Doyle’s Farm Produce Pty Ltd atf Claredale Family Trust v Murray-Darling... [2026] NSWSC 1036 | Case summary

Doyle’s Farm Produce Pty Ltd atf Claredale Family Trust v Murray-Darling... [2026] NSWSC 1036 | Case summary | Zoe

At a glance

Source facts

Court

Supreme Court of NSW

Decision date

2026-08-31

Before

Faulkner J

Catchwords

TORT – Negligence – duty of care – statutory authority – Murray-Darling Basin Authority established under the Water Act 2007 (Cth) – river system to be operated in the public interest – Authority’s functions include executing the agreement between States for sharing the water between the States – Authority required to operate in the face of conflicts – duty alleged to be owed to private irrigators – statutory framework does not erect or facilitate a sufficient relationship between the Authority and private irrigators – duty of care to some private irrigators would distort the focus of the statutory framework – alleged duty would be incongruous with the statutory framework TORT – Negligence – duty of care – pure economic loss – novel case – assumption of responsibility – salient features – reasonable foreseeability – actual foresight – identification of the risks of harm – control over the risks of harm and vulnerability – no specific reliance – no indeterminacy – equivalent duties of care would be owed to others – no assumption of responsibility – salient features do not warrant imposition of duty of care to avoid pure economic loss TORT – Negligence – alleged breach of duty in the conduct of river operations – whether failure to take precautions which a reasonable person in the Authority’s position would have taken – no breach of duty TORT – Negligence – alleged breach of duty in the conduct of commercial negotiations for access to privately-owned infrastructure – no breach of duty TORT – Negligence – causation of loss – factual causation – identification of counterfactual river operations – expert evidence TORT – Negligence – causation of loss – factual causation – determination of what the person who suffered harm would have done – admissibility of evidence – s 5D(3)(b) of the Civil Liability Act 2002 (NSW) – “statement made by the person suffering the harm” – farming business conducted through corporate trustee of family trust – family member involved in the management of business authorised to swear affidavit on behalf of corporate trustee – evidence inadmissible under s 5D(3)(b) TORT – Negligence – quantum of loss – price elasticity of demand for water – expert evidence – the price of hay PRIVATE INTERNATIONAL LAW – Choice of law – tort – Authority established to operate independently of the States – Authority’s place of work in the Australian Capital Territory – negligence alleged to arise from exercise of power to “give directions for the release of water” – directions variously received and acted upon in New South Wales and Victoria – negligence committed in the Australian Capital Territory – substantive issues in the case to be determined in accordance with the law of the Australian Capital Territory PRIVATE INTERNATIONAL LAW – Civil Law (Wrongs) Act 2002 (ACT) – Chapter 4 Negligence – does not apply of its own force to claim against the Authority – applies by virtue of s 64 of the Judiciary Act 1903 (Cth) PRIVATE INTERNATIONAL LAW – Choice of law – procedural issues governed by law of the forum – Court exercising federal jurisdiction – s 79 of the Judiciary Act 1903 (Cth) – whether s 5D(3)(b) of the Civil Liability Act 2002 (NSW) applicable to a case governed by Chapter 4 of the Civil Law (Wrongs) Act 2002 (ACT) – “the determination of factual causation” – whether s 5D(3)(b) applies to the determination of “factual causation” for the purposes of s 45(1)(a) of the Civil Law (Wrongs) Act 2002 (ACT) – s 5D(3)(b) applicable for the purposes of s 79 of the Judiciary Act 1903 (Cth) WORDS AND PHRASES – “statement made by the person suffering the harm” – “the determination of factual causation”

Source

Original judgment source is linked above.

Catchwords

TORT – Negligence – duty of care – statutory authority – Murray-Darling Basin Authority established under the Water Act 2007 (Cth) – river system to be operated in the public interest – Authority’s functions include executing the agreement between States for sharing the water between the States – Authority required to operate in the face of conflicts – duty alleged to be owed to private irrigators – statutory framework does not erect or facilitate a sufficient relationship between the Authority and private irrigators – duty of care to some private irrigators would distort the focus of the statutory framework – alleged duty would be incongruous with the statutory framework TORT – Negligence – duty of care – pure economic loss – novel case – assumption of responsibility – salient features – reasonable foreseeability – actual foresight – identification of the risks of harm – control over the risks of harm and vulnerability – no specific reliance – no indeterminacy – equivalent duties of care would be owed to others – no assumption of responsibility – salient features do not warrant imposition of duty of care to avoid pure economic loss TORT – Negligence – alleged breach of duty in the conduct of river operations – whether failure to take precautions which a reasonable person in the Authority’s position would have taken – no breach of duty TORT – Negligence – alleged breach of duty in the conduct of commercial negotiations for access to privately-owned infrastructure – no breach of duty TORT – Negligence – causation of loss – factual causation – identification of counterfactual river operations – expert evidence TORT – Negligence – causation of loss – factual causation – determination of what the person who suffered harm would have done – admissibility of evidence – s 5D(3)(b) of the Civil Liability Act 2002 (NSW) – “statement made by the person suffering the harm” – farming business conducted through corporate trustee of family trust – family member involved in the management of business authorised to swear affidavit on behalf of corporate trustee – evidence inadmissible under s 5D(3)(b) TORT – Negligence – quantum of loss – price elasticity of demand for water – expert evidence – the price of hay PRIVATE INTERNATIONAL LAW – Choice of law – tort – Authority established to operate independently of the States – Authority’s place of work in the Australian Capital Territory – negligence alleged to arise from exercise of power to “give directions for the release of water” – directions variously received and acted upon in New South Wales and Victoria – negligence committed in the Australian Capital Territory – substantive issues in the case to be determined in accordance with the law of the Australian Capital Territory PRIVATE INTERNATIONAL LAW – Civil Law (Wrongs) Act 2002 (ACT) – Chapter 4 Negligence – does not apply of its own force to claim against the Authority – applies by virtue of s 64 of the Judiciary Act 1903 (Cth) PRIVATE INTERNATIONAL LAW – Choice of law – procedural issues governed by law of the forum – Court exercising federal jurisdiction – s 79 of the Judiciary Act 1903 (Cth) – whether s 5D(3)(b) of the Civil Liability Act 2002 (NSW) applicable to a case governed by Chapter 4 of the Civil Law (Wrongs) Act 2002 (ACT) – “the determination of factual causation” – whether s 5D(3)(b) applies to the determination of “factual causation” for the purposes of s 45(1)(a) of the Civil Law (Wrongs) Act 2002 (ACT) – s 5D(3)(b) applicable for the purposes of s 79 of the Judiciary Act 1903 (Cth) WORDS AND PHRASES – “statement made by the person suffering the harm” – “the determination of factual causation”

Judgment

Source: NSW Caselaw (www.caselaw.nsw.gov.au). Converted to plain text by ZoeLaw. Not the official version.

Medium Neutral Citation: Doyle’s Farm Produce Pty Ltd atf Claredale Family Trust v Murray-Darling Basin Authority [2026] NSWSC 1036

Hearing dates: 11–28 August, 1–5, 10–19 September, 8–9 October, 10–14 November 2025

Date of orders: 31 August 2026

Decision date: 31 August 2026

Jurisdiction: Common Law

Before: Faulkner J

Decision: (1) The proceedings on the Fifth Further Amended Statement of Claim filed on 30 September 2025 brought by the Plaintiffs on behalf of the Group Members described in paragraphs 8 and 9 of the Fifth Further Amended Statement of Claim be dismissed.
(2) The Plaintiffs pay the Defendant’s costs.
(3) Any party which wants to apply for a different or further costs order has liberty to apply, such liberty to be exercised within 21 days by email to my Associate.

Catchwords: TORT – Negligence – duty of care – statutory authority – Murray-Darling Basin Authority established under the Water Act 2007 (Cth) – river system to be operated in the public interest – Authority’s functions include executing the agreement between States for sharing the water between the States – Authority required to operate in the face of conflicts – duty alleged to be owed to private irrigators – statutory framework does not erect or facilitate a sufficient relationship between the Authority and private irrigators – duty of care to some private irrigators would distort the focus of the statutory framework – alleged duty would be incongruous with the statutory framework
TORT – Negligence – duty of care – pure economic loss – novel case – assumption of responsibility – salient features – reasonable foreseeability – actual foresight – identification of the risks of harm – control over the risks of harm and vulnerability – no specific reliance – no indeterminacy – equivalent duties of care would be owed to others – no assumption of responsibility – salient features do not warrant imposition of duty of care to avoid pure economic loss
TORT – Negligence – alleged breach of duty in the conduct of river operations – whether failure to take precautions which a reasonable person in the Authority’s position would have taken – no breach of duty
TORT – Negligence – alleged breach of duty in the conduct of commercial negotiations for access to privately-owned infrastructure – no breach of duty TORT – Negligence – causation of loss – factual causation – identification of counterfactual river operations – expert evidence TORT – Negligence – causation of loss – factual causation – determination of what the person who suffered harm would have done – admissibility of evidence – s 5D(3)(b) of the Civil Liability Act 2002 (NSW) – “statement made by the person suffering the harm” – farming business conducted through corporate trustee of family trust – family member involved in the management of business authorised to swear affidavit on behalf of corporate trustee – evidence inadmissible under s 5D(3)(b) TORT – Negligence – quantum of loss – price elasticity of demand for water – expert evidence – the price of hay PRIVATE INTERNATIONAL LAW – Choice of law – tort – Authority established to operate independently of the States – Authority’s place of work in the Australian Capital Territory – negligence alleged to arise from exercise of power to “give directions for the release of water” – directions variously received and acted upon in New South Wales and Victoria – negligence committed in the Australian Capital Territory – substantive issues in the case to be determined in accordance with the law of the Australian Capital Territory PRIVATE INTERNATIONAL LAW – Civil Law (Wrongs) Act 2002 (ACT) – Chapter 4 Negligence – does not apply of its own force to claim against the Authority – applies by virtue of s 64 of the Judiciary Act 1903 (Cth) PRIVATE INTERNATIONAL LAW – Choice of law – procedural issues governed by law of the forum – Court exercising federal jurisdiction – s 79 of the Judiciary Act 1903 (Cth) – whether s 5D(3)(b) of the Civil Liability Act 2002 (NSW) applicable to a case governed by Chapter 4 of the Civil Law (Wrongs) Act 2002 (ACT) – “the determination of factual causation” – whether s 5D(3)(b) applies to the determination of “factual causation” for the purposes of s 45(1)(a) of the Civil Law (Wrongs) Act 2002 (ACT) – s 5D(3)(b) applicable for the purposes of s 79 of the Judiciary Act 1903 (Cth) WORDS AND PHRASES – “statement made by the person suffering the harm” – “the determination of factual causation”

Parties

Applicant/Plaintiff:

Doyle’s Farm Produce Pty Ltd atf Claredale Family Trust

Respondent/Defendant:

Murray-Darling Basin Authority

Legislation Cited (28)

Water Act 2007(Cth)

Civil Liability Act 2002(NSW)

Civil Law (Wrongs) Act 2002(ACT)

Judiciary Act 1903(Cth)

Cited:** Australian Capital Territory (Self-Government) Act 1988(Cth)

Civil Liability Act 2002(WA)

Civil Procedure Act 2005(NSW)

Interpretation Act 1987(NSW)

Legislation Act 2001(ACT)

National Parks and Wildlife Act 1974(NSW)

Productivity Commission Act 1998(Cth)

Public Governance, Performance and Accountability Act 2013(Cth)

Public Services Act 1999(Cth)

Uniform Civil Procedure Rules 2005(NSW)r 14.14(1).

Cases Cited: A.I. McLean Pty Ltd v Hayson [2008] NSWSC 927
Anchorage Capital Master Offshore Ltd v Sparkes (2023) 111 NSWLR 304; [2023] NSWCA 88
Armory v Delamirie (1722) 1 Stra 505; (1772) 93 ER 664
Australian Competition and Consumer Commission v Liquorland (Australia) Pty Ltd and Woolworths Limited [2006] FCA 826
Australian Securities and Investments Commission v Edensor Nominees Pty Ltd (2001) 204 CLR 559; [2001] HCA 1
AvWest Aircraft Pty Ltd v Clayton Utz [2018] WASC 167
Badenach v Calvert (2016) 257 CLR 440; [2016] HCA 18
Bevan v Coolahan (2019) 101 NSWLR 86; [2019] NSWCA 217
Brodie v Singleton Shire Council (2001) 206 CLR 512; [2001] HCA 29
Brookfield Multiplex Ltd v Owners Corporation Strata Plan 61288 (2014) 254 CLR 185; [2014] HCA 36
Bryan v Maloney (1995) 182 CLR 609; [1995] HCA 17
Byrne v Australian Airlines Ltd (1995) 185 CLR 410; [1995] HCA 24
Caltex Refineries (Qld) Pty Ltd v Stavar (2009) 75 NSWLR 649; [2009] NSWCA 258
Central Highlands Regional Council v Geju Pty Ltd [2018] 3 Qd R 550; [2018] QCA 38
Cessnock City Council v 123 259 932 Pty Ltd (2024) 281 CLR 39; [2024] HCA 17
Chamberlain v Ormsby [2005] NSWCA 454
Chappel v Hart (1998) 195 CLR 232; [1998] HCA 55
Commonwealth v Amann Aviation Pty Ltd (1991) 174 CLR 64; [1991] HCA 54
Commonwealth v Mewett (1997) 191 CLR 471; [1997] HCA 29
Cooper Brookes (Wollongong) Pty Ltd v Commissioner of Taxation (1981) 147 CLR 297; [1981] HCA 26
Crimmins v Stevedoring Industry Finance Committee (1999) 200 CLR 1; [1999] HCA 59
Cullen v New South Wales [2026] HCA 19
Dansar Pty Ltd v Byron Shire Council (2014) 89 NSWLR 1; [2014] NSWCA 364
Deputy Commissioner of Taxation (Cth) v Moorebank Pty Ltd (1988) 165 CLR 55; [1988] HCA 29
DFX v Coventry City Council [2021] EWHC 1382 (QB)
Dow Jones & Co Inc v Gutnick (2002) 210 CLR 575; [2002] HCA 56
Doyle’s Farm Produce Pty Ltd v Murray Darling Basin Authority [2025] NSWSC 1070
Doyle’s Farm Produce Pty Ltd v Murray Darling Basin Authority (No 2) (2021) 106 NSWLR 41; [2021] NSWCA 246
Duchess of Argyll v Beuselinck [1972] 2 Lloyd’s Rep. 172
East Metropolitan Health Service v Ellis (by his next friend Ellis) [2020] WASCA 147
Electricity Networks Corporation v Herridge Parties (2022) 276 CLR 271; [2022] HCA 37
Ellis v Wallsend District Hospital (1989) 17 NSWLR 553
Farriss v Axford [2023] NSWCA 255
Gates v City Mutual Life Assurance Society Ltd (1986) 160 CLR 1; [1986] HCA 3
Geju Pty Ltd v Central Highlands Regional Council [2016] QSC 279
Glavinic v Commonwealth [2023] ACTSC 361
Graham Barclay Oysters Pty Ltd v Ryan (2002) 211 CLR 540; [2002] HCA 54
Gunns Limited v State of Tasmania [2015] TASSC 52
Hazlett v Presnell (1982) 149 CLR 107; [1982] HCA 58
Hill v Van Erp (1997) 188 CLR 159; [1997] HCA 9
Hunt Leather Pty Ltd v Transport for NSW [2025] HCA 53
HXA v Surrey County Council; YXA v Wolverhampton City Council [2024] 3 All ER 341
Ibrahimi v Commonwealth of Australia (2018) 366 ALR 341; [2018] NSWCA 321
John Pfeiffer Pty Ltd v Rogerson (2000) 203 CLR 503; [2000] HCA 36
Jones v Dunkel (1959) 101 CLR 298; [1959] HCA 8
JP SPC 4 and another v Royal Bank of Scotland International Ltd [2022] 4 All ER 431
Lym International Pty Ltd v Marcolongo [2011] NSWCA 303
M & E M Hull Pty Limited v Thompson [2001] NSWCA 359
Maguire v Simpson (1977) 139 CLR 362; [1977] HCA 63
Mal Owen Consulting Pty Ltd v Ashcroft (2018) 97 NSWLR 1163; [2018] NSWCA 135
Malec v JC Hutton Pty Ltd (1990) 169 CLR 638; [1990] HCA 20
Mallonland Pty Ltd v Advanta Seeds Pty Ltd (2024) 418 ALR 639; [2024] HCA 25
McLennan Asset Services Pty Ltd v Macquarie Investments Ltd [2017] NSWSC 1811
Minister for the Environment v Sharma (2022) 291 FCR 311; [2022] FCAFC 35
Murphy v Overton Investments Pty Ltd (2004) 216 CLR 388; [2004] HCA 3
N v Poole Borough Council [2020] AC 780
Neal v Ambulance Service of New South Wales [2008] NSWCA 346
Pabai v Commonwealth of Australia (No 2) [2025] FCA 796
Paul v Cooke (2013) 85 NSWLR 167; [2013] NSWCA 311
Perez de la Sala v Perez de la Sala [2026] NSWCA 160
Perre v Apand Pty Ltd (1999) 198 CLR 180; [1999] HCA 36
PPK Willoughby Pty Ltd v Baird [2019] NSWSC 704
PPK Willoughby Pty Ltd v Baird [2021] NSWCA 312
Project Blue Sky Inc v Australian Broadcasting Authority (1998) 194 CLR 355; [1998] HCA 28
Provincial Insurance Australia Pty Ltd v Consolidated Wood Products Pty Ltd (1991) 25 NSWLR 541
Pyrenees Shire Council v Day (1998) 192 CLR 330; [1998] HCA 3
Queensland Bulk Water Supply Authority (t/as Seqwater) v Rodriguez & Sons Pty Ltd (2021) 393 ALR 162; [2021] NSWCA 206
Re Metal Storm Limited (in liq) (receivers and managers appointed) (No 2) [2019] NSWSC 1682
Richtoll Pty Ltd v WW Lawyers Pty Ltd (in liq) [2016] NSWCA 308
Rizeq v Western Australia (2017) 262 CLR 1; [2017] HCA 23
Roads and Traffic Authority of NSW v Dederer (2007) 234 CLR 330; [2007] HCA 42
Rodriguez & Sons Pty Ltd v Queensland Bulk Water Supply Authority trading as Seqwater (No 22) [2019] NSWSC 1657
Rogers v Whitaker (1992) 175 CLR 479; [1992] HCA 58
Roo Roofing Pty Ltd v Commonwealth of Australia [2019] VSC 331
Rosenberg v Percival (2001) 205 CLR 434; [2001] HCA 18
Samm Property Holdings Pty Ltd v Shaye Properties Pty Ltd (2017) 345 ALR 633; [2017] NSWCA 123
Sellars v Adelaide Petroleum NL; Poseidon Ltd v Adelaide Petroleum NL (1994) 179 CLR 332; [1994] HCA 4
Seltsam Pty Ltd v McGuiness (2000) 49 NSWLR 262; [2000] NSWCA 29
Sibley v Kais (1967) 118 CLR 424; [1967] HCA 43
Solomons v District Court of New South Wales (2002) 211 CLR 119; [2002] HCA 47
Southern Properties (WA) Pty Ltd v Executive Director of the Department of Conservation and Land Management (2012) 42 WAR 287; [2012] WASCA 79
Strong v Woolworths (2012) 246 CLR 182; [2012] HCA 5
Sullivan v Moody (2001) 207 CLR 562; [2001] HCA 59
Sutherland Shire Council v Heyman (1985) 157 CLR 424; [1985] HCA 41
Tabet v Gett (2010) 240 CLR 537; [2010] HCA 12
Tabet v Mansour [2007] NSWSC 36
Tame v New South Wales (2002) 211 CLR 317; [2002] HCA 35
Tekin v Stratford [2026] NSWCA 54
Timberland Property Holdings Pty Ltd v Bundy [2005] NSWCA 419
Tooheys Ltd v Commissioner of Stamp Duties (NSW) (1961) 105 CLR 602; [1961] HCA 35
Transport for NSW v Hunt Leather Pty Ltd (2024) 115 NSWLR 489; [2024] NSWCA 227
UGL Rail Pty Ltd v Wilkinson Murray Pty Ltd [2014] NSWSC 1959
University College Cork – National University of Ireland v Electricity Supply Board [2018] IECA 82
Vairy v Wyong Shire Council (2005) 223 CLR 422; [2005] HCA 62
Voli v Inglewood Shire Council (1963) 110 CLR 74; [1963] HCA 15
Voth v Manildra Flour Mills Pty Ltd (1990) 171 CLR 538; [1990] HCA 55
Ward v The Queen (1980) 142 CLR 308; [1980] HCA 11
Weber v Greater Hume Shire Council (2019) 100 NSWLR 1; [2019] NSWCA 74
Wingecarribee Shire Council v Lehman Brothers Australia Ltd (in liq) [2012] FCA 1028
Woods v Multi-Sport Holdings Pty Ltd (2002) 208 CLR 460; [2002] HCA 9
Wyong Shire Council v Shirt (1980) 146 CLR 40; [1980] HCA 12
X (Minors) v Bedfordshire County Council [1995] 2 AC 633

Like most rivers in Australia, the Murray River no longer flows freely from its numerous sources in southern New South Wales and northern Victoria down to the ocean off the coast of South Australia. Over the last century many dams, weirs and locks have progressively been constructed which are now operated as a coherent whole to exercise a measure of control over the water in the River.
Since 2008, the Murray-Darling Basin Authority has been the entity which has had primary responsibility for the exercise of that control. I will refer to the Authority as the “Defendant”. The Defendant was established by an enactment of the Commonwealth Parliament called the Water Act 2007 (Cth). I will refer to it as the “Commonwealth Statute”. To some extent the purpose and functions of the Defendant are to be determined by reference to the Commonwealth Statute but there is a larger picture. On 3 July 2008 the governments of New South Wales, Victoria, South Australia, Queensland, the Australian Capital Territory and the Commonwealth entered into an Agreement about how the water of the Murray River is to be shared amongst the States and Territory. As will be seen in more detail below, a central element of their agreement is the appointment of the Defendant as an independent expert to execute the provisions of the Agreement for the sharing of the water. The Agreement is an important part of the regulatory framework in which the Defendant operates and by reference to which its purpose and functions are to be understood.
Depending upon where they are located, the dams and weirs along the Murray River are generally owned and operated by different State water authorities. One of the powers which the States have agreed to give to the Defendant is the power to direct the release of water from some of the dams and weirs. By exercising the release power, the Defendant is able to move water about the river system so as to meet the entitlements of the States when and where they require.

Andrew Bishop was a Senior River Operator within the River Operations Team. His Affidavits contain the most comprehensive explanation of the events surrounding the river operation decisions which are now the subject of the negligence allegations, both in 2017/2018 and 2018/2019. The Affidavits are lengthy and address a large volume of complex material ranging from high-level concepts right down to some line items in spreadsheets. Mr Bishop was cross-examined across four days, some of which was combative.
The Plaintiffs submit that Mr Bishop was a poor witness. They give four reasons but none of them survives scrutiny. First, Mr Bishop accepted in cross-examination that he was an “advocate” for the Defendant’s decisions. The passage of the cross-examination where Mr Bishop accepted the word “advocate” occurred on day 7 of the hearing. [2] From the full passage it is clear that Mr Bishop’s idea of advocacy is that he was prepared to defend a decision which he believed to be correct but not in a way that was contrary to the true facts. Mr Bishop’s use of the word “advocate” was not intended to convey the partisan meaning which made all the lawyers sit up when they heard it. His use of the word does not warrant caution being given to his evidence.
Secondly, on four occasions Mr Bishop laughed when responding to a question in cross-examination (leaving aside the grimace when asked about the preferred style of address of his colleague, Dr Davis). In order to understand the reason for Mr Bishop’s laughter it is necessary to see how each such instance came about. The Plaintiffs came to Court with a theory that a release of an environmental flow of water from the Goulburn River (known as a “Goulburn Pulse”) had to be supplemented (“topped up”) by water stored in Lake Victoria to replace water diverted by irrigators (“consumptive use”) as the pulse flowed past their properties on its 15-day passage to the South Australian border. The Defendant’s modelling of the future storage level of Lake Victoria made an allowance for other consumptive use from the Murray River system but not for extra consumptive use referrable to a Goulburn Pulse. The result, as reasoned by the Plaintiffs, was that the Defendant’s modelling of future storage levels of Lake Victoria was too low in the 2017/2018 and 2018/2019 water years. This simple explanation will make more sense after other details are given below about the operation of the Murray River system. For current purposes it is sufficient to know that the Plaintiffs’ theory had a superficial efficacy but it is in fact wrong. The correct position is that diversions for consumptive use by irrigators occur (and at the same volume) whether or not a Goulburn Pulse is in motion. Mr Bishop explained the correct position as follows:

The Plaintiffs submit that the chuckling shows that Mr Bishop had an emotional involvement in the substratum of the case. They further submit that the chuckling bespeaks a person so intolerant of a contrary perspective that he found it difficult to listen to that perspective. The chuckling is said negatively to impact the weight and reliability which the Court can attribute to Mr Bishop’s evidence. I reject those submissions. In particular, Mr Bishop was not asked to listen to a “contrary perspective”. With prolonged repetition, he was asked to accept a proposition which he knew to be wrong and illogical.
The third reason given by the Plaintiffs for their submission that Mr Bishop was a poor witness (“unreliable”) is an instance when Mr Bishop gave emphatic evidence about the Defendant not calling on the Goulburn inter-valley trade (which is generally referred to as “IVT”) account early in the water year. The Plaintiffs submit that the evidence has been shown to be wrong because there was one occasion when Dr Davis authorised such a call in June 2018. I reject the submission. Mr Bishop’s emphatic evidence was directed to calling on the Goulburn IVT account for the operational purposes of addressing the risk of system shortfall in the face of the competing risk of spill. It is true that Dr Davis did authorise a call on the Goulburn IVT account in June 2018 (and also in May 2018) but that was for a different purpose: it was for a specific environmental programme called the “River Murray Increased Flows”. It was irrelevant to the point which Mr Bishop was making. When the June 2018 call was brought to Mr Bishop’s attention he said that “there would be some reason, potentially for this” and he was right. The details which are relevant to these matters are addressed below.
The fourth reason given by the Plaintiffs for why Mr Bishop was a poor witness is that he is said to have initially rejected a proposition about pre-empting a transfer of water from the Menindee Lakes to Lake Victoria, and then accepted the proposition after being shown an Affidavit by another member of the River Operations Team (Mr Waterworth). There is no substance in this reason. When the proposition was first put to Mr Bishop he said:

The Plaintiffs also point out that Mr Bishop gives extensive evidence about the decision by the Defendant to commence overbank transfers of operational water on 3 October 2017 even though Mr Bishop was on leave from 25 September until 16 October 2017. It is true that Mr Bishop was not involved in any decision-making process on the day when the Defendant gave the direction for the water to be released (3 October 2017 – see below) or in the week leading up to that day. Mr Bishop says in his First Affidavit that he was on leave and does not claim otherwise. It is clear from his Affidavit that his evidence about the overbank transfers in 2017 is based on the contemporaneous documentary record. With that clarity, the Plaintiffs’ criticism of Mr Bishop’s evidence on this topic is undermined by their artificial portrayal of the decision-making process within the River Operations Team. This is addressed in detail below. For current purposes, it is sufficient to say that up until he went on leave Mr Bishop personally conducted the Defendant’s regular modelling which was the integral basis for the River Operations Team’s decision to commence overbank transfers on 3 October 2017. Mr Bishop was well placed to explain the decision and to have an opinion about the reasonableness of the decision.
There is no doubt in my mind that Mr Bishop was an honest witness. He too sought to give answers which were as correct and accurate as possible. It was often a difficult task. Some of the cross-examination did not make allowance for the detail, complexity and nuance in the operation of the Murray River system. There were extended periods when the questioner and the witness were at cross-purposes, especially on the topic of the operation of the Defendant’s Annual Operating Plans (see further below). Sometimes Mr Bishop was presented with a binary choice between two incorrect options. I observed that Mr Bishop found the cross-examination process frustrating and tiring. He sometimes struggled to reduce complicated answers to short, precise and coherent statements. I am nonetheless satisfied that at all times Mr Bishop tried to be responsive to what he was asked. There was no sign that he tried to anticipate where a line of questioning might be going. He was not evasive. He made concessions when appropriate but would not be persuaded to make concessions when not. His evidence is to be accepted. My conclusion is reinforced by the fact that the Plaintiffs have not identified any instance of a deficiency in Mr Bishop’s evidence despite the length of the cross-examination.

