RIVER RESEARCH AND APPLICATIONS River Res. Applic. 19: 377–395 (2003) Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/rra.745 FLOW RESTORATION AND PROTECTION IN AUSTRALIAN RIVERS ANGELA H. ARTHINGTONa,b* and BRADLEY J. PUSEYb a b Co-operative Research Centre for Freshwater Ecology and Management, Centre for Catchment and In-Stream Research, Griffith University, Nathan, Brisbane, Queensland 4111, Australia Co-operative Research Centre for Tropical Rainforest Ecology and Management, Centre for Catchment and In-Stream Research, Griffith University, Nathan, Brisbane, Queensland 4111, Australia ABSTRACT Since 1857 new Australians have constructed many thousands of weirs (3600 in the Murray–Darling Basin alone) and floodplain levee banks, 446 large dams (>10 m crest height) and over 50 intra- and inter-basin water transfer schemes to secure water supplies for human use. Flow regulation has changed the hydrology of major rivers on three temporal sales–the flood pulse (days to weeks), flow history (weeks to years) and the long-term statistical pattern of flows, or flow regime (decades or longer). The regulation of river flows is widely acknowledged as a major cause of deteriorating conditions in many Australian river and floodplain ecosystems. In response to mounting environmental concerns, all states, territories and the Commonwealth Government have committed the nation to the principles of ecologically sustainable development and a process of national water reform. Rivers and wetlands are now recognized as legitimate ‘users’ of water, and jurisdictions must provide water allocations to sustain and where necessary restore ecological processes and the biodiversity of water-dependent ecosystems. Progress in the protection and restoration of river and wetland water regimes has been significant, with over half of mainland aquatic systems designated to receive water allocations of some sort. However, exactly how much water they will receive or retain is unclear from the data available. Moreover, the ecological outcomes and benefits of water allocations are not yet apparent in most aquatic ecosystems, with the exception of certain waterbird breeding events, the disruption of algal blooms in weirs and improved fish passage. After reviewing these issues, this paper addresses two vital questions: How much water does a river need? and How can this water be clawed back from other users? Studies conducted to date in Queensland rivers suggest that around 80–92% of natural mean annual flow (and other ecologically relevant hydrological indicators) may be needed to maintain a low risk of environmental degradation. In the Top End of the Northern Territory, some rivers are maintained at 80% of their natural flow, whereas two-thirds of various flow indicators has been proposed as the restoration target for the River Murray, and 28% of natural mean annual flow has been negotiated for the Snowy River in Victoria. To validate these estimates, ecologists are seeking opportunities to turn river restoration projects into long-term hypothesis-driven experiments in ecological restoration, and the funding, time and institutional support to do so. The paper ends with some suggestions to advance the water reforms and achieve higher levels of water allocation for the environment.Copyright # 2003 John Wiley & Sons, Ltd. key words: Australia; flow restoration; environmental flows; flow regulation; water reforms; experimental research; water pricing, water trading INTRODUCTION In honour of our late colleague, Professor W.D. (Bill) Williams, it is appropriate to begin this paper about flow protection and restoration in Australia with two observations from his early texts on the ecology of Australian inland waters. The direct effects of man on the inland aquatic environment have undoubtedly had important consequences for the distribution and abundance of much of the aquatic biota, consequences which, moreover, are probably still a long way from the full attainment (Bayly and Williams, 1973: Inland Waters and their Ecology). Almost every Australian lake, river, stream and pond now bears the imprint of man, or will shortly do so. And, like it or not, the future of the Australian inland aquatic environment is for man to determine (Williams, 1980: An Ecological Basis for Water Resource Management). Two decades later, Boulton and Brock (1999) concluded their text (Australian Freshwater Ecology: Processes and Management) with an overview of water issues, commenting: *Correspondence to: A. H. Arthington, CCISR, Griffith University, Nathan, Brisbane, Queensland 4111, Australia. E-mail: a.arthington@griffith.edu.au Copyright # 2003 John Wiley & Sons, Ltd. Received 28 April 2002 Revised 10 December 2002 Accepted 16 December 2002 378 A. H. ARTHINGTON AND B. J. PUSEY In Australia, there has been a shift in emphasis from concern largely about water quantity and supply for humans’ needs (engineering solutions) to protecting water quality and sustaining natural ecosystems (ecological solutions). We now recognise the central role of water regime in most wetland processes . . . . Today, in 2003, Australians have a much keener appreciation of the remarkable features of the continent, the scarcity, fragility and value of its freshwater ecosystems and the importance of maintaining and restoring river and wetland water regimes. Before reviewing progress with flow protection and restoration in Australia, it is important to appreciate the extreme climatic variability characterizing much of the continent, and how challenging their new conditions were for European colonists accustomed to seasonal variability in rainfall but unacquainted with a climate that could swing between drought in one year and floods the next. As a consequence of the dire need for reliability of water supplies and food, many weirs, dams and levee banks were constructed and river flows were regulated by these structures or modified by water abstraction and diversions. The legacy from these and other interventions in rivers and floodplain wetlands is the modification of water regimes and severe ecological impacts. This paper gives a brief summary of these impacts and an introduction to the wider literature on the state of Australia’s inland waters. In 1993, pressured by growing environmental, social and economic concerns, the Commonwealth Government and all Australian states and territories committed the country to the principles of ecologically sustainable development and a wide-ranging programme of water reforms, including economic restructuring of the water industry and formal allocations of water to the environment. Rivers and wetlands have been recognized as legitimate ‘users’ of water, and jurisdictions have been defining and implementing water allocations to sustain and restore ecological processes and the biodiversity of water-dependent ecosystems. Progress in the protection and restoration of river water regimes is recounted in this paper, through which run two vital questions: How much water does a river need? and How can this water be clawed back from other users? AUSTRALIA’S CLIMATE AND HYDROLOGY Australia (Figure 1) is a relatively small (8.42 106 km2) continent lying between latitudes 10 410 S (Cape York) and 43 390 S (southern Tasmania), a land of extensive plains and plateaux with only 2% of the land area above Figure 1. Map of Australia showing locations of major rivers mentioned in the text Copyright # 2003 John Wiley & Sons, Ltd. River Res. Applic. 19: 377–395 (2003) 379 AUSTRALIAN RIVERS 1000 m a.s.l and 59% lying between 200 m and 1000 m (Jennings and Mabbutt, 1986). It lies in the mid-latitude high pressure belt of the southern hemisphere and its climate is dominated by eastward moving anticyclonic cells with a low moisture content (Nix, 1981). Australia is the driest inhabited continent, receiving on average only 455 mm of rainfall annually; 66% of the continent gets <500 mm of annual rainfall (Nix, 1981). Evapotranspiration losses are high and mean annual runoff is only 57 mm, yielding the world’s lowest percentage of rainfall as runoff (mean of 12% compared to 33% for North America). Rainfall and runoff are not only lower in Australia than on any other continent but also more variable, both spatially and temporally. Spatial variations in runoff separate the continent into two distinct areas, a humid coastline and an arid interior, with the greatest proportion of total runoff occurring in the northern and northeastern coastal areas (88% from only 26% of the land area), and the least recorded in arid and semi-arid regions (75% of the continent receives <12.5 mm of annual runoff). Extreme temporal variations in runoff are also characteristic of much, but not all of Australia (Finlayson and McMahon, 1998; Smith, 1998). According to a range of hydrological indicators, arid-zone rivers such as Cooper Creek and the Diamantina River have the most variable flow regimes on Earth (Puckridge et al., 1998). Groundwater plays an important role in Australia’s water budget. Recent assessments suggest that some 25 780 Gl of rainwater finds its way into groundwater aquifers each year (SEAC, 1996). Groundwater, emerging at or near the surface, provides the base flows that sustain life in many arid-zone and other rivers during dry periods, supplies Australia’s unique mound springs and cave systems, as well as sustaining many types of plant community. Australia’s climate is changing under the influence of global warming. Studies by the Commonwealth Scientific and Industrial Research Organisation (CSIRO) suggest that annual average temperatures will rise by 1–6 C over most of Australia by 2070 (Whetton, 2001). Warmer conditions will lead to increased evaporation rates, decreased moisture, increased moisture stress, and either wetter or drier conditions, with significant effects on the number of flood and drought years. Daily rainfall variability and extreme events will change but not uniformly across the continent, leading to both increases and decreases in stream flows. FLOW REGULATION AND ITS ECOLOGICAL CONSEQUENCES IN AUSTRALIA The uneven spatial distribution and extreme temporal variability of surface and groundwater supplies in Australia have played a pivotal role in its settlement and development. Water control schemes were constructed almost immediately upon European colonization to secure reliable water supplies for settlements, mining and irrigated agriculture. From 1857 to the present, new Australians have constructed a multitude of flow regulation structures, including many thousands of weirs (3600 in the Murray–Darling Basin alone), locks and floodplain levee banks, 446 large dams (>10 m crest height) and over 50 intra- and inter- basin water transfer schemes. Australia now has the highest per capita water storage capacity in the world (8.8 107 Ml or 106 l), with 79% of this stored water used for irrigation purposes, the highest proportional usage of all OECD countries (SEAC, 1996). Most of this water is held in a few very large storages, the ten largest holding about 50% of total capacity. Details of dams and other interventions in rivers and floodplain wetlands can be found in Kingsford (2000). Australia’s legacy from the era of dam construction and floodplain modifications is highly modified flow regimes and degraded rivers. Flow regulation has changed the hydrology of rivers on three scales of temporal variation (Walker et al., 1995): the flood pulse (days to weeks), flow history (weeks to years) and the long-term statistical pattern of flows, or flow regime (decades or longer). Total river flows and their temporal patterns are fundamentally altered in many of the continent’s great rivers, including the River Murray, the major headwater tributaries of the Darling in New South Wales, the Burdekin, Fitzroy, Proserpine, Pioneer, Burnett, Mary, Brisbane and Logan rivers of Queensland, many coastal rivers in New South Wales, the Snowy River in Victoria, the Ord River in Western Australia, the Gordon and other Tasmanian rivers, as well as many small stream systems throughout the continent. Within-channel flows have been markedly reduced in streams and rivers, seasonal flow patterns have been lost or inverted and many floodplains and wetlands are much less frequently flooded than before dam construction, whereas some wetlands and temporary streams that naturally dried periodically are now permanently wet, or have lost their natural seasonal pattern of wet and drier months (Walker, 1985; Walker and Thoms, 1993; Gehrke et al., 1995; Gippel and Stewardson, 1995; Maheshwari et al., 1995; Kingsford, 2000; Sheldon et al., 2000; Thoms and Sheldon, 2000). The hydrological variability that typifies the natural flow regime of most Australian rivers (Pusey et al., 1993, 1995; Lake, 1995; Pusey and Arthington, 1996; Finlayson and McMahon, 1998; Copyright # 2003 John Wiley & Sons, Ltd. River Res. Applic. 