The Living Murray
Foundation Report on the significant ecological assets targeted in the First Step Decision

Your location »  5 Information base for the Chowilla Floodplain and Lindsay–Wallpolla Islands system » Contents » Foundation Report Home » MDBC Home

Printer friendly PDF (1.4Mb)
Download pdf in sections:
Part A (496Kb)
Part B (574Kb)
Part C (400Kb)
Part D (238Kb)

5 Information base for the Chowilla Floodplain and Lindsay–Wallpolla Islands system

5.1 Introduction

The Chowilla Floodplain and Lindsay-Wallpolla Islands significant ecological asset (SEA) comprises three separate locations: Lindsay Island in Victoria; Wallpolla Island in Victoria; and the Chowilla Floodplain, which spans South Australia and New South Wales (DEH, 2003) and the NSW section of the Chowilla Floodplain (Figure 1.1).

The Riverland Ramsar Wetland incorporates the Chowilla and Ral Ral Anabranch on the northern side of the River and the Murtho floodplain on the southern side-all in South Australia. The Chowilla area comprises the most ecologically significant section of the Riverland Ramsar Wetland (DEH, 2003) and is the focus of maintenance and restoration efforts at this site under the First Step Decision. The Ral Ral and Murtho areas are important, but this Foundation Report concentrates on the Chowilla Floodplain area of the Riverland Ramsar Wetland, which has a larger available information base.

Although Chowilla Floodplain, Lindsay Island and Wallpolla Island sites each have their own unique characteristics, at the scale of the First Step Decision they are grouped into one ecological asset (see Figure 5.1). To help structure the information, the discussion of this asset is in two parts: Part A Chowilla Floodplain (located downstream of Lindsay Island, with the bulk of its area in South Australia) (Figure 5.2) which has some characteristics that are distinctly different from the Lindsay-Wallpolla system. Part B dealing with the Lindsay Island and Wallpolla Island systems as a single group (even though they are separated by around 40 river kilometres).

 

The First Step Decision Interim Ecological Objective for the Chowilla Floodplain and Lindsay-Wallpolla Islands is to maintain high biodiversity values of the Chowilla Floodplain (Table 1.2). The expected outcomes are: high-value wetlands maintained; current area of River red gum maintained; and, at least 20% of the original area of Black box vegetation maintained. (Table 1.2). Various management opportunities to achieve these objectives will be explored later in this chapter. The following sections describe the characteristics of the SEA, exploring the links between the biophysical condition of the SEA and hydrological and other factors.

Lindsay-Wallpolla Islands

The Lindsay-Wallpolla Islands are particularly important sites given that they support a number of Murray cod and other native fish nurseries, a diversity of landforms (including wetlands, billabongs and a system of anabranches) and a range of species, particularly fish and birds. Flood flows from the River Murray are fundamentally important to the environmental condition of the Lindsay-Wallpolla system. The health of the system has been threatened by river regulation, mainly with respect to reduced frequency of medium-sized floods (the seasonal pattern of river flow is largely unchanged).

Lindsay and Wallpolla Islands are formed on the southern side of the River Murray (Victoria) by a series of anabranches that leave and then rejoin the river, leaving the islands situated between the anabranch channels and the main stem. Wallpolla Island covers an area of 9,200 ha and Lindsay Island has an area of 15,000 ha. The main anabranch forming Wallpolla Island is Wallpolla Creek (Figure 5.3) and the main anabranch forming Lindsay Island is Lindsay River (Figure 5.4). Lake Wallawalla is a shallow, permanent riverine lake located off the lower Lindsay River and is part of Lindsay Island. Potterwalkagee Creek is another anabranching channel that is located between Lindsay and Wallpolla Islands, and it forms Mulcra Island (2,156 ha). Thus, Lindsay Island and Wallpolla Island are linked by Mulcra Island, and although it has ecological values, Mulcra Island is not part of the Significant Ecological Asset.

The islands are located in the far northwest of Victoria just downstream of Mildura-Wentworth. Wallpolla Island is located downstream of Wentworth Weir (Lock 10) and upstream of Lock 9 (Figure 5.2). Lindsay Island is located just downstream of Lake Victoria and between Lock 6 and Lock 7 (approximately 700 km from the Mouth) (Figure 5.2). Frenchmans Creek flows from the River Murray to Lake Victoria, diverting some water from Wallpolla Island, while the Rufus River runs from Lake Victoria into the Lindsay Island system. One of the largest channels within this system is Mullaroo Creek, which diverges from the River Murray just upstream of Rufus River, crosses the island from east to west, and then joins the Lindsay River (SKM & Roberts, 2003).

Chowilla Floodplain

Located near the NSW, Victorian and South Australian border, the Chowilla Floodplain is on the northern side of the River Murray (Figure 5.2). Most of the floodplain lies in South Australia, with a section extending into New South Wales. It covers 17,700 ha, forming the largest floodplain complex in the lower River Murray (MDBC, 2002).

The Chowilla Floodplain is located in a semi-arid environment, with mean annual rainfall approximately 260 mm/year, and average evaporation 1,960 mm/year. These wetlands and the floodplain are dependent on the River Murray and a system of more than 100 km of anabranch creeks for inundation. Figure 5.5 illustrates just one example (Lake Limbra, which is located on the north of the floodplain) of the mosaic of different vegetation communities and hydrological environments that make up the Chowilla Floodplain.

The site is downstream of the main water management headworks and large diversions of the River Murray, which means typical flows each month are considerably less than under natural conditions. Although the Chowilla Floodplain retains much of its natural character and attributes (Sharley & Huggan, 1995, p. 9; MDBC, 2002), there are large areas, and an increasing proportion, of vegetation that is in poor condition (Overton & Jolly, 2003). Declining health means reduced capacity of the area to sustain plant and animal communities and to contribute to the ecology of the broader river system. The current extended period of low flow is causing considerable deterioration in ecosystem health.

Vegetation condition on Chowilla Floodplain is related to both flow regime and groundwater level and salinity. Other factors such as grazing by sheep, feral and over-abundant native animals also influence vegetation condition.

There is limited capacity to achieve significant environmental improvement on the Chowilla Floodplain using the existing water management infrastructure (principally Lock 6). This is because much of the significant ecosystem areas are elevated well above the level of the weir pool. The groundwater and salinity factors (see Overton & Jolly (2003) for a review) distinguish Chowilla Floodplain (in the Sunraysia zone) from the significant ecological assets further upstream such as Barmah-Millewa Forest and Gunbower, Koondrook-Perricoota forests. These upstream assets still utilise groundwater, but they are generally underlain by fresh groundwater (Jolly et al., 1994).

 

Part A-Chowilla Floodplain

5.2 Value and condition of Chowilla Floodplain

5.2.1 Conservation significance of Chowilla Floodplain

The Chowilla Floodplain and associated anabranch system (Figure 1.1 and Figure 5.6) has long been recognised as having high conservation values. The Chowilla Floodplain in South Australia is part of the Riverland Wetland Complex, which was listed under the Ramsar Convention in 1987 (ANCA, 1996, pp. 494-6). It is also listed on the national and state directories of important wetlands. The Chowilla Floodplain is part of the Bookmark Biosphere Reserve (Department of the Environment and Heritage, 2004b), and is part of the network of international Biosphere Reserves coordinated by the UNESCO Man and the Biosphere Program (Crabb, 1997, p. 64).

 

Chowilla Floodplain contains the largest remaining area of natural riverine Red gum forest in the lower River Murray. It contains the full range of the region's riverine vegetation communities within the Murray Scroll Belt sub-region, and 11 of the 12 vegetation communities within the Black box floristic zone, found in the Riverina Biogeographical Region (Margules et al., 1990). The dominant native vegetation in Chowilla Floodplain includes River red gum, Black box, River cooba and Lignum (Figure 5.8). Chenopods are common. Grassland communities occur in Chowilla Floodplain, including the perennial Rats tail couch, a river bank and river edge species (Young et al., 2001, p. 198). The survey of O'Malley and Sheldon (1990, p. 33) rediscovered Wavy marshwort, an aquatic plant that was considered extinct in South Australia at the time.

A threatened flora assessment was recently conducted (Robertson, 2003) for the Riverland Ramsar Wetland site and also including a 5 km buffer area on the northern, eastern and southern boundaries plus related habitat area downstream to Overland corner. A total of 154 plant species were recorded, with ten having the status Vulnerable and 27 having the status Rare under the South Australian rating system. Under the Murray Flora Region rating system, two had the status extinct, nine were vulnerable, five were threatened, 47 were rare, 75 were uncertain and five were not assessed but of possible significance. Robertson (2003) also data compiled information on vegetation communities and plant associations, largely focused on sourced from the earlier survey of O'Malley and Sheldon (1990). South Australian threatened plant associations at Chowilla are Closed Herbland in billabongs found at Nil Nil and Queen Bend (note: Nil Nil is on private land), Low Shrubland represented by degraded examples on low floodplain areas, ephemeral community common found on low floodplain and Low Shrubland found in Coombool Swamp. Robertson (2003) provided grid reference locations for all threatened plants in the Riverland Ramsar Wetland, so this information can be used in GIS applications.

The River red gum Woodlands of the Chowilla Floodplain area support one of the highest diversities of birds of any terrestrial vegetation association in South Australia (Carpenter, 1990; Sharley & Huggan, 1995). The Black box Woodlands provide habitat for many Mallee birds and floodplain and riverine species (Carpenter, 1990; Sharley & Huggan, 1995). The permanent wetlands are important drought refuges for waterbirds and are used for breeding by some species; the ephemeral wetlands are important for waterbird breeding (Carpenter, 1990; Sharley & Huggan, 1995, p. 83).

Chowilla Floodplain has a high diversity of terrestrial and aquatic habitats, supports populations of rare or endangered species, has fish breeding habitat, supports populations of breeding waterbirds, has habitats not well represented elsewhere, and has a relatively low level of direct human disturbance (it has been indirectly impacted by regulation and salinity effects).

The first attempt at formal identification of sites of high significance was by O'Malley and Sheldon (1990, pp. 200-5). Four of the five high conservation sites they identified were close to or within anabranch channels, or close to the Lock 6 Weir pool. The permanently high water levels, which are the result of regulation, have maintained biologically favourable hydrological conditions for some species by providing water and flushing salts from the soil (or had a neutral effect on the local hydrology), and O'Malley and Sheldon (1990) recommended that these conditions be maintained. In contrast, the group `Coombool Swamp, Lake Limbra, Gum Flat, Lake Littra and surrounds' (Figure 5.7) comprises elevated sites distant from the river and anabranch, and O'Malley and Sheldon (1990) were concerned that the current flooding regime may be inadequate. Coombool Swamp (312 ha) is a large deflation basin ringed by Black box Woodland and is filled from Monoman Creek (Roberts et al., 2003, p. 4).

 

A pilot study to develop a framework for identification and classification of sites of high conservation significance is being conducted by the Department for Environment and Heritage (DEH), South Australia. This framework uses a GIS model, and is based on an approach outlined in James and Saunders (2001). The biological data are limited to spatially mapped information such as plant communities, threatened species and vegetation health. Using a GIS, the concept involved was to add together, within a grid structure, the biological assets of ten separate layers weighted on their priority value, to provide a final conservation rating for any one grid cell. The conservation ratings were expressed as a percentage of the maximum value attainable. Based on this information, decisions could be made (in conjunction with local knowledge and available resources etc), regarding those areas that would be higher priority for conservation and management. The distribution of three conservation value classes is described in Smith and Dominelli (2005).

 

5.2.2 Geomorphology and hydrology of Chowilla Floodplain

Inundation of the Chowilla Floodplain occurs when the river flow to South Australia exceeds 33,100 ML/day at the South Australian border. The relationship between flow in the River Murray and extent of inundation was initially developed by Sharley and Huggan (1995) (Figure 5.9), but this could now be revised using the DEH GIS-based vegetation model and digital elevation model, or using the River Murray Flood Inundation Model [for a description see RMCWMB (2004)].

Lake Victoria (Figure 5.1) is used principally to provide a reliable water supply for the Lower Murray region in South Australia and to mitigate and augment flood peaks as required (MDBC, 2002). Water enters Lake Victoria upstream of Lock 9, often with highly turbid floodwaters from the Darling River, and water is discharged via the Rufus River downstream of Lock 7. The entitlement flow to South Australia varies between 3,000 ML/day and 7,000 ML/day, depending upon the time of the year (MDBC, 2002). This amount cannot be delivered solely from the upper storages of the River Murray due to the channel capacity constraints of the Barmah Choke (MDBC, 2002), so the water stored in Lake Victoria and/or Menindee Lakes is used to meet the shortfall. During the El Niño episode of the 1980s, the demand for irrigation water was such that flows to the lower Murray were comprised mainly of Darling River water stored and released from Lake Victoria (Walker et al., 1992). In wetter seasons during the 1990s, Lake Victoria was filled with lower turbidity water originating from the Murray-Murrumbidgee catchment (Jensen, 1998; Thoms et al., 2000, pp. 68-9).

 

Depending on circumstances, river operators have the option of utilising the storage capacity of Lake Victoria to mitigate floods to South Australia. In 2000, Lake Victoria was, for the first time, used to enhance a flood peak (DWLBC, 2002; Gippel, 2003), in combination with raising of the weir pool level of Lock 5, which is the lock immediately downstream of Chowilla Floodplain. Thus, the operation of Lake Victoria and Menindee Lakes offers potential to boost naturally occurring flood peaks, and the area of flooding, or extend the duration of naturally occurring floods to Chowilla Floodplain. The turbidity and salinity of the water delivered from these storages will vary depending on its original source and how long it has been in storage.

The construction of Lock 6 in 1930 on the River Murray towards the downstream end of the Chowilla Floodplain (Figure 5.22) resulted in permanently higher water levels on the adjacent floodplain area, higher groundwater levels, and continuous flows of water through the Chowilla Floodplain anabranch system (Sharley & Huggan, 1995).

The soils of the Chowilla Floodplain were described in detail by Hollingsworth et al. (1990). The soils consist generally of a layer of alluvial grey cracking clay, known as Coonambidgal Clay, up to 5 m deep, overlying an unconsolidated alluvial sand deposit, known as Monoman Sand, approximately 30 m deep (Dawes et al., 1998). The boundary between these layers is often unclear, with transitional material of varying clay content up to 1 m thick (Dawes et al., 1998). Since the installation of Lock 6, groundwater levels have risen from the Monoman Sand formation to between 2 and 4 m from the surface in the Coonambigdal Clay (Dawes et al., 1998). In some sections of Chowilla Floodplain the groundwater is now about 2-3 m higher than it was under natural conditions, so that it is well within the tree root zone (Roberts et al., 2000, p. 42; Sharley and Huggan, 1995, pp. 9, 22, 109; Walker et al., 1996). The mechanism is illustrated by Figure 5.10. The area is one of natural discharge of saline groundwater, but the higher groundwater levels caused by Lock 6 have exacerbated the salinity problems (Sharley & Huggan, 1995, p. 9). Also, flushing of salt from the floodplain soils occurs less frequently as a consequence of the reduction of flood frequency.

 

5.2.3 Vegetation classifications used at Chowilla Floodplain

The Chowilla Floodplain supports a diverse range of vegetative communities. For MFAT modelling of Zone E, Roberts et al. (2003) included Werta Wert and Coombool Swamp in the floodplain sites. Werta Wert is fringed by Red gums with a Spike rush understorey and Black box and Lignum Woodland dominate the adjacent floodplain. Coombool Swamp is ringed by Black Box woodland. The modelling included four following floodplain vegetation classes: River red gum Woodland, Black box Woodland, Lignum Shrubland and Rats tail couch Grassland. Wetlands were represented by Ribbon Weed. These classes of vegetation are consistent with those used in other literature concerning Chowilla Floodplain. However, there are other significant vegetation communities at Chowilla Floodplain, including Mallee eucalypts, River red gum Forest, River cooba (also known as Eumong, Acacia stenophylla) tall Shrubland, Murray (or Cypress) Pine (Callitris preissie), Cane Grass (Eragrostis australasica), Chenopod and Samphire (Sharley & Huggan, 1995, p. 81). Lignum occurs in extensive pure stands, or in association with River cooba. River cooba can occur in pure stands, but is usually found as an understorey component to River red gum and Black box.

O'Malley (1990) surveyed the Chowilla Floodplain vegetation communities using quadrat analysis to indicate patterns of plant community composition across a gradient of physical parameters. Analysis of the data produced six floristic groupings (based on similarity of floristic composition). These groupings were: Floodplain Black box (perhaps also including River red gum, Lignum and River cooba); Blackbush/Hopbush sand based communities; Lake-Bed Herbfield; River red gum Forest; Weedy Lagoon; and, Aquatic Herbfield. Margules and Partners et al. (1990) mapped the vegetation communities along the River Murray, and this classification scheme has since been widely used. The classifications used by Sharley and Huggan (1995) for Chowilla Floodplain are based on this scheme. Overton (CSIRO, Land and Water) also produced a vegetation community map for Chowilla Floodplain (cited in Overton & Jolly, 2003). More recently, Overton and Jolly (2003) developed a new map for Chowilla Floodplain (Figure 5.11) based on the mapping of Margules et al., (1990), Overton Mapping, aerial photography, satellite data and field observations.

