15.4.1 Savanna woodlands

Printer-friendly version

(a)  Savanna climate and vegetation


Figure 15.21 A typical savanna woodland at the wetter end of its distribution near Darwin, Northern Territory. Annual rainfall is between 1600 and 1800 mm, trees are tall (12 – 16 m) and stand density is high. (Photograph courtesy G. Duff, school of Biological Sciences, Northern Territory University, Darwin)

Savannas cover 30% of mainland Australia and are charac-terised by a continuous grassy understorey with a discon-tinuous canopy of trees and shrubs (Figure 15.21). Grasses are either annual or perennial, but almost invariably C4 plants. The woody component is dominated by eucalypts and/or acacias, although in some regions (for example, on clay soils in the Barkly and Victoria River Districts) other genera such as Terminalia dominate. Tree cover can vary from a few scattered trees and shrubs to more than 50% cover. The basal area of woody species ranges from 0.25 m2 ha–1 to at least 10 m2 ha–1.

Standing biomass and species composition of savanna woodlands between 11°S and 18°S within the Northern Territory vary according to plant-available moisture and soil type (Figure 15.22). Over this latitudinal range, mean annual rainfall declines from approximately 1600 mm to 400 mm. Temperature extremes also increase inland. In parts of the more arid interior (e.g. Tennant Creek, at the southern extent of the savannas), temperatures range from below 0ºC to above 40°C.


Figure 15.22 Tree basal area increases according to annual rainfall and soil clay content (taken here as a surrogate for soil nutrient resources). (Based on unpublished data compiled by D. Eamus and G. Duff)

Plant-available moisture and nutrients are acknowledged as the primary determinants of savanna structure and function (i.e. physiological and phenological behaviour). Secondary determinants are fire (Chapter 19) and grazing. Fires are frequent, annual in many areas, and cover more than 50% of the northern end of the savannas in any one year. However, these fires are of much lower intensity than those which occur in southern Australia (typically 1000 W m–2 compared with 100000 W m–2 in an ‘Ash Wednesday’ episode). In these low-intensity savanna fires only the grassy understorey is burnt. In the relatively tiny area of savanna which escapes fire for prolonged periods (decades), vegetation structure changes. Grass density is significantly reduced and mid-storey shrub layer density increases.

Scattered within the matrix of the savanna are pockets of fire-sensitive vegetation such as monsoon rainforests. In the Northern Territory alone there are approximately 15000 such patches, typically only a few hectares in extent.

Savannas are tropical ecosystems and occur where climate is strongly seasonal, with pronounced wet and dry seasons. In the Top End of the Northern Territory, for example, more than 90% of the average annual rainfall occurs between November and April. At least six months of any year will be rainless. Temperatures are uniformly high with daily maxima generally above 30°C. Atmospheric vapour pressure deficit (VPD) shows a seasonal cycle, with minimum VPD occurring from December to March and afternoon maxima of 5 kPa between July and September.

(b)  Response to moisture


Figure 15.23 In a typical year on a savanna woodland, rainfall is confined to six months. During the wet season, the vapour pressure deficit (VPD) is) low and soil water availability is high. During the dry season, rainfall is essentially absent,VPD is higher and soil water availability declines to a minimum. Throughout the year mean daily temperature varies only by 2-4°C. During the wet season, stomatal conductance (gs) is high during day-time, with little variation between morning and afternoon. By contrast, in the dry season, gs peaks to a lower value compared to the wet season and shows a marked decline each afternoon. Virtually all savanna tree species in northern Australia follow this pattern. (Based on various sources compiled by D. Eamus and G, Duff)


Figure 15.24 Pre-dawn leaf water potential is a good indicator of soil water availability because leaf water potential continues to increase at night-time until the gradient between soil and leaf water potential is zero. In the wet season when soils are at or above field capacity, pre-dawn leaf water potential remains close to zero, declining gradually during the dry season as soil water availability diminishes. Evergreen trees continue to transpire through-out the dry season and experience lower pre-dawn water potentials than deciduous trees, which because of their loss of canopy can maintain a higher water potential than evergreen species (Based on various sources compiled by D. Eamus and G. Duff)


Figure 15.25 Stomata close in response to increased leaf to air vapour pressure difference, and all trees examined so far in the wet-dry tropics show such behaviour. Starting with comparably high stomatal conductance gs) under mild conditions, Acacia auriculiformis is hypersensitive to increasing leaf to air vapour difference compared to Eucalyptus tetrodonta, so that carbon gain per day per unit leaf area will diminish earlier in A. auriculiformis as the dry season advances. Differences in canopy volume per tree can, however, vary in the opposite direction to gs sensitivity, so that carbon fixed per unit tree will be comparable, and results in co-dominance on given sites (Based on original data in Cole 1994, and redrawn by D. Eamus and G. Duff from Berryman et al. 1994)

The wet season is characterised by high soil water availability and low VPD. The dry season is characterised by low soil water availability and high VPD. These differences are reflected in diurnal patterns of stomatal conductance (gs) (Figure 15.23), which in turn relate to seasonal patterns of pre-dawn leaf water potential (Ψleaf) (Figure 15.24). Data were obtained in a study at Solar Village, an unburnt savanna woodland 20 km south of Darwin. Two features can be discerned: the wet season gs is approximately double the value measured in dry season and stomata remain fully open for much of the day during the wet season. In the dry season, however, gs peaks mid-morning and declines approximately linearly from midday through to the evening. Both features typify savanna species.

