Australian Acacias have spread to many parts of the world. In South Africa, species such as A. mearnsii and A. dealbata are invasive. Consequently, more effort has focused on their clearing. In a context of increasing clearing costs, it is crucial to develop innovative ways of managing invasions. Our aim was to understand the biophysical properties of A. mearnsii in grasslands as they relate to grass production and to explore management implications. Aboveground biomass (AGB) of A. mearnsii was determined using a published allometric equation in invaded grasslands of the north Eastern Cape, South Africa. The relationships among the A. mearnsii leaf area index (LAI), Normalized Difference Vegetation Index (NDVI) and AGB were investigated. The influence of A. mearnsii LAI and terrain slope on grass cover was also investigated. Strong linear relationships between NDVI, LAI and AGB were developed. Acacia mearnsii canopy adversely impacted grass production more than terrain slope (p < 0.05) and when LAI approached 2.1, grass cover dropped to below 10% in infested areas. Reducing A. mearnsii canopy could promote grass production while encouraging carbon sequestration. Given the high AGB and clearing costs, it may be prudent to adopt the 'novel ecosystems' approach in managing infested landscapes.

No abstract

Accurately measuring evapotranspiration (ET) is essential if we are to derive reasonable estimates of production and water use for semi-arid savannas. Estimates of ET are also important in defining the health of an ecosystem and the quantity of water used by the vegetation when preparing a catchment-scale water balance. We derived ET 0 from an automatic weather station 30 km west of Skukuza, Kruger National Park, South Africa using the Penman-Monteith equation, and then used the MODIS LAI to inform the model of canopy phenological dynamics. This result was compared with 173 days of ET measurements from the eddy covariance (ET ec ) system near Skukuza in 2007 as well as from the ET recorded by a large-aperture scintillometer at the same site in 2005. The model compared favourably with both sets of measured data and, when used independently of the eddy covariance data, ET MODIS predicted an annual ET of 378 mm in 2007 for the semi-arid savanna around the Skukuza flux site.