Climate-change threats to native fish in degraded rivers and floodplains of the Murray–Darling Basin, Australia

Additional keywords: conceptual models, native fish, regionalisation, riparian vegetation, water-resource development.

Climate change, with associated changes in land use, atmospheric CO₂ concentration, nitrogen deposition and acid rain, as well as introductions of exotic species, is considered one of the most important determinants of current declines in global biodiversity (Sala et al. 2000). The challenge for ecologists is to predict the likely responses of species and communities to climate change over and above the background natural ecological variability (Verschuren et al. 2000) and, in heavily modified areas, above and beyond the effects of human development.

Humans already withdraw ~50% of available freshwater resources globally (Szöllosi-Nagy et al. 1998), with requirements increasing almost 10-fold during the 20th century (Biswas 1998). The effects of this development on aquatic ecosystems have been severe, evidenced by the numerous global and regional assessments that depict accelerating losses of biodiversity and declines in ecosystem function (Vörösmarty et al. 2000). These impacts in freshwater ecosystems stem from changes to the natural flow regime (Poff and Zimmerman 2010), loss of river–floodplain connectivity (Tockner et al. 2008) and channelisation and construction of in-stream barriers, to name a few. The added effects of climate change in already stressed ecosystems will only further exacerbate the decline in freshwater biota (Pittock and Finlayson 2011).

In Australia, most of the larger rivers in the south-east of the country suffer from extensive agricultural and rural development and a consequent decline in river ecosystem health (Mercer and Marden 2006). The most notable are the rivers of the Murray–Darling Basin (MDB) located in the south-east of Australia. The MDB occupies ~14% of the continent's surface area (1.07 million km²), produces ~40% of the country's annual agricultural production (Crabb 1997) and annually contributes ~AU$18 billion worth of produce to the national economy (MDBA 2010). As a focus for intensive agricultural production, there is intense interest in the condition of aquatic ecosystems in the MDB, as well as the possible impacts of climate change (Davies et al. 2010).

There has been widespread degradation of both riverine and floodplain biota in the MDB, with large tracts of floodplain forest transitioning to terrestrial ecosystems (Pittock and Finlayson 2011). In their assessment of fish species, hydrology and macroinvertebrates, Davies et al. (2010) found 20 of the 24 river basins to be in poor or very poor condition. The impacts of flow modifications, including thermal pollution, have been implicated in the demise of native fish in the MDB-regulated rivers, because of their impacts on physiology, spawning and movement (Gehrke and Harris 2001; Growns 2008). Native fish numbers are now only 10% of pre-European levels (MDBC 2004) because of factors such as changes to the natural flow regime, habitat degradation and barriers to biological connectivity (Pratchett et al. 2011). Assessing how climate change will affect fish species in the MDB must consider these previous and continuing impacts.

Much in-stream fish habitat has been lost as a result of wood removal across the MDB (Koehn et al. 2004). Disruption of riparian–riverine linkages and the loss of riparian and floodplain flora will continue to result in impacts to MDB fish, aside from the direct impacts of river regulation. In the present review, we begin by exploring the effects of water resource and climate change on riparian and in-stream vegetation in the MDB. We then assess how climate change might alter native fish assemblages across the MDB by focussing on 18 individual species, spanning the majority of fish within any local assemblage within the MDB (largely on the basis of distribution maps from Lintermans (2007)). These impacts are presented in a regional context to demonstrate how responses are likely to differ across climate zones and levels of water-resource development.

Projected changes to temperature, rainfall and evaporation, particularly in inland Australia, will alter the frequency and magnitude of heavy rain events, floods and droughts, and affect runoff, soil moisture and salinity on a regional basis (Pittock 2003). In the northern and southern MDB, declines of between 8% and 12% of median runoff, with <5% change in the central, eastern and southern upland regions, are predicted (PMSEIC 2007). The frequency of drought is predicted to increase by 20–40% by 2030 (compared with 1975–2000 average) (PMSEIC 2007). Most global climate models indicate that future winter rainfall is likely to be lower across the MDB, particularly in the southern MDB (CSIRO 2008), which is significant because most of the rainfall and runoff occurs in winter in this region.