Apart from the Defendant’s own officers, the only witness who gave evidence directed to the reasonableness of the decisions made by the Defendant was Dr Anthony Ladson. Dr Ladson was called by the Plaintiffs. Dr Ladson is a highly qualified and experienced hydrologist with expertise in the science of the movement, distribution and management of water. His vocational experience includes giving advice to the Defendant over a number of years, including in relation to studies about managing the Barmah Choke. In 2009/2010 Dr Ladson contributed to the Defendant’s Annual Operating Plan, although the evidence does not reveal the nature of his contribution (or indeed the format and content of the Annual Operating Plan in 2009/2010).
Notwithstanding that past work for the Defendant, Dr Ladson made clear that he has no relevant experience as a river operator. He accepted that there may be aspects of river operations which are not hydrological, such as environmental, political and social considerations. There are a number of disputes about whether Dr Ladson’s is qualified to comment on a particular matter and, if so, what weight ought to be given to his evidence. Given the terms of the Fifth Further Amended Statement of Claim (which is generally particularised by reference to sections in Dr Ladson’s reports), the Plaintiffs’ opening of the case and the sheer volume of Dr Ladson’s reports, it is remarkable how infrequently the Plaintiffs refer to Dr Ladson’s evidence in their closing submissions. It is therefore convenient to address any residual issues arising from Dr Ladson’s expertise as those issues arise below.
The Defendant makes a further submission about Dr Ladson which should be addressed at the outset. It arises from the circumstances in which Dr Ladson was retained by the solicitors for the Plaintiffs. The proceedings were commenced in 2019. There is no dispute that Dr Ladson was first retained in 2020. He said that his first task was to investigate the operation of the Murray River system and explain it to the lawyers. The fruit of that early investigation is apparent in the first 180 pages of Dr Ladson’s First Report dated 22 December 2023 which contain a detailed explanation of the Murray River system which is general in nature and is not tailored to any particular issue in the case. At some stage Dr Ladson was asked to prepare a report but it is not known when. Dr Ladson said variously that he did not initially anticipate that he would be asked to prepare a report for the Court and that he never thought that he would be doing anything other than preparing an independent report. Dr Ladson downloaded a copy of the Expert Witness Code of Conduct in October 2020. In a covering Affidavit affirmed on 22 December 2023 Dr Ladson stated that he has been provided with and read a copy of the Expert Witness Code of Conduct, that he understands it and that he has complied with it in preparing his First Report.
The Defendant submits that there is evidence that Dr Ladson was first retained in early 2020 and that he performed extensive investigation work before he downloaded the Expert Witness Code of Conduct and before he was instructed to prepare an expert report for the purposes of being an independent witness. On the other hand the Plaintiffs submit that the evidence does not demonstrate that Dr Ladson was retained before October 2020 and that he was retained from the outset as an independent expert witness.
There is a lack of clarity in the evidence about what actually happened which seems surprising in the modern world of documented communications. The evidence which has been put before the Court suggests that the facts on the ground did not reflect a formality of instructions. Dr Ladson described his initial role as being a “consultant”. That description ought to be accepted whether or not the Plaintiffs’ solicitors actually used that word at that time. Despite the propositions agreed to in re-examination, the clear tenor of Dr Ladson’s evidence in cross-examination was that there was a point when his role transitioned from being a consultant to the Plaintiffs’ lawyers into being an independent expert witness. That point is probably not susceptible to identification by a precise date. It may not have been a point so much as a process. It is not possible to say how much investigation work and report writing had occurred when the transition was complete.
It is, however, not necessary to make any findings about these details. Whilst the Defendant drew attention to the issue it did not make any submissions as to its impact on the admissibility or reliability of Dr Ladson’s evidence. His three reports were adduced into evidence without objection (other than two isolated paragraphs), as was Dr Ladson’s contribution to a Joint Report. The absence of objection may be explained by the fact that some of the evidence about Dr Ladson’s retainer emerged in cross-examination. Even so, the Defendant does not take the point any further in its closing submissions.
I observed Dr Ladson to be an impressive witness within the area of his expertise. In cross-examination he was diligent to observe the limits on his expertise. He made concessions where appropriate. He was constructive and accommodating throughout what the Plaintiffs fairly describe as lengthy cross-examination. Despite any handicap from the circumstances in which he was first retained, I have no reason to think that Dr Ladson did not successfully make the transition to an expert witness subject to the Expert Witness Code of Conduct.
One further matter warrants mention. Dr Ladson’s understanding of the legal significance of hindsight appears to have developed as his involvement with this case has progressed. Some of Dr Ladson’s earlier analysis of what could have been done differently has been informed by hindsight and is not therefore relevant to any issue in the case. Caution is therefore necessary when considering any conclusory views expressed by Dr Ladson, to the extent that they are still relied upon by the Plaintiffs.

(a) Furthest upstream is Dartmouth Dam which has an active storage capacity of 3,785 GL, which means that 3,785 GL can be stored in the Reservoir on top of the water which lies too low to be released (which is called “dead storage”). Dartmouth Reservoir is narrow and deep. Those features, combined with its relatively high elevation, mean that the evaporation loss of water from Dartmouth Reservoir is relatively low. (b) Next downstream is Hume Dam which has an active storage capacity of 2,982 GL. It takes approximately 25-26 days for water to flow from Hume Dam to the South Australian border. Hume Reservoir is also relatively narrow, deep and elevated. It has a relatively low evaporation rate, although not as low as Dartmouth Reservoir. (c) The Menindee Lakes are not located on the Murray River but on the Darling River north of the confluence. Altogether the Menindee Lakes have an active storage capacity of 1,1251 GL. It takes approximately 14 days for water to flow from Weir 32 (the relevant outlet from the Lakes) to the South Australian border. The Menindee Lakes are broad and shallow. Located in western New South Wales they have a very high evaporation rate. There is evidence that in October 2017 water was being lost to evaporation at the rate of 1.5 GL per day. The Menindee Lakes are unique in the sense that the States have agreed that the water stored in them will not always be available for use in the operation of the Murray River system and, when the water is available, Victoria will generally be entitled to an equal share even though the Lakes are located entirely in New South Wales. (d) Lake Victoria is the last storage before the South Australian border. It has a total storage capacity of 677 GL of which 577 GL is active storage. It takes about a day for water to flow from Lake Victoria to the South Australian border. The evaporation rate at Lake Victoria is lower than for the Menindee Lakes but higher than for Hume Reservoir. Lake Victoria is subject to some regulations for the purposes of preserving its cultural and historic significance to the local Indigenous communities which means that there are restrictions on the quantity of water which may be stored in the Lake at different times of the year.

The narrowest reach of the Murray River is called the “Barmah Choke”. It is located in the Barmah-Millewa Forest which is about a third of the way from Hume Dam to the South Australian border. Yarrawonga Weir is located just above the Forest. Yarrawonga Weir is the last river infrastructure by which the flow of water into the Barmah Choke can be regulated.
For operational purposes the two main tributaries into the Murray River are the Goulburn River (from Victoria) and the Murrumbidgee River (from New South Wales). They converge with the Murray River below the Barmah Choke but above the confluence with the Darling River. It takes approximately 16 days for water to flow from the confluence of the Goulburn River to the South Australian border.
As explained in more detail below, the States have agreed that South Australia is entitled to receive a specified quantity of water each month of the year. For example, South Australia is entitled to receive 159 GL each January. In broad terms, any excess water which flows across the border is for the benefit of South Australia and at the expense of New South Wales and Victoria which are generally entitled to an equal share of the residue. The significance of that arrangement is that Lake Victoria is core to the operation of the Murray River system. It performs the essential functions of providing a store of water from which to meet South Australia’s upcoming monthly entitlements and also a reservoir into which excess approaching water may be diverted and stored should that accord with the sharing arrangement in the Agreement. The extent to which Lake Victoria is able to perform those functions at any particular time depends on it having enough stored water or, as the case may be, not too much.
Especially in summer, there are periods of high demand for water below the Barmah Choke, including the demand for water to meet South Australia’s monthly entitlement (which is highest in the summer months). In addition to water stored in Lake Victoria, the demand may be met by releasing water from the Menindee Lakes (if available for use in the operation of the Murray River system) or calling for water from the Goulburn River and/or the Murrumbidgee River in accordance with the separate systems for regulating those rivers.
Otherwise, water must be transferred from above the Choke (i.e. from Hume Dam) by a bulk transfer of water. The water transferred in this way is called “operational water”. There are three ways to transfer operational water from above to below the Barmah Choke. First, it may be transferred within the channel capacity of the Murray River. Some of the documents refer to this as “in-channel transfer”. The extent to which this is possible depends on how much water needs to be transferred and the constraint of the Choke. Secondly, it may be transferred at a rate which is above the channel capacity of the Choke. This is overbank transfer which is referred to above.
The third way is to use the privately owned MIL infrastructure to bypass the Choke. Use of the MIL infrastructure depends on MIL’s agreement which was not forthcoming for the 2017/2018 water year and for some of the 2018/2019 water year. The circumstances in which the Defendant did not have access to the MIL infrastructure and the steps taken by the Defendant to get access is the subject of a separate complaint by the Plaintiffs.

Before turning to the parties’ submissions on choice of law it is necessary to set out some detail about the physical location and operation of the Murray River system. Apart from the workplace of the Defendant’s River Operations Team, about which the Defendant itself expresses a faint reservation, none of the following is controversial.
The Defendant does not dispute that its headquarters is in Canberra in the Australian Capital Territory. When they filed the Statement of Claim, the Plaintiffs were required to give an address for the Defendant which they did by specifying the address of its headquarters in the Central Business District of Canberra. All the officers of the Defendant who subsequently made Affidavits in these proceedings, including the Chief Executive who verified the Defence, gave the headquarters address as the address of the deponent. The witnesses evidently conformed to the common practice in commercial cases of not giving their private residential addresses but their work addresses instead. The Plaintiffs have identified a number of references in the cross-examination of the Defendant’s officers where they referred to the Australian Capital Territory as their place of work. The inference to be drawn is that Canberra is (and was at the relevant time) the place of work of the River Operations Team. It is (and was) also the place of work of Mr Reynolds whose conduct is the primary focus of the separate complaint about the negotiations for access to the MIL infrastructure in 2017 and 2018.
The Murray River is entirely within New South Wales. [15] There can be disputes about precisely where the southern bank of the river stops and Victoria begins, [16] but that is not an issue in this case. There are a number of dams, weirs and other features of the Murray River which are relevant.
The relevant dam which is situated furthest upstream is Dartmouth Dam which is located on a tributary of the Murray River called the Mitta Mitta River. Dartmouth Dam is in Victoria. Pursuant to cll 52, 56 and Schedule A of the Agreement in Schedule 1 to the Water Act 2007 (Cth), Dartmouth Dam is operated by Victoria. In practice it is operated by Goulburn Murray Water. GMW is a Victorian statutory corporation established under the Water Act 1989.
Downstream from Dartmouth Dam is Hume Dam on the Murray River itself. Hume Dam is located in New South Wales. Behind the dam sits Hume Reservoir which is bisected by the border between New South Wales and Victoria. Pursuant to the Agreement, Hume Dam is operated jointly by New South Wales and Victoria. Directions for releases of water from Hume Dam are given by the Defendant to a duty officer from WaterNSW which is a New South Wales government agency. The duty officer passes the direction to an officer of the private operator of the local hydro-electric power station. The power station is located in New South Wales. Water released from Hume Dam generally flows through the power station and into the Murray River. Hume Dam is operated in harmony with Dartmouth Dam such that a release of water from Hume Dam may sometimes result in a release from Dartmouth Dam.
Downstream from Hume Dam is Yarrawonga Weir. The Weir is located mostly in New South Wales but the southern-most end protrudes into Victoria. Behind the Weir sits Lake Mulwala which is bisected by the border between New South Wales and Victoria. Pursuant to the Agreement, Yarrawonga Weir is operated by Victoria. Directions for releases are given by the Defendant to a duty officer from GMW who is located at the GMW office on the Victorian bank of the Murray River. The directions are then passed to the private operator of the local hydro-electric power station which is located in Victoria. Water which is released at Yarrawonga Weir generally flows through the power station up to 13,000 ML/day and then returns to the Murray River. Above 13,000 ML/day the released water flows over the Weir (which is mostly in New South Wales).
Water leaves the Murray River and flows into the MIL infrastructure via the Mulwala Canal which commences on the New South Wales side of Lake Mulwala. The MIL infrastructure is in New South Wales. It is operated by MIL which, as already observed, is a private irrigation company.
Downstream from Yarrawonga Weir is the Barmah-Millewa Forest. It is bisected by the border between New South Wales and Victoria. More precisely, the Millewa Forest is on the New South Wales bank of the Murray River and the Barmah Forest is on the Victorian bank. There are regulators on either side of the border through which water can be released into the Forest(s) (see below). The Barmah Choke is located within the Forest and is part of the Murray River.
The Goulburn River flows into the Murray River downstream of the Barmah Choke. The Goulburn River is in Victoria. The flow from the Goulburn River is primarily regulated by Eildon Weir which is in Victoria and which is operated by GMW.
The Darling River flows from New South Wales into the Murray River at Wentworth. Upstream from Wentworth, the Darling River is joined by the Menindee Lakes which are in New South Wales. The Lakes are operated by New South Wales, both when the Defendant has access to the stored water and when it does not.
Downstream from Wentworth is Lake Victoria which is in New South Wales. It is operated by South Australia.

The Defendant submits that there is no single rule to determine the location of a tort and cites the tenth edition of Nygh*’*s Conflict of Laws in Australia for the proposition that the tendency in the cases is to stress the place to which the activity of the defendant was directed, rather than the place where the activity complained of originated. The Defendant refers to a number of cases of negligent misstatement where the Court has found that the statement was in substance made at the place to which it was directed and where it was intended to be acted upon. The Plaintiffs rely on Voth v Manildra Flour Mills as authority for that approach.
Where the tort arises from an omission of the defendant, the Defendant submits that the location of the tort is the place where the omission assumed significance.
Even if the Defendant made its decisions in the Australian Capital Territory and from there communicated (or omitted to communicate) the directions which are alleged to constitute the breaches of duty, the Defendant submits that the directions were intended to be acted upon elsewhere. Although much of the analysis of the Defendant’s operational decisions is addressed to releases from Yarrawonga Weir, the Defendant’s choice of law submissions focus on releases from Hume Dam because releases at Yarrawonga Weir are said to be an intermediate part of bulk transfers of operational water from Hume Reservoir to Lake Victoria. The Defendant submits that the operational decisions which are said to be negligent were directed to releases from Hume Dam in the 2017/2018 water year and that it was at Hume Dam that the Defendant’s failure to direct earlier and greater releases in the 2018/2019 water year assumed significance. The Defendant submits that the releases from Hume Dam occurred in New South Wales. It further points out that the Menindee Lakes and the MIL infrastructure are in New South Wales. Whilst it accepts that the Goulburn River is in Victoria, it submits that the Defendant’s failures to call on the Goulburn IVT account assumed significance when water from the Goulburn River failed to flow into the Murray River in New South Wales.
The Defendant submits that the fact that the Defendant’s River Operations Team is located in the Australian Capital Territory is essentially fortuitous.
The Defendant ultimately makes alternative contentions about the location of its alleged torts, namely New South Wales or Victoria. The Plaintiffs submit that the need to state alternatives demonstrates variability in the Defendant’s reasoning which is a flaw.

This is a case where the quality of the Defendant’s conduct is an integral part of the tort upon which the Plaintiffs sue. It is the failures to take certain precautions which gives the Defendant’s conduct the negligent quality which is central to the Plaintiffs’ case. The failures are addressed in more detail below but it is instructive to contextualise and summarise them here for the purposes of ascertaining where the Defendant’s conduct occurred.
Both parties agree that the heart of the Plaintiffs’ case is the Defendant’s exercise of the release power under cl 98 of the Agreement in Schedule 1 to the Water Act 2007. Clause 98 permits the Defendant to “give directions” for the release of water from certain storages on the Murray River. In passive language, cl 98 further provides that water must be released “in accordance with any such directions”. The Defendant’s power to give directions is to be understood in the context of the other provisions in the Agreement which address the circumstances in which the power may be exercised, including the requirement not to exercise the release power in a manner that has the potential to have a material effect on State water entitlements (cl 30(1)), the requirement to exercise the release power in accordance with the Objectives and Outcomes Document (cl 30(2)) and the requirement to consult the Basin Officials Committee in certain circumstances (cl 33).
In the 2017/2018 water year the Defendant gave directions on 3 October 2017 and on the following days. As finally articulated, the Plaintiffs case is focussed not so much on the ultimate giving of the directions but on the anterior decision to give the directions. Either way, the Defendant’s conduct is said to have been negligent because of seven specified failures which are said generally to have occurred before the directions were given. They include the failure to conduct modelling using the Fortran Assessment Model, the failure to document the decision, the failure to consult with the Water Liaison Working Group (also known as “WLWG”), the failure correctly to interpret the Bureau of Meteorology’s climate outlook and the failure appropriately to use the Daily Operations Spreadsheet. The Plaintiffs also rely upon the failure to prioritise releases from the Goulburn IVT account and failures in directing releases from the Menindee Lakes.
In the 2018/2019 water year the Plaintiffs’ case is focussed on the failure of the Defendant to decide in June 2018 to depart from the 2018/2019 Annual Operating Plan by making earlier and greater releases from Hume Reservoir/Yarrawonga Weir. They also rely on the failure to decide to call on the Goulburn IVT account in June, July and August 2018.
The allegation of negligence arising from the negotiations about access to the MIL infrastructure is directed to decisions made by the Defendant over an extended period whilst the negotiations were taking place.
The inference to be drawn is that the Defendant’s conduct which gives rise to these allegations occurred in the Australian Capital Territory. The workplace of the River Operations Team and Mr Reynolds was in the Australian Capital Territory which in the ordinary course is where the conduct would have occurred. The relevant time period is 2017 and 2018 which was before the disruption to work practices more recently seen in the wake of the Covid-19 pandemic. There is no suggestion that the Defendant’s officers worked from home, or that their homes were across the border in Queanbeyan or Collector or some other New South Wales town. The Defendant does not undermine the inference that the conduct occurred at the Canberra workplace by identifying two isolated occasions when officers of the Defendant attended meetings which were held in regional centres, nor by observing that some of the duties of the River Operations Team required river operators occasionally to visit other locations. Nor is it correct for the Defendant to submit that the place where the Defendant’s conduct occurred is fortuitous just because a Saturday telephone call might just as easily have been made from the sidelines of a sporting fixture across the border in New South Wales. The Defendant’s submissions do not reflect the professional, collaborative and focussed nature of the work of the River Operations Team. The work practices of the River Operations Team are addressed in more detail below.
The Plaintiffs’ case is that the River Operations Team ought to have made reasonable decisions about the exercise of the release power (and calls on the Goulburn IVT account and releases from the Menindee Lakes). The work which the Plaintiffs submit ought to have been undertaken as part of making reasonable decisions, such as further or different modelling with the Fortran Assessment Model and/or the Daily Operations Spreadsheet, record keeping and consultation with others, would properly have been undertaken by the River Operations Team at its workplace. The same applies to the negotiations about access to the MIL infrastructure where the alleged breach focusses not so much on individual communications with MIL or WaterNSW but on the commercial decisions made by the Defendant in relations to the negotiations.

Even if the Defendant’s conduct occurred in the Australian Capital Territory the burden of the Defendant’s submissions is that the choice of law ought to be determined not where the Defendant’s conduct occurred but at the place to which it was directed and the place where it assumed significance. For the purposes of identifying the applicable law, the Defendant’s submissions are unpersuasive for a number of reasons. First, it makes sense to refer to a negligent misstatement being “directed to” a particular location because receipt of the statement and action upon it are part and parcel of the wrong occurring. Until there is receipt and action, there is simply a statement albeit one which is incorrect and made without reasonable care. Like the accountant’s negligence in Voth v Manildra Flour Mills, the present case is not comparable to negligent misstatement cases. For the purposes of determining where the tort was committed, the Defendant’s negligent decisions were acts complete in themselves. [20]
Secondly, the Defendant’s argument depends on the content of the alleged negligent conduct being reduced functionally to the directions given (or not given) under cl 98 of the Agreement. Insufficient regard is had to the anterior decision-making process which is the substance of the alleged negligence. It may just as cogently be said that the River Operations Team’s decision making was “directed to” the Defendant’s own duty officer who was responsible for communicating a direction to the operator of the relevant storage. The Defendant’s duty officer was located in the Australian Capital Territory. The Defendant’s submissions do not grapple with the substance.
Thirdly, the Defendant’s argument further depends on the content of the alleged negligent conduct being reduced geographically to directions given for releases from Hume Dam in New South Wales. Bulk transfers from Hume Reservoir to Lake Victoria lie at the heart of the Plaintiffs’ case but they are not the totality of the case. There are other essential parts. The releases which are most central to the Plaintiffs’ allegations occurred at Yarrawonga Weir with all its geographic intricacy. The measurement of the amount of water releases each day at Yarrawonga Weir is the touchstone for the alleged negligence as made plain by paragraphs 36, 141 and 178 of the Fifth Further Amended Statement of Claim. In both the 2017/2018 water year and the 2018/2019 water year another essential element of the Plaintiffs’ case is the failure to call on the Goulburn IVT account. On the Defendant’s own analysis, decisions about the Goulburn IVT account were directed to and assumed significance in Victoria, not New South Wales. That is where GMW would have released water if directed by the Defendant. The Plaintiffs rightly submit that the Defendant’s attempt to relocate the place to which its conduct was directed to the confluence of the Goulburn River and Murray River is contrived and cannot be accepted. In substance, the Defendant’s decisions were multifaceted. On the Defendant’s own analysis they were simultaneously directed to locations in both New South Wales and Victoria. The Defendant is not able to avoid that conclusion by focussing on Hume Dam. These circumstances underscore the centrality of the anterior decisions and the relatively peripheral nature of the multiple steps subsequently taken to give effect to the decisions.
Fourthly, the contention that the Defendant’s decisions were directed to, or assumed significance in, locations where water was physically released is incongruous with the nature of the Plaintiffs’ case. The Plaintiffs sue for pure economic loss suffered by reason of reduced water allocations to them. Their case is not like the Queensland Floods case [21] where the release (or non-release) of water caused physical damage to the property of the claimants. The physical release of water at Yarrawonga Weir (or Hume Dam) and its subsequent loss overbank in the Barmah-Millewa Forest is an important part of the Plaintiffs’ case but it is an overstatement to say that it is the gravamen of the case. The Plaintiffs say that the consequence of the water lost overbank was, first, less water in the storages of the Murray River system and, next, a lower allocation of water to each of New South Wales and Victoria. Under the statutory framework, the Defendant is responsible for accounting for water and for making allocations to the States (Subdivisions C and D of Division 1 of Part XII of the Agreement – see below). There is nothing to suggest that allocations to the States were not made by the Defendant at its workplace in Canberra. Making the allocations to the States was the Defendant’s final involvement in the process by which the Plaintiffs claim to have suffered loss.
In this more complete context the place to which the Defendant’s decisions were directed is not as simple or singular as the Defendant would have it. As stated in cl 4(2) of the Objectives and Outcomes Document, the general objective of the Defendant’s decisions was to operate the Murray River system efficiently and effectively in order to deliver State water entitlements, and to maximise the water available to the Southern Basin States after providing for operating commitments in the Murray River system. From the Defendant’s point of view, its decisions were ultimately “directed to” the allocations to be made to the States which happened in Canberra. That is where the Defendant’s decisions assumed significance.
After the States received their allocations they, not the Defendant, made allocations to the Plaintiffs which in the ordinary course would have occurred in New South Wales (or in Victoria for some Group Members). The choice of law enquiry is not, however, about the place where the Plaintiffs suffered loss, even for a case of negligence where loss is the gist of the action.
Fifthly, and related to the two preceding points, as the Water Act 2007 makes plain the management of the water resources of the Murray-Darling Basin is a matter of national interest. The relevant parts of the Water Act 2007 are set out later in this Judgment. The Defendant is a central part of the structure which has been put in place to achieve the purposes of that Statute, including by acting as an expert which is independent of the States. The Defendant’s role is reminiscent of John Dixon J’s reference in Roo Roofing to “centralised” negligence occurring within a national scheme. [22] The political centralisation of the making of the Defendant’s decisions which are the subject of this case is an integral part of the structure. Whilst independence from the States does not necessarily entail physical separation of decision-making entities, it naturally follows within the Australian Federation with its national capital territory. The national character of the structure is consistent with the decisions of the Defendant assuming their significance at the centralised place where they are made and not simultaneously at multiple other locations.
There is some substance in the analogy which the Plaintiffs seek to draw with Pabai. As Wigney J said, each case turns on its own facts and it is generally not appropriate to reason on the basis of factual analogies. [23] Nonetheless, Pabai is an illustration of the law found to be applicable to a tort which was said to have been committed by the Commonwealth making decisions in Canberra. The Defendant rightly points out that the decisions in Pabai concerned policy matters but the distinction is one of degree. [24] The decisions in Pabai included decisions about the funding and construction of seawalls. Like the present case, Pabai focussed on the central making of the decisions more than the remote implementation of them. The law of the Australian Capital Territory was held to be applicable.

As they did in a previous hearing in these proceedings, both parties submit that the Defendant is the Commonwealth for the purposes of s 64. [25] The Defendant submits that therefore s 64 operates to pick up the ACT Liability Statute and apply it to the parties’ rights in this case. [26]
The Plaintiffs put forward two arguments as to why that submission should not be accepted. First, the Plaintiffs submit that s 64 is inconsistent with s 27 of the Australian Capital Territory (Self-Government) Act 1988, and the latter provision is to prevail as the more recent and more specific enactment of the Commonwealth Parliament. The Plaintiffs rely on the High Court’s decision in Deputy Commissioner of Taxation (Cth) v Moorebank Pty Ltd, [27] where it said that the provisions of the Judiciary Act 1903, including s 64, enjoy no special authority among the statutes of the Commonwealth and may be expressly or impliedly amended or repealed by subsequent Commonwealth legislation. The application of s 64 to cases falling within the provisions of subsequent legislation will be excluded to the extent that such application would be inconsistent with the subsequent legislation. [28]
I do not accept that there is any inconsistency. Section 27 is directed to the power of the Australian Capital Territory Legislative Assembly to make enactments which bind the Crown in right of the Commonwealth. It does not address the circumstances in which any such enactment may be applied by force of the Judiciary Act 1903 to determine the rights of the parties to a suit to which the Commonwealth or a State is a party. As a matter of construction, there is no inconsistency between the two provisions.
The Plaintiffs’ second submission is that the nature and function of s 64 is to fill what would otherwise be lacunae or gaps in the law in a suit to which the Commonwealth or a State is a party. [29] It is submitted that the common law of Australia is complete in itself and that there is no gap which needs to be filled by the various provisions of the ACT Liability Statute. The submission does not need to be considered beyond Chapter 4 of the ACT Liability Statute which governs negligence. As far as Part 4 is concerned, I do not accept the Plaintiffs’ submission. There is no dispute that s 64 may extend to substantive rights. [30] The provisions of Chapter 4 are intended to have effect as modifications to the common law of negligence in accordance with their terms. They are not intended to be meaningless by filling gaps which do not exist. It is often submitted, as the parties submit in the present case, that the operation of the provisions in Part 4 will not give rise to a result which is different to that which would have occurred under the common law. Even if that is true, the provisions nonetheless provide the framework within which the parties’ rights are to be determined. They have work to do no less in a case to which the Commonwealth or a State is a party than in any other case. The application of Part 4 to the rights of the parties in these proceedings conforms to the nature and function of s 64.
The Plaintiffs alternatively submit that the question of the operation of s 64 of the Judiciary Act 1903 ought not to be decided because none of the provisions of the ACT Liability Statute will affect the outcome of the case. Once it is accepted that the applicable law of the wrong is the Australian Capital Territory and not New South Wales or Victoria, the Defendant agrees that the operation of s 64 is not significant. The Defendant nonetheless submits that the question ought to be decided.
Even if no different result would have followed had the unmodified common law of negligence been applied it is still necessary to decide this case in accordance with the correct governing principles. The application of s 64 determines what those principles are and it is therefore appropriate that the question be decided. In any event, as explained below the applicability of the Part 4 provisions about negligence is indirectly material to the question of the admissibility of some of the Plaintiffs’ hypothetical evidence. For these reasons, I find that in this case the provisions of Part 4 of the ACT Liability Statute apply by virtue of s 64 of the Judiciary Act 1903.

As the Plaintiffs submit, the most grammatical construction of s 5D(3)(b) gives it an operation as one part of a coherent whole by which the exclusion of evidence applies in the circumstances described in the chapeau and “factual causation” takes its meaning from s 5D(1)(a). That construction accords with the convention of defined terms taking their meaning from the definition in the statute in which they are used. It should, however, be noted that the attribution of meaning to “factual causation” is not by way of an explicit definition, of which examples are to be found in s 5. Section 5D(1)(a) describes a concept and concludes the description with “(factual causation)”.
Beyond these matters there is little to commend the Plaintiffs’ construction. Because the ACT Liability Statute also uses the phrase “factual causation”, no violence to the text and only a minimal imposition on the context is necessary for the reference to “factual causation” in the s 5D(3) chapeau to include the attribution of meaning to that term by s 45(1)(a) of the ACT Liability Statute. Not only is the phrase the same but the meaning attributed to it is materially the same in both statutes. That is borne out in this case where both parties agree that it does not matter whether the NSW Liability Statute or the ACT Liability Statute (or the Wrongs Act 1958 (Vic)) governs the question of causation.
Nor is there any risk that the operation of s 5D(3)(b) will be distorted if it is applied to proof of causation under the ACT Liability Statute. The purpose of s 5D(3)(b) is to prevent a trial judge placing any weight on the sort of evidence described in the provision. [40] Whether or not a civil liability statute applies, the vice which is sought to be addressed by the measure is the same. Even where the plaintiff is an honest witness, it is notorious that the plaintiff’s own evidence about what he or she would have done may be unreliable. [41] An inference drawn from the objective facts is to be preferred. These are common law principles which accord with purpose of s 5D(3). That purpose applies equally whether the trial judge is making a determination of factual causation for the purposes of s 5D or s 45 of the ACT Liability Statute. There is no rational reason why the New South Wales legislature should be taken to have intended that s 5D(3)(b) not apply in the circumstances of this case.
As a variation on the same theme, the Plaintiffs make the further submission that s 79(1) will not pick up a law which is an integral part of a State legislative scheme if to do so would give the law an altered meaning. Reference is made to Commonwealth v Mewett where Gummow and Kirby JJ referred to a State limitation period which would be given an altered meaning if it were to be picked up without other parts of the State statute which allowed for an extension of the limitation period in certain circumstances. [42] That principle has no application in this case. As already observed, there is nothing in the application of s 5D(3)(b) which would be altered if it were to be picked up and applied to a determination of factual causation for the purposes of s 45(1) of the ACT Liability Statute. Apart from generalities the Plaintiffs do not identify any such alteration.
Whilst the text of a State statute may mean that there will be no cases where it is applicable when the Court is exercising federal jurisdiction, there is sufficient flexibility in the operation of s 79 of the Judiciary Act 1903 to pick up some State laws which are expressed by reference to the State context. [43] Section 5D(3) of the NSW Liability Statute is one such law even if “factual causation” is construed narrowly.
For these reasons s 5D(3)(b) of the NSW Liability Statute is picked up and applies to these proceedings by force of s 79(1) of the Judiciary Act 1903. The Defendant’s specific objections to the Plaintiffs’ evidence are addressed below in the context of the parties’ submissions on quantification of loss.

