Watching the tide roll away – contested interpretations of the nature of the Lower Lakes of the Murray Darling Basin

Additional keywords: ecological condition, estuaries, Murray Darling Basin, palaeolimnology, Ramsar wetlands.

The Murray–Darling Basin (MDB) is one of Australia’s largest drainage divisions, spanning more than one million square kilometres. It is also its most productive, supporting over 40% of the nation’s agricultural domestic product. This productivity has been underpinned by an extremely high level of water abstraction to drive intensive irrigation agriculture at great cost to water-dependent ecosystems (Kingsford 2000). Originally to sustain navigation, but ultimately to guarantee water supplies, the rivers of the MDB became highly regulated after the commissioning of many dams and weirs, mostly after 1922.

The initial efforts at irrigation were limited and the greatest volumes of abstraction were focussed in the upper reaches of the Murray and Goulburn Rivers (Mallen-Cooper and Zampatti 2018). Nevertheless, concern about the impact of even this low level of abstraction on the flows entering South Australia was sufficient to warrant a presentation and debate within the colonial South Australian Parliament as early as 1886. For example, the Honorable Jon Rankine is quoted as saying in the House of Assembly in 1887 that ‘Many people imagined that there would be nothing to fear from only flood waters being taken, but this was a great mistake. All the floodwaters were required to drive out the salt water so as to keep the lakes and a portion of the lower river fresh for a few months in the year’ (Sim and Muller 2004, p. 26). This dispute intensified and became an important distraction to the nation’s Federation signed in 1901. The Interstate Royal Commission on the River Murray (Davis et al. 1902) was to resolve this issue and, by 1915 the River Murray Commission was instituted to manage cross-state contests over water (Smith 1998).

The Coorong (~35 000 ha) and Lower Lakes, comprising Lake Alexandrina (65 000 ha) and Lake Albert (16 800 ha), lie at the mouth of the Basin, where the River Murray discharges into the Southern Ocean, and are a lagoon and estuarine complex (Fig. 1). Geomorphically, the Coorong estuary and the Lower Lakes are separated by the last interglacial shoreline, the Woakwine Range, but are connected hydrologically by a series of channels, of which the Goolwa Channel, the continuation of the River Murray, carried most of the river water to the sea. To preserve freshwater resources in Lake Alexandrina from the penetration of tidal waters entering through these channels, an array of barrages was commissioned and completed in 1940. These are situated near the mouth of the River Murray and acted to raise the level of the Lake and hold back incoming tides. They also slowed river flow, leading to the accumulation of salt-laden sediments within the lagoonal complex. This resulted in the gradual formation of Bird Island, a tidal delta south of Hindmarsh Island, and the northward migration and ultimate closure of the river mouth (Bourman et al. 2000). The high evaporation rate typical of the region heightens the risk of the Lower Lakes drying. The commissioning of the barrages therefore left river water as the main source of water to maintain lake levels.

Fig. 1.  Lake Alexandrina is a large terminal lake near the mouth of the River Murray, South Australia. Letters mark the locations of barrages.

In 1985 the Coorong and Lower lakes were listed under the Ramsar Convention on the basis of, inter alia, their significant fish and waterbird populations and cultural (indigenous) significance. At the time the Lower Lakes were described as being mostly fresh (Department of Environment and Heritage 2000). Consistent with this, a report entitled A Fresh History, funded by a regional government agency, defined Lake Alexandrina as being predominantly fresh (Sim and Muller 2004). This report was largely based on documentary and anecdotal evidence and focussed particularly on the observation that, in 1901, the lake had become salty for the first time. This was attributed to abstraction upstream for irrigation in the eastern States, particularly in neighbouring Victoria.

Actions to restore the ecosystems of the MDB were in full swing by the late 20th century, with several reviews and audits conducted (Kingsford 2000). Among the outcomes include the identification of the volumes required to return the system to ecological health variously calculated at between 3200 and 7600 GL (e.g. Jones et al. 2002; Young et al. 2011; Williams 2017). By 2006 water levels in Lake Alexandrina had fallen due to drought and abstraction upstream and the Commonwealth Government had notified the Ramsar Secretariat of likely change in ecological character under Article 3.2 (Pittock et al. 2010). The risk of acidification through the oxidation of sulfidic sediments drew proposals for engineering solutions to protect to freshwater status of the lower lakes (Souter et al. 2013), but also drew calls for 700 GL of flow specifically to maintain a substantially freshwater and fluctuating lake system (Kingsford et al. 2009 cited in Pittock et al. 2010). The decision, made in the MDB Plan of 2012, to return 2750 GL year⁻¹ to the river, through water buybacks, irrigation renewal and water transfer efficiencies, was made in no small part because it was effectively argued that Australia had an international commitment to restore Lake Alexandrina and retain it as a freshwater lake (Department of Environment and Heritage 2009) and to restore the Coorong lagoon by ensuring the river mouth remains open. Federal and State Governments agreed to allocate a further 450 GL where it could be demonstrated that this would accrue no socioeconomic hardship to communities in the Basin (Murray Darling Basin Authority 2012).

