Flow characteristics of rivers in northern Australia: Implications for development

Summary

Annual, monthly and daily streamflows from 99 unregulated rivers across northern Australia were analysed to assess the general surface water resources of the region and their implications for development. The potential for carry-over storages was assessed using the Gould-Dincer Gamma method, which utilises the mean, standard deviation, skewness and lag-one serial correlation coefficient of annual flows. Runs Analysis was used to describe the characteristics of drought in northern Australia and the potential for ‘active’ water harvesting was evaluated by Base Flow Separation, Flow Duration Curves and Spells Analysis. These parameters for northern Australia were compared with data from southern Australia and data for similar Köppen class from around the world. Notably, the variability and seasonality of annual streamflow across northern Australia were observed to be high compared with that of similar Köppen classes from the rest of the world (RoW). The high inter-annual variability of runoff means that carry-over storages in northern Australia will need to be considerably larger than for rivers from the RoW (assuming similar mean annual runoff, yield and reliability). For example, in the three major Köppen zones across the North, it was possible (theoretically) to only exploit approximately 33% (Köppen Aw; n = 6), 25% (Köppen BSh; n = 12) and 13% (Köppen BWh; n = 11) of mean annual streamflow (assuming a hypothetical storage size equal to the mean annual flow). Over 90% of north Australian rivers had a Base Flow Index of less than 0.4, 72% had negative annual lag-one autocorrelation values and in half the rivers sampled greater than 80% of the total flow occurred during the 3-month peak period. These data confirm that flow in the rivers of northern Australia is largely event driven and that the north Australian environment has limited natural storage capacity. Hence, there is relatively little opportunity in many northern rivers to actively harvest water for on-farm storage, particularly under environmental flow rules that stipulate that water can only be extracted during the falling limb of a hydrograph. Streamflow drought severity, the product of drought length and magnitude, was found to be greater in northern Australia than in similar climatic regions of the RoW, due to higher inter-annual variability increasing the drought magnitude over the course of normal drought lengths. The high likelihood of severe drought means that agriculturalists seeking to irrigate from rivers in northern Australia should have especially well developed drought contingency plans.

Introduction

The catchments of northern Australia generate approximately 60% of Australia’s surface runoff, yet most rivers in the region remain undeveloped (NLWRA, 2002). This lack of water harvest in the North is largely because settlement and intensive land use since Europeans arrived in 1788 have been focused on the so-called ‘well-watered’ more familiar temperate climes of southern Australia (Blainey, 1966, Courtenay, 1977). This is not to say that the North has not received attention. There is a long history of attempts to develop cultivated agriculture in the tropical north of Australia – most of which have been judged unsuccessful (Davidson, 1965, Kelly, 1966, Davidson, 1969, Lacey, 1979, Woinarski and Dawson, 1997). Failure to understand and appreciate the northern environment is often cited as a factor in the failure of agricultural developments (e.g. Woinarski and Dawson, 1997).

Recently there has been renewed interest in developing the water resources of northern Australia (PMC, 2007) fuelled partly by widespread perceptions of abundant water resources in that region (Gehrke, 2005), and partly by droughts and declining rainfall trends in southern Australia. There is also now general recognition that in some catchments in southern Australia, water is over-allocated (Fullagar, 2004). However, proposals to utilise water from rivers of northern Australia are highly contentious. These rivers have iconic status to many Australians and hold cultural value to indigenous populations (Jackson, 2004). They also support fisheries that have considerable commercial, recreational and cultural value (Griffin, 1987, Robins et al., 2005) and a rapidly growing tourism industry. There is considerable debate about whether and how the North’s water resources should be utilised for purposes other than current environmental uses (e.g. Camkin et al., 2006). This study seeks to inform that debate on the future of water in the North by providing a general assessment using data available on the water resources of the region.

Regional scale analyses of the hydrological characteristics of river systems in northern Australia are few and limited to certain sub-regions. Horn (2000) made a qualitative assessment of the ground and surface water resources of Cape York (Fig. 1) with reference to many rivers in that region, while McMahon et al. (2005) provided a comprehensive analysis of the hydrology of the Lake Eyre Basin, an episodically filled, intermittent lake in central Australia whose upper catchment lies within the geographic scope of this study (Fig. 1). More recently, Moliere et al. (2006) applied multivariate cluster analysis to classify 28 streams within the Fitzroy (WA), Daly, and Flinders catchments (Fig. 1). This analysis identified four distinct streamflow patterns in these three catchments (perennial, seasonal, dry seasonal and seasonal-intermittent) and of the five variables examined, the coefficient of variability of annual flows and the mean annual number of zero flow days were able to explain 94% of the flow pattern.

Most other studies have been at the catchment scale and focused primarily on fluvial sedimentation and erosion (e.g. Prosser et al., 2002, Fielding et al., 2004), particularly with respect to the geomorphologic regimes of northern estuaries (e.g. Coleman and Wright, 1978, Chappell, 1993, Wolanski and Chappell, 1996), sediment and nutrient delivery to the Great Barrier Reef (e.g. Moss et al., 1992, Neil et al., 2002, Brodie and Mitchell, 2005, McKergow et al., 2005), ecological assessment and water requirement (e.g. Pusey and Arthington, 1996, Erskine et al., 2003, Braimbridge and Malseed, 2007), baseflow and groundwater in limestone and dolomitic rocks in the Northern Territory (e.g. Tickell, 2002, Cook et al., 2003, Jolly et al., 2004) or existing irrigation schemes (e.g. Ruprecht and Rodgers, 1999). The applicability of results from continental scale analysis of Australian and global streamflow datasets to northern Australia is limited by a lack of flow data for this region (e.g. McMahon, 1977, McMahon, 1982, McMahon et al., 1987, Haines et al., 1988, Finlayson and McMahon, 1992, McMahon et al., 1992, Dettinger and Diaz, 2000, Peel et al., 2001, Peel et al., 2002a, Peel et al., 2004a).

To address the data deficiencies in past regional scale analyses, a database of flow records from 99 unregulated rivers across northern Australia (each with 10 years or more of continuous record) has been assembled. In this paper, these data are used to gain an understanding of the key features of the surface water resources of northern Australia, by performing a series of analyses at annual, monthly and daily time-steps. It is anticipated that this work will provide a regional scale context for more detailed hydrological investigations in the North in the future. Results from these analyses are compared with results from the rest of the world (RoW)¹ for the same Köppen classes and from southern Australia (i.e. represented by Köppen classes Csa, Csb, Cfb, Cfc, BSh and BSk) to ascertain whether river flow in northern Australia is different or not. To provide an appropriate basis for comparison and wider interpretation of the results of our analyses, we have stratified the results by Köppen climate type (Köppen, 1936, Peel et al., 2007).

In the next section, we provide a brief overview of the study area including details of the data used in the analysis. This is followed by an analysis of annual, monthly and daily streamflow characteristics under the headings of estimating reservoir capacity, reservoir management and water harvesting. Finally, a brief discussion on the applicability of the results across northern Australia is provided.