Although the median 2030 climate projections suggest an overall reduction of 11% in average surface-water availability across the MDB, there are strong regional variations, with reductions of 3% in the Paroo catchment, 12% for the Murray River at Wentworth and 21% in the Wimmera (CSIRO 2008). These changes in runoff- and surface-water availability must be set against a backdrop of historical levels of water use in the basin of ~50%, which have already caused widespread changes in the nature of riverine and floodplain ecosystems.

River red gums are the iconic trees of the MDB and contribute significant structural woody habitat and finer organic material (Koehn et al. 2004) that provide shelter from predators, food and nutrients. River red gums form extensive forests within the Murray and Murrumbidgee catchments (e.g. Barmah–Millewa Forest) and fringe most of the main rivers and tributaries throughout the basin (Roberts 2001). The species is found across a large temperature gradient and can utilise a range of water sources, including surface- and groundwater sources (Thorburn et al. 1994). Their distribution, density and condition, however, are closely linked to flooding, particularly flood frequency (Roberts 2001). Consequently, changes in flood frequency are usually associated with changes in river red gum condition. In the Barmah Forest, for example, river red gum trees have historically withstood an absence of flooding for up to 18 months during droughts in 1982 and perhaps also during droughts of a similar duration in 1904, 1915, 1944 and 1967 (Bren et al. 1987). The extreme drought in the lower MDB since the mid-1990s, however, has stressed or killed a large number of river red gums.

Impacts of water-resource development on river red gums have occurred as a result of changes in flood frequency, duration and magnitude (Cunningham et al. 2007). Where parts of the floodplain have been permanently inundated, large stands of river red gum have perished. Individual trees and river red gum forests in other locations are stressed because of a lack of flooding of sufficient duration and frequency. Reductions in flooding associated with river regulation and water abstraction have probably impaired recruitment. Increased groundwater salinity, also associated with river regulation in parts of the MDB, has further contributed to declining river red gum health (Overton et al. 2006).

Climate-change impacts on river red gums are most likely to manifest themselves through changes in the frequency and magnitude of floodplain inundation events. Trees and forests already stressed by water-resource development may exhibit mass mortality as a result of the increased stress associated with further reductions in flooding caused by climate change. Increased duration of dry spells, along with higher temperatures, may reduce survivorship of seedlings. Such impacts to river red gums, leading to less structural woody habitat, will have significant implications for fish, such as Murray cod (Maccullochella peelii) (Koehn 2009) and trout cod (Maccullochella macquariensis) (Nicol et al. 2007). Reduced inputs of organic matter from finer material such bark and leaves will also have an impact on food supply for small-bodied fish and cascade upwards through food webs, whereas the loss of stream-side canopy may increase algal growth and further exacerbate rising water temperatures as a result of reduced shading.

Lignum shrubland occurs throughout the MDB, with large stands on the floodplains of the lower Warrego River, Barwon–Darling River and northern Darling tributaries (Roberts 2001). In the eastern MDB, it fringes the more ephemeral or temporary waterbodies. Lignum is very tolerant of drought conditions and can survive as a leafless shrub for several years on low rainfall with little to no flooding (Craig et al. 1991; Capon 2003). After long dry spells, lignum responds quickly to either flooding or rainfall through rapid leaf growth and flowering (Capon et al. 2009). Lignum shrubland is particularly significant in the MDB as breeding habitat for colonial waterbirds, especially where it comprises large, mature shrubs in intermediately flooded areas (Capon et al. 2009).

Impacts of water-resource development on lignum shrubland in the MDB are likely to have been minimal in comparison with the widespread historic clearing of this species. Lignum is intolerant of prolonged flooding (Roberts 2001; Capon 2003), so reductions in flood frequency and duration associated with water-resource development do not appear to have resulted in significant declines in the health of remaining lignum stands. Instead, this may have allowed lignum to increase its range into areas that were previously flooded more frequently or for longer durations, such as in the terminating wetlands of the northern MDB dryland rivers.

Higher air temperatures and reductions in rainfall and flood frequency and duration associated with climate-change scenarios may alter the character and condition of lignum stands throughout the MDB and, therefore, their value as habitat. Lignum recruitment appears to be closely linked to hydrology (Capon et al. 2009) and further alterations to flooding patterns may change the population structure of lignum stands as well as facilitating the further encroachment into channels and open water areas by seedlings. Such impacts are likely to have implications for fish populations in the MDB by altering habitat and food-web structure, via flow-on effects of impacts to piscivorous waterbirds that rely on lignum shrublands for habitat and via changes to inputs of organic material (e.g. lignum leaves).