Today, the Basin-wide Environmental Watering Strategy (Murray Darling Basin Authority 2014), based on Lester et al. (2011), provides minimum expected outcomes for end-of-basin flows. These include:

• barrage flows of at least 2000 GL year⁻¹ on a three-year rolling average of 95% of the time and a two-year minimum of 600 GL year⁻¹;

• water levels in the lower lakes to be above sea-level and at +0.4 AHD (Australian height datum) 95% of the time;

• salinity in Lake Alexandrina to be less than 1000 μS cm⁻¹ 95% of the time, and less than 1500 μS cm⁻¹ all the time;

• the Murray River mouth to be open 90% of the time to an average depth of 1 m (Murray Darling Basin Authority 2014, pp. 23, 24).

These prescriptions essentially preclude the ingress of seawater from the Southern Ocean and oblige State and Commonwealth Governments to provide environmental flows of sufficient volume to the end of the MDB to maintain Lake Alexandrina in a predominantly freshwater state and to keep the river mouth open.

Two strands of evidence informed the decision to keep Lake Alexandrina fresh: (1) an interpretation based largely on historical and anecdotal evidence (Sim and Muller 2004) and (2) a new interpretation of palaeolimnological evidence by Fluin et al. (2009), which differs significantly from an earlier interpretation by Fluin (2002) and Fluin et al. (2007) that the lake was estuarine. In this paper I argue that the palaeolimnological evidence provides scientific evidence that may inform management decisions and that the best interpretation of that evidence is that Lake Alexandrina has an estuarine history, as presented by Fluin (2002) and Fluin et al. (2007), and now verified in Helfensdorfer et al. (2019). The crux of the matter, therefore is that the foundation for the claim in subsequent published literature that the Lower Lakes were always fresh relies on the findings of Fluin et al. (2007), which does not provide that interpretation. Further, the foundation for the claim for a freshwater history by various levels of government relies on Fluin et al. (2009), which neither addressed nor rebutted the findings of an estuarine history made in Fluin et al. (2007). The now generally accepted, but erroneous, position that Lake Alexandrina has been fresh for 7000 years provides an important foundation for the need for additional freshwater flows down the Murray River, as agreed in the MDB Plan.

Lakes and estuaries contain sediments that accumulate more or less continuously over time (Weckstrom et al. 2017). Buried within these sediments are chemical and biological remains that reflect the nature of the wetland at the time of sediment deposition. The collection of sediment cores, the subsampling of sediments and the identification of these remains allows for the condition of the wetland to be inferred. In estuaries this usually allows for a ~7000-year history to be outlined as this was the point in geological history that sea levels last stabilised and geomorphic evolution of present-day coastal wetlands commenced. Diatoms are particularly useful fossil bioindicators of estuarine condition (Taffs et al. 2017) owing to their abundance, diversity and close association with, and widespread calibration to, water chemistry (e.g. Gell 1997). Importantly also, owing to their preference for waters of particular salinities and ionic compositions, they can be used to diagnose the influence of salinity of marine origin (thalassic) compared with that from non-marine (athalassic) sources.

Barnett (1994) analysed sediment cores from Lake Alexandrina and concluded that the lake was estuarine from 7000 years ago, but despite being variable on account of climate variability, overall was fresher in the modern period as the barrages, commissioned in 1940, had restricted previous marine incursions. Bourman and Barnett (1995) observed that before barrage construction in the 1930s saline water was present sporadically far upstream and that the estuarine Lakes Albert and Alexandrina became fresh after the barrages were in place. Further, Bourman et al. (2000, p. 141) stated that ‘barrage construction on the beach facies of the last interglacial shoreline transformed the estuary into freshwater lakes with permanently raised water levels and reduced the tidal prism by 90%’. The observations for an estuarine history, and even the intrusion of saline water ‘far upstream’, are now vindicated by the hydrodynamic modelling of Helfensdorfer et al. (2019), which suggests that estuarine conditions extended 200 km north of Lake Alexandrina during the Holocene high sea-level conditions (7000–5500 years BP) and, even under lower sea levels, Lake Alexandrina remained estuarine throughout the 7000 years before the local hydrology being altered through barrage construction in the mid 20th century.

