Natural Resources Access Regulator v Lidokew Pty Ltd - [2023] NSWLEC 130 - NSWLEC 2022 case summary — Zoe

NSWNSWLEC

Natural Resources Access Regulator v Lidokew Pty Ltd

[2023] NSWLEC 130

Land and Environment Court (NSW)|2022-10-18|Before: Duggan J

View original source

At a glance

Source facts

Court

Land and Environment Court (NSW)

Decision date

2022-10-18

Before

Duggan J

Source

Original judgment source is linked above.

Judgment (56 paragraphs)

[1]

TABLE OF CONTENTS Nature of proceedings Facts Relevant legislative provisions Essential elements of the offence - Water Take Charges Water demand of the cotton crop for each water year Evidence Prosecution evidence Dr Wayne Meyer Antecedent soil moisture at beginning of cropping Defendant's evidence Prosecutor's submissions Defendant's submissions Findings on crop water requirements Crop growing period Antecedent crop moisture levels Rainfall on Havana North WFIE The IrriSAT computer programme The ±15% rate of uncertainty Rainwater available Evidence Prosecutor's evidence Professor Albert Van Dijk Mr Andrew Falkenmire Defendant's evidence Mr James Purcell Prosecutor's submissions Defendant's submissions Findings on rainfall Stored water availability Evidence Prosecutor's evidence Mr Robert Day Defendant's evidence Mr James Purcell Prosecutor's submissions on stored water availability Defendant's submissions on stored water availability Findings on stored water availability Submissions on evidence as to guilt on amount of water taken Prosecutor's submissions Defendant's submissions Findings on evidence of guilt on amount of water taken The Metering Charges Essential elements of the offence - The Metering Charges Evidence Prosecution evidence Mr Andrew Judge Defendant's evidence Prosecutor's submissions on the Metering Charges Defendant's submissions on the Metering Charges Evidence of fair wear and tear Operate properly and proper operation Evidence does not prove meters were not operating properly or a failure to ensure proper operation The error rates vary as between the meters No evidence of any failure to maintain for s 91H of the WM Act charge Honest and reasonable mistake Defences - s 91M of the WM Act Findings on the Metering Charges Conclusion and orders

[2]

Nature of proceedings

Lidokew Pty Ltd (the Defendant) has entered a plea of not guilty to a number of charges brought against it by Grant Barnes, Chief Regulatory Officer, Natural Resources Access Regulator (the Prosecutor). These proceedings relate to whether the Prosecutor has established that the Defendant committed one or more of the offences charged.
By Amended Summonses filed on 24 and 25 June 2021, the Prosecutor alleges that at the property comprising Lot 1 of Deposited Plan 1128755, located at 1563 Doreen Lane, Wee Waa in the state of New South Wales, known as 'Havana North' (Havana North or the Property), the Defendant committed the following six offences in contravention of the Water Management Act 2000 (NSW) (WM Act): 1. Charges 1-3 (proceedings 2020/88368, 2020/88369 and 2020/88370), relating to three Davies Shephard meters fitted to three separate pumps, in that: 1. In contravention of s 91I(2) of the WM Act, between about 29 January 2019 and 29 April 2019 inclusive, the Defendant took water from a water source to which Pt 3 of Ch 3 of the WM Act applied by means of a metered work while its metering equipment was not operating properly or was not operating; or 2. In the alternative, in contravention of s 91H(2) of the WM Act, between 29 January 2019 and 4 June 2019 inclusive, the Defendant failed to ensure the proper operation of the metering equipment that had been installed in connection with a water management work; (the Metering Charges). 1. Charges 4-6 (proceedings 2020/187127, 2021/181935 and 2021/181936) in that between about 1 July 2016 and 30 June 2017 (Charge 4), 1 July 2017 and 30 June 2018 (Charge 5), and 1 July 2018 and 30 June 2019 (Charge 6) respectively: 1. In contravention of s 60C(1)(b) of the WM Act, the Defendant took water from a water source to which Pt 3 of Ch 3 of the WM Act applied otherwise than in accordance with the water allocation for the access licence by which the taking of water from that water source was authorised, and knew or had reasonable cause to believe that the taking of the water was not in accordance with the water allocation; or 2. In the alternative, in contravention of s 60C(2) of the WM Act, the Defendant took water from a water source to which Pt 2 of Ch 3 of the WM Act applied otherwise than in accordance with the water allocation for the access licence by which the taking of water from that water source was authorised; (the Water Take Charges).
The alleged offence under s 60C(1)(b), being the primary alleged offence in the Water Take Charges, is a Tier 1 offence pursuant to the WM Act.
The Metering Charges, being the alleged offences under s 91I(2) and (in the alternative) under s 91H(2), as well as the alternative Water Take Charge pursuant to s 60C(2), are Tier 2 offences pursuant to the WM Act.

[3]

