Managing Agricultural Systems in a Variable, Non-Stationary Climate: Part II. Natural Resource Management and Grazing Systems
Authors: Steven Crimp, Neil Flood, Greg McKeon and Mark Howden
Final Report on research undertaken for Land and Water Australia Project QNR31 Managing Climate Variability Program (August 2004)
[Acrobat versions available at the end of this page]
Hard copies are available by emailing to rouseabout@nrm.qld.gov.au
Abstract
Australia's grazing lands are important to it's Agricultural economy as they occupy over 50% of the continent and support approximately 120 million sheep and 25 million cattle. In Queensland, grazing of native pastures is the major land use (85% of State) and contributes approximately $2.5 billion in gross agriculture value (i.e. approximately 40% of total agriculture).
Analyses of national climate variability have revealed that climatic events (e.g. drought, flood and extreme temperatures) are increasingly falling outside the long-term historical experience, as a consequence of anthropogenic climate change. As current agricultural practices have been strongly shaped by historical climate conditions, predicted changes in future climate, resulting from enhanced greenhouse gas concentrations, may result in loss of productivity and environmental degradation.
In order to ensure that natural resource and related on-farm management strategies take into account both changes in variability and non-stationarity, detailed climatic analysis quantifying historical trends and projected changes/impacts are required. These analyses need to be both scalable and transferable - principles that have been carried through the analyses undertaken for this project.
Trend analyses have been performed on rainfall and temperature data in Queenslandís Burnett district, with the data used to simulate historical time sequences of deep drainage and run-off as well as on-farm management issues such as pasture productivity/condition and heat stress in cattle.
The results showed that annual rainfall for the Burnett study region (made up of an East, North, Central and South sub-region) has declined over the period 1890 to 2002. Of the four stations examined in the East sub-region half demonstrated statistically significant declining trends (i.e. P<0.05), all the stations (4) in the North sub-region demonstrated statistically significant declining trends (i.e. P<0.05) and the Central and South sub-regions demonstrated 2 and 1 statistically significant declining trends respectively. The analysis of both historical maximum and minimum temperatures revealed a consistent pattern of warming across the study region although only the warming trends in minimum temperature were statistically significant (i.e. P<0.05).
The simulations of annual pasture production, revealed consistent declining trends across all stations and sub-regions except for the driest station in the Central sub-region (Auburn). While the simulation results returned a consistent declining trend in pasture production for all subregions analysed, there was no instance where the linear trends were statistically significant or explained more than 1% of the production variability. Similarly the simulation of deep drainage returned declining trends across all 16 stations examined in the Burnett study region, although these declines were almost non-existent in both the Central and South sub-regions for the analysis period 1890 to 2002. The simulated runoff data returned three instances where positive trends occurred. These instances occurred in the East, Central and South sub-regions in predominantly drier locations. Again, even though both positive and negative trends were simulated no statistically significant results were returned.
Projections of change in deep drainage, runoff, pasture production and heat stress were made through modification of the historical climate data. Projections of change in average temperature and rainfall were used to modify a range of other climate inputs such as daily measures of solar radiation, vapour pressure and evaporation. The climate change experiments were limited by the coarse resolution of the climate projections, with individual climate model projections varying little for the entire study region. The greatest variation was as a result of inter-model differences with the climate models used producing a possible +6% to -6% change in rainfall by 2030 and +19% to -19% change in rainfall by 2070 (in response to 0.85 and 2.55oC of global warming).
Even though both positive and negative changes in rainfall were projected for the Burnett Study region and the CO2 fertilisation effects were considered, the average simulated pasture production still demonstrated small declines (less than 1%) in productivity by 2030 (although productivity changes were highly variable between stations). By 2070 declines in pasture production of -4.5% were simulated.
Projected changes in deep drainage in response to enhanced greenhouse gas concentrations were greater than for pasture production with a relatively consistent pattern of decline across almost all of the 16 locations simulated. Stations in the South sub-region did show a number of instances were deep drainage increased in 2030, but declined by 2070.
Similarly runoff demonstrated a consistent pattern of decline across the East, North and Central sub-regions, with the South sub-region demonstrating both increases and decreases. The runoff values highlighted the water-limiting environment of both the South and Central sub-regions with runoff and deep drainage relatively small compared to the wetter North and East sub-regions.
The impact of climate change on heat stress was large with significant increases experienced across all sub-regions. The impacts were greatest in the Central and North sub-regions with over 100% increase in heat stress simulated by 2070.
FullReport (2Mb) (updated 13:38, 5 Dec 2008)
| 
;)