A core taken, but not analysed by Barnett (core LA 2), and another collected by Gell and Fluin in 1996 near the river entrance (core LA 1), formed the basis of analyses in the 2002 PhD thesis of Fluin (Fluin 2002). The evidence derived from six of the 30 cores collected from the Coorong by Gell and colleagues was reported to the South Australian Department of Water, Land and Biodiversity Conservation in Gell and Haynes (2005) and is detailed in a book chapter by Gell (2017). The interpretation of the interactions between the River Murray, the Lower Lakes and the Coorong were presented at the Past Global Changes ‘Salinity, Climate Change and Salinisation’ workshop in Mildura in September 2004 (Gell et al. 2007). This conference paper was subsequently published as Fluin et al. (2007) on which I led the interpretation of the records from the Coorong. The published records of fossilised diatoms in the two Lake Alexandrina cores reveal changes in key indicator taxa over the last 7000 years. Non-contiguous subsampling leaves out much of the record but the data reveal gradual changes, attributed to long-term changes in regional rainfall, until the commissioning of the barrages, as discussed below.

Critical to the interpretation of the fossil record is the understanding of the ecological preferences of the taxa preserved in the sediments. This is largely achieved through the collection of modern diatom specimens and the calibration of the relative abundance of these to measured water quality parameters. This was achieved for diatoms from inland salt lakes (Gell 1997) and was completed for some Australian estuaries (Haynes et al. 2011; Saunders 2011; Logan and Taffs 2013). As diatoms are ecologically conservative and largely cosmopolitan, interpretation of Australian fossil sequences can benefit from the ecological preferences identified from databases developed elsewhere across the world. In summary, the salinity preferences, for example, of most widespread species of diatoms, are well known internationally, and the information is readily available in a series of papers by a range of different authors.

It is clear from this international and Australian evidence that the taxa considered marine or estuarine in Fluin et al. (2007), Cyclotella striata, Paralia sulcata and Thalassiosira lacustris, are indeed reflective of marine conditions (see Appendix 1 for detailed review of the known ecology of these taxa). None were recorded from inland salt lakes (Gell 1997) and so their presence reflects saline or subsaline conditions influenced by waters of marine (thalassic) origin. Other key taxa in the LA 2 record were Staurosirella (syn. Fragilaria) pinnata and Pseudostaurosira (syn. Fragilaria) brevistriata. Both were recorded in inland lakes with weighted average salinity optima of 3.9 g L⁻¹ and 1.9 g L⁻¹ respectively. Gell and Haynes (2005) and Fluin et al. (2007) reported Staurosirella pinnata to be abundant in the upper sediments of both the north and south lagoons of the Coorong, which is, and has been in historic times, saline to hypersaline. It was also recorded to be abundant in lakes of 6.3 g L⁻¹ and above in Gell (1997). On the basis of this evidence S. pinnata appears to be highly salt tolerant (eurysaline); P. brevistriata, on the other hand, is regarded as an obligate freshwater taxon. While these diatoms are broadly tolerant, these inferred preferences are used to summarise the change in condition of Lake Alexandrina, as revealed from core LA 2, and presented in Fig. 2.

Fig. 2.  Summary diagram of the diatom stratigraphy of core LA 2 (based on Fluin et al. 2007). The ‘% marine’ curve = ∑% C. striata, P. sulcata and T. lacustris; ‘% eurysaline’ = S. pinnata; ‘% fresh’ = ∑ P. brevistriata, A. granulata; ‘% other’ = ∑all other taxa on Fig. 2) (Fluin et al. 2007). Note variable scales in species plots.

In reference to the high incidence of thalassic (marine) taxa in the early part (~5000–7000 years BP) of the Lake Alexandrina record, Fluin et al. (2007, p. 130) stated ‘The presence of Thalassiosira lacustris, Cyclotella striata and Paralia sulcata indicate marine influence at this time’ and ‘the change in diatom community [after 5000 years BP] is likely to represent a decrease in lake level and increased penetration of seawater, possibly associated with the variable, dry climate phase after the mid-Holocene wet phase’.

The return to regional wet conditions is reflected in the passage ‘The decline in Thalassiosira lacustris above 160 cm [~2200 years BP] marks a further increase in freshwater river input conditions, perhaps influenced by the increases in precipitation’ (Fluin et al. 2007, p. 130).

