In the Australia and New Zealand region, droughts are closely related to major drivers of year-to-year and decadal variability such as ENSO, Indian Ocean SSTs, the Antarctic Circumpolar Wave (White and Peterson, 1996; Cai et al., 1999; White and Cherry, 1999), and the Interdecadal Pacific Oscillation (Mantua et al., 1997; Power et al., 1998; Salinger and Mullan, 1999), as well as more or less chaotic synoptic events. These are all likely to be affected by climate change (see Sections 12.1.5 and 12.2.3, and TAR WGI Chapters 9 and 10).
Using a transient simulation with the NCAR CCMO GCM at coarse resolution (R15) (Meehl and Washington, 1996), Kothavala (1999) found for northeastern and southeastern Australia that the Palmer Drought Severity Index indicated longer and more severe droughts in the transient simulation at about 2xCO2 conditions than in the control simulation. This is consistent with a more El Niño-like average climate in the enhanced greenhouse simulation; it contrasts with a more ambivalent result by Whetton et al. (1993), who used results from several slab-ocean GCMs and a simple soil water balance model. Similar but less extreme results were found by Walsh et al. (2000) for estimates of meteorological drought in Queensland, based on simulations with the CSIRO RCM at 60-km resolution, nested in the CSIRO Mk2 GCM.
A global study by Arnell (1999), using results from an ensemble of four enhanced greenhouse simulations with the HadCM2 GCM and one with HadCM3, show marked decreases in runoff over most of mainland Australia, including a range of decreases in runoff in the Murray-Darling basin in the southeast by the 2050s of about 12-35%. HadCM3 results show large decreases in maximum and minimum monthly runoff. This implies large increases in drought frequency.
The decrease in rainfall predicted for the east of New Zealand by downscaling from coupled AOGCM runs for 2080 and the corresponding increase in temperature are likely to lead to more drought in eastern regions, from East Cape down to Southern Canterbury. Eastern droughts also could be favored by any move of the tropical Pacific into a more El Niño-like mean state (see Table 3-10). The sensitivity of New Zealand agriculture and the economy to drought events was illustrated by the 1997-1998 El Niño drought, which was estimated to result in a loss of NZ$618 million (0.9%) in GDP that year. A drought in north and central Otago and dry conditions in Southland associated with the 1998-1999 La Niña resulted in a loss of about NZ$539 million in GDP (MAF, 2000).
Recurring interest in Australia in policies on drought and disaster relief is evidence of a problem in managing existing climate variability and attempts to adapt (O'Meagher et al., 1998). Present variability causes fluctuations in Australian GDP on the order of 1-2% (White, 2000). Drought and disaster relief helps immediate victims and their survival as producers (e.g., QDPI, 1996) but does not reduce costs to the whole community and in fact may prolong unsuitable or maladapted practices (Smith et al., 1992; Daly, 1994), especially if there is climatic change. Farm productivity models are being used to simulate past and present farm production and to assess causes of and management options for coping with drought (Donnelly et al., 1998). This is contributing to the fashioning of drought assistance and advisory policies.
The potential impact of drought on the Australian economy has declined, in relative economic terms, over time in parallel with the decline in the importance of agriculture to the economy (ABARE, 1997; Wilson and Johnson, 1997). In 1950-1951, the farm sector constituted 26.1% of GDP, whereas currently (1997-1998) it constitutes 2.5%. Similarly, the contribution of the farm sector to Australian exports has fallen from 85.3% (1950-1951) to 19.6% (1997-1998), with a reduction in the total farm sector labor force of about 6%. This despite the fact that farm production has increased over the same period. Thus, drought remains an important issue throughout Australia for social, political, geographical, and environmental reasons (Gibbs and Maher, 1967; West and Smith, 1996; Flood and Peacock, 1999).
Stehlik et al. (1999) studied the impact of the 1990 drought on more than 100 individuals from 56 properties in central Queensland and northern NSW to document the social experiences of dealing with drought. They conclude that there is strong evidence that the impact of the extended drought of the 1990s is such that rural Australia will never be the same again: "There is a decline in population: a closing down of small businesses, fewer and fewer opportunities for casual or itinerant work, more and more producers working off-farm' and a reduction in available services."
A change in climate toward drier conditions as a result of lower rainfall and higher evaporative demand would trigger more frequent or longer drought declarations under current Australian drought policy schemes, which rely on historical climate data and/or land-use practices on the basis of an expectation of historical climatic variability. A major issue for operational drought schemes is the choice of the most relevant historical period for the relative assessment of current conditions (Donnelly et al., 1998).