5.2.4 Floodplain vegetation distribution and hydrological regime

The original scientific work on hydrology-vegetation relationships at Chowilla by Sharley and Huggan (1995), has since been built upon by other researchers. Of particular note is the work undertaken in partnership between the DEH (SA) and CSIRO Division of Land and Water. The work on flow and groundwater was recently compiled, and modelling results provided, in Overton and Jolly (2003).

The vegetation communities are distributed across the floodplain according to hydrological, soil and salinity gradients, so that floods of various magnitudes impact on these communities to differing extents (Figure 5.12). Sharley and Huggan (1995) tabulated the flood frequencies of the major floodplain vegetation communities of Chowilla Floodplain before regulation (Table 5.1). There is a range of flood magnitudes associated with each community (Figure 5.12). The range of flood magnitudes required to inundate River red gum and Black box are plotted on Figure 5.13. As an example, the discharge required to inundate 50% of the area of River red gum is 68,000 ML/d, and 85,000 ML/d for Black box woodland.

Sharley and Huggan (1995) also listed the approximate river flow conditions associated with the commencement of filling of some sites of high conservation significance (Table 5.2). These `commence to fill' values were based on field observations, supported by interpretation of satellite images of actual floods. Roberts et al. (2003, p. 29) cited somewhat different `commence to fill' values for Werta Wert and Coombool Swamp than those given by Sharley and Huggan (1995) (Table 5.2). The more recent REG E values (Roberts et al., 2003) were based on manual inspection of the Floodplain Inundation Model used in South Australia. For Werta Wert the `commence to flow' values of Roberts et al. (2003, p. 29) were River red gum Woodland and Open plains Swamp at 45,000 ML/d, Lignum at 50,000 ML/d, and Black box Woodland at 60,000 ML/d.

 

 

The flooding requirement of River red gum has been studied in detail at Barmah Forest in Victoria (Dexter, 1978), and based on this and other information, CPBR and ANH (2004) provide a comprehensive profile of the attributes of this species. Shallow and short duration (one month) floods are less effective than long-duration floods, because they only water trees close to the channel and soil moisture rapidly returns to pre-flood levels (Young, 2001, p. 191). The critical time in regeneration is seedling establishment, rather than germination (which does not strictly depend on flooding). The optimum time for floods to recede is spring/early summer (Dexter, 1978). Seedlings cope with summer heat stress by accessing soil water (Jolly et al., 1994; Roberts & Marston, 2000).

On the Chowilla Floodplain, the climate is considerably drier than in the Barmah-Millewa Forest area, soils have lower organic matter, and floods are less frequent. Red gum is restricted to a narrow riparian fringe lining the main channel and it has an open woodland form, or forms dense patches on low-lying spots (Roberts & Marston, 2000). The total area of the Floodplain covered with River red gum is 752 ha (4% of total area) (Sharley & Huggan, 1995). Here, River red gums use water according to their location, with floodplain trees using mostly groundwater, and trees adjacent to waterways sourcing 30-50% of water from surface channels (Thorburn & Walker, 1994).

The natural flooding frequency of River red gum communities at Chowilla Floodplain was once per year to once every two years (Table 5.1). The area of River red gum flooded increases non-linearly with increasing flood magnitude (Figure 5.14). Most of the River red gum Forest and Woodland communities occur in areas of the floodplain that are inundated by River Murray floods of less than around 83,000 ML/d (Figure 5.31). Note that the area inundated of Red gum Forest and Red gum Woodland is similar for River Murray floods of 83,000 ML/d, 94,000 ML/d and 102,000 ML/day, suggesting that most of it is inundated at a flow of between 71,000 ML/d and 83,000 ML/d.

River red gums are tolerant of saline conditions to some extent, with an upper limit generally 20 dS/cm (MDBC, 2003), but the upper tolerance in the Chowilla floodplain area is thought to be 30 dS/m (Overton & Jolly, 2003). Groundwater salinity is potentially a limiting factor for River red gum. The groundwater table lies 1 to 5 m deep at 35 to 85 dS/m across the majority of the floodplain, but it is less saline, at 5-18 dS/m, close to the river where most Red gum is located (Noyce & Nicholson, 1993; Overton and Jolly, 2003, p. 14). Soil type and hydraulic properties also strongly condition vegetation distribution (Overton & Jolly, 2003).

 

The flood discharges shown on the graph were selected from those with known inundation areas, and separated by around 10,000 ML/d. This relationship could be revised using the new DEH GIS-based vegetation model and digital elevation model.

Like River red gum, Black box can grow in a range of environmental conditions. In Australia, Black box is less widespread than River red gum, with the semi-arid inland floodplains of the lower River Murray being important sites. The soils of the Black box floodplains are mostly heavy clays, in contrast with the more varied and more sandy soils of the River red gum zones. Soils of Black box zones are also more saline than those of River red gum zones (Smith & Smith, 1990; Overton & Jolly, 2003). Three plant families dominate the understorey: grasses, daisies; and chenopods. The chenopods are characteristically salt tolerant and dominate in the higher elevation areas (Smith & Smith, 1990). Lignum shrub is often found associated with Black box communities, either as understorey on lower elevation areas, or as dense thickets adjacent to Black box Woodlands (Smith & Smith, 1990). Black box Woodlands are low and very open on the outer floodplain.

Black box Woodland covers an estimated 3,384 ha of the Chowilla Floodplain (44% of total) (Sharley & Huggan, 1995). Here, Black box trees vary in height from 3 to 21 m depending on the flood frequency they experience. Compared to River red gum, Black box is less tolerant of floods, and more tolerant of droughts, which reflects its position higher on the floodplain (Young, 2001, p. 192). Studies at Chowilla Floodplain show that Black box is opportunistic with finding water, sourcing it from the surface soils, as well as shallow (to 3 m) and moderately saline groundwater (Overton & Jolly, 2003). Black box trees on the Chowilla Floodplain are able to tolerate an upper limit of groundwater salinity of 55 to 58 dS/m (Overton & Jolly, 2003; DWLBC, no date). A hybrid between Black box and Eucalptus graculis, a mallee species that grows in areas adjacent to the floodplain (known as the `Green box' variant), seems to be able to withstand highly saline areas (Harper, 1997; DWLBC, no date).

Although Black box is salt tolerant, vigour decreases even at moderate salinities, especially where the water table is high, leaving only a shallow depth of aerated soil in which to grow (DWLBC, no date). The area of Black box flooded increases non-linearly with increasing flood magnitude (Figure 5.14). The natural flooding frequency of Black box at Chowilla Floodplain was once every two to ten years (Table 5.1) and duration of two to five months. The natural flooding frequency of Lignum was once every two to eight years (Table 5.1).

5.2.5 Health of floodplain vegetation

The ecological health of Chowilla Floodplain has long been affected by many factors, including livestock grazing, timber cutting to fuel river boats, and construction of Lock 6 (Crabb, 1997, p. 64). River regulation has reduced flood frequency and the area is also subject to extended drought. The health of the vegetation communities on the Chowilla Floodplain has been declining over recent years. Some key findings from past studies are:

• Tree dieback events were reported anecdotally in the periods 1965-70, and

1985-89. These two periods were characterised by low rainfall and lack of floods exceeding 1,500 GL/month (equivalent to a daily average of 50,000 ML/day) at the South Australian border. These are the same conditions that have prevailed from 1996 to 2003, when tree death has again been recorded (MDBC, 2003, p. 22) (Figure 5.15).

• Vegetation surveys conducted on Chowilla Floodplain during 1988-89 found a high incidence of dieback among River red gum and Black box woodland associations (O'Malley & Sheldon, 1990, p. 33). There was a suggestion that the areas more distant from channels were the worst affected.

• The regeneration rates of the key riverine vegetation species are low, and the trees lack vigour on the floodplain below Wentworth (Margules & Partners et al., 1990; Smith & Smith, 1990; MDBC, 2003).

• Surveys conducted in February 2003 indicated that approximately 80% of River red gum trees on the River Murray floodplain in South Australia were stressed to some degree. In the area between Wentworth and Renmark (which includes Chowilla Floodplain) more than half of all trees were stressed or dead (MDBC, 2003, p. 16).

• Approximately 45-55% by area of the tree communities are currently in poor health (Overton & Jolly, 2003) (Figure 5.16).

 

The cause of the tree stress at Chowilla Floodplain are a combination of three main factors: prolonged drought, reduced flooding, and soil salinisation (O'Malley & Sheldon, 1990, p. 33; MDBC, 2003, pp. 20-2; Overton & Jolly, 2003) (Figure 5.15). Salinisation of floodplain soils was blamed for the dieback of Black box observed in the 1980s and 1990s (Jolly et al., 1993; Slavich et al., 1999a). The massive River red gum dieback event currently occurring (MDBC, 2003) is a case of the reduced flooding being compounded by an increase in soil salinisation. This has reduced the amount of fresh water available to River red gums, which they typically source from having roots into permanent and ephemeral wetlands and creeks.

Regulation has reduced the frequency of suitable floods for watering Black box communities from once every two to ten years to once every five to twenty years (Sharley & Huggan, 1995). Dieback of Black box has occurred in highly saline areas, while elsewhere the tree communities are in poor condition (Overton & Jolly, 2003, p. 8). There is also a lack of regeneration as a result of a reduced flood events. Based on known flooding frequencies and groundwater depth, Sharley and Huggan (1995) considered that 70% of Black box Woodland at Chowilla Floodplain is at risk of degradation in the long-term. Taylor (1993) estimated that 40% of Black box was in poor condition, while this estimate was recently revised to 45% to 55% (Overton & Jolly, 2003).

5.2.6 Birds

Based on field surveys of terrestrial and wetland habitats, Carpenter (1990) described the distribution, abundance and use of habitats by birds of the Chowilla Floodplain. A total of 134 bird species were recorded, including 37 waterbirds. For the waterbirds, eight species were recorded breeding. Based on data from other surveys, a total of 170 bird species have been recorded, and at least 33 others are known to occur in similar habitats nearby. Twenty of these species are considered to be rare, vulnerable or endangered in South Australia. The Barking owl, Azure kingfisher, Western calamanthus and Spotted bowerbird are either presumed extinct or have virtually disappeared from the area (Carpenter, 1990). Carpenter (1990) was of the opinion that Chowilla Floodplain has outstanding importance for bird fauna in South Australia.

Black box Woodland supported the highest diversity of bird species, mainly ground foragers and hollow nesting species. Stock grazing, soil salinisation and introduced predators would, therefore, have the greatest potential to affects this avifauna (Carpenter, 1990). River red gum Woodland provides habitat for canopy feeders, and therefore a decline in health or coverage of red gum would impact on some species of conservation significance: (e.g., Square-tailed kite, Blue-faced honeyeater, Little friarbird, Regent parrot and Striped honeyeater). The Regent parrot (Eastern) is listed as vulnerable nationally and in South Australia. It breeds in hollows in River red gum trees close to or over water. Chowilla Floodplain supports a breeding population of over 100 pairs that have been recorded nesting at ten locations adjacent to the main river channel.

The nine wetlands surveyed by Carpenter (1990), which were all at least partially inundated at the time of survey, supported over 5,000 waterbirds, with the majority (75%) being Australian Grey Teal. Carpenter (1990) was of the opinion that reduced flooding has caused a decline in waterbird numbers (especially Musk duck, Dusky moorhen, Eurasian coot and Buff-banded rail). Permanent lagoons with reedbeds and the shallow margins of anabranches provide refuge for waterbirds during drought periods.

5.2.7 Mammals

Brandle and Bird (1990) provided baseline information on the distribution, abundance and conservation significance of mammal species present on the Chowilla Floodplain. Seventeen native and eight introduced species were recorded during the survey, and include a high diversity of species. Several areas were identified as especially important in terms of mammal habitats. The most important were those areas low floodplain subject to more frequent inundation, particularly those areas with cracking clays. It was considered important for these areas to have fringing vegetation such as woodland or lignum to provide suitable cover for mammals during flood periods when they must vacate the cracks.

A comparison of the list of mammal species collected in 1988 with those collected by an expedition in the nineteenth century shows at least 15 extinctions (Brandle & Bird, 1990).

Brandle and Bird (1990) highlighted four feral species as serious, but potentially manageable, threats to mammals. These were rabbits, feral goats, feral pigs and foxes.

5.2.8 Reptiles and amphibians

Bird and Armstrong (1990) surveyed reptile and amphibian species on Chowilla Floodplain, with particular emphasis on riverine and floodplain habitats with populations that are at risk from the effects of flow regulation. Seven species of frog were recorded. This included the Southern bell frog, which although is nationally listed as vulnerable, is regionally common with populations recorded at most of the large wetlands throughout the site. Of the 28 reptile species known from Chowilla Floodplain, 11 are classed as dependent either on the wetlands or on the fringing riverine woodlands and clayey floodplains. Of these, five species (three species of tortoises, a species of skink and the Tiger snake) are restricted to wetlands.

Bird and Armstrong (1990) reported observations by local herpetologists that frog numbers declined from 1960. Several agents of change have been suggested, including river regulation, increased salinity and turbidity, climatic change and an increase in environmental pollutants (Bird & Armstrong, 1990). Accompanying the decrease in frog abundance has been a dramatic decline in Tiger snake numbers.

5.2.9 Fish

In a survey of fish at Chowilla Floodplain, Lloyd (1990) observed that the fish community was similar to those at other locations along the lower River Murray. However, the extensive development of floodplain macrohabitats at Chowilla has allowed for more obvious habitat differentiation among species. The anabranches of Chowilla Floodplain provide the only significant spawning grounds and nursery habitat in South Australia for Golden perch, Murray cod and Silver perch (Lloyd, 1990).

Two fish species present at Chowilla Floodplain are listed as nationally vulnerable. These are: Murray hardyhead, with limited site surveys having located a single population adjacent to Lake Littra on the Chowilla Floodplain; and Murray cod which occurs in the main river channel, and which during floods may move into anabranch system and flooded lentic channels.

Although there are few historical data, a decline in the populations of many native fish, including Murray cod, Golden perch and River blackfish has been reported for the lower River Murray under regulated flow conditions (Walker, 1985; Walker & Thoms, 1993). Regional extinctions are well advanced for five native species in the lower Murray, and another two are threatened (Lloyd & Walker, 1986; Lloyd et al., 1989). The lower Murray fish community is severely depleted, with 29% of 55 native species absent and others present in very low numbers. Native fish represent only about 5% of the total fish biomass (Wittington et al., 2000), but this partly reflects the high numbers of Carp (Gehrke et al., 1995).

Flow regulation is implicated in declines in native fish in the lower Murray because floods are vital for reproduction in most species, while less variable flows favour alien species (Walker & Thoms, 1993). The main problem is reduced opportunities for recruitment because of the elimination of small floods, but sudden changes in water levels below weirs can strand fish eggs (Lloyd et al., 1989). Walker et al. (1992) considered the effects of flow regulation on fish populations in the lower Murray to be significant, because floods enhance spawning and recruitment in several species, and suitable floods are now less frequent and less prolonged. For example, no recruitment of Murray cod was recorded in the River Murray in South Australia for the period 1975-89 (Jensen, 1996). In the Murray system, flood peaks are strongly related to spawning in Golden perch and Silver perch. Reproduction in other fish, such as Rainbow fish and Hardyheads are greater in flood peaks, but the reason is not well understood. (Schiller & Harris 2001) have suggested that floods may benefit fish by dispersing eggs and larvae and by signalling the arrival of a highly productive period in the river when abundant food will be available.

Despite regulation, significant areas of the Chowilla Floodplain are still occasionally inundated. However, Thoms and Walker (1989) suggested that the rate of recession of flows in the lower Murray under the regulated flow regime might be too fast for the plants and animals to adapt and subsequently young fish risk being stranded if the flood recedes too quickly.

5.2.10 Macroinvertebrates

In a study of macroinvertebrates in seven different microhabitats at 13 sites representing six different macrohabitats on the River Murray and its floodplain at Chowilla, Lloyd and Boulton (1990) and Boulton and Lloyd (1991) found a relatively large number of taxa. Billabongs were thought to serve as refuges for many lentic taxa that rely on regular inundation to survive. In an experimental study using soil samples from different parts of the Chowilla Floodplain, Boulton and Lloyd (1992) found that a high diversity and biomass of invertebrates emerged from samples that were taken from annually flooded areas, while the sediment which flooded once every 22 years produced only protozoans. This result points to the risk of reducing the reserve of invertebrates through reduction of flood frequency associated with regulation.