In a novel case such as this, the existence, nature and scope of a duty of care to avoid pure economic loss is to be determined by a judicial evaluation of the salient features of the case which tend for or against the duty. The evaluation is not an exercise in going through a checklist, nor is it an instinctive synthesis of the salient features of the particular case taken as a whole. The evaluation is a chain of reasoning linking the salient features of the case to the conclusion that the alleged duty is or is not owed as a matter of principle. The evaluation of the salient features in the particular case is a means of facilitating the incremental and analogical development of the common law from relevant precedents. [57]
In the abstract, the features of a case which may be salient to the question of duty are not capable of exhaustive statement. [58] The parties made submissions about the features of this case which they contend are salient. Having regard to those submissions, the following paragraphs describe some features which have been found to be salient in other cases.
A feature which has been found to support the imposition of a duty of care to avoid pure economic loss is where the defendant had actual knowledge of the risk to a person or class of persons of the particular type of loss that eventuated, and the magnitude of that loss. [59] In Perre v Apand McHugh J said that the knowledge may also be constructive for some kinds of economic loss. [60]
Another feature which has been found to support the imposition of a duty of care to avoid pure economic loss is where the defendant had control over the activity which caused the loss. [61] It is necessary to consider the degree and nature of the control exercised by the defendant over the risk of harm which eventuated. [62]
Another feature which has been found to support the imposition of a duty of care to avoid pure economic loss is where the plaintiff was vulnerable to the economic loss which eventuated because he or she was unable to protect him or herself from it. To be a feature of the case which is supportive of the imposition of a duty of care, the vulnerability must be special in the sense that the plaintiff is in a different position to the general public. [63] A capacity of the plaintiff to protect him or herself has been found to tell against the imposition of a duty of care being owed. [64]
A feature which has been found to support the imposition of a duty of care to avoid pure economic loss is where the plaintiff has relied on the defendant and where the defendant is aware of the reliance. [65] Again, the reliance must be special in the sense that general reliance that a public entity will perform a public function is not sufficient. [66]

A feature which has been found to tell against the existence of a duty of care to avoid pure economic loss is where the duty would give rise to indeterminant liability. Whilst indeterminacy is sometimes referred to as another salient feature, sight must not be lost of its importance. The passage set out above from the Mallonland case shows that the law of negligence’s different treatment of pure economic loss is in part reflective of a concern about indeterminacy. Indeterminant liability will “ordinarily deny” [70] or be a “valid reason to reject” [71] the existence of a duty of care.
A liability will be indeterminate in the relevant sense if it cannot be realistically calculated having regard to the amount, the time and the class to which the liability may be incurred. [72] The touchstone is the inability to calculate, not the size of the potential class nor the magnitude of the total loss once the calculation has been done. If there is sufficient certainty about the qualifications of class membership and the quantum of the loss which each member may suffer, the defendant’s liability will not be indeterminant even though the class is very large and the total loss is substantial. [73]
A finding that there is a particular class with an ascertainable membership to which the defendant owes a duty of care will not be denied just because there may be other classes to which the defendant would owe similar duties of care but which have memberships which cannot be ascertained. Whilst the defendant’s overall liability may be indeterminant in such a case, the defendant’s liability to the particular class will not. In Perre v Apand, [74] McHugh and Hayne JJ held that the defendant owed a duty of care to potato growers within 20km of the relevant land but rejected a further duty of care to others who were indirectly affected by the economic loss suffered by those potato growers. The further duty of care would have given rise to an indeterminant liability because there was no rational basis to limit the persons theoretically within the class. The rejection of the further duty of care did not, however, preclude the duty of care owed to the potato growers.
Indeterminacy is antithetical to the defendant’s actual (or constructive) knowledge of the risk to a class of persons and of the magnitude of the potential loss. It may also be antithetical to reasonable foreseeability which is essential. The defendant cannot know of or even foresee that which is indeterminant in the relevant sense.

In a case such as this where the Defendant is a statutory authority, consideration of whether a duty of care is owed and, if so, the contents of the duty starts with the statutory framework in which the Defendant operated. [78] The terms, scope and purpose of the statutory framework must be examined closely. The examination must identify the functions, duties and powers of the statutory authority. Consideration must be given not only to the powers which were exercised in the immediate case but also those which were not. [79] Consideration must also be given to any other duties imposed on the statutory authority and whether those duties may conflict with the alleged duty of care. [80]
The need for such an examination is a pertinent example of the reason why the alleged duty of care has to be considered prospectively and not by reference to the particular breach which is said to have occurred in the immediate case. After the event the power which was exercised and the circumstances in which it was exercised are known. If the alleged duty was owed to a particular class of persons, the statutory authority may also have had to take reasonable care to avoid the risk of harm to that class when deciding to exercise other powers, or not to exercise other powers from which other consequences may have flowed. All the potential consequences of the alleged duty need to be considered.
The ultimate question is whether, in all the circumstances, the statutory regime erects or facilitates a relationship between the statutory authority and a class of persons which displays sufficient characteristics answering the criteria for intervention by the tort of negligence, such that the common law imposes a duty of care as to the manner in which the statutory authority exercises the power which was exercised. If so, the common law duty of care operates alongside the rights, duties and liabilities created by statute. [81] In answering the question, the relationship which is erected or facilitated must be considered as a whole and not only by reference to an isolated aspect of the relationship. [82]
As stated above the duty which is imposed by the common law must not be incompatible with the powers and duties imposed by the statute. [83]

The statutory regime which established the Defendant and the legal framework in which it has subsequently operated has some features and complexity not commonly (or at all) found in other statutory regimes.
The statute which lies at the centre of the legal framework is the Water Act 2007. It is an enactment of the Commonwealth Parliament. The measure enacted by the Commonwealth Statute which is most (but not exclusively) relevant to the current case is the establishment of the Defendant as a statutory authority to operate in accordance with the Murray-Darling Basin Agreement.
The text of the Agreement is set out in Schedule 1 to the Commonwealth Statute. The Agreement takes its legal effect not from the reference to it in the Commonwealth Statute but from the consent of the parties to the Agreement which are the Commonwealth and the States and Territory directly concerned with the Muray-Darling Basin. The Agreement contains the primary level of regulation of the Murray-Darling Basin.
The Agreement in turn provides for other documents to be prepared and approved in which more detailed regulation is set out. Amongst those other documents, and of importance in the current case, is the Objectives and Outcomes Document. I will generally refer to this as the O&O Document.
There is a further layer of regulation which is relevant in this case which is to be found in the statutes regarding the distribution of water to the individual holders of water entitlements, such as those who carry on businesses of irrigated agricultural production in the relevant parts of New South Wales and Victoria. The Defendant is a creature of the Commonwealth Statute but the question whether it owed the alleged duty of care cannot be determined without considering the State regulatory regimes. In the Statement of Claim the persons to whom the duty is alleged to have been owed and the risks of harm it is said the Defendant had a duty to avoid are defined by reference to water entitlements under the State regimes. In New South Wales, the relevant statute is the Water Management Act 2000 which I will refer to as the NSW Statute. In Victoria it is the Water Act 1989 which I will refer to as the Victorian Statute.
Before turning specifically to the establishment and functions of the Defendant, it is necessary to address the framework generally.

The purpose of the Basin Plan is to provide for the integrated management of the Basin water resources in a way that promotes the objects of this Act. [84] A detailed process was prescribed for the development of the Basin Plan which has now been completed. The Defendant had a role, namely to prepare the Basin Plan for adoption by the Commonwealth Minister for Water after a process of consultation with the concerned States and Territory. The Basin Plan was signed into law by the Minister on 22 November 2012. It is now a legislative instrument. [85]
The Basin Plan is arranged into 13 Chapters which broadly take their content from a list of matters specified in s 22 of the Commonwealth Statute. Chapter 6 addresses the long-term average sustainable diversion limits on taking water from the Murray-Darling Basin. For each region, cl 6.04 specifies the number of gigalitres which may be taken each year. The Defendant may review the sustainable diversion limits (cl 6.06) and propose adjustments (cl 7.01). Chapter 8 sets out an environmental watering plan. Chapter 11 deals with the amount of water necessary to meet critical human needs. Chapter 12 sets out detailed water trading rules. Chapter 13 specifies a programme for monitoring and evaluating the effectiveness of the Basin Plan, for which the Defendant is primarily responsible.
The subject of critical human needs is addressed in Part 2A of the Commonwealth Statute. It provided that the Basin Plan had to be prepared having regard to the fact that the governments have agreed that the highest priority water use is the core human consumption requirements in urban and rural areas and those non-human consumption requirements that a failure to meet would cause prohibitively high social, economic or national security costs as far downstream as Wellington in South Australia. [86] Provision is also made for monitoring conditions. In extreme conditions, provision is made for departure from the usual water sharing arrangements to ensure that critical human needs are met. [87]

The consequence of this provision is that each of New South Wales and Victoria is entitled to all the water from any tributary on its own territory if the tributary enters the Murray River below Doctors Point. Water in the Menindee Lakes, despite being located entirely within New South Wales, is shared equally between New South Wales and Victoria until the storage level falls below 480 GL and then not again until after the storage level rises above 640 GL (cl 95). Otherwise, the practical outworking of the sharing arrangement is that the water stored in Dartmouth Reservoir, Hume Reservoir, Lake Victoria, the Menindee Lakes (when included) and otherwise in the Murray River is shared half each by New South Wales and Victoria. The water delivered to South Australia is deducted equally from the water allocated to New South Wales and Victoria (cl 96). There are mechanisms for the sharing arrangements to be suspended in times of insufficiency for critical human needs (Division 2) and extreme or unprecedented circumstances (Division 3).
As more fully addressed below, the provisions in the Agreement for sharing water between the States are followed by Subdivision C which is entitled “Control by Authority”. The first provision in Subdivision C is cl 98 which empowers the Authority to give directions for water to be released from the River Murray storages. The Defendant’s power under cl 98, which is referred to as the “release power”, is addressed below.
The agreement about sharing the water between the States cannot operate without agreement on the rules to account for the storage, transfer, loss and use of the water in the river system. Subdivision D of Division 1 of Part XII contains the parties’ agreement about the accounting rules. The rules address the circumstances in which a State is taken to have used part of its entitlement (cl 109). They also provide for the deduction from each State’s entitlement of any water “lost by evaporation or other means” (cl 110). For example, evaporation from a storage which includes water deferred by South Australia is deducted first from South Australia’s entitlement (cl 110(1A)). Special rules are made for accounting for water released from each of Dartmouth Reservoir (cl 118), Hume Reservoir (cl 119), Menindee Lakes Storage (cll 120 and 121) and for upper River Murray storage spills (cl 117). There is also provision for periods of special accounting when there is a declared shortage of water (cl 106).
The Defendant has a central role in accounting for the water in the river system which is addressed below.

The scheme of these provisions is to impose a prohibition on the Defendant making any decision in relation to river operations if the decision has the potential to have a material effect on State water entitlements. Such a decision requires the approval of the BOC. The BOC’s approval may be provided in advance by virtue of approval in the O&O Document. If the O&O Document does not approve the decision, the Defendant is required to seek the BOC’s approval by a referral under cl 33(3).
The Agreement does not specify when a decision will have the potential to have a material effect on State water entitlements. Clause 31(4) contemplates that the O&O Document will specify criteria. As set out below, the O&O Document specifies a number of specific matters which the Defendant must refer to the BOC but it does not specify general criteria for materiality.
Other matters addressed by Part V of the Agreement are preparation of an annual corporate plan for the purposes of s 35 of the Public Governance, Performance and Accountability Act 2013 (Cth) and preparation of an annual work plan which sets out the Defendant’s activities relating to the Agreement for the next four years.
The physical infrastructure which is critical to the regulation of the flow of water in the river system is addressed in Part VIII of the Agreement. Certain works are listed in the Schedule to the Agreement, including Dartmouth Dam, Hume Dam, the Lake Victoria works, Yarrawonga Weir and thirteen other weirs and locks in New South Wales, Victoria and South Australia. The Defendant does not own any of the works but has a role in planning the relevant State authority’s management and operation of the works. Management includes maintenance, repair, renewal and replacement (cl 53(2)). The Defendant must prepare an asset management plan and submit it for approval, first to the BOC and then to the Ministerial Council (cl 53). The Defendant may determine procedures for the operation of the works (cl 66) with which the State authority must comply (cl 68). Provision is made for the construction of new works.
The Defendant performs important functions in the implementation of the States’ agreed water sharing arrangement. Whilst the Defendant does not own or directly operate storages such as Hume Dam and Yarrawonga Weir, cl 98 of the Agreement provides:

The Plaintiffs submit that a duty to take reasonable care to avoid pure economic loss may arise where there has been an assumption of responsibility without the requirement separately to consider the salient features of the case. The Plaintiffs rely on Edelman J in the Mallonland case at [61] and [66] and a number of English cases [108] for the proposition that an undertaking by a defendant to take on a task for a plaintiff (or for a class of which the plaintiff is a member) may be implied not only from the defendant’s conduct but also from its “office”. It is submitted that there may be an assumption of responsibility from the performance of a statutory obligation or the operation of a statutory scheme.
The Plaintiffs submit that in this case the Defendant has expressly or impliedly undertaken to perform river operations with due care, amongst other things, to protect the Plaintiffs from the risk of economic loss that arises from inefficient river operations. As submitted, the Plaintiffs identify the relevant function of the Defendant to be the requirement in cl 29(1)(b) of the Agreement to give effect to high-level decisions of the BOC. The high-level decisions are said relevantly to be found in the O&O Document. The Plaintiffs’ case is that the Defendant owed them a duty of care when giving effect to the objectives and outcomes in the O&O Document.
In support of their argument the Plaintiffs emphasise a number of matters. First, it is submitted that the BOC made policy decisions. The Defendant’s implementation of those decisions may be characterised as operational because the Defendant made decisions which were a product of administrative direction, expert or professional opinion, technical standards or general standards of reasonableness. That perspective is said to be supported by the requirement that the Defendant refer certain matters to the BOC for approval in accordance with cl 33 of the Agreement and cll 13-17 of the O&O Document.
Next, the Plaintiffs submit that the Defendant’s river operations functions are subject to “empirical, hydrological standard[s]” which are apt to supervision by the Court, although the Plaintiffs do not identify a precise or measurable content of the standards. The Plaintiffs instead refer to numerous statements in the documents about efficiency, effectiveness, maximisation, minimisation of waste and reasonable practicability. The Plaintiffs submit that the Defendant’s conduct is amenable to review by the Court and therefore it is capable of being subject to a duty of care.
The Plaintiffs further submit that the provisions of the O&O Document recognise and formalise a direct relationship between the Defendant and irrigators such as the Plaintiffs when the Defendant undertakes its river operations functions. The Plaintiffs rely on statements which refer to people using the Murray River system and stakeholders. The Plaintiffs particularly rely on cl 14(1)(b) of the O&O Document which required the Defendant to refer to the BOC any decision which would likely require a State to reduce the volume of allocations it has previously announced as being available to water entitlement holders. The Plaintiffs further submit that a direct relationship with irrigators arises from the “interconnected statutory framework” which includes the NSW Statute and the Victorian Statute by which the States allocate water to individual water entitlement holders. The Defendant’s obligation to maximise the water available to the States is to be seen in that light. A relationship between the Defendant and irrigators is said further to be created and facilitated by the objects of the Commonwealth Statute which refer to optimisation of economic outcomes and the use of markets to achieve efficiency. Reference is also made to the extrinsic materials which recognise the importance of the Murray River to irrigated agriculture in Australia and the national leadership and integrated management to be provided by the Defendant as an “independent expert”. [109]
The Plaintiffs submit that the release power in cl 98 of the Agreement is central to the Defendant’s river operations functions. By reference to what Beech-Jones J said in [20] of Chapter 11 of his Honour’s Judgment in the Queensland Floods case, the release power is not a mere statutory power because the non-exercise of the release power is the positive exercise of a power to store water which would otherwise be released.
By reason of these matters, the Plaintiffs submit that the Court ought to find that the Defendant assumed responsibility to the Plaintiffs when it exercised the release power.
Whilst much attention was given to the Defendant’s exercise of the release power in 2017 and 2018, there is one aspect of the Plaintiffs’ case which does not arise from an exercise of the release power. Prior to 10 September 2018 the Defendant operated the river system without the benefit of access to the infrastructure privately owned by MIL. The Plaintiffs submit that the Defendant’s duty of care included an obligation to obtain access to the MIL infrastructure. The Plaintiffs submit that this part of their case falls within the pleading. The Defendant’s function of giving effect to high-level decisions of the BOC as set out in the O&O Document included the specific objectives and outcomes in cll 5.1 and 12.9 of Annexure A to the O&O Document. Those specific objectives and outcomes addressed bypassing the Barmah Choke with a controlled flow which “may be directed” through the MIL infrastructure in certain circumstances. The specific objectives and outcomes contemplated discussions and agreement with MIL for that purpose. The Plaintiffs do not identify a specific power of the Defendant in respect of which there was an assumption of responsibility for this part of their case, although under s 173 of the Commonwealth Statute the Defendant has the power to do anything that is necessary or convenient to be done for or in connection with the performance of its functions. The Plaintiffs appear to contend that the Defendant assumed a responsibility to them when negotiating with MIL about access to the MIL infrastructure.
In addition to the statutory framework, the Plaintiffs submit that the assumption of responsibility is further demonstrated by contemporaneous communications from the Defendant to irrigators and the general public in which the Defendant repeatedly referred to its awareness of and responsibility variously to farmers, communities, irrigators, agricultural industries, entitlement holders and stakeholders. The Plaintiffs also rely on two of the Defendant’s internal documents (a draft reference manual and an email between two employees of the Defendant on 3 September 2018) which contained expressions of responsibility to irrigators, together with the subjective views expressed by the witnesses from the Defendant who gave evidence in these proceedings. The Plaintiffs emphasise that some of the views expressed by officers of the Defendant were expressed in terms of irrigators ordering water from the States which then collated the orders and passed them on to the Defendant. The Plaintiffs also rely on some minutes of contemporaneous meetings where similar expressions of responsibility are recorded, although the meetings were not meetings of the Defendant but meetings of the BOC (albeit attended by some officers from the Defendant).
The Plaintiffs see further support for an assumption of responsibility by reason of charges levied on irrigators which charges were calculated by reference to the costs incurred by the Defendant in performing its functions.
The fact that the Agreement addresses allocations to the States and not to individuals or even classes of individuals is said by the Plaintiffs to be a “lawyerly device” which is not reflective of reality.

The Defendant submits that a novel duty of care is to be assessed having regard to the salient features of the case. For a statutory authority that assessment must begin with the statutory framework.
The Defendant makes a number of submissions about the statutory framework. It submits that the statutory framework establishes a relationship between the Defendant and the States but not between the Defendant and the Plaintiffs. The Defendant emphasises that the only entitlements specified in the Agreement are entitlements of the States. The individual entitlement holders do not feature anywhere in the statutory framework. It submits that the water entitlements of the Plaintiffs are creatures of the State statutes and that the allocation of water to those water entitlements depends upon the exercise of discretion by State ministers and agencies. Any protection for individual entitlement holders is through the States’ relationships with the Defendant. The Defendant submits that there are multiple different holders of water entitlements within each State as between which water is allocated by the State in accordance with its own policy interests. The Defendant further submits that accountability is intended to operate at the political level, not the legal level.
The Defendant further submits that as a matter of practice there is no relationship between the Defendant and individual water entitlement holders. Whilst it accepts that the orders for water placed by the States with the Defendant reflect the orders placed by entitlement holders with each State, the Defendant has no interest in the identity of individual entitlement holders. The States are responsible for retail water accounting. The Defendant is reliant on data from the States as to diversions of water from the Murray River system, which data is essential to the conduct of the Defendant’s river operation activities.
The Defendant submits that the Plaintiffs’ demarcation between high-level policy decisions and operational decisions ought not to be accepted and that there are aspects of the decisions which the Defendant makes which are not neatly described as either policy or operational. In any event, the Defendant submits that the fact that the decision may be characterised as operational does not mean that a duty of care will exist. More pertinently, the Defendant submits that its decisions are not subject to an empirical, hydrological standard and that they are not apt for supervision by the Court.
The Defendant submits that the O&O Document provided a mechanism for resolving conflicts between the various objectives and outcomes, namely consultation with the States through the WLWG, the RMOC and the BOC. Such a regime does not allow for a duty of care to be owed by the Defendant to one class of users when there may be a conflict with a duty owed to others.
The Defendant submits that incompatibility with the statutory framework is fatal to an alleged duty of care and that that is the position in the present case. At the very least the imposition of a duty of care to avoid pure economic loss to the Plaintiffs would distort the Defendant’s performance of its statutory functions. Indeed, it would subject the Defendant to conflicting and irreconcilable obligations. The Defendant submits that it is already subject to actual conflicts, including as between the States, between the environmental and other water entitlement holders and between different classes of water entitlement holders. Some specific examples of situations are given which the Defendant submits are actual conflicts.
Beyond incompatibility with the statutory framework, the Defendant accepts that pure economic loss to the Plaintiffs was reasonably foreseeable. The Defendant nonetheless denies that it has control over the source of the risk of harm which it identifies as suffering “economic loss as a result of receiving less, or insufficient, water”. The Defendant says that the primary source of that risk of harm is low rainfall. The Defendant reasons that the Plaintiffs are not vulnerable because there are measures available to them to manage the risk of low rainfall, including purchasing temporary water, building dams, levies and sinking bores etc.
The Defendant further submits that the duty of care alleged by the Plaintiffs would give rise to an indeterminant liability. It points out that the definition of Group Member in paragraph 8 of the Statement of Claim extends beyond the holders of water entitlements to include contractual counterparties and entities which “utilised” the water entitlements of others. The Defendant submits that it is not apparent how it could have identified all class members or how it could estimate each person’s business or trading losses. The Defendant submits that if the duty of care is owed to the Plaintiffs and Group Members, a similar duty would be owed to others who use the Murray River for economic purposes, and others located elsewhere in the Murray-Darling Basin.
The Defendant submits that assumption of responsibility is not a free-standing basis for the imposition of a duty of care. But even if it is, the analysis must begin with the statutory framework which in this case does not erect or facilitate the necessary relationship between the Defendant and the Plaintiffs.
The Defendant submits that the salient features of this case are such that the Defendant did not owe a duty to avoid pure economic loss to the Plaintiffs.

The outline of the statutory framework which is set out above demonstrates the intention that the Murray River system be operated in the public interest. The intention is expressed in the long title of the Commonwealth Statute, where reference is made to the national interest, and is apparent from the objects specified in s 3.
The purpose of the statutory framework for which the Defendant was established and in which it functions is further to be understood in light of the fact that the subject matter which is to be regulated is the water of the Murray-Darling Basin. The statutory framework proceeds from the premise that it is a valuable and scarce resource. Those features of the resource are consistent with an intention that the river system be operated efficiently but it is difficult to give meaning to the concept of efficiency without an understanding of who is intended to benefit from the efficiency. The statutory framework addresses itself to the benefits to be had from the water in the Murray River system in terms which are uniformly generalised. Reference is made to types of activity rather than to those who undertake the activity: economic, social, environmental, cultural, irrigation, industrial, recreational, stock and domestic use, critical human needs and national security. Some of the identified benefits are susceptible to a degree of measurement but some not at all. Some of the benefits are non-financial, either entirely or in part. Apart from critical human needs, the statutory framework does not generally attempt to prioritise or even compare one kind of benefit to another.
Harmoniously with referring to the benefits with a high degree of generality, the statutory framework contemplates that the enjoyment of the various benefits will be at a level which is collective. An example which is pertinent is the object of the Commonwealth Statute which is specified in s 3(d)(iii) which refers to “the net economic returns to the Australian community”. Nowhere in the statutory framework, including the O&O Document, is there a reference to the enjoyment of benefits by any particular group or a particular class of individuals.
Indeed a prominent feature of the statutory framework is the absence of any reference to persons in an individual capacity and the absence of any reference to individual water entitlements. Leaving aside the NSW Statute and the Victorian Statute (see below), the Plaintiffs refer to only one place in the voluminous regulatory documentation where reference is made to holders of water entitlements, namely cl 14(1)(b) of the O&O Document (there is a variation on the same theme in cl 15(1)(b)). Clause 14(1)(b) required the Defendant to refer a matter to the BOC after first seeking advice from the RMOC in the following circumstance:

The structure which has been adopted to serve the public interest is for the States to agree upon a regime for the water of the Murray-Darling Basin to be shared amongst the States and for there further to be agreed accounting rules by which that water sharing regime will be implemented. The agreed water sharing regime and accounting rules are set out in the Agreement. Each year they were also addressed in cll 11 and 12 of the O&O Document. The terms of the Agreement and the O&O Document make clear that the role of the States is central and substantive. For example, the express purpose of the Agreement is to give effect to the Basin Plan, the Commonwealth Statute “and State water entitlements”. The term “river operations” is defined in the Agreement by reference to “sharing water between State Contracting Governments”. The Agreement establishes not only the Ministerial Council (Part III) but also the BOC (Part IV). No less than the Ministerial Council, the BOC is populated by members who “represent” the States and Territory (cl 18). Matters are to be referred to the BOC if they have the potential to have “a material effect on State water entitlements”. Under cl 102A the initial requirement to meet critical water needs is to be distributed by the Defendant amongst the States. Under cl 102C the Defendant may arrange for one State to advance water to another. The Defendant’s general objective in cl 4(2)(a)(i) of the O&O Document was to operate the River Murray system efficiently “in order to deliver State water entitlements” and in cl 4(2)(a)(ii) to maximise the water available “to the Southern Basin States”. The general outcome specified in cl 4(2)(b)(iii) was delivery to the States of “their authorised water orders”. All entitlements to receive water are expressed as entitlements of the States. [114] There are many other provisions which are directed to the interests of the States.
Specifically, and relevantly for this case, the interests of the States are the touchstone by which the Defendant is to exercise its power to direct the release of water from upper River Murray storages. In giving directions, the Defendant must have regard to maintaining supply to South Australia of the quantities of water to which it is entitled (cl 98(3)(i)) and facilitating the exercise of New South Wales and Victoria of their respective rights to use water from the upper River Murray as they require (cl 98(3)(iv)). As will be seen, there is a direct conflict between those two matters. Whilst the Agreement provides a regime for the management of that conflict, a conflict subsists at the practical, operational level by virtue of the unpredictable, uncertain and dynamic nature of the Murray River system.
The central role of the States is apparent from the Second Reading Speech of the Commonwealth Minister for the Environment, Heritage and the Arts on 25 September 2008 when the Commonwealth Statute was amended to recognise the Agreement. Hansard records the Minister’s statement that the States had referred powers to the Commonwealth as part of a new area of cooperative arrangements to face the challenges of water scarcity and water security. The Ministerial Council and the BOC were said to have been established so that the States would have a major role in putting the Basin Plan into effect.
Within that framework the Defendant has been established as an expert which is independent of the States to deliver water to the States in accordance with their agreed sharing regime and utilising their agreed accounting rules. The Minister’s Second Reading Speech went on to refer to the establishment of the Defendant as an “independent expert” agency, first to prepare the Basin Plan and then to ensure that the water resources are managed in an integrated and sustainable way. The independence to which the Minister referred is independence as between the States. It is necessary that the Defendant have that independence because the States have interests which provide the context in which Defendant was established to operate and between which there is potential for conflict.
The independence of the Defendant is also to be understood in the context that by consent to the Agreement the States have in substance forfeited important controls to the Defendant. The States own and operate various dams and weirs by which the Murray River system is regulated but, if the Defendant directs a release of water, s 98(1) provides that water must be released in accordance with that direction.
The statutory framework does not require or even acknowledge the existence of the allocation systems of the individual States (apart from the references in cl 14(1)(b) of the O&O Document). The position of individual holders of water entitlements is left to regulation by each State. Allocation of water in New South Wales is currently (and was at the relevant time) the subject of the NSW Statute and in Victoria it is currently (and was at the relevant time) the subject of the Victorian Statute. The Defendant submits that allocation within each State involves an exercise of discretion by Ministers and government agencies but that should not be overstated. For another purpose (causation), the Defendant accepts that each State has a system of allocation to water entitlement holders which is deterministic or mechanistic once the allocation to the State by the Defendant has occurred. Nonetheless, each State’s system is a matter for it. New South Wales and Victoria’s respective systems are similar but not identical. There are differences between each State’s population, geography, infrastructure, environment and climate. Victoria has classes of water entitlements for which New South Wales does not have an equivalent. For the purposes of determining allocations to water entitlement holders, Victoria currently makes a different (less pessimistic) assumption about the probability of future rainfall to that made by New South Wales. Each State’s system gives rise to water being received by entitlement holders in different quantities. For example, by the first week of October 2017, the holders of general security access licences in New South Wales had been allocated 29% of their entitlements whereas the holders of equivalent entitlements in Victoria had been allocated 100%.
Apart from New South Wales and Victoria, South Australia has different entitlements under the water sharing arrangement and its own system of allocation to water entitlement holders. Under cl 91 of the Agreement South Australia had certain rights to defer receipt of some of its monthly entitlements. On 2 May 2018 there was a meeting of the WLWG at which the South Australian representatives said that there may be a change in the approach to the exercise of South Australia’s deferral rights following the recent state election. Any impact on the quantum or timing of allocations to South Australian water entitlement holders was a matter for the South Australian government.
The role created for the Defendant within the statutory framework runs as far as allocation of water to the States and fulfilment of water orders from the States but no further. The Defendant has no role in the allocation or other use of water by each State. The Defendant does not know what water has been diverted by individual holders of water entitlements other than what the States report back to the Defendant for the purpose of the determination of future allocations to the States. Pursuant to cl 6(2)(a) of the O&O Document the Defendant’s obligation to achieve its objectives and outcomes was subject to the States providing timely and accurate data about allocations to their water entitlement holders.
The absence of a role for the Defendant in the allocation of water to individual water entitlement holders is incongruent with the Defendant owing the Plaintiffs a duty of care. The tension between the focus of the Defendant’s statutory role and the alleged duty of care is apparent from the way the Plaintiffs have sought to find a home for the alleged duty amongst the Defendant’s statutory functions and powers. The Plaintiffs ultimately submit that the specific function of the Defendant to which the duty “attaches” is the function in cl 29(1)(b) of the Agreement “to give effect to any high level decision of the [BOC] in relation to river operations”. It is a strange choice (not that there is a better one) because cl 29(1)(b) refers to another entity which is not the Plaintiffs. As explained above, the BOC is comprised of members who represent the Commonwealth, the States and the Territory. The members must act unanimously. Pursuant to cl 26(1)(c) of the Agreement the BOC’s functions relevantly include exercising responsibility for high-level decision making in relation to river operations, including setting objectives and outcomes to be achieved by the Defendant in relation to river operations. The definition of “river operations” is set out above, from which the focus on the States is apparent. The requirement that the Defendant give effect to the BOC’s objectives and outcomes may indicate an intention that the Defendant owes a duty to the BOC as a representative committee of the States, or directly to the States, but not a duty owed to the Plaintiffs.