Emergent macrophytes, including grasses, sedges, reeds and rushes, are widespread across the MDB's floodplains, wetlands and waterway fringes. Prominent and ecologically significant species include grasses such as couch (Cynodon dactylon), Moira grass (Pseudoraphis spinescens), water couch (Paspalum distichum) and common reed (Phragmites australis) as well as bulrushes (Typha spp.) and sedges such as Eleocharis spp. and Cyperus spp. Most species exhibit some tolerance to both inundation and drying, although life-history responses vary widely, reflected by species distributions (Roberts 2001). Some species (e.g. Moira grass and water couch) require frequent, seasonal inundation, whereas others (e.g. common reed) can persist through considerable periods of drought both as mature plants and as persistent rhizomes in the soil. In contrast, bulrushes grow in water to 2 m deep and, although they can tolerate water regimes that vary from permanently wet to seasonally or periodically dry, they do not tolerate permanent flooding over 2 m deep and can only tolerate dry conditions for short periods after the growing season (Roberts 2001).

The impacts of water-resources development are likely to have varied amongst species in this group, depending on their particular environmental tolerances and life histories. Anecdotal evidence, for instance, suggests that common reed was historically more widespread in terminal floodplains of the MDB and the decline in the extent of this species may reflect reductions in flood frequency, as well as grazing pressure, in such areas (Roberts 2001). Moira grass plains on floodplains of the Murray River are also threatened as a result of water-resources development, both as a direct result of changes in flood frequency, duration and timing on the life history of this species, as well as indirectly as a result of their encroachment by both river red gum and giant rush because of altered flooding patterns (Bren 1992; Roberts 2001).

Further reductions in the duration and frequency of flooding, which are likely under many climate-change scenarios, will have a major impact on many plant species in this group, and further reductions in the extent of Moira grass plains and stands of common reeds are of particular concern. Higher air temperatures, combined with lower humidity, have the potential to affect germination and regeneration of these species through reduced soil moisture. Survival of persistent propagules (e.g. rhizomes) may also decline as a result of declining soil moisture and flood frequency. Bulrushes are also likely to be affected by altered flooding patterns associated with water-resources development and climate change.

Although Typha stands may well expand in certain regions of the MDB under climate-change scenarios, favoured by warm, nutrient-rich conditions, in other areas where water levels and soil moisture decline and the periods between flooding are extended, Typha stands may disappear. Such loss of grasses and reeds, particularly in terminating and off-channel wetlands, are likely to have a significant impact on smaller-bodied fish species, such as pygmy perch and carp gudgeons, that rely on them for both cover and food. Furthermore, plants such as reeds and grasses that fringe river channels also provide significant organic inputs to fish food webs; losses would reduce fish abundance and diversity (Fig. 1).

Fig. 1.  Conceptual model of some possible impacts of river regulation and climate change on fish the diversity and abundance of species in the Murray–Darling Basin. Figure adapted and simplified from Pusey and Arthington (2003) and Boulton and Brock (1999).

Environmental filters have been used to predict changes in fish species presence on the basis of their physiological and habitat requirements (Poff 1997; Bond et al. 2011). However, patchy information on environmental tolerances of many MDB fish taxa limits predictions of impacts from climate change. Many MDB fish species appear to be ecological generalists (flexible breeding strategies, diet and habitat), given their wide distribution across varying climatic and geographical regions (Lintermans 2007). Therefore, the main climate-change drivers for generalist fish within the MDB will relate to flow (and therefore changes in habitat availability) and to physiological impacts felt through increased water temperatures influencing spawning times and reduced oxygen levels impairing fitness and survival (Fig. 2; Pankhurst and Munday 2011). Given the over-riding impacts of current levels of water-resource development on fish throughout the MDB, our predictions of climate-change impacts acknowledge how individual species have already been affected, and will continue to be so, by water-resource development (MDBC 2004).

Fig. 2.  Conceptual model of the predicted impact of climate change (mediated though changes in precipitation, increased temperature and increased evaporation) on the diversity and abundance of fish species. Δ = change, ↑ = increase.