Facts

On the first day of the hearing, the Prosecutor provided to the Court a document titled "Key Factual Matters" (Exhibit C) prepared by the Prosecutor and taken to be admissions by the Defendant (Tcpt, 4 October 2022, p 5(27-29)) as extracted below: The Defendant 1. Lidokew Pty Ltd ACN 002 942 289 (the defendant) is an Australian company registered on 4 June 1985. Mr Gary Phelps (DOB: 20/01/1953) is the defendant's sole director and secretary. The defendant is a proprietary company limited by shares, with a share capital comprising 77 ordinary shares. Havana North Pty Ltd is the majority shareholder, holding 75 ordinary shares in the defendant. The remaining 2 ordinary shares are held by Mr Phelps. The Property 2. The alleged offences the subject of these proceedings occurred at a property (Property) located at 1563 Doreen Lane, Wee Waa NSW (Lot 1 DP 1128755) known as "Havana North". At all relevant times, the defendant was the registered owner of Havana North. Havana North is between 1,212 and 1,225 hectares in size. The following image is an aerial map showing the boundaries of the Property and the location of Dams 1, 2, 3 and 4 and bore sites at Havana North: 3. Havana North is located near Wee Waa in Northern NSW, approximately 41 kilometres north-west of Narrabri. The relevant Local Government Area is Narrabri Shire. The defendant uses the property primarily for cotton production. 4. At all relevant times, the defendant was entitled to take bore water from the Lower Namoi Groundwater Source by using three metered works installed at Havana North. At all relevant times, the use of that groundwater source was regulated by the Water Sharing Plan for the Upper and Lower Namoi Groundwater Sources 2003 (WSP). Application of Part 3 of Chapter 3 of the Act 5. A proclamation made by the Governor under ss 55A and 88A of the Water Management Act 2000 (NSW) (WM Act) and published in the Government Gazette No 127, 27 October 2006, page 8907 (Proclamation), declared that from 27 October 2006, Part 2 and Part 3 of the WM Act applied to, amongst others, water sources to which the WSP applied. 6. By reason of the Proclamation, at all relevant times, Part 2 and Part 3 of Chapter 3 of the WM Act applied to the Lower Namoi Groundwater Source. Water access licence and approvals 7. At all relevant times, the defendant held the following water access licence and approval granted under the WM Act: A. water access licence 12400, reference number 90AL806354 (WAL 12400), for 494 units of water from the Lower Namoi Groundwater Source; and B. combined water supply works and water use approval 90CA806356 for the construction and use of three Bores (works 1-3) to obtain water from the Lower Namoi Groundwater Source for the purpose of irrigation (Combined Approval 90CA806356). WAL 12400 nominates Combined Approval 90CA806356 as the only authorised extraction points. 8. The location of the three bores is set out in the image above at paragraph 2. Water Allocation 9. During the period 1 July 2016 to 30 June 2017, the defendant's water allocation for the purposes of s 60C the WM Act was 988ML (being the "Annual Use Limit" for WAL 12400). 10. During the period 1 July 2017 to 30 June 2018, the defendant's water allocation for the purposes of s 60C the WM Act was 988ML (being the "Annual Use Limit" for WAL 12400). 11. During the period 1 July 2018 to 30 June 2019, the defendant's water allocation for the purposes of s 60C the WM Act was 988ML (being the "Annual Use Limit" for WAL 12400). 12. In each of the three relevant water years, ie 1 July 2016 to 30 June 2017, 1 July 2017 to 30 June 2018 and 1 July 2018 to 30 June 2019, respectively, the water allocation was derived from an accrual of 494ML (494 units with an Available Water Determination of 1.0ML/unit) and a carryover from the previous water year of at least 494ML. 13. In each of the three relevant water years, ie 1 July 2016 to 30 June 2017, 1 July 2017 to 30 June 2018 and 1 July 2018 to 30 June 2019, respectively, the water allocation able to be extracted was capped at 988ML by virtue of cl 34(9) of the Water Sharing Plan for the Upper and Lower Namoi Groundwater Sources 2003 which provides for a maximum usage of 2ML per unit of aquifer licence share component in any given water year (plus or minus any water trades, of which there were none). The metered works 14. There are three metered works at Havana North relevant to these proceedings, which correspond with the location of groundwater bore sites: A. the Diesel Pump (extraction site 17637) located at GPS co-coordinates S 30.05973 ̊ and E 149.30707 ̊, B. the Electric Pump (extraction site 105552) located at GPS co-ordinates S 30.05702 ̊ and E 149.31534 ̊, and C. the House Pump (extraction site 17636) located at GPS co-ordinates S 30.04956 ̊ and E 149.30904 ̊. 15. Each of those three pumps is an approved water supply work under combined approval 90CA806356, of which the defendant is the approved holder. 16. At all relevant times, on the pump at each of the three bore sites, there was an installed Davies Shephard Water Meter, the purpose of which was to measure the volume of water taken.

[5]

