Dye, PF, Jarmain, C, Oageng, B, Xaba, J & Weiersbye, IM 2008, 'The Potential of Woodlands and Reed-Beds for Control of Acid Mine Drainage in the Witwatersrand Gold Fields, South Africa', in AB Fourie, M Tibbett, I Weiersbye & P Dye (eds), Mine Closure 2008: Proceedings of the Third International Seminar on Mine Closure, Australian Centre for Geomechanics, Perth, pp. 487-497, https://doi.org/10.36487/ACG_repo/852_44
(https://papers.acg.uwa.edu.au/p/852_44_Dye/)
Abstract: The Mine Woodlands Research Project was initiated in 2002 at AngloGold Ashanti Limited to quantify the 
effectiveness of trees in minimizing acid mine drainage (AMD) from mine tailings storage facilities (TSFs), 
and in rehabilitating polluted soils. We report on the increase in evapotranspiration (ET) expected as a 
result of tree planting on AMD in the 500–700 mm rainfall “Highveld” zone of central South Africa.  
The TSFs are situated within transformed (degraded) grassland and open savannah woodland, with 
Phragmites australis reed-beds occurring on perennial seepages and drainage lines. Highveld grassland is 
characterized by seasonal growth and shallow root systems, and a review of catchment water balances and 
ET measurements indicates that its annual ET in the central Witwatersrand (mean annual precipitation = 
650 mm) approximates only 566 mm. This suggests that grassland around TSFs would be ineffective in 
preventing seepage from entering groundwater and streams. However, some tree species colonize the 
periphery of TSFs, and could potentially consume enough seepage to lower water tables to below the 
polluted zone. The difference in annual and seasonal ET between the various vegetation types is an 
important indicator of the potential of such `engineered woodland buffers’. Hourly ET from a P. australis 
reed-bed in AMD was measured for one year, using an energy balance micrometeorological technique. 
Despite seasonal growth and winter dormancy, annual reed-bed ET totaled 1170 mm. The heat-pulse 
velocity (HPV) technique was used to monitor year-round sap flow in stems of the native tree Rhus lancea 
(Karee) growing in AMD on the same mine. Annual transpiration only, estimated on a canopy area basis, 
was 976–1211 mm (average 1096 mm). Annual ET of Eucalyptus camaldulensis (River Red Gum) on sites 
with freely available, good-quality groundwater was simulated with the 3-PG forest plantation model. 
Simulated annual ET varies between 917 and 1454 mm over a 14 year rotation (mean = 1277 mm), peaking 
at 8 to 10 years of age, and declining thereafter. The significance to AMD control of a conservative increase 
in ET (400 mm) arising from tree-planting was then estimated using seepage rate data from a large TSF. 
From the results we conclude that a strategic and relatively small-scale planting of trees around tailings 
dams, either indigenous woodlands or alien tree plots, would potentially lead to a significant increase in 
mean annual ET, and thereby reduce AMD pollution of deeper groundwater and streams. Where P. australis 
reed-beds occur on seepage, little increase in ET would accrue from tree planting, and such reed-beds 
should be conserved and managed appropriately to promote maximum leaf area development and 
transpiration rate.