Acknowledging the prevalence of athalassic (non-marine) taxa, Fluin et al. (2007, p. 132) stated: ‘The Holocene diatom assemblages of Lake Alexandrina reflect relatively freshwater conditions with longstanding and major inputs from the River Murray. Marine water indicators were never dominant in Lake Alexandrina’.

The 2007 paper did, however, clearly enunciate the post-barrage change to freshwater conditions when it stated that: ‘The barrages completely separate both lakes from the Coorong, with infrequent fresh water flowing through the barrage gates. As a result, Lake Alexandrina is presently [my emphasis] a large, predominantly fresh water system with no salt water input’, and ‘The greatest change to the diatom flora is again near the surface, at 30 cm, mostly attributable to a strong increase in Pseudostaurosira brevistriata coinciding with the estimated time boundary for the onset of river regulation. Further this increase is associated with a small decrease in Staurosirella pinnata that may be attributable to the barrages controlling tidal flux to the Lake favouring Pseudostaurosira brevistriata, which has a lower salinity tolerance than Staurosirella pinnata’ (Fluin et al. 2007, p. 130).

In summary, Fluin et al. (2007) concluded that water column salinity in the large terminal Lake Alexandrina was only moderately influenced by tidal inflow of seawater, particularly over the past ~2000 years. It is now (i.e. today) largely fresh as a result of isolation by a series of barriers completed by 1940. Unequivocally, Fluin et al. (2007) stated that, before regulation, Lake Alexandrina was tidal, with the balance between marine and river influence attributable to the regional hydroclimate as revealed in the water balance records of the ‘rain gauge’ lakes of western Victoria. So, although it was more influenced by fresh water over the last 2000 years, Lake Alexandrina continued to be influenced by seawater and was therefore, by definition, an estuary (see Tagliapietra et al. 2009). The greatest change in the entire record, as revealed by the CONISS dendrogram, was after the commissioning of the barrages whereupon the diatom flora reflected unprecedented, freshwater conditions. Not unexpectedly, the interpretation of Fluin et al. (2007) is consistent with that presented in Fluin (2002).

Coincident with the compilation of the palaeolimnological research of Fluin and coworkers on Lake Alexandrina, the River Murray Catchment Water Management Board published the document ‘A Fresh History of the Lakes: Wellington to Murray Mouth, 1800s to 1935’ (Sim and Muller 2004, p. 1) that concluded that ‘Prior to European settlement, Lakes Alexandrina and Albert at the terminus of the River Murray were predominately [my emphasis] fresh …’. Further, this document (Sim and Muller 2004, p. 1) stated that ‘Contrary to what many believe today, saltwater intrusions into the Lake environment were not common until after 1900 when significant water resource development had occurred in the River Murray system’. Clearly, this conclusion is in stark contrast to that made from palaeolimnological data produced in Fluin (2002) and published in Fluin et al. (2007).

The interpretations of Sim and Muller (2004) are founded mostly on anecdotes, particularly those coming from the time of the Federation Drought at the turn of the 19th to 20th centuries, but more generally across the years ~1820–1935. As such, it provides a synthesis of the commentary within South Australia as to the changing nature of Lake Alexandrina and the lower reaches of the River Murray as perceived by local residents. Sim and Muller (2004, p. 1) noted that ‘Short-lived intrusions of saltwater would occur during periods of low flow down river resulting in a lowered level of water in the lakes. Even in times of these low flows, it would appear that only small areas of the Lakes were affected’ and that ‘Saline invasions were more common after 1900 and the development of irrigation works because reduced river flows could not hold back the sea’ (Sim and Muller 2004, p. 1).

The years around 1900 were characterised by one of the more significant droughts in documented history, and the reconstruction of hydroclimate since 1788 (Gergis et al. 2012) reveals it to have coincided with a substantial shift in the Pacific Decadal Oscillation relative to 230 years of variability. This was conceded thus: ‘Irrigation schemes began at the same time as a long lasting, widespread drought that further diminished the amount of water in the river system’ (Sim and Muller 2004, p. 1).