The data of Bennison et al. (1989) showed that macroinvertebrate numbers downstream of Lake Victoria were lower by a factor of ten compared with upstream of the lake. In the 1990s, the lower turbidity during the summer/autumn period (due to a higher proportion of clearer Murray-Murrumbidgee water) resulted in a two- to three-fold increase in macroinvertebrate numbers, mostly crustaceans, especially the shrimp (Paratya australiensis) and prawns (Macrobranchium australiense) (Thoms et al., 2000, p. 68; Jensen, 1998, p. 224).

5.2.11 Aquatic macrophytes

Roberts and Ludwig (1990) surveyed the abundance and distribution of aquatic macrophyte communities at Chowilla Floodplain, covering main river channel, fast and slow flowing anabranches, billabongs and backwaters. Nearly all the plants recorded in the survey are commonly reported from the River Murray system, and have been reported in previous surveys from this area.

Low species richness along anabranches was attributed to high flows, and grazing by stock, particularly goats. Some of the banks on the faster flowing anabranches were actively eroding, making conditions unsuitable for establishment of plants. Banks along the slower flowing anabranches were mostly bare and affected by stock grazing (Roberts & Ludwig, 1990).

5.2.12 Permanent and Semi-permanent Wetlands, including anabranch creeks

O'Malley and Sheldon (1990) surveyed some wetland sites in Chowilla Floodplain. A tree dieback event was underway at Chowilla Floodplain at the time of the 1990 survey, and Werta Wert was one of the few areas of where healthy River red gum and Black box Open Woodland was recorded away from channel edge localities. Significant regeneration of these eucalypt species was also noted. Seven native plant species of conservation significance in South Australia were recorded. Billabongs are regionally rare, and the billabongs at Werta Wert are likely to support diverse communities when filled (O'Malley & Sheldon, 1990).

Queen Bend is a low floodplain site that includes a series of periodically inundated billabongs close to the River Murray. O'Malley and Sheldon (1990) found no evidence of eucalypt dieback at this site. Nine native plant species of conservation significance in South Australia were recorded, including two that were previously considered to be extinct regionally or state-wide.

Islands in anabranch creeks surveyed by O'Malley and Sheldon (1990) were found to support wetland areas of considerable biological significance. The anabranches have high aquatic diversity, ranging from permanent fast-flowing anabranches to slightly brackish, slow-flowing anabranches, backwaters and temporary billabongs. Twelve native plant species of conservation significance in South Australia were recorded. In the region between the River Murray and the inner creeks, Hypurna Creek and Slaney Creek, where surface water gradients induced by Lock 6 have resulted in freshening of groundwater, the vegetation has in the past been reported to be exceptionally healthy (O'Malley & Sheldon, 1990; Sharley & Huggan, 1995, p. 83).

Lakes Limbra and Littra, Coombool Swamp and Gum Flat (Figure 5.7) are temporary wetlands with a flooding frequency of once every two to three years (Lake Littra) to once every five years (Coombool Swamp and Lake Limbra) or less (Gum Flat). Except for Gum Flat, these areas were dry at the time of the survey by O'Malley and Sheldon (1990), but they were considered to be very important for macroinvertebrate populations and as fish breeding areas when flooded. The survey at Gum Flat revealed high populations of waterbirds, and several breeding pairs (Carpenter, 1990). Nine native plant species of conservation significance in South Australia were recorded in these temporary wetlands (O'Malley & Sheldon, 1990).

The health of wetlands at Chowilla Floodplain have been degraded by grazing and trampling by stock, high rabbit, goat and kangaroo populations, introduction and dispersal of exotic plant species, and river regulation (O'Malley & Sheldon, 1990). Temporary wetlands now flood less frequently. Prior to regulation, anabranch creeks were dry for the greater part of the year, but now they are permanently flowing. O'Malley and Sheldon (1990) pointed out that the deep, cool, fast-flowing habitat offered by these sites is extremely rare along the River Murray, and is also important for a number of species. This was sufficient grounds for O'Malley and Sheldon (1990) to recommend maintaining this current (regulated) flow regime.

5.2.13 Geomorphology of River Murray channel in the zone containing Chowilla Floodplain

Under the regulated regime, the banks in the lower Murray are subject to increased duration of near bankfull flow and more rapid changes in water level (Thoms & Walker, 1991). Thoms and Walker (1989) reported bank retreat of over two metres per year as a result of apparent block failure and bank scour. The channel cross-section at locks 3 and 4 and 8 to 10 has stabilised 30 to 40 years after regulation, while at locks 5 to 7 erosion has continued (Thoms & Walker, 1992). It was proposed that the main process of erosion was large-scale bank slumping accelerated by extremely rapid flood recessions. Walker et al. (1992) found that the banks immediately downstream of Lock 4 steepened in their cross-section slope from a range of 65º to 72º during 1988, to a range of 72º to 81º during 1989.

Smaller, more frequent changes in water level associated with routine weir operations may be undermining the toe of the banks, so that they are vulnerable to slumping following the more significant falls in water level (Walker et al., 1992). The steepening of river banks has acted to limit vegetation growth, degrading the littoral zone as a resource for fish (Walker et al., 1992). The channel is developing a stepped gradient associated with the weirs. Thoms and Walker (1992) explained this in terms of bed degradation downstream of weirs and deposition of material in the weir pools. The ecological significance of these geomorphological changes for Chowilla has not been investigated in detail, but such changes are known to alter instream physical habitat conditions, and erosion of levees on banks can alter the `commence to flow' threshold for wetlands.

5.2.14 River salinity

Approximately 40% of salt in the lower River Murray enters within South Australia, and a quarter of salt load entering the river in this zone is thought to be related to the presence of weirs (Walker, 1985). The saline groundwater is forced into the river by the hydraulic pressure of the artificially elevated water levels in the weir impoundments. The average groundwater inflow is about 5 ML/d resulting in a salt load to the River Murray of about 120 tonnes/day, or 45,000 tonnes per year (Chowilla Working Group and Chowilla Reference Group, 1992). The effect of this groundwater related salinity on the River Murray is reduced when possible by coordinating operation of releases of fresher water from Lake Victoria and the Menindee Lakes Storage on the lower Darling River (MDBC, 2002).

Since the construction of Lock 6, the surface soil layers of the Chowilla Floodplain have accumulated an additional 60,000 ML of saline water, containing about 2 million tonnes of salt (Sharley & Huggan, 1995, p. 29). This saline water, now stored at shallow depth, impacts on the root zone of vegetation, lake beds and stream beds. Floodwaters impose pressure on the shallow groundwater causing saline groundwater to drain to waterways in the recessions following floods (Sharley & Huggan, 1995, p. 29; Overton & Jolly, 2003, p. 61). Peak salt loads in Chowilla Creek as high as 1,800 tonnes/day can result (Overton & Jolly, 2003, p. 61). While large floods can produce large salinity differences when the flood recedes, Sharley and Huggan (1995, p. 99) noted that flows of less than 40,000 ML/d produce salinity difference of less than 50 EC units. Recently, Overton and Jolly (2003, p. 63) demonstrated a linear relationship between salt load and river flow for Chowilla for 18 observed floods, and found that the effect of elevated salt loads can last for up to two years.

5.2.15 Human use values

Traditional owners

Prior to European settlement the Maraura, Ngintait and Erawirung people occupied the Chowilla area, with a history of occupation dating back some 12,000 years, although it is uncertain whether occupation has been continuous over that period. A preliminary cultural heritage survey undertaken in 1991-92 indicated that, in contrast to the floodplain further downstream, the Chowilla Floodplain anabranch area is very rich in Indigenous cultural heritage, with sites which include burials, artefact scatters, middens, hearths, scarred trees and other isolated artefacts. The sand dunes and lunettes on the floodplain are particularly rich sites.

Grazing

Chowilla Station has been operated as a wool-growing business since 1865. The floodplain portion (12 062 ha) of the Station owned by the South Australian Government (DWLBC, no date) is managed according to a Game Reserve Management Plan overseen by the DEH. Monitoring programs have been in place for a number of years, to measure the impacts of grazing on floodplain vegetation. However analysis of grazing impacts on the floodplain vegetation is difficult to determine due to the complexity of the influencing factors on the floodplain. Degraded lunettes adjacent to intermittent lakes on Chowilla and nearby Calperum Stations are evidence of high grazing pressure in the 1860s (O'Malley and Sheldon 1990, p. 205). A historic overland stock route which once passed through Chowilla Floodplain would have also resulted in over-grazing (Harper, 1997).

Total grazing pressure is likely to be more significant than just domestic stock grazing pressure. Populations of feral herbivore species such as pigs, goats, European rabbits and European hares exist at Chowilla Floodplain. High density grazing in Chowilla Floodplain by a combination of domestic, feral and/or native herbivores or just one herbivore species can reduce habitat value, introduce and disperse exotic plant species, alter the species composition and greatly suppress regrowth of native vegetation (O'Malley & Sheldon, 1990, pp. 205-8).

Under current management, Chowilla Station carries relatively conservative numbers of sheep, and grazing of the anabranch islands is confined to a light summer grazing regime (O'Malley & Sheldon, 1990, p. 205). The Chowilla vegetation survey conducted in the late 1980s by O'Malley (1990) revealed 22% of the plant species present were introduced, which was significantly lower than the 33% recorded for the entire Murray River floodplain by Margules et al. (1990). Also, the presence of recently established seedlings of River cooba, River red gum, Blackbush and Black box at Chowilla was indicative of light grazing pressure.

Irrigation

Small areas of irrigated lucerne or clover have occasionally been established, with the Kulcurna/Tareena Pastures Station being the most significant, with a licence to irrigate 101 ha (Sharley & Huggan, 1995, p. 76). Currently there is about 15 ha of irrigation on Chowilla Floodplain (Tony Herbet, DEH, pers. comm., June 2004). There are no irrigation areas adjacent to Chowilla Floodplain.

Recreation

The Chowilla Flodplain area is one of South Australia's most important recreational areas due to the weir pool at Lock 6 maintaining high water levels through the anabranches, creating good fish habitat as well as year round boating opportunities.

The Chowilla Floodplain area is also recognised as being the most valuable of all areas in South Australia for the canoeing component of outdoor educational programs for secondary schools, tertiary educational classes and youth agencies. The Chowilla Floodplain experiences more than 2,500 camping nights per year, and the area is a popular fishing and hunting site for locals (Harper, 1997).

A significant proportion of visitation to Chowilla Floodplain is via pleasure craft using the main stream of the River Murray anabranch creek systems. State laws enable unrestricted public access to waterways from the main river channel regardless of the type of land tenure and/or ownership.

The majority of vehicle-based access is centred on the Chowilla Game Reserve and Regional Reserve, which is an integral part of the Bookmark Biosphere Reserve. It comprises 17,508 ha (Game reserve) and 17,614 ha (Regional Reserve) of floodplain and wetland.

Commercial fishing

In the 1850s when commercial fishing commenced in the River Murray, Golden perch and Murray cod were the two main species harvested. Since this time catch rates in the River Murray have declined and have been reported as the lowest per square kilometre of floodplain of any major river in the world. In the past catches were on a par with those of other rivers (Chowilla Working Group, 1991).

The Chowilla Floodplain area is important for native fish because Lock 6 causes more regular small overbank floods than elsewhere on the lower Murray, thus permitting both survival and breeding. As a natural fish hatchery, Chowilla Floodplain is likely to be important for re-stocking downstream areas. In the late 1980s, the Chowilla anabranch provided about 15% of the South Australian commercial catch from the river fishery (Chowilla Working Group, 1991)

Certain species of native fish are protected in the River Murray in SA. There is a closed season for fishing Murray cod from 1 September to 31 December inclusive. Catfish, Silver perch and Murray crayfish are totally protected. Gear restrictions also exist with mesh nets, gill nets and bait net fishing currently banned (DEH, pers. comm., June 2004).

Management of community uses of Chowilla Floodplain

The Chowilla Community Consultation Program was established in July 1990 to consider a wide range of management issues in the Chowilla Floodplain, including recreation and fishing (Chowilla Working Group, 1991, p1). A new program of community engagement and consultation has since replaced this program. Originally the Chowilla Community Consultation Program recommended provision of facilities that would encourage containment of vehicle-based campers. Recreational fishing and game hunting are managed under the State Fishing Acts. Open days for waterfowl hunting are also permitted under the established practices used in Game Reserves under the National Parks and Wildlife Act (Sharley and Huggan 1995). DEH undertakes active management of recreational use, development and maintenance of visitor facilities and public access to the Game Reserve.

5.2.16 Summary of knowledge of the condition of Chowilla Floodplain

The high conservation value of the Chowilla Floodplain has long been recognised, evidenced by Ramsar listing of the Riverland Wetland Complex (which contains Chowilla Floodplain) in 1987. Chowilla Floodplain has a high diversity of terrestrial and aquatic habitats, supports populations of rare or endangered species, has fish breeding habitat, supports populations of breeding waterbirds, has habitats not well represented elsewhere, and has a relatively low level of direct human disturbance. Despite its high conservation values, surveys have demonstrated significant environmental changes in the Chowilla Floodplain. The ecological health of Chowilla Floodplain has long been affected by many factors, including a history of livestock grazing, timber cutting to fuel river boats, construction of Lock 6 and regulation of flows in the River Murray.

Several tree dieback events have been reported in historical times, and these have been attributed to three main factors: prolonged drought, reduced flooding, and soil salinisation. River regulation has reduced the frequency of suitable floods for watering Black box communities from once every two to ten years to once every five to twenty years, and there is also reduced opportunities regeneration as a result of less frequent flood events. Surveys show that currently 45% to 55% of Black box communities are in poor condition. Reduced wetland flooding has caused a decline in some waterbird numbers. Frog numbers declined from 1960, with river regulation, increased salinity and turbidity, climatic change and environmental pollutants being implicated as possible causes. The anabranches of Chowilla Floodplain provide the only significant spawning grounds and nursery habitat in South Australia for Golden perch, Murray cod and Silver perch. Although there are few historical data, declines in the populations of these native fish have been reported for the lower River Murray under regulated flow conditions. Low species richness of aquatic macrophytes recorded along anabranches was attributed to high flows, and grazing by stock, particularly goats. Some of the banks on the faster flowing anabranches were actively eroding, making conditions unsuitable for establishment of plants. The health of wetlands at Chowilla Floodplain have been degraded by grazing and trampling by stock, high rabbit, goat and kangaroo populations, introduction and dispersal of exotic plant species, and river regulation. Temporary wetlands now flood less frequently. Prior to regulation, anabranch creeks were dry for the greater part of the year, but now they are permanently flowing.

Since the construction of Lock 6, the surface soil layers of the Chowilla Floodplain have accumulated an additional 60,000 ML of saline water, containing about 2 million tonnes of salt. This saline water, now stored at shallow depth, impacts on the root zone of vegetation, lake beds and stream beds.

The knowledge base for Chowilla Floodplain is extensive and comprehensive, as described in a large body of literature, but there is scope to improve on this. The various data on the ecological components of the Chowilla Floodplain (i.e., birds, fish, mammals, amphibians, macroinvertebrates, algae, vegetation) have not yet been compiled into a single database and summarised in terms of relevant and consistent measures of diversity, extent, abundance, and long-term trends. This would be a worthwhile exercise that would highlight weaknesses, inconsistencies, or gaps in data, leading to more streamlined monitoring. Such monitoring data are useful for describing relative environmental condition, and trends.

5.3 Factors impacting on environmental values of Chowilla Floodplain

5.3.1 Degree of change to flows due to river regulation

The Chowilla Floodplain is downstream of:

• locks 11 to 7;

• Lake Victoria outlet (Rufus River);

• the Darling River confluence with the River Murray;

• major headworks storages on the River Murray (Hume and Dartmouth dams); and the Lower Darling River and the Menindee Lakes, and

• three of the largest point diversions from the River Murray [Mulwala Canal, Yarrawonga Main Channel and National Channel (Torrumbarry)], and is directly impacted by the Weir pool.

The distribution of the median daily flow for each month under natural and current conditions at the South Australian border is shown in Figure 5.17. The area between the curves is a measure of volumetric flow change (mainly diversions and losses). The mean flow in the River Murray at the South Australian border has been reduced to 46% of the natural volume (Maheshwari et al., 1993, p. 17; Maheshwari et al., 1995), while the median flow is 39% of natural (Gippel et al., 2002; Gippel & Blackham, 2002). Figure 5.17 also shows the entitlement flow to South Australia in each month.

By the time the River Murray enters South Australia, the flow seasonality is similar to that of the pre-regulation regime as many tributaries have entered the river system (Thoms et al., 2000, p. 116) (Figure 5.17).