An overall decline in gs between wet and dry seasons (Figure 15.23) reflects declining soil water availability. Pre-dawn leaf water potential is a useful indicator of plant-available soil moisture, and is correlated with daily maximal gs. As soil water potential declines, pre-dawn leaf water potential declines and daily maximum gs declines. In addition the leaf to air vapour pressure difference (LAVPD) increases in the dry season, but also increases between morning and afternoon in both wet and dry season. During the wet season when soil water availability is high and the increase in LAVPD between morning and afternoon is small, gs remains high because water uptake by the roots is able to meet the demands of transpiration. In contrast, in the dry season when soil water availability is low and the increase in LAVPD during the day is considerably larger, gs peaks in the morning, when evapo-transpirational demand is lower and soil water content higher compared to the afternoon.

Sensitivity of gs to LAVPD is a common feature of savanna species, as shown in Figure 15.25 for Eucalyptus tetrodonta and Acacia auriculiformis. The form of the relationship and the sensitivity of gs to LAVPD differs between species, as does a threshold LAVPD above which a rapid decline in gs occurs.

Changes in leaf water potential are not a primary cause of differences in gs between seasons, since identical leaf water potential can be measured in the wet season and early dry season, but gs is substantially lower in the early dry season than the wet season. Furthermore, Cole (1994) was able to show on relatively wet sites occupied by A. auriculiformis that pre-dawn and minimum leaf water potential differences between seasons were small, but significant declines in gs between morning and afternoon or between wet and dry seasons nevertheless occurred. This reduction in gs was associated with increased VPD.

Seasonal decline in plant-available moisture is also reflected in measurements of pre-dawn leaf water potential (Figure 15.24) and diurnal changes in leaf water potential. Pre-dawn water potential shows little change during the wet season when soils are saturated and values range between -0.1 and -0.4 MPa. However, during the dry season, substantial declines occur as soil water content decreases and extraction of water occurs from deeper soil.

(c)  Phenology

Close inspection of savannas reveals four main phenological groups. These are species which are (1) evergreen, (2) partly deciduous, (3) semi-deciduous and (4) fully deciduous. Ever-green species retain a full canopy throughout the year (e.g. Eucalyptus miniata); partly deciduous species exhibit a decline in canopy cover only briefly and to levels not below 50% of their full canopy (e.g. Eucalyptus bleeseri); semi-deciduous species lose more than 50% of their canopy during the dry season (e.g. Erythrophleum chlorostachys); and fully deciduous species are leafless for at least two months during the dry season (e.g. Terminalia ferdinandiana). Species in Top End savannas are more or less evenly distributed between these four cat-egories. The fact that there is no one optimum or dominant strategy indicates that a variety of factors interact to determine deciduous behaviour, and different species respond to these factors to differing degrees.

Being evergreen is costly. Efficiency of water and nutrient use is lower than in deciduous species. So why be evergreen?

The benefit of maintaining a canopy throughout the year is the maintenance of a positive carbon balance for every month of the year. Thus evergreen species fix carbon during the dry season, and they also have a canopy which can take advantage of improved conditions (early rain and decreased VPD) at the end of the dry season without the lag associated with developing new leaves. Furthermore, the ratio of costs to ‘payback’ is positively correlated with leaf longevity. Thus, if a leaf is costly in terms of construction and maintenance, and the return on investment (assimilation of carbon) is low per unit time (the cost/benefit ratio is large), then leaf lifespan must be longer (evergreen) so that the leaf lives long enough to pay back (in carbon fixation) the investment of carbon and other nutrients made.

Deciduousness also has costs, principally the loss of carbon and other nutrients associated with leaf fall and the loss of ability to fix carbon for six months of every year. The benefits of deciduousness include avoidance of extreme foliar drought and hence a reduced need to develop and maintain more extensive root systems, and avoidance of sustained herbivory.

In evergreen species, leaves either live longer or are turned over continually. Longer lived leaves need greater protection against herbivory and consequently invest more resources in secondary compounds which serve to make foliage less palatable due to fibre and tannins, or even poison-ous due to secondary metabolites such as cyanogenic glycosides. There would be less pressure for protection in deciduous species.

Evergreen and deciduous species thus differ in phenology but coexist in nature due to comparable outcomes for carbon balance from expensive but long-lived leaves of evergreens and short-lived, less expensive leaves of deciduous species that fix carbon for about six months each year. All four pheno-logical groups recognised in the wet–dry tropics of Australia (Williams et al. 1997) are equally represented, confirming that no single strategy confers a unique advantage.