19: 377–395 (2003) 380 A. H. ARTHINGTON AND B. J. PUSEY Puckridge et al., 1998), and the water regimes of many wetlands have been altered at a range of temporal and spatial scales (Maheshwari et al., 1995; Walker et al., 1995; Arthington et al., 2000; Brizga, 2000; Kingsford, 2000; Thoms and Sheldon, 2000). Flow regulation is widely acknowledged to be a major cause of deteriorating conditions in many Australian river and floodplain ecosystems (Cullen and Lake, 1995; Lake, 1995; Walker, 1985; Walker et al., 1995; Kingsford, 2000; Bunn and Arthington, 2002). Ecological changes in regulated river systems include massive loss of wetlands (90% of floodplain wetlands in the Murray–Darling Basin, 50% of coastal wetlands in New South Wales and 75% of wetlands on the Swan Coastal Plain in southwest Western Australia; Bunn, 1999; Kingsford, 2000), decline of riparian forests (Kingsford, 2000; Leslie, 2001), invasion of dewatered river channels and wetlands by vegetation (Bren, 1992), changes in aquatic plant community structure in regulated river reaches and weirs (Blanch et al., 2001; Mackay et al., 2001), population and species diversity declines of invertebrates, fish and waterbirds (Harris, 1984; Doeg et al., 1987; Gehrke et al., 1995; Kingsford, 2000; Quinn et al., 2000) and several invertebrate extinctions (Sheldon and Walker, 1997). Seven lacustrine invertebrates disappeared when Lake Pedder in Tasmania was flooded as part of a hydropower scheme, and leeches collected from River Murray wetlands by Victorian hospitals at up to 30 000 per annum were seldom seen after the 1970s (Kingsford, 2000). More conspicuous impacts of flow regulation have been massive blooms of toxic cyanobacteria in storages and rivers, most notably the 1000 km long bloom in the Barwon–Darling River in 1991, so severe as to necessitate the declaration of a state of emergency and trucking of water to rural homesteads (Bowling and Baker, 1996). Australia’s regulated rivers have also been prone to invasions of exotic species of plants such as water hyacinth (Eichhornia crassipes), Hymenachne, willows (Salix spp.), and fish (Arthington et al., 1990), especially carp (Cyprinus carpio; Gehrke et al., 1995) and tilapia (Arthington and Blühdorn, 1994). As Cullen and Lake (1995) so aptly state, we have degraded our rivers and wetlands in ways that prevent them from sustaining natural aquatic ecosystems and their high levels of endemicity, and are replacing them with simplified systems of lower diversity and many exotic species that will be much less useful to humans in the future. Bunn and Arthington (2002) review Australian examples of four key flow-related processes underlying the decline of aquatic biodiversity in regulated rivers (habitat alterations and loss, disruption of life history processes and recruitment, loss of longitudinal and lateral connectivity, and susceptibility to invasions by exotic species). All of these processes combine to alter the natural distributions of species and the dynamics of aquatic populations, as well as aquatic community structure, ecological interactions (parasitism, competition and predation) and ultimately, food web structure and energy flow, as demonstrated so often in other parts of the world (Sparks, 1995; Ward and Stanford, 1995; Lundqvist, 1998; Ward et al., 2001; Naiman et al., 2002). WATER REFORM IN AUSTRALIA Scientists and environmental groups began agitating about the impacts of large dams and the degradation of rivers in the early 1970s when remote Lake Pedder was flooded and the wild rivers of Tasmania were threatened by hydropower developments. At that time, the management objective was to provide ‘maximum supplies with minimum wastage of water’ (Water Conservation and Irrigation Commission, 1971), and a wastewater mentality prevailed. Australia’s large catchments were termed ‘drainage divisions’, rivers were said to supply ‘effluent’ creeks and water flows to floodplains, wetlands, aquifers or to the sea were called ‘surplus flows’ and were regarded as ‘wasted’. Twenty years later Australia adopted the National Strategy for Ecologically Sustainable Development, and in 1994, in response to growing environmental concerns and the need for economic restructuring of the water industry, all three tiers of government in Australian (Federal or Commonwealth, State/Territory and Local) committed to a process of national water reform. The Council of Australian Governments (CoAG), representing all governments, drove the water reform process, agreeing to ‘implement comprehensive systems of water allocations and entitlements backed by separation of water property rights from land title and clear specification of entitlements in terms of ownership, volume, reliability, transferability and, if appropriate, quality’ (CoAG Communique, 25 February 1994, Hobart, Attachment A Water Resource Policy, Item 4a). Underlying the CoAG water reforms is the expectation that implementation of water pricing, trading and greater efficiency of use will lead to improvements in land and water management and in all forms of environmental quality, particularly the ecological integrity of river and floodplain ecosystems (Smith, 1998). Copyright # 2003 John Wiley & Sons, Ltd. River Res. Applic. 19: 377–395 (2003) AUSTRALIAN RIVERS 381 To give policy guidance to jurisdictions and water managers responsible for implementing the reforms, 12 National Principles for the Provision of Water for Ecosystems (ARMCANZ, ANZECC, 1996) were formulated (Table I). The basic premise of the principles is that rivers and wetlands are legitimate ‘users’ of water and that the provision of water allocations is critically important to ecological sustainability. The goal of the principles is ‘to sustain and where necessary restore ecological processes and biodiversity of water dependent ecosystems’ by means of environmental water allocations that are legally recognized and founded on the best scientific information available (ARMCANZ, ANZECC, 1996, p. 5). The principles further require that all aspects of water management must consider the implications for water-dependent ecosystems; monitoring systems must be established to assess the ecological outcomes of water allocations, and they must be amenable to adjustment on the basis of new knowledge and understanding gained through monitoring and research. Under the Australian Constitution, State and Territory Governments have responsibility for the management of the surface and groundwater resources of the continent (except for areas of Commonwealth land). Thus to implement the CoAG water reforms, all jurisdictions have had to introduce new water legislation and make adjustments to policy and institutional structures enabling consideration of the provision of environmental water allocations. Early histories of these developments can be found in Allan and Lovett (1996) and Smith (1998). They have led to diverse legislative and institutional arrangements, accompanied by regional and quite variable approaches to the determination of water allocations for the environment, reflecting different geographic, climatic and ecological conditions, past policies, water management practices and commitments to water users, and scientific as well as methodological developments. ENVIRONMENTAL FLOW METHODS Methods to determine environmental water allocations (formerly termed ‘in-stream flows’, now generally called ‘environmental flows’ or ‘water regimes’ (to accommodate wetlands and groundwater systems) have been explored in Australia since the 1980s (e.g. Richardson, 1986; Hall, 1989). Several texts, conference publications and literature reviews (Gordon et al., 1992; Arthington and Pusey, 1993; AWWA, 1994; Tharme, 1996; Arthington and Zalucki, 1998; Dunbar et al., 1998; Tharme, 2000, this issue) describe and compare individual methods used in Australia to define the flows required for maintenance of geomorphological processes and channel morphology, aquatic and riparian vegetation, invertebrates, freshwater fish, other river-dependent vertebrates, wetlands and estuaries. Many of these methods are still in use. However, since 1992, several broader methodologies have been developed to address the water requirements of the ‘riverine ecosystem’ (Arthington et al., 1992) rather than the needs of just a few aquatic taxa (e.g. fish species of significance for recreational fishing and/or conservation, riparian and wetland vegetation, waterbirds). Arthington and Pusey (1993) describe Australia’s progression from ‘rules of thumb’ to the present-day emphasis on the water requirements of ecosystems. Methodologies that consider the entire riverine ecosystem include the holistic approach (Arthington et al., 1992; P. M. Davies et al., unpublished report, 1996; Pettit et al., 2001), and the building block methodology (Arthington and Long, 1997; Arthington and Lloyd, 1998; King and Louw, 1998), the expert panel method (Swales and Harris, 1995), the scientific panel method (e.g. Thoms et al., 1996), the flow restoration methodology (Arthington et al., 2000), the benchmarking methodology (Brizga, 2000) and DRIFT (downstream response to imposed flow transformation) (King et al., this issue). Each of these methods is described in Arthington (1998) and/or Tharme (2000). The ‘natural flows paradigm’ (Poff et al., 1997) and basic principles of river corridor restoration (Ward et al., 2001) underpin all of these ecosystem methodologies (now termed ‘holistic’ methodologies by Tharme, 1996, this issue). Their common objective is to maintain or partially restore important characteristics of the natural (or modelled unregulated) flow regime (i.e. the quantity, frequency, timing and duration of flow events, rates of change and predictability/variability) required to maintain or restore the biophysical components and ecological processes of in-stream and groundwater systems, floodplains and downstream receiving waters (e.g. terminal lakes and wetlands, or estuaries). Arthington et al. (1998) divided these methodologies into two main types: ‘bottom-up’, that is, methods that ‘construct’ a modified flow regime by adding flow components to a baseline of zero flow (this category includes the holistic approach and the South African building block methodology); and ‘top-down’ methodologies. The latter type addresses the question: How much can we modify a river’s flow regime before the aquatic ecosystem begins to noticeably change or becomes seriously degraded? Copyright # 2003 John Wiley & Sons, Ltd. River Res. Applic. 