In that context, there is no support in the Commonwealth Statute, the Agreement, the O&O Document or in the State statutes for the Plaintiffs’ contention that a relationship has been erected between them and the Defendant. Nor can it be said that the statutory framework facilitates any such relationship, and certainly not one which displays sufficient characteristics to warrant the imposition of a common law duty of care to avoid economic loss to the Plaintiffs. The Plaintiffs submit otherwise but the provisions relied upon are insubstantial and fall well short of that which is required.
The Plaintiffs submit that cl 14(1)(b) of the Agreement demonstrates an intention that there be a direct relationship between the Defendant and the Plaintiffs but for the reasons given above there is no support in cl 14(1)(b) for such a relationship. It is further to be observed that the reference in cl 14(1)(b) to “water entitlement holders” was not limited to the Plaintiffs and Group Members in this case but included all holders of water entitlements including those in South Australia, those in other parts of the Murray-Darling Basin, traders of water entitlements who do not use them for agriculture and environmental water holders. There is nothing to distinguish the Plaintiffs and Group Members from all other water entitlement holders.
The Plaintiffs also place reliance on s 174 of the Commonwealth Statute which provides that the Commonwealth must pay to the Defendant any amount paid by the Defendant to discharge a liability arising from an act or omission in the bona fide execution of the powers vested in the Defendant. Section 174 contemplates that the Defendant may incur a liability from an act or omission which occurs in the execution of its powers but it is a generic provision. It does not reveal anything about the sort of liability that the Defendant may incur (if any) and to whom (if anyone).
The Plaintiffs also rely upon cl 145 of the Agreement which provides a mechanism for a payment made by the Commonwealth in respect of a liability of the Defendant to be shared with the States equally. Clause 145 contemplates that the Defendant may incur a liability, but it is another generic provision. Indeed, when the detail of the States’ agreement to contribute to liabilities of the Defendant is considered it undermines any expectation that the Defendant might incur a liability to a particular class of water entitlement holders by reference to their geographic location, such as the Plaintiffs and Group Members in this case. Under cl 145 New South Wales, Victoria and South Australia must contribute even if the Defendant’s liability arises from an execution of power which did not occur in that State and did not relate to that State (cf the more limited obligation of Queensland (cl 37) and the Australian Capital Territory (cl 38)). The contribution regime may give rise to anomalous results if the Defendant incurs a liability as an incidence of a relationship with water entitlement holders located in a particular State. The present case illustrates the point. The Plaintiffs contend that the Defendant owed a duty of care to certain holders of water entitlements located in New South Wales and Victoria. If the Defendant is liable to pay compensation to those water entitlement holders, South Australia will be required by cl 145 to contribute equally to that liability.
The Plaintiffs do not refer to any other provision of the statutory framework to support a finding that it erects or facilitates the necessary relationship between the Defendant and them.
Of course the benefits to be obtained from the water of the Murray-Darling Basin concerns people as individuals and as members of classes. Different individuals and classes will have different interests in the various benefits, some more immediate than others. In light of the NSW Statute and the Victorian Statute, the Plaintiffs and other holders of water entitlements in those two States will be affected by the performance of the Defendant’s functions. It is not to be supposed that the statutory framework has been put in place in ignorance of that reality. Nonetheless the present task is to examine the statutory framework to ascertain whether it is intended to erect or facilitate a relationship between the Defendant and the Plaintiffs which warrants intervention by the tort of negligence, and specifically a duty of care to avoid pure economic loss to the Plaintiffs and Group Members.
The Plaintiffs’ submission that the interposition of a role for the States is a lawyerly device proceeds from a hindsight perspective of the events which have happened and finds no support in statutory framework. The Court’s task in considering the statutory framework is objectively to discern intention, in the first instance of the Commonwealth Parliament and then of the States and Territory acting directly as parties to the Agreement and indirectly through the BOC’s approval of the O&O Document. Especially in a Federation where the relevant river runs through (or along) four states, there is nothing to suggest that the statutory framework has been adopted to avoid the Defendant owing a duty of care it would otherwise owe to individual water entitlement holders in the various States.

There is another aspect of the statutory framework which cannot be reconciled with the Defendant being subject to a duty to avoid economic loss to the Plaintiffs. Inherent in the waters of the Murray-Darling Basin being a valuable and scarce resource is conflict amongst those concerned with the waters’ use. The Defendant is required to conduct its functions in the face of conflict. Decisions made by the Defendant may benefit some interests and prejudice others.
Recognising the difficulty which potential conflict poses to subjecting the Defendant to the alleged duty of care, the Plaintiffs submit that all conflicting claims to the scarce waters of the Murray-Darling Basin have been “resolved conclusively” by the Agreement. They specifically rely upon cl 88 which provides for South Australia’s entitlement and cl 94 which provides for the sharing of the residual water between New South Wales and Victoria. The Plaintiffs submit that the result is that the Defendant has been left with a purely operational role. They submit that the efficient operation of the river system will serve that conclusive resolution of conflicts. There is therefore room for a common law duty of care. They say that such a duty would enhance the statutory framework’s expressed concern for accountability. The imposition of the alleged duty does not therefore conflict with the statutory framework.
There are two difficulties with this argument. First, the only conflicting claims which have been addressed by the statutory regime are claims amongst the States. The statutory framework is not concerned with and does not address conflicting claims between different classes of water entitlement holders, whether the difference be geographic, reliability of access, type of use, temporal or other. The Defendant is required to perform its functions in the face of all such conflicting claims.
Secondly, the premise that there has been a conclusive resolution of all conflicting claims cannot be accepted even amongst the States. The Plaintiffs’ submissions to the contrary do not pay sufficient regard to the nature of the task required to operate the Murray River system and the inherently unpredictable, uncertain and dynamic environment in which that task is to be performed. When those matters are properly taken into account, it is clear that an important aspect of the Defendant’s functions is to manage competing risks in the delivery of State water entitlements. The parties to the Agreement have agreed a framework within which the Defendant is to manage the risks but they have not eliminated all potential for conflict in the performance of the Defendant’s functions.

One key area of conflict which is relevant to the Defendant’s functions arises from the competing risks inherent in meeting South Australia’s water entitlement each and every month. The conflict is central to the issues in the present case. In addition to informing the determination of whether the Defendant owed the Plaintiffs a duty of care, the conflict is also relevant to the allegations that the duty of care was breached. It is appropriate for the issue to be addressed in some detail.
The Agreement provides that South Australia “is entitled to receive” a specified number of megalitres of water each month (cl 88) measured as the water flows past specified locations along the Murray River and the Lindsay River close to the border with New South Wales (cl 89). South Australia’s monthly entitlement to receive is subject to South Australia’s right to defer (cl 91). It is further subject to the South Australian member of the BOC making a request to the Defendant that there be a limited variation between months, to which request the Defendant “may” accede (cl 90). New South Wales and Victoria are generally entitled to half the water in the Murray River system after South Australia’s entitlement has been deducted (cl 94). Water which flows over the border in excess of South Australia’s entitlement therefore increases the water received by South Australia at the expense of the water remaining in the system for allocation to New South Wales and Victoria (cll 92 and 93). Apart from cl 90, these provisions are expressed in mandatory terms.
The Plaintiffs specifically submit that this regime means that there is no conflict between the States because “South Australia’s entitlement must be met”. The simplicity of that submission accords neither with the practical reality of implementing the Agreement nor with the full terms of the Agreement.
Taking the practicalities first, receipt of South Australia’s entitlement might be implemented with precision month after month if there was an unlimited reservoir of water immediately adjacent to the South Australian border which could be accessed by a tap, turned on and off at will. [115] The reservoir would need to be unlimited not just as to the water available to be released from it but also as to its capacity to store any additional water which flows down. In reality the achievement of South Australia’s entitlement depends upon a large and complex river system which behaves in an inherently dynamic, imprecise and uncertain way. Key aspects of the river system which affect the receipt of water by South Australia include natural flows in the river, inflows resulting from rain, diversions for use, conveyance losses, evaporation losses, the location and capacity of upstream storages, flow times from storages, channel capacity and temperature.
Future rainfall and inflows into the river system upstream of the South Australian border are unknowable within time frames which are critical to the operation of the river. Winter months are generally wetter than summer months but rainfall cannot be predicted with reliability beyond four days into the future and not at all beyond eight days. In accordance with the law of gravity, it is possible for water to flow to the South Australian border from other parts of the river system. The primary storage of operational water is Hume Reservoir. It generally takes 25-26 days for water to flow from Hume Reservoir to the South Australian border. Another potential source of water is the Menindee Lakes from which it takes approximately 14 days for water to flow to the South Australian border. The water in the Menindee Lakes is not always available for the purpose of supplying South Australia’s entitlement and the volume of water which can be transferred from that source is significantly less than from Hume Reservoir. The last storage before the South Australian border is Lake Victoria from which it generally takes one day for water to flow to the border. Evaporation rates are materially higher at Lake Victoria than at Hume Reservoir. There is therefore an efficiency cost in transferring bulk water from Hume Reservoir to Lake Victoria for the purpose of having enough water waiting closer to the border to meet South Australia’s entitlements over upcoming months.
It follows from the uncertainties and imprecisions in these aspects of the river system that meeting South Australia’s monthly entitlement involves inescapable and irreducible risks. The evidence refers to two risks. First, there is the risk of system shortfall which is the risk that there will not be enough water in the river and the storages close enough to the border to ensure that South Australia receives next month’s entitlement under cl 88. Secondly, there is the risk of spill which is the risk that there will be too much water in the river and not enough spare capacity in the storages (particularly Lake Victoria) to ensure that water in excess of South Australia’s entitlement for the next month does not flow across the border.
The risk of system shortfall and the risk of spill are in direct conflict. Whilst the quantity of water currently stored in Lake Victoria may mean that there is a very low risk that South Australia will not receive its monthly entitlement in the immediately upcoming month or months, the risk of system shortfall increases for months further into the future. Subject to what happens with rainfall across the river system, it may be possible to keep the risk of system shortfall at virtually zero by regularly transferring water from Hume Reservoir to ensure that Lake Victoria is at a constantly high level. However, unforeseen and unforeseeable rainfall may occur after water has left Hume Reservoir but before it arrives at Lake Victoria with the result that there is not enough capacity to store all the excess water in Lake Victoria before it can be released to South Australia in accordance with future monthly entitlements. Therefore a measure taken to reduce the risk of system shortfall may increase the risk of spill.
In view of the fact that the quantity of water South Australia is entitled to receive in any month is not reduced if it has received a greater quantity than it is entitled to receive in a previous month (cl 93), and New South Wales and Victoria are entitled to share the water in the river system after South Australia has received its entitlement (cl 94), the three States have conflicting interests in relation to the balance between the risk of system shortfall and the risk of spill. There may be costs to New South Wales and Victoria from reducing the risk of system shortfall for the benefit of South Australia. Those costs may not be limited to running an increased risk of spill. In particular, transfers of water from Hume Reservoir result in material reductions in the water in the river system available for allocation to New South Wales and Victoria through conveyance losses and increased evaporation.
The foregoing description is a simplification. There are other aspects of operating the river system which directly affect the receipt of South Australia’s monthly entitlement. Access to the water in the Menindee Lakes is subject to minimum storage levels which have been set partly in response to concerns for local areas of human habitation. There are four lakes in the Menindee Lake system and there are different considerations to take into account as between releasing water from each. The storage level at Lake Victoria is subject to some seasonal limits which have been imposed to serve local Indigenous historic and cultural interests in the stability of the lake foreshore. There are channel capacity restraints between Hume Reservoir and the South Australian border which have to be accommodated, sometimes in coordination with holders of environmental water entitlements. Inter-valley trading accounts may provide alternative sources of water with different lead times but they require accounting adjustments, possibly between New South Wales and Victoria. These matters may affect the balance between the risk of system shortfall and the risk of spill.
The fundamental challenge which confronted the Defendant each year is apparent from the following high-level description of its work taken from the Affidavit dated 30 June 2022 by Mr Bishop:

It is against that practical background that the full terms of the Agreement are to be considered and in particular the statement in cl 88 that South Australia is entitled to receive the quantity of water specified for each month. There is no doubt that the Defendant has a role in South Australia’s receipt of its entitlements. Although not a party to the Agreement, the Defendant must comply with any requirements of the Agreement when performing its functions and duties and exercising its powers: s 18E(2) of the Commonwealth Statute.
As already observed, cl 29(1)(b) of the Agreement specifies that one of the Defendant’s functions is to give effect to high-level decisions of the BOC in relation to river operations. The BOC’s high-level decisions were found in the O&O Document. The O&O Document expressly allowed for the uncertain environment in which the river system is operated. That was apparent in a number of places in the O&O Document. The Defendant’s first general objective in cl 4(2)(a)(i) was to operate the Murray River system “efficiently and effectively” in order to deliver State water entitlements. That objective was internally contradictory to the extent that it applied to South Australia’s entitlement under cl 88 of the Agreement. The requirement “effectively” to deliver South Australia’s water entitlement each and every month may have made it necessary to undertake transfers of water from Hume Reservoir which were not “efficient” compared to continuing to store the water in Hume Reservoir. Such transfers may have increased the risk of spill if, for example, it rained before the water arrived at Lake Victoria. More efficient river operations may have reduced the risk of spill but been less effective to deliver South Australia’s water entitlement.
The transfer of water from Hume Reservoir to Lake Victoria was the subject of specific objectives and outcomes specified in Appendix A to the O&O Document. The relevant specification was as follows:

These provisions provided the Defendant with a more detailed framework within which to perform the task of delivering South Australia’s monthly entitlement by transferring water from Hume Reservoir to Lake Victoria close to the border. Nonetheless the specific objective in cl 12.5(b) and the specific outcome in cl 12.5(c) each required a balance, particularly between “conserving” and “contributing”. The matters noted in the interpretation clause in cl 12.5(d) included some subjective and imprecise language, such as “maximising water availability”, “minimising undesirable losses”, “as long as reasonably possible”, “enough water in transit”, “continue to contribute” and “when it becomes clear”. This language, together with the subject matter being addressed, reveals an intention that there be an element of judgement exercised by the Defendant when performing its functions.
Further, as already observed, the specific objectives and outcomes in cl 12.5 of Annexure A were expressly tempered by reference to the uncertain environment in which the Defendant operated. Clause 6 of the O&O Document is set out above. It referred to “high variability and uncertainty” and provided that the Defendant was to do “only what is reasonably practicable in all the circumstances”. Specific uncertainties were listed of which weather conditions, rainfall and evaporation rates were the first.
If, after considering the cl 6 uncertainties, the Defendant concluded that it may be unable to achieve an objective or outcome (cl 7) or it considered that there may be a conflict (cl 8) the Defendant had to refer the matter to the WLWG. If so advised by the WLWG, the Defendant then had to refer the matter to the BOC after consultation with the RMOC. These mechanisms allowed for the possibility that a matter may not be referred to the BOC but left with the Defendant (and the WLWG) even though the Defendant had concluded that it may be unable to achieve an objective or outcome or that there may be a conflict. It would depend upon the advice received from the WLWG. The O&O Document did not address the circumstances in which the WLWG would or would not advise the Defendant to refer a matter to the BOC.
The annual report which the Defendant was required to prepare under cl 12(2) of the O&O Document required it to indicate “the degree to which any general or Specific Objectives and Outcomes specified under cll 4 or 5 were achieved by the [Defendant]” which suggests that a degree of non-achievement was expected.
Despite the mandatory language in cl 88 of the Agreement, the O&O Document read as whole did not require the Defendant to ensure that South Australia received its water entitlement each and every month, but only to do that which was reasonably practicable. The O&O Document did not operate to remove all uncertainties from the delivery of South Australia’s water entitlement. This is unsurprising given the nature of the task. The Defendant’s obligation to give effect to the BOC’s high-level decisions in the O&O Document was attended by risks and, with those risks, conflicts between the interests of the States. It was contemplated that there may be circumstances where conflicts would be left with the Defendant to address with the WLWG. There may be occasions when there would be departures from the objectives and outcomes in the O&O Document.
The Defendant submits that the imposition of a duty of care to avoid economic loss to the Plaintiffs and Group Members as a result of inefficient operations of the river would conflict with the Defendant’s objective of operating the Murray River system efficiently and effectively in order to deliver South Australia’s water entitlement. I accept that submission. Such a private duty would introduce into the Defendant’s role of independently balancing the risk of system shortfall and the risk of spill across the South Australian border a concern for the individual holders of water entitlements in some (but not all) parts of New South Wales and Victoria. At the very least, the duty would distort the independent role which the Defendant is intended to perform in executing the provisions of the Agreement concerning sharing water between the States.
Indeed, the conflict runs deeper. If, for the reasons given by the Plaintiffs, the Defendant owed a duty of care to the holders of water entitlements in some parts of New South Wales and Victoria, there is no apparent reason why the Defendant would not also have owed a duty of care to the holders of water entitlements in South Australia. The holders of South Australian water entitlements may have been prejudiced if the Defendant did not effectively deliver one or more of South Australia’s monthly entitlement under cl 88 of the Agreement. The Defendant’s concern for efficiency in New South Wales and Victoria might have been at the expense of effectiveness in South Australia, and vice versa.
The conflict arising from the competing risks entailed in delivering South Australia’s monthly entitlement under cl 88 is not theoretical. It is directly relevant to the events which occurred in 2017 and 2018 and which have given rise to the claims in these proceedings. As is apparent from the consideration below of the particular allegations of negligence, at the most fundamental level the decisions made by the Defendant which are now the subject of the Plaintiffs’ complaints were made in the context of the Defendant having to balance the risk of system shortfall in the delivery of South Australia’s entitlements and the risk of spill across the South Australian border. The statutory framework contemplates that the Defendant would make those decisions free of a concern for the private interests of any particular individual or class of individuals.

Whilst the Plaintiffs’ case focusses on particular decisions in the course of the Defendant performing its cl 29(1)(b) function, most specifically exercises of the release power at Yarrawonga Weir in October 2017 and June 2018, regard must be had to all the Defendant’s functions and powers when considering whether the alleged duty of care was owed. [116] If a duty of care was owed, the Defendant would have been required to fulfil it prospectively.
There is evidence of the Defendant actually making decisions which addressed the interests of other persons at the expense of the Plaintiffs. As explained in more detail below, on 4 October 2017 the Defendant directed a release of water from the Menindee Lakes so that South Australia’s monthly entitlement would in part be supplied from that source. WaterNSW implemented the release on 7 October 2017. The evidence shows that it would have been possible for the Defendant to direct the release at an earlier date and at a greater rate of increase. Had that happened, the need to transfer water from Hume Reservoir to Lake Victoria may have been reduced with a consequential reduction in losses to the water in the river system ultimately available for allocation to New South Wales and Victoria (and thence to water entitlement holders such as the Plaintiffs). In choosing the commencement date for the release and the rate of increase the Defendant had regard to the interests of irrigators on the lower Darling River (not the Plaintiffs) and other community interests around the Menindee Lakes. The commencement of the release was delayed to allow lower Darling River irrigators to move pumping equipment to higher ground ahead of the Darling River rising. There may also have been some river crossings which would be cut by the rising waters which would cause problems for the locals unless they were given time to prepare. The rate of release conformed to advice from the WLWG on 26 September 2017 to “balance the requirement to maximise resource availability, local community sensitivities and concerns and environmental needs”. The commencement date of 7 October 2017 and the rate of increase reflected a balance between conflicting interests.
Another instance of actual conflict arose not from geography nor from the use to which different entitlement holders put water allocations, but from temporal considerations. The negotiations in 2018 for access to the MIL infrastructure are addressed below. The Plaintiffs submit that the alleged duty of care obliged the Defendant to achieve an outcome from the negotiations which was for the immediate benefit of the holders of water entitlements at the time the negotiations occurred (and on behalf of whom the Plaintiffs now sue). As will be seen, such a duty may have conflicted with the interests of holders of water entitlements in the future in achieving a negotiated outcome which provided long-term benefits for the efficient operation of the Murray River.
Another instance of actual conflict is apparent from the Plaintiffs’ own allegations of breaches of the alleged duty of care. The particular allegation has now been abandoned, but the Plaintiffs originally pleaded that the Defendant breached the alleged duty of care in September 2017 and July/August 2018 when it released water from other sources to “top up” the quantity of water from the Goulburn River which had been ordered by environmental water entitlement holders. The top up was necessary because the quantity ordered by the environmental water holders would not be diminished by conveyance and other losses by the time the water arrived at the South Australian border. [117] The releases from other sources reduced the quantity of water available for allocation to New South Wales (and thence to the Plaintiffs). At the heart of this alleged breach is the contention that the Defendant ought to have preferred the interests of the Plaintiffs to those of the environmental water holders. Whether or not the Plaintiffs ought to have been preferred was a matter for the hierarchy of decision-making established by the statutory framework and not for the imposition of a common law duty of care owed to some water entitlement holders and not to others. The fact that this particular allegation of breach is not pressed does not detract from the fact that, had the Defendant owed the duty of care for which the Plaintiffs still contend, the Defendant would have had to take the duty into account when making decisions like those made in September 2017 and July/August 2018.
Another instance of potential conflict exists by reason of restrictions on trade of water from above the Barmah Choke to below. As explained more fully below, such trade is generally prohibited because of the capacity constraints of the Choke, but the prohibition can be relaxed by the Defendant in exceptional circumstances. Prohibiting trade is intimately connected with exercising the release power because the prohibition is intended to enable the Defendant to release more operational water from Hume Reservoir and Yarrawonga Weir for transfer to Lake Victoria. Relaxation of the trade restriction would likely be in the interests of water entitlement holders located above the Choke because it would permit them to sell to a larger market, and also to those below the Choke looking to buy. On the other hand, it may not be in the interests of water entitlement holders generally if it causes greater losses to the water in the river system for allocation to New South Wales and Victoria making it necessary inefficiently to transfer water through the Choke in excess of channel capacity. In exceptional circumstances a decision about relaxing the restriction lies with the Defendant. Imposing the duty of care for which the Plaintiffs contend would require the Defendant to take the economic interests of the Plaintiffs into account when making the decision.
A further source of potential conflict arises from cl 90 of the Agreement which provides as follows:

(a) water in the Menindee Lakes may be drawn for the benefit of lower Darling irrigators but at the expense of the town water supply to Broken Hill; (b) generally releasing water from the Menindee Lakes instead of Hume Reservoir may benefit Murray River irrigators at the expense of lower Darling irrigators; (c) operating decisions by the Defendant may result in water being reallocated between New South Wales and Victoria under cl 116 of the Agreement – an example of this is where the Defendant decides to call on water from the Goulburn IVT account (100% for Victoria’s account) instead of from Hume Reservoir (shared between each State’s account) for the purposes of filling Lake Victoria, with the result that Victoria’s share of the water in Lake Victoria will be higher and may breach the s 116 threshold for a reallocation to New South Wales (were such a situation to occur there may be a conflict within the class constituted by the Group Members in these proceedings which includes water entitlement holders in both New South Wales and Victoria) – as recorded in the minutes of the WLWG meeting on 10 October 2017, an instance of this appears to have occurred in October 2017 when the timing of the Defendant calling on the Goulburn IVT account (and hence the quantity and timing of the water being transferred through the Barmah Choke to Lake Victoria) was influenced by the risk of an “internal spill” in the accounts of New South Wales and Victoria (see below); (d) there may be competition for channel capacity in the lower Goulburn River between water called from the Goulburn IVT account for operational purposes by the Defendant and water which environmental water holders wish to release from the Goulburn River – were that to occur the Defendant may have to choose between reducing the quantity of water flowing out of the Goulburn River for environmental purposes and increasing the quantity of water flowing through the Barmah Choke in order to have sufficient water arrive at Lake Victoria in time to ensure that South Australia receives its upcoming monthly entitlement; (e) the Defendant’s operating decisions earlier in the water year may subsequently give rise to seasonal competition for channel capacity between the Defendant’s need to transfer operational water and environmental water entitlement holders’ desire to use their water at an environmentally optimal time; (f) the different routes by which water may be transferred through the Barmah-Millewa Forest have competing merits, namely ecological benefits on the one hand and efficiency of return flows for further use of the water downstream on the other; and (g) different operating decisions by the Defendant today may have different impacts on the erosion of the Barmah Choke with potential consequences for the Defendant’s options for performing its functions in the future and also for environmental water holders to use water for the benefit of the environment – this is another example of a possible temporal conflict between the interests of the current holders of water entitlements and those who come to hold them in the future.

Another aspect of the statutory framework which may be relevant to the imposition of a duty of care is the provision expressly made (or not made) for the accountability of the Defendant. The relevant provisions are set out above. The Commonwealth Statute and the Agreement provide for review of the Basin Plan but do not specifically address accountability by the Defendant for the performance of its functions.
Clause 4(6)(a) of the O&O Document specified general objectives in relation to communication and information management, one of which was that the Defendant be accountable for its actions in accordance with the Agreement. The general outcome which was relevant to that objective was for the Defendant to provide appropriate, timely and accurate information about its river operations. The information was to be provided to the Ministerial Council, the BOC, the RMOC, the WLWG, other relevant committees, other stakeholders with an interest in the Defendant’s river operations and the public. The breadth of those to whom the information was to be provided, including the public, indicates that the kind of accountability contemplated by the O&O Document was not legal accountability. Legal accountability specifically to the Plaintiffs might be said to be incongruous with the objective of accountability “in accordance with the Agreement” because the Agreement is expressed in terms of obligations to the States and not to individual water entitlement holders.
The Defendant’s objective in cl 4(4)(a)(ii) was to contribute to the economic activities and values of people using the Murray River system. The related outcome was the maintenance of “productive relationships” with such people, which falls well short of the Defendant having a legal obligation to take reasonable care to protect them from economic loss. To some extent that may be contrasted with the Defendant’s objective in cl 4(4)(a)(i) which was to contribute to the safety of communities along the River Murray and the related outcomes of limiting damage to downstream communities when managing flooding and mitigating events that may adversely affect the quality of water available for urban, irrigation, industrial, environmental, recreational and stock and domestic use. In any event, there was no reference in the O&O Document to recompense or compensation for any person affected by the Defendant’s river operations.
The O&O Document provided both for annual self-assessment of the Defendant’s river operations (cl 12(2)) and for annual review by an independent committee, albeit one comprised of experts appointed by the Defendant (cl 12(3)). Whilst the efficacy of the accountability which might be achieved by those two mechanisms might be questioned, the mechanisms were approved by the BOC for whose ultimate benefit the reviews would be undertaken. Clauses 12(2) and (3) were expressed in constructive language. In particular, cl 12(2)(b)(ii) required the Defendant’s report to “indicate the degree to which any general and Specific Objectives and Outcomes specified under clauses 4 or 5 were achieved by the [Defendant]” which indicates that some level of non-achievement was expected. Clauses 12(2) and (3) suggest that the purpose of the reviews was to improve future performance rather than to allocate blame for past performance. That purpose might have been undermined by a common law duty of care if it discouraged a full and frank assessment of the previous year. On the other hand, it might be said that a duty to act with reasonable care could reinforce the goal of improvement.
There are aspects of the express measures for accountability which are in tension with an intention that the Defendant owe a common law duty of care to the Plaintiffs, especially the express focus on accountability to the States. The focus on improvement for the future is reinforced by the imperfect knowledge of the river system upon which the statutory framework was premised (see below). On the other hand, the Defendant was not intended to undertake an inscrutable task behind closed doors. On the whole, the fact that some measures of accountability have been put in place is a largely neutral consideration, but the specified measures do not positively require the imposition of a duty of care.

One further observation about the statutory framework is that it proceeds from the clear premise that no one, including the Defendant, has complete knowledge of all aspects of the Murray-Darling Basin and the Murray River system. Ongoing research is expected and the Defendant is given a central role in undertaking that research. One of the objects of the Commonwealth Statute is the collection and analysis of information about Australia’s water resources and the use and management of water in Australia. The functions of the Defendant include conducting research and investigations into the Basin water resources and developing measures for the equitable, efficient and sustainable use of the Basin water resources. Now that it has been prepared, the Defendant has a role in monitoring and evaluating the Basin Plan. Subdivision G of Division 1 of Part 2 of the Commonwealth Statute provides for the regular periodic review of the Basin Plan in the future. Subdivision H provides for an annual review of the Basin Plan’s effectiveness. Part VII of the Agreement provides for the Defendant to carry out surveys, investigations and studies regarding the desirability and practicality of works or measures for the equitable, efficient and sustainable use of the water of the Murray-Darling Basin. The Defendant’s general objective in cl 4(5)(b)(iii) of the O&O Document includes improvement of the knowledge and practices of environmental watering. These are examples of many such provisions which contemplate a continuous process of developing the management of the Murray-Darling Basin.
The matters which are expected to be the subject of research by the Defendant include operational matters. Of importance to the present case is the Defendant’s function “to develop, in consultation with the Basin States, an integrated water model for the Murray-Darling Basin”. [119] The word “develop” suggests an ongoing process. Modelling is critical to the operation of a river like the Murray River. Complicated models are required which, because of the unique attributes of the Murray-Darling Basin, cannot be obtained but must be developed.
This observation about the nature of the Murray River is pertinent for three reasons. First, it is another aspect of the dynamic, uncertain and risky environment in which the Defendant was established to function. Secondly, it supports the conclusion that the Defendant is not intended to be accountable in a way which inhibits future learning and improvement. Thirdly, it is independently relevant to the question whether the Commonwealth Parliament and the States intended that the Defendant be subject to a common law duty of care. The fact that the Defendant must perform its functions without perfect knowledge does not mean that there are no aspects of its activities which may be subject to a duty of care. The duty for which the Plaintiffs contend is a duty to take only that care which is reasonable. Many professionals, including medical professionals, are required to give advice and undertake procedures with imperfect knowledge about the condition they are treating. Nonetheless, it is another feature of the statutory framework which is in tension with the broad duty of care alleged by the Plaintiffs in paragraph 86 of the Statement of Claim.

There is some artificiality about finding an undertaking by an entity which is a creature of the same statutory regime from which the undertaking is said to be implied. It may be said that responsibility is not so much assumed by the statutory entity as thrust upon it. Nonetheless, there may be an assumption of responsibility as an incidence of the statutory office. [120] Albeit in a property damage case, the High Court has referred to a statutory authority entering into the exercise of its statutory powers as the occasion for a duty of care to arise. [121]
The difficulty which the Plaintiffs confront with the contention that there has been an assumption of responsibility by the Defendant, and which they have not overcome, is apparent from the formulation of the Defendant’s express or implied undertaking which the Plaintiffs submit the Court ought to find. They contend that there was an undertaking by the Defendant to undertake its river operations functions “with due care inter alia to protect the Plaintiffs and Group Members”. [122] The Plaintiffs have stopped short of contending that the Defendant undertook to conduct river operations for the Plaintiffs and Group Members. That is understandable. The Defendant is established to operate the Murray River system in the public interest. There are many different interested persons, including the Australian nation as signatory to some international treaties. The statutory framework is centred on the interests of the States. If for anyone, the Defendant has undertaken to conduct river operations for the public or for the States. So much follows from the terms of the statutory framework set out above.
The plurality said in the Mallonland case that the best understanding of assumption of responsibility requires the defendant to undertake the relevant task for the plaintiff. [123] Whilst the words of a judgment are not to be construed like a statute, there is a material difference between what the Plaintiffs submit and what the plurality said is required. An undertaking to perform a task for another person impresses the performance of the task with the objective of ensuring that person’s interests rather than merely recognising the other person’s interests as an incident of performing the task for some other objective. It is a defendant’s undertaking to perform the task for the plaintiff which gives rise to the relationship which warrants the imposition of a duty of care. [124] The essential relationship is absent when the defendant does not undertake to perform the task for the plaintiff, even if an incident of the defendant’s undertaking to perform the task for someone else is that it will do with due care to protect others, including the plaintiff.