What Sim and Muller (2004) neglected to report were the findings of the 1902 Interstate Royal Commission that were contrary to their position. The Commissioners represented three States (Victoria, New South Wales and South Australia) and sought counsel from across the MDB, and not just from South Australia. Davis et al. (1902) reported many observations, including one which stated that ‘One effect of a deep entrance channel would be to increase the saltness of the lakes, which, after a strong north-west or westerly gale, are brackish; the salt water being forced up channels as far as Wellington’ (Davis et al. 1902, p. 33). Further, they reported the observations that ‘When the winds shift to the south-east it is again blown out of the lake, a greater quantity running out under these circumstances than during any river flood’ (Davis et al. 1902, pp. 33, 34) and that of the master of a trading boat who is quoted as saying ‘he had known the water of the lakes as salt in past years’ (Davis et al. 1902: p. 34). Ultimately, the Commissioners concluded that ‘Apart from verbal statements, the evidence of facts is against the hypothesis that there has been any increase in saltness in the Murray Lakes by reason of diversion of water from the river channel’ (Davis et al. 1902, p. 34).

Sim and Muller (2004) do not report on these contrary views expressed in Davis et al. (1902). Given that the 1902 Interstate Royal Commission would have been regarded, less parochially, as the authoritative document of the time, it is clear that Sim and Muller (2004) represents a small subset of the views available. For Davis et al. (1902) to come to a conclusion, on balance, that is so markedly different to that given in Sim and Muller (2004) suggests that the reporting in the latter is not reflective of the full range of views held at the time. We now see that the conclusions reached by Sim and Muller (2004) are not only in disagreement with the palaeolimnological evidence, but are also not consistent with the wide-ranging conclusions of the 1902 Interstate Royal Commission.

From 2010 a new report (Fluin et al. 2009) was posted on the South Australian Government website. Entitled ‘An Environmental History of the Lower Lakes and The Coorong’, this 22-page report on the palaeolimnology of Lake Alexandrina and the Coorong lagoon was produced by three of the five authors of the 2007 publication. Neither Gell nor Hancock, coauthors on the 2007 publication, were authors of the 2009 report.

Using the same diagrams and descriptions as Fluin et al. (2007), Fluin et al. (2009) concluded: ‘There is no evidence in the 7000 year record of substantial marine incursions into Lake Alexandrina’, yet they also stated that ‘There were substantial alterations to the diatom community in Lake Alexandrina following European settlement and particularly after barrage installation’. In contradiction to the evidence presented in Fluin et al. (2007), they asserted that: ‘Over the 7000 year record, there are minimal numbers (generally <10%) of estuarine diatoms’ and that ‘… estuarine conditions have essentially been absent from this section (LA 1) of the lake (<5%)’.

The interpretation of the LA 2 record in Fluin et al. (2009) can be found in the 2007 paper but with two significant changes (Table 1). Specifically, the words ‘marine influence’ (Fluin et al. 2007, p. 130) are altered to ‘minor marine influence’, and ‘increased penetration of seawater’ (Fluin et al. 2007, p. 130) are altered to ‘increased penetration of more brackish water’. Both alterations diminish the interpretation of a tidal influence on the ecological character of Lake Alexandrina mounted in Fluin et al. (2007). The second alteration creates confusion as the term brackish cannot be qualified, it meaning salty waters, usually the result of freshwater mixing with seawater (e.g. Bayly 1967). So, the use of the terms ‘minor’ and ‘brackish’ serve to diminish the role of the ocean as a source of lake water salinity.

Table 1.  Alterations to passages found in Fluin et al. (2007) and Fluin et al. (2009)

Fluin et al. (2009) does not present, nor review, the interpretation made in Fluin et al. (2007) and merely cites it as previous work. Second, it offers no new palaeolimnological evidence from Lake Alexandrina that qualifies the record from core LA 2, nor does it provide new knowledge of the preferences of the key species that would support a new interpretation. Fluin et al. (2009) does, however, provide supplementary evidence in the form of an additional record of core RS 1. Although they include it under the results for Lake Alexandrina, the core was actually taken downstream from Goolwa, ~28 km from the lake and only ~7 km from the river mouth. It was taken in the Goolwa Channel, a section of the River Murray that typically carried 70% of the total flow to the mouth. This core reflects a mix of thalassic (Cyclotella striata, Paralia sulcata and Thalassiosira lacustris) and freshwater (Pseudostaurosira brevistriata) taxa before weir construction, and an increase in freshwater species (Epithemia spp., Rhopalodia gibba) after. The authors conclude that these shifts represent a change from ‘marine/hypersaline waters to brackish post-barrage installation’, at one point, yet that ‘pre-barrage diatoms are estuarine – marine – but with an absence of typical marine diatoms’ in the summary. Therefore, the LS 1 record throws no more light onto the interpretation of LA 2 other that to reinforce the conclusions that the taxa considered as indicative of marine influence in Fluin (2002) and Fluin et al. (2007) were so, as they were also common in sediments closer to the ocean.

