Defining drought, categorising current droughts, and assessing grassland and rangeland sustainability in a quantitative and scienti:fic manner are important national issues for Australian State and Commonwealth governments, landholders and agribusiness. A challenge for ecologists and modellers of Australia's grasslands and rangelands is to integrate biological models, geographic information systems, satellite imagery, economics, climatology and visual high-performance computing into readily available products that can provide monitoring and prediction advice in near real-time.The QDNR systems approach to the management of native grasslands recognises that drought occurs at a regional scale, and that impacts on livestock and natural resources can be forecast using simple models of soil water, plant growth and animal performance. Our vision for a comprehensive Australian Gmssland and Rangeland Assessment System (Aussie GRASS) is one that consists of the best combination of minfall analyses, seasonal climate forecasts, satellite and terrestrial monitoring, and simulation models of relevant biological processes. This will provide a rational basis for largescale management decisions by graziers, extension workers, land resource managers, bureaucmts and politicians. Aussie GRASS products are currently used within the Queensland government for drought declaration assessments and applications for Drought Exceptional Circumstances.The Aussie GRASS national spatial modelling framework allows agricultural simulation models to be run at a continental scale on a 0.05 degree 05 km) grid. The simulation model currently in use by the Aussie GRASS project is the GRASP pasture model developed for tropical native pastures in Queensland by QDPI and QDNR. In the latest Aussie GRASS project, other regional models are being examined for their applicability to areas such as the southem winter perennial grass zone, chenopod shrublands or the high minfall temperate zone. 330 The Queensland version of the Aussie GRASS model is currently used to produce data for a monthly report -A Summary of Seasonal Conditions in Queensland. Model outputs are used in conjunction with recorded and forecast rainfall, satellite imagery, Southern Oscillation Index and current drought declarations to build a comprehensive picture of the current and future seasonal conditions impacting on primary producers. Other numerous outputs from the model can be produced and tailored as required.

This study examined the potential to simulate the quality, as indicated by nitrogen concentration, of the diet of sheep grazing the Mitchell and mulga grasslands of western Queensland. Development of this simulation capability will allow pasture growth and animal production models to be more easily coupled. Modifications and optimisation of an existing beef cattle diet selection model, in conjunction with a single sward pasture model, accounted for 69.1% (P < 0.001) and 41.9% (P < 0.001) of variation in sheep dietary nitrogen concentrations observed from grazing trials on Mitchell and mulga grasslands, respectively. Failure to simulate some of the higher recorded dietary nitrogen concentrations was probably associated with high forb content in the diet. Examination of the results indicated that development of pasture growth models which simulate major pasture species, or groups of species (e.g. perennial grasses, annual grasses, browse, forbs, legumes), would appear to be necessary before diet selection models will be better able to explain the variation in dietary quality observed in grazing animals.