There is no support in the statutory framework for the conclusion that the Defendant undertook the performance of its river operations functions for the Plaintiffs and Group Members as a class which is distinct from the world at large. The necessary relationship would have been incongruous with the Defendant performing its functions in the public interest by delivering the water entitlements of the States in accordance with their agreed sharing arrangement. These conclusions are demonstrated by specific reference to the function of the Defendant to which the Plaintiffs contend the duty of care attached, namely the function specified in cl 29(1)(b) of the Agreement. It cannot be said that giving effect to high-level decisions of the BOC in relation to river operations was a function which the Defendant undertook for the Plaintiffs and Group Members. The term “river operations” is defined in cl 2 of the Agreement which is set out at [238] above. Each part of the definition directs attention to the States as the entities which own and operate the works on and adjacent to or connected to the river and the entities to which the Defendant provides services. In particular, river operations includes the Defendant’s activities relating to “the execution of the provisions of the Agreement concerning sharing water between the [States]”.
Contemporaneous communications in which the Defendant described its awareness of and responsibilities to groups which include the Plaintiffs are not capable of contradicting the intention of the statutory framework, objectively ascertained. That is especially so when there is no allegation that any of the Plaintiffs relied on any such communication. The meagre evidence of the Defendant’s two internal documents, one a draft reference manual and the other an email by an employee, and the subjective views of the Defendant’s witnesses in these proceedings do not take the Plaintiffs’ case any further. A description of a process by which a State entity collates water orders from individual entitlement holders and passes them on to the Defendant does not accurately represent the framework established by the Commonwealth Parliament and the States’ Agreement.
The fact that irrigators in New South Wales and Victoria paid charges by reference to the Defendant’s operating costs does not support the Plaintiffs’ case when the evidence is examined. On the contrary, it further demonstrates that the relationship erected by the statutory framework is between the Defendant and the States and that there is no direct relationship between the Defendant and the Plaintiffs and Group Members. The Commonwealth pays the Defendant’s costs under s 174 of the Commonwealth Statute. Under cl 145 of the Agreement the States have agreed effectively to reimburse the Commonwealth. The evidence about the practical payment of the costs is limited, but in New South Wales it appears that the State pays its share to the Commonwealth from consolidated revenue. Since 2014 the New South Wales Treasurer has required WaterNSW to pay a dividend to the State in an amount equal to the amount paid for New South Wales’s share of the Defendant’s costs. The charges subsequently made to individual water entitlement holders were levied by WaterNSW (not the State) for the purpose of funding the dividend. There is no suggestion that the Defendant had any role in recovering its own costs or any say in how charges made by WaterNSW were calculated or described. There is no evidence that the Plaintiffs or Group Members paid any charges directly to the Defendant. The Plaintiffs and Group Members are not in any legal or common sense the Defendant’s customers.

There is a threshold dispute between the parties in this part of the case about the precise identity of the risks of harm. The Plaintiffs plead the risks of harm in paragraphs 77 to 78 of the Statement of Claim which are set out above. The burden of the pleading is that the risks were sourced in a reduced supply of irrigation water from the Murray River. For its part, the Defendant says that the source of the risks was insufficient (or less) water for the conduct of irrigated agricultural production for which rainfall is overwhelmingly the most significant factor. A similar argument was initially advanced by the flood engineers in the Queensland Floods case but (“wisely”) not persisted with. [125]
The Defendant’s submissions about the identity of the risks of harm ought to be rejected. They do not sufficiently reflect the nature of irrigated agriculture and the roles respectively played by the parties in that field of human activity. It may be accepted that the availability of water is a fundamental aspect of agricultural production and that rainfall is overwhelmingly the most significant factor. However, the purpose of irrigation is to make available water which would not otherwise be supplied by rainfall. Irrigation is one of the measures available to farmers to manage the risks of having insufficient (or less) water. By these proceedings the Plaintiffs seek to complain about interference with their interests in being able to use that management measure. They do not complain about the condition which it is the purpose of the measure to manage.
In any event, the issue is not to be resolved by the Defendant’s abstract references to agricultural operations. The Plaintiffs have pleaded their case by reference to interference with their interest in the supply of irrigation water from the Murray River. They are entitled to bring that case. If otherwise appropriate, the common law may impose a duty of care to protect the Plaintiffs’ interests in being able to use irrigation from the Murray River to make available water which would not otherwise be supplied by rainfall. The Defendant’s submissions which focus on rainfall as the primary source of water for agriculture are addressed to a case not brought by the Plaintiffs.
Once the risks of harm have been identified it is possible to analyse the salient features of the case. [126]

As a general principle the plaintiff’s vulnerability is likely to be the corollary of the defendant’s control. [127] It is so in the present case. The Plaintiffs and Group Members were unable to protect themselves from interference with their interest in receiving irrigation water from the Murray River in order to manage the risks of having insufficient (or less) water for their agricultural production. The Defendant adduced a body of evidence about measures which are sometimes available to agricultural enterprises to manage the risks of low rainfall. In the context of considering the duty of care, the fact that there were a number of measures other than irrigation to manage the risk of insufficient (or less) water is no answer to the vulnerability of the Plaintiffs and Group Members to loss of one measure in particular.
In one sense the Plaintiffs’ and Group Members’ vulnerability was no different to that of any other user of water from the Murray River system. A reduced allocation to New South Wales and Victoria would result in a reduced allocation to all holders of water entitlements, including environmental water holders. Although it might not be measured in dollars, a reduced water allocation to environmental water holders had value.
However, there is an aspect of the vulnerability of the Plaintiffs and Group Members which distinguished it from the vulnerability of the general public to inefficient exercises of the release power by the Defendant. Unlike environmental water entitlement holders and some other people interested in the waters of the Murray River system, the Plaintiffs and Group Members were vulnerable to a particular type of harm, namely pure economic loss.
The Plaintiffs and Group Members were in a similar position to irrigators in South Australia whose agricultural production might also be affected by the way in which the Defendant performed its river operations functions. Were they to be adversely affected, they too would suffer economic loss. However, the positions are not identical. Given cl 88 of the Agreement, South Australia would receive its entitlements unless there was inefficiency of sufficient magnitude to result in a system shortfall. In the language of cl 4(2)(i) of the O&O Document, South Australian irrigators were vulnerable not so much to the Defendant being inefficient as ineffective. South Australian irrigators were vulnerable to a different sort of risk than that to which irrigators in New South Wales and Victoria were exposed.

The first issue is whether the duty of care alleged by the Plaintiffs would give rise to an indeterminate liability on the part of the Defendant. For the reasons correctly submitted by the Plaintiffs, the answer to that question is no. In its terms, the alleged duty would have been owed to the Plaintiffs and Group Members. The definition of Group Members is multifaceted. Not only does it include persons who directly held water entitlements of specified classes within a specified date range, but also third parties who “held” water entitlements under contractual arrangements plus third parties who “utilised” allocations of water under entitlements held by others. The inclusion of third parties may make the membership of the class unwieldy but it does not make it unascertainable. At any given time during the specified date range the total quantity of water entitlements was fixed and ascertainable. Just by way of example, the evidence shows that in the 2018/2019 water year the total general security water entitlements issued in New South Wales for the lower Murray region entitled the holders to a share of 373 GL, of which 120 GL were held by environmental water holders and 253 GL were held by others. The total quantity of water entitlements operates as a definite limitation on the class just as the geographic location and elevation of the flooded properties did in the Queensland Floods case. In these proceedings the third parties are evidently included as Group Members not to increase or duplicate the members of the class but more accurately to identify the entities which were financially affected by the water allocations which occurred. It is true that the nature and extent of the loss suffered by each Group Member will be difficult to calculate as a whole, but there are likely to be sufficient parameters to govern the calculations so as to ensure that the quantum of the Defendant’s liability is not unascertainable.
The fact that the alleged duty of care does not strictly speaking suffer from indeterminacy does not improve the case for imposing the duty of care. It merely excludes a reason for which a duty of care would generally be rejected. [128]

The second issue which has been raised in this area is related but different. If the Defendant is held to owe a duty of care to protect the Plaintiffs and Group Members from economic loss, why would it not also owe an equivalent duty of care to others? A decision that a duty of care is owed to some who suffer financial harm as a consequence of negligent conduct but not to others is only just if it is capable of being explained on a rational basis. [129] On the Plaintiffs’ reasoning, there would seem to be no reason why a duty of care to avoid pure economic loss would not also be owed to South Australian irrigators, and also to lower Darling irrigators. Further, the Plaintiffs allege that one consequence of the Group Members receiving lower allocations was that the price of temporary water was higher during the relevant period. They claim to be entitled to compensation for loss caused by higher water prices. If the Group Members are entitled to compensation for that loss, why not also other farmers who did not hold water entitlements but who were nonetheless affected by higher water prices? And why ought not there be compensation for entities like superannuation funds and private investors who did not undertake agricultural activities but who traded in water and water entitlements if they suffered a financial loss? There is also the question of the Group Members’ customers and other suppliers, and other businesses more broadly affected by the financial fortunes of Murray River irrigated agriculture. The statutory framework within which the Defendant operates has a wide focus, concerned as it is with the public interest and the “net economic returns to the Australian community”.
It would not automatically follow from the imposition of the duty of care in favour of the Plaintiffs and the Group Members that the Defendant owed duties of care to other individuals or groups. The claim by each individual or group would have separately to be examined in accordance with the principles set out above. It may be that some other duties are precluded by some of the limiting factors such as reasonable foreseeability. Having said that, there does seem to be an acceptance by the Plaintiffs that the Defendant would have owed duties of care to others. The Plaintiffs submit that there was an assumption of responsibility by the Defendant because it expressly or impliedly undertook to perform its river operations functions with due care “inter alia” to protect irrigators against the risk of economic harm that arises from inefficient river operations.
The prospect of duties of care being widely owed to others is not strictly a question of indeterminacy because the duty alleged to be owed to the Plaintiffs and Group Members will not itself give rise to an unascertainable liability. The duty of care owed to the Plaintiffs and Group Members is not to be rejected simply because duties of care owed to others may in combination give rise to indeterminant liability. [130]
The prospect that duties of care would be widely owed to others is nonetheless relevant. The prospect that the Defendant must also have owed duties of care more widely underlines the incongruity between the statutory framework given the conflicted environment in which the Defendant was established to function. A concrete example of this is the conflict which would potentially arise if equivalent duties of care were owed to South Australian irrigators and lower Darling irrigators. There can be no rational basis to explain why the duty of care would be owed to one such group and not others in which case conflicting duties seem inevitable.
The prospect that duties of care may be owed more widely also detracts from the Plaintiffs’ contention that there has been an assumption of responsibility. It is more difficult to find that there has been an undertaking by the Defendant to take on the task of operating the river for the Plaintiffs and Group Members when there is no rational reason then to conclude that there have not been undertakings to take on the same task for other groups, including groups with potentially conflicting interests.
The prospect of the Defendant owing duties of care to other groups beside the Plaintiff and Group Members reinforces the conclusion that no duty of care was owed to the Plaintiffs and Group Members.

The first case is the Queensland Floods case. The Plaintiffs naturally emphasise the similar factual matrix and the conclusion that the operator of the two Queensland dams owed a duty of care to people downstream to avoid flood damage to their property. The defendant, a statutory authority, had control over the flow of water in the river, which meant that it had significant control over the risk of flooding. The people with property downstream were relevantly vulnerable. To the defendant’s knowledge, the risk of flooding was foreseeable. The class of persons to whom the duty was owed was very large but not indeterminant. The duty of care was found not to be inconsistent with the statutory regime within which the defendant operated the dams. There was no appeal from the trial judge’s finding that a duty of care was owed. [133]
I accept the Defendant’s submission that any analogy which might otherwise have been drawn from the Queensland Floods case is undermined by the key fact that the risk of harm which the flood engineers were found to have a duty to avoid was damage to property. As stated above, the law will prima facie impose a duty of care if damage to property is reasonably foreseeable. The fact that the risk of harm was damage to property was central to the trial judge’s conclusion that a duty of care was owed in Queensland Floods case. It directly affected the analysis of each feature of the case. [134]
There are a number of other points of distinction which warrant emphasis. First, the relevant function of the defendant in the Queensland Floods case was flood mitigation, not water supply. The contrast between those two functions reflects the contrast between property damage and economic loss. It is one thing to find that a duty of care is owed to avoid damage to property when engaged in flood mitigation. It is another thing altogether to find that a duty of care is owed to avoid pure economic loss when engaged in the Defendant’s functions of operating the Murray River system efficiently and effectively in order to deliver State water entitlements.
Secondly, a prominent feature of the reasoning of the trial judge in the Queensland Floods case was the control which the defendant had over the source of the risk of harm and the vulnerability of the plaintiffs. [135] The prominence makes sense where the risk of harm was damage to property from flooding. When analysing these features of the case the trial judge considered analogies with cases about fire (Pyrenees Shire Council v Day), a collapsed bridge (Brodie v Singleton Shire Council), asbestos (Crimmins), shipwreck (Ibrahimi v Commonwealth), flooding (UCC v ESB) [136] and food poisoning (Graham Barclay Oysters). In the present case, I have found that the Defendant also had control over the source of the risks of harm and that the Plaintiffs were vulnerable. Those features are salient but they do not have the same prominence because there was no risk of damage to the Plaintiffs’ persons or property. A closer analogy is Southern Properties (WA) Pty Ltd v Executive Director of the Department of Conservation and Land Managemen**t [137] where an authority conducting a prescribed burn for the purpose of managing the risk of future bushfires was held not to owe a duty of care to protect local grape growers from the risk of smoke damage to the crops.
Thirdly, there is no comparison with the relationship erected by the statutory regime between the defendant and the plaintiffs in the Queensland Floods case. [138] The defendant was specifically engaged in flood mitigation and the plaintiffs were the downstream property owners who were at risk from the flooding. The relationship was direct, specific and, it might be said, the focus of the regulatory framework in which the defendant undertook its flood mitigation tasks. By contrast, the statutory framework in which the Defendant functions is centred on the interests of the States with which the Defendant has a direct relationship. The Plaintiffs and Group Members held water entitlements to be used for only one of several uses, and in one location only. They are not the focus of the statutory framework in the same way as were the plaintiffs in the Queensland Floods case.
Fourthly, and following from the previous point, given the plaintiffs’ more focussed and specific place in the regulatory framework, the imposition of a duty of care on the defendant to protect the plaintiffs from property damage was consistent with and not incongruous with the regulatory framework.138 [139] There was not the same potential for a distortion of the focus of the regulatory framework as there is in this case.
Fifthly, the location of the relevant property downstream of the dams meant that not only was the membership of the class to which the duty was owed in the Queensland Floods case ascertainable, there was no suggestion of equivalent duties of care being owed to other classes. [140] There was not the same potential for conflicting duties as there is in the present case, nor was there the potential for duties to be owed widely by the defendant.
Sixthly, the trial judge recognised in the Queensland Floods case that a duty of care might be owed even if the flood engineers were thereby exposed to some conflicting duties. Depending on the conflicts, they might be accommodated by informing the question whether the care which the flood engineers had taken was reasonable.. [141] In the Queensland Floods case the duty of care was owed by the flood engineers to all members of the class of downstream property owners, all of whom would benefit from flood mitigation being undertaken with reasonable care even though particular flood mitigation measures might cause damage to some who would not have suffered any damage if the flood mitigation was undertaken otherwise. [142] Some of the conflicts to which the Defendant is exposed in the present case are different. That is especially so for the conflicts which arise from the requirement that the Defendant balance the risk of system shortfall and the risk of spill in order to achieve the objective to operate the Murray River system efficiently and effectively in order to deliver State water entitlements.
Another significant point of difference with the Queensland Floods case is that the statutory framework within which the defendant operated was materially different. In particular, there was far greater prescription for the performance of flood mitigation to which the defendant’s duty of care attached, the prescription variously arising from the development consent for the dams, the conditions of the licensing of their operation and the Manual. [143] The tasks of the defendant more ready answered the description “operational” and was more apt for supervision by the Court. The contrast with the Defendant’s function to “give effect to any high level decision” of the BOC is stark.
Given the nature of the damage suffered by the plaintiffs and the different statutory framework, the Queensland Floods case is a poor analogy.

It was reasonably foreseeable that the Plaintiffs and Group Members may suffer pure economic loss if the Defendant did not undertake its river operations functions efficiently. That alone does not provide a basis for imposing the alleged duty of care on the Defendant.
In addition, the Defendant had actual knowledge of the risks of the economic harm alleged by the Plaintiffs, it relevantly exercised control over the source of the risks and the Plaintiffs were vulnerable because they were unable entirely to protect themselves from the risk of receiving a reduced supply of water from the Murray River with which to manage the risks of having insufficient (or less) water with which to conduct their irrigated agriculture activities. It may also be accepted that the Plaintiffs and Group Members were reliant on the Defendant to undertake river operations efficiently, although no more so than other groups.
It should also be accepted that a feature of the case is that the imposition of the alleged duty of care would not strictly speaking expose the Defendant to an indeterminate liability although that is not a reason for imposing the duty of care so much as the exclusion of a particular reason for rejecting it.
There have been cases where these features have been found to support the imposition of a duty of care to avoid pure economic loss to a plaintiff. However, the judicial evaluation of the features of a case is not an exercise in going through a check-list, nor is it an exercise of instinctive synthesis. These features mean that the imposition of a duty of care would not suffer from the problem that it might expose the Defendant to unforeseen and unforeseeable liability, nor to a liability arising from risks which have a source in something that the Defendant cannot control, nor to an indeterminant liability. However, the imposition of the alleged duty of care would nevertheless be incongruous with the statutory framework, at the very least by distorting the performance of the Defendant’s functions. The incongruity would arise not just from the Defendant’s potential liability to the Plaintiffs and Group Members but also from the Defendant’s potential liability to other individuals and groups, some of whom may have conflicting interests. The incongruity further arises from the difficulty in identifying the content of the duty with which the Defendant would be obliged to comply were it to be imposed. Were the duty of care to be imposed, the Defendant would be subject to the duty in circumstances where there has been no assumption of responsibility by the Defendant and where there has been no reliance by the Plaintiffs and Group Members different to the general reliance on a public entity to perform a public function. A judicial evaluation of all the features of the cases leads to the conclusion that the Defendant did not owe a duty of care to avoid pure economic loss to the Plaintiffs and Group Members.
There is nothing in the cases relied upon by the Plaintiffs which warrants a different conclusion. The Plaintiffs’ contention that they were owed the alleged duty of care is to be rejected. It follows that the proceedings are to be dismissed.

As previously noted, the Menindee Lakes are operated by WaterNSW.
The Menindee Lakes are a shared resource, in the sense that there are times when the Defendant has access to the water stored in the Menindee Lakes for the purposes of delivering State water entitlements and times when it does not. The sharing of the resource comes from cl 99(1) of the Agreement which provides that the Defendant may not direct that water be released from the Menindee Lakes Storage after its volume falls below 480 GL and before it next exceeds 640 GL. Outside those parameters the water belongs to New South Wales.
As already observed, cl 12.2 of Appendix A to the O&O Document provided that the Menindee Lakes had to be operated in harmony with Lake Victoria. Specifically, when the Defendant had access to water in the Menindee Lakes, bulk transfers of the water to Lake Victoria “should continue to be triggered” whenever Lake Victoria was below a specified level. The essential reason for that requirement was that evaporation losses from the Menindee Lakes are materially higher than those from Lake Victoria: cl 12.2(b). The Plaintiffs submit that the harmony operations balanced the advantages of reduced system losses from evaporation against the risk of spill from Lake Victoria into South Australia. That accords with the evidence.
For present purposes, if the Defendant had access to the Menindee Lakes in the month of October bulk transfers “should continue to be triggered” if Lake Victoria was below 680 GL. In fact, Lake Victoria’s maximum capacity is only 677 GL which means that in October bulk transfers from the Menindee Lakes were always triggered so long as there was more than 640 GL in the Menindee Lakes. They would continue to be triggered until the Menindee Lakes fell below 480 GL.
Another aspect of the river system is that the rate of evaporation from the Menindee Lakes is not only higher than Lake Victoria but also much higher than Hume Reservoir. All other things being equal, system losses can be reduced by transferring water to Lake Victoria from the Menindee Lakes in preference to water from Hume Reservoir. Clause 12.5(d) of the Annexure A to the O&O Document contained the following note:

For the purposes of preparing the AOP each year, the Defendant selected six hypothetical scenarios. As described in the 2017/2018 AOP Document, each scenario assumed a specific quantity of “annual River Murray system inflows”. The assumptions for each scenario were in fact more detailed, including assumptions about inflows for each month for each geographical part of the river system which altogether comprised an annual system inflow assumption.
Individual scenarios were (generally) designated by the Defendant by reference to actual inflows recorded for the previous 125 years and a calculation of annual exceedance probability. For example, one scenario was designated “very dry” for which it was assumed that annual inflows would have an annual exceedance probability of 95%. Thus, the “very dry” scenario was based on an assumption that inflows would occur in the current water year which would be less than the inflows that had actually occurred in 118 of the last 125 years. A scenario was designated “dry” which had an annual exceedance probability of 90%. A scenario was designated “moderate” which had an annual exceedance probability of 75%.
Designating scenarios by this method meant that there were at least 100 different scenarios available for selection. As each scenario included constituent sub-assumptions about the timing and location of the annual inflows across the entire river system, there were in fact many more different scenarios which could be selected.
From that range the Defendant selected six scenarios for the purposes of preparing the AOP. The Defendant did not select the same six scenarios each year. For each scenario of assumed inflows, the Defendant then made further assumptions about numerous other factors, such as conveyance (or transmission) losses (seepage and evaporation en route), South Australia’s storage rights, use of water for consumptive purposes and use of environmental water (as to quantity, timing, location and method). These are only examples of the assumptions made in the AOP. There are aspects of the Murray River system which are not directly touched by the issues in the present case but which were nonetheless part of modelling and about which assumptions were made and affected the output of the modelling, such as flows in the Murrumbidgee River tributary and water transferred from the Snowy Hydro Scheme.
The assumptions for each scenario were then input into the Fortran Assessment Model which is bespoke software developed by the Defendant and its predecessors to model the behaviour of the Murray River system. Rules to simulate the natural behaviour of the river system have been developed and coded within the model. The coding also includes rules designed to simulate decisions about how the river system will be operated, such as minimum release rates from various dams and weirs. In accordance with the coded rules and based on the input assumptions, the Fortran Assessment Model’s functionality includes identifying optimal river operations (including releases of water from the various storages) in order to meet demands (including delivery of South Australia’s water entitlement). [161]
The output of the Fortran Assessment Model is generated at twelve monthly intervals for each scenario, starting with June and ending with next May. Once the Fortran Assessment Model has been run and an output is produced, the output is reviewed by the Defendant’s River Operations Team and subjected to some manual adjustments to reflect certain operational matters which the Fortran Assessment Model does not take into account. Examples of matters which may be the subject of post-processing manual adjustments are environmental water demands and rules relating to releases of water from different parts of the Menindee Lakes. Another matter which may be subject to the post-processing manual adjustments is bulk transfers of operational water, including from Hume Reservoir to Lake Victoria and from the Menindee Lakes to Lake Victoria. Manual adjustments may be made to the rates of release of water at Yarrawonga Weir. Some of these manual adjustments involve an exercise of judgement by the River Operations Team. The judgement may be informed by a process of trial and error with the various assumptions.
The product of the entire process is then collated into a spreadsheet which is sometimes referred to as the “AOP Spreadsheet”. The Defendant then prepares a document which includes text and graphs to communicate the information in the AOP Spreadsheet. The document is sometimes referred to as the “AOP Document”. The terminology is not unimportant. For instance, cl 10(1)(d) of the O&O Document specified that the Defendant would carry out its functions “in accordance with the Annual Operating Plan”. The Defendant submits that cl 10(1)(d) should be construed as referring to the AOP Document alone, not the AOP Spreadsheet. Having regard to the nature of conducting river operations “in accordance with” the AOP (see below) there is no reason to limit cl 10(1)(d) in that way. The more efficacious and plausible meaning is that the Defendant would carry out its functions having regard to all available information including the information in the AOP Spreadsheet.
Dr Davis (Senior Director of the River Operations Team) described the entire process of preparing the AOP as “quite cumbersome”. I infer that is an example of Dr Davis’s understated turn of phrase. He said that the River Operations Team has to check all the assumptions and make sure everything is all right each time the process is undertaken. The Defendant’s practice was to undertake the process not only in July and October as required by the O&O Document but to repeat the process at regular (approximately monthly) intervals. Each time the process was repeated the River Operations Team would substitute such inputs as had been assumed on the previous occasion which were now known from more recently observed data. For example, when originally prepared and adopted in July, each scenario in the AOP included assumptions about inflows for all future months starting with August. At the end of August the Defendant could repeat the AOP process for all scenarios but modified to use the August inflows which had actually occurred in the meantime.
According to Mr Bishop (Senior River Operator) the purpose of repeating the process may simply be to update the information available to the Defendant from a more up-to-date AOP. Another purpose may be to use different assumptions and/or make different adjustments to compare outcomes. Repeatedly running the Fortran Assessment Model and comparing outputs could be used to assess risks or test options for future operational decisions. This is the trial and error process referred to above.
The parties do not agree on the meaning and effect of the text, numbers and graphs contained in the AOP Document and the AOP Spreadsheet. The issue is addressed below as part of the analysis of the Plaintiffs’ contention that the Defendant failed to act in accordance with the AOP in the 2017/2018 water year.

It is convenient at this point to describe the structure of the Defendant’s organisation which is relevant to this case.
As set out above, the Defendant has multiple functions, not all of which are relevant to the present case. The conduct of the Defendant which is relevant to the 2017/2018 water year was undertaken by the River Operations Team. The Executive Director of the Division which included the River Operations Team was Alan Dreverman. He was replaced by Mr Reynolds from 31 August 2017.
At the relevant time Dr Davis was the Senior Director of the River Operations Team. As Senior Director he had broad oversight of river operations and supervised the River Operations Team. The Senior Director also chaired meetings of the WLWG.
From 1 June 2017 the Director of the River Operations Team was Jacqueline Hickey. She was responsible for managing day-to-day river operations.
The day-to-day river operations are undertaken by river operators. At all relevant times Andrew Bishop and John Waterworth were Senior River Operators. The other River Operators included Thomas Zouch (who from February 2018 was an Assistant Director), William Lucardie (who was a “less experienced” member of the Team), Kerry Greenwood and Laura McCann (who was also a “less experienced” member of the Team).
Each week a river operator is designated as the Duty Operator. The Duty Operator is responsible for maintaining the Daily Operations Spreadsheet as described below. The Duty Operator is also responsible for communicating the Defendant’s directions to release water to the State water authority which operates the dam or weir in question.
Whilst a Duty Operator is designated each week, the evidence makes clear that the River Operations Team is aptly named. It is a “small team”. Ms Hickey said that it is made up of “collective experienced river operators”. The evidence of Dr Davis, Ms Hickey and Mr Bishop, supported by the contemporaneous documents, demonstrates that the work of the River Operations Team is collaborative. Some of the collaboration is structured, such as the weekly Team review meeting (usually on a Tuesday morning), but much of it is informal. Dr Davis saw Ms Hickey pretty much every day. Consultation within the Team is continuous, including when the Fortran Assessment Model is run and AOP Spreadsheets and AOP Documents are prepared, updated and analysed. A river operator may be responsible for preparing a particular document but more senior members of the Team, including Dr Davis and Mr Reynolds (and before him Mr Dreverman), reviewed all the documents which have become central to the present case. They did so multiple times. Where assumptions have to be adopted, statements have to be published or judgements have to be made, collective consideration is given to the task. Multiple members of the Team attended WLWG meetings and presented information, including all of the meetings which are relevant to this case.
The Plaintiffs have sought to paint a picture of the Duty Operator making operational decisions unilaterally and idiosyncratically, almost furtively, without input from or even the knowledge of other members of the River Operations Team let alone senior managers. The Plaintiffs submit that on 29 September 2017 the Duty Operator was “left to [her] own devices” to make decisions which were significant, unusual, risky and dangerous. That picture finds no support in the evidence.

The Fortran Assessment Model and the AOP are not the only tools used by the Defendant to make decisions about operating the Murray River system. The Daily Operations Spreadsheet is another tool.
The Daily Operations Spreadsheet is not produced using the Fortran Assessment Model. It is an excel spreadsheet in which formulae have been developed and embedded by the Defendant. It requires the input of observed data about the river system such as inflows, losses and storage levels. It also requires the input of assumptions about future matters, including uses (diversions) and inflows. Using historic data from (approximately) the last three months and assumed inputs for (approximately) the next three months, the Daily Operations Spreadsheet processes and depicts the behaviour of the river for the upcoming (approximately) three months.
As the name suggests, the Daily Operations Spreadsheet is maintained on a daily basis. Maintenance of it is part of the duties of the Duty Operator. It is updated each morning by replacing the assumed inputs from the preceding day with newly received observed data. The Defendant has adduced evidence about the updating process. It is a structured and semi-automated routine. The Defendant automatically receives observed data from the collection infrastructure generally maintained by State water authorities and the Bureau of Meteorology. The Duty Operator extracts certain information from a wider body of data and inputs it into the Daily Operations Spreadsheet. The new data is subjected to two layers of quality assurance by the Duty Operator. Every month or so the Daily Operations Spreadsheet is stepped forward a whole month by deleting the oldest month of observed data and introducing a new future month of assumed inputs. In that sense it is a living document.
An input which is important to the functionality of the Daily Operations Spreadsheet is losses through evaporation and seepage. The Daily Operations Spreadsheet includes observed data about evaporation and seepage losses which are updated each day. The Defendant’s evidence is that assumptions about short-term future losses are based on a calculation using an equation that trends future losses based on observed losses. Assumptions about long-term future losses are trended to seasonal loss behaviour. Those assumptions can be adjusted where, for example, higher temperatures are forecast. Data from Bureau of Meteorology forecasts are used for rainfall and temperatures. The assumptions about inflows are conservative. They allow for any actual or short-term (four days) forecast rainfall but then recede back down to 90% AEP.
The Daily Operations Spreadsheet is used on a day-to-day basis by the river operators in the River Operations Team, not just the Duty Operator. Like any spreadsheet, someone can use it to test particular operations such as increasing the flow rate at Yarrawonga Weir or changing the date upon which such an increase might start. For example, assumptions about future release rates can be changed and changed again to see how the change will affect downstream storage levels modelled in the spreadsheet.
It is the practice of the River Operations Team to use the Daily Operations Spreadsheet as part of the process of making decisions about changes to release rates.
Unlike the Fortran Assessment Model, the Daily Operations Spreadsheet has the functionality to produce outputs at daily intervals.
The practice of the River Operations Team is to “backup” the Daily Operations Spreadsheet each day to a shared location where it can be accessed by all members of the Team. The backup is overwritten the next day. If a particular river operator uses the Daily Operations Spreadsheet to test possible changes, in the ordinary course not all (or any) of the iterations of the spreadsheet generated by that process are saved.