Third, in their summary of Lake Alexandrina Fluin et al. (2009) suggest that ‘in terms of salinity, it would seem that there was a continuum in Lake Alexandrina prior to barrage construction’. This represents the typical nature of estuaries where there is a gradient from seawater near the mouth to fresh inland. However, their map of the distribution of ‘estuarine’ and ‘freshwater’ states (their fig. 6) identifies clearly delineated zones and no ‘continuum’. Here, the freshwater zone lies north of a line that commences just seaward of core LA 2, a site which, elsewhere in the report, they concede was ‘estuarine’ and ‘slightly brackish’ before the barrages.

Fourth, the new interpretation has a political context. It is apparent that Fluin et al. (2009) recognise the management context of their account when they observe in the project background that the national media have reported that many people have justified opening the barrier between the systems on the basis of a past estuarine baseline. Fluin et al. (2009) appear to give support to those resisting this particular management approach when they conclude ‘some calls to allow sea water into Lake Alexandrina have been justified by an assumption that sea water would have penetrated into the Basin in the past. However, as is summarised in [their] fig. 6, this is not the case and the majority of the Lake has been fresh for its entire history’. This position, which is at odds with the conclusions reached by Fluin (2002) and Fluin et al. (2007), comes with no new evidence, other than that provided by core RS 1, which was taken from the Goolwa Channel, a site that was clearly estuarine. The sum of marine taxa presented in Fig. 2 here is derived from the Fluin et al. (2007) data and shows thalassic diatoms accounting for 5–35% of the total for ~7000 years, up until the commissioning of the barrages in 1940. As diatom productivity has been observed to be greater in the upper reaches of an estuary (Arndt et al. 2007), it is possible that these relative abundance values under-represent the influence of tidal waters on the salinity regime operating in Lake Alexandrina. Further, the remaining diatoms are mostly composed of valves of a euryhaline taxon, which suggests variable, saline conditions. By justifying the preclusion of the inflow of seawater as a means of managing lake acidification, which involves invoking an entirely fresh history for the lower lakes, the position posited by Fluin et al. (2009) lends weight to the calls for increased river flows as the management solution as proposed by Kingsford et al. (2009) and adopted in the environmental watering strategy (Murray Darling Basin Authority 2014). Further insight into the significance of the new interpretation may be gained from a quote from a local resident from the region cited in Gross et al. (2012, p. 59): ‘The incentive for returning large volumes of water as environmental flows is reduced in an ‘estuarine’ perspective of the lakes’. It is to these water resources management issues that I now turn.

By 2010 there were available two different interpretations from palaeolimnological evidence of the prebarrage state of Lake Alexandrina, one published in Hydrobiologia stating that it was an estuary with the relative influence of sea and river water mediated by climate (Fluin et al. 2007, based on Fluin 2002 and consistent with Barnett 1994), and the other on the South Australian Government website stating that most of the lake was fresh for its entire history (Fluin et al. 2009). This has paved the way for considerable confusion in the scientific literature and is revealing of the way that scientific conclusions are reported, even recycled, among users of that information.

A brief exploration of the Scopus website (www.scopus.com) on 10 September 2017 showed that Fluin et al. (2007) had been cited on 44 occasions. Of these, 17 were self cites; two were in outputs too obscure to retrieve, eight were on matters of diatom ecology or related to determinations of sedimentation rate and eight pertained specifically to the Coorong and not the two Lower Lakes. Of the remaining nine citations, seven clearly misrepresented the findings reported in Fluin et al. (2007). Of these, Brookes et al. (2015) note the lake to have been estuarine before the barrages, but erroneously stated that the palaeolimnological evidence (i.e. Fluin et al. 2007) revealed it to be ‘predominantly fresh’. As Table 2 reveals, most of these publications used the evidence from the palaeolimnological record to attest to a permanent freshwater history, even though the original 2007 paper, while suggesting a fresher condition under wetter climates, and a freshwater state after regulation in 1940, never stated that Lake Alexandrina was either ‘predominantly fresh’ or ‘a freshwater refuge’. The continuous presence of marine species and the dominance of the euryhaline Staurosirella pinnata indicate that, at the LA 2 site, Lake Alexandrina was always an estuary and could not be defined as predominantly fresh until the construction of the barrages.

