TS-CRC Student project - Soil properties and plant water use of tropical savannas

Northern Territory University, Darwin: Completed

Georgina Kelley

Commenced: 1997

Summary | Objectives | Background | Results | More information |

Georgina Kelley at the research site

George Kelley gets a lift to her study site


This project was part of the first-round TS-CRC's research project, Water and Carbon Exchange of Savannas and concentrated on the way in which soil properties and plant physiological characteristics affected fluxes of water in the soil-plant-atmosphere continuum. Study sites were contained within the Howard River East catchment Northern Territory, in conjunction with other activities of the Centre's research project.

For more information on the PhD project, go to the the first-round CRC project Water and Carbon Exchange of Savannas.


The objectives of the project included:

  1. To monitor both soil water and groundwater levels in the catchment on a regular basis and to determine seasonal changes in both these measures. This work is conducted in collaboration with Water Resources Division of Lands, Planning and Environment.
  2. The daily and seasonal monitoring of plant water status in three major ecosystems of the savanna: eucalypt woodland, Melaleuca swamp and monsoon vine forest.
  3. To monitor the storage of water in the stems of trees on a daily and seasonal basis, in order to determine the extent to which trees rely on water stored within their trunks for transpiration.
  4. To quantify physical properties of the soils of the Howard River East catchment, in order to determine how these properties affect the availability and use of water by plants.


Savannas, a vegetation system comprising both trees and grasses in varying proportions, cover more than 10 per cent of the global land surface and approximately 25 per cent of the Australian continent. Globally they are increasingly subject to rising levels of human habitation and development. Australian savannas are relatively pristine and consequently provide a rare opportunity to investigate the structure and function of this ecosystem during early stages of development.

Australian savannas are largely restricted to tropical and sub-tropical regions and experience a highly seasonal rainfall of between 500 and 1900 mm per year. Climate is dominated by summer monsoons and a wet season of five or six months duration.

Development of groundwater resources in northern Australia is increasing as population and horticultural requirements for water increase. Within the extensive savanna of northern Australia, smaller patches of wet monsoon forest and Melaleuca swamps are found, typically in lower-lying areas with access to perennial water.

George’s work investigated aspects of the physiological and hydrological properties (soil and plant) of trees in these three contrasting vegetation assemblages. In addition, the PhD investigated sources of water used by trees using stable isotope analyses of xylem sap, soil water and groundwater.

The heat pulse technique was used to examine tree water use in several dominant tree species of open-forest savanna, Melaleuca swamp and monsoon forest in wet and dry seasons.


Transpiration (per unit leaf area) did not differ between savanna and Melaleuca swamp. However, in the savanna, stand water use was aseasonal (0.9 mm d -1 ) but in the Melaleuca swamp forest stand water use was significantly larger in the wet season (1.7 mm d -1 ), compared to the dry season (1.0 mm d -1 ). Transpiration did not appear to be reduced by waterlogging in the wet and early-dry seasons. Water use was also seasonal in the monsoon forest (1.5 mm d -1 in the wet season, 0.4 mm d -1 in the dry season). Changes in leaf area and stomatal conductance may cause seasonal patterns of water use in swamp and monsoon forests. Annual water use was 315 mm in the savanna, 508 mm in the Melaleuca swamp forest and 568 mm in the monsoon forest.

Whole tree hydraulic conductance ( K ) was measured for savanna and Melaleuca swamp forest trees. There was no significant difference in K (per unit sapwood area) between seasons or communities. This suggests that trees in the two communities exhibit similar hydraulic properties despite significant differences in soil type and soil water availability.

Groundwater use by Melaleuca trees was assessed by comparing deuterium and oxygen isotope composition of xylem sap, soil water and groundwater. Soil water was predominantly used in the wet and early dry season (despite inundation during this time) although groundwater was used towards the end of the dry season. Dependence of Melaleuca swamp forest and possibly monsoon forests (similarly located in low-lying areas) on groundwater has important implications for management of groundwater resources.

Time domain reflectometers (TDR) were used to measure changes in soil water content at several depths throughout over a 2 year period in the savanna. During the wet season both overstorey (trees ) and understorey (predominantly grasses) utilised water in the soil lying above a lateritic duricrust layer at 1.2 m depth. Following cessation of rains, trees used water from above and below the duricrust layer; during the later 1999 dry season water below the duricrust was used exclusively. Water stored in the upper 5 m of soil was sufficient to supply the entire volume of water transpired by trees in the dry season.

Groundwater dependence by savanna vegetation appears unlikely. However, Melaleuca swamp forests and monsoon forests are more likely to be dependent on groundwater during the dry season. This must be an important consideration for future management of catchments and groundwater resources.


Brochure: Sustainable use of groundwater in the Top End
Sustainable use of groundwater in the Top End. Based on TS-CRC project: Carbon and water exchange in savannas [pdf 553.9 kb]