19: 377–395 (2003) 382 A. H. ARTHINGTON AND B. J. PUSEY Table I. National Principles for the Provision of Water for Ecosystems (ARMCANZ, ANZECC, 1996) GOAL The goal for providing water for the environment is to sustain and where necessary restore ecological processes and biodiversity of water-dependent ecosystems. PRINCIPLES Basic premise of principles Principle 1 River regulation and/or consumptive use should be recognised as potentially impacting on ecological values. Determining environmental water provisions Principle 2 Provision of water for ecosystems should be on the basis of the best scientific information available on the water regimes necessary to sustain the ecological values of water-dependent ecosystems. Provision of water for ecosystems Principle 3 Environmental water provisions should be legally recognized. Principle 4 In systems where there are existing users, provision of water for ecosystems should go as far as possible to meet the water regime necessary to sustain the ecological values of aquatic ecosystems whilst recognizing the existing rights of other water users. Principle 5 Where environmental water requirements cannot be met due to existing uses, action (including reallocation) should be taken to meet environmental needs. Principle 6 Further allocation of water for any use should only be on the basis that natural ecological processes and biodiversity are sustained (i.e. ecological values are sustained). Management of environmental water allocations Principle 7 Accountabilities in all aspects of management of environmental water provisions should be transparent and clearly defined. Principle 8 Environmental water provisions should be responsive to monitoring and improvements in understanding of environmental water requirements. Other uses Principle 9 All water uses should be managed in a manner which recognizes ecological values. Principle 10 Appropriate demand management and water pricing strategies should be used to assist in sustaining ecological values of water resources. Further research Principle 11 Strategic and applied research to improve understanding of environmental water requirements is essential. Community involvement Principle 12 All relevant environmental, social and economic stakeholders will be involved in water allocation planning and decision-making on environmental water provisions. Copyright # 2003 John Wiley & Sons, Ltd. River Res. Applic. 19: 377–395 (2003) AUSTRALIAN RIVERS 383 The benchmarking methodology (described in Arthington, 1998; Brizga, 2000) is the only top-down methodology applied in Australia for assessing the risk of environmental impacts due to water resource development. It is used routinely in Queensland and is beginning to be used in Western Australia. This methodology is designed to link information on alterations of natural flow regimes with the geomorphological and ecological consequences of flow regime change, by evaluating the condition of a range of river reaches subject to various degrees of flow regulation and water resource development. The flow regime of the study river is described in terms of a set of key flow indicators or statistics of geomorphological and/or ecological relevance (derived from theoretical, empirical and conceptual models). A suite of ‘benchmark models’ is then developed, each model linking the percentage change in a particular flow statistic from its natural (pre-regulation) value with the observed geomorphological and ecological impacts of flow regulation. These models are used to develop a risk assessment framework that guides the evaluation of the potential environmental impacts of future scenarios of water resource development and management (Brizga, 2000). DRIFT (downstream response to imposed flow transformation) also embodies a top-down approach enabling predictions of the ecological impacts of flow regime change (King et al., this issue). In contrast, the flow restoration methodology combines a ‘bottom-up’ and ‘top-down’ approach (Arthington, 1998). It was designed to define more precisely how the flow regime of a regulated river could be changed to achieve defined geomorphological and ecological objectives and an overall shift in the ecological characteristics of the riverine ecosystem towards the natural (pre-regulation) state (Arthington et al., 2000). It has particular relevance to rivers regulated by large dams but could be applied to any river system regulated by infrastructure (e.g. by inter-basin transfer, weirs or levee banks) and to situations of run-of-the-river water abstractions or alterations to wetland water regimes and groundwater tables. One of its special features is that alternatives to flow restoration (e.g. physical habitat restoration) are evaluated when some elements of the pre-regulation flow regime cannot be restored fully for practical or legal reasons. Within each of these holistic, ecosystem approaches, the particular methods, tools and models employed to define the flow requirements of specific ecosystem components vary largely, it seems, as a function of the spatial scale of the study and the time and resources available to undertake it. Such methods may include wetted perimeter analysis, the physical habitat module of IFIM (the instream flow incremental methodology) (Bovee, 1982; Milhouse et al., 1989), the flow events method describing the statistical properties of flow components judged to be of geomorphological or ecological importance (Stewardson and Cottingham, 2002), wetland and riparian water budget analysis (e.g. Pettit et al., 2001) and the characterization of flow pulses designed to mobilize sediments, maintain water quality and substrate composition, inundate channel benches, backwaters and floodplain terraces, provide cues that initiate biological events (e.g. fish migration, spawning, flowring and seed set) and drive ecological processes (energy flow, material exchanges between river channels, floodplains or the downstream estuary; see Loneragan and Bunn, 1999). Although such methods (some involving modelling approaches) are available, many environmental flow studies, short on time and resources, have been forced to rely heavily on the theoretical literature, conceptual models and best professional judgement rather than undertaking new field work or modelling activities. This is particularly the case for many applications of ‘expert’ and ‘scientific panel’ methods, which have become a way of fast-tracking environmental flow evaluations in many circumstances (Cottingham et al., 2002). Environmental flow studies based on a firmer foundation of knowledge (and predictive models) describing the relationships between hydrology and ecological processes in particular rivers are relatively rare in Australia. Applications of the holistic approach in Western Australia (P. M. Davies et al., unpublished report, 1996; Pettit et al., 2001) and in Queensland rivers (for example, the Brisbane River (Arthington et al., 2000) and Logan/Albert River (Arthington and Long, 1997; Arthington and Lloyd, 1998)) are exceptions, drawing on several years of field data. These better grounded studies are rare for one major reason: the past limited commitment of funding agencies to long-term ecological research on rivers and their floodplains in relation to the hydrological regime. Most river research in Australia has been conducted within three-year (or even shorter) funding programmes. The Commonwealth’s ‘Environmental Flows Initiative’, an R&D programme established in 1999 with strong support from scientists and government water managers, has lacked institutional commitment. Recently, Land and Water Australia (the former Land and Water Resources Research and Development Corporation) has resolved to fund environmental flows research (N. Schofield, personal communication, 2002), after a hiatus of some years, giving renewed hope for a more consistent and supportive research programme. Copyright # 2003 John Wiley & Sons, Ltd. River Res. Applic. 19: 377–395 (2003) 384 A. H. ARTHINGTON AND B. J. PUSEY In data-poor settings, expert panel assessments continue to propose probability statements based on ‘rules of thumb’. These serve their immediate purpose (i.e. setting some sort of ‘cap’ on further water diversions, or providing interim targets for flow restoration). However, they may also serve to undermine the past decade of efforts to place environmental flow assessments on a firm scientific foundation, based on understanding the ecological roles of the natural flow regime and using this understanding to produce quantitative descriptions of restorative and protective flow regimes. In over-committed river basins, such as the Murray–Darling, benchmarking could easily be applied, based on the many regulated and degraded sections of this river, followed by application of the flow restoration methodology to define scientifically sound flow restoration targets and scenarios in key reaches and tributaries. One of the features of water management for the environment is the insistence of different jurisdictions that because all rivers are different, it is not possible simply to adopt environmental flow methods and guidelines from other jurisdictions. Thus, a high number of different methods has been developed and used in Australia (37 in total; Tharme, this issue). The general failure of most jurisdictions to learn from the experiences (successful or not) of others typifies the overall water reform programme. In Australia, there has been only one national conference on environmental flows with a published proceedings (AWWA, 1994) since the introduction of the water reforms, and despite the wealth of published and grey literature freely available, and dozens of journal and conference papers, we see much re-invention of the environmental ‘water wheel’. Thus, the state of Victoria is only now proposing its definitive method (called FLOWS) for the estimation of environmental flows, even though the very first of such studies was conducted in that state in the 1980s (e.g. Hall, 1989). In an effort to apply more quantitative methods in Victoria, Stewardson and Cottingham (2002) have developed the flow events method. This provides a means of quantifying the components of flow that ecologists recommend should be retained or restored for specific ecological purposes, with a particular emphasis on maintaining variability. A final point regarding methods is that the techniques of risk assessment so well-developed and effectively applied in relation to water quality management in Australia (National Water Quality Management Strategy) have not been employed to any extent in environmental flow studies. This is surely a conundrum given the pivotal definition of ecological water requirements: ‘descriptions of the water regimes needed to sustain the ecological values of water-dependent ecosystems at a low level of risk’ (ARMCANZ, ANZECC, 1996). The benchmarking methodology appears to be the only structured process to provide a risk assessment framework (based on the assessment of river response to past flow regulation) within which future scenarios of water use can be evaluated in terms of their probable impact on aquatic ecosystems. A best practice framework (Arthington et al., 1998) proposed a logical sequence of activities leading up to, and following, an environmental flow assessment, within which any of the methods or methodologies outlined above could be applied. It includes a process of peer review and a monitoring and feedback loop intended to allow adjustments of environmental flows as new knowledge becomes available through monitoring and research. Some form of peer review process seems necessary and important to ensure scientific rigour, credibility and confidence in water allocation recommendations that may be very costly, practically and politically difficult to achieve, and fraught with social and legal implications. Because doubts have been raised about the scientific foundations of some environmental flow recommendations, scientists are now being pitted against consultants employed by the proponents of new water resource developments, and litigation has already occurred to secure access to water for irrigation purposes (e.g. in the Condamine–Balonne system of the upper Murray–Darling Basin). However, no amount of challenge or litigation will solve the fundamental problem that ecological science cannot yet predict how any stream or river, or any one component of it such as the fish assemblage, or even a single aquatic species, will respond to alterations in the flow regime. PROGRESS WITH FLOW PROTECTION AND RESTORATION IN AUSTRALIA Flow protection There is widespread agreement that it is far cheaper for society to prevent degradation of rivers and their floodplains in the first place than it is to restore degraded aquatic ecosystems. For example, the Snowy River Restoration Project (see below) will cost at least $643 million to achieve total quantifiable benefits of only $61 million (Pigram, Copyright # 2003 John Wiley & Sons, Ltd. River Res. Applic. 