Table 2.  Details of seven papers that have used Fluin et al. (2007) to argue for a fresh history of the Lower Lakes

As the debate pertaining to the restoration of the Murray–Darling Basin has intensified, a composition was submitted to The Conversation, an independent Australian online media outlet whereby scientists can air controversial topics for the wider public to read (Finlayson et al. 2017). The Conversation encourages feedback via a Comments section, and the article drew a particularly misguided post (https://theconversation.com/we-need-more-than-just-extra-water-to-save-the-murray-darling-basin-80188):

‘This has been studied using remains of diatoms, which neatly signal whether environments are saline, brackish or fresh, and they show unambiguously that for the last 7000 years Lake Alexandrina was a freshwater environment with only a few brief incursions of saltwater during extreme drought events (which over a 7000 time-span, you will have a few of). So yes, the lakes were indeed predominately fresh [my emphasis] before white man modification/water extraction. A dry read but one such paper detailing this evidence is: Fluin J, Gell P, Haynes D, Tibby J, Hancock, G. (2007). Paleolimnological evidence for the independent evolution of neighbouring terminal lakes, the Murray Darling Basin, Australia. Hydrobiologia 591: 117–134’.

Remarkably, the person posting the comment concluded that the Lower Lakes were predominantly fresh, which was not the actual conclusion of the paper he/she cited.

In summary, the discussion above shows that a range of writers, including scientists and members of the wider public, appear to have used Fluin et al. (2007) to lend authority to a position they themselves had surmised existed in the lower lakes over the past ~7000 years. While members of the wider public may find website reports easily accessible, most authors of scientific papers have ready access to peer-reviewed journals and Hydrobiologia is a well recognised means of disseminating such work. So, how did they get it so wrong? There are several possible explanations:

• They read Fluin et al. (2007) and concluded that the authors said, or intended to say, that Lake Alexandrina was ‘predominantly fresh’ thereby exhibiting ‘confirmation bias’ (sensu Berger and Johnston 2015);

• They assumed that the condition of the lake was ‘predominantly fresh’ and used Fluin et al. (2007) as an authority without checking it first to confirm that interpretation;

• They took the opinion of Sim and Muller (2004) that the lake was ‘predominantly fresh’ but sought, or were required under review, to cite peer-reviewed evidence to that effect, and failed to check the original 2007 article;

• They had read Fluin et al. (2009) which provided them with the conclusion that the lake was ‘predominantly fresh’, but cited Fluin et al. (2007) as the authority anyway.

Irrespective of the means by which ‘predominantly fresh’ became the accepted natural condition of the Lower Lakes, it seems that the science community has, perhaps inadvertently, presented a conclusion that is at odds with the original representation of the published palaeolimnological evidence, whether that be from Fluin et al. (2007) or even from previous interpretations from Barnett (1994), Bourman and Barnett (1995) or Bourman et al. (2000).

Despite the published paper by Fluin et al. (2007) clearly observing a variable, estuarine state before the barrages and the onset of fresh conditions after their commissioning, government agencies have used and cited, exclusively, the conclusions of the unpublished Fluin et al. (2009) in considerations on the natural state of Lake Alexandrina. For example, the South Australian Department of Environment and Heritage (2009, p. 4) stated: ‘The diatom record in lakebed sediments provides strong evidence that the Lower Lakes have been predominantly freshwater for the last 7000 years and that seawater ingressions, when they did occur, did not extend northwards of Point Sturt’, which is seaward of core site LA 2. This position was released in a fact sheet (Government of South Australia 2010) which stated that ‘Diatoms found in 7000 years of sediments indicate the majority of Lake Alexandrina was freshwater in all years’. These documents were consolidated in Lester et al. (2011), which identified the water requirements for Lake Alexandrina. None of these documents cited or addressed the observations of an estuarine history as described in Fluin et al. (2007), nor the observations of Barnett (1994); Bourman and Barnett (1995) and Bourman et al. (2000). The permanently fresh interpretation from Fluin et al. (2009) was then adopted in the environmental watering strategy (Murray Darling Basin Authority 2014) that set targets for Lake Alexandrina as being ‘full’ and ‘fresh’ (<1000 EC) 95% of the time, maintained by the release of at least 2000 GL year⁻¹ over the barrages. In conclusion, the new interpretation and its widespread adoption have acted to obfuscate the original conclusions of several earlier, published palaeolimnological studies and have appeared to strongly influence the socio-political decision-making process that has laid down an outcome of great ecological, social and economic consequence for water management in Australia.