Essential elements of the offence - Water Take Charges

Section 60C of the WM Act creates two offences: first, that provided for in s 60C(1) which includes an element of mens rea; second, that provided for in s 60C(2) which creates a strict liability offence.
As to the s 60C(1) charge the essential elements of the offence are: 1. That a person takes water from a water source; 2. The relevant water source is one to which Ch 3 of the WM Act applies to the water source; 3. An access licence authorises the taking of water from the water source; 4. Water is taken otherwise than in accordance with the water allocation authorised by the access licence; and 5. The person knew or had reasonable cause to believe that the taking of the water was not in accordance with the water allocation.
The s 60C(2) charges has each of the essential elements as set out in (1)-(4) above but does not have as an essential element that referred to in (5).
In this case, the Defendant has admitted the elements set out in (1)-(3) above, the question therefore is whether, in respect of the s 60C(1) charge the Prosecutor has relevantly established the elements in (4) and (5), and if not whether, for the purposes of the alternate charge in s 60C(2) it has established the essential element in (4).
As to the Water Take Charges the Prosecutor bears the onus of establishing beyond reasonable doubt each of the essential elements of that charge.
The Prosecutor identified by way of brief outline of its case at pars 8-10 of its written closing submissions the manner in which it put its case as follows: 8. The prosecutor's case is based on the evidence of a number of experts who, between them, estimate the volume of water that would have been required to meet irrigation needs within the relevant water years (Dr Meyer) and assess the potential sources of water capable of meeting those requirements from time to time (Professor Van Dijk and Robert Day). Essentially, the evidence of those experts seeks to model the system of water use on the Property - the inputs and outputs - for the relevant periods of the charges. 9. The effect of the expert evidence considered collectively is that the irrigation requirements for the crops grown on the Property (i) could not have been met through rainfall, stored water and bore water use as recorded on the meters and (ii) would have required more bore water than the annual use limits that applied to WAL 12400 provided. In each of the relevant water years, substantially more water would have been required to produce the crops than that provided for by rainfall and bore water usage (as metered) and that additional water must have come from the bores (there being no other available source of water on the Property). 10. The recorded meter use figures are so aberrant from the volume of water that must have been taken from the bores that the only reasonable inference is that the defendant knew that the meters were substantially under-reporting the take of water. Even on the defendant's own estimate of "5 or something" megalitres per hectare "on average" (factoring in efficiency losses), the crop sizes it was sowing would have required more than its annual use limit in each of the water years. At the least, the defendant had reasonable cause to believe that it was taking water otherwise than in accordance with its water allocation.
The Prosecutor also accepted that its case relied upon circumstantial evidence to establish the amount of water actually taken and for establishing the Defendant's state of mind.
Where a case relies upon circumstantial evidence a number of principles are relevant. In Shepherd v The Queen (1990) 170 CLR 573 at 579-580 Dawson J (with whom Toohey and Gaudron JJ agreed and with whom Mason CJ expressed general agreement) identified the approach to circumstantial evidence as follows: Circumstantial evidence is evidence of a basic fact or facts from which the jury is asked to infer a further fact or facts. It is traditionally contrasted with direct or testimonial evidence, which is the evidence of a person who witnessed the event sought to be proved. The inference which the jury may actually be asked to make in a case turning upon circumstantial evidence may simply be that of the guilt of the accused. However, in most, if not all, cases, that ultimate inference must be drawn from some intermediate factual conclusion, whether identified expressly or not. Proof of an intermediate fact will depend upon the evidence, usually a body of individual items of evidence, and it may itself be a matter of inference. More than one intermediate fact may be identifiable; indeed the number will depend to some extent upon how minutely the elements of the crime in question are dissected, bearing in mind that the ultimate burden which lies upon the prosecution is the proof of those elements. For example, with most crimes it is a necessary fact that the accused was present when the crime was committed. But it may be possible for a jury to conclude that the accused was guilty as a matter of inference beyond reasonable doubt from evidence of opportunity, capacity and motive without expressly identifying the intermediate fact that the accused was present when the crime was committed. On the other hand, it may sometimes be necessary or desirable to identify those intermediate facts which constitute indispensable links in a chain of reasoning towards an inference of guilt. Not every possible intermediate conclusion of fact will be of that character. If it is appropriate to identify an intermediate fact as indispensable it may well be appropriate to tell the jury that that fact must be found beyond reasonable doubt before the ultimate inference can be drawn. But where - to use the metaphor referred to by Wigmore on Evidence, vol. 9 (Chadbourn rev. 1981), par. 2497, pp. 412-414 - the evidence consists of strands in a cable rather than links in a chain, it will not be appropriate to give such a warning. It should not be given in any event where it would be unnecessary or confusing to do so. It will generally be sufficient to tell the jury that the guilt of the accused must be established beyond reasonable doubt and, where it is helpful to do so, to tell them that they must entertain such a doubt where any other inference consistent with innocence is reasonably open on the evidence. As I have said, the prosecution bears the burden of proving all the elements of the crime beyond reasonable doubt. That means that the essential ingredients of each element must be so proved. It does not mean that every fact - every piece of evidence - relied upon to prove an element by inference must itself be proved beyond reasonable doubt. Intent, for example, is, save for statutory exceptions, an element of every crime. It is something which, apart from admissions, must be proved by inference. But the jury may quite properly draw the necessary inference having regard to the whole of the evidence, whether or not each individual piece of evidence relied upon is proved beyond reasonable doubt, provided they reach their conclusion upon the criminal standard of proof. Indeed, the probative force of a mass of evidence may be cumulative, making it pointless to consider the degree of probability of each item of evidence separately.