19: 377–395 (2003) AUSTRALIAN RIVERS 385 2000). Nevill (2001) has analysed the legislative and institutional background to biodiversity conservation in Australia, and the nation’s progress towards the identification of important aquatic areas for conservation in protected reserves. Even the streams that flow through areas protected by national parks, fauna reserves and Ramsar sites are not protected from all human disturbances. Streams in the Wet Tropics World Heritage Area of northern Queensland are becoming enriched with nutrients emitted by tourists, water is being extracted for resort use and translocated fish are treated as a major recreational resource rather than a threat to native biota. Not a single Australian river system is fully protected by the existing national reserve system, nor does the Ramsar Convention afford protection to many of Australia’s most significant wetlands (McComb and Lake, 1990). Nevill (2001; also P. Cullen, unpublished presentation, 2001) advocates setting up a National Heritage River System, starting with rivers of high conservation value identified by the states and territories in their water allocation programmes, and/or already protected by conservation reserves. These rivers may include arid-zone rivers (Georgina, Diamantina and Cooper Creek), rivers within the Wet Tropics World Heritage Area of Queensland, the Lake Eyre Basin, the Clarence River in New South Wales, the Ovens River in Victoria, rivers within the World Heritage Area of southwestern Tasmania, the Fitzroy River and Shannon River in Western Australia, and the East Alligator River and others in the Northern Territory. Many wetlands recognized as priority areas for conservation (Arthington and Hegerl, 1988; McComb and Lake, 1990), for example the Barmah–Millewa River Red Gum Forest and other significant wetlands on the River Murray, groundwater wetlands in Western Australia and floodplain wetlands in Queensland, as well as many others, could be linked into a national network of aquatic reserves. Quite apart from their heritage values, these rivers and wetlands will serve as the major sources of propagules and colonists for degraded rivers and wetlands that have already lost much of their biological diversity. Flow restoration Restoration is defined by Middleton (1999) as ‘returning a site to a condition similar to the one that existed before it was altered, along with its predisturbance functions and related physical, chemical and biological characteristics. The goal of restoration is to establish a site that is self-regulating and integrated within its landscape, rather than to re-establish an aboriginal condition . . . ’. Rehabilitation is a different process in that it seeks to improve the condition of an area but not necessarily in the direction of the pre-existing undamaged state (Bradshaw, 1997). Environmental flow determinations for Australian regulated rivers are consistent with Middleton’s definition of restoration. They aim to identify and restore flow components that will move a river in the direction of its natural condition, or the best representation of this that can be identified. Many river rehabilitation works have been undertaken in the past, and even today, some aspects of water and river management are more aligned with the goals of rehabilitation sensu Bradshaw (1997). For instance, water management rules have been defined and experiments conducted to achieve the suppression of algal blooms in weirs (e.g. Maier et al., 2001) with the prime objective of preventing the outbreak of potentially toxic cyanobacterial blooms, rather than re-establishing natural communities of algae and periphyton. In southern rivers, flows and habitat have been managed in such a way that exotic trout will continue to flourish and sustain recreational fishing. The CoAG reforms require water allocations to ‘stressed rivers’, wetlands, floodplains, terminal lakes, estuaries and ecosystems dependent primarily upon groundwater. The following provides a brief summary of progress and notable examples of flow restoration studies. Data from the National Land and Water Resources Audit (NL&WRA, 2000) suggest that 26% of rivers and a similar proportion of groundwater systems are presently over-allocated, that is, they cannot sustain the current levels of water use and still maintain ecological values. This is probably a very conservative figure as the methods used to define and estimate ‘sustainable levels of water diversions’ (NL&WRA, 2000) have differed from one state to another, and the amounts of water considered adequate to sustain river and groundwater ecosystems are poorly documented in the Audit reports and remain for the most part obscure. The Commonwealth National Competition Council has recently reported ‘significant progress with land and water reforms’ which cost Australia $2.2 billion annually. All states and territories have introduced new water legislation giving the environment formal, legal water rights, and progress reports suggest that over 50% of Australia’s rivers, wetlands and groundwater systems have been evaluated and will receive some form of environmental water allocation. Copyright # 2003 John Wiley & Sons, Ltd. River Res. Applic. 19: 377–395 (2003) 386 A. H. ARTHINGTON AND B. J. PUSEY The rest of this section of the paper presents several case studies that illustrate the diversity of approaches being used in Australia to restore river flow regimes and aquatic ecosystems, the amounts of water that may be involved and the anticipated ecological outcomes. Snowy River Restoration Project The Snowy River rises in southeastern Australia and flows to the sea at Marlo on the Victorian coast. In 1974, the Snowy Mountains Hydro-electric Scheme was completed at a cost of $800 million, and began to divert over 1 106 Ml of water annually (representing 99% of mean daily flow) from the upper catchment westward to the Murray–Darling Basin. This inter-basin transfer scheme underwrites $1542 million worth of agricultural products from the Murray and Murrumbidgee valleys (Davies et al., 1992; Pigram, 2000). With Jindabyne Dam in place, large floods in excess of 40 000 Ml day1 continue to pass downstream with their frequency unaltered by dams and diversions, but floods in the range of 20 000 Ml and 30 000 Ml day1 occur much less frequently than in the freeflowing river. Their loss has caused a range of geomorphological and ecological impacts exacerbated by changes in catchment land-use. The channel below Jindabyne Dam has become narrow, increasingly vegetated, silted up and less diverse, physically and biologically (Pigram, 2000). Using an expert panel approach and reviews of existing ecological data, scientists have recommended 28% of mean annual flow below Jindabyne as the overall target for restoration of the Snowy’s flow regime. Details of the seasonal distribution of flows, and the volume and duration of particular flow events, are still being decided, but the general principle is to apply a holistic approach and mimic important features of the pre-regulation flow regime. Environmental flows are expected to achieve improvements in thermal regime, channel structure, longitudinal connectivity, dispersal and migration of biota, triggers for fish spawning and the aesthetics of the degraded riverine environment. A monitoring programme has been put in place to ‘benchmark’ the existing condition of the river before the flows are restored and to track the anticipated ecological outcomes. The Murray–Darling Basin and the cap on water diversions The Murray–Darling Basin spreads over an area of 1.073 106 km2 or about 14% of the total area of Australia, and across parts of four Australian states (New South Wales, Victoria, Queensland and South Australia) and all of the Australian Capital Territory. These governments and the Federal Government share responsibility for managing the natural resources of the basin. Some form of inter-state cooperation in the management of the basin’s water resources has been in place since 1902, and from 1985, the Murray–Darling Basin Commission (MDBC) has facilitated institutional arrangements (Close and McLeod, 2000). The aquatic ecosystems of the Murray– Darling Basin have been more affected by flow regulation and water diversions than any other Australian catchment (Kingsford, 2000). Impoundments can store 103% of the mean annual runoff and 81% of total divertible water is extracted, primarily for irrigated agriculture. The mean annual flow at the Murray mouth is only 39% of natural. The very first environmental flow provision in the basin was an allocation of 100 Gl year1 for watering the Barmah–Millewa Forest (65 000 ha) on the River Murray. This forest is the largest contiguous stand of river red gum (Eucalyptus camaldulensis) in the world and is listed under the Ramsar Convention. About 54 waterbird species breed in the forest and a further ten species are non-breeding vagrants or migrants listed in Japan Australia Migratory Birds Agreement (JAMBA) and the China Australia Migratory Birds Agreement (CAMBA). Natural flooding creates suitable nest sites and foraging areas and the fecundity of colonially nesting waterbirds is contingent upon the timing, magnitude, duration and frequency of flooding, and the rate of water level recession (Leslie, 2001). By reducing the number and duration of floods, river regulation has reduced the frequency of successful breeding episodes by more than 80% compared to the natural precedent (Leslie, 2001). The first environmental flow of 100 Gl to the forest in 1998 achieved negligible ecological benefit, and modelling studies have subsequently shown that accumulating this environmental flow provision for a number of years, to allow a larger periodic release, would achieve better environmental outcomes than releasing 100 Gl each year (NRE, 1999). Environmental water releases have now been tailored to satisfy the known water requirements of a range of wetland biota, especially the breeding of waterbirds. In 2000–2001, the first effective natural flood since 1993 was enhanced by the largest environmental water release yet made in Australia (a total of 341 Gl, released in three Copyright # 2003 John Wiley & Sons, Ltd. River Res. Applic. 