Examples of Australian government involvement in rural industries that have been subject to decline in commodity prices over several decades (e.g., wool) suggest that the industries will be supported until the cost to the overall community is too high and the long duration or high frequency of drought declarations is perceived as evidence that the drought policy is no longer appropriate (Mercer, 1991; Daly, 1994). In the case of wool, the shift of government policy from that of support to facilitation of restructuring has involved a judgment about future demand and therefore prices (McLachlan et al., 1999) and has only occurred after an extended period of low prices (Johnston et al., 1999). With a change in climate toward drier conditions, drought policy probably would follow a similar path.
The New Zealand Government response to drought comes under Adverse Climatic Events and Natural Disasters Relief policy that was released in 1995. Government responds only when rare climatic or natural disasters occur on a scale that will seriously impact the national or regional economy and the scale of the response required is beyond the capacity of local resources. The policy is to encourage industry/community/ individual response, rather than reliance on government support.
Science has a major role in assessing the probability that recent and current climatic conditions could be the result of natural variability or increased GHGs. At best, these assessments are presented in probabilistic terms (e.g., Trenberth and Hoar, 1997). The public and its representatives will have to judge what constitutes evidence of anthropogenic effects and to what extent future projections and their impacts should be acted on. Because of their impact, future droughts provide a very public focus for assessing the issues of climate change compared to natural variability. Appropriate land-use and management practices can be reassessed by using agricultural system models with CO2 and climate projections from GCMs (Hall et al., 1998; Howden et al., 1999f; Johnston et al., 1999). However, political judgments between the alternatives of supporting existing land use or facilitating reconstruction are likely to require greater certainty with regard to the accuracy of GCMs than is currently available (Henderson-Sellers, 1993).
One source of adaptation is seasonal and long-lead climate forecasting. This is one area in which climate science already is contributing to better agricultural management, profitability, and, to some extent, adaptation to climate change (Hammer et al., 1991, 2000; Stone and McKeon, 1992; Stone et al., 1996a; Johnston et al., 1999). Indeed, empirical forecasting systems already are revealing the impact of global warming trends (Nicholls et al., 1996b; Stone et al., 1996b), and these systems already are adapting to climate change through regular revision and improvements in forecasting skill.
Cropping, horticulture, and forestry in Australia and New Zealand are vulnerable to invasion by new pests and pathogens for which there are no local biological controls (Sutherst et al., 1996; Ministry for the Environment, 1997). The likelihood that such pests and pathogensparticularly those of tropical or semi-tropical originwill become established, once introduced to New Zealand, may increase with climate warming.
Indepth case studies are being conducted in Australia to test the performance of pest impact assessment methodologies for estimating the vulnerability of local rural industries to pests under climate change (Sutherst et al., 1996). In New Zealand, pests that already are present may extend their ranges and cause more severe damage. For example, because of the reduced incidence of frosts in the north of New Zealand in recent years, the tropical grass webworm (Herpetogramma licarisalis) has increased in numbers and caused severe damage in some pastures in the far north.
The vulnerability of horticultural industries in Australia to the Queensland fruit fly Bactrocera (Dacus) tryoni under climate change was examined by Sutherst et al. (2000). Vulnerability was defined in terms of sensitivity and adaptation options. Regional estimates of fruit fly density, derived with the CLIMEX model, were fed into an economic model that took account of the costs of damage, management, regulation, and research. Sensitivity analyses were used to estimate potential future costs under climate change by recalculating costs with increases in temperature of 0.5, 1.0, and 2°C, assuming that the fruit fly will occur only in horticulture where there is sufficient rainfall or irrigation to allow the crop to grow. The most affected areas were the high-altitude apple-growing areas of southern Queensland and NSW and orange-growing areas in the Murrumbidgee Irrigation Area. Apples and pears in southern and central NSW also were affected. A belt from southern NSW across northern Victoria and into South Australia appeared to be the most vulnerable.
Adaptation options were investigated by considering, first, their sustainability under present conditions and, second, their robustness under climate variability and climate change. Bait spraying is ranked as the most sustainable, robust, and hence most promising adaptation option in boh the endemic and fruit fly exclusion zones, but it causes some public concern. The sterile insect technique is particularly safe, but there were concerns about costs, particularly with large infestations. Exclusion is a highly effective approach for minimizing the number of outbreaks of Queensland fruit fly in fly-free areas, although it is vulnerable to political pressure in relation to tourism. These three techniques have been given the highest priority.
Other reports in this collection