Grazing for livestock production is the major land use in Australia in terms of area. The components of livestock production examined in this report were wool and meat from both cattle and sheep. Grazing enterprises cover a wide range of climates and vegetation zones, and hence represent a complex challenge in terms of calculating the impact of climate change. The major objective of the project was to develop a simple model that allowed the impact of changes in temperature and rainfall on livestock production to be calculated by the Garnaut Review’s economic modelling team. Building on a previous approach (Crimp et al. 2002), we have used the existing spatial distribution of plant growth attributes and livestock management (stocking rate, sown pastures) to develop relationships to explain the effect of -

*        spatial climate variation on livestock carrying capacity; and

*        $ value of production.

In deriving a final model for simulating national carrying capacity, eleven models were initially developed and compared. The eleven models were based on combinations of simulated pasture variables from the soil-water balance and pasture growth model GRASP (Rickert et al. 2000, McKeon et al. 2000), alternative treatments of the arid zone and either inclusion or exclusion of a livestock production index.

To parameterise the GRASP model for each climate location, pasture growth attributes were drawn from the AussieGRASS national modelling framework (Carter et al. 2000). On this basis, GRASP was used to calculate pasture growth and an index of the length of the growing season. In turn, these variables were converted to livestock carrying capacity -
*        (beef number equivalents per hectare); and
*        $ production per hectare using relationships derived from Statistical Division data.
The Statistical Divisions that constitute the arid zone of Australia (72% of area, 28% of livestock numbers) were treated in two ways, either as 22 climate locations, or as 76 climate locations. We reasoned that splitting the arid zone and including a greater number of locations in the simulation produced a more robust model for simulating national carrying capacity from climate inputs.

Eleven different models were compared using sensitivity tests with various climate change scenarios. The selected model supplied to the Garnaut Review’s economic modelling team was regarded as a conservative representation of the effects of climate change in the arid and semi-arid zones, whilst providing reasonable explanation of existing spatial variation in livestock carrying capacity (LCC) in higher rainfall zones in southern Australia. The selected model included the likely effects of temperature increase on the value of livestock production.

The limitations to the approach include: uncertainties with regard to the effects of CO2 on pasture growth and animal diet quality; the representation of seasonal distribution of rainfall in climate change scenarios; and representation of any future adaptations which are likely to have a strong socio-economic component as well as a climatic component.

The main findings of the project were:

1. Simulated variables ‘pasture growth’ and ‘length of the growing season’ accounted for a high (>70%) proportion of spatial variability in livestock carrying capacity (cattle and sheep expressed as beef equivalents).
2. The explanation of spatial variation was improved by accounting for additional factors such as tree density (expressed as tree basal area).
3. The simulated variable ‘length of the growing season’ better accounted for variation in livestock carrying capacity in the more climatically favourable regions of south-eastern Australia.

In conclusion, a simple model capable of simulating the impact of climate change on livestock carrying capacity and production across Australia was developed and supplied the model equations to the Garnaut Review’s economic modelling team. As part of model development the uncertainties associated with the application of the model and those areas where further research is required were also identified.