19: 377–395 (2003) AUSTRALIAN RIVERS 387 parcels of water). This carefully crafted environmental flow strategy enabled at least 15 000 breeding pairs of 20 or more waterbird species to breed successfully in the forest, and one endangered egret species bred for the first time since 1975 (Leslie and Ward, 2002). While the river red gum forests at Barmah may be reasonably well-watered in some years, others further down the Murray River such as the wetlands at Chowilla have been deprived of their natural flood flows, and many other parts of this river system have been adversely affected by flow regulation (Kingsford, 2000). In 1995, following an extensive audit of water use, the Murray–Darling Ministerial Council introduced a basin-wide cap on further water diversions. This ‘cap’ is intended to limit the volume of water diversions to 1993/1994 levels of development. New developments are possible under the cap provided that the water for them is obtained by improving water use efficiency or by purchasing water from existing developments (Close and McLeod, 2000). The cap has been hailed as a major innovation in water management with few international precedents. However, ecologists warn that while the cap is an essential step in slowing ecosystem decline, it cannot be expected to result in marked improvements in riverine health because it has been set at a level of water extraction that contributed to the current degradation of the river system, (Whittington et al., 2000). A recent review of environmental flow allocations for the River Murray calculated that a 40% reduction in the existing cap, or 4000 Gl more water basin-wide, would be needed to restore critical flow volumes and flood pulses of ecologically relevant size and frequency, and significantly improve the ecological condition of the river system (G. Jones et al., unpublished report, 2001). This project used a risk-based approach to evaluate various scenarios of ecosystem recovery, proposing that a regulated river will tend to shift towards a healthy condition if various important flow ‘indicators’ (hydrological statistics) can be restored to two-thirds of their pre-regulation level. The ‘twothirds natural’ level was intended as a possible target for restoration of the ‘working’ (i.e. regulated) River Murray, not a level for ‘acceptable degradation’ or ‘sustainable diversion’ of minimally impacted rivers. It has since been reformulated as the two-thirds rule, expressed as follows (Cullen, 2001): ‘Any river where more than 1/3 of the median flow is extracted is likely to be seriously damaged. In such catchments, there should be an annual clawback of water for the environment until this level is reached or, a sustainable level of extraction is determined through sound research. Compensation should only be paid where there is a genuine legal right to this water’. Cullen’s rule refers to only median annual flow; however, G. Jones et al. (unpublished report, 2001) applied the two-thirds criterion to a range of flow statistics describing important characteristics of the flow regime (magnitude, timing, frequency, duration and rate of rise and fall of flood pulses, seasonal flow patterns, and flow variability) with known or presumed relevance to river ecosystem components. So the proposed two-thirds rule is less of a blunt instrument than might appear. In northern valleys of New South Wales where irrigated cotton is the major crop, environmental water allocations have reduced diversions in the Gwydir Valley to 11% below cap ¼ 37 000 Ml year1 and in the Macquarie Valley to 12% below cap ¼ 54 000 Ml year1. These reductions in water diversions are restoring about 10% of river water on average to the valleys of the upper Murray–Darling system. Despite the small amount of water involved, environmental flow management targets have been established, based around 12 broad River Flow Objectives (Chessman and Jones, 2001). These objectives range from protecting low channel flows and pool water levels and mimicking the natural frequency, duration and seasonal nature of drying periods in temporary streams, to maintaining or restoring the natural inundation patterns and distribution of floodwaters supporting natural wetland and floodplain ecosystems. As well as aiming to mimic natural patterns of river flow in the low-flow range, reservoir water is sometimes stored to create a ‘bank’ of water that can be used for specific environmental purposes, variously referred to as an ‘environmental contingency allowance’ or a ‘wildlife allocation’ (Chessman and Jones, 2001). The former bank ‘credits’ have been used to flush blooms of cyanobacteria, the latter to sustain waterfowl breeding events. This type of release may mimic a natural flow event, or be used more as an intervention in the natural pattern of river flows to achieve what is seen to be a worthwhile environmental goal. Benchmarking flow regulation impacts In Queensland, environmental flow determinations currently apply the benchmarking methodology (Brizga, 2000) to develop a risk assessment framework for the evaluation of future water infrastructure proposals in each catchment. Benchmarking in the Burnett Basin has produced two risk levels, the first defining ‘the benchmarked Copyright # 2003 John Wiley & Sons, Ltd. River Res. Applic. 19: 377–395 (2003) 388 A. H. ARTHINGTON AND B. J. PUSEY level above which assessed sites are more likely to have undergone no/minor change from reference condition due to water resource development’ and the second defining ‘the benchmarked level below which assessed sites are more likely to have major/very major change from reference condition due to water resource development for at least one ecosystem component’ (Brizga, 2000). A key point about the risk levels established for Queensland river systems is the inference that a substantial proportion of the natural flow (about 80–90% of mean annual flow) may need to be retained, without any change in seasonal patterns of flow, in order to maintain ecological values at a low level of risk. The principle that water flowing to the sea is not ‘wasted’ has also been established by demonstrating strong correlations between the magnitude of summer discharge from coastal rivers and commercial fish and crab/ prawn catches in the Logan estuary (Loneragan and Bunn, 1999). The freshwater flow requirements of estuaries are now considered on a routine basis as part of water resource plans in coastal catchments, and the techniques developed in Queensland are also being applied in Western Australia. Brisbane River flow restoration study The flow restoration methodology (Arthington et al., 2000) was developed to advise on the environmental flow requirements of the Brisbane River downstream from Wivenhoe Dam. This storage was constructed primarily to impound flood waters. The dam also provides the major source of potable water to Brisbane city by delivering a steady flow of water (averaging 650 Ml day1) to a treatment station at Mt Crosby, and a small amount of electricity is generated at Splityard Creek Hydropower Station fed by water from Lake Wivenhoe. The study, described in Arthington et al. (2000), required evaluation not only of environmental flow options, but also a review of the operations of all water release rules, particularly those pertaining to flood management (governed by an Act of Parliament), and advice on restoring an inefficient fish ladder at Mt Crosby, as well as an assessment of the potential environmental impacts of electricity generation using the daily flow releases from Lake Wivenhoe. This study demonstrated how the operation of Lake Wivenhoe as a ‘translucent dam’ (i.e. allowing some flows to pass through as and when they occur) could restore not only ecologically important flow magnitudes but also flood pulses and flow variability at various temporal scales (daily, flood events, monthly, and from year to year). Hydrological simulations showed clearly which particular environmental flows could and could not be achieved with the dam operating as a flood mitigation storage and major source of water for Brisbane. Modified dam operating rules, piped water delivery to Mt Crosby, physical channel rehabilitation to replace riffle habitats drowned out by elevated low flows, reconstruction of the fish ladder, and tapping into alternative sources of water for Brisbane, were all explored as part of the project’s recommendations, as were the implications for environment flows of future growth in Brisbane’s urban and rural populations. The Brisbane River study demonstrates rapid progress from simplistic rule-of-thumb guidelines based on hydrological indicators (dismissed in the 1980s by Richardson (1986) and by Arthington and Pusey (1993)) and expert panel approaches involving desktop studies and observation of water releases from dams (admittedly intended to be the start of a more holistic process; Swales and Harris, 1995). Flow restoration studies in Western Australia The holistic approach (Arthington et al., 1992) is being used in Western Australia in flow restoration studies. These identify the important geomorphological and ecological features of aquatic and riparian systems that are dependent upon flow and/or groundwater, and from an understanding of flow–ecology relationships, specific ecological water requirements are defined for in-stream and riparian systems (P. M. Davies et al., unpublished report, 1996; Pettit et al., 2001). Particular issues for Western Australia are groundwater systems and wetlands (Froend and McComb, 1994) and mitigating the downstream ecological effects of water-supply impoundments located on the Darling Scarp. Permanent streams originating on the scarp have very high quality water and contain ancient Gondwanic fauna. Their lower reaches flow across the Swan Coastal Plain, where one of the major allocation issues is maintaining flow permanence and ensuring water quality in pools, particularly dissolved oxygen concentrations over the dry summer months, and thus the capacity of pools to provide summer refugia for fish and invertebrates (P. M. Davies et al., unpublished report, 1996). Modelling techniques are being used to predict dissolved oxygen concentrations in pools based on pool volume, in-flow and out-flow, rates of gross primary production and respiration, day length and water temperature (P. M. Davies et al., unpublished report, 1996; P. Davies, University of Western Australia, personal communication 2002). Copyright # 2003 John Wiley & Sons, Ltd. River Res. Applic. 19: 377–395 (2003) AUSTRALIAN RIVERS 389 HOW MUCH WATER DOES A RIVER NEED? Rivers and their floodplains need their natural flow regime in all of its spatial and temporal variability to maintain their natural ecological integrity and long-term evolutionary potential (see reviews in Arthington et al., 1992; Lake, 1995; Walker et al., 1995; Poff et al., 1997; Richter et al., 1997; Puckridge et al., 1998; Ward et al., 2001; Bunn and Arthington, 2002). A water regime that is near to the natural unregulated state is the target for National Heritage Rivers, Ramsar wetlands and aquatic systems in World Heritage Areas and national parks. Such a target is not impossible, although it will involve some flow restoration activities, and other drivers of ecological processes in rivers and their floodplains (i.e. catchment and riparian vegetation, sediment regime and water quality) must also be kept intact to achieve long-term protection of these systems. In an ideal world, environmental water requirements would be defined, and alternative water resource developments or restoration scenarios evaluated, by means of quantitative predictive models of the relationships between hydrology and the ecological processes governing biological diversity and ecosystem integrity. Such models are not yet available for any Australian river system and many of the available modelling techniques are prohibitively expensive, particularly at a ‘whole-of-catchment’ scale. Scientists can certainly demonstrate from research in unregulated rivers that some fish species spawn during periods of low and relatively stable spring–summer flows (e.g. Milton and Arthington, 1984; Humphries and Lake, 2000; Pusey et al., 2001) whilst for others spawning is triggered by flow pulses, and that waterbird breeding and recruitment in floodplain wetlands are dependent upon the timing, duration and frequency of floods (Leslie, 2001). However, water managers need to know how far these flow attributes can be altered from their natural levels before fish fail to spawn, or recruitment of waterbirds falls off drastically. Likewise, those with the responsibility for river