The first months of the 2017/2018 water year (June and July) had inflows which were exceeded by 94-95% of historic inflows. Conditions turned wetter in August 2017 with inflows which were exceeded by 61% of historic inflows. In combination the total inflows for June to August 2017 were exceeded by 78% of historic inflows.
In view of the recent rain and the rain forecast at that time, in early-August 2017 the Defendant stopped making bulk transfers of water from Hume Reservoir to Lake Victoria. On 10 August 2017 the Defendant met with the WLWG. On 28 August 2017 the Defendant again ran the Fortran Assessment Model using observed data for inflows since the 2017/2018 AOP was originally adopted. An updated AOP Spreadsheet and AOP Document were produced. On 30 August 2017 a further meeting with the WLWG occurred to which the Defendant reported that it had not made any decision about future bulk transfers of water from Hume Reservoir to Lake Victoria.
The quantity of water stored in Lake Victoria rose during September 2017. By 28 September the Lake was virtually full at 670 GL. The rise was not caused by September rain but by inflows from August rain finally reaching the Lake. September 2017 was a very dry month. The Defendant issued a Weekly Report on 4 October 2017 which stated that September had been the driest September since records commenced 118 years ago.
A meeting with the WLWG took place on 11 September 2017 at which bulk transfers from Hume Reservoir to Lake Victoria were discussed. The Defendant reported that it had made no key decisions.
On 20 September 2017 the River Operations Team again ran the Fortran Assessment Model using data about the inflows actually observed since the end of August. Mr Bishop either undertook the task or assisted another river operator, Ms Greenwood, to do so. The process was not undertaken for all six scenarios but only for the “moderate” scenario and the “dry” scenario. In the case of the “dry” scenario, two versions were prepared, one which assumed that all transfers of bulk water from Hume Reservoir to Lake Victoria would be within the channel capacity of the Barmah Choke and one which assumed that some transfers would be made above the capacity of the Barmah Choke. Test results were produced for all three bases.
The basis which assumed future inflows in accordance with the “moderate” scenario assumed that there would be overbank transfers and that they would commence in October 2017. The assumption was that 25 GL of water would be transferred overbank in October with further overbank transfers in November and December. On those assumptions, the output from the model showed that there would be 656 GL of water in Lake Victoria as at 31 December 2017 which was expected to be enough to meet South Australia’s entitlements over summer and autumn.
One basis of the test which assumed inflows in accordance with the “dry” scenario assumed that transfers above the capacity of the Barmah Choke would occur in November and December. It was also assumed that the Defendant would draw 210 GL of water from the Goulburn IVT account plus 110 GL from the Murrumbidgee IVT account. On those combined assumptions, the output from the model showed that the quantity of water stored at Lake Victoria would decline through October, November and December to a level of 603 GL as at 31 December 2017. At that modelled storage level, it was expected that there would still be sufficient water in Lake Victoria to meet South Australia’s entitlements over the coming summer and autumn.
On the other hand, if inflows were assumed in accordance with the “dry” scenario and no overbank transfers were assumed, and it was assumed that the Defendant would draw 225 GL of water from the Goulburn IVT account plus 80 GL from the Murrumbidgee IVT account, the output from the model showed that there would be only 472 GL in Lake Victoria at the end of the year. That storage level would not be enough to meet South Australia’s entitlements from summer onwards. There would be a system shortfall.
The three outputs from Fortran Assessment Model were provided to the WLWG.
On 22 September 2017 Dr Davis went on leave until 26 September 2017. On 25 September 2017 Mr Bishop went on leave until 16 October 2017.
On 26 September 2017 officers of the Defendant again met with the WLWG. Dr Davis was still on leave so the meeting was chaired by Ms Hickey. Four other river operators from the River Operations Team went to the meeting, including Ms Greenwood, Ms McCann and Mr Waterworth. Nine members of the WLWG attended, including representatives from New South Wales, Victoria and South Australia. The minutes of the meeting record that the River Operations Team was “considering the likelihood of overbank flows commencing around mid-October” and that “overbank flows downstream of Yarrawonga are likely to be required for at least 6 weeks from Oct-Nov”. Another statement recorded in the minutes was that transfer of water from Hume Reservoir to Lake Victoria “will need to commence”. The following “Key Authority operational decision” is recorded:

Although Chapter 4 of the ACT Liability Statute governs the Plaintiffs’ claims, the parties agree that there is nothing in the provisions of the Statute which affects the application of the common law principles in this case.
As set out above, assuming reasonable foreseeability of the not insignificant harm, there is no dispute that a breach of the duty of care to protect the plaintiff from the risk that the harm will occur if the defendant fails to take a precaution which a reasonable person in his or her position would have taken. [162]
If the defendant takes the precautions which a reasonable person would have taken there is no breach, even if the plaintiff nonetheless suffers the foreseeable harm. The duty is one of reasonable care, not a duty to ensure that the harm does not occur. [163]
When considering whether a precaution was one which a reasonable person would have taken, regard must be had to the probability that the harm would happen if precautions were not taken, the likely seriousness of the harm, the burden of taking precautions to avoid the risk of harm and the social utility of the activity creating the risk of harm. [164] Regard may also be had to any conflicting responsibilities the defendant has. [165]
The duty of care is not owed at large. It is not owed in the abstract. [166] In order to determine whether a particular precaution was one which a reasonable person would have taken it is necessary to keep in mind the person to whom the duty was owed and the risk of harm from which the defendant had a duty to protect that person. [167]
The question of breach must be considered from the perspective of the defendant at the time the breach is alleged to have occurred. Things known to the defendant at the time, or which ought to have been known, may be taken into account. [168] The breach must be considered without the benefit of hindsight.
A particular application of the general prohibition on using hindsight to determine breach relates to the relative importance of multiple factors which pertained at the time the defendant failed to take a precaution. The advantages of hindsight include knowing which of many factors have turned out to be important and which have not. [169] In Rosenberg v Percival, Gleeson CJ said at [16]:

The cause of action upon which the Plaintiffs sue is negligence. It is important to bear that in mind because of the statutory and contractual framework in which the Defendant acted.
As set out above, a prominent feature of the case is the Murray-Darling Basin Agreement and the documents sequentially contemplated by the Agreement, namely the O&O Document and the 2017/2018 AOP. The Plaintiffs do not, however, sue the Defendant for breach of contract. Nor could they because the Plaintiffs are not a party to the Agreement. Nor is the Defendant. Under s 18E(2) of the Commonwealth Statute the Defendant must in performing its functions and duties and exercising its powers comply with any requirements of the Agreement. So too, under cl 30(2) of the Agreement, the Defendant must carry out its functions in relation to river operations in accordance with objectives and outcomes specified in the O&O Document but the Defendant is not a party to the O&O Document. Nor are the Plaintiffs. Merely because the Defendant has failed to act in accordance with a particular provision of the Agreement or the O&O Document does not prove breach of the alleged duty of care owed to the Plaintiffs.
The principle is true even if breach of the Agreement or the O&O Document is viewed as a breach of a statutory provision by virtue of s 18E(2) of the Commonwealth Statute.
Whilst breaching a statutory duty may be a matter to take into account when considering whether a particular duty of care has been breached, the former does not automatically lead to the latter. There is a different cause of action where breach of a statutory duty causes a particular kind of loss to a person within a class for whose benefit the statutory duty was imposed. The touchstone for that cause of action is a legislative intention that the statutory duty protect a particular class of persons’ interests for which a private cause of action will lie. [170] The Plaintiffs bring no such claim in these proceedings.
These matters warrant emphasis because a number of the Plaintiffs’ final particulars of negligence are framed as failures by the Defendant variously to comply with specific provisions of the O&O Document or the 2017/2018 AOP. To some extent the O&O Document and the 2017/2018 AOP identify the task which the Defendant was undertaking, [171] but the question remains throughout whether the Defendant failed to act with reasonable care to protect the Plaintiffs from pure economic loss. [172]

Apart from the members of the Defendant’s River Operations Team, no witness was called who has experience with river operations.
As explained above, the Plaintiffs called Dr Ladson who was a hydrologist with expertise in the science of the movement, distribution and management of water. Dr Ladson does not have experience as a river operator. He accepted that there are aspects of operating a river about which he is not qualified to express an opinion, such as environmental, political and social considerations. There are some passages in Dr Ladson’s reports which are expressed in conclusory terms. Dr Ladson frankly accepted that he is not qualified to express an opinion about what a river operator acting with reasonable care would have done in any particular situation. Any evidence from Dr Ladson which is directed to reasonableness or a similar concept is limited to the hydrological aspects of the decision or action in question. Whether Dr Ladson’s views are nonetheless relevant to the alleged breaches by the Defendant’s River Operations Team depends on whether or not there are non-hydrological considerations which were relevant to the particular decision or action.
The Defendant submits that it is a large step to make a finding of negligence in relation to the operation of a very complex system unsupported by evidence from a person qualified to operate that system about what a reasonable operator would have done. The standard of reasonable care required of the Defendant is that of a reasonable river operator as described above. [177] The role which expert evidence plays in determining whether a professional defendant has failed to meet the standard of care depends on the nature of the breach, including the nature of the precaution which it is alleged the defendant failed to take. For some breaches, for example where it is rudimentary and obvious, it may be possible to demonstrate a failure to meet the standard of care without evidence from other persons with experience in exercising the relevant special skill. [178]
Whilst there are differences, the present case has similarities with cases where there is a complaint about the diagnosis and treatment performed by a medical professional. For both water years, the Plaintiffs contend that the Defendant did not act with reasonable care when identifying the condition of the Murray River system, specifically Lake Victoria, and deciding upon a course of action to address that condition. It has been said that for such a case, evidence from an experienced professional is of very considerable significance, but that the nature of the particular risk and its foreseeability are not exclusively within the province of the relevant profession. They are often matters of common sense. [179]
In any event, it remains the Plaintiffs’ onus to demonstrate that the Defendant failed to act with reasonable care in the particular circumstances of this case.

The Plaintiffs’ allegations of breach in the 2017/2018 water year commence with the proposition that the Defendant’s decision to release more than 10,000 ML of water at Yarrawonga Weir on 3 October 2017 was a significant and unusual decision. That proposition is important context in which the reasonableness of the decision is to be assessed. Sometimes explicitly and sometimes implicitly, the Plaintiffs contend that care should have been taken commensurate with the gravity of the decision.
There is no dispute that transferring operational water through the Barmah Choke results in more water being lost when the transfer is at a magnitude which exceeds the channel capacity which forces some water to go overbank. Not all the water which goes overbank is lost. The process is referred to as an “overbank transfer” of water to the river system below the Choke. From the total amount of water which goes overbank there are return flows into the Murray River below the Choke. There are particular paths (flood runners, or occasionally “creeks”) through which overbank water is directed (by forest regulators) in an attempt to maximise the return flows. Nonetheless, some water does not return to the River and is “lost” due to evaporation, seepage and transpiration. The amount of lost water depends on the conditions, including the moisture content of the floor of the Forest at the time when the water goes overbank. The Plaintiffs describe the losses from transferring water overbank as “very high” or “very significant”, but context is needed to make those descriptions meaningful. So too the Plaintiffs’ submission that overbank transfers are a “last resort”. In any event, there is no dispute that a river operator acting with reasonable care will avoid a transfer of operational water above the capacity of the Barmah Choke to the extent reasonably possible.
In the abstract the Plaintiffs submit that the decision on 3 October 2017 to transfer operational water overbank through the Barmah Choke was “rare”, “largely unprecedented” and “exceptional”. They point out that a decision to transfer water above the channel capacity of the Barmah Choke had happened only three times in the previous 47 years, namely in 1994/1995, 2002/2003 and 2015/2016. The primary facts may be accepted but the characterisation is based on a false comparison. At all relevant times prior to June 2017 the Defendant (and its predecessors) had access to the MIL infrastructure as a means of bypassing the Barmah Choke. Dr Ladson refers to evidence of the MIL infrastructure being used in 1983. The 2017/2018 water year was only the second time that the Defendant was confronted with the task of managing the risks of system shortfall without the MIL infrastructure. In Section 8.3.1 of the 2017/2018 AOP, the lack of access to the MIL infrastructure was stated to be the reason why overbank transfers were assumed to be necessary later in the year. The decision to conduct the overbank transfers in 2017/2018 did not have the historic significance for which the Plaintiffs contend.
Nor was it a decision of particularly great magnitude. Hindsight shows that the quantity of water lost on the floor of the Barmah-Millewa Forest as a result of the overbank transfers between 4 and 16 October 2017 was about 55 GL. [180] Whilst it may be accepted that every drop matters, that figure requires context when examining the decision which led to that outcome, especially for the purpose of considering breach of a duty to exercise care tempered by reasonableness. When preparing the 2017/2018 AOP, the Defendant took the 2006/2007 water year as a relevant reference point because it was a recent year which had been “dry”. In the 2006/2007 water year, 750 GL of water was lost from seepage and evaporation when transferring water from Dartmouth Dam to the South Australian border. For the purposes of 2017/2018, the Defendant decided to assume higher losses to take account of a number of factors, including the assumed need for two months of overbank transfers through the Barmah Choke. Thus the Defendant decided to assume that in the 2017/2018 water year there would be transfer losses of between 850 GL and 1,050 GL depending on the inflows during the year. [181] As indicated, there is no suggestion that the Defendant acted negligently in making that assumption when preparing the 2017/2018 AOP. Having made that non-negligent assumption, there is no suggestion that it would have been negligent for the Defendant to have operated the river system during the 2017/2018 water year in such a way that between 850 GL and 1,050 GL was lost from transferring water from Dartmouth Dam to the South Australian border.
In any event, the loss figures in the AOP were merely assumed figures, adopted at the commencement of the water year for modelling purposes. The Defendant knew that the actual conveyance losses would be different. The 2017/2018 AOP contained the following statement:

These facts also show that the loss of water from overbank transfers were losses which in the ordinary course would be (and actually were) expected to happen. There is no dispute that the way in which water is transferred, including the time of year, makes a difference to how much water is lost en route, but there is no lossless option available to the Defendant once it is decided to transfer water.
When considering making overbank transfers at the end of September 2017 the Defendant would not have known exactly how much water would be lost as a result of the decision it was considering. It would have known that it was generally less efficient to transfer water overbank than transferring water in-channel. It would have known that the 2017/2018 AOP had been prepared on the assumption that at least 100 GL of water might be lost above the amount lost in the 2006/2007 water year. As Mr Reynolds said in cross-examination, the Defendant knew that losses occur all the time when conveying water not just when there are overbank transfers. There is nothing to suggest that the Defendant was concerned about incurring dramatically higher losses when deciding to increase the release rate at Yarrawonga Weir above 10,000 ML/day, or that a river operator acting with reasonable care would have seen it any differently.
In that objective context, it is necessary to bear in mind the precise nature and content of the decision which the Plaintiffs now criticise. The 2017/2018 AOP was prepared on the assumption that there would be some overbank transfers in the 2017/2018 water year if inflows corresponded to those assumed for the “very dry”, “dry” and “moderate” scenarios. That information is communicated in the graph of the rate of release of water at Yarrawonga Weir which is set out at [543] above. The brown, yellow and grey lines all depicted above 10,000 ML/day for two months. For those three scenarios the AOP was based on the assumption that the overbank transfers would be necessary to meet lower system requirements including filling Lake Victoria over summer. It was assumed that there would be two months of overbank transfers. It was assumed that losses between Dartmouth Dam and the South Australian border would be 100 GL higher than in 2006/2007 mainly because of the need for overbank transfers. As for timing, it was assumed that the overbank transfers would occur in November and December. The Plaintiffs make no criticism of the 2017/2018 AOP or any of these assumptions. In other words, the Plaintiffs accept that the Defendant acted with reasonable care by working on these assumptions about overbank transfers when the AOP was finalised and adopted in early-August 2017.
At all times after the 2017/2018 AOP was finalised and adopted the Defendant continued to work on the assumption that overbank transfers might occur. The updates to the AOP on 20 September 2017 were prepared by a process which specifically tested the alternative assumption that there be no overbank transfers for the “dry” scenario. That test indicated that Lake Victoria would have insufficient water to meet South Australia’s entitlements by around February and March 2018. In other words, having regard to the inflows which had occurred in June, July and August, the Defendant checked whether overbank transfers were still necessary and discovered that they might be. The Plaintiffs make no criticism of that work.
Properly understood, the Plaintiffs’ case is not that the Defendant was negligent by deciding whether to make overbank transfers. In substance, the Plaintiffs accept that the Defendant had already made that “decision” in a non-negligent way when the 2017/2018 AOP was prepared and adopted. Instead, the Plaintiffs contend that the Defendant was negligent in deciding when to commence overbank transfers. To be even more precise, the Plaintiffs accept that it was not negligent for the Defendant to decide to commence overbank transfers in October 2017 as that was the basis upon which the testing with the Fortran Assessment Model was undertaken on 20 September 2017 for the “moderate” scenario. The negligence is said to arise from the decision to commence on 4 October rather than, say, 15 October.
The precise identification of the decision now criticised relates directly back to the gravity of the decision. As explained above, hindsight shows that 55 GL of water was lost as a result of the overbank transfers which were actually made. However, there is no evidence that a river operator would have expected that more water would be lost if the overbank transfers commenced on 4 October rather than on 15 October 2017. There is no basis to suggest that the risk of losing more water would have been a relevant factor in the decision about when to commence overbank transfers, at least not as between 4 October and 15 October 2017.
This is one of the aspects of the Plaintiffs’ case which demonstrates that it is informed by hindsight. The only reason that the decision as to timing is now said to be significant (and the cause of an unnecessary loss of 55 GL of water) is that it started to rain in mid-October 2017. When that happened, the Defendant formed the view that the overbank transfers were not, after all, required. Subsequent events have revealed that a delay from 4 October to 15 October 2017 would therefore have made a difference to the amount of water lost in the 2017/2018 water year. The River Operations Team did not have the benefit of hindsight when it made its decision to commence on 4 October 2017.
A fact made clear by the evidence as a whole is that at all times, both before and after the 2017/2018 AOP was finalised, the River Operations Team gave continuous and active consideration to transfers of bulk operational water from Hume Reservoir to Lake Victoria, both as to timing and magnitude. Mr Bishop described a continuous process of “steering” Lake Victoria throughout the year, suggestive of constant attention and care. On the evidence as a whole, that suggestion rises to the level of an inference, both plausible and probable. A recurrent theme of the evidence of Mr Bishop, Ms Hickey and Dr Davis was the need to address the never-ending balance between having not enough water in Lake Victoria (the risk of system shortfall) and having too much (the risk of spill).
The Plaintiffs have drawn out isolated comments made by the Defendant’s river operators in contemporaneous documents, affidavits and cross-examination and submit that a certain situation or perspective pertained at a particular time. The submissions do not pay sufficient regard to the nature of the work performed by the River Operations Team. To the knowledge of the river operators, the variables within the complex system were inherently uncertain and constantly changing. Whether expressly (as was often the case) or by necessary implication, statements made by the river operators about the future was qualified by a degree of uncertainty.
A striking feature of communications by the River Operations Team about future events, including the Team’s own expectations, intentions and plans, is that river operators use language which is consistently imprecise as to timing. Many documents were referred to during the hearing from which this is apparent. It was also apparent from the five days of oral evidence given by Mr Bishop, Ms Hickey and Dr Davis. Listening to the cross-examination left me with the distinct impression that river operators are open-minded about future matters and are reluctant to be pinned down, whether by training or experience or otherwise. This is not a criticism. Their language testifies to a premium being placed on flexibility.
The flexible outlook is exemplified by the River Operations Team’s communications about the commencement of overbank transfers in 2017. At all times prior to October 2017 the Defendant communicated the commencement date in imprecise language. The 2017/2018 AOP used the phrase “from late spring into early summer”. The minutes of the WLWG meeting on 26 September 2017 used phrases such as “around mid-October” and “for at least 6 weeks from Oct-Nov”. As late as 5pm on 29 September 2017 Ms Greenwood’s email referred to commencement “sooner than previously forecast”. She did not specify a date. These are just examples.
The significance of these observations is that the decision to release water in excess of the channel capacity of the Barmah Choke on 4 October 2017 was an incremental step in a continuous process by the River Operations Team. The process had begun as long ago as July 2017 with the adoption of the 2017/2018 AOP, and even before that with the modelling at the end of May from the preceding year. The decision in early-October was not a stand-alone decision. Given what had gone before, it was not unusual or significant, much less rare, risky or dangerous. As I have said, the process leading up to 29 September 2017 is not criticised by the Plaintiffs. The question for decision in the present case is whether the Defendant was negligent when taking the incremental step of fixing the date to commence overbank transfers on 4 October 2017 rather than a date later in October.
No concession was made by any of the Defendant’s witnesses which contradicts these findings. The Plaintiffs submit that a passage from Mr Reynolds’s cross-examination shows that the decision to make overbank transfers was an “unusual and risky decision”. The characterisation may fairly be attributed to the words of the question but not Mr Reynolds’s answer. [182] More generally, the importance, rarity and significance of “overbank transfers” was put to the Defendant’s witnesses and generally accepted, but the questioning occurred at a high level of generality which did not grapple with the chronology and context set out above. In the abstract, it may be accepted that a decision to transfer water overbank through the Barmah Choke is an important, significant and unusual decision, but that fact does not inform the issues in the case when considered in isolation and without context.
In their Reply Submissions the Plaintiffs say that the decision to commence overbank transfers on 4 October 2017 was “urgent”. They also refer to the Defendant “immediately…abandon[ing] its plans”. These submissions are contrary to the evidence. At all relevant times the Defendant was engaged in a steady continuation of a process which had been ongoing for months.

Relatedly, the Plaintiffs present their case as a complaint about a decision to commence overbank transfers and that the decision was made by Ms McCann who was the Duty Operator during the week at the end of September 2017. The burden of their submissions is that Ms McCann was solely responsible for the decision. The Plaintiffs contend that Ms McCann made her decision on Friday 29 September 2017. The principal relevance of that date is that it supports the contention that Ms McCann made the decision at a time when she was unassisted by senior management of the River Operations Team. On 29 September each of Mr Reynolds, Dr Davis and Mr Bishop were on leave. It is submitted that Ms McCann was “left to [her] own devices” when she made her decision. The implicit proposition is that Dr Davis did not have any discussions about the commencement date when he was in the office the previous day, nor did Mr Reynolds when he returned the next working day (Tuesday 3 October 2017) even though Ms McCann did not issue the direction to the operator of Yarrawonga Weir until 2:45pm on 3 October 2017. The actual overbank release of water did not commence until 4 October 2017.
This is a part of the Plaintiffs’ case which has an air of unreality about it. It follows from the Plaintiffs’ mischaracterisation of the decision to bring forward the overbank transfers by a little over a week as an urgent, risky and dangerous decision. The argument does not make allowance for the work which had already occurred nor the nature of the ongoing work of steering Lake Victoria through the year. It does not make allowance for the collaborative practices of the River Operations Team.
Turning to the evidence, the Daily Operations Sheet maintained by the Duty Operator (Ms McCann) contains the record for 29 September 2017 which is set out at [581] above. At 5pm Ms Greenwood sent the email referred to at [578] above. No direction to increase the release of water at Yarrawonga Weir was issued on 29 September 2017, nor over the long weekend on 30 September, 1 October and 2 October when Ms McCann was on duty and, perhaps, really left to her own devices.
In the ordinary course, on the morning of Tuesday 3 October 2017 the River Operations Team would have held its weekly Team meeting, about which Mr Bishop’s unchallenged evidence was as follows:

The next day Dr Davis returned to work. A diary appointment for 4 October 2017 suggests that he met with Ms Hickey at 2pm to “catch up on events of the past week”. This was the day upon which water started to flow at Yarrawonga Weir above 10,000 ML/day. Another instruction was given further to increase the flow rate to 12,000 ML/day. It may safely be inferred that Dr Davis was informed of what had happened that day and what was planned for the next day. There is no evidence of any disagreement by Dr Davis. On the contrary, on 5 October 2017 another member of the River Operations Team (Mr Lucardie) was the Duty Operator and recorded an instruction further to increase the flow rate to 12,500 ML/day. Mr Lucardie did so again on 6 October and again on 7 October. In these proceedings Dr Davis has expressed the view that the decision to transfer water overbank was appropriate and that he held that view at the time. Apart from an error about the climate outlook, there is no basis not to accept that evidence.
Ms Hickey was responsible for the Team and was at work on all relevant days. Her evidence was that she communicated with the Duty Operator and other members of the Team on a daily basis. She “supervised the Duty Operator who made day to day decisions to adjust or maintain release rates from upper River Murray storages within a release range contemplated by the O&O Document, the AOP or as discussed with the WLWG and the Senior Director of Operations to meet demands within the River Murray System”. Under Ms Hickey’s supervision the Duty Operator communicated with the external operators of infrastructure, like GMW at Yarrawonga Weir.
As relates specifically to the events in the 2017/2018 water year, Ms Hickey said that she discussed the commencement of overbank transfers with other members of the Team although she does not specifically recall who. She says that the Duty Operator consulted her about the commencement of overbank transfers on 4 October 2017 and that she agreed that it would be suitable to make those releases. She did not regard herself as “the decision-maker”.
Dr Davis said in his Affidavit that he had discussions with members of the River Operations Team, including Ms Hickey, about making overbank transfers. He was not precise about the dates but they appear to include 28 September 2017. Dr Davis did not say what was said in any discussion but he said that as result of them and other considerations he formed the view that overbank transfers would be appropriate.
Dr Davis was cross-examined about the identity of “the decision-maker”. Responding to propositions put to him, Dr Davis said that Ms Hickey was probably not the decision-maker. His offer to expand was not taken up. He said it could have been Mr Reynolds. It was put to him that it wasn’t himself to which he responded:

As for Mr Reynolds, he gave evidence that he formed the view that overbank transfers were appropriate but did not provide any specifics. In cross-examination he did not believe that he “exercise[d] the release power” in cl 98 of the Agreement, by which he meant that he did not “provide the specific direction”. He thought that any decision to release water would have been made with the knowledge of Dr Davis or himself.
Having regard to the evidence, on the balance of probabilities the decision to increase the rate of release of water at Yarrawonga Weir above 10,000 ML/day was not made until 3 October 2017 at or shortly before 2:45pm when Ms McCann issued the direction to the operator of Yarrawonga Weir. On the balance of probabilities, the decision was made collectively by Mr Reynolds, Ms Hickey and other members of the River Operations Team. Others involved in the discussions which resulted in the decision include Dr Davis (on 28 September 2017), Ms Greenwood and Ms McCann. Mr Lucardie’s involvement is demonstrated by the further instructions issued by him after he took over as Duty Operator on 5 October 2017. There may have been others.
The Plaintiffs’ focus on the precise identity of “the decision-maker” informs at least three issues in the case: the (lack of) experience brought to bear in the decision-making process, the absence of a sub-delegation of the release power by Mr Reynolds and/or Dr Davis and the Defendant’s failure to call Ms McCann and other members of the River Operations Team. The first issue is not supported by the evidence. The delegation issue is not a matter of substance in a claim for damages for loss caused by alleged negligence. It has negligible relevance to the issue of breach and is rendered sterile by the issue of causation. Delegation is addressed further below.
In relation to Jones v Dunkel there is no explanation as to why Ms McCann (nor Ms Greenwood nor Mr Lucardie) was not called by the Defendant to give evidence. The Plaintiffs submit that the Defendant’s failure to explain why these officers were not called engages the principle in Jones v Dunkel. They identify three matters about which evidence might have been given, namely the misreading of the Climate Outlook, the modelling conducted using the Daily Operations Spreadsheet on 29 September 2017 and the dead storage level at the Menindee Lakes. These issues are addressed as they arise below.
In this context, each of the seven precautions which the Defendant failure to take can now be addressed.

In cross-examination Dr Davis conceded that “that patch of white that comes in there just below the Australian Capital Territory, that’s all the Murray Valley” was an “average” and “not a dry outlook”.
Ms Hickey’s evidence is that she reviewed the Climate Outlook dated 28 September 2017 which indicated that “conditions were likely to be warmer”.
For a continent which Ms Hickey described as “one of the most variable hydrologies”, each of the maps in the Climate Outlook was depicted in small scale. The features of the Murray-Darling Basin including the rivers, valleys and mountains were not marked, although the border between New South Wales and Victoria is a proxy for the path of much of the upper Murray River. The text addressed future rainfall and temperatures in very general terms. Within those limitations, there was nothing in either the maps or the text which proves that either Dr Davis, Ms Hickey or any other member of the River Operations Team misread the information about temperature in the Climate Outlook dated 28 September 2017.
As for future rainfall, Dr Davis was ready to concede that the relevant map in the Climate Outlook did not depict a dry outlook. Given that concession, it is clear that an error has been made somewhere. The Plaintiffs submit that the error was made in 2017 when Dr Davis read the Climate Outlook, which error he has now faithfully recorded in his Affidavit in these proceedings. By far the more plausible explanation is that the error was made in 2022 when Dr Davis was proofreading page 133 of his 214-page Affidavit and did not notice that it said “dry outlook” when it should have said “hot outlook”. Even with the limitations referred to above, the map of future rainfall depicted most of New South Wales and Victoria as pale green or white which, for the reasons given by Dr Ladson, was not a dry outlook. If Dr Davis read it otherwise in 2017 he made an error which was obvious and inexplicable. On the other hand, an error in paragraph 599(e) of his Affidavit is explicable because the cross-reference to paragraph 596 is a refence to hotter conditions, not drier conditions. If Dr Davis intended to say “hot” rather than “dry”, the inclusion of the cross-reference makes sense. When asked about paragraph 599(e) in cross-examination he said that it should be corrected. The Plaintiffs have not proved that Dr Davis misread the Climate Outlook dated 28 September 2017.
Even in her absence, there is no basis for a finding that Ms McCann misread the Climate Outlook. The premise of the contrary contention is that Dr Davis and Ms Hickey misread the Climate Outlook and Ms McCann has not been called to say that she read it differently to them. The submission collapses given my finding that neither Dr Davis nor Ms Hickey misread the Climate Outlook.
The Plaintiffs have not proved that any person from the Defendant misread the Climate Outlook dated 28 September 2017. The first alleged failure is not proved.

Two questions which assumed prominence in the case are What is the Annual Operating Plan? and What does it mean to act in accordance with it? The first question requires an inquiry into the nature and effect of the AOP. Only then is it possible to answer the second question. Having read the 2017/2018 AOP Document and the 2018/2019 AOP Document, having reviewed the AOP Spreadsheets upon which they are based and having listened to the cross-examination of the Defendant’s river operators, the answers to these questions are not simple.
The Plaintiffs have not expressly addressed these questions in their closing submissions, written or oral. Their submissions are descriptive of the process by which the AOP was prescribed by the O&O Document and then prepared, and of the document (and spreadsheets) which resulted. The Plaintiffs do not squarely address the meaning and effect of the AOP as it relates to the decision on 4 October 2017 to increase the flow of water at Yarrawonga Weir above 10,000 ML/day. Their case has to be inferred from the conclusions for which they contend, namely that the AOP did not “provide for” overbank transfers until November 2017 and by doing otherwise the Defendant “departed from” and failed to “adhere to” the AOP. The Plaintiffs’ case is therefore to be understood as the contention that, upon the adoption of the AOP, the rates of release of water at Yarrawonga Weir which were depicted in the AOP for each month of the water year were thereafter a prescription of what the Defendant had to do. The obligation was subject to review or formal amendment of the AOP.
Whether that contention is to be accepted requires a consideration of the text of the O&O Document which contained the compliance requirement and to which the AOP Document expressly referred. Regard must also be had to the subject matter of river operations. The text of the AOP Document itself must be reviewed. On this issue, regard may also be had to the expert evidence and the evidence of river operators who prepared and used the AOP. Only then can the nature and effect of the AOP be understood.