The impact of climate change on invasive species will also be relevant to native fish. Invasive species are expected to undergo a mix of responses to climate change, reflecting the diverse range of climatic conditions under which they originally evolved (Rahel and Olden 2008). For example, Bond et al. (2011) predicted a contraction in the distribution of brown (Salmo trutta) and rainbow trout (Oncorhynchus mykiss) in the southern part of the MDB. The reduction of these species, especially in upland regions, would be expected to benefit many native species. Morrongiello et al. (2011) suggested that gambusia may benefit from increased water temperatures, again, especially in the south. Any increase in the distribution and abundance of this species is likely to be negative, especially for small-bodied native fish. Common carp (Cyprinus carpio) has opportunistic recruitment patterns and high rates of population growth that allow it to persist in suboptimal conditions (Koehn 2004). Hence, this fish species is unlikely to be affected by climate change and will probably increase in some places, as a result of increased temperature and thus breeding and growth response when periodic recruitment opportunities (occasional floods) occur.

On the basis of the conceptual model in Fig. 2, we developed five derivative models for different groups of MDB fish species. Fish were grouped according to ecological similarities in diet, breeding strategies (such as spawning flexibility and type of spawning cue), habitat associations (e.g macrophytes, snags, open water) and temperature tolerances (e.g cold water and warm water). For brevity, only two of these models are presented here to demonstrate how we assessed climate-change impacts on species diversity and abundance.

The first model predicts the abundance of Murray cod (Fig. 3). This model demonstrates two pathways of effect, as follows: (1) decreased precipitation leading to reduced water levels and flow, affecting spawning and recruitment, and (2) increased temperature influencing the timing of spawning and physiological tolerance, ultimately reducing Murray cod abundance (Table 1). Overall, Murray cod will decline across the MDB, especially in the northern regions where reduced flows will exacerbate impacts of loss of hydrological connectivity and reduce thermal (summer) refugia because of the loss of deep pools (Table 1). There may be some localised increases in populations where stream temperature rises partially negate impacts of cold-water pollution below bottom-release dams (Table 1). Additionally, there are likely to be longer-term indirect impacts throughout the MDB where riparian and floodplain trees are lost, given they provide significant habitat and food for Murray cod (Table 1).

Fig. 3.  Simplified conceptual model of climate-change impacts on Murray cod in the Murray–Darling Basin. ↑ = increase, ↓ = decrease.

Table 1.  Predicted effects of climate change on the main drivers of abundance and distribution of specific native fish species in the Murray–Darling Basin (MDB)
↓ = decreased distribution and abundance, ↑ = increased distribution and abundance. Where there is both a positive and negative response this will be region-specifi

Table 1 cont.   Predicted effects of climate change on the main drivers of abundance and distribution of specific native fish species in the Murray–Darling Basin (MDB)
↓ = decreased distribution and abundance, ↑ = increased distribution and abundance. Where there is both a positive and negative response this will be region-specifi

Table 1 cont.   Predicted effects of climate change on the main drivers of abundance and distribution of specific native fish species in the Murray–Darling Basin (MDB)
↓ = decreased distribution and abundance, ↑ = increased distribution and abundance. Where there is both a positive and negative response this will be region-specifi

Our second example (Fig. 4) demonstrates some of the key drivers influencing five medium- to large-bodied fish species that are either widely distributed across the whole MDB (yellowbelly (Macquaria ambigua), eel-tailed catfish (Tandanus tandanus) and silver perch (Bidyanus bidyanus)) or across the northern MDB (Hyrtl's tandan (Neosilurus hyrtlii) and spangled perch (Leiopotherapon unicolor)) (Table 1). Increased temperature will directly influence predator–prey relationships by affecting lower levels of the food web, including primary productivity, potentially creating more niches for these fish. Changes in precipitation and evaporative loss will reduce the amount of water in the river, impairing connectivity and recruitment of these species. In the northern MDB, increased temperatures will allow species such as yellowbelly, spangled perch and Hyrtl's tandan with high temperature tolerances and opportunistic breeding strategies to take advantage of the high primary productivity (Table 1). Silver perch and eel-tailed catfish are unlikely to gain much benefit from such changes in the northern MDB because the drier climate will have a further impact on hydrology, particularly connectivity (Table 1). In the southern MDB, however, these two species will probably benefit from increases in stream temperature, given how much they have been affected by cold-water pollution (Table 1).

