Climate Change 2001:
Working Group II: Impacts, Adaptation and Vulnerability
Other reports in this collection

12.5. Food and Fiber

12.5.1. Introduction

Some 60% of Australia is used for commercial agriculture (Pestana, 1993). Only 2% is used for broad acre and intensive crop production; 4% is sown to pastures. Soil and topography are major constraints on cropping, which produces about 50% of the gross value of farm production—the rest being divided equally between meat production and livestock products such as wool and milk.

Russell (1988) identifies climatic "frontiers" affected by climate change and interdecadal variability in rainfall—namely, the inland limit to agriculture (mean annual rainfall <300 mm in the south and <750 mm in the north), the southern limit of effective summer rainfall; and the lower rainfall limit of high value crops (on the order of 1,000 mm). Temperature also is limiting, with some temperate crops held south of their high-temperature (northern) limit, and other more tropical crops held at their low-temperature (southern or high-altitude) limit. Large areas of the interior and west are desert or very arid rangelands with low yields; much of this land is now returned to Aboriginal management.

There is great interannual variability, especially in the interior and more northern regions, associated mainly with ENSO, convective rainfall, and tropical cyclones. Australia is known as a land of droughts and flooding rains. Secondary factors such as wildfires also account for losses of fodder, animals, and farm infrastructure (sheds, fences, machinery), and hail causes significant crop losses. Accordingly, drought and disaster relief policies are matters of ongoing concern (O'Meagher et al., 1998, 2000; Pittock et al., 1999), as is sustainability in the face of economic pressures and global change (Abel et al., 1997).

In New Zealand, pastoral agriculture provides more than 40% of the country's export earnings (Statistics New Zealand, 1998). Dairy farming is the major activity in wetter areas; sheep dominate hilly and drier areas, and beef cattle are widely distributed throughout the country. Pastures are highly productive—composed largely of introduced grass and nitrogen-fixing legume species—and support high stock numbers. Rainfall in New Zealand generally is not strongly seasonal, but high evapotranspiration rates in the summer make pastoral agriculture in the east vulnerable to largely ENSO-related variability in summer rainfall.

12.5.2. Pastoral Farming

Howden et al. (1999d) have summarized and updated work by Hall et al. (1998), McKeon et al. (1998), and Howden et al. (1999a,b). They find that although CO2 increase alone is likely to increase pasture growth, particularly in water-limited environments, there also is strong sensitivity to rainfall, so that a 10% reduction in rainfall would counter the effect of a doubled CO2 concentration. A 20% reduction in rainfall at doubled CO2 is likely to reduce pasture productivity by about 15% and live-weight gain in cattle by 12% and substantially increase variability in stocking rates, reducing farm income. The latest scenarios, which have substantial reduction in rainfall in many parts of Australia, would tend to reduce productivity.

Howden et al. (1999d) also found that doubled CO2 concentrations are likely to increase the deep drainage component under pastures, which may increase the risk and rates of salinization where the potential for this problem exists. Doubled CO2 and increased temperature would result in only limited changes in C3 and C4 grass distributions (Howden et al., 1999a).

In New Zealand, productivity of dairy farms might be adversely affected by a southward shift of undesirable subtropical grass species, such as Paspalum dilatatum (Campbell et al., 1996). At present, P. dilatatum is recognized as a significant component of dairy pastures in Northland, Auckland, Waikato, and the Bay of Plenty. A "user-defined" management threshold for the probability of finding this grass in dairy pasture is predicted by a climate profile technique (Campbell and Mitchell, 1996). This can be considered the point at which adaptive management changes are regarded as necessary. This technique was applied with the IS92a (mid) and IS92e (high) climate change scenarios and the CSIRO4 GCM pattern, using the CLIMPACTS integrated assessment model (Kenny et al., 1995, 2000). Results indicate a significant southward shift in the probability of occurrence of P. dilatatum with global warming; more southerly geographic thresholds are reached at later dates, but 25-30 years earlier with the higher emissions scenario.

Comprehensive assessment of the response of dairy cattle to heat stress in NSW and Queensland was carried out by Davison et al. (1996). Physiological effects of heat stress include reduced food intake, weight loss, decreased reproduction rates, reduction in milk yields, increased susceptibility to parasites, and, in extreme cases, collapse and death. Heat stress can be reduced by the use of shade and sprinklers, and thresholds for their use can be determined. Jones and Hennessy (2000) applied this adaptation to the Hunter Valley in NSW, using probabilistic estimates of temperature and dewpoint changes resulting from climate change for the IS92 range of scenarios to 2100. They then estimated the probabilities of given milk production losses as a function of time and calculated the economic benefits of provision of shade and sprinklers. They conclude that heat-stress management in the region would be cost-effective. However, such adaptation may not be as cost-effective in a hotter or more humid climate.

Howden and Turnpenny (1997) and Howden et al. (1999e) also have looked at heat stress in beef cattle. They find that heat stress already has increased significantly in subtropical Queensland over the past 40 years (where there has been a warming trend) and that it will increase further with greenhouse-induced global warming. They suggest a need for further selection for cattle lines with greater thermoregulatory control, but they point out that this may be difficult because it may not be consistent with high production potential (Finch et al., 1982, 1984).



Other reports in this collection