Current knowledge about soil and water interactions related to below ground processes in reconstructed ecosystems on land highly disturbed by mining activities is very limited. It is not clear what influence trees and the associated grass understorey vegetation have on the water table and on the moisture status throughout the profile, nor has the effect of water or lack of it on the trees and grasses been adequately addressed. As water availability is an important problem in this system, the present studies were done to 1) develop and validate appropriate techniques of measuring water use of trees in the glasshouse prior to its use in the field, 2) to quantify the relative contributions of trees and grass understorey to tailings dam evapotranspiration in the wet and dry seasons and 3) to determine the contribution of the water table to evapotranspiration from the tailings dam.
The problem was approached by determining water relation responses of the native trees species used to revegetate the 310 ha Kidston Gold Mine (KGM) tailings dam to varying and repeated levels of water stress. Water use was monitored by the stem heat balance technique of measuring sap flow. An experiment was conducted under glasshouse conditions and water was withheld for periods of 5 or 6 days at various times to impose a range of water stress. Repeated water stress treatments for all of the species (E. camaldulensis, E. maculata, E. citriodora, A. salicina) did not appear to alter their water use characteristics.
The calibration test of the stem heat balance and the heat pulse velocity techniques against the absolute loss in weight prior to their use to measure the water use of various tree species, proved to be highly accurate. However, because of the collar type design of the stem heat balance technique, the method continuously heats the tree stems. When used continuously over 6 days it caused constriction and severe damage to bark tissue of sensitive trees species, resulting in the death of these sensitive species. This therefore limits the usefulness of the technique with native trees.
Field experiments were conducted on a 11 ha revegetation trial site at the southern end of the 310 ha KGM tailings dam. The mine is located 1.5 km from the abandoned town of Kidston in the Georgetown mining district of Queensland, approximately 280 km west-north-west of Townsville and 260 km south-west of Cairns, at latitude 18˚ 53˚ S and longitude 144˚ 09˚ E. The experimental design consisted of four plots in each of the driest and wettest areas and at two intermediate locations. A weather station within the main tailings dam recorded the climatological data. The two intermediate plots were intensively instrumented for measurement of soil water, the occurrence and duration of fluctuations in depth of the water table, and of plant water relations. Tree water use was measured directly by the heat pulse technique. The grass understorey water use was estimated by the change in soil moisture content beneath the grass canopy in an area where the influence of trees was negligible.
The tailings area on the 11 ha site was found to have a water table which fluctuated between 150 - 217 cm from wet to dry season. The tailings consisted of randomly distributed layers of fine and coarse materials and distinct layers of lower hydraulic conductivity were shown to result in periodic perched water tables at the depth of reduced hydraulic conductivity.
Water use by trees was normalised and expressed as daily sap flow relative to the projected canopy area. In the drier portion (IW plot) of the trial area, water use was higher than in the wetter part (IE plot). In the latter, water use was affected by the prolonged waterlogged soil conditions, resulting in frequent leaf senescence and reduced growth. Over the one year study in 1999 total mean daily transpiration rate, expressed as litres per square metre of canopy and measured by the heat pulse method, was 6.29 and 3.22 litres day-1 m-2 on the IW plot, and 3.46 and 2.16 litres day-1 m-2 on the IE plot. Mean Epan during the measurement period in the dry season was 6.8 and 5.1 mm day-1 during the wet season. Differences in the water use of trees in the experimental plots were due to the combined effects of tree species and the detrimental effects of waterlogged soil conditions during the wet season in the IE plot.
Average daily water use was strongly correlated (0.80**) with the stem diameter at probe implantation site and with leaf area (0.68**). Use of the stem diameter at probe implantation site was the best parameter with which to estimate the tree water use of the plant community, given the accuracy and ease of measurement compared with projected canopy area.
Tree canopy covered an area of 15.1% and 12.6% of the total areas in the IW and IE plots respectively. The remaining areas were occupied by grass understorey dominated by buffel grass (Cenchrus ciliaris). The grass density differs between the plots. The water balance of the system (expressed as fluxes relative to total plot area) indicated that on average, trees in the IW plot transpired 0.65 and 0.96 mm day-1 and in the IE plot 0.38 and 0.55 mm day-1 during the 1999 dry and 2000 wet seasons. Average tree water transpiration from the water table accounted for 89.96% of evapotranspiration in the dry season and 77.68% in the wet season, the proportion increasing as the water table height increased during the wet season. The corresponding estimated average grass understorey contributions to evapotranspiration for the dry and wet seasons as measured by the neutron moisture meter (NMM), were 0.032 and 0.174 mm day-1 considered an underestimate when compared with the biomass method (dry biomass / water use efficiency (WUE) of buffel grass). Earlier water balance studies under trees and pasture irrigated with secondary treated sewage effluent indicated not much difference in their transpiration rate. Assuming grasses are transpiring at approximately the same rate as trees, the mean water use from a plant community on the dry part of the site (IW plot) would be 0.67 mm day-1 and on the wetter part (IE plot) it would be 0.26 mm day-1 during the wet season. In the dry season, grasses in the IW plot senesce and do not transpire while the IE plot grasses continued to grow and transpire at a rate of 0.052 and 0.011 mm day-1 in the IW and IE plots respectively. The transpiration rate of the grass understorey, as measured from the NMM, is a gross underestimate due to the presence of perched and main water tables.
The low hydraulic gradient across the area represents an estimated discharge rate of 2.6 mm day-1 towards the lowest portion, at the northeast comer of the site. There was a general drop of about 0.5 - 5 cm during the dry season in September 1999 and a variable rise and fall during the wet season, of the water levels in the piezometers. Mean decrease in water table depth was 0.85 mm day-1 during the wet season and a corresponding 4.25 mm day-1 drop during the dry season. Whether this fall is due to the discharge under gravity, to tree water use or both, requires further investigation.