As for the text of the AOP Document itself, some of the most pertinent provisions and graphs are set out above. The AOP Document must nonetheless be read as a whole. A number of general features are apparent. First, it is clear that the intended audience for the AOP Document was the general public. The AOP Document had the form and content of a document which was prepared to provide information to the public about the topics addressed in the document. It described itself as “published” and commenced with information about contacts and copyright. It concluded with a direction to the Defendant’s website for further information. The substantive parts were presented in a variety of formats, including text, tables and graphs. Most if not all of the information would have already been known to the Defendant and maybe also to the members of the WLWG.
Secondly, a prominent aspect of the text is that future matters were generally referred to as assumptions which the Defendant had adopted. The Document did not express expectations, estimates or forecasts other than to expose the reasoning for the adoption of a particular assumption. The AOP Document explicitly stated that future events would be different to those which had been assumed.
Thirdly, there were a large number of assumptions, some of which cascaded from others. The document conveyed complexity and interrelationship between the many aspects of the river system.
Fourthly, the document contained express statements about future uncertainty and variability, which statements were unambiguous and repeated. There were multiple disclaimers against reliance on the information in the document.
Finally, and following on from the preceding features, the AOP Document addressed planning for the future only in the sense that it described assumptions about future river operations which were expressed to be derivative of other assumptions. Nowhere did the AOP Document set out a “plan” in the sense of a prescription of what the Defendant was going to do. It described a landscape within which the Defendant would make future decisions about what to do.
Specifically in relation to the rate at which water would be released at Yarrawonga Weir, the AOP Document included the graph which is set out again here:
For each month and for each of the six scenarios, a value was taken from the relevant AOP Spreadsheet and depicted in the graph. With the greater precision made possible by the Spreadsheet, the figure for October 2017 was 10,000 ML for each of the “dry” and “moderate” scenario. For November and December 2017 it was 15,000 ML. Consistent with the language used elsewhere in the AOP Document, these values were said to be the rates of release assumed for the relevant scenarios. The six scenarios between them were described as “the likely range of conditions and operations that are most likely in 2017–18”. Other statements made in the AOP Document about the uncertainty of future operations are set out above. Nowhere did the text of the AOP Document express the intention of the Defendant to cause water to be released at Yarrawonga Weir at 10,000 ML/day in October and at 15,000 ML/day in November 2017, not even a conditional intention. Nowhere in the AOP Document did the Defendant express a commitment to that course of action.
A document which operated in the manner just described was not perhaps aptly described as a “plan”. In 2019 the AOP was renamed. It is now called the “Annual Operating Outlook”. The meaning and operation of a document is not necessarily dictated by the label given to it. Nor does the use of the word “plan” in cl 10 of the O&O Document require the AOP to be given an operation which departs from the specific contents of cl 10(1)(b)(iv). If contrary to the foregoing, the Defendant was required by cl 10 to adopt a prescriptive document which bound the Defendant’s future operations, the Defendant failed to do so. The 2017/2018 AOP is not such a document. The evidence demonstrates that it would not have been reasonably possible to prepare such a document because there were too many variables and too many unknowns.

Although written for the general public, there are parts of the AOP Document which are technical in the sense that the meaning and effect of those parts may be assisted by expert explanation. This includes the graph set out at [685]. The horizontal axis of the graph is a monthly timeline. If the graph was consulted in September 2017, what information (if any) would it have conveyed about the actual flow of water at Yarrawonga Weir for October 2017? Would it have been different if the graph was consulted in October 2017? More pertinently, what information would it have conveyed to a river operator acting with reasonable care in the circumstances which pertained in this case?
Apart from the Defendant’s river operators, no witness was called to explain the meaning of the graph (or the AOP) based on specialised knowledge based on the witness’s training, study or experience.
The Plaintiffs’ expert, Dr Ladson, is a highly qualified and experienced hydrologist, particularly in relation to the management of waterways. As noted above, in 2009/2010 Dr Ladson contributed to the AOP, although it is not known if it was in the same form. Dr Ladson prepared reports in which he reviewed the 2017/2018 AOP and the decision made by the Defendant in early-October 2017 to release more than 10,000 ML/day water at Yarrawonga Weir on 4 October 2017. Dr Ladson made a number of criticisms about that decision. However, nowhere in Dr Ladson’s evidence does he explain the meaning or the effect of the graph set out at [685], nor any of the other similar graphs in the 2017/2018 AOP or in the 2018/2019 AOP. Dr Ladson said that the AOP “is intended to guide river operations” but nowhere does Dr Ladson say that, having regard to the 2017/2018 AOP, the Defendant was in some way required to release water at Yarrawonga Weir at 10,000 ML/day in October 2017, or that the release of more water would be contrary to the graph (or any other part of the AOP).
In his Second Report Dr Ladson accepted that the AOP’s do not determine the actions that must be undertaken by the Defendant. Dr Ladson said that a reason for that was that some operational aspects are not included in the AOP planning documents but are added directly in the Daily Operations Spreadsheets.

Another aspect of the use made of the AOP by the River Operations Team highlights a further problem with the Plaintiffs’ contention that the Defendant was obliged to comply with the rates of release at Yarrawonga Weir depicted in the AOP for each month of the water year. For each future month, the AOP depicted six rates of release, one for each scenario. For some months the rates of release were the same or similar for two or more of the scenarios. For some months they were all different. The Plaintiffs’ contention that the Defendant was required to comply with a particular rate of release in a particular month presupposes an anterior determination that one of the six scenarios “applies” or dictates the operation of the river at that time.
A need to determine and apply a single scenario at any given time demonstrates the gulf between the Plaintiffs’ case and the true nature and effect of the AOP. For the purpose of planning future operations, fixing upon a single scenario to the exclusion of the other five would be antithetical to the way the AOP works. The AOP contemplates multiple scenarios (as expressly required by cl 10(1)(b)(iv)(G) of the O&O Document) which illustrate what might happen to the river system depending on what inflows occur throughout the remainder of the water year. Beyond the next four, possibly eight, days the River Operations Team does not know what the rainfall and inflows will be for the year, nor which of the six hypothetical scenarios will be most closely matched. This is especially so in the early months of the water year.
The AOP is intended to depict a range. The Plaintiffs do not submit that the range selected by the Defendant for the purposes of its operational planning was not reasonable, nor is there any evidence upon which such a submission could be made. There is nothing to say that the Defendant should only have selected four scenarios rather than six. Having selected a range depicted by six scenarios, there is nothing to say that it was not reasonable for the Defendant to consider the whole of the range, especially in the early months of the water year or, pertinently, in September and October. At one extreme, the range was depicted by the “extreme” dry scenario for which it was assumed that inflows would be lower than had ever previously occurred since records commenced 125 years ago. There is no basis to suggest that it was not reasonable for the Defendant to take that possibility into account for the purposes of performing its functions within the statutory framework. At the other extreme was the “wet” scenario. In 2017 the river system had experienced inflows which conformed to or exceeded those assumed for the wet scenario in 37 of the last 125 years. It was reasonable for the Defendant also to have regard to that end of the range when making decisions in September and October.
Mr Bishop explained the approach in paragraph 183 of his Affidavit dated 30 June 2022:

In the light of these limited facts, the Defendant’s “very dry” scenario assumed annual inflows for the entire river system which would be exceeded by the annual inflows for 95% of the previous 125 years. The Defendant’s “dry” scenario had an annual exceedance probability of 90%. The Defendant’s “moderate” scenario had an annual exceedance probability of 75%. As at October 2017, it was not possible to determine which scenario would be the one which the actual inflows would most closely match by the end of the water year in May 2018.
The Plaintiffs submit that the Defendant made the relevant decision to release water at Yarrawonga Weir above 10,000 ML/day “in the context of” the “dry” scenario. They do not explain why the “dry” scenario was applicable in October 2017, either generally or for the specific purpose of commencing overbank transfers from Hume Reservoir to Lake Victoria. They do not address how the determination of the applicable scenario was to be made generally. If one were merely to take the annual exceedance probability of the inflows for the entire system to date and match it to the closest scenario, 78% AEP inflows for June to August 2017 most closely matched the “moderate” scenario, not the “dry” scenario.
Starting on 1 June, in any given water year the first three months are generally the wettest. However, without further explanation it is not obvious why the inflows which actually occur in the first three or four months will determine the total inflows which will occur for the whole water year. More precisely for current purposes, it is not obvious why the inflows which occurred in June to September 2017 would have been indicative of the inflows which would occur in October, November and December 2017 (which was the time by which it was intended that Lake Victoria be full, as referred to in the Lake Victoria Operating Strategy).
Mr Bishop said in his Affidavit that, having regard to the inflows which had occurred in June to August 2017, the river conditions and river operations “were between the conditions and operations assumed in the moderate and drier end May 2017 AOP Spreadsheet”. [190] As set out above, on 20 September 2017 the River Operations Team undertook some updates to the AOP. They did not run all six scenarios. They only ran the “dry” and “moderate” scenarios. They evidently had made an assessment that updates of these two scenarios would be of the most assistance in making future planning decisions.
Dr Davis accepted in cross-examination that the “dry” scenario applied as at the end of September 2017, albeit in general terms. In context, his evidence should be taken as meaning no more than that the annual exceedance probability of the inflows to the relevant date for the entire system most closely matched the inflows assumed for the “dry” scenario. It was not a concession that the Defendant had an obligation to proceed on the basis that annual inflows for the 2017/2018 water year would be less than 90% of historic inflows (but not less than 95%).
The need to make a determination about which scenario “applied” further undermines the Plaintiffs’ contention that the Defendant was obliged to comply with the AOP in the sense that it had to release water at Yarrawonga Weir at a rate specified for one of the six scenarios.
Nonetheless, the Plaintiffs’ complaint about non-adherence to the AOP can conveniently be considered on the assumption that the “dry” scenario applied at the end of September and the beginning of October 2017.

For the “dry” scenario, the 2017/2018 AOP contemplated that 10,000 ML/day would be released at Yarrawonga Weir in October 2017. On each of 1 October and 2 October 9,941 ML was actually released. On 3 October 9,871 ML was released and on 4 October 10,810 ML. Thereafter the rate of release exceeded 10,000 ML each day, all of which was operational water until 13 October. The 2017/2018 AOP contemplated that 15,000 ML/day would be released at Yarrawonga Weir in November 2017.
Prepared using the Fortran Assessment Model, the AOP did not have the functionality to specify the release of water for each individual day of the month. The figure for any given month was therefore an average figure. So much was specified as a note to the graph set out at [543] above. There is no dispute that a change in the rate of release from one month to the next cannot be implemented suddenly. It must be implemented gradually with incremental changes made each day leading up to (or down to) the new rate of release. Nor is there any dispute that there may be disadvantages in permitting a river to have a constant rate of flow over a prolonged period, such as erosion of the banks. It follows that where a monthly rate of release was specified in the AOP it was contemplated that there may be days when the quantity of water which was actually released would exceed the specified rate of release.
The Defendant submits that the mere fact that operational water in excess of 10,000 ML was released at Yarrawonga Weir each day between 4 October and 13 October 2017 does not demonstrate that the Defendant did not comply with an obligation that the average rate of release be 10,000 ML/day in October 2017. It submits that some latitude was implicit and, without more, a failure to comply cannot be found.
Whilst the submission might be accepted for an isolated day of any given month, it is to be rejected on the facts of the present case. There are two reasons for this. First, the increase in the flow of water on 4 October 2017 was not an increase made by the Defendant pursuant to a plan that the daily release rate for the whole of October would still average out at 10,000 ML. Secondly, I accept the Plaintiffs’ submission that an increase above 10,000 ML at Yarrawonga Weir was not merely a change in quantity. Below Yarrawonga Weir the Barmah Choke imposed a capacity constraint which would be exceeded above 10,000 ML/day. The increase in flow on 4 October 2017 therefore had a qualitative difference to that contemplated by the 2017/2018 AOP.
If, contrary to the foregoing, the 2017/2018 AOP imposed an obligation on the Defendant to release water at Yarrawonga Weir at 10,000 ML/day during October 2017, the Defendant did not comply with that obligation.

If contrary to the foregoing I had found that the Defendant failed to act in accordance with the AOP because it commenced overbank transfers in October 2017 instead of November 2017, I would make the further finding that that failure was not a breach of the duty of care alleged to be owed to the Plaintiffs.
In considering this particular of breach, and many of those which follow, sight must not be lost of the principles set out at [599]-[603]. The issue which the Court has to decide is not whether the Defendant has or has not acted in accordance with the 2017/2018 AOP. The Court has to decide whether the Defendant failed to act with reasonable care to protect the Plaintiffs from pure economic loss. The two issues are different. The answer to the first does not determine the answer to the second.
The Plaintiffs argue this part of the case as if their claim is for breach of a contract to which they and the Defendant were counterparties. They submit that not acting in accordance with the AOP was a failure to take “appropriate precautions” against the risk that the Plaintiffs would suffer pure economic loss. The submission can be understood as a reference to the requirement that the Defendant take a precaution against the risk of harm (i.e. economic loss to the Plaintiffs) which a reasonable person in the position of the Defendant would have taken for the purposes of s 43(1)(c) of the ACT Liability Statute. Beyond that submission, the Plaintiffs do not address how a failure to act in accordance with the AOP was a breach of the Defendant’s duty of care to the Plaintiffs.
As set out above, the Agreement records the agreement between the States about sharing the water resources of the Murray-Darling Basin between the States. The O&O Document and the AOP are derivative parts of that agreement. Assuming the AOP imposed prescriptive obligations on the Defendant, adherence to some of those obligations may have been to the financial detriment of irrigators in New South Wales and Victoria especially where they relate to taking of measures to minimise the risk of system shortfall so that South Australia received its entitlements under cl 88 of the Agreement. To put it the other way, a failure to adhere to some of the obligations in the AOP may have been to the financial benefit of irrigators in New South Wales and Victoria.
The point is demonstrated by the facts in this case. If the Defendant had adhered to a requirement that there be no overbank transfers until November 2017, the advent of the mid-October rainfall meant that there was no need for any overbank transfers in the 2017/2018 water year. Yet at the end of October 2017 total inflows for June, July, August, September and October were less than 84% of historic flows. On the Plaintiffs’ approach, the river system was between the “dry” and “moderate” scenarios and the AOP “required” that overbank transfers still be made in November and December 2017 (see the graph at [685] above). Given the rainfall in October, adherence to that requirement may have been to the Plaintiffs’ financial detriment.
By way of further example, the Plaintiffs alleged that the Defendant’s duty of care required the Defendant to depart from the 2018/2019 AOP because in the circumstances of the 2018/2019 water year adherence to the AOP was to their financial detriment.
These are just examples. They do no more than illustrate the modest proposition that a failure to act in accordance with the AOP did not automatically constitute a breach of the Defendant’s alleged duty of care to the Plaintiffs. In each instance it is still necessary for the Plaintiffs to prove that the Defendant failed to act with reasonable care to protect the Plaintiffs against pure economic loss.

The parties’ Written Submissions contain a debate about the terms of cl 10 of the O&O Document and the Defendant’s obligations in relation to the results of the tests. It is not clear whether the Plaintiffs contend that the runs of the Fortran Assessment Model on 20 September 2017 constituted a formal update of the AOP in accordance with which the Defendant was thereafter required to carry out its functions pursuant to cl 10(1)(d) of the O&O Document. They would be right not to. One need merely observe that three different spreadsheets resulted from the testing and that it would not have been possible for the Defendant to adhere to all three. It would not even have been possible for the Defendant to adhere to the testing for the “dry” scenario because there were two results based on inconsistent assumptions about overbank transfers. The O&O Document required a review and update of the AOP in October (after the wettest months of the year). The testing which occurred in September was a different exercise. The output from the testing was given to the members of the WLWG and discussed at the meeting on 26 September 2017. As a separate item of operations planning, the minutes record that a draft updated AOP would be provided in the following weeks.
The Plaintiffs submit that pursuant to cl 12.5(d) of Appendix A to the O&O Document the Defendant was obliged to be “guided by” the September testing when making a decision about the timing of commencement and rates of transfer of operational water from Hume Reservoir to Lake Victoria. The obligation in cl 12.5(d) (and also cl 12.1(d)) is directed to the “Monthly Operating Plan”. The Defendant’s witnesses explained that the “latest run” of the Fortran Assessment Model constituted the Monthly Operating Plan. It is doubtful that the testing which was undertaken in September 2017 answered that description. It did not address all six of the scenarios. For the “dry” scenarios there were two runs based on inconsistent assumptions. They could not both be the latest run.
In any event, the obligation in cl 12.5(d) was no more than to be “guided by” the latest run of the Fortran Assessment Model. Clause 12.5(d) required the Defendant to be guided by a number of things of which the latest run of the Fortran Assessment Model was only one. The Defendant was also to be guided by the Lake Victoria Operating Strategy and the WLWG. There was no provision in the O&O Document which obliged the Defendant to “adhere to” the test results obtained on 20 September 2017.

Given the fact that the tests result for the “moderate” scenario assumed that overbank transfers would occur in October 2017, the Plaintiffs’ argument that the Defendant failed to adhere to the mid-September testing depends upon an anterior finding that the “dry” scenario applied. The difference between the “dry” scenario and the “moderate” scenario was not material for the 2017/2018 AOP which was approved in July 2017 but it was important for the testing conducted on 20 September 2017.
For the reasons given above, the Plaintiffs have not proved that the inflows which had occurred as at the end of September obliged the Defendant to adhere to the test results for the “dry” scenario as opposed to the “moderate” scenario. The Defendant evidently considered the test results for both scenarios which assisted it to make future decisions. The effect of the Plaintiffs’ submissions is that by the end of September 2017 the Defendant was obliged completely to ignore the test result for the “moderate” scenario, based as it was on the assumption that overbank transfers would commence in October 2017. There is no basis for that suggestion. The Defendant had gone to the trouble of carrying out tests for the “moderate” scenario. Inflows up to the end of August 2017 had hovered between the “dry” scenario and the “moderate” scenario and it was not known which way they would go for the remainder of the year. The Defendant was entitled to have regard to all the test results which it had obtained.
The fact that two scenarios were tested also highlights the flaw in the Plaintiffs’ contention that the Defendant was obliged to adhere to the test results obtained on 20 September 2017. How does one “adhere to” a test result, let alone multiple test results?
Having regard to the purpose for which the Fortran Assessment Model was run on 20 September 2017, there is no basis to find that the Defendant failed to be guided by the results, including the result which assumed the “dry” scenario and overbank transfers. The tests were conducted for a proper purpose. The results were evidently considered and discussed with the WLWG on 26 September 2017. The Plaintiffs have not proved that anything further was required of the Defendant acting with reasonable care.

Following on from the second problem, a further problem with the Plaintiffs’ submission is that it does not address the facts. The Plaintiffs submit that the matter being considered at the end of September 2017 was the storage level of Lake Victoria as at 31 December 2017. They say it was a multi-month matter best addressed by the Fortran Assessment Model which had better long-term functionality. The submission is oversimplified and cannot be accepted.
As at 28 and 29 September 2017 both the Fortran Assessment Model and the Daily Operations Spreadsheet demonstrated that Lake Victoria would have insufficient water as at 31 December 2017. Both models demonstrated that overbank transfers would be required to reduce the risk of system shortfall and a failure to deliver South Australia’s entitlements over summer. By 29 September 2017 the matter being considered was the precise date upon which to commence overbank transfers. The Fortran testing had been undertaken on 20 September 2017 using the assumption alternatively that overbank transfers would commence in October (if “moderate” conditions were assumed) or November (if “dry”). That was as far as the Fortran Assessment Model could take it. The Fortran Assessment Model lacked the functionality to provide any more precision.
The Plaintiffs say that even if overbank transfers could reasonably have been commenced in October 2017 it should have been in “mid” October and not on 4 October. By 29 September 2017 the Defendant was considering the precise date when the release of water at Yarrawonga Weir would increase above 10,000 ML/day. Although that decision would be informed by the year-end storage level of Lake Victoria, it is too simplistic to say that it was a multi-month decision for which the Fortran Assessment Model was the best tool. The daily interval of the Daily Operations Spreadsheet provided necessary functionality which the Fortran Assessment Model did not have. Mr Bishop said that the Daily Operations Spreadsheet was “effectively a model of the system in and of itself that’s on a daily time-step, with considerably more detail and ability to manipulate things in a short time-scale”.
In any event the Plaintiffs’ contention that the Daily Operations Spreadsheet lacks multi-month functionality is not correct. It is true that the Daily Operations Spreadsheet did not provide forecasts for the whole of a 12-month period, but it did provide forecasts for approximately three months in advance. At the end of September 2017 the Daily Operations Spreadsheet was capable of forecasting the storage level of Lake Victoria as at 31 December 2017.
In cross-examination the Defendant’s witnesses readily agreed that the Fortran Assessment Model and the AOPs produced using it were critical tools for the conduct of the Defendant’s river operations functions. It may be accepted that the Fortran Assessment Model was indispensable in the sense that it had functionality which none of the Defendant’s other tools had. However, no Defendant witness made a concession about the superiority of the Fortran Assessment Model in relation to the decision about which day of the month to commence overbank transfers, either generally or specifically in the circumstances which pertained at the end of September 2017.
A further problem with this part of the Plaintiffs’ case is that their submissions do not reflect the evidence about the nature, content and operation of the Daily Operations Spreadsheet. The Plaintiffs sought to portray the Daily Operations Spreadsheet as a spreadsheet which is created anew each day, populated each time with both inputs and formulae “adopted” by whoever happened to be the Duty Operator that day, for whatever reason that person thinks best. It was contrasted with the Fortran Assessment Model which is a bespoke computer model.
Ms Hickey accepted that the Fortran Assessment Model is a “computer” model although she resisted the suggestion that the Daily Operations Spreadsheet is not. It is an excel spreadsheet which performs its computations within that format. The evidence establishes that the Daily Operations Spreadsheet is no less bespoke or computerised than the Fortran Assessment Model. The Fortran Assessment Model performs calculations upon inputs using formulae which have been coded into the programme which operates the model. It has been developed over time by programmers at the Defendant (and its predecessors) specifically for the purpose of operating the Murray-Darling Basin. The Daily Operations Spreadsheet is the same, except that the formulae have been set up using excel software. The Defendant’s practice of maintaining the Daily Operations Spreadsheet is described above. The Duty Operator does not make a new spreadsheet each day. Each day the Duty Operator takes the “living document” from the previous day and updates the inputs with actual data received the previous day. Whilst the Duty Operator is responsible for updating the spreadsheet each day, it is a shared resource. The River Operations Team can adjust the loss rates, but there is no basis in the evidence for the Plaintiffs’ suggestion that that might be done each day in an isolated and idiosyncratic way by the individual Duty Operator.
The Plaintiffs seek further to undermine the reasonableness of the Defendant’s use of the Daily Operations Spreadsheet by pointing out that on 26 September 2017 the WLWG was told that information presented to it was “subject to rain events”. It is not clear from the minutes precisely what was said to be subject to rain events. In any event, the statement was no more than the general caveat which would be expected and which, if not express, would have been implicit in any forward-looking statement made by a river operator. Any such statement applied equally to the 2017/2018 AOP and the results of the testing carried out on 20 September 2017. In any event, the Plaintiffs have not proved that there was a rain event, either actual or forecast, which warranted any previous assumptions to be abandoned or any approval from the WLWG to be revisited. To the extent that the Plaintiffs rely on the Climate Outlook which was issued by the Bureau of Meteorology on 28 September 2017 there is no evidence that a river operator acting with reasonable care would have had regard to the rain information in that document given the Bureau’s limited ability to forecast rain beyond eight, or even four, days.

There is no dispute that the loss rates assumed for the Daily Operations Spreadsheet on 28 and 29 September 2017 were higher than those assumed when the 2017/2018 AOP was prepared using the Fortran Assessment Model or when the testing was undertaken on 20 September 2017.
The mere fact that the loss rates assumed for the Daily Operations Spreadsheet at the end of September 2017 were different to those assumed when using the Fortran Assessment Model does not mean that they were not loss rates which a river operator acting with reasonable care could have assumed. The Plaintiffs have put forward some reasoning to demonstrate otherwise but none of the reasons are material. The Plaintiffs point out that the Fortran Assessment Model modelled the full year, but they have not demonstrated how that made a difference to the use of assumed loss rates in the modelling which actually occurred on 28 September and 29 September 2017, especially when the period being modelled was only to the end of 31 December 2017. The superiority of the relevant loss formulae or assumptions in the Fortran Assessment Model is not demonstrated merely by the submission that the Fortran Assessment Model is “bespoke hydrologic modelling software”. As explained above, the Daily Operations Spreadsheet is no less “bespoke”, “hydrologic”, “software” or “modelling”. There is no evidence to prove the asserted superiority of one model over the other, either generally or for the specific purpose of undertaking modelling in September 2017 of the storage level of Lake Victoria as at 31 December 2017. The potential for human intervention was the same for each model, having regard to the post-processing manual adjustments performed on the outputs from the Fortran Assessment Model. Both models incorporated a degree of human judgement, in each case made collaboratively by the River Operations Team.
The Plaintiffs refer to a passage in Mr Bishops’ First Affidavit where he says that loss estimates in the Daily Operations Spreadsheet are less accurate than in the Fortran Assessment Model. The relevant passage of Mr Bishop’s Affidavit was directed to a different issue. He was explaining how each model uses different data to calculate losses which have already occurred. Historic losses were not relevantly applicable to the forecast of future losses and the year-end storage level of Lake Victoria in the modelling being undertaken in September 2017.
Although this is largely a hydrological issue, there is no evidence from Dr Ladson to support the proposition that the future loss rates which were assumed in the Daily Operations Spreadsheet were unreliable. He identifies the difference between the loss assumptions for each model. He does not say that it was not reasonable for the Defendant to have used the loss assumptions which it did use for the Daily Operations Spreadsheet on 28 September and 29 September 2017.
One apparent reason why the loss rates used in the Daily Operations Spreadsheet may have been expected to be more reliable than those used for the 2017/2018 AOP and for the Fortran testing on 20 September 2017 is that they were based on more up-to-date data. The AOP loss rates were assumed in July and the loss rates assumed for the 20 September testing (if different) were assumed in mid-September. Mr Bishop explained that rates for short-term future losses in the Daily Operations Spreadsheet were in part derived from data about recent actual losses. They were also derived in part from forecast temperatures, about which the Bureau of Meteorology had issued a Climate Outlook on 28 September 2017 and which was to some extent useful information for the Defendant (cf future rain). As referred to above, temperature affects losses through evaporation and through the demand for water by those entitled to take water from the Murray River. As at 28 and 29 September and 3 October 2017, the Daily Operations Spreadsheet was more up to date and, to that extent, may have been a more reliable forecast of year-end Lake Victoria storage levels.
The issue addressed in the preceding paragraph does not warrant a finding that the Daily Operations Spreadsheet was a model which was superior to the Fortran Assessment Model either generally or for any particular purpose. The point is that any comparison between the two models is complex and the Plaintiffs’ reliance upon a departure from the loss assumptions used for the Fortran Assessment Model is not a sure foundation for a finding that it was negligent to use the Daily Operations Spreadsheet as the Defendant did.
Each party put forward competing figures to demonstrate that the losses which were actually experienced in October, November and December 2017 turned out to be (or not to be) greater than the higher loss rates assumed in the Daily Operations Spreadsheet at the end of September 2017. Equally informed by hindsight, these figures are not relevant to the question whether the assumption of any particular loss rates was reasonable at the end of September and early-October 2017. They do however show that the actual losses were very large and the difference between the various figures are also very large. In that context, the Plaintiffs’ emphasis on the 273% difference between one particular loss rate assumed for the mid-September Fortran testing compared to what is said to be the equivalent figure in the Daily Operations Spreadsheet on 28 September 2017 is deprived of some of the superficial significance it might otherwise have had.
The Plaintiffs’ final submission on this issue is to point out that Ms McCann was the Duty Operator who prepared the Daily Operations Spreadsheet on 28 and 29 September 2017. They submit that Ms McCann has not been called to explain the inputs or other factors affecting Lake Victoria’s storage, including the loss rates which she “apparently manually inputted or configured in the Excell document”. The failure to call Ms McCann is not explained. The Plaintiffs submit that an inference may readily be drawn that the Daily Operations Spreadsheet which was prepared on 28 September and carried through to 29 September 2017 “did not address the radically higher loss rates” compared to the loss rates assumed for the modelling which was undertaken on 20 September 2017 using the Fortran Assessment Model. They further submit that the inference is available that “the loss rates and inputs in any such testing that may have been done were not reasonable or supportable”.
The force of the Jones v Dunkel submission depends upon the Plaintiffs’ view of the Duty Operator preparing each day a new Daily Operations Spreadsheet by idiosyncratically adopting assumptions about numerous inputs, alone and unsupervised. The submission assumes that once prepared, the Daily Operations Spreadsheet is not loaded onto the shared drive where it is available to and used by other members of the River Operations Team, like Ms Greenwood who evidently consulted it and emailed a graph from it to the Barmah-Millewa Operations Advisory Group at 5pm on 29 September 2017. The submission also has in it the unspoken premises that the assumptions about losses in the Daily Operations Spreadsheet had previously conformed to the assumptions about losses used for the testing with the Fortran Assessment Model which was undertaken on 20 September 2017, and that a radical change was made to those assumptions on 28 September 2017. It is that context the Plaintiffs submit Ms McCann should have been called to explain.
Mr Bishop’s unchallenged evidence about the River Operations Team systematically and routinely updating of the Daily Operations Spreadsheet using data newly received, and the role of the Duty Operator in that quotidian process, undermines the Plaintiffs’ Jones v Dunkel submission. Ms McCann did not have the agency in the process for which the Plaintiffs contend, nor was the Daily Operations Spreadsheet susceptible to the degree of interference by the Duty Operator which is implicit in the Plaintiffs’ submission. In particular, apart from that mechanically consequent on the input of data from the previous day, it has not been shown that any change was made to the assumptions about losses in the Daily Operations Spreadsheet on 28 or 29 September 2017 by Ms McCann or anyone else. It is not known when the loss assumptions diverged from those used for the Fortran Assessment Model or indeed whether there was ever a divergence from a previously common position. The Plaintiffs’ submission about preparation of the Daily Operations Spreadsheet on 28 and 29 September 2017 does not rise above speculation.
More importantly, the Plaintiffs bear the onus of proof. Whilst a Jones v Dunkel inference may arise against a defendant, in the absence of evidence from Dr Ladson (or anyone else) that the loss assumptions in the Daily Operations Spreadsheet were not such that a river operator acting with reasonable care could have used, there is no real issue raised in the case which calls for an explanation by the Defendant. Having explained the system for the preparation of the Daily Operations Spreadsheet and the loss assumptions used in it, and there being nothing to suggest a departure from that system on 28 or 29 September 2017, the Defendant’s failure to call the Duty Operator who implemented the system on those days is explicable.
The Plaintiffs have sought to discharge their onus of proof on this alleged breach of the Defendant’s duty of care not by calling expert evidence but by arguing the objective facts. They have not demonstrated that there was a defect in the modelling in the Daily Operations Spreadsheet on 28 or 29 September 2017 by reason of the assumed loss rates. They have not proved on the balance of probabilities that the Defendant acting with reasonable care would not for that reason have relied on the Daily Operations Spreadsheet.