Fig. 4.  Simplified conceptual model of climate-change impacts on yellowbelly, eel-tailed catfish, silver perch, Hyrtl’s tandan and spangled perch in the Murray–Darling Basin. ↑ = increase, ↓ = decrease.

Of the remaining species presented in Table 1, we grouped them as follows: Group A, including carp gudgeons and bony bream; Group B, including river blackfish, two-spined blackfish, trout cod, mountain galaxiids and Macquarie perch; and Group C, containing Australian smelt, un-specked hardyhead, southern pygmy perch, flat-headed gudgeon and Murray–Darling rainbowfish. Group A fish tolerate an extremely wide range of environmental conditions and their abundance is often strongly linked to levels of primary productivity (Table 1). The increase in temperature resulting from climate change should, therefore, increase the food base of these fish directly and any change to flows is likely to have no major influence on them (Table 1).

The distributions of Group B fish all extend into the cooler upland streams of the MDB (Lintermans 2007) and these fish could be loosely termed as ‘cold-water tolerant’. Increased stream temperatures are likely to have a direct physiological effect on these fish, particularly where they are located at their high temperature limits (e.g. river blackfish and mountain galaxiids in the northern MDB, Table 1). In the southern MDB, increased temperatures may lead to increased abundances in local populations of fish such as trout cod, two-spined blackfish and Macquarie perch affected by cold-water pollution (Table 1). Such changes in temperature regime could affect larval recruitment success of species such as trout cod and Macquarie perch with short breeding seasons, by affecting the timing of zooplankton emergence, for example (Table 1). It is likely that all of the species within this group have been largely affected by the presence of exotic trout, so the expected decline of trout (Bond et al. 2011) because of increased stream temperatures may lead to more native species in these rivers (Table 1).

Group C comprises smaller-bodied fish, often found in wetlands and off-river channel waterbodies often associated with macrophytes (Table 1). The reduction since European settlement of species such as Murray–Darling rainbowfish, un-specked hardyhead and southern pygmy perch can be largely attributed to a loss of connectivity within riverine ecosystems (Table 1). Distributions of these species will become further fragmented as flow diminishes as a result of climate change. Losses of floodplain vegetation (e.g. rushes and reeds in floodplain wetlands) will also further affect fish such as flat-headed gudgeons and southern pygmy perch associated with macrophytes (Table 1). Increased temperatures will also have an impact on small-bodied fish with short breeding seasons and relatively narrow diet ranges, such as Australian smelt, also found in many floodplain habitats (Table 1). These fish are likely to be affected because any shift in seasonal thermal regime is likely to disrupt the synchrony between the seasonal peak in smelt larvae and specific species and size classes of their zooplankton prey.

Given that the impacts of climate change are predicted to vary with latitude as well as with the degree of flow modification, we classified the MDB into 10 regions on the basis of their latitudinal position and the degree of flow management (Fig. 5). Ecological communities will tend to vary over an elevation gradient, so we used the largely elevation-driven regionalisation boundaries in the Sustainable Rivers Audit (Davies et al. 2008) to differentiate uplands from lowlands and to distinguish the alpine region. By using this classification and the predicted effects of climate change on individual species (Table 1), we can predict regional changes in fish assemblages throughout the MDB.

Fig. 5.  Location of the Murray–Darling Basin within Australia and its subsequent classification into different regions on the basis of latitudinal position, altitude and degree of flow management.

Both of these regions, located in the northern MDB, with rivers ultimately flowing into the Darling River (Fig. 5), represent highly regulated rivers because of the presence of water storages (Table 2). There are significant effects of cold-water pollution in these regions, owing to bottom-release off-takes in water storages. Therefore, any warming as a result of climate change will confer an advantage to those fish species that have been reduced in their range and abundance from cold-water pollution. It would be expected that both range and abundance of Murray cod and river blackfish could increase slightly in both of these regions because increased water temperatures partially mitigate effects such as reduced spawning and growth from thermal pollution (Table 1). Although increased temperatures could also facilitate the range expansion of generalist fish species such as yellowbelly, eel-tailed catfish, silver perch and bony bream, these highly regulated rivers are likely to have less water overall. As a result, drought refugia will dry out faster, leaving fewer suitable habitats across the landscape and probably cancelling out any gains made though increased water temperature.