management need to be convinced that restoring 10% or 20% of flow magnitude, or five small floods instead of one larger one, will have measurable ecological benefit (Bunn and Arthington, 2002). The benchmarking methodology applied in Queensland water resource planning appears to offer the most tractable method for making probability statements about the ecological implications of altering a river’s flow regime by specified amounts compared to the natural regime. While the risk assessment models of the benchmarking methodology are superficially similar to the outputs of rule-of-thumb methods such as the Montana method (Tennant, 1976) (i.e. ecologically significant levels of change are related to a set of key flow indicators) there are substantial differences in how the outputs are derived. In particular, benchmarking provides a framework which enables locally applicable models of river condition versus hydrological change to be developed. It also provides a structured approach for incorporating scientific understanding of the complex role of flow in ecosystems that goes well beyond the simple relationships originally recognized by the Montana method. However, until we have a better understanding of how flow–ecology relationships may vary across the individual subcatchments of large river basins with distinctly different flow regimes (see Pusey and Arthington, 1996; Pusey et al., 1998), from basin to basin and between biogeographic regions (e.g. Pusey et al., 2000), the benchmarking methodology must be applied with caution. It should be used only as a ‘holding position’ whilst research and monitoring strengthen our understanding of how rivers have in the past, and will in the future, respond to change in their individualistic flow regimes. In the interim, something must be done, and done quickly, to prevent further deterioration of river condition in most areas of Australia, especially in areas dominanted by irrigated agriculture. The studies conducted to date in Queensland rivers suggest that at least 80–90% of natural flows (as measured by a core set of hydrological indicators) may be needed to maintain a low risk of environmental degradation, based on the Fitzroy, Barron and Burnett River benchmarking studies. In the Top End of the Northern Territory, some rivers are maintained at 80% of their natural flow, whereas in the case studies reviewed above, two-thirds of various flow indicators has been proposed as the restoration target for the River Murray, and a value of only 28% of natural mean annual flow has been negotiated for the Snowy River. It may well be that some rivers can maintain their fundamental ecological integrity with less than 80–90% of natural flows, but this has yet to be demonstrated. Advocating only 66% of a series of flow statistics for a regulated river with significant water infrastructure may help the River Murray to recover some of its natural values, but must not be interpreted as a national ‘target flow’ that could pre-empt more systematic analyses of environmental flow requirements and more generous flow allocations. Furthermore, it must not detract from options for achieving different levels of restoration in various tributaries and reaches of large rivers, and certainly should not be used as a level for sustainable diversion of minimally impacted rivers. Copyright # 2003 John Wiley & Sons, Ltd. River Res. Applic. 19: 377–395 (2003) 390 A. H. ARTHINGTON AND B. J. PUSEY We propose as an immediate alternative to the two-thirds rule: that river-specific benchmark models be developed throughout Australia using well-established quantitative field techniques for the assessment of river condition, and further, that the risk assessments made possible using benchmarking should be used to ‘cap’ further water diversions until more robust estimates of flow requirements can be determined through research. We made a similar recommendation to cap water diversions ten years ago (Arthington and Pusey, 1993). If a national benchmarking programme is put in place, it will be essential for scientists to communicate the message that benchmark levels must be treated as hypotheses rather than as fixed targets for flow restoration and protection. Furthermore, ecologists must be given opportunities to test these hypotheses through short- and long-term experimental research and monitoring, and given the time, funding and institutional support to do so. Flow restoration experiments Conceptual models describing the processes underlying anticipated restoration trajectories abound in the literature and reports from environmental flow studies, giving wide scope for hypothesis-driven flow restoration research and monitoring as an explicit component of flow restoration projects. Although there are significant experimental design issues (few suitable reference systems and limited opportunities for replication), turning flow restoration projects into experiments in restoration ecology is long overdue (Kingsford, 2000; Lake, 2001; Bunn and Arthington, 2002). We recommend the establishment of a series of demonstration flow restoration projects in focus catchments around Australia with significant problems due to flow regulation and realistic opportunities for flow restoration. Each case study could be planned and managed according to a standard format for monitoring and assessment of critical variables related to flow and flow-driven ecological processes. The monitoring and assessment programme should be designed to determine the outcomes of particular changes in key flow characteristics (described by key flow statistics) representing the major types of changes being made in flow regimes in different areas of Australia. The opportunity to build hypothesis testing into flow restoration projects has been taken up in some jurisdictions. The New South Wales IMEF (Integrated Monitoring of Environmental Flows) programme has been designed to link four components: a flow rule or rules; the River Flow Objective (RFO) that the flow rule is designed to achieve; an intended environmental outcome of the RFO; and the biophysical mechanism by which the application of the rule is expected to lead to the outcome. As the authors of IMEF state: ‘Hypothesis testing would enable the expected links between the flow regimes and ecology to be either confirmed or reformulated, regardless of whether or not a long-term ecological improvement is recorded’ (Chessman and Jones, 2001). The Co-operative Research Centre (CRC) for Freshwater Ecology is conducting a long-term flow restoration project in the Campaspe River, where the effects of operating Lake Eppalock as a ‘translucent’ dam (allowing some of the in-flow to the dam to be released downstream as and when it arrives) are being monitored. At present minimal releases are made from the lake outside of the summer irrigation season. The proposed new flow regime will allow 25% of Lake Eppalock’s incoming water to be released between May and October. The objective of the experiment is to restore some of the natural seasonality and variability that existed before the river was impounded and regulated, and by doing so, promote recovery of invertebrate and fish populations (Humphries et al., 1999; Humphries and Lake, 2000). The new flow regime would not compromise existing users’ security of supply because Lake Eppalock has to reach 64% capacity before any environmental water releases are made. Long-term studies of flow restoration (such as the Campaspe Project) can be disrupted by natural variations of rainfall and runoff, and other external influences. To date, the planned experimental flow release from Lake Eppalock has not been made because of water shortages in the catchment. Short-term and small-scale experimental water releases, such as flows designed to trigger fish migration and spawning, offer a more tractable research alternative, even though the range of issues that can be investigated in this way is limited. King et al. (1998) demonstrated that experimental releases from the Clanwilliam Dam on the Olifants River, South Africa, resulted in spawning of the Clanwilliam yellowfish, but suitable water temperatures were also required to ensure egg survival and larval development. Very little manipulative research of this type has been conducted in Australia’s regulated rivers. Long-term ecological research Commitment to restoration experiments has been difficult to achieve in Australia and this remains one of two large gaps in river research. The second major gap is long-term research (Franklin, 1989) on flow-related Copyright # 2003 John Wiley & Sons, Ltd. River Res. Applic. 19: 377–395 (2003) AUSTRALIAN RIVERS 391 ecological processes in rivers that have not been regulated or modified by major catchment disturbances. Research in these systems is essential to provide a baseline understanding of river ecosystem functioning in relation to flow regime, flow history and flow events, and predictive models of flow regulation impacts (Kingsford, 2000; Bunn and Arthington, 2002). Such research is also needed to be able to predict restoration trajectories (Petts, 1987; Lake, 2001; Ward et al., 2001), and to predict the impacts of climate change on river ecosystems as their flow regimes change in response to rising temperature, altered rainfall distributions, increasing evaporation rates and greater water stress. Competing for environmental water allocations will be all the more difficult under future climatic and hydrological scenarios, and aquatic ecologists are not well-equipped scientifically to deal with these added pressures, nor have Australians in general faced up to the ‘adaptation challenge’ (Whetton, 2001) that increased climate variability will present. Saving water for environmental protection How can we ‘claw back’ water for rivers? Cities use 8–10% of Australia’s stored water and up to 80% of such water is used for watering lawns and gardens (Smith, 1998). Most cities have implemented water metering systems and water pricing, and there is a significant and growing shift towards economical native plant gardens and ground covers rather than the European style of brilliant green, over-fertilized and over-watered lawns (Smith, 1998). However the efficiency gains so produced are likely to be absorbed by future population increases, and the drift of urban settlements onto former agricultural holdings. City dwellers are well-disposed to environmental protection measures and many cities and towns exact some form of environmental levy as part of their land ratings. However, saving water in urban areas is only a minor part of the solution to Australia’s water crisis. The agriculture sector uses 79% of stored water. Must this extravagant level of water use continue? Recent research indicates that about 75% of Australia’s agricultural production comes from 25% of farms, representing those that are profitable in the longer term, largely, but not only, because they have access to ample water for irrigation, whereas 25% of farms are marginal in terms of profits and 50% are not viable. Under the CoAG water reforms one might expect that water allocations should be provided only to ecologically sustainable agricultural systems. This may be the trend of the future under the reforms and more enlightened community, economic and political agendas. However, many of the most profitable agricultural systems, for example cotton and sugar, have devastating effects on the aquatic and surrounding riparian/terrestrial environment, manifest in the ‘irrigation syndrome’ and clearly evident in the cotton-growing areas of the Murray–Darling Basin (Arthington, 1996) and around sugar cane plantations in the Burnett catchment and further north (Arthington et al., 1997). If fully implemented and regulated by governments, ecologically sustainable agricultural practices could result in significant savings of water to