In considering a circumstantial case the whole of the circumstances established by the evidence are to be considered and weighed as a whole and not by a piecemeal approach to each particular circumstance. As was stated in Chamberlain v The Queen (No 2) (1983) 153 CLR 521 (Chamberlain) at 536 (per Gibbs CJ and Mason J): It follows from what we have said that the jury should decide whether they accept the evidence of a particular fact, not by considering the evidence directly relating to that fact in isolation, but in the light of the whole evidence, and that they can draw an inference of guilt from a combination of facts, none of which viewed alone would support that inference. Nevertheless the jury cannot view a fact as a basis for an inference of guilt unless at the end of the day they are satisfied of the existence of that fact beyond reasonable doubt. When the evidence is circumstantial, the jury, whether in a civil or in a criminal case, are required to draw an inference from the circumstances of the case; in a civil case the circumstances must raise a more probable inference in favour of what is alleged, and in a criminal case the circumstances must exclude any reasonable hypothesis consistent with innocence (see Luxton v. Vines (49); and Barca v. The Queen (50)). The statement by Lord Wright in Caswell v. Powell Duffryn Associated Collieries, Ld. (51), that "There can be no inference unless there are objective facts from which to infer the other facts which it is sought to establish" is obviously as true of criminal as of civil cases. The process of reasoning in a case of circumstantial evidence gives rise to two chances of error: "first from the chances of error in each fact or consideration forming the steps and second from the chance of error in reasoning to the conclusion": Morrison v. Jenkins (52). It seems to us an inescapable consequence that in a criminal case the circumstances from which the inference should be drawn must be established beyond reasonable doubt. We agree with the statement in Reg. v. Van Beelen (53), that it is "an obvious proposition in logic, that you cannot be satisfied beyond reasonable doubt of the truth of an inference drawn from facts about the existence of which you are in doubt".
It is to be accepted that the Prosecution must exclude all reasonable hypotheses consistent with innocence. As was stated in R v Baden-Clay (2016) 258 CLR 306 at [46]-[47] (per French CJ, Kiefel, Bell, Keane and Gordon JJ): 46 The prosecution case against the respondent was circumstantial. The principles concerning cases that turn upon circumstantial evidence are well settled (33). In Barca v The Queen (34), Gibbs, Stephen and Mason JJ said: "When the case against an accused person rests substantially upon circumstantial evidence the jury cannot return a verdict of guilty unless the circumstances are 'such as to be inconsistent with any reasonable hypothesis other than the guilt of the accused': Peacock v The King (35). To enable a jury to be satisfied beyond reasonable doubt of the guilt of the accused it is necessary not only that his guilt should be a rational inference but that it should be 'the only rational inference that the circumstances would enable them to draw': Plomp v The Queen (36); see also Thomas v The Queen (37)." 47 For an inference to be reasonable, it "must rest upon something more than mere conjecture. The bare possibility of innocence should not prevent a jury from finding the prisoner guilty, if the inference of guilt is the only inference open to reasonable men upon a consideration of all the facts in evidence" (38) (emphasis added). Further, "in considering a circumstantial case, all of the circumstances established by the evidence are to be considered and weighed in deciding whether there is an inference consistent with innocence reasonably open on the evidence" (39) (emphasis added). The evidence is not to be looked at in a piecemeal fashion, at trial or on appeal (40).
The Prosecutor's case relied upon establishing the amount of water actually taken in each relevant water year from inferences to be drawn from expert evidence relating to: 1. The water demand of the cotton crop in each water year as identified by Dr Wayne Meyer; 2. The amount of water available in each water year from rainfall and runoff as determined by Prof Albert Van Dijk; and 3. The amount in each water year of any available stored water as determined by Mr Robert Day.
This evidence was then to be measured against the water allocation authorised by the licence for each water year, and would, on the Prosecution case establish the inference that the crop could not have been grown without taking more water than that allocated.
The Prosecutor also accepted that whilst all three circumstances were relevant in order to establish that it had proven its case to the relevant standard, the evidence of Dr Meyer as to crop water demand was an indispensable inference to be drawn. If the crop water demand amount was not accepted, the Prosecution's circumstantial case would fail - as it could not establish a water demand to which the rainfall and stored water could be applied: Tcpt, 14 October 2022, p 553(40)-554(48).
Accordingly, it is necessary that I determine whether the Prosecutor has proven beyond reasonable doubt its circumstantial case having regard to the evidence including the three relevant areas of expert evidence that it relies upon to draw the relevant inferences.