 

Flooding of the Chowilla Floodplain was a regular event prior to regulation (Sharley, 1992; Thoms & Walker, 1993; DWLBC, 2002a). The change in flood frequency between natural and current conditions is illustrated in Figure 5.18 by the change in the pattern of colours. The four thresholds in Figure 5.18 were selected simply to represent a range of flood magnitudes. These thresholds may have some particular ecological significance (see Figure 5.13), but they were not selected on that basis. To summarise the change in flooding frequency for these thresholds:

• flows >33,100 ML/d that occurred in 95% of years
now occur in 48% of years;

• flows >52,000 ML/d that occurred in 77% of years
now occur in 29% of years;

• flows >78,000 ML/d that occurred in 45% of years
now occur in 12% of years;

• flows >90,000 ML/d that occurred in 37% of years
now occur in 10% of years; and

• maximum inundation of the Chowilla Floodplain now occurs less than once every 25 years compared to the natural frequency of once every eight years.

It should be noted that Figure 5.18 is based on the average daily flow for each month, calculated by dividing the monthly flow by the number of days. The actual peak flood flow experienced in each month is almost certainly higher than these monthly average threshold values. Thus, the analysis displayed in Figure 5.18 is only indicative of changed flooding patterns.

 

The River Murray in the Chowilla Floodplain area now has prolonged periods of low flow. This is illustrated by the fact that South Australia's entitlement flow is received on average in 55% of months (Thoms et al., 2000). Low flows (<5,000 ML/day) occur 66% of the time under regulation, but prevailed for only 7% of the time under natural flow conditions (Wittington et al., 2000).

Changes to the frequency and duration of all but the largest floods have occurred. The largest impact has been on medium sized flows (both in frequency and duration). (Table 5.8; Figure 5.36) (Sharley, 1992; Thoms & Walker, 1993; Sharley & Huggan, 1995; DWLBC, 2002a). Flow durations for floods less than 100,000 ML/day have decreased by up to two months per year after regulation (Table 5.9; Figure 5.37).

For floods greater than 100,000 ML/day there appears to have been no change in flood duration. Large floods are rare, with a single event of 250, 000ML/day having been recorded once in the past 100 years (Sharley & Huggan, 1995, p. 94). The effect of regulation on floods of this size (>100 000ML/day) is minimal as the river breaks its bank, with water spilling out extensively on the floodplain.

5.3.2 Influence of weirs

<¸

In some areas of Chowilla Floodplain the water table is now around 2-3 m higher than it was prior to regulation (Akeroyd et al., 1998), due to the mound effect of the weir associated with Lock 6 (Sharley & Huggan, 1995). Other floodplains downstream of Chowilla and upstream of weirs are affected in similar ways, including Reny Island upstream of Lock 5 and Pike River upstream of Lock 4 (Woodward-Clyde, 2000). Other impacts of weirs on Chowilla Floodplain are as for Lindsay-Wallpolla Islands (see Section 5.3.3).

5.3.3 Salinity and Groundwater

Salinisation of floodplain soils is a major factor in the decline of the health of floodplain trees in the Chowilla area. Salinisation is linked to both groundwater and flooding processes. The primary cause of salinisation is increased rates of evapotranspiration (and hence movement of salt up into the plant root zone) due to reduced flooding frequency and rising groundwater (Overton & Jolly, 2003, p. 1). The combination of a semi-arid climate with surface clay soils of low permeability means that, generally, there is little leaching of salt between floods (Overton & Jolly, 2003, p. 15). Regular floods are important because they recharge the soil and groundwater, and flush salt that has accumulated through the dry period from the tree root zones (Jolly et al., 1993; MDBC, 2003, p. 13; Overton & Jolly, 2003). During non-flood growing periods, trees extract water from both the soil and the capillary zone of the groundwater, leaving behind the salt (Jolly et al., 1993; MDBC, 2003, p. 26; Overton & Jolly, 2003). Rises in the naturally saline groundwater are due to the effects of river regulation from Lock 6 (Overton & Jolly, 2003, p. 1).

Sharley and Huggan (1995, p. 37) concluded that rising saline groundwater was the primary cause of degradation of River red gum and Black box communities at Chowilla Floodplain, although currently it appears that drought is the main factor (Overton & Jolly, 2003, p. 15). Sharley and Huggan (1995) found that in the higher elevation areas where flood frequency was less than once every ten years, the degradation was worsened through lack of flushing of salts from the soil. Currently, about 35-50% of the floodplain area does not receive flooding at least once every ten years (i.e., not flooded at 82,640 ML/d in the River Murray), while before regulation the entire floodplain was flooded on average once every ten years (Sharley & Huggan, 1995, p. 94).

5.3.4 Other factors

While flow regulation, and changes to salinity and groundwater appear to be the main agents of environmental change in the Chowilla area, other factors such as grazing by sheep, feral and over-abundant native animals also influence vegetation condition, but they have less impact on tree condition. In the past, large quantities of timber were cut from the floodplain to fuel river boats and the Renmark irrigation pumps (Crabb, 1997, p. 64).

The current drought is causing tree stress (MDBC, 2003). While droughts are a natural phenomenon on the River Murray, they have a more serious impact if the severity of the drought is increased by regulation, or if flow regulation and other factors have placed the vegetation in a more stressed condition than it would otherwise be. This appears to be the case at Chowilla Floodplain.

5.4 Opportunities to meet objectives for this site-Chowilla Floodplain

5.4.1 Introduction

Floodplain salinisation, reduced flood frequency, rapid flood recession, and stable weir pool levels are the main hydrologically related causes of environmental problems at Chowilla Floodplain. Actions that could improve the environmental condition of Chowilla Floodplain include increasing the frequency and duration of flows (either by enhancing natural floods by manipulating releases from Lake Victoria, Menindee Lakes or Lock 6, or pumping onto the floodplain or into wetlands), and groundwater/salinity management (Overton & Jolly, 2003, p. 64-8). Also, improved land management activities such as reducing the grazing pressure from stock, feral animals and over-abundant native species would probably lead to improved environmental health.

5.4.2 Integrated catchment and salinity management

In 1992, after a period of intensive research and public consultation, an integrated resource management plan for Chowilla Floodplain was developed by South Australia focusing on saline groundwater management and rehabilitation of floodplain biodiversity. Between 1993 and 1997, $1 million was spent on Stage One of the rehabilitation of Chowilla Floodplain, which focused on wetlands rehabilitation, revegetation, protection of high conservation sites and management of recreational access.

In 2002 the South Australian Department of Water, Land and Biodiversity Conservation implemented the Chowilla Integrated Natural Resource Management Project as a multi-disciplinary project jointly funded by the National Action Plan for Salinity and Water Quality, the Natural Heritage Trust and the Murray-Darling Basin Commission (DWLBC, no date; DWLBC, 2004). The objectives of the project are to:

• enhance and where possible, restore the environmental values of the floodplain, including meeting the obligations under the Ramsar Convention. Action is linked to saline groundwater management activities, maximisation of environmental flow opportunities, managing the effects of weir pools to enhance wetland health, managing grazing pressure and protecting cultural heritage sites;

• remove saline groundwater from the Chowilla Floodplain with the aim of minimising its effect on salinity levels in the River Murray and maximising benefits to the biodiversity of the floodplain; and

• involve all sectors of the community in the enhancement and restoration of the health of the Chowilla Floodplain.

5.4.3 Enhancement of natural floods to increase their peak or duration using water releases from storages

Recommendations for flood enhancement to improve the condition of floodplain vegetation have been around for some time (e.g. Jolly et al., 1993), but these have been tempered by the practical reality that, for Black box Woodland communities at least, there are few opportunities to generate the very large floods required to flood the high elevation areas. Also, floods do not remove salts from all areas of the floodplain, and there is the disbenefit of generating stream salinity (Walker et al., 1996; Akeroyd et al., 1998; Slavich et al., 1999b). Despite this, Sharley and Huggan (1995) made suggestions on enhancing River Murray flow peaks, through releases from Lake Victoria, to extend the area and frequency of floodplain vegetation watered on the Chowilla Floodplain wetlands. This work estimated the area flooded for different flow rates, likely environmental benefits, and possible negative impacts. Although salinisation was perceived to be the main cause of degradation of vegetation, Sharley and Huggan (1995, p. 37) recommended reinstatement of more floods to the current flow regime based on the positive response of Chowilla Floodplain's severely salinised vegetation (especially Black box and Lignum communities) following floods in 1990 and 1992.

A recent study of the health of River red gum in the lower River Murray (MDBC, 2003) supported the recommendation of Sharley and Huggan (1995) that inducing significant flood events was an appropriate management response to mitigate tree decline. The most efficient way to create an effective flood is to enhance a natural river flood event with well-timed releases from one or several storages. The `piggy-back' release of water from Lake Victoria, Menindee Lakes or even upstream storages on key tributaries would increase the extent of inundation of floodplain woodlands and wetlands at Chowilla Floodplain. However, the impact of deliberate floods on other river users would also have to be considered (Sharley & Huggan, 1995). One favourable aspect of implementing this form of management action at Chowilla Floodplain is that, due to travel times of flows, there is normally at least several days to plan a piggybacking release from storage for preparation (as opposed to the Barmah Millewa Forest, where following a flood in the Ovens River, decisions on release have to be made almost immediately).

Overton and Jolly (2003) recently compiled the findings of efforts to date on modelling the impacts from future management scenarios on vegetation health on the Chowilla Floodplain. The modelling examined both groundwater lowering and enhanced flooding options. For sites with the water table less than 3 m deep, water table lowering may provide slightly better benefits than the modest increases in flooding, depending on delivery and management. For sites with water tables 3 to 4.5 m deep the relative benefits of these two approaches will be variable, depending on soil type, salinity and elevation. Groundwater discharge is close to zero for groundwater tables deeper than 5 m (Overton & Jolly, 2003, p. 66).

Increased flooding frequency will improve vegetation health without lowering groundwater by changing the balance of leaching to discharge. A large percentage of the Chowilla Floodplain area requires up to 100 days of inundation per five years to prevent salt accumulation, but there are other large areas that require from 100 to 1,800 days of inundation per five years (Overton & Jolly, 2003, p. 66). In the flood enhancement approach (the implementation of which still relies on the occurrence of natural floods) the risk remains that extended droughts will cause repeated dramatic declines in vegetation health (Overton & Jolly, 2003, p. 66). Also, water returning to the river from the floodplain can carry high loads of salt, with the higher the flood peak, the higher the salt load. Another limitation is that the high areas of the Chowilla Floodplain require large floods for inundation. Not only are large floods infrequent, but their top-up will potentially require large volumes of environmental water in order to achieve the required level for the required number of days. Overton and Jolly (2003, p. 66) suggested that such high elevation areas could be considered sacrificial areas.

In October 2000 a flow release of about 100 GL was made from Lake Victoria, successfully flooding several low-lying wetlands upstream of Lock 5, just downstream of Chowilla (DWLBC, 2002a). This was not `new' environmental water, rather, the release was used to re-shape the hydrograph. This flow release was made in conjunction with a trial weir raising using stop logs on Lock 5. The ecological responses to this managed flow event, which could become a routine management action under the First Step Decision, is discussed later in this chapter.

5.4.4 Complementary structural and operational measures

The Living Murray Environmental Works and Measures Program is a seven year, $150 million program to deliver structural and operational works and measures to improve the health of the River Murray by making the best use of water currently available (MDBC, 2003b; MDBMC, 2004). A number of projects completed under past programs, and others already commenced under the Environmental Works and Measures Program, address flow management issues within the Chowilla Floodplain. The total cost of these structural and operational works and measures has not been firmly estimated, but the Environmental Works and Measures Program has provided an indicative budget allocation of $18.8 million for the entire Chowilla Floodplain and Lindsay-Wallpolla Islands SEA (MDBMC, 2004, p. 36). The final budget allocation will be determined through the suite of project development activities already commenced within the program.

Enhancement of flow management structures

The project `Environmental enhancement of the Chowilla Floodplain' (being undertaken by Department of Water, Land and Biodiversity Conservation, South Australia) as part of the Environmental Works and Measures Program will investigate the options to enhance and extend the existing network of ecologically sensitive flow management structures at Chowilla Floodplain (MDBMC, 2004). Flowing environments cannot be restored to the main channel of the River Murray in the weir pools, but flowing environments can be restored in anabranch creeks. Opportunities to restore flowing environments exist wherever channels straddle weirs, including Slaney Creek and Pipeclay Creek. These channels are possible locations for regulator modifications, as well as Boat Creek and Bank E. Not limited in scope, the options investigated may include the provision of additional or refined regulators and construction of levees to control or manage flows through the floodplain. The project will also inform the construction of a salt interception scheme to maximise floodplain health, and investigate a range of flow and salinity management scenarios in terms of floodplain health and fish habitat outcomes.

Experimental watering of the Chowilla Floodplain

The project `Experimental watering of the Chowilla Floodplain' (undertaken by the Department of Water, Land and Biodiversity Conservation, South Australia) was recently completed as part of the Environmental Works and Measures Program (MDBMC, 2004, p.38; DWLBC, 2004). This pilot project, undertaken on the Monoman Island Horseshoe, trialled one of the tools under investigation for general use on the River Murray SEAs (manipulated flooding using mobile pumping infrastructure). The implementation and results of the experiment are described later in Section 5.8 of this report.

Modification of locks and weirs

Relatively constant river stage height is not ecologically ideal. The ability to manipulate weir and lock levels provides the operational flexibility to artificially enhance flood events by raising weir levels to create ponding or flow retention (Ohlmeyer, 1991; Blanch et al., 1996), or alternatively to lower weir levels to reinstate a natural drying phase to wetlands that have been artificially inundated for long periods of time as a result of flow regulation. Blanch et al. (2000) predicted that an increase in amplitude of river level fluctuation during low flows from the current 0.1-0.2 m to 0.2-0.5 m would be sufficient to reinstate water regimes suitable to the majority of plant species they surveyed.

In addition, the ability to manipulate locks and weirs will greatly enhance the opportunity to regulate floodplain connectivity by altering the extent of inundation. It should be noted, however, that raising of weir pools can only provide benefits to a limited area upstream of the weir pool, due to backwater effects, whereby the increase in water level declines in the upstream direction. A trial weir raising using stop logs on Lock 5 was undertaken in October 2000. The ecological responses to this are discussed later in this report.

The `Tristate weir pool manipulation project', as part of the Living Murray Environmental Works and Measures Program, is investigating the capacity to raise water levels at weirs based on current load capacities, estimating costs and developing options associated with structural adjustment of weirs, and determining the structural modifications required for various weir raising options (MDBMC, 2004, p. 43). Specific investigations are underway in South Australia to review the function and operation of all the banks and weirs (including Lock 6) and there is potential for some of these structures to be modified or removed to enable more effective delivery of water to critical areas on the floodplain (DWLBC, 2004).

Improved flow management in Lindsay-Wallpolla Islands system

Flows through Lindsay-Wallpolla Islands system have a direct affect on flows into the Chowilla Floodplain, with particular effects on salinity. Flows through the Lindsay-Wallpolla system are currently controlled by a structure at the top of Mullaroo Creek and, as this structure will need replacement in the future, investigations into the options for provision of a more flexible structure to enable more targeted flow management through this part of the system have commenced as part of the Environmental Works and Measures Program (see Section 5.4.3).

Lake Victoria outlet modifications

Improved flow management in Lindsay-Wallpolla Islands system

Flows through Lindsay-Wallpolla Islands system have a direct affect on flows into the Chowilla Floodplain, with particular effects on salinity. Flows through the Lindsay-Wallpolla system are currently controlled by a structure at the top of Mullaroo Creek (see previous Section 5.4.3).

Lake Victoria outlet modifications

To increase the peak flow magnitude and duration of floods in the Lower River Murray, feasibility studies have investigated the options for improved flexibility in the operation and outlet capacity for Lake Victoria.

For Lake Victoria, the option to manipulate environmental flows through provision of additional outlet capacity and capacity of downstream Rufus River is being investigated, including investigations of the capacity to fill Lake Victoria on a rising flood and the need for additional riverbank protection works to prevent erosion. Another potentially more costly option under investigation is the provision of a regulator on the southern bank of the inlet channel to Lake Victoria.

Despite achievement of an increase in channel capacity, the potential to use releases from Lake Victoria to top up floods may be limited by the small difference in elevation between the Lake and the River Murray. However, this requires further investigation.

Menindee Lakes outlet modifications

Another option for increasing peak flow magnitudes and duration of floods in the Lower River Murray and to increase flood frequency is increasing the Menindee Lakes outlet capacity (from 5,000 to 10,000 ML/d) and modifying operation of the Lakes. As the Lakes provide a significant storage volume, the ability to manipulate higher outflows from the Menindee Lakes would greatly enhance the flexibility to manage flows through the Chowilla Floodplain.