be restored for environmental purposes. However the battles over water now taking place in such catchments as the Condamine–Balonne and the Burnett suggest that sustainable agricultural systems may be a long while off. The CoAG reforms also provide the principles and the platform for implementing full-cost pricing of water to provide an incentive for efficiencies in water use across all sectors, but particularly in the agricultural sector, where levels of wastage are known to be very high (Postel, 2000). Water savings achieved under government-funded water use efficiency programmes should be returned to the environment, and a percentage of the water acquired via water trading could also be returned. Environmental water banks and water trading The above mechanisms for retrieval of water for rivers and wetlands are likely to take a long time to achieve in Australia, and whilst the various institutional, economic, social and political adjustments are taking place, rivers will continue to degrade. Many rivers, especially those in rural areas, need immediate attention. How can more water for the environment be acquired now? We can see no alternative but for the environment to be a participant in the water market. It would seem fanciful to suggest that, just as farmers are compensated for loss of property, crops and income due to disastrous floods and droughts, the environment, now a legal ‘user’ of water in Australia, should be compensated for its past water losses to off-stream use. However, we suggest the establishment of ‘water banks’, not reservoirs full of water to be used only for environmental purposes (although shares in reservoir water volumes will play a part in providing more water for the environment), but financial banks, operated by ‘river managers’ Copyright # 2003 John Wiley & Sons, Ltd. River Res. Applic. 19: 377–395 (2003) 392 A. H. ARTHINGTON AND B. J. PUSEY with the power and legal right to buy and sell water on behalf of the environment. To resource these water banks, penalties for failure to meet environmental flow targets would be paid to the river managers. Several potential mechanisms to raise funds for the water banks are already well-established in other areas of natural resource management and conservation. As a nation, we could establish a system giving citizens the option to make voluntary contributions on each water bill to fund water banks in rural and other critical areas. Finally, raising large national and international donations to water banks is a very real option, that is already taking effect through the activities of organizations such as World Wildlife and the American Nature Conservation Agency. Backed by such arrangements and adequate start-up funding, environmental water trading could be a viable future scenario for water management in Australia. ACKNOWLEDGEMENTS We thank Dr Jackie King from the University of Cape Town, South Africa, for inviting this paper and her many forms of support and encouragement over the past decade. This paper represents a contribution from the Cooperative Research Centre for Freshwater Ecology and the Co-operative Research for Rainforest Ecology and Management, and we thank both centres for funding support. We also acknowledge in-kind and intellectual contributions from Griffith University and our colleagues there, in particular Professor Stuart Bunn, Mark Kennard, Steve Mackay, Steve and Fiona Balcombe, Fran Sheldon, Glen Wilson, Christopher Irons, Maria Barrett and Lacey Shaw. Thanks are due to Sue Capon for drawing Figure 1. Finally, we are grateful to our many colleagues from other institutions who responded to an informal survey. Many will recognize their contribution in the preceding text. REFERENCES Arthington AH. 1996. The effects of agricultural land use on tributaries of the Darling River, Australia. GeoJournal 40(1–2): 115–125. Arthington AH. 1998. Comparative Evaluation of Environmental Flow Assessment Techniques: Review of Holistic Methodologies. LWRRDC Occasional Paper 26/98. Land and Water Resources Research and Development Corporation (LWRRDC): Canberra. Arthington AH, Blühdorn DR. 1994. Distribution, genetics, ecology and status of the introduced cichlid, Oreochromis mossambicus, in Australia. In Inland Waters of Tropical Asia and Australia: Conservation and Management, Dudgeon D, La P (eds). Mitteilungen (Communications), Societas Internationalis Limnologiae 24: 369–386. Arthington AH, Hegerl EJ. 1988. The distribution, conservation status and management problems of Queensland’s athalassic and tidal wetlands. In The Conservation of Australian Wetlands, McComb AJ, Lake PS (eds). Surrey Beatty and Sons: Sydney; 59–101. Arthington AH, Lloyd R (eds). 1998. Logan River Trial of the Building Block Methodology for Assessing Environmental Flow Requirements: Workshop Report. Centre for Catchment and In-Stream Research and Department of Natural Resources: Brisbane, Queensland. Arthington AH, Long GC (eds). 1997. Logan River Trial of the Building Block Methodology for Assessing Environmental Flow Requirements: Background Papers. Centre for Catchment and In-Stream Research and Department of Natural Resources: Brisbane, Queensland. Arthington AH, Pusey BJ. 1993. In-stream flow management in Australia: methods, deficiencies and future directions. Australian Biologist 6: 52–60. Arthington AH, Zalucki JM (eds). 1998. Comparative Evaluation of Environmental Flow Assessment Techniques: Review of Methods. LWRRDC Occasional Paper 27/98. Land and Water Resources Research and Development Corporation (LWRRDC): Canberra. Arthington AH, Hamlet S, Blühdorn DR. 1990. The role of habitat disturbance in the establishment of introduced warm-water fishes in Australia. In Introduced and Translocated Fishes and their Ecological Effects, Pollard DA (ed.). Bureau of Rural Resources Proceedings No. 8. Australian Government Printing Service: Canberra; 61–66. Arthington AH, Bunn SE, Pusey BJ, Bluhdorn DR, King JM, Day JA, Tharme RE, O’Keeffe JH. 1992. Development of an holistic approach for assessing environmental flow requirements of riverine ecosystems. In Proceedings of an International Seminar and Workshop on Water Allocation for the Environment, Pigram JJ, Hooper BP (eds). The Centre for Water Policy Research, University of New England: Armidale, Australia. Arthington AH, Marshall J, Rayment G, Hunter H, Bunn S. 1997. Potential impacts of sugarcane production on the riparian and freshwater environment. In Intensive Sugar Cane Production: Meeting the Challenges Beyond 2000, Keating BA, Wilson JB (eds). CAB International: Wallingford, UK; 403–421. Arthington AH, Brizga SO, Kennard MJ. 1998. Comparative Evaluation of Environmental Flow Assessment Techniques: Best Practice Framework. LWRRDC Occasional Paper 25/98. Land and Water Resources Research and Development Corporation (LWRRDC): Canberra. Arthington AH, Brizga SO, Choy SC, Kennard MJ, Mackay SJ, McCosker RO, Ruffini JL, Zalucki JM. 2000. Environmental Flow Requirements of the Brisbane River Downstream From Wivenhoe Dam. South East Queensland Water Corporation and Centre for Catchment and InStream Research: Brisbane. Copyright # 2003 John Wiley & Sons, Ltd. River Res. Applic. 19: 377–395 (2003) AUSTRALIAN RIVERS 393 Allan J, Lovett S. 1996. Institutional Impediments to Managing Environmental Water Provisions. Bureau of Resource Sciences: Canberra. ARMCANZ, ANZECC 1996. National Principles for the Provision of Water for Ecosystems. Occasional Paper SWR No 3. ARMCANZ and ANZECC: Canberra. AWWA. 1994. Proceedings of the Environmental Flows Conference. Australian Water and Wastewater Association (AWWA): Artarmon, Victoria. Bayly IAE, Williams WD. 1973. Inland Waters and Their Ecology. Longman, Australia: Camberwell, Victoria. Blanch S, Newton S, Balrd K. (eds). 2001. The Way Forward on Weirs. Inland Rivers Network: Sydney. Boulton AJ, Brock MA. 1999. Australian Freshwater Ecology: Processes and Management. Gleneagles Publishing: Glen Osmond, South Australia. Bovee KD. 1982. A Guide to Stream Habitat Analysis Using the Instream Flow Incremental Methodology. Instream Flow Information Paper 21, FWS/OBS-82/26. US Department of Fisheries and Wildlife Services. Bowling LC, Baker PD. 1996. Major cyanobacterial bloom in the Barwon-Darling River, Australia, in 1991, and underlying limnological conditions. Marine and Freshwater Research 47: 643–657. Bradshaw AD. 1997. What do we mean by restoration? In Restoration Ecology and Sustainable Development, Urbanska KM, Webb NR, Edwards PJ (eds). Cambridge University Press: Cambridge; 8–14. Bren LJ. 1992. Tree invasion of an intermittent wetland in relation to changes in the flooding frequency of the River Murray, Australia. Australian Journal of Ecology 17: 395–408. Brizga SO. 2000. Burnett Basin Water Allocation Management Plan: Proposed Environmental Flow Performance Measures (2 vols). Department of Natural Resources: Brisbane. Bunn SE. 1999. The challenges of sustainable water use and wetland management. In Water: Wet or Dry? Proceedings of the Water and Wetlands Management Conference. Nature Conservation Council of NSW: Sydney; 14–22. Bunn SE, Arthington AH. 2002. Basic principles and ecological consequences of altered flow regimes for aquatic biodiversity. Environmental Management 30: 492–507. Chessman B, Jones H. 2001. Integrated Monitoring of Environmental Flows: Design Report. NSW Department of Land and Water Conservation: Parramatta, New South Wales. Close AF, McLeod AJ. 2000. The Murray-Darling Cap-sharing a natural resource. Water. Journal of the Australian Water Association (July/ August) 27(4): 37–42. Cottingham P, Thoms MC, Quinn GP. 2002. Scientific panels and their use in environmental flow assessment in Australia. Australian Journal of Water Resources 5: 103–111. Cullen P. 2001. The future of flow restoration in Australia. WaterShed (September): 1–2. Cullen P, Lake PS. 1995. Water resources and biodiversity: past, present and future problems and solutions. In Conserving Biodiversity: Threats and Solutions, Bradstock RA, Auld TD, Keith DA, Kingsford RT, Lunney D, Sivertsen DP (eds). Surrey Beatty and Sons: Sydney; 115–125. Davies BR, Thoms M, Meador M. 1992. An assessment of the ecological implications of inter-basin water transfers, and their threats to river basin integrity and conservation. Aquatic Conservation Marine and Freshwater Systems 2: 325–349. Doeg TJ, Davey GW, Blyth JD. 1987. Response of the aquatic macroinvertebrate communities to the dam construction on the Thomson River, southseastern Australia. Regulated Rivers: Research and Management 1: 195–209. Dunbar MJ, Gustard A, Acreman M, Elliott CRN. 1998. Overseas Approaches to Setting River Flow Objectives. R&D Technical Report W6B (96)4. Institute of Hydrology: Wallingford. Finlayson BL, McMahon TA. 1988. Australia vs the world: a comparative analysis of streamflow characteristics. In Fluvial Geomorphology of Australia, Warner RF (ed.). Academic Press: Sydney; 17–40. Franklin JF. 1989. Importance and justification of long term studies in ecology. In Long Term Studies in Ecology: Approaches and Alternatives, Likens G (ed.). Springer Verlag: New York. Froend RH, McComb AJ. 1994. Distribution, productivity and reproductive phenology of emergent macrophytes in relation to water regimes at wetlands of South-western Australia. Australian Journal of Marine and Freshwater Research 45: 1491–1508. Gehrke PC, Brown P, Schiller CB, Moffatt DB, Bruce AM. 1995. River regulation and fish communities in the Murray-Darling River system, Australia. Regulated Rivers: Research and Management 11: 363–375. Gippel CJ, Stewardson MJ. 1995. Development of an environmental flow strategy for the Thomson River, Victoria, Australia. Regulated Rivers: Research and Management 10: 121–135. Gordon ND, McMahon TA, Finlayson BL. 1992. Stream Hydrology: An Introduction for Ecologists. John Wiley and Sons: Chichester. Hall DN. 1989. Preliminary Assessment of Daily Flows Required to Maintain Habitat for Fish Assemblages in the LaTrobe, Thomson, Mitchell and Snowy Rivers, Gippsland. Technical Report Series No. 85, Arthur Rylah Institute for Environmental Research, Victorian Department of Conservation, Forests and Lands: Heidelberg, Victoria. Harris JH. 1984. Impoundment of coastal south-eastern Australia, and a review of its relevance to fish migrations. Australian Zoologist 21: 235– 250. Humphries P, Lake PS. 2000. Fish larvae and the management of regulated rivers. Regulated Rivers: Research and Management 16: 421–432. Humphries P, King AJ, Koehn, JD. 1999. Fish, flows and floodplains: links between freshwater fishes and their environment in the MurrayDarling River system, Australia. Environmental Biology of Fishes 56: 129–151. Jennings JN, Mabbutt JA. 1986. Physiographic outlines and regions. In Australia—A Geography, Vol. 1, The Natural Environment, Jeans DN (ed.). Sydney University Press: Sydney; 80–96. King JM, Louw MD. 1998. Instream flow assessments for regulated rivers in South Africa using the Building Block Methodology. Aquatic Ecosystem Health and Management 1: 109–124. Copyright # 2003 John Wiley & Sons, Ltd. River Res. Applic. 