[9]

Dr Wayne Meyer

Dr Wayne Meyer described his primary field of expertise as being "in irrigation water and soil management, crop physiology and soil science". He holds a Bachelor of Agricultural Science (majoring in agronomy) and a PhD in water uptake of plant root systems.
Dr Meyer engaged in the task of estimating the amounts of water required to produce the cotton crops grown on Havana North in each of the water years 2016/2017, 2017/2018 and 2018/2019. He then calculated how much of this water would be required to be acquired from sources other than direct rainfall on those crops.
Dr Meyer concluded that for each nominated growing year the following: 1. 2016/2017 crop was grown on 320ha with a nominal start date of 7 December 2016 and finish date of 17 April 2017. The crop was estimated to require 9.1ML/ha or at least 2,912ML of irrigation water to be available; 2. 2017/2018 crop was grown on 324ha with a nominal start date of 5 December 2017 and finish date of 14 April 2018. The crop was estimated to require 10.6ML/ha or at least 3,693ML of irrigation water to be available; 3. 2018/2019 crop was grown over two different periods and was split by an east and a west field: 1. The east field was 179ha with a nominal start date of 28 November 2019 and a finish date of 28 March 2020. The crop was estimated to require 11.2ML/ha or at least 2,005ML of irrigation water; 2. The west field was 157ha with a nominal start date of 10 December 2018 and a finish date of 20 April 2019. The crop was estimated to require 10.2ML/ha or at least 1,601ML of irrigation water.
In order to undertake his estimate of water required Dr Meyer used a methodology explained in Exhibit 5: John Hornbuckle et al, "IrriSAT Technical Reference" (2016), https://irrisat-cloud.appspot.com/doc/IrriSAT_Technical_Reference.pdf, accessed 27 November 2023 (IrriSAT Technical Reference). Crop water requirements (ETc) can be estimated considering a climatic parameter called reference evapotranspiration (ETo), which represents the evapotranspiration from a standardized vegetated surface, and a crop factor called crop coefficient (Kc) that relates ETc to ETo by the equation (Allen et al., 1998): ETc = ETo Kc Where: ETc crop evapotranspiration [mm] ETo reference evapotranspiration [mm] Kc crop coefficient [-] ETo is the rate that an extensive surface of green, well-watered grass of uniform height actively growing and completely shading the ground, evaporates water. As soil factors do not affect this reference surface, the only factors affecting ETo are climatic parameters (solar radiant energy, air temperature, humidity and wind speed). These climate variables can be obtained from weather data.
Dr Meyer sourced daily weather data from the SILO site, using the dataset for the Wee Waa Agricultural Research Station (CSIRO Myall Vale, station 53022) (Myall Vale Weather Station). The Myall Vale Weather Station was approximately 33.2km from Havana North. The data sourced by Dr Meyer included daily maximum and minimum temperatures, rainfall and the calculated value of ETo that used observed values of humidity (vapour pressure) and incoming solar radiant energy (sunlight). That was the basis of the ETo figure used by Dr Meyer.
Dr Meyer determined the crop coefficient (or Kc) value using the IrriSAT weather-based irrigation scheduling tool. As Dr Meyer explained: The IrriSAT methodology uses satellite images to determine the Normalized Difference Vegetation Index (NDVI) for each field, from which the plant canopy size can be determined and a specific crop coefficient (Kc) can be estimated. … When a particular area of interest such as an irrigated field is identified it can be marked with a red outline and the value of the specific crop coefficient (Kc) can be derived for each pixel in the marked area using the relationship between NDVI and Kc as described by Hornbuckle et al. (2016). Kc has been shown to be closely related to the canopy ground cover fraction (i.e. light interception) which can be estimated from remote sensing measurements of the Normalized Difference Vegetation Index (NDVI).
Dr Meyer described the next step in the following terms: With values of reference evapotranspiration ETo determined from the daily weather data and then localized crop evapotranspiration (ETc) determined from ETo and the satellite images it is then possible to estimate the crop water use. It is assumed that applied irrigation was managed so that the additional water from rainfall was accounted for. It is important that the duration of water use by the cotton crop is identified i.e. the date when significant crop water use starts and the date when significant crop water use finishes.
Dr Meyer determined his estimates for the start and end dates of the crops by reference to normalized difference vegetation index (NDVI) distribution observable from satellite imagery.
Once the length of the season was determined, Dr Meyer generated daily values of ETc over the course of the cropping season to arrive at a figure of total crop water use. The volume of water that needs to be applied as irrigation water is the difference between the rainfall over the course of the cropping season and the total crop water use figure.
As Dr Meyer explained, a further determination must then be made so as to arrive at the volume of irrigation water required. The level of irrigation efficiency on the farm must be determined to assess the volume of source water required to service the crop water requirements. Dr Meyer adopted a Whole Farm Irrigation Efficiency figure (WFIE) of 70% (which accounts for water losses due to evaporation, seepage, deep percolation and the like). The consequence of the application of a WFIE is that the total amount of water required to apply the amount for crop water demand to the crop itself must be increased to take account of these losses - that is, on Dr Meyer's estimate, an additional water volume of 30% is required.
Dr Meyer applied this method to each of the relevant water years. Dr Meyer expected that his estimates of crop water use had an uncertainty of ±15%.
By way of example, in the 2016/2017 water year Dr Meyer nominated a crop starting date of 7 December 2016 and a crop finishing date of 17 April 2017. This period of 132 days did not reflect the time from planting to harvest as it omits the period over which seedlings are emerging and evaporating only relatively small amounts of water. The total irrigation requirements are estimated by Dr Meyer over the course of the 132-day period to have been 2,912ML, with an error rate of ±15%. The lower bound of that figure (that is, 15%) is 2,475.2ML. Dr Meyer's figure was based on total rainfall of 185mm over the relevant period of 132 days. Dr Meyer undertook a similar exercise in each of the other water years the subject of the charges.
As Dr Meyer's calculations included an amount for rainfall contribution the amount of the total irrigation water requirement of 2,475.2ML would need to be derived either from: one or more of the bores on the Property; via stored water; or via rainfall harvested into storages on the Property.
Dr Meyer observed that for the relevant water years the crop yield for irrigated cotton was less than the average yield reported in the Cotton Seed Distributors Ambassador Network Program (CSD). He considered that the comparability between the actual yield indicates that his estimate of irrigated water requirements is valid and consistent with that needed for a fully irrigated and reasonably well managed crop.
He also compared the yield for the dry land crops (that is, not irrigated) grown on Havana North. Again, the actual yield for each relevant water year was below the average reported by CSD. He considered that such a yield was explicable as the crop water demand would have exceeded the rainfall and indicates insufficient water available from rainfall and water stored in the soil profile to meet crop needs.
Dr Meyer was cross-examined on his evidence.
As to his use of rainfall data Dr Meyer gave evidence in cross-examination that: 1. He utilised the data from the Myall Vale Weather Station for rainfall data; 2. Looking at data from Narrabri Airport, Molly, Bellata (Aberfeldie), Pendennis and Wee Waa (George Street) Weather Stations shows a variation in rainfall in the general locality; 3. Whilst the volume of water varies from month to month on a statistical basis the variability is not significant, however, the variability may be "made up" in periods outside the growing period for a cotton crop and his evidence is the amount of rainfall during the growing season for cotton; 4. It can be inferred from the rainfall information available that across this landscape the rainfall in the area on any given day in any given month does not fall uniformly across the landscape; and 5. As there is no rainfall gauge at the Property it cannot be known what the rainfall was at the Property, but it can be said that it is unlikely to be significantly different from the known rainfall - but it was accepted that it could be higher or lower.
