Dr Greg McKeon, CRC for Greenhouse Accounting, Queensland Department of Natural Resources and Water
prepared for the 2006 Australian State of the Environment Committee, 2006
Agricultural land uses can be ranked in terms of water requirements that range from extensive grazing in arid environments to irrigated pastures, sugar cane and horticulture in coastal high rainfall locations. Year-to-year variability in rainfall places a potential stress on the viability and vulnerability of particular land uses. For example, dryland cropping has had fluctuating fortunes at its margins in western New South Wales (Condon 2002, p. 189). Not surprisingly, the 'forecasting' of rainfall at annual to generational time scales has been seen as a way of improving land and water use planning, and has been the 'holy grail' of Australian climate science.
During the last 100 years there has been increased understanding of the components of the climate system. Some of the year-to-year variability is the result of the 'chaotic' nature of the climate system (Burroughs 2003, p. 53). Major fluctuations have also occurred in the 'forces' that drive the climate system. These include natural forces, such as solar variability and volcanic eruptions (Burroughs 2003, p. 88), and the human-induced impacts listed previously (Burroughs 2003, p. 92-100). Variation in these natural and human-induced 'forcings', in combination with chaotic climatic processes, results in a complex global climate system in which 'cause and effect' are not easily identified.
As well as variability in sea surface temperatures, atmospheric pressures and rainfall at ENSO time scales (three to seven years), there have also been consistent patterns of variability in global sea surface temperatures and pressures at longer time scales (for example, decadal and multi-decadal, Allan 2000). Of particular importance for Australian rainfall, are the longer-term (15–20 years) fluctuations in basin-wide Pacific Ocean sea surface temperatures (Mantua et al. 1997, Power et al.1999, Lindesay 2003). The newness of this scientific understanding is reflected in: (1) the various names for this multi-decadal feature of the Pacific Ocean referred to as the Pacific Decadal Oscillation (PDO), or Inter-decadal Pacific Oscillation (IPO); (2) the difficulty of explicitly separating possible long-term (about 20 years) sea surface temperature fluctuations from the chaotic clustering of El Niño and La Niña events; and (3) assessing the impact of human-induced term climate change.
The interaction of the La Niña phase of ENSO and the 'cool' phase of the PDO/IPO (Mantua and Hare 2002) has been associated with above-average annual rainfall, particularly in eastern Australia (for example, early 1890s, 1916–18, mid-1950s, early 1970s and perhaps the late 1990s). The understanding of the behaviour of the PDO/IPO is still the subject of scientific debate and, as yet, no climate forecasting capability exists at decadal or multi-decadal time scales (Power et al.2003). Currently, available climate forecasts are at seasonal and annual time scales with probabilistic forecasts reflecting the uncertain non-deterministic behaviour of the climate system.
Climate change projections over the next 100 years are derived from simulations using a number of global climate models, each with different mathematical representations of the climate system and different capabilities in terms of available computing resources. More importantly, the simulations also necessarily include a range of predictions or scenarios of future global economic development, population growth and technological advances. Hence, there can be a wide range of plausible outcomes in terms of regional changes in rainfall and temperature. In addition to these uncertainties, regional climates are also subject to uncertain biophysical feedbacks from the impact of climate changes on vegetation and landscape hydrology. Thus, decision makers now face the problem of integrating knowledge of historical variability, seasonal and annual outlooks, and a range of long-term climate change projections.
The major gap in climate forecasting capability is at the decadal and multi-decadal scale. Recent studies have emphasised the importance of climate variability at decadal and multi-decadal time scales that affect important regional climate elements such as seasonal rainfall and severe storm frequency (such as cyclones and hurricanes). Furthermore, reconstructions of regional rainfall records for north-east Queensland for a few hundred years before the instrumental record (Lough 2003), indicate that major decadal and multi-decadal periods have occurred that may have been drier than have been experienced in the last 100 years (the main period of agricultural and urban development). Thus, climate science, through the study of historical records and future projections, is providing evidence to support the changing perceptions of Australia's rainfall variability. Instead of a 'random and unpredictable' climate that varies about a constant long-term (stationary) average, land managers should now expect rainfall (and other climate elements) to vary on time scales that may require changes in existing land use and business practices (choice of crops, livestock carrying capacity, water allocation and insurance risk assessment). The alternative of allowing or supporting inappropriate land and water use, or livestock carrying capacity, has in the past led to resource degradation (Condon 2002, McKeon et al.2004) and impeded structural adjustment (McColl and Young 2005) by maintaining enterprises not suited for surviving climatic and economic variability.
The history of government policy on land use in Australia provides examples of successes and failures in managing for climate uncertainty. In some cases, the lack of understanding of rainfall variability at decadal and multi-decadal time scales has led to false expectations and inadequate planning. The uncertainty resulting from such variability has been compounded in Australia by the fact that agricultural and urban settlements were occurring before or at the same time that meteorologists, farmers and graziers were collecting the necessary data to measure climate variability. These data were necessary to formulate appropriate land use recommendations (such as extensive grazing in contrast to dryland cropping) and to indicate what government support was likely to be required through drought and flood events. Despite a lack of knowledge of the forces driving climatic variability, the substantial economic contribution that agriculture (including pastoralism) has made to the development of the national economy is testament to the successful adaptation that has occurred over the last 200 years (Burroughs 2003, p. 152). Ironically, with current and expected changes in climate, resource-use planners may now be in a similar position of climatic uncertainty as those who were developing land and water-use policy one hundred years ago. Thus the current challenge for decision-making is to integrate the emerging capability from climate science, with all its transparent uncertainty, to make better decisions at both individual time scales (individual farmers' and graziers' lifetimes or business life cycles) and at societal (multigenerational) time scales. Examples of current approaches to meet this challenge are presented later in this paper (Box 2).
Links to another web site
Links to data in the DRS
Opens a pop-up window