19: 377–395 (2003) 394 A. H. ARTHINGTON AND B. J. PUSEY King JM, Cambray JA, Dean Impson N. 1998. Linked effects of dam-released floods and water temperature on spawning of the Clanwilliam yellowfish Barbus capensis. Hydrobiologia 384: 245–265. King JM, Brown C, Sabet H. (2003). A scenario-based holistic approach to environmental flow assessments for rivers. River Research and Applications 19: 619–639. Kingsford RT. 2000. Ecological impacts of dams, water diversions and river management on floodplain wetlands in Australia. Austral Ecology 25: 109–127. Lake PS. 1995. Of floods and droughts: river and stream ecosystems of Australia. In River and Stream Ecosystems. Ecosystems of the World, Cushing CE, Cummins KW, Minshall GW (eds). Elsevier: Amsterdam; 659–694. Lake PS. 2001. On the maturing of restoration: linking ecological research and restoration. Ecological Management and Restoration 2: 110–115. Leslie DJ. 2001. Effect of river management on colonially-nesting waterbirds in the Barmah–Millewa Forest, south-eastern Australia. Regulated Rivers: Research and Management 17: 21–36. Leslie DJ, Ward KA. 2002. Murray River environmental flows 2000–2001. Ecological Management and Restoration 3: 221–223. Loneragan NR, Bunn SE. 1999. River flows and estuarine ecosystems: implications for coastal fisheries from a review and a case study of the Logan River, southeast Queensland. Australian Journal of Ecology 24: 431–440. Lundqvist, J. 1998. Avert looming hydrocide. Ambio 27: 428–433. Mackay SJ, Arthington AH, Werren G. 2001. Ecological impact of weirs in the Pioneer Catchment, Queensland. In The Way Forward on Weirs, Blanch S (ed.). Inland Rivers Network: Sydney; 39–58. Maheshwari BL, Walker KF, McMahon TA. 1995. Effects of river regulation on the flow regime of the River Murray, Australia. Regulated Rivers: Research and Management 10: 15–38. Maier HR, Burch MD, Bormans M. 2001. Flow management strategies to control blooms of the Cyanobacterium, Anabaena circinalis, in the River Murray at Morgan, South Australia. Regulated Rivers: Research and Management 17: 637–650. McComb A, Lake PS. 1990. Australian Wetlands. Angus and Robertson: North Ryde, New South Wales. Middleton B. 1999. Wetland Restoration, Flood Pulsing, and Disturbance Dynamics. Wiley: New York. Milhous RT, Updike MA, Schneider DM. 1989. Physical Habitat Simulation System Reference Manual—Versions 2. Instream Flow Information Paper 26. USD 1 Fish Wildlife Service Biological Report 89. Milton DA, Arthington AH. 1984. Reproductive strategy and growth of the crimson-spotted rainbow fish, Melanotaenia splendida fluviatilis (Castelnau) (Pisces: Melanotaeniidae) in south-eastern Queensland. Australian Journal of Marine and Freshwater Research 35: 75–83. Naiman RJ, Bunn SE, Nilsson C, Petts GE, Pinay G, Thompson LC. 2002. Legitimizing fluvial ecosystems as users of water. Environmental Management 30: 455–467. Nevill J. 2001. Freshwater Biodiversity: Protecting Freshwater Ecosystems in the Face of Infrastructure Development. Water Research Foundation of Australia, Australian National University: Canberra. Nix HA. 1981. The Environment of Terra Australis. In Ecological Biogeography of Australia, Vol. 1, Keast A (ed.). W. Junk: The Hague; 103– 133. NL&WRA (National Land and Water Resources Audit). 2000. Water Resources in Australia. National Land and Water Resources Audit, c/o Land & Water Australia, Canberra, ACT (Australian Capital Territory), published on behalf of the Commonwealth of Australia. http: audit. ea.gov.au/ANRA/water/docs/national/Water_Contents.html NRE. 1999. Entitlements to the Murray: Outcomes of Work to Define How Victoria’s River Murray Water is to be Shared. Department of Natural Resources and Environment: Victoria. Pettit NE, Froend RH, Davies PM. 2001. Identifying the natural flow regime and the relationship with riparian vegetation for two contrasting Western Australian rivers. Regulated Rivers: Research and Management, 17: 201–215. Petts GE. 1987. Time-scales for ecological changes in regulated rivers. In Regulated Streams. Advances in Ecology, Craig JF, Kemper JB (eds). Plenum Press: New York; 257–266. Pigram JJ. 2000. Viewpoint—Options for rehabilitation of Australia’s Snowy River: An economic perspective. Regulated Rivers: Research and Management 16: 363–373. Poff NL, Allan JD, Bain MB, Karr JR, Prestegaard Kl, Richter BD, Sparks RE, Stromberg JC. 1997. The natural flow regime: a paradigm for river conservation and restoration. BioScience 47: 769–784. Postel SL. 2000. Entering an era of water scarcity: the challenges ahead. Ecological Applications 10: 941–948. Puckridge JT, Sheldon F, Walker KF, Boulton AJ. 1998. Flow variability and the ecology of large rivers. Marine and Freshwater Research 49: 55–72. Pusey BJ, Arthington AH. 1996. Variability of flow regimes in the Burdekin River catchment. Proceedings of the 23rd Hydrology and Water Resources Symposium. Institution of Engineers, Australia: Barton ACT; 213–219. Pusey BJ, Arthington AH, Read MG. 1993. Spatial and temporal variation in fish assemblage structure in the Mary River, south-east Queensland: the influence of habitat structure. Environmental Biology of Fishes 37: 355–380. Pusey BJ, Arthington AH, Read MG. 1995. Species richness and spatial variation in fish assemblage structure in two rivers of the Wet Tropics of Northern Queensland, Australia. Environmental Biology of Fishes 42: 181–199. Pusey BJ, Arthington AH, Read MG. 1998. Freshwater fishes of the Burdekin River, Australia: biogeography, history and spatial variation in assemblage structure. Environmental Biology of Fishes 53: 303–318. Pusey BJ, Kennard MJ, Arthington AH 2000. Discharge variability and the development of predictive models relating stream fish assemblage structure to habitat in north-eastern Australia. Ecology of Freshwater Fishes 9: 30–50. Pusey BJ, Arthington AH, Bird J, Close PG. 2001. Reproduction in three species of rainbowfishes (Melanotaeniidae) from rainforest streams in north-eastern Queensland, Australia. Ecology of Freshwater Fish 10: 75–87. Copyright # 2003 John Wiley & Sons, Ltd. River Res. Applic. 19: 377–395 (2003) AUSTRALIAN RIVERS 395 Quinn GP, Hillman TJ, Cook R. 2000. The response of macroinvertebrates to inundation in floodplain wetlands: a possible effect of river regulation? Regulated Rivers: Research and Management 16: 469–477. Richardson BA. 1986. Evaluation of in-stream flow methodologies for freshwater fish in New South Wales. In Stream Protection: The Management of Rivers for Instream Uses, Campbell IC (ed.). Water Studies Centre, Chisholm Institute of Technology: Victoria; 143–167. Richter BD, Baumgartner JV, Wigington R, Braun DP. 1997. How much water does a river need? Freshwater Biology 37: 231–249. SEAC (State of the Environment Advisory Council). 1996. State of the Environment Australia 1996. CSIRO Publishing: Melbourne. Sheldon F, Walker KF. 1997. Changes in biofilms induced by flow regulation could explain extinctions of aquatic snails in the lower River Murray, Australia. Hydrobiologia 347: 97–108. Sheldon F, Thoms MC, Berry O, Puckridge J. 2000. Using disaster to prevent catastrophe: referencing the impacts of flow changes in large dryland rivers. Regulated Rivers: Research and Management 8: 95–299. Smith DI. 1998. Water in Australia: Resources and Management. Oxford Sheldon University Press: Melbourne. Sparks RE. 1995. Need for ecosystem management of large rivers and floodplains. BioScience 45: 168–182. Stewardson MJ, Cottingham P. 2002. A demonstration of the flow events method: environmental flow requirements of the Broken River. Australian Journal of Water Resources 5: 33–47. Swales S, Harris JH. 1995. The Expert Panel Assessment Method (EPAM): a new tool for determining environmental flows in regulated rivers. In The Ecological Basis for River Management, Harper DM, Ferguson AJD (eds). John Wiley and Sons: Chichester; 125–134. Tennant DL. 1976. Instream flow requirements for fish, wildlife, recreation and related environmental resources. Fisheries 1: 6–10. Tharme RE. 1996. Review of International Methodologies for the Quantification of the Instream Flow Requirements of Rivers. Water Law Review: final report for policy development, South African Department of Water Affairs and Forestry. Freshwater Research Unit: University of Cape Town: Pretoria. Tharme RE. 2000. Review of International Methodologies for the Quantification of the Instream Flow Requirements of Rivers. Water Law Review: final report for policy development, South African Department of Water Affairs and Forestry. Freshwater Research Unit: University of Cape Town: Pretoria. Tharme RE. 2003. A global perspective on environmental flow assessment: emerging trends in the development and application of environmental flow methodologies for rivers. River Research and Application 19: 397–441. Thoms MC, Sheldon F. 2000. Water resource development and hydrological change in a large dryland river: the Barwon-Darling River, Australia. Journal of Hydrology 228: 10–21. Thoms MC, Sheldon F, Roberts J, Harris J, Hillman TJ. 1996. Scientific Panel Assessment of Environmental Flows for the Barwon-Darling River. New South Wales Department of Land and Water Conservation: Sydney. Walker KF. 1985. A review of the ecological effects of river regulation in Australia. Hydrobiologia 125: 111–129. Walker KF, Thoms, MC. 1993. Environmental effects of flow regulation on the River Murray, South Australia. Regulated Rivers: Research and Management 8: 103–119. Walker KF, Sheldon F, Puckridge JT. 1995. An ecological perspective on dryland rivers. Regulated Rivers: Research and Management 11: 85– 104. Ward JV, Standford JA. 1995. Ecological connectivity in alluvial river ecosystems and its disruption by flow regulation. Regulated Rivers: Research and Management 11: 105–119. Ward JV, Tockner U, Uehlinger U, Malard F. 2001. Understanding natural patterns and processes in river corridors as the basis for effective river restoration. Regulated Rivers: Research and Management 117: 311–323. Water Conservation and Irrigation Commission. 1971. Water Resources of New South Wales. Australian Government Printer: Sydney. Whetton P. 2001. Will climate variability change in the future? Climag 9: 1–2. Whittington J, Cottingham P, Gawne B, Thoms M, Walker K. 2000. Ecological sustainability of rivers of the Murray-Darling. In Review of the Operation of the Cap. Companion Paper No. 1. Murray-Darling Basin Ministerial Council: Canberra. Williams WD. 1980. An Ecological Basis for Water Resource Management. Australian National University Press: Canberra. Copyright # 2003 John Wiley & Sons, Ltd. River Res. Applic. 19: 377–395 (2003)