As to his identification of the nominal start and finish dates of the cotton crops Dr Meyer indicated in cross-examination that: 1. He did not establish these dates from information from the Defendant; 2. He utilised satellite imagery to determine when a green leaf is emerging from the crop as the start date - rather than the date of sowing the crop - and the finish date is the date where "brown off" of the green leaf is observed; 3. From a combination of the best the satellite imagery available he derives an NDVI established from identifying the different wavebands per pixel utilsing the IrriSAT computer programme; and 4. The satellite images are taken on a weekly basis so to identify a particular day a process of inference must be undertaken.
With respect to the water consumption calculation Dr Meyer gave evidence that: 1. The calculation in an ideal world would be calculated from daily data. However, in this case only having weekly data, a degree of assumption must be undertaken which is sufficient based upon his experience; 2. He had not been to the Property nor spoken to the Defendant; 3. Whilst a particular farm crop does not develop uniformly the IrriSTAT computer programme accounts for that by a statistical analysis; 4. The IrriSAT computer programme requires a selection of a variable for crop height to determine the ETo. He was not provided with any data on crop height but the IrriSAT computer programme does not require a determination of when a crop reaches a particular height but rather whether as a crop it is a high or low height crop; 5. Cloud cover on particular days will influence ETo due to net radiation at the crop surface. He had no details for the Property and utilised those derived from the Myall Vale Weather Station as, in his opinion, across the landscape the net radiation would be very similar on a daily basis; 6. He applied to the water usage figure an uncertainty rate of ±15% based upon his experience and it should not be lower or higher. There is no basis upon which this figure is derived apart from experience. The effect of that uncertainty rate will vary depending upon the rainfall data selected. 7. He applied a WFIE of 70%. This figure was derived from a paper by Guy Roth et al, "Water-use efficiency and productivity trends in Australian irrigated cotton: a review" (2013) 64 Crop & Pasture Science 1033 (the Roth Paper) which measured a range from a high efficiency rate of 80% to a low efficiency rate of 43% with an average of 70%. The Roth Paper showed some efficiencies exceeded 90%. However, such figure related to individual efficiencies whereas he considered WFIE; and 8. Exhibit 8, a Department of Industries fact document entitled "Benchmarking water productivity of Australian irrigated cotton" (2019): https://www.dpi.nsw.gov.au/__data/assets/pdf_file/0006/1185288/Benchmarking-Water-Productivity-of-Australian-Cotton.pdf, accessed 27 November 2023 (2019 Fact Sheet) indicates that there is increasing WFIEs year on year and he accepted this as a true statement. The Roth Paper considered farm efficiency of 70% in 2006 to 2008 and of around 83% and 81% in 2012/2013 and 2017/2018 respectively. He cannot categorically rule out a WFIE in this case of 80% because he has not been asked to undertake the empirical exercise on the ground and has no firsthand experience of the efficiencies of the Property.
Dr Meyer identified for the 2016/2017 water year a requirement of 9.1ML/ha was the whole of farm source water requirement. He considered Exhibit 7, a publication entitled "Water Use Efficiency in the Cotton Industry" (2016): https://web.archive.org/web/20170215112203/http://cottonaustralia.com.au/cotton-library/fact-sheets/cotton-fact-file-water, accessed 27 November 2023 (2016 Fact Sheet) prepared by Cotton Australia with which he was familiar. He accepted that there was a trend of improving water efficiency among cotton farmers of about 4% per annum. That document identified an average irrigation requirement of 7.8ML/ha being different to his calculation because it related to irrigation requirement rather than the whole of farm source water requirement. He also accepted that the Roth Paper identified a 6-7ML/ha requirement but did not accept that such figures were comparable to his figure as his figure was whole of farm source water requirements, which would be higher than the irrigation requirement. He was unable to identify exactly how much higher in this case.
He accepted that the ginning records of bales of cotton provided to the gin as recorded in the evidence was unable to permit a calculation of irrigation water either used or required.
As to the cotton yield for each of the water years Dr Meyer has stated in his report: 124. The calculations made using the IrriSat software takes account of the actual condition of the crop that is recorded by the satellite instruments. This means that if for some reason, the crop leaf canopy does not grow well this will be reflected in the lower crop coefficient (Kc) value that are used to calculate crop evapotranspiration and hence the irrigation water required. The estimated amount of water required to produce the cotton yields on Havana North for 2016-17 was 2912ML, for 2017-18 it was 3693ML and for 2018-19 was 3606ML. 125. If the crops, especially those in the latter two seasons had shown evidence of being better managed then the water required would have been greater. The poor yield (7.4 bales/ha) of the 2018-19 irrigated crop is likely to have resulted from a relatively poor crop establishment, probably insufficient fertilizer and perhaps insufficient irrigation water. It is also likely that the continuous cropping of cotton on the same fields during all three seasons resulted in poor growing conditions from plant nutrient insufficiency, disease buildup and or weed competition.
He was cross-examined on this evidence and stated at Tcpt, 11 October 2022, p 333(5)-334(6): Q. At paragraph 125 of your expert report, if I could ask you to turn to that paragraph, at court book 1329, one of the things you suggest for the 2018/19 crop year, if I could refer to it-- A. Yes. Q. --as that, having regard to the yield per hectare that's identified in the ginning records, based on the known planning area of 7.4 bales per hectare-- A. Yes. Q. --you ascribe that to a number of matters. You say relatively poor crop establishment, probably insufficient fertiliser, and perhaps insufficient irrigation water. Do you see that? A. Yes. Q. And, of course, not having measured any of those variables, you're not expressing a view there as to whether any of those things were in fact the case, you're just, I suggest, identifying some common causes for a reduction in bales per hectare. That's correct? A. No, that's not correct. I - I have - the evidence associated with those statements is - comes from, firstly, looking at the satellite imagery. The evidence of relatively poor crop establishment comes from the - you can see that from the satellite imagery. You can see during the growing season that rather than the crop being a very robust green, because it's got adequate fertiliser, that there is period during that season when there was very poor greening and yellowing on the crop. Very likely from insufficient fertiliser And thirdly, the evidence that you've had three cotton crops grown in a row, without a break crop is unusual in terms of good cotton crop management. And usually, because of the need for break crops and so on, that's unusual practice, and was generally known to lead to poor yield from subsequent crops. Q. But of course, not having been to Havana North, you've got no knowledge of the amount of fertiliser used as in terms of the amount bought and applied to these fields. Do you? A. I don't have direct evidence of that information, no. Q. Similarly, in relation to the manner of sowing and crop establishment and the techniques used by the farmer over that growing season, you've got no actual objective information about that. Do you? A. I have no objective information about the methods of establishment or so on. I can see the result of the establishment and the growth of the crop coming from the satellite imagery. Q. And indeed, one of the things you identify a paragraph 125, is that the amount of bails per hectare, 7.4. Relatively low the locality or region for that year, that 2018, 2019 year. That could be something that results from insufficient irrigation water. And that's something that you've expressly said at paragraph 125 of your report. A. That's correct. It could be from that. You don't know exactly what the cause is, but from the evidence of the crop, clearly the crop was not thriving and there were a number of causes that could be - could have caused that.