Table 2.  Overall changes in fish assemblage in 10 regions, owing to the effects of climate change and relevant hydrological variation from water-resources development (WRD)
Assemblage changes are predicted from the potential impacts on individual species given in Table 1

Table 2 cont.  Overall changes in fish assemblage in 10 regions, owing to the effects of climate change and relevant hydrological variation from water-resources development (WRD)
Assemblage changes are predicted from the potential impacts on individual species given in Table 1

The rivers of the southern MDB comprise a mixture of regulated and unregulated rivers. Most predicted changes in this region relate to increased water temperatures, although in the regulated rivers, effects will be further exacerbated by reduced flows (Table 2). These increases are likely to affect cold water-tolerant species such as river and two-spined blackfish and trout cod, reducing their breeding activities and thus overall abundance (Table 1). However, balanced against these predicted reductions in native species will be the likely displacement of exotic trout (Bond et al. 2011), which currently has a significant impact on native riverine food webs as a major predator of invertebrates and fish (Lintermans 2007). Such an impact on trout could facilitate increases in vulnerable species such as mountain galaxias and the two blackfish species.

The Alpine region contains many headwater streams of the Murray and Murrumbidgee rivers in the eastern MDB (Fig. 5). For the regulated rivers, the impacts of water-resource development would outweigh any small changes to flow regime brought about by climate change. Therefore, the impacts of climate change will be mostly restricted to increases in stream temperatures that alter the distributions of the cold water-tolerant species, especially two-spined blackfish (Table 2). It is likely that temperature increases will reduce river blackfish distributions at lower altitudes and affect the spawning time and food availability for Macquarie perch. As with the southern uplands, we predict that significant impacts on exotic trout will greatly benefit native fish species such as mountain galaxias and the two blackfish species through reduced competition and predation.

The dryland region in the upper western corner of the MDB (Fig. 5) consists largely of the unregulated semiarid rivers. While there are no major diversion weirs or water storages on these rivers, there are many low-level weirs that currently impose the largest threat to fish migration and breeding patterns in the region (Table 2). The fish of these rivers are mostly ecological generalists, adapted to cope with the highly variable flows of these rivers (Gehrke et al. 1995; Balcombe et al. 2006, 2011). It is unlikely that climate change will substantially change these fish assemblages, given the ability of the species to cope with extremes of climate. However, where the impact of climate change interacts with the effects of barriers on fish migration (such as low flows), the fish species that migrate as part of their breeding cycle (e.g. Hyrtl's tandan, Balcombe and Arthington 2009; Kerezsy et al. 2011) could be affected in some rivers. There are likely to be some localised impacts of less flow reaching terminal wetlands (climate change added to water-resource development), affecting floodplain and wetland vegetation, especially encroachment from lignum. This could enhance the local abundance of fish species, including exotic species such as carp and gambusia, because of the presence of added structural habitat and increased organic input when these systems are flooded.

Originating in the northern and eastern upland region, the Darling tributaries region refers only to the lowland section of these rivers (Fig. 5). All are affected by water-resource development to varying degrees because of the presence of headwater storages, low-level weirs and significant levels of water abstraction for irrigation (Table 2). Climate change is likely to lead to fewer refuge pools (remaining pools will also be shallower) during dry periods, which will have a flow-on effect for fish that are close to their upper thermal limits, such as Murray cod. Generalist species such as eel-tailed catfish, yellowbelly and bony bream may increase in abundance in response to declines in Murray cod. Balanced against this is the potential expansion of the distributions of exotic fish, particularly of common carp and gambusia, which are established and significant competitors for food resources in these rivers (Gehrke et al. 1995; Balcombe et al. 2011). The combined effects of reduced flows and reduced connectivity from development will also have an impact on species, including carp gudgeons, Murray–Darling rainbowfish and un-specked hardyheads, that use more marginal waterbodies such as anabranches and billabongs (Lintermans 2007).