It is odd that the grey-literature report (Fluin et al. 2009), which occurred only on a potentially transient webpage, was the preferred reference upon which to base these decisions, compared with a series of peer-reviewed papers in international journals with full archiving information and access. Now, this decision to set water quality targets, based on a new interpretation of a fresh history, and so to prioritise allocation of scarce environmental water to Lake Alexandrina, is challenged in the literature by Helfensdorfer et al. (2019). This new published paper not only affirms Lake Alexandrina to have been estuarine for the last 7000 years (i.e. prebarrages) but provides evidence for the tidal influence reaching more than 200 km inland before 5500 years BP. The interpretations found in the reports of Sim and Muller (2004) and Fluin et al. (2009) are at odds with the published, peer-reviewed observations of Barnett (1994), Bourman and Barnett (1995), Bourman et al. (2000), Fluin et al. (2007) and now Helfensdorfer et al. (2019). Despite this, it is the interpretations from the former alone that have been adopted by government to derive a highly significant water allocation decision.

In the acute national contest under the allocation of water under the MDB Plan of 2012, we have evidence for peer-reviewed, published science being supplanted by multiple reports in the grey literature (e.g. Sim and Muller 2004; Fluin et al. 2009) without the provision of new evidence, valid reinterpretation or the establishment of any shortcomings of the original scientific research. Although Fluin et al. (2009) included new data from core RS 1 in the Goolwa channel, some 28 km downstream, as part of new evidence of the salinity history of Lake Alexandrina, it provided a revised interpretation of Lake Alexandrina, and core LA 2 specifically, without providing any new evidence from Lake Alexandrina sensu stricto. This opened a contest of interpretations that continues on internet posts and blogs today, and likely influenced decisions made under the MDB Plan, the largest investment by government into the rehabilitation of a degraded river system in Australia.

Postmodern deconstruction of scientific evidence dispels the myth that science is ever entirely objective. Head (1995), for example, neatly portrayed the likely inherent biases in the condition of the author of the well known and highly influential book ‘The Future Eaters’ (Flannery 1994), the celebrated version of the coevolution of the Australian landscape and its people. Being human, it is nigh impossible for a scientist to remain absolutely objective, and society always ought to contextualise the author or speaker when absorbing the evidence put forward by scientists, particularly when a given management position is advocated. However, while a single position may be advocated, even based on strong epistemic values, for many practitioners it remains that conservation biologists may have licence, and even a duty, to advocate from non-epistemic evidence (Boon 2019) given the ‘crisis’ nature of the discipline. The wider community, however, might be justified in seeking to know when a scientist is drawing exclusively on empirical evidence, albeit from a personal context, and when he/she is advocating a position to counter the forces leading to ‘the current state and trajectory of the natural world’ (Boon 2019).

The warning for us in this story about water management and pre-European ecological states in the MDB is that it is critical that published interpretations be revised only in association with either (1) evidence of important errors of fact or interpretation in the original papers, or (2) the provision of new evidence that supplants earlier findings. Both are valid grounds for the taking of a new or revised position, and indeed the very progress of scientific understanding requires that this type of revision takes place. When it comes to presenting scientific evidence, however, scientists should realise that the political process has no conscience. Because of this reality, we must preserve the reputation of our science by reporting it objectively and making clear when we are advocating from non-epistemic evidence. Then, as members of a civil society, we may apply our right to participate in the challenging processes that make decisions that affect people, their places and the environment.

In the case of water allocations in the MDB, this case-study shows the values of maintaining the correct and transparent use of scientific information in a complex aquatic/terrestrial system in which the following conditions hold:

1. There is considerable competition for fresh water in the MDB and this had led to political conflict. Essentially, there is not enough water to meet irrigation demand, much less maintain a biodiverse and functional ecosystem.

2. This conflict led to a program of national regulation of water flows to maintain the river’s natural environment and provide for human use, the 2012 MDB Plan. This approach is, however, merely the most recent manifestation of a debate that has gone on for well over a century and has included several Royal Commissions.

3. The compromises reached in the 2012 MDB Plan do not satisfy all users and there is a national debate on just how much water should be allowed to flow to the mouth of the river for both environmental needs and use by South Australians.

4. The amount of data available to inform management decisions about water use and availability is very restricted and is subject to a certain degree of interpretation.

5. Considerably more research is needed before any resolution of the likely state of the MDB prior to European settlement, or even before the implementation of large-scale river diversions and irrigation schemes in the early parts of the 20th century, can be understood. Even then it is possible that any such state would differ today regardless of human uses of the river.

6. The consistent and transparent use of scientific information, subject to peer-review and published in the scientific literature, is critical to the resolution of these problems.

The author declares no conflict of interest.