[11]

Defendant's evidence

The Defendant adduced expert evidence from Mr James Purcell, Chartered Professional Civil Engineer with experience in water resources and irrigation engineering. He gave evidence that his specialty was the investigation, design and construction project management of irrigation infrastructure and in the last 22 years, the measurement and management of water use efficiency on irrigation farms. He has also developed and used software programs relating to the measurement of water balance on irrigation farms.
Mr Purcell gave evidence that he had undertaken an inspection of the Property on 7 September 2022.
Mr Purcell had considered the evidence of Dr Meyer. As to the evidence of Dr Meyer, Mr Purcell disputed the water needs for each cotton crop as identified by Dr Meyer.
Mr Purcell identified the seasonal crop water requirements as dependent upon climate and the length of the season. To calculate that requirement Mr Purcell noted: 5.0 Seasonal Crop Water Requirements at Havana North. 5.1 The actual water volume consumed by the crops at Havana North through evapotranspiration will vary slightly each season depending on the climate and the length of the season. 5.1 Crop Water Requirements 5.1.1 Calculation of the crop water requirement for any season and crop requires basic crop and climate data. 5.1.2 To calculate the crop water requirements for cotton at Havana North, first it is necessary to determine whether the crop was fully irrigated and disease free. For those conditions it is possible to calculate the crop water requirements based on plant emergence date, daily reference evapotranspiration at Havana North, crop factors and defoliation or crop end date. 5.1.3 Daily reference evapotranspiration is readily available from the SILO Australia climate data www.longpaddock.qld.gov.au 5.1.4 Provided the crop is fully irrigated and disease free the most reliable method of calculating the crop water requirement, in my opinion, is to multiply the measured/calculated reference evapotranspiration by the crop factors as described in Food and Agriculture Organization of the United Nations - Irrigation and Drainage Paper 56 titled "Crop Evapotranspiration - Guidelines for Computing Crop Water Requirements". 5.1.5 This publication provides the crop factors which basically converts the reference evapotranspiration which is for a grass reference crop to say cotton at Havana North. 5.1.6 If it appears that the crop has been water stressed or not reaching full potential, then it is possible to determine the "crop factors" using the IrriSAT satellite imagery technique. 5.1.7 This technology uses satellite images of the field in question to determine the Normalised Difference Vegetation Index (NDVI) which in turn can produce a crop coefficient for that time. By multiplying the reference evapotranspiration by the IrriSAT calculated coefficient, the daily crop water use can be determined.
Mr Purcell considered that Dr Meyer's use of a WFIE of 70% could not be reliably accepted as: 8.1 The seasonal calculated crop water requirements is the volume of water that the crop used in the season. The volume of water that the irrigation farm needed to have available to be able to provide the crop water requirements is going to be larger than the crop water requirements because of the water losses in the farm infrastructure while storing and delivering the water to the crop. 8.2 The water losses at Havana North would be from: • Seepage and evaporation in the dams; • Seepage and evaporation in the supply channels; • Seepage and evaporation in the drains from the fields back to the dams; • Evaporation from the surface of the water on the fields during irrigation; • Over irrigation or deep percolation of water below the crop root zone; and • Any accidental loss of water due to infrastructure failure or overtopping of channel or drains. 8.3 The Whole Farm Irrigation Efficiency (WFIE) shows the amount of irrigation water that was used by the plant as a percentage of total water inputs to the farm (Roth et al. 2013). If there were no losses on the irrigation farm, then the WFIE would be 100%. 8.4 The selection of a WFIE of 70% by Wayne Meyer in Paragraph 48 of his Expert Report seems to be based on the stated improvement in average WFIE from 57% to 70% in the paper by Roth et al. 2013. 8.5 WFIE is usually calculated based on measured or calculated water losses on farm and calculated crop water use. 8.6 As none of Havana North's water losses have been calculated or measured nor have the total available water inputs been measured or calculated, the selected value of 70% for WFIE may or may not be reasonable. 8.7 In my opinion, it is very difficult to justify the selection of an average value for WFIE for Havana North particularly without completing a detailed site inspection.
Mr Purcell was cross-examined on this evidence. Mr Purcell did not resile from his opinion that the 70% WFIE adopted by Dr Meyer was an unreliable assumption. He stated that he had developed a water track programme (in conjunction with others) that was designed to do a whole of farm water balance. The determination of such a balance required an examination of farm infrastructure as well as farm management practices. Absent such an analysis adopting a WFIE was "making a judgment call on some unknown basis": Tcpt, 13 October 2022, p 500(40)-(41).
He stated that he had used the IrriSAT computer programme in his practice. It was not his first choice of a programme to use for determining water usage requirements for a crop, his first choice being the use of FAO56 (being the document referred to at [5.1.4] of his report and recited at [54] above). The IrriSAT computer programme was his second choice. He considered it to be an inferior tool because it is an indirect measure of Kc utilising infrared from satellite images.

[12]

