certain high fire and low grazing threshold large amounts of grass lead to regular hot fires destroying all seedlings eventually excluding trees.

Functionally, arid and humid savannas are very different (Walker, 1985). Arid savannas have more affinity with semi-desert vegetation in their adaptation to water limitation, growth is less predictable and responds closely to rainfall events. Humid savanna vegetation has more adaptations for fire and low nutrients and trees are functionally more similar to forest vegetation. Savanna vegetation also shows adaptations for herbivory (particularly thorns) and light/shade. Many plants have underground stems and complex root systems to cope with high environmental stresses. Table 2 captures the main functional attributes that reflect current environmental conditions for Southern Africa (Scholes et al, 1997).

Grasses with a C₄ photosynthetic pathway tend to be dominant, particularly in hotter environments (Huntley, 1982; Menaut et al, 1985). Grasses in wetter African sites tend to be perennial and tufted, with fewer, larger seeds, while in more arid sites they are rhizomatous or stoloniferous, producing many, small seeds (Scholes et al, 1997). Some may be fairly deep rooted with some roots even approaching 2 m. Annual grasses tend to have shorter roots than perennials and are more competitive in upper soil layers responding faster to low and erratic rainfall patterns where there is not so much deep percolation. On the other hand, perennials can respond quicker to the first rains, mobilising root stores before annuals start their growth from seed. On the whole, grasses are able to respond quickly to rainfall events with rapid leaf flush and high productivity, while leaf loss during dry periods reduces losses due to respiration.

Trees have access to deep water enabling pre-rain leaf growth by up to three months before the first rains (Tybirk et al, 1992) and late leaf shedding which enables a longer growing season than grasses (Scholes & Walker, 1993). It was initially thought that trees essentially used deeper soil water than grasses, and this resource partitioning allowed the co-existence of these two lifeforms (Walter, 1971). However, it is now known that some trees, particularly in more arid areas, have shallow roots that extend laterally (particularly on clay soils where percolation is low) and may not have many deep roots (e.g. Menaut, 1983; Knoop & Walker, 1985; Belsky et al, 1989; Coughenour et al, 1990; Le Roux & Bariac, 1998). Besides water, this takes advantage of nutrient concentration in upper soil layers, including that washed out from fire ash in the early rains before grasses become established. In arid areas, deciduous trees are more common as leaf loss reduces water stress. Deciduous trees in humid savannas are thought to have most of their roots in the upper soil layers, thus they shed their leaves in the dry season. Evergreen trees have extensive root systems enabling them to use water from deep soil layers, thus maintaining relatively high transpiration and photosynthetic rates during the dry season (Vareschi, 1960).

There are few empirical data on allocation of primary production in savanna plants. Herbaceous plants are short-lived and almost all aerial parts die back after reaching maturity to be replaced at the start of the next growing season. Perennials allocate more primary production to belowground structures, dropping their leaves and increasing root production at the onset of the dry season. Annuals, on the other hand, have much higher seed production and may produce seed several times throughout the growing season to avoid drought, fire and herbivory. Woody plants invest more production to long-lived structures (the bole, branches and coarse roots). They also use a higher proportion of primary production for maintenance respiration (Illius et al, 1996). The fraction of NPP allocated belowground for all plants is affected by nutrient and water availability, fire, herbivory and competition, increasing with stress and disturbance. Belowground allocation is typically 40-80% in grasses compared with 20-60% in trees (Scholes & Hall, 1996).

Interactions, both competitive and facilatory, between the tree and grass components are very complex, variable and poorly understood (Scholes & Archer 1997). Trees and grasses both compete for limited resources, and grass biomass/NPP typically decreases as tree biomass/density increases. Several studies show a concave curvilinear change (Donaldson & Kelk, 1970; Walker et al, 1972; Beale, 1973; Dye & Spear, 1982; Scanlan & Burrows, 1990; Teague & Smit, 1992). The inverse relationship between an index of "treeness" and grass production shows the steepest decline when woody biomass is low (Figure 1). It has been shown at one site that the degree of curvature decreases as the productive potential of the site (i.e. water and nutrient availability) increases, suggesting that it is related to the degree of resource competition (Scanlan & Burrows, 1990). Convex relationships also occur (Aucamp et al, 1983). Trees can create a more suitable microhabitat for grass species and low tree densities have been seen to increase grass production in arid areas by providing shade (reducing evaporation), and nutrient concentration due to leaf litter and root decay and faeces of sheltering animals and birds (Weltzin & Coughenour, 1990; Belsky et al, 1989, 1993; Pugnaire et al, 1996).

In most situations, mature trees out-compete grasses for light, water and nutrients, yet grasses out-compete small shrubs and tree seedlings (reducing establishment) and they increase the likelihood of fires which kill small trees (Knoop & Walker, 1985; Scholes & Archer, 1997). This competitive asymmetry can lead to structural instability (Scholes & Hall, 1996). Often some degree of tree clumping takes place adding further complexity with conditions often very different between the under-canopy and inter-canopy areas (Weltzin & Coughenour, 1990, Belsky, 1989, 1993; Veetas, 1992; Mordelet & Menaut, 1995).

Many thanks to Bob Scholes (CSIR, South Africa), Xavier Le Roux (INRA, Clermont-Ferrand, France), Jonathan Scurlock (ORNL, USA) and Joe Scanlan (Department of Natural Resources, Queensland, Australia) for providing information and making corrections to the manuscript. Dale Kaiser & Sonja Jones (ORNL, USA) for calculating tropical % of the Olson et al (1983) "grasslands" category.

Sadly, David Hall passed away in August 1999 before this chapter was published. His knowledge and love of savannas was only surpassed by his eagerness to learn and teach.

     If you have entered this site from UNEP.NET or SCOPE, then feel free to use the link below to view the full site.