In order to understand this part of the Plaintiffs’ case, some further factual context is necessary. Lake Victoria’s storage capacity is 677 GL. The modelling in the Daily Operations Spreadsheet on 28 September 2017 showed that the storage level of Lake Victoria would be just below 400 GL on 31 December 2017. The lake would not be full.
In their Reply Submissions the Plaintiffs submit that the Defendant has not explained what was “wrong” with a forecast year-end storage level for Lake Victoria of less than 400 GL. The question is not whether anything was wrong with the forecast storage level but what would have been the view of a river operator acting with reasonable care. On that issue, the evidence is clear. Given the high downstream demand for water over summer and the limited options for transferring water to Lake Victoria during the summer months, the Defendant’s policy was to have Lake Victoria full on 31 December 2017 so as to reduce the risk of system shortfall and not being able to deliver South Australia’s entitlements. The 2017/2018 AOP, which the Plaintiffs do not criticise, contemplated that Lake Victoria would be full or close to full in December in all scenarios other than the “extreme dry” scenario. That contemplation is depicted in the graph at [549] above. The Guidance Document specified that one aim of the Defendant was to have Lake Victoria full by the end of December. As the Plaintiffs submit for other aspects of their case, a river operator acting with reasonable care would have had regard to each of these documents.
The Defendant further points out that Dr Ladson approached his analysis of the circumstances which confronted the Defendant at the end of September 2017 on the basis that a forecast Lake Victoria year-end storage level of 400 GL was deficient. In cross-examination he accepted that it was “unacceptable”. Given the Defendant’s submissions about the widespread unreliability of Dr Ladson’s evidence, the Plaintiffs say that the Defendant’s reliance on Dr Ladson’s evidence on this topic is cherry-picking. That may be right but even so the more important question is whether Dr Ladson’s view on this issue ought to be accepted. It accords with the Defendant’s policy and the documents referred to in the preceding paragraph. It does not suffer from the difficulties with some of Dr Ladson’s other evidence, including infection by hindsight and an inability to address non-hydrological matters. The Plaintiffs do not refer to any substantive reason why Dr Ladson’s evidence ought not to be accepted on this particular issue.
By the end of September 2017 it was open to a river operator acting with reasonable care to consider that water needed to be transferred to Lake Victoria. In the language used by the parties at the hearing, something needed to be done.

To consider the Defendant’s complaint about the Plaintiffs’ use of hindsight for this part of their breach case, it is essential to identify precisely the risk of harm for which the Plaintiffs contend because the identification is necessary to assess what would have been a reasonable response. [191] It is not the Plaintiffs’ case that the Defendant breached its duty of care by delaying unnecessarily the release of water from the Menindee Lakes and thereby causing a loss of water to the overall system and, ultimately, the Plaintiffs. For example, it is not alleged that the system-wide water resources were lower than they would have been because more water was lost to evaporation from the Menindee Lakes than would have occurred had the water been transferred to Lake Victoria as early, as quickly and at as high a rate as possible. Nor do the Plaintiffs put forward a case based on loss caused by the Defendant’s failure to call on the Goulburn IVT account in October 2017. The opportunity to call on the Goulburn IVT account is said to be a circumstance in which the alleged breach of duty occurred, but the failure to do so is not a freestanding claim of negligence.
The Plaintiffs’ case is that the Defendant breached its duty of care when it exercised the release power and commenced overbank transfers at Yarrawonga Weir on 3 October 2017. It says that a river operator acting with reasonable care would not have commenced the overbank transfers on that date because there were two measures which he or she would have taken instead.
An essential part of the Plaintiffs’ case, not only for breach but also for causation, is the proposition that a river operator acting with reasonable care would have formed a view at the end of September or in early-October 2017 that the two alternative measures would be sufficient to address the forecast “unacceptable” storage level of Lake Victoria. Absent that view, it was not negligent for a river operator to have commenced the overbank transfers on 3 October 2017 to minimise the risk of system shortfall.
For the purposes of his First Report, Dr Ladson was instructed to consider alternative measures which could have been used instead of commencing overbank transfers in early-October. In response, Dr Ladson expressed the opinion that there were alternative measures, namely faster releases from the Menindee Lakes and a call on the Goulburn IVT account in early-October 2017. Dr Ladson identified specific parameters for each alternative measure as to timing and volume. Had the first measure been taken in accordance with his parameters (faster transfers from the Menindee Lakes), Dr Ladson calculated that the storage level of Lake Victoria would have been increased by up to 37.8 GL. Had the second measure been taken (call on the Goulburn IVT account), Dr Ladson calculated that the storage level of Lake Victoria would have been increased by up to 29 GL. In combination the calculated increase was up to 66.8 GL. The Plaintiffs rely on Dr Ladson’s analysis to prove that a river operator acting with reasonable care would have used the two alternative measures instead of commencing overbank transfers on 3 October 2017.
The calculated gains from the alternative measures (66.8 GL) would not have been seen on 28 September 2017 to be sufficient to ensure that Lake Victoria would be filled having regard to the forecast figures taken from the modelling in the Daily Operations Spreadsheet. Even when confined to the immediate period between 1 and 18 October, at that time the modelling showed that the forecast decline in the storage level of Lake Victoria would be 87 GL.
Confronted with that analysis in cross-examination, Dr Ladson accepted that his opinion about the sufficiency of the two alternative measures was based on hindsight. To identify the alternative measures, Dr Ladson had not considered the problem which prospectively confronted the Defendant at the end of September 2017. Instead he used historic records to work out the actual decline in the storage level of Lake Victoria between 2 October and 18 October 2017. It was 43 GL. He then identified the parameters for the alternative measures to supply that deficiency. On the basis of that analysis the Plaintiffs now submit that on 1 October 2017 (or three days before) a river operator acting with reasonable care would have ordered the release of water from the Menindee Lakes at 2,500 ML/day. The 2,500 ML figure has no relevance to the facts in the case other than Dr Ladson’s calculation of how much water would have been required to eliminate the fall in Lake Victoria’s storage level which subsequently occurred. Dr Ladson’s period ends on 18 October 2017 because that is when it started to rain.
Even if the two alternative measures were available at the end of September 2017, the Plaintiffs have not proved that a river operator acting with reasonable care would have formed the view that overbank transfers were not required. In the face of a forecast 87 GL decline of Lake Victoria’s storage level by 18 October 2017 and the 260 GL deficiency forecast for the end of the year, the two alternative measures identified by Dr Ladson would not have been viewed as capable of addressing the reasonably foreseeable risk of Lake Victoria having insufficient water on 31 December 2017. In the language of s 43(1)(c) of the ACT Liability Statute, and as reflected in MFI-38, neither alternative was a precaution which a reasonable person in the position of the Defendant would have taken, either alone or in combination.
The Plaintiffs’ argument based on the sixth alleged failure by the Defendant fails at the outset.
In fact, for the reasons which follow neither of the two alternative measures was available at the end of September 2017.

The proposal referred to in this passage was one where bulk transfers of operational water would not be made from the Menindee Lakes to Lake Victoria, or not as fast as might otherwise occur, so that more water would be held back with which to sustain minimum flows in the lower Darling River for a longer period of time. Whilst the higher evaporation rate would mean that holding water back in the Menindee Lakes would reduce the overall availability of operational water to the States (estimated to be 30-50 GL) it might be possible to make an agreement with an environmental water holder to offset the reduced water by supplying an equivalent amount elsewhere in the Murray River in exchange for the environmental benefits of prolonged flows in the lower Darling River.
On 27 September 2017 an officer of WaterNSW sent an email to an officer of the NSW Environmental Water Holder. Two officers from the Defendant were copied in, Ms Hickey and Ms Greenwood. The emails addressed the proposal. The email attached modelling which suggested that delay of the bulk transfers would extend the period of minimum flows for two months. If the delay was implemented, the overall system was modelled to lose 50 GL of water, which the NSW Environmental Water Holder was requested to make good through the accounting process. The WaterNSW officer said he was keen to discuss further “in the near future, as [the Defendant is] making preparations to call water from Menindee in the next week or so”. There is no evidence that WaterNSW received a reply before 4 October 2017.
On 29 September 2017 Ms Hickey sent an email to the members of the WLWG in which she stated that the Defendant was planning to commence transfers of water from the Menindee Lakes to Lake Victoria “within the next fortnight”. She said that the releases would be no greater than 1,800 ML/day and the rate of increase would be managed to ensure compliance with the maximum rate permitted under cl 10.2 of the O&O Document. The emails stated:

Attached to the email were some media talking points given “sensitivity of the local community regarding flows in the system, especially as summer approaches”.
As at 30 September 2017 the storage level in the Menindee Lakes was 658 GL which meant that the Defendant was permitted to transfer water to Lake Victoria under cl 95 of the Agreement. Approximately 175 GL was available before the 480 GL minimum was reached and the remaining water in the Menindee Lakes would cease to be available to the Defendant. As set out above, on Sunday 1 October 2017 680 GL became the new specified level for Lake Victoria, up from 500 GL on 30 September 2017. Under cl 12.2 of Appendix A to the O&O Document a harmony release of water from the Menindee Lakes was “triggered”.
Monday 2 October 2017 was a public holiday. On 3 October 2017 Mr Reynolds returned from leave and had a morning meeting with Ms Hickey about the Menindee Lakes. As at that date no agreement has been reached between WaterNSW and the environmental water holders about a delay in transferring water from the Menindee Lakes to Lake Victoria. As it turned out, an agreement was reached during November 2017.
On 3 October 2017 the Defendant and WaterNSW worked on a joint press release about the release of water from the Menindee Lakes.
At 3:22pm on 4 October 2017, Adam McLean, who described himself as “Duty Officer”, sent an email to the members of the River Operations Team, including Ms Hickey, in which he stated that he had “placed an order” with WaterNSW to target a flow of 1,800 ML/day at Weir 32, “commencing as soon as practicable (and following the normal rates of rise)”. WaterNSW had responded by telling the Defendant that the releases would commence on 7 October in order to give the locals time to move their pumps and make other preparations.
At 5:45pm on 4 October 2017 Ms Hickey sent an email to the members of the WLWG in which she informed them that an order for bulk transfers from the Menindee Lakes had been placed and that the transfers would commence on 7 October 2017. She stated that “the flows will be ramped up over several days in accordance with the max rates of rise permitted in the lower Darling”. Mr McLean was sent a copy of the email.
On 5 October 2017 the Defendant and WaterNSW issued a joint press release. The press release stated that WaterNSW would start releasing water from the Menindee Lakes on 7 October 2017. It also stated that leaving a greater proportion in the top lakes would provide greater security for local water users, including the Broken Hill town supply.
As for the flow of water which actually occurred, there are different figures in different places in the evidence but the burden is the same. Using the figures in the Ready Reckoner, [192] up until 6 October 2017 WaterNSW released water from the Menindee Lakes around the minimum daily rate of about 370 ML/day. On 7 October it was increased to 406 ML/day. On 8 October it was increased to 640 ML/day. Each day thereafter it was further increased by approximately 300 ML until 12 October when 1,804 ML was released. From 13 October onwards the daily rate hovered around 1,800 ML/day.

The Plaintiffs’ essential complaint is that the Defendant ought to have directed WaterNSW to commence the release of water from the Menindee Lakes from 1 October 2017 (not from 4 October 2017) and that the direction should have been given in a way which ensured that notice was given to the locals three days before 1 October 2017. Had that happened, water would have flowed from the Menindee Lakes on 1 October 2017 (not on 7 October 2017).
In the Plaintiffs’ favour it may be assumed that there is always an advantage for the river system as a whole to have as much water transferred from the Menindee Lakes to Lake Victoria as soon as possible, that advantage being the loss of less water to evaporation. However, there are also disadvantages. Leaving aside local political sensitivities, retention of water in the Menindee Lakes provides a larger pool with which to sustain a minimum flow down the lower Darling River for longer (the minimum flow was specified in cl 10.3 of Annexure A to the 2017/2018 O&O Document to be 200 ML/day during October and 300 ML/day in November and December). That is so even though more water will be lost to evaporation whilst it is held back, waiting to be released at the minimum flow rate. It may be inferred from cl 10.3 of the O&O Document and the minutes of the WLWG meeting on 26 September 2017 that there is a recognised environmental benefit from extending the period of minimum flow in the lower Darling River. Section 8.6.1 of the 2017/2018 AOP referred to benefits for native fish. There is also evidence of an operational benefit for the supply of water to lower Darling River irrigators, especially in Dr Davis’s Second Affidavit. In the three years preceding 2017 there had been a number of occasions when the Darling River had ceased to flow which had caused difficulties for delivering water to irrigators.
It is not necessary to know exactly what these benefits are or how they might be evaluated or quantified. Nor is it necessary to weigh them against the advantage of reduced evaporation losses (even if that were possible). The clear inference from the WLWG minutes is that a delay in the transfer of water to Lake Victoria was considered to be an “opportunity”. The Defendant and the WLWG recognised the potential for the benefits and sought to realise them through discussions with the relevant environmental water holders. The Plaintiffs emphasise that no agreement with the environmental water holders was in place on 1 October 2017. That is correct but irrelevant. The discussion at the WLWG meeting contemplated delay in the face of the prospect of an agreement. It is not necessary to know whether the prospect was good or bad, only that the WLWG considered it worth the wait.
The WLWG minutes make clear that the timing for the commencement of bulk transfers to Lake Victoria was intricately connected with the pursuit of discussions with the environmental water holders. There is no evidence to support a finding that it was not reasonable for the Defendant to delay the commencement of bulk transfers, especially when the matter was discussed at the WLWG meeting in the terms recorded in the minutes. Dr Ladson does not say that it was not reasonable for the Defendant to delay. This is one particular circumstance where Dr Ladson accepted that, whilst he was qualified to express a view about increased evaporation from a hydrological perspective, he was not qualified to express a view about environmental issues or the other issues which the WLWG and the Defendant took into account. He could not weigh one against the other.
Nor is there a basis to find that it was not reasonable for the Defendant to delay in giving notice of the upcoming transfer of water, so long as there was a prospect of reaching an agreement which might result in the transfers not occurring. The bulk transfers of operational water would raise the level of the lower Darling River. The Defendant’s Reference Manual contemplated that the Defendant would give notice. A purpose of giving notice was to give locals time to move pumps and stock to higher ground. Expense might be incurred in making preparations. The transfers might also cut some low-lying river crossings. The BOC had set some objectives and outcomes for the Defendant to achieve in relation to communications and information management. [193] It may be accepted that the Defendant’s usual practice was to pre-empt such transfers by giving notice to the locals three days before the earliest commencement date (i.e. 1 October). It is also clear that there were local political sensitivities and it was thought important that communications with the public be handled with care. The Defendant could reasonably have considered that it was appropriate to delay making an announcement which might subsequently turn out to be unnecessary.
The Plaintiffs point out that Ms Hickey does not particularly remember the issue of a potential agreement and does not identify it as a contemporaneous reason for deciding the commencement date for the transfers from the Menindee Lakes. Nor does Mr Reynolds. Given the clarity of the minutes and the coherence of the other documentary evidence, including the subsequent email communications, the lack of oral evidence does not warrant a different conclusion to that set out above. The contemporaneous documents are in any event the best guide.
There is nothing in the terms of the O&O Document, the 2017/2018 AOP or any other document upon which the Plaintiffs’ rely which required the Defendant to commence bulk transfers from the Menindee Lakes on 1 October 2017 irrespective of any environmental consideration, any off-setting operational benefit hoped to be achieved from an agreement with the environmental water holders and the views of the WLWG evident from the meeting on 26 September 2017.
The highwater mark of the Plaintiffs’ case is the statement in cl 12.2(d)(i) of the O&O Document that bulk transfers “should…be triggered” on 1 October 2017. The Plaintiffs submit that the statement is to be construed as an obligation to commence bulk transfers on that date. For its part, the Defendant submits that “triggered” implies something less than a mandatory start date. The Defendant’s construction is to be preferred having regard to the context of the O&O Document as a whole which seemed rarely to intend to constrain the Defendant’s conduct as rigidly as the Plaintiffs contend for cl 12.2(d)(i). In any event, cl 6(2)(g) of the O&O document stated that any obligation in cl 12.2.(d)(i) was qualified by reasonable practicability having regard to uncertainty about the environmental consequences of particular activities. Further, cl 7(1) permitted the Defendant to refer potential non-compliance with cl 12.2 to the WLWG for advice. The issue was discussed at the WLWG meeting on 26 September 2017. The Defendant’s subsequent conduct conformed to that contemplated by the WLWG’s “advice”. Whether or not there remained a technical breach of cl 12.2 of the O&O Document, there is no basis to say that the Defendant did not act with reasonable care.
That is enough to determine this part of the claim against the Defendant. However, there is a more fundamental flaw in the Plaintiffs’ argument. The environmental benefit from maintaining a minimum flow in the lower Darling is demonstrated by the fact that in November 2017 the environmental water holders considered it sufficiently valuable to agree to “pay” for the benefit. The payment was an accounting adjustment by which there would be a reduction in their entitlement to environmental water elsewhere in the river system. The accounting adjustment essentially meant that additional operational water would become available elsewhere in the system. It appears to have been a sort of swap of operational water lost through evaporation in the Menindee Lakes for environmental water elsewhere in the system. The implications from the swap for the conduct of irrigated agricultural production are not known but there is reason to consider that the implications may have depended on the location of the properties of various irrigators. Specifically, prolongation of minimum flows in the lower Darling River would be a benefit to irrigation in that region, but not elsewhere. Additional operational water in another part of the system may have benefited irrigation elsewhere including by freeing up channel capacity with which to avoid future overbank transfers.
This demonstrates the general proposition, described at [599]-[603] above, that the Plaintiffs do not prove a breach of a duty of care to protect the Plaintiffs from pure economic loss simply by demonstrating a failure to adhere to a provision of the 2017/2018 AOP, the O&O Document, the Agreement or even the Commonwealth Statute. In any particular circumstance, the question is whether the Defendant failed to act with reasonable care to protect the Plaintiffs from pure economic loss. On the Plaintiffs’ case they had a potential economic interest in an early transfer of water from the Menindee Lakes to Lake Victoria in October 2017. Given the prospect of an agreement with environmental water holders, they may also have had an economic interest in a delay. Protection of the Plaintiffs’ economic interests was not the same as adherence to any particular document.
This particular alleged breach is also a singular demonstration of why the Defendant did not owe the Plaintiffs a duty of care. In the Queensland Floods case the trial judge found that the flood engineers owed a single duty to all the relevant owners of downstream property to exercise reasonable care in the conduct of flood operations to avoid or minimise the risk of harm to property and the risk of interference with use and enjoyment of property. [194] It was accepted that individual owners of downstream properties might be affected differently by decisions made by the flood engineers, but any conflict between individuals to whom the single duty was owed could be accommodated at the stage when consideration was given to breach of the duty. As Mason J said in Wyong Shire Council v Shirt, a matter to be taken into account when considering whether the defendant acted with reasonable care is any other conflicting responsibility which the defendant may have. [195]
The situation confronted by the Defendant when considering whether and how quickly to transfer operational water from the Menindee Lakes to Lake Victoria in September 2017 was different. The Defendant had to make a multifaceted decision which might affect a number of interests. Unlike the flood engineers in the Queensland Floods case, the Defendant had more than one objective. Given the environmental considerations, the interests of irrigators on the lower Darling River (including their economic interests), the interests of South Australia, the position of the WLWG and the prospect of greater operational flexibility which an agreement with environmental water holders might provide, it cannot be said that the Defendant owed to everyone who might be affected a single duty to take reasonable care in exercising its river operations functions. The conflicts to which the Defendant might be subject could not be accommodated as part of the enquiry into whether the Defendant breached a single duty by failing to act with reasonable care.
More particularly, the decision about transferring operational water from the Menindee Lakes to Lake Victoria is a concrete demonstration of the intention of the Commonwealth legislature and the States that the Defendant be free to make the decision without being concerned to ensure that the outcome would produce the best financial outcome for the Plaintiffs. Were it otherwise, the Defendant’s ability to operate the Murray-Darling Basin in accordance with the purposes of the Commonwealth Statute and the Agreement would be distorted. The Plaintiffs do not allege that the Defendant owed a single duty to all those who might be affected by the exercise of its river operations functions. They allege a duty owed only to them and the Group Members. That duty would be incongruous with the statutory framework.
There is one further issue raised by this part of the case, namely the uniformity of interest as between those Group Members who held New South Wales water entitlements and those who held Victorian water entitlements. The evidence does not make clear the identity of the environmental water holder (or holders) with which the WLWG hoped an agreement might be made. Given the environmental interests related to the lower Darling River plus the fact that WaterNSW approached the NSW Environmental Water Holder about the proposed arrangement on 27 September 2017, it is likely that the NSW Environmental Water Holder was the entity which eventually agreed to swap environmental water elsewhere in the Murray River system. If so, the overall economic outcome for holders of New South Wales water entitlements might have been improved by such an arrangement but at the expense of the holders of Victorian water entitlements. Under the proposed arrangement water would be lost from the Menindee Lakes when it was a resource which was shared by New South Wales and Victoria and offset by water received from the NSW Environmental Water Holder. The NSW Water Environmental Water Holder presumably held water entitlements issued by New South Wales, in which case the offset accounting may have increased the water available to New South Wales to allocate to all New South Wales water entitlement holders, including the Plaintiffs and some (but not all) Group Members.
The evidence does not address the details of the arrangement proposed on 26 September 2017 and finally agreed in November 2017 in order to make any findings of fact about whether the arrangement benefitted the Plaintiffs and all Group Members uniformly. It is clear, however, that the Defendant may have been confronted by a further layer of conflict. It is another demonstration of why the Defendant was not intended to owe a duty of care to protect any particular class of water entitlement holder from economic loss.

The Plaintiffs do not specify what events must occur before a river operator acting with reasonable care may transfer water from Hume Reservoir. Do the Plaintiffs say that no water may be released from Hume Reservoir until the Menindee Lakes are exhausted by the departure of the last release? Or do they accept that water may be released from Hume Reservoir so long as it is forecast to arrive at Lake Victoria after all water is released from the Menindee Lakes? If the latter, water may be released from Hume Reservoir many days before water is released from the Menindee Lakes given the longer travel time.
The mere fact that water was first released overbank at Yarrawonga Weir on 4 October 2017 and water was not physically released from the Menindee Lakes until 7 October 2017 does not demonstrate that the Defendant did not transfer water from the Menindee Lakes “in preference to” transfers from Hume Reservoir. That simplistic proposition was put to Mr Bishop in cross-examination and he rejected it. Nor can it be said that the Defendant did not give preference to Lake Menindee just because the direction for overbank flows from Hume Reservoir was given on 3 October 2017 and the order for releases from the Menindee Lakes was given on 4 October 2017.
In cross-examination of the Defendant’s witnesses, the proposition was put to them that overbank transfers from Hume Reservoir (via Yarrawonga Weir) commenced on 4 October 2017 whilst bulk transfers from the Menindee Lakes did not commence until 7 October 2017. The Defendant’s witnesses accepted the chronology, but they did not accept a failure thereby to give “preference” to transfers from the Menindee Lakes beyond that which was assumed in the question.
The minutes of the WLWG on 26 September 2017 make clear that the Defendant considered, and the WLWG evidently agreed, that it was necessary to raise the storage level of Lake Victoria by transferring water from both sources. Water from the Menindee Lakes alone would not be sufficient. In those circumstances, the Plaintiffs have not proved that there was a failure to give preference to transfers from the Menindee Lakes merely because of the sequence in which the transfers subsequently occurred.

The Plaintiffs submit that the Defendant breached its duty of care because of the speed with which the release rate was accelerated up to the planned release rate of 1,800 ML/day. Clause 10.2(d) of the O&O Document relevantly specified “target maximum rates of rise” of 500 ML/day. On 4 October 2017 Ms Hickey wrote in her email to the WLWG that the releases would increase “in accordance with the max rates of rise permitted in the lower Darling”. Despite that, the rate of increase in the release rate which occurred from 7 October 2017 was roughly 300 ML/day. Slower increases in the release rate delayed water being transferred out of the Menindee Lakes with the result that more water was lost to evaporation than would have been lost had the release rate been increased more quickly. Dr Ladson has calculated the loss to be 2 GL.
Accepting the general proposition that faster increases in the release rate will reduce evaporation losses, Dr Davis gave evidence that there were three reasons why slower increases in the release rate may nonetheless be preferable in some cases. His reasons are: the need to draw down the bottom lakes within the Menindee Lakes in preference to the top lakes, the need to manage the environmental impacts caused by rapid increases in the release rate and the need to be sensitive to community concerns about releases from the local water source at a time of impending drought.
By way of elaboration on the first reasons, in his First Affidavit Dr Davis described the topography of the Menindee Lakes. Gravity dictates that water may only be released from the bottom lakes at a slower rate than from the top lakes because the bottom lakes have lower elevation relative to the point of entry into the Darling River. The effect of Dr Davis’s evidence is that a release of water from the Menindee Lakes in the first instance has to be at a low rate so that the water from the bottom lakes is used first as contemplated by cl 10.4 of the O&O Document.
There is no sufficient basis to find that the Defendant breached its duty of care because the rate of increase was below the “target maximum” in cl 10.2(d). There are a number of reasons for this.
First, there is no basis in the evidence for the implicit premise of the Plaintiffs’ contention, namely that the Defendant was obliged to ensure that the rate of increase was the maximum rate. As a matter of construction, cl 10.2(d) did not impose that obligation. By specifying a rate of increase to be the “maximum” there was implicit permission to increase more slowly. Qualification of the maximum by the word “target” did not oblige the Defendant to aim to achieve the maximum. The maximum was still the maximum and rates below it were still permitted. Specifying the maximum to be a target may have meant that rates above the maximum were also permitted (so long as they occurred whilst trying to get down to the maximum) but it did not mean that rates below the maximum were not permitted. There is nothing in Ms Hickey’s email at 5:45pm on 4 October 2017 which demonstrates otherwise.
Secondly, apart from cl 10.2(d) the Plaintiffs have not proved that a river operator acting with reasonable care would have implemented the fastest allowed rate of increase. Minimisation of evaporation losses would be served by a faster increase but at the expense of one or more of the three countervailing considerations identified by Dr Davis. It is not possible on the evidence to weigh the competing considerations and say that a river operator acting with reasonable care would not have done what the Defendant did.
Thirdly, cl 10.4 of the O&O Document made clear that a relevant consideration for the rate of increase was the impact of releases on the volume of water in the four individual lakes which make up the Menindee Lakes. On this matter, the O&O Document expressly stated that the Defendant should be guided not just by the WLWG but by WaterNSW. The importance of WaterNSW’s role is not to be overlooked. When storage in the Menindee Lakes falls below 480 GL, control reverts back to WaterNSW and would remain there until the storage again reached 640 GL. The four lakes have different attributes, not only in relation to accessibility but also in relation to evaporation. The condition of the four lakes when storage reached 480 GL would affect their subsequent operation and was presumably an important matter to WaterNSW. A feature of Dr Davis’s evidence about the operation of the four lakes is the extent to which the Defendant worked with WaterNSW in relation to the details of operational decisions concerning the Menindee Lakes. When asked about the rate of increase of the release, Dr Davis said:

Doyle’s Farm Produce Pty Ltd atf Claredale Family Trust v Murray-Darling Basin Authority

At a glance

Court

Decision date

Before

Catchwords

Source

Catchwords

Judgment

Parties

Legislation Cited (28)

TABLE OF CONTENTS

JUDGMENT

Introduction

The Murray River

Operation of the Murray River by the Murray-Darling Basin Authority

Overbank transfers in 2017 and 2018

Irrigators in New South Wales

Pleadings

Representative proceedings

Common questions

Preliminary issues

WITNESSES

Plaintiffs’ lay witnesses

Defendant’s lay witnesses

Mr Reynolds

Dr Davis

Ms Hickey

Mr Bishop

Mr Foreman and Mr Korn

Mr Johnson

Specific submissions about acceptance of a witness’s evidence

Jones v Dunkel

Expert witnesses

Dr Ladson

Other expert witnesses

RAIN

Rainfall and rain forecasting

Short-term rain forecasts

Three-month climate outlook

Four seasons

Temperature

RIVER OPERATIONS FUNDAMENTALS

APPLICABLE LAW

Introduction

Principled approach

State boundaries

Plaintiffs’ submissions on choice of law

Defendant’s submissions on choice of law

Determination – the tort was committed in the Australian Capital Territory

The Defendant’s conduct occurred in the Australian Capital Territory

Not where the decisions were directed nor where they assumed significance

Summary of where the tort was committed

Nonetheless the Civil Law (Wrongs) Act 2002 (ACT) does not apply directly

Civil Law (Wrongs) Act 2002 applies by virtue of s 64 of the Judiciary Act 1903 (Cth)

Section 5D(3)(b) of the Civil Liability Act 2002 (NSW) applies

DUTY OF CARE

The first issue

Plaintiffs’ pleaded case

Duty to avoid a risk of pure economic loss

Principled approach to determining a duty to avoid pure economic loss

Duty to avoid pure economic loss is not generally owed

Duty to be determined by the salient features of the case

The role of assumption of responsibility

The role of indeterminacy

The role of incongruency

Principled approach to determining a duty owed by a statutory authority

Structure of the next part of these reasons

Statutory framework

Overview of regulation

The Commonwealth Statute

Establishment of the Defendant

The Agreement between the Commonwealth, the States and the ACT

The Ministerial Council and the Basin Officials Committee

Water Liaison Working Group

Functions and powers of the Defendant

Objectives and Outcomes Document

Accountability of the Defendant

The NSW Statute and the Victorian Statute

Plaintiffs’ submissions on duty of care