Provision of mobile pumping infrastructure

As indicated in Figure 5.15, there are climatic sequences in which large areas of the Chowilla Floodplain would not receive water from the river for over a decade, resulting in the death of individuals of many species including River red gum trees. Where the alternative may be no watering, one choice is the provision of mobile pumping infrastructure to enable target watering of critical parts of the floodplain. While the required pumping capacity is significant and may best be provided through a combination of mobile pumping equipment and on-ground permanent earthworks, the potential benefits for floodplain and wetland vegetation, waterbirds and fish may be significant.

The project `Mobile pumping infrastructure' (to be undertaken by the Murray-Darling Basin Commission) will begin in July 2005 as a complementary project as part of the Environmental Works and Measures Program (MDBMC, 2004, p. 51). The project will investigate the environmental benefit and economic feasibility of using mobile pumps to maintain and/or enhance floodplain health and to recommend a sustainable solution to maintaining floodplain health in the longer term. The mobile pumping infrastructure could potentially also benefit other significant assets including the Gunbower and Koondrook-Perricoota Forests and Hattah Lakes, as required.

Salinity management

Options to reduce floodplain salinity at Chowilla have focused principally on lowering the floodplain water table, reducing groundwater flow towards the river and increasing floodplain inundation (Sharley & Huggan, 1995). The saline water table beneath the floodplain can be lowered directly by groundwater pumping. An extensive network of groundwater wells are being investigated at Chowilla Floodplain (DWLBC, 2004). The groundwater wells will reduce the discharge of groundwater to wetlands and creeks and reduce the potential for evaporative concentration of salt in the soil. A drilling and pump testing program has recently been completed on the Chowilla floodplain at Gum Flat in South Australia (and Tareena Bong in New South Wales) (Howles & Marsden, 2003). The results of the investigation program indicated that the hydrogeology of the sediments underlying the Chowilla floodplain is more complex at the local scale than previous investigations had indicated. Aquifer hydraulic parameters, determined from the results of the pump tests, have been used for input to a numerical groundwater model of the region to allow broadscale planning of a conceptual wellfield design (Howles & Marsden, 2003). There is a need to integrate salinity interception and floodplain inundation options to maximise the benefits of improved flows (Overton & Jolly, 2003).

Overton and Jolly (2003) demonstrated an example of the use of the DEH conservation value in combination with environmental risk modelling (in this case, areas that do not have very shallow groundwater (<1.5 m), and areas with flooding requirements over 70,000 ML/d), to prioritise areas for rehabilitation. This produced a map of preliminary identification of the areas in the Chowilla Floodplain that would be most benefited by groundwater extraction (Figure 5.21). In reality, the first stage for prioritising areas would be to identify the best areas from the biodiversity value, and then determine what management actions will rehabilitate or maintain these areas (Overton & Jolly, 2003, p. 68).

5.4.5 Links between ecological objectives and management opportunities-Chowilla Floodplain

The ecological health of Chowilla Floodplain has been affected by many factors, including a history of livestock grazing, timber cutting to fuel river boats, construction of Lock 6 and regulation of flows in the River Murray. Several tree dieback events have been reported in historical times, and these have been attributed to three main factors: prolonged drought, reduced flooding, and soil salinisation. It has been suggested that reduced flooding has negatively impacted waterbirds, frogs and fish. Thus, while other factors are known to affect ecological condition of the Floodplain, flow regime is of fundamental importance. Floodplain salinisation, reduced flood frequency, rapid flood recession, and stable weir pool levels are the main hydrologically-related causes of environmental problems at Chowilla Floodplain.

Increasing flooding frequency (and/or increasing flood duration) by topping up naturally-occurring floods will improve vegetation health. Investigations are underway into modification of Menindee Lakes and Lake Victoria structures and operations to enable more efficient top-up of natural floods. These floods will be mainly aimed at watering River red gum, but a proportion of Black box will also be targeted. Proposed structural works may include the provision of additional or refined regulators and construction of levees to control or manage flows through the floodplain. Specific investigations are underway to review the function and operation of all the banks and weirs (including Lock 6) and there is potential for some of these structures to be modified or removed to enable more effective delivery of water to critical areas on the floodplain. Mobile pumping technology has already demonstrated environmental benefits in terms of groundwater salinity and levels, and maintenance and/or enhancement of River red gum tree health.

5.5 Examples of implementation of environmental flows

5.5.1 Experimental watering of River red gums on the Monoman Island Horseshoe, Chowilla Floodplain

Background

The Monoman Island horseshoe is an old anabranch of Chowilla Creek and has numerous River red gums of varying ages lining its banks. This creek would normally start to flood at a River Murray flow of approximately 60,000 ML/day. Prior to river regulation, a flood of this magnitude would occur six in ten years. Now, it floods on average 21 times every one hundred years, and the last time the creek was fully inundated was 1993, although it was partially inundated in 2000 (DWLBC, 2004).

As part of The Living Murray Works and Measures Program (MDBMC, 2004), an experimental watering of River red gums on the Monoman Island Horseshoe, Chowilla Floodplain, was undertaken in May 2004. The objectives of the project were to:

• investigate the response of River red gums to an artificially created flood regime; and

• identify the logistical and administrative processes required to implement watering schemes (in this case, flooding of relict meanders) on floodplains to achieve environmental outcomes (DWLBC, 2004).

Three coffer dams were constructed to contain water in the creek and 141 ML of water was pumped into the creek bed to fill and maintain a height of 19 m AHD for four weeks. This is equivalent to a natural flood of 73,000 ML/d (DWLBC, 2004). The approximate area flooded was 6.5 ha. The water used to fill the creek came from the licence held by the Minister for Environment and Conservation (#2095), which is part of his share of the Loxton rehabilitation works (DWLBC, 2004).

Results to date

Within one week of flooding, decreases in groundwater salinity of as much as 23,000 EC (from 23,700 EC to 394 EC) were measured. While salinity of the groundwater varies across the site, it was also very high (ranging from 41,000 EC to 15,000 EC; averaging around 25,000 EC). Likewise, the magnitude of the response also varied, although all piezometers showed significant decreases in salinity levels as a result of the flooding (DWLBC, 2004).

Prior to flooding, groundwater was in the lower end of the root zone for River red gums (4.5 to 5.0 m below surface). After the experimental flooding, groundwater levels rose, with most areas showing at least a 0.3 m rise.

The response in vegetation to the flooding was significant. While all of the trees on the horseshoe are stressed and many have lost leaves, within three weeks of flooding, those trees which had at least some leaves remaining showed signs of new growth (DWLBC, 2004) (Figure 5.22; Figure 5.23). Not all of the trees present have shown signs of response, especially those that had already lost all of their leaves.

Monitoring of tree health will continue in the long term, and a decision to re-flood the creek bed will be taken if it is considered the most appropriate action to maintain floodplain vegetation values (DWLBC, 2004).

5.5.2 Flood peak enhancement, combined with weir raising at Lock 5 (located downstream of Chowilla Floodplain)

Background

In the Chowilla Floodplain and Lindsay-Wallpolla Islands systems there have been no past observations of the ecological response to the types of flow management that would become routine under the First Step Decision. A trial of flood enhancement combined with weir surcharging was undertaken at Lock 5, just downstream of Chowilla in October 2000, and this is a reasonable guide to the responses that would be likely at Chowilla if a similar event was implemented there.

The October and December 2000 natural floods provided opportunities to enhance flows for the purpose of increasing the area of floodplain inundation in South Australia, and thereby stimulating growth in floodplain ecosystems (RMW & SA Water, 2000; RMW, 2000a). This event was particularly significant in demonstrating a greater emphasis on environmental values and was the first time that the Commission agreed to operate works to increase the peak flow to South Australia. This contrasts with past practice where Lake Victoria has been used to reduce flood peaks. Monitoring of environmental, community and infrastructure impacts was conducted by South Australian agencies (RMW & SA Water, 2000; RMW et al., 2000; DWLBC, 2002a).

A trial release of approximately 100 GL of water for flood peak enhancement was made, combined with weir raising at Lock 5. Lock 5 is approximately 52 km downstream of the downstream boundary of the Chowilla Floodplain. This was not a new environmental allocation, but rather an opportunistic river operation. However, the fact that the water was replaced in Lake Victoria (or Menindee Lakes) is irrelevant. If there had been 100 GL in Lake Victoria in an environmental allocation, the same result would have been achieved. This is analogous to having 100 GL in an environmental account in Lake Victoria to water part of the Riverland, or to produce other environmental benefits downstream (e.g., in the Coorong). If an opportunity to top-up a flood to Chowilla Floodplain did not arise in spring or early summer, then other downstream benefits could be targeted rather than allowing much of the water to evaporate or seeping away.

During the October 2000 event, the flow to South Australia was enhanced by an appropriately timed release of 9,700 ML/day from Lake Victoria (DWLBC, 2002a), so that a natural peak of 32,000 ML/day at the South Australian border reached a peak of 42,050 ML/day (RMW, 2000a). This enhanced flood resulted in increased flooding at Chowilla, but the specific effects were not monitored.

More widespread inundation was achieved by surcharging of the upstream pool level of Lock 5 near Renmark by about 0.5 m using stop logs on the weir (Figure 5.24). To ensure the stability of the weir at Lock 5 during the trial, it was necessary to also surcharge Lock 4 weir pool, by about 0.4 m, thereby raising the water levels downstream of Lock 5 and maintaining a safe upstream to downstream head difference (DWLBC, 2002a).

The Department of Water, Land and Biodiversity Conservation conducted a monitoring program during the October 2000 trial with a focus on the response of groundwater and surface water to the changed river operating conditions. The main objectives were to determine the extent of water level rise in the river channel and floodplain, and to determine the salinity change in the river, wetlands and backwaters. Fish movement and plant growth were also monitored (DWLBC, 2002a).

In the larger December 2000 event, the Commission agreed to enhance the peak to South Australia by about 9,000 ML/day using a release from Lake Victoria to achieve a peak of 63,400 ML/day on 17-18 December (RMW, 2000b; RMW et al., 2000). This peak discharge included a contribution of about 11,000 ML/day from the Darling River resulting from flood operation of Menindee Lakes.

Hydrological response

As a result of the October 2000 event, the area of floodplain that was inundated was increased by about 10%, increasing the area flooded by about 2,400 ha. The original peak of 32,000 ML/d would have inundated about 850 ha of Chowilla Floodplain (5% of the floodplain area); enhancement to 42,050 ML/day would have inundated about 1,500 ha of Chowilla Floodplain (9% of floodplain area). Weir surcharging did not affect Chowilla Floodplain, which is located beyond the upstream extent of the backwater.

The enhancement of flow during the December 2000 flood peak increased the area of floodplain inundated on the Chowilla Floodplain wetlands from 18% to 27% of the total area, inundating 4,600 ha, [as predicted by the inundation model of Sharley & Huggan (1995) (see Figure 5.9)]. On this occasion there were insufficient arrangements in place to provide additional floodplain inundation by surcharging of upstream pool level at Lock 5 and other weirs.

Lock 5 is located at river kilometre 562. While the peak enhanced flow at this location was 42,050 ML/d, the weir surcharging produced a river stage that was equivalent to about 70,000 ML/d in the proximity of the weir, with the backwater becoming insignificant 40 km upstream, or at river kilometre 602 (DWLBC, 2002a). The downstream boundary of the Chowilla Floodplain lies at river kilometre 614 (Sharley & Huggan, 1995), so Chowilla Floodplain did not benefit from the weir surcharging. The hydrograph of the event is shown in Figure 5.25.

Environmental responses

Since the construction of Lock 6, the surface soil layers of the Chowilla Floodplain have accumulated additional saline water. The October 2000 enhanced floodplain inundation event appeared to have a minimal impact on River water salinity. During the trial the River water salinity rose by around 20 EC units for 30 days, which was comparable to the rise recorded in a similar-sized flow event in 1998 (DWLBC, 2002a).

DWLBC (2002a) found that from a groundwater perspective, there were no environmental costs. Groundwater levels generally mirrored those of the river and creek levels. With two exceptions, groundwater EC units at all sites that had good hydraulic connection to the river was lowered as a result of the trial. This has the effect of providing lower salinity water to terrestrial floodplain vegetation. This benefit of the 0.5 m weir raising was likely to be short-lived and localised, with much of the degraded floodplain vegetation not receiving any real benefit. Most of the degraded Black box communities lie distant from the rivers and creeks, and at a higher elevation, and therefore require a flow of at least 60,000 ML/day. The flow enhancement and Lock 5 surcharge did create water levels similar to that of a 60,000 ML/day flow, but the effect was limited to the first 18 km of river upstream of Lock 5 (approximately 30% of the pool length). There were some concerns raised about longer-term raising of weir pool levels (for periods of months) possibly causing increased salt accumulation (DWLBC, 2002a).

It is often assumed that native fish (whether larvae, juvenile or adult) use the inundated floodplain as habitat during flooding periods, even though there is little documented evidence to support this view (Humphries et al., 1999). The trial therefore sought to monitor fish movement during flood flows. No adult fish were recorded attempting to move into the inundated wetlands during the trial. Some smaller natives were observed in large numbers trying to move into both the temporary and permanent inlet to Lake Merreti. The trial appeared to coincide with the local major annual Carp spawning event, and the raised water levels appear to have improved their breeding potential (DWLBC, 2002a).

The plant study, undertaken by the University of Adelaide, indicated that there was no response of aquatic plants, and only a limited response of amphibious species, suggesting low seedbank viability, or heavy grazing (Siebentritt et al., 2003). As anticipated, flood-tolerant and flood-dependent species grew and germinated and flood-intolerant species senesced. The floodplain perennial shrub Tangled lignum flowered at flooded elevations. There was a change in the floristic composition in ground-dwelling species as a consequence of the enhanced flood, with a shift towards a more amphibious species. However, these species accounted for less than 1.4% of the abundance and cover. It was assumed that the duration of the enhanced flow was inadequate to generate a response from aquatic plants (DWLBC, 2002a).

The Chowilla Floodplain vegetation inundation models of Sharley and Huggan (1995) predict that the October flood enhancement (from 32,000 ML/day to 42,050 ML/d) caused inundation of Saltbush (chenopods) and Open Plain swamp communities that would not have received water from the original 32,000 ML/d event (Table 5.4). In terms of absolute gain in area inundated, Saltbush (chenopods), River red gum communities, Open Plain swamps and Wetlands and Waterbodies were the main beneficiaries of the enhancement (Figure 5.26). The additional flooding achieved through weir surcharging probably increased inundation of some vegetation communities upstream of Lock 5, but the effect did not reach the Chowilla Floodplain.

As a result of the December event, the peak of which was enhanced from 54,400 ML/day to 63,400 ML/day, the main beneficiaries in terms of absolute gain in area inundated were Open Plain swamps, Black box communities, Lignum and River red gum communities (Figure 5.26 and Table 5.4).

The success of an enhanced flood event such as the Lock 5 trial rests upon recruitment of new individuals, and not merely the survival, growth and reproduction of established plants. While some new vegetative growth did establish in this case, recruitment of individuals from new seed was not observed. With respect to vegetation, the value of these small, occasional interventions in environmental flow management may be to maintain existing communities rather than restore degraded ones (Siebentritt et al., 2004). The main vegetation communities to benefit from flood enhancement, and the size of the benefit, depend on the magnitude of the flood event that is being enhanced. At Chowilla Floodplain, Black box Lignum and Open Plain swamps, in particular, receive much greater benefit from enhancement of larger floods (Figure 5.21). MDBC (2003, p. 21) reported expert advice that the two-month long inundation provided by the late-2000 to early-2001 flood on the Chowilla Floodplain was probably insufficient to provide adequate replenishment of groundwater sources for River red gum communities.

Preliminary review of social impacts associated with the enhanced flood

Tree health deterioration and loss on the Chowilla Floodplain is a major concern for many people (MDBC, 2003, p. 29). Flood enhancement partially addresses this `big picture' concern, but the implementation of such actions raises other short-term and local scale social issues.

Prior to and during the Lock 5 flood enhancement trial, numerous stakeholder issues were raised (DWLBC, 2002a). A professional fisherman, whom operated between Locks 5 and 6, reported that fishing was the best it had been since 1974 during the trial. Downstream of Lock 6 it was reported that fish had stopped moving with the suggestion that this was due to poor water quality. Concern was also expressed about the rise in water levels causing access problems (Figure 5.44) and breaching of levee banks. Houseboat owners expressed concern about stranding and the use of Paringa Bridge.

Calls to information lines regarding the trial were predominantly from Renmark and Paringa residents. The flood recession generated as many, if not more, inquiries than the peak of the event (DWLBC, 2002a).

A local community representative commented that the enhanced flow rate and Lock 5 surcharge used in the trial were insufficient to achieve real ecological benefits for Chowilla Floodplain, which was too distant from the well-watered areas (DWLBC, 2002a).