Prosecutor's submissions

Dr Meyer's determination of crop water demand and irrigation water requirements should be accepted. He utilised an appropriate methodology for determining crop water demand utilising acceptable available objective evidence.
The criticism of Dr Meyer's approach would not be accepted as: 1. The criticism of his adopted error rate was not established. As was stated by Dr Meyer in rejecting the Defendant's suggestion that the error rate should be higher, in the order of 20% (Tcpt, 11 October 2022, p 320(17-21)): No. The - the figure I gave - I gave you that figure, the 15% is on my experience through, again, 20 years of measuring ETc and ETo, and looking at how well we do in terms of using that relationship between ETo and ETc using the Kc value. And as it indicates that variability is plus or minus 15%. So, in general terms, it could be slightly greater or slightly less of that sort of order. There does not appear to have been any basis for the suggested error rate of 20% and it was roundly rejected by Dr Meyer having regard to his substantial experience in the measurement of ETc and ETo. 1. The criticism of the adoption of a 70% WFIE should not be accepted. Whilst Dr Meyer agreed that he could not "categorically rule out" that the WFIE was as high as 80% he did not however resile from his opinion that 70% was the appropriate figure (Tcpt, 11 October 2022, p 324(11-15)): No, I don't agree with that because I - from my - looking at the data which is - particularly the data which is contained in some of those previous tables on the Roth paper, the 80% is an exception and the - the majority of the - of the whole farm irrigation efficiencies tend to be in the lower end, and the 70% that I chose as the average value is an - is an upper value, a conservative value for a reasonably well run - a pretty well run irrigation system and a whole farm basis. The Court would find that Dr Meyer was well justified in adopting a figure of 70% having regard to: 1. His evidence that individual application losses can range from 80%-90%; 2. His evidence that WFIE takes into account more than simply individual application losses; 3. The evidence of Mr Purcell that WFIE takes into account (in addition to features of individual application loss) a range of matters from which there can be water losses; 4. The Roth Paper that observes that the largest loss is from evaporation from on farm water sources, which position was supported by Mr Phelps in his ROI where he observed that the evaporation rate he was experiencing in late 2019 was "scary". Such evaporation rates are also indicated by the SILO data for the relevant years; and 5. Poor farm management practices including three successive years of cotton cropping; lack of use of soil water probes to maximise water efficiency and the storage of water in water storages where evaporation can occur.
The criticism of Dr Meyer with respect to his adoption of the assumption that the water stored in the soil profile at the start of the cropping season would not be accepted. His explanation as to why he made such assumption should be accepted as he explained at Tcpt, 11 October 2022, p 332(45)-333(3) that: And the soil profile will only hold a certain amount of water because it will drain through. Not all of that water will be available for the subsequent crop… if it was to refill the soil profile, then that water would be available, and as I indicated in my report, make the assumption that the amount of water at the start and the end of the season is the same, in the absence of any other information, that's a reasonable assumption. The Court would conclude that this is a reasonable assumption to make given that the soil would be exposed to irrigation over the course of the cropping season. Indeed, there is no evidence to contradict the reasonableness of that assumption. Further, the poor performance of the dryland crops in 2016/2017 and 2018/2019 is indicative of either a lack of moisture in the soil profile when the crops were planted or poor seasonal rainfall.
Ultimately, the Court would accept and act on Dr Meyer's estimates as to the volume of water applied to the crops in each of the relevant water years. In the context of the criminal standard, the Court may wish to proceed on the conservative approach of adopting the lower bound of Dr Meyer's estimate (that is, -15%). That is of course not his actual estimate, but it is a figure he regards as plausible.

[13]

Defendant's submissions

The evidence of Dr Meyer as to crop water demand and irrigation water requirements should not be accepted as a proper foundation for determination of those factors. The exercise undertaken by Dr Meyer is based upon assumptions without on the ground investigation and is insufficient to found the necessary inferences that the Prosecution seeks to draw. In short, just about every variable that needs to be measured to work out what amount of water was needed from the bores is missing and unquantified in the evidence. The only actual measurement of any relevant water in this case is that taken from the bores.
By virtue of the analysis undertaken by Dr Meyer the determination of crop water demand and irrigation water demands are very sensitive to the assumptions applied in the analysis. A variation in any one of the assumptions may have a significant effect on the ultimate calculation. As these are assumptions made, rather than direct evidence, the making of the assumption must be appropriately justified. Dr Meyer's assumptions have not been justified and should not be accepted. In particular, the following assumptions made by Dr Meyer should not be accepted: 1. The use of a 70% WFIE - Mr Purcell gives evidence that the WFIE of the farm is uncertain without it being calculated from on the ground investigation, assessment and measurement. Dr Meyer accepted that it was not impossible for a WFIE figure of 90% to be reached and the Roth Paper at p 1042 has some examples of farms achieving this level of irrigation efficiency. Dr Meyer has not measured the WFIE at Havana North and has not undertaken any empirical exercise on the ground and could not categorically rule out a WFIE figure of 80% or above. 2. Rainfall - The use of the Myall Vale Weather Station is inappropriate in a locality where rainfall is variable. The evidence suggest that a single weather station as used by Dr Meyer is an inappropriate means to determine the rainfall to be used to determine the amount of irrigation water required to meet crop demand. The evidence supports a finding that the rainfall in the locality is highly variable such that the adoption of a single weather station as a proxy for measurements at Havana North is highly problematic. 3. The application of a nominal start date - Dr Meyer selected a nominal start date that cannot be exactly known just from the satellite imagery rather than the actual start date for the start of the crop year. The dangers of the estimate and computer algorithm based remote analysis approach. 4. The assumption that the soil moisture level would be the same at the beginning and end of the growing period. Dr Meyer had not measured the antecedent soil moisture at the start of the growth season. There is evidence of significant rainfall at local stations prior to his December start dates. The assumption of Dr Meyer that there was the same level of soil moisture at the start and end of the season, rather than, for example, a high soil water content start point which the crop was able to draw on during the early part of the season, is not supported by any evidence at all. For both the 2017 and 2018 start dates there was significant rainfall in the days before the nominal start of Dr Meyer's crop season: 3 December 2017 Wee Waa (George Street), 47.6mm; and 22 November 2018, 30.1mm. The nominal start date for the 2017/2018 crop was 5 December 2017, and for the eastern fields for 2018/2019 was 28 November 2018. The likely presence of significant soil moisture at the start of both these seasons from rainfall has not been considered by Dr Meyer and is a significant omission. 5. The error rate - Dr Meyer's determination of his error rate was not properly substantiated. In any event, the consequences of his error rate are such that the difference in the determination of the amount of water required for crop demand and irrigation water will vary considerably.
The unreliability of Dr Meyer's assessment of water demands was also evidenced by error or concessions made in his own evidence, including: 1. Dr Meyer agreed only after much questioning with the obvious proposition that the ginning records showing the bales per hectare of cotton do not provide a water usage figure, changing his earlier evidence initially denying this proposition; 2. Dr Meyer's acceptance that the low yield in say 2018/2019 of cotton from Havana North could be the result of insufficient irrigation water, a concession that undermines the entire Prosecution case which is based on seeking to prove excessive water use; 3