The Darling lowland region in the mid-western MDB (Fig. 5) has been significantly affected by water-resource development, with major abstractions and many low-level weirs (Table 2). Because of the high levels of water-resource development and the associated decline of the fish assemblages in the region, climate-change impacts are not likely to cause too many further changes. However, further reductions in flow and increased temperatures may decrease available drought refugia for Murray cod. Furthermore, bony bream may increase its abundance in such refuge habitats in response to fewer predators and because of its ability to capitalise on increased algal resources associated with high water temperatures. These increases in primary productivity in degraded habitats could also favour common carp (Koehn 2004), placing further stress on native fish. Lignum may encroach into drier sections of rivers such as wetlands, which could exacerbate impacts of exotic fish.

The Murray region in the lower MDB (Fig. 5) is a highly regulated riverine system with many weirs and locks. Flows are managed via release from major water storages. Fish stocks have been largely affected by cold-water pollution for significant lengths of river below water storages. As a result, increased temperatures are likely to ameliorate some of these impacts and thus increase the abundance and range of species such as trout cod, Murray cod, yellowbelly and silver perch (Table 2). Given the lack of connectivity of this highly modified and regulated river with off-channel and floodplain habitats, any further reductions to flow will mean fewer floodplain connections and further degradation of fish such as yellowbelly, silver perch, Australian smelt, flat headed gudgeons, un-specked hardyheads and pygmy perch that use floodplains and associated waterbodies (Table 1). Less floodplain inundation will also have an impact on riparian vegetation, especially river red gum. This will have longer-term effects on the habitat use and food resources of native fish. However, fewer floodplain connections may have an impact on common carp recruitment because inundated floodplains provide a significant source of carp recruits in the MDB (Crook and Gillanders 2006).

The southern rivers of the MDB that feed into the mid- and lower Murray (Fig. 5) are mostly regulated with storages on all but the Ovens River (Table 2). The likely reduced flows and reductions in flood frequency will have an impact on in-channel and floodplain habitats. Although weirs are likely to maintain refuge pools in regulated river reaches, overall flow reductions may exacerbate the loss of longitudinal and lateral connectivity. With less flow, there is also likely to be less macrophyte habitat, leading to fewer small-bodied species such as pygmy perch, carp gudgeons and flat headed gudgeons as they become more vulnerable to predation (Table 2). In headwater streams, increased temperatures will drive an upstream contraction of cold water-tolerant species such as river and two-spined blackfish and trout cod. Again, barrier effects may impede such movements in more developed catchments.

The lower Murray region represents the termination of the Murray–Darling system (Fig. 5) and is highly regulated via locks and weirs (Table 2). Consequently, the assemblages of riverine fish are already highly degraded and subtle changes to either temperature or flow will probably not result in much impact (Table 2). Nonetheless, increased water temperatures combined with less flow, and thus less lateral connectivity, are likely to further affect the already degraded habitats for wetland fish species such as pygmy perch, flat headed gudgeons, un-specked hardyheads and Murray–Darling rainbowfish because of loss of macrophytes and increased salinity (Table 1). Given the high temperature tolerance of gambusia, it is likely to flourish in these wetland habitats and further affect small-bodied native fish species. Further declines in river red gum woodlands will also have flow-on effects to the riverine fish assemblages because of reduced structural woody habitat, carbon inputs and food resources. Given the raft of impacts already known for this region, the future looks bleak for several species that are restricted in distribution and abundance, including river blackfish, purple-spotted gudgeon (Mogurnda adspersa) and Murray hardyhead (Craterocephalus fluviatilis).

Most of the regions of the MDB contain highly altered river systems. The Lower Murray region is characterised by weirs and locks that constrain longitudinal connectivity throughout the river and lateral connectivity with associated floodplains. In the northern uplands, reaches show significant effects of thermal pollution and flow regulation from water storages, again leading to riverine systems that have been largely altered. As such, the fish assemblages of the MDB are highly degraded and unlikely to improve significantly under current management regimes. Climate change will further stress these already ‘degraded’ fish assemblages. However, in some regions the change will be limited given the already degraded environment, whereas in others there could be an improvement for native fish, given that temperature increase is likely to at least partially ameliorate some of the effects of thermal pollution and increase stress on resident exotic taxa.

The effects of climate change on fish assemblages should not be considered in isolation from existing water-resource development in any river catchment, especially in heavily altered systems such as the MDB, where historical levels of water-resource use exceed the likely impacts of climate change. In this respect, future management plans that aim to improve the condition of rivers in the MDB must consider both the added pressure of climate change as well as the current stress associated with agricultural and human water use.