The growth of vegetation, noise of frogs, and return of waterfowl during the flood enhancement trial were perceived as indicators of ecological benefit by the local community (DWLBC, 2002a).

Part B - Lindsay-Wallpolla Islands

5.6 Value and condition of Lindsay-Wallpolla Islands

5.6.1 Conservation significance of Lindsay-Wallpolla Islands

Wallpolla Island is a State Forest while Lindsay Island is part of the Murray Sunset National Park. The Victorian Government and the Mallee Catchment Management Authority have identified both the Lindsay and Wallpolla Islands as high ecological value areas. Lindsay Island, Wallpolla Island and Lake Wallawalla are listed under the Directory of Important Wetlands (ANCA, 1996) and are nationally significant. The anabranches of the islands are also important native fish breeding habitats (Meredith et al., 2002). Although the areas of permanent and semi-permanent wetland in each site are small, they support species that are of national, state and local importance (ANCA, 1996). Lake Wallawalla is considered to be a `high value' wetland system in that it supports a range of significant flora and fauna as well as a number of Indigenous heritage sites (SKM, 2003a). The Lindsay and Wallpolla Islands sites are highly significant for Indigenous people.

Descriptions of the natural resources of the Lindsay-Wallpolla Island system can be found in Beovich (1994, 1995), ANCA (1996), Egis (2001) and SKM and Roberts (2003). SKM and Roberts (2003) noted that there have been few ecological or biodiversity studies in this area, with most reports recycling information from earlier studies. However, the data that are available clearly indicate that this system is of high conservation significance for Victoria. Egis (2001) reported that Lindsay and Wallpolla Islands support two plant species of national significance and 51 plant species of state significance. There are 80 plants that are of regional significance: 35 are found on both islands; 11 on Wallpolla Island only; and 34 on Lindsay Island only. SKM and Roberts (2003) were of the opinion that the apparently higher number of species on Lindsay Island can be attributed to a greater number and more intense field studies having been conducted there.

According to the department of sustainability (DSE) database, Lindsay and Wallpolla Islands have a total of 62 VROT plant species. About 29 of these VROT species are rare in Victoria because their habitat is rare in Victoria, which adds to the importance of Lindsay-Wallpolla Islands at the state level. SKM and Roberts (2003) estimated that only ten of the VROT species have flow-specific requirements, with most of these occurring in River red gum forests, or on moist soils of billabongs or river banks.

The Directory of Important Wetlands in Australia (ANCA, 1996) listed 17 birds in Lindsay Island and three in Wallpolla Island. For Lindsay Island, Egis (2001) reported five species listed under JAMBA (Department of Foreign Affairs 1995a) and CAMBA (Department of Foreign Affairs 1995b), as well as three listed under CAMBA only. Significant bird species that have been found at wetland sites within the Lindsay-Wallpolla Islands that are connected at 60,000 ML/d flow in the River Murray include the Pied cormorant, Royal spoonbill, Musk duck, Freckled duck, Great egret, Caspian tern and the Nankeen night heron (SKM & Roberts, 2003).

The higher diversity of listed bird species in Lindsay Island can be attributed to the presence of Lake Wallawalla, which attracts regionally significant numbers of waterbirds when flooded (SKM & Roberts, 2003). When flooded, Lake Wallawalla is an important area for waterbirds, with 34 species being recorded. At times this site can support up to 5% of the Victorian population of Maned duck.

A total of 14 fish species have been recorded from active channels of the Lindsay and Wallpolla Islands (Douglas et al., 1998; Meredith et al., 2002; SKM & Roberts, 2003). The fish community is representative of the River Murray fish community, which is listed under the Victorian Flora and Fauna Guarantee Act. Four fish species are identified as significant, having been listed under the Act. They are Silver perch, Flyspecked hardyhead, Crimson spotted rainbow fish and Murray cod (SKM & Roberts, 2003, p. 49). The study of Meredith et al. (2002) on Lindsay Island recorded Flat headed gudgeon in Mullaroo Creek for the first time, as well as the first record for Lindsay Island of the Murray hardyhead, which is listed as a vulnerable species. The anabranch creeks of the Lindsay Island provide important habitat for Murray cod breeding nurseries (Meredith et al., 2002). The diversity of habitats in this system probably explains the high fish diversity (SKM & Roberts, 2003).

5.6.2 Regional hydrology, and flow paths within Lindsay-Wallpolla Islands

The Lindsay-Wallpolla system is influenced by a number of locks and weirs on the River Murray (Locks 6 though to 10) that affect flow through the anabranch system (Figure 5.2). Lock 10 (Wentworth Weir) is just upstream of Wallpolla Island and is influenced by flows from the Darling River and Menindee Lakes as well as the Upper Murray. Lock 9, located immediately downstream of Wallpolla Creek, influences flow into the creeks and billabongs of the Wallpolla Island system. The interconnecting anabranches of Wallpolla Island form a number of wetlands. Dedmans Creek and Frenchmans Creek are the most active creeks in the Wallpolla system, being active all the time under regulated flows. Other creeks that are active under higher flow conditions include Finnigans Creek, Sandy Creek, Moorna Creek, Milky Creek and Thompsons Creek. Moorna Lagoon is permanently inundated by regulated flows (SKM & Roberts, 2003; SKM, 2003b).

The Lindsay River anabranch separates from the River Murray 10.5 km upstream from Lock 7 and rejoins the river about 7 km east of the South Australia-Victoria border (Figure 5.2). Lock 6 is located downstream of this confluence and the weir pool has a significant influence on the lower reaches of Lindsay River and Mullaroo Creek. Lock 7 weir pool controls flow into Mullaroo Creek and Lindsay River upstream of this point (SKM & Roberts, 2003). Lindsay Island and the associated anabranches are intersected by a number of smaller streams and billabongs. Mullaroo Creek and Toupnein Creek to a lesser extent (due to its smaller spatial coverage) are inundated under regulated low flows from the River Murray (SKM & Roberts, 2003; SKM, 2003b).

Figure 5.28 shows an idealised cross-section of Lindsay Island, and it is also representative of Wallpolla Island. The profile shows the major landforms and flow paths from the River Murray into Mullaroo Creek and Lindsay River. Groundwater recharge is a significant part of the regional hydrology of the Lindsay-Wallpolla Islands. Measurements of groundwater in the Channel Sands on Wallpolla Island indicate that the groundwater table is located at a depth of 2 to 5 m and has a salinity of 1 to 8 dS/m, although there was one recording of 40 dS/m (SKM, 2004a). RWC (1992) reported salinities in the Channel Sands near Lake Wallawalla in the range 22 dS/m to 41 dS/m.

Lake Wallawalla is a quasi-circular deflation basin located on the southern point of the system with a surface area of 828 ha. It is a significant source of groundwater recharge to the Lindsay Island system (Figure 5.28). It is separated from the main floodplain area of Lindsay River by a levee that impedes the natural flood path (the Mail Route Road). The Mail Route Road acts as an impediment to medium-sized floods, with several small culverts allowing limited flow (SKM, 2003a). Under natural conditions, this would have made a continuous connection with the floodplain. Black box communities surround the lake, with extensive sand dune deposits to the south-east. During flood inundation, Lake Wallawalla becomes a wetland and is a significant site for bird breeding (SKM, 2003a).

Of the water that flows from the River Murray into the Lindsay-Wallpolla Island systems, some is lost to groundwater recharge and evaporation, but the majority is returned to the River Murray downstream. The current flow paths are a product of the action of past fluvial processes. The soils of the area are relatively susceptible to fluvial-induced erosion over time. The floodplain gradient is relatively flat and lends itself to the development of a range of landscape types that are associated with anabranching river systems (SKM & Roberts, 2003).

Landforms that characterise the Lindsay Island system include meanders, meander scrolls, multiple terraces, lunettes, sand dunes, billabongs and wetlands (Ahern, 1999). Beovich (1993) recognised three main landscapes: the current river floodplain, high alluvial plains, and lunettes.

The flow path of the Lindsay River is such that initial flooding results in inundation of the existing anabranches that are not active under regulated flows, after which the banks and floodplain become inundated. Larger floods result in inundation of previously abandoned anabranches and ancestral floodplains.

5.6.3 Flow thresholds and flow capacity for Lindsay-Wallpolla Islands and Lake Wallawalla

For each flow path a flow threshold and channel capacity determines the timing and extent of flooding for the anabranch creeks, floodplains, wetlands, billabongs and abandoned ancestral channels. SKM and Roberts (2003) identified five flow thresholds (also known as stages) at which flooding of the various landscape features in the Lindsay-Wallpolla system occurs (Table 5.5).

The Murray Wetlands Working Group carried out a mapping project to examine flow thresholds of various wetlands using aerial photographs (Green & Alexander, 2003). A number of seasonal flow events were examined to indicate the landscape features that may be inundated for a particular flood event (Figures 5.7 to 5.9). Regulated River Murray flows result in Wallpolla Creek being subject to continuous low flows (of approximately 9,000 ML/d) (Figure 5.29). The 1995 flood with a peak of 66,000 ML/d resulted in most of the anabranches joining up to the Wallpolla Creek, while higher floods (up to and above 140,000 ML/day) result in the backwaters, abandoned channels and meander cut-offs being filled.

Potterwalkagee Creek, which forms Mulcra Island, flows continuously at low flows of less than 5,000 ML/d (Figure 5.30). Further downstream, off the River Murray, Lindsay River also flows continuously at low flows. At a peak flow of 40,700 ML/d the anabranch channels between Lake Victoria and Lindsay Island flood. Flow spills into smaller anabranches of the Lindsay River at 50,150 ML/d and onto the elevated floodplains at 122,500 ML/d. Downstream, Toupnein Creek is inundated, but ponded, at regulated low flows (Figure 5.31). Toupnein Creek begins to flow at Stage II flows (Table 5.5). At higher River Murray flows, Toupnein Creek, the Lindsay River and Mullaroo Creek break their banks to flood Lindsay Island (as occurred in the 1993 flood where the peak flow was 110,962 ML/day).

Lake Wallawalla is considered part of the Lindsay Island wetland system and fills from spill-over from the Lindsay River during medium-sized flood events. The late-2000 flood event, having a peak flow of 63,427 ML/d, connected the lake to the river. Lake Wallawalla also receives flows from groundwater sources. The groundwater level between the Lindsay River and Lake Wallawalla is affected by operation of the locks. Operation of the locks elevates surface water levels and increases seepage to groundwater (SKM & Roberts, 2003).

Analysis of the historical flow record under natural conditions indicates that Lake Wallawalla floods once every three to five years on average and requires two years for water to evaporate or drain away. In addition, the inundation and drying times for the lake are heavily influenced by operation of the locks along the River Murray and structures that block flows along the anabranches (such as the Mail Route Road) (SKM, 2003a).

Aerial photographs of past flooding events provide an indicator of the extent of anabranch and wetland inundation. Figure 5.32 and Figure 5.33 show flooding events for 1988, 1993, 1995 and 2000 for Wallpolla Island (no imagery was available for Lindsay Island at that time). The images show the site before and after a flood event (not during).

The assumption used for interpretation of the images is that the time taken between `before' and `after' the flood is short enough that water would not have substantially drained from the area, so maximum extent of inundation is assumed in the `after' images. The extent of inundation may have been greater if images were captured during the flood event (rather than `after'). However, generally, the elapsed time is short enough to allow an indicative snapshot as to the maximum extent of inundation for a given flood event.

The 1988 flood event had a peak of 49,168 ML/d (Figure 5.32). This resulted in an increase in floodplain inundation of Frenchmans Creek into Lake Victoria. However, there was no change in flow for Wallpolla Creek evident. Flows to Moorna Lagoon (the abandoned meander cut-off to the north of the River Murray) remained unchanged, although it should be noted that Moorna Lagoon is permanently inundated by regulated low flows.

In 1995 and 2000 flood events, with peaks of 66,600 ML/day and 57,600 ML/day respectively, extended the flooded area of Frenchmans Creek from that which was flooded in 1988 (Figure 5.11). However no significant additional flooding occurred in Wallpolla Creek or associated floodplains during the 1995 and 2000 flood events.

The 1993 flood event, which had a peak flow of 110,962 ML/d was large enough to flood the anabranch creeks, floodplains, abandoned meander cut-offs and backwater billabongs of Wallpolla Island (Figure 5.32). In this case, the connectivity between the River Murray and backwater lagoons was established. This event was large enough to flood the entire system and the elevated `Open Area' wetlands (see following section).

5.6.4 Vegetation distribution at Lindsay and Wallpolla Islands

The main vegetation types at Lindsay and Wallpolla Islands consist of: River red gum; Black box forest communities; and Lignum shrublands and grasses (associated with wetland habitat). For MFAT modelling, Roberts et al. (2003) included Lindsay Island as one of the floodplain sites. The configuration considered Lake Wallawalla and its `halo' (i.e., a circular fringe) floodplain of River red gum woodland as a waterbird complex, and assessed only two types of floodplain vegetation (Rats tail couch grassland and Lignum shrubland). These grasslands are mapped as Open Areas in River Murray Mapping, but include some small patches of lignum (Roberts et al., 2003). These classes of vegetation are consistent with those used in other literature concerning Lindsay and Wallpolla Islands.

River red gum provides key habitat for several species of special significance and River red gum forest has the highest number of significant taxa within Lindsay Island (Ahern, 1999). Black box woodland provides habitat for significant fauna, such as Carpet python and Apostle birds. This community has a high diversity of shrubs, is likely to be significant for plant diversity and to support invertebrates and frogs. The riparian fringes of Black box woodland contain rare Emu bushes (SKM & Roberts, 2003, p. 38). Lignum shrublands make up only a small proportion of the areas of Lindsay-Wallpolla Islands, but it is a distinctive and important type of wetland that is rare in Victoria (SKM & Roberts, p. 56). Lignum shrubland supports various species of conservation significance, including Paucident planigale, Red-naped snake, and probably the Tessellated gecko. When flooded, macrophytes develop in the inter-shrub area and provide habitat for shallow-water feeding waterbirds, and frogs (SKM & Roberts, 2003, p. 39).

Beovich (1994) and SKM and Roberts (2003) identified a fourth vegetation class termed 'Open Areas'. This class constitutes a large percentage of the area of both Wallpolla Island (3,087 ha) and Lindsay Island (6,912 ha). Open Areas often become wetlands during high flood events. These areas lack woody perennials and often appear bare when dry. Aquatic and semi-aquatic herbs will grow after the areas are inundated by a flood event. The history of flooding will determine the composition of species found in the Open Areas. Extensive grasslands of Rats tail couch occur on heavy grey clay areas. These were considered by Beovich (1994) to be regionally significant on account of their rarity.

The distribution of riparian vegetation communities classes for Wallpolla Island is shown in Figure 5.34 and for Lindsay Island in Figure 5.35.

5.6.5 Floodplain vegetation distribution and hydrological regime

Each of the three major vegetation community types (River red gum, Black box and Lignum) varies in floristic composition and in terms of seasonal water requirements. Hydrological variables that are important for plants to survive in a wetland system in the short term include flood depth, flood duration, and flooding frequency. The detailed response of vegetation to water depth varies significantly between communities. The hydrological needs of these vegetation communities have not been studied in detail at Lindsay and Wallpolla Islands, but relevant information can be obtained from other sites in the River Murray system (refer to chapters on Barmah-Millewa Forest, Gunbower, Koondrook-Perricoota forests and Hattah lakes SEAs).

The season and timing of flooding has a significant impact on the floristic composition of the wetland community given that germination and subsequent growth respond to variables (including climate) at a predefined point in the season. Roberts and Marston (2000) demonstrated that if flooding changes from winter to spring/summer, a change in species to favour those that grow in summer might occur. Similarly, in the absence of any seasonal flooding, some species will find it difficult to replenish seed stocks. Thus flood timing and frequency are just as important as the volume of water required for wetland inundation.

The Lindsay-Wallpolla Islands contain anabranches, some of which flow continuously (such as Mullaroo Creek and Dedmans Creek) while the other smaller anabranch creeks are ponded and flow seasonally. Thus, ideal watering conditions for the Lindsay-Wallpolla system require variation in low flow frequency and duration, as well as larger seasonal floods that inundate the floodplains and fill backwaters, such as cut-off meanders, open area wetlands and other small channels.

River red gum communities

River red gum are particularly well adapted to the variability of water availability from season to season. Studies from the Barmah-Millewa Forest show that ideal watering conditions occur once every two years with complete drying of the soil in between events (Chesterfield, 1986). However, the length of time they will survive during dry spells will depend on access to shallow aquifer water. Studies have shown that the species can survive for up to four years when permanently inundated (although younger specimens will die during this time) (Briggs & Maher, 1983; Bren, 1987). Although River red gum are resilient to variable hydrological conditions, frequent seasonal flooding provides the best conditions for their germination and growth. Sheet flooding is not believed to be essential for maintenance of River red gum, but it is vital for seedling establishment (SKM, 2003a). Thorburn et al. (2004) argued that groundwater, or the presence of a shallow aquifer, may be equally, if not more important, than surface water characteristics for River red gum health.

Studies from the Barmah-Millewa Forest show that winter-spring watering is ideal for River red gum, with the duration of flooding between four and seven months providing the best conditions for growth. Summer flooding, while providing adequate water for growth of small trees and maintenance of established trees, may alter the understorey community (Robertson et al., 2001).

River red gum covers an approximate area of 3,955 ha of the Lindsay Island and 1,730 ha of Wallpolla Island (SKM & Roberts, 2003). River red gums are located on the low-lying reaches of the River Murray that run though the west of the Lindsay-Wallpolla Islands. Isolated areas of River red gum are found along Mullaroo Creek, the Lindsay River and Wallpolla Creek. The fringes of Lake Wallawalla also support isolated stands. At low flows, water flows into the stands via creeks and small `active' channels of the Mullaroo, Lindsay and Murray rivers. At high flows, the stands become inundated, resulting in organic matter being washed into the anabranch creeks.

Analysis of inundation imagery by SKM and Roberts (2003) indicated that 10-70% of the Lindsay-Wallpolla combined vegetation community is inundated between 65,000 - 90,000 ML/d (Figure 5.36). Analysis of pre-regulation Lock 10 flows by SKM (2003b) showed that the normal frequency of these events before regulation was once every 1.3 to 1.6 years. Floods exceeding 56,000 ML/d ranged from one to six months in duration and were separated by dry spells not normally exceeding two years. Moderate to high flows provide a valuable vector for dispersing River red gum seed downstream. Although high flows are important, continuously high flows can be just as damaging as long droughts due to excessive waterlogging of root systems. Long periods of high flows can lead to channel erosion resulting in root exposure and tree falls (SKM & Roberts, 2003).

 

Black box Woodland communities

Black box is normally located at a higher elevation in the landscape than River red gum, due to its lower flood tolerance. The Black box community is healthiest when flooded for a period of between four and six months every four to five years. However, surveys from other areas of the River Murray system show that prolonged dry periods of up to 12 to 18 months can result in a significant loss of Black box (Roberts & Marston, 2000). As for River red gum, the most successful periods of regeneration of Black Box occur after large-scale flooding. In western New South Wales and on the lower Darling River, germination is most successful if it occurs during the cooler months-May to October (SKM, 2003a). Black Box seedlings do not have any specific adaptations for flood tolerance, and seedlings can tolerate flooding for about one month (SKM, 2003a).

Black box Woodland covers an approximate area of 4,797 ha of Lindsay Island and 5,423 ha of Wallpolla Island. It is the most common vegetation community on Wallpolla Island and the second most common vegetation community on Lindsay Island. Shrubs such as Moonah, Nitre goosefoot and Lignum grow in the understorey. Being located higher on the floodplain, Black box Woodland is not flooded as frequently as the River red gum community. The southern areas of Mullaroo Creek and west of Lake Wallawalla support degraded Black box communities. For the Lindsay Island, a flood event between 63,300 ML/day and 79,900 ML/day is the threshold for Black box inundation (Figure 5.36). On Wallpolla Island, the threshold for flooding occurs between 89,000 ML/day and 118,000 ML/day (Figure 5.36).

Lignum Shrubland

The most common shrub located within the Lindsay-Wallpolla Islands is Lignum, although Nitre goosefoot is a common plant on Lindsay Island as understorey in Black box woodland. Lignum is extremely tolerant of waterlogging, moderately tolerant of salinity, and vulnerable to drought (Craig et al. 1991). It appears to grow most vigorously when soil moisture is high, and when soil pH and conductivity are relatively low (Craig et al., 1991). Maintenance of Lignum populations requires flooding every three to ten years, with more frequent flooding required more in saline areas (Craig et al., 1991). Although Lignum is tolerant of waterlogging, it cannot survive in continuously wet conditions, requiring a phase of complete drying to cracking between floods (Roberts & Marston 2000).

Lignum Shrubland occurs on the low-lying areas of the floodplain on cracking grey or yellow-grey clay soils, distant from the continuously flowing anabranches such as Mullaroo Creek. As a result, medium to large floods are required to water Lignum Shrubland (SKM & Roberts, 2003). Based on River Murray Mapping, approximately 60 ha of Lindsay Island supports Lignum compared to 405 ha on Wallpolla Island, although this is likely to be a slight underestimate (SKM & Roberts, 2003).

The flow threshold for flooding of Lignum on Wallpolla Island is 88,580 ML/day to 118,000 ML/day in the River Murray at Lock 10. For Lindsay Island, Lignum is completely inundated at 79,000 ML/day in the River Murray. As Lignum in this system is known to occur on readily drained flat-sloping areas, it follows that here it is directly influenced by flood duration. Understorey species are often aquatic sedges and grasses that are short lived when drying out occurs after a flood event (also known as `opportunistic' wetland herb species) (SKM & Roberts, 2003).

Open Areas

Open Areas are characterised by a lack of woody perennials, and where vegetation does occur, chenopod species (e.g., bluebushes and saltbushes) are the major vegetation type (Beovich, 1994). Short-lived grasses and herbs also occur in the Open Areas, particularly after inundation by a large flood. Extensive Rat's tail couch grasslands noted by Beovich (1993; 1994) are assumed to be a post-flooding vegetation phase (SKM & Roberts, 2003). In response to large floods, these areas can also be classified as intermittently flooded temporary wetlands. Flood depth of the Open Areas is on average 0.4 m, but can be up to 1.0 m. Herbs, grasses and sedges make up a diverse range of wetland species post-flood. The openness of these areas is also conducive to foraging by birds and waterbirds when the areas are inundated.

Open Areas begin to flood at approximately 65,000 ML/day in the River Murray, but are not fully inundated until 115,000 ML/d. About 50% of the Lindsay-Wallpolla Island Open Areas are inundated at 75,000 ML/d (SKM & Roberts, 2003, p. 57). Flooding occurs from Mullaroo Creek, the Lindsay River and Wallpolla Creek.

5.6.6 Health of floodplain vegetation

Various factors affect the health of River red gum communities in the Lindsay-Wallpolla Islands system. Removal of fallen branches for firewood simplifies the habitat and removes shelter. Environmental weeds such as Lippia are a threat to biodiversity. Introduced plants such as Willows and Dodder affect functional characteristics. High levels in weir pools can result in waterlogging that leads to a decline in tree vigour or tree death. However, in terms of tree condition, SKM and Roberts (2003, p. 51) found only two instances of dieback, located in the extreme east of Lindsay Island. At all other sites inspected there was no evidence of decline in River red gum Forest or Woodland. Compared to trees, the understorey plants have short roots that are unlikely to be accessing groundwater, so it is likely that the composition of the understorey has changed over the past 80 or more years in response to the altered flooding regime (SKM & Roberts, 2003, p. 51).

Black box trees are drought hardy and tolerant of moderate saline conditions, responding in pulses to floods. Thus, reduced flood frequency is likely to reduce growth rate and canopy development in established trees (SKM & Roberts, 2003, p. 53). Recruitment may not be so severely affected, because germination and seedling establishment can also occur after heavy rain. SKM and Roberts (2003) speculated that grazing pressure was as important as flow regulation in limiting recruitment of Black box. The condition of Black box Woodland on Lindsay-Wallpolla Islands was considered by SKM and Roberts (2003, p. 53) to be generally good. However, some patches along the southern side of Mullaroo Creek and west of Lake Wallawalla are in poor condition, but the cause is unknown. The extensive cover of halophytic vegetation suggests that the surface soil is saline. An isolated groundwater salinity measurement of 40 dS/m has been recorded, which is still lower than the upper tolerable limit of Black box at Chowilla of 55 dS/m (Overton & Jolly, 2003). Grazing is likely to have affected the recruitment of dominant shrubby species (SKM & Roberts, 2003, p. 38). Unlike the trees, the ephemeral herbs and grasses in the Black box association have chronically reduced opportunities to germinate and complete their lifecycle (SKM & Roberts, 2003, p. 53).

Lignum Shrublands and Open Areas are threatened by reduced flooding frequency due to regulation, and mechanical damage from stock and vehicles. SKM and Roberts (2003, p. 56) predicted that the quality and extent of Lignum Shrublands would decline with time and that the occurrence of opportunistic terrestrial plants would increase. It is not known if this trend has already begun.

5.6.7 Wetlands

On Lindsay Island, most wetlands are small (75% being between 1 and 10 ha in area), shallow, and have a typical inundation duration of four to eight months. There is one large wetland-Lake Wallawalla. The small wetlands are typically freshwater meadows or shallow freshwater marshes (SKM & Roberts, 2003). Medium-sized wetlands are more common on Wallpolla Island, which also has a number of deep freshwater marshes. Wallpolla Island has more wetlands than Lindsay Island, with 184 compared to 124 respectively. The Wallpolla Island system also has a greater occurrence of lentic wetland forms (characterised by scroll swales and cut-off meanders).

Low-lying wetlands connected at 60,000 ML/d flow in the River Murray (many of which are impounded by the weir pools), have extensive littoral zones and fringes that support a variety of significant flora and fauna (SKM & Roberts, 2003, p. 43). However, they also act as refuges for carp. Stable water levels can cause notching of the bank and possibly lead to bank erosion.

Wetlands on high level flow paths and flood runners that are inundated by flows in the range 60,000 - 115,000 ML/d provide important habitat during their ponded phase, and they may also be important during the non-flooding phase. The contribution that these acts make to the condition of the asset warrants further investigation. Regulation of flow means that they are dry for longer than natural. Further investigation is required about the health of high-level wetlands flooded at discharges above 115,000 ML/d.

5.6.8 Waterbirds

Specific water requirements of wetland birds vary from species to species, but some general observations can be made. Cormorants and pelicans prefer to nest around wetland areas that are permanently inundated, but permanent and stable inundation of wetland systems is detrimental to successful breeding in the long term (Briggs & Thornton, 1999).

Kingsford (1998) has also noted that watering of wetland habitat has a direct influence on waterbird breeding success. Wetlands need to be inundated long enough for birds to breed and raise young. A sudden drop in water level too early in the breeding season often results in birds abandoning their nests. While this is an important factor in the short term, the key determinant of successful breeding in the long term is the frequency of wetland inundation (Leslie, 2001). The most critical stress factor in this case is the impact of a prolonged period of low-flows or within channel flows, whereby the dry period exceeds the reproductive lifespan of wetland bird species.

Herons, ibis, spoonbills, and bitterns require five to eight months inundation of wetlands to complete the breeding cycle. The ideal time for the flooding is winter/spring (Briggs & Thornton, 1999). Flooding that occurs later in the season will require seven to ten months of wetland inundation to achieve the same breeding outcome. Briggs et al. (1994) suggested that since the onset of river regulation, sustained wetland flooding for the time required to complete opportunities are limited.

In contrast to other bird types, ducks, geese and swans do not necessarily require a long period of wetland inundation for successful breeding and are less affected by the seasonal timing of flooding. However, Crome (1986) and Briggs et al. (1997) have shown that each of these species is characterised by an increase in breeding response after flooding occurs.

The diversity of waterbirds in the Lindsay-Wallpolla Islands system is high, and the abundance is high during flood recession periods.

5.6.9 Fish

The anabranches contain ideal habitat for Murray cod, including variable morphology of the creeks, the existence of fast and slow flowing hydraulic habitat, and snags along the river edges (Figure 5.4; Figure 5.37). A negative factor is the existence of a number of regulators within the Lindsay-Wallpolla Islands, which, in their present form, act as barriers for fish movement through the anabranch system (Figure 5.38).

There are three types of active channel, based on hydrology. Ponded creeks are connected to the weir pool (Lock 6 for Lindsay Island and Lock 9 for Wallpolla Island). The major creeks are ponded for about two-thirds of their length. Anabranches are also ponded. Intermittently-flowing creeks begin to flow as discharge in the River Murray increases up to about 60,000 ML/d. At higher flows the infrequently-flowing creeks begin to flow.

It appears that fish diversity in the Lindsay-Wallpolla Islands system is high (SKM & Roberts, 2003, p. 49). For a comprehensive study of fish species and their reproductive success within the Lindsay-Wallpolla Islands refer to Meredith et al. (2002).

 

5.6.10 Salinity

The significant interaction between groundwater and surface water of the Lindsay-Wallpolla Islands means that increasing surface water flows onto the islands presents a risk of increased mobilisation of salt towards the surface. Dudding and Evans (2001) investigated the relationship between surface water flooding and salt discharge from groundwater, establishing a positive linear relationship for both Islands. Estimates of the volume of salt likely to be mobilised into the River Murray system for various flow scenarios have not yet been made. Currently, Mullaroo Creek is used to divert water from the Lindsay River to ensure that the salinity level at Lindsay Point is low enough to allow irrigation.

For Wallpolla Island, flows into Wallpolla Creek are influenced by Lock 9. According to Sluiter and Parsons (2000), flows in Wallpolla Creek have not spilled onto the floodplain since 1996. Water quality data indicates that salinities are significantly higher in the isolated pools of Wallpolla Creek than river reaches downstream of Finnigans Creek that are connected to the weir pool. This is potentially indicative of saline groundwater interaction with the surface water tributaries of Wallpolla Island. The appearance of isolated patches of halophytic plants associated with dieback of Black box trees (see Section 5.2.6) also suggests the possibility of soil salinisation.

5.6.11 Human use values

Lindsay Island is part of the Murray Sunset National Park, and as such, supports only recreational activities such as camping and bushwalking. As Wallpolla Island is a State Forest it supports limited logging activities and recreational activities.

The traditional owners of the Lindsay-Wallpolla Islands are the Wergaia and Latji Latji people. Studies carried out in the Mallee Zone, and including the Lindsay and Wallpolla Islands, indicates a diverse range of Indigenous site types (SKM, 2003a). Most of these sites occur around floodplain areas or correspond to areas where flooding would have occurred under the natural flow regime. Edmonds (2003) suggested that occupation for the Lindsay-Wallpolla Islands by Indigenous peoples was semi-continuous and can be dated to 21,000 years BP.

Two archaeological assessments have been carried out for Wallpolla Island. As a result, 15 sites are registered on the Aboriginal Affairs Victoria Register (Edmonds, 2003). These sites include burial grounds, middens and isolated artefacts. A review of Indigenous values and archaeology of Lake Wallawalla can be found in SKM 2003a). The assessment by Kelton (1966) concluded that Lake Wallawalla is significant as a cultural landscape, comprising a broad range of Indigenous sites of varying levels of integrity.

Kelton (1996) surveyed Lindsay Island for Indigenous sites of significance. The survey, conducted along floodplain, riverbank and dune areas, found scarred trees to be the most numerous of the sites (42 located in Lindsay Island) followed by hearths (16), campsites (5) and shell middens (3). A number of buriak sites and isolated artefacts were also found. However, Kelton (1996) suggested that the number of significant sites the Lindsay Island area is likely to be much greater than the survey results indicate. Lindsay Island has a high cultural significance to the traditional owners.

5.6.12 Summary of knowledge of the condition of Lindsay-Wallpolla Islands

The high conservation value of the Lindsay-Wallpolla Islands has long been recognised, evidenced by their listing under the Directory of Important Wetlands. The area is rich in flora and fauna, including threatened species, and is highly significant for Indigenous people. Despite its high conservation values, surveys have demonstrated some environmental changes in the area.

In general, surveys show that River red gum has not significantly declined, with only two cases of tree dieback reported. Floodplain creeks that are connected to the weir pool are now wetter than under natural hydrological conditions, and this appears to offer an advantage to River red gums, expanding their range. Black box trees are drought hardy and salt tolerant and respond positively to flood pulses. Thus, reduced flood frequency is likely to reduce growth rate and canopy development in established trees. In some cases germination and seedling establishment can also occur after heavy rain. Grazing pressure is probably as important as flow regulation in limiting recruitment of Black box. Lignum Shrublands and Open Areas are threatened by reduced flooding frequency due to regulation, and mechanical damage from stock and vehicles. Lignum is drought tolerant, but extending the length of the dry period between floods is likely to reduce the quality of the habitat it provides for fauna. Regulation of flow means that they are dry for longer than natural, which is likely to reduce the opportunity for plant seeds and animals to propagate. The diversity of waterbirds in the Lindsay-Wallpolla Islands system is high, and the abundance is high during flood recession periods. The anabranches contain ideal habitat for Murray cod. A negative factor is the existence of a number of regulators within the Lindsay-Wallpolla Islands, which, in their present form, act as barriers to fish movement through the anabranch system. It appears that fish diversity in the Lindsay-Wallpolla Islands system is high.

The knowledge base for Lindsay-Wallpolla Islands is extensive and comprehensive, but there is scope to improve on this. The various data on the ecological components of the Islands (i.e., birds, fish, mammals, amphibians, macroinvertebrates, algae, vegetation) have not yet been compiled into a single database and summarised in terms of relevant and consistent measures of diversity, extent, abundance, and long-term trends.

Uncertainty can be reduced through improved knowledge of process, which can be achieved through hypothesis-driven investigations that will likely have specific data collection requirements.

5.7 Factors impacting on environmental values of Lindsay-Wallpolla Islands

5.7.1 Degree of change to flows due to river regulation

The climate of the Lindsay-Wallpolla system is semi-arid with an annual average rainfall of approximately 250 mm. October is the wettest month. Rainfall helps to maintain vegetation between events but flooding from the River Murray is critical for regeneration.

The Lindsay-Wallpolla Islands are downstream of:

• the Darling River confluence with the River Murray;

• Mildura (Lock 11);

• Wentworth Weir (Lock 10);

• major headworks storages on the River Murray (Hume and Dartmouth dams) and the Lower Darling River and the Menindee Lakes; and

• three of the largest point diversions from the River Murray (Mulwala Canal, Yarrawonga and Torrumbarry).

Flows in this section of the River Murray follow a seasonal pattern similar to that experienced prior to river regulation, with maximum flows in winter and spring, and minimum flows in autumn (Maheshwari et al., 1995) (Figures 5.16 and Figure 5.40).

Regulation has resulted in a reduction in the frequency and magnitude of small- to medium-sized floods as well as change in the overall volume delivered to the islands (Box 5.1). The effects of regulation in this section of the river can be summarised as causing:

• a reduction in flow volumes (both high and low flows);

• long periods of low regulated flows; and

• for larger flow events, reduced flood frequency, duration and size.

Box 5.1 - Statistics on the change of flow regime to the Lindsay-Wallpolla Islands. Source: unless stated, SKM and Roberts (2003), based on an analysis of the 109 year flow record.
Reduced flood frequency	Downstream of Wentworth Weir, under current conditions, the frequency of flood events with peaks greater than 10,000 ML/d has been reduced. Small flood events that are above regulated flow (peaks between 20,000 ML/d to 40,000 ML/d) now occur approximately 56 - 86 times per year rather than approximately 106 - 117 times per year. However, it is the frequency of events larger than this (classed as 'medium-sized' events) that have suffered the greatest impact under current conditions. This is due to the dampening effect of lock and weir operation. For these events (peaks between 50,000 ML/day to 100,000 ML/d) frequency has been at least halved under current conditions. For floods with peaks >100,000 ML/d the frequency is less than one third of the natural frequency. In contrast to the reduced frequency of floods with peaks >10,000 ML/d, it should be noted that the frequency of small in-channel events with a peak of <10,000 ML/d has doubled.
Reduced flood duration	For the Lindsay-Wallpolla Islands, the greatest change in event duration has been for flows up to 10,000 ML/d. The median duration of events below 10,000 ML/d has been reduced from 258 days to 46 days. Flows above 20,000 ML/d and below 115,000 ML/d suffered the greatest decrease in the duration of events as a result of river regulation. Larger flood events (greater than approximately 115,000 ML/d) have been relatively unaltered.
Reduced annual volume and changes to seasonality	For flows up to approximately 20,000 ML/d, the seasonal pattern has largely been retained, with the exception that more low flows occur in winter than could be expected under natural seasonal conditions. For higher events, up to 60,000 ML/d, seasonality has shifted, now occurring one to two months later in the season. For larger events (> 60,000 ML/d) the onset of flooding has been slightly delayed. At the South Australian border, current median flow is 39% of natural (Gippel et al., 2002). Mean flow is 46% of the natural volume (Maheshwari et al., 1993, p. 17).

Wallpolla Creek (including Mullaroo Creek) is influenced by flows from Lock 9 as well as inflows from the Darling and Murray at Lock 10. The weir pool at Lock 7 backs flow up the River Murray, which flows into Mullaroo Creek. Lock 9 is located just downstream of Frenchmans Creek, which is the inlet for Lake Victoria. It influences flow into Potterwalkagee Creek, an anabranch of the River Murray, located between Wallpolla and Lindsay Islands. Lock 8 is the primary regulator for flows into the Lindsay River.

A comparison of natural and current hydrological conditions at Lock 10 (Figure 5.39) shows that flows in spring and summer are considerably less than under natural conditions. At Lock 8 (downstream of Wallpolla Island) the effect is even greater (Figure 5.40). Winter flows have also been reduced.

Flow thresholds corresponding to flows at which particular landforms (and the habitat they support) are inundated were selected for more detailed analysis. Flows of less than 20,000 ML/d (refer to table) are regulated low flows that keep `active' channels flowing. Active channels include the Lindsay River, Wallpolla Creek and Mullaroo Creek. Flow spills out of the `active channels' and into the smaller anabranches at between 20,000 ML/d and 35,000 ML/d (Stage II flows). At flows greater than this, water spills onto the floodplain. The threshold for connectivity between the anabranches and floodplain is reached at a flow of approximately 65,000 ML/d (flows less than this are Stage III flows). High flows above this (Stage IV) result in progressive connectivity until complete inundation of the Islands are achieved at flows above 115,000 ML/d (Stage V flows).

Analysis of the occurrence of flow classes at Wentworth Weir (Lock 10) (Figure 5.41) shows that the number of large flood events (i.e., >65,000 ML/d) has been reduced over spring and summer. Also, the period of time that low flows are experienced has increased. Low flows that fill `active' channels are now the most common. A reduction in the number of medium-sized floods has also occurred (20,000 ML/d to 35,000 ML/d).

Analysis of the occurrence of flow classes at Lock 8 (Figure 5.42) shows a similar result to that for Lock 10 at Wentworth. At Lock 8 the peaks of flood events (between 35,000 ML/d - 65,000 ML/d for example) at Wentworth are attenuated by the influence of weir pools associated with Locks 9 and 8. For example, a flood event of 65,000

ML/d at Wentworth may only be as large as 35,000 ML/d downstream at Lock 8. The impact of the locks and weirs on river regulation means that Lindsay Island suffers from greater occurrence of low flow events than Wallpolla Island, although the difference is small. In any case, losses to Wallpolla Island will reduce the total water available to flow into Lindsay Island.

5.7.2 Changes in the frequency and duration of flow events

The flow thresholds for each of the active channels in the Lindsay-Wallpolla system are provided in Table 5.6. The thresholds indicate flow levels in the River Murray above which the channel capacity of each of the named anabranches is exceeded. When the threshold is exceeded, flow spills out into the floodplain. Thus, the flow thresholds are a useful indicator as to how often wetlands in the area are likely to be inundated. Table 5.6 indicates that while the flow thresholds for Upper Wallpolla Creek have not been altered, for Moorna Creek and Moorna Lagoon they have been significantly reduced, resulting in permanent inundation of the lagoon with low flows. The same is true for Mullaroo Creek, which has had its flow threshold reduced from 23,000 ML/d to 4,000 ML/d resulting in continuous low flow events under current conditions. SKM (2003b) analysed the impact of these changes on the hydrology of the individual wetlands. In general, regulation has resulted in a reduction in the frequency and duration of flow events in most channels (e.g., Upper Wallpolla Creek and Lindsay River), while others which have a lowered intake threshold experience a higher proportion of flows than under natural conditions (e.g., Mullaroo Creek at Lindsay Island and Dedmans, Moorna and Milky Creeks).

SKM and Roberts (2003) analysed the frequency and duration of flood events at Lock 10, and related this to the key vegetation communities. Flood frequency has declined for the entire flood discharge range (includes all vegetation types) (Figure 5.43). While duration has not been affected for floods over 90,000 ML/d, so that much of the Black box and the high-level wetlands are not affected in this respect (Figure 5.44).

Regional system hydrological thresholds as defined in Table 5.9 were used in a more detailed flow duration analysis to gain a better understanding of system-wide impacts. For Wallpolla Island, the River Murray flow at Lock 10 was used for the analysis. For the lower Wallpolla Island and Lindsay Island, Locks 9 and 7 were used. A comparison of current and natural flow conditions (Table 5.7) indicates that the duration of flows greater than 20,000 ML/d has changed from 62% of the time under natural conditions to 28% of the time under current conditions at upper Wallpolla Island, and a similar degree of change occurred for Lock 7, which influences flow through Mullaroo Creek in the Lindsay Island system. There has been a similar relative reduction in the duration of flows in the medium- and large-sized flood event classes (35,000 ML/d and 65,000 ML/d respectively).

Table 5.8 compares, for natural and current conditions, the number of events with duration greater than 14 days for Locks 10, 9 and 7. For flow events of <20,000 ML/d, the number of flow events has fallen to a lesser extent than the other flow thresholds. The higher the flood threshold, the greater is the relative reduction in flood frequency.

Table 5.9 compares, for natural and current conditions, the median duration of events for those events that exceed a 14-day duration. The median duration is a reasonable indicator of the length of seasonal of flooding. For the Lindsay-Wallpolla system, a similar pattern is evident for flow at Locks 10, 9 and 7. Under natural conditions, the median duration of the smaller flood events was around five months. Under current conditions, this is more likely to be three months duration or less. For floods that result in overbank flow (of 35,000 ML/d) the total time for floodplain inundation has been reduced by a similar magnitude. Apart from floods of 65,000 ML/d at Lock 10, high floods have changed little in duration (Table 5.9), which is consistent with the data in Figure 5.44.

5.7.3 Other factors

While flow regulation, and weirs have certainly altered the hydrology of the Lindsay-Wallpolla Islands area, SKM and Roberts (2003) could only infer that this had affected the health of the system. Trees showed only minor evidence of dieback, and fish and waterbirds were in good condition. River red gums have adapted to reduced flood frequency to some extent by relying more on groundwater. Indeed, some areas of River red gum have been advantaged by the availability of groundwater from the effects of the weir pools. SKM and Roberts (2003) speculated that grazing pressure was as important as flow regulation in limiting recruitment of Black box. The communities of shallower-rooted floodplain plants are likely to have changed in composition since regulation began affecting the hydrology around 80 years ago, but this cannot be demonstrated through a visual inspection.

Grazing is likely to have affected the recruitment of dominant shrubby species, and Lignum is threatened to some extent by mechanical damage from stock and vehicles. Removal of fallen branches for firewood degrades the habitat of the forest and woodland floor.

5.8 Opportunities to meet objectives for this site-Lindsay-Wallpolla Islands

5.8.1 Introduction

A number of recommendations have been made to restore flows to enhance streamflow habitat for the Lindsay-Wallpolla Islands (SKM & Roberts, 2003). These were developed at three spatial scales: whole-of-river operations; regional system (pertaining to local area inflows from the River Murray); and `within system' features to utilise existing flows (such as structures and levees located along the anabranches of the Lindsay-Wallpolla Islands).

5.8.2 Use of environmental water

Water from Lake Victoria could be used to ensure water spills onto the floodplain for the required duration. For floods less than 50,000 ML/day (as limited by Lake Victoria outlet constraints, but this is under review), water could be supplied from Lake Victoria through Rufus Creek to `top up' smaller floods for the western end of Lindsay Island system (Wallpolla Island is just upstream of Lake Victoria). This option could potentially be used to increase frequency or duration of medium-sized events that allow connectivity of the anabranches.

Increased flooding in the Lindsay-Wallpolla Islands will result in increased recharge, which in turn causes increased discharge of saline groundwater and hence salt load to the surface water system. For a given flood height, increasing flood frequency is more likely to have an adverse impact than increasing the flood duration (SKM & Roberts, 2003). Increasing the flood frequency is likely to raise groundwater levels, while increased flood duration could have the beneficial effect of creating a flushed zone of good quality groundwater (SKM & Roberts, 2003).

5.8.3 Structural and operational measures

Regional scale options

To a large extent, the flows in the Lindsay-Wallpolla Islands are regulated by the operation of Locks 6, 7, 8 and 9. It may be possible to change the ways the locks are operated to introduce more variability into the low flows. Weir pool `drawdown' offers potential for larger variations in water level (SKM, 2004f), but this has not been investigated experimentally at Lindsay-Wallpolla Islands.

Floodplain and channel specific options

SKM and Roberts (2003) investigated a number of options to increase flow variability in the active anabranch channels of the Lindsay-Wallpolla Islands. The study explored five types of options for increasing floodplain watering and flow variability and also suggested that fish passage be improved.

Existing weirs and regulators on channels can be used to manage the flow through a range of active channel habitats. For fixed crest weirs and earthen banks, the addition of a regulator will allow control over the flow regime, and thus give greater operational flexibility. Regulators on wetlands would have a similar goal of providing operational flexibility in order to exclude excessive flooding and influence flow timing and flood inundation. Levees could be used in conjunction with weirs to pond water, but the negative effect of interfering with the distribution of waters from natural floods was considered by SKM and Roberts (2003) to be sufficient reason not to pursue this option.

The study of SKM and Roberts (2003) advised against the pumping of water onto wetlands systems to simulate flooding. The main cause for concern was the high ongoing cost and intensive management required. Despite this warning, pumping is carried out in the Chowilla and Lindsay-Wallpolla as an emergency measure for isolated wetlands that have not been watered for a long time. Pumping could also potentially be used to water specific areas of the floodplain, such as Black box, which is located on high elevation areas and is flooded too infrequently under the regulated flow regime. However, such areas are not natural basins, so a levee would have to be constructed to pool the water. All options that involve changing the flow in active channels can potentially have complex effects on salt loads, and this would need assessment on a case-by-case basis. Changing flows so that currently dry channels receive regular flow will likely have a negative impact.

SKM and Roberts (2003) recommended further exploration of a number of specific options for management of water on Lindsay-Wallpolla Islands at the local-scale: modification of the existing fixed crest weir on Mullaroo Creek to include a regulator (with fish passage); allowing the flow regime to be manipulated independently of the River Murray; replacement of the current small diameter pipe at the western branch of the upper Lindsay River; construction of a weir at the western end of Wallpolla Creek to hold back the Lock 9 weir pool and create a lentic system with a more natural variable water level; construction of a regulator on Horseshoe Lagoon to control inflows from Finnigans Creek; building a regulator at the junction of Toupnein Creek to manage water levels in Websters Lagoon independently of Finnigans and Toupnein Creek; allowing reinstatement of a more natural wetting and drying cycle; build a regulator and weir at both ends of Moorna Creek in order to change the flow regime to be more ephemeral; and additional regulators for Finnigans, Moorna and Dedmans Creeks. More recently, assessment of the benefits, cost and feasibility of these regulator options has been undertaken.

The Living Murray Environmental Works and Measures Program is a seven year, $150 million program to deliver structural and operational works and measures to improve the health of the River Murray by making the best use of water currently available (MDBC, 2003b; MDBMC, 2004). A number of projects completed under past programs, and others already commenced under the Environmental Works and Measures Program, address flow management issues within the Lindsay-Wallpolla Islands. The project `Improved flow management of the Lindsay-Wallpolla system' (undertaken by Mallee Catchment Management Authority, Victoria) aims to enhance the floodplain ecosystem of the Lindsay and Wallpolla Islands, including the provision of a diversity of structural and physical habitats, an increase in the diversity and distribution of native fish, and an increase in the abundance and diversity of aquatic and floodplain vegetation (MDBMC, 2004, p. 37). The work includes development of a digital elevation model and hydraulic model of the floodplain and channels, and investigation, design and construction of recommended works, including: Mullaroo Creek, Upper Lindsay River; Lake Walla Walla; Horseshoe and Websters lagoons; and Wallpolla, Finnegans and Upper Lindsay creeks (MDBMC, 2004, p. 37).

5.8.4 Links between ecological objectives and management opportunities-Lindsay-Wallpolla Islands

The connection between the ecological condition of Lindsay-Wallpolla Islands and flows to the islands has been established. Other factors are known to affect ecological condition of the islands, but flow regime is of fundamental importance. Reduced frequency and duration of medium-sized flood events in spring, and increased length of the dry period between floods are the main hydrological factors implicated in observed environmental changes in the islands. Grazing pressure from native animals is another factor implicated in degradation of the vegetation. Regulators act as barriers to fish movement through the anabranch system. The proposed opportunities for managing flows in the Islands aim to reverse or partially reverse the effects of flow regulation.

The construction of new regulators and modification of existing structures (including modified operating rules) will increase flow variability in the active anabranch channels of the Lindsay-Wallpolla Islands, increased floodplain watering, more natural wetting and drying cycle for wetlands, and improved fish passage (fishways are to be provided). These measures will help maintain the currently high biodiversity values of Lindsay-Wallpolla Islands.

Next page » 6 Information base for the Murray Mouth, Coorong, and Lower Lakes

Previous page » 4 Information base for the Hattah Lakes

Foundation Report Home | MDBC Home | Copyright
Contents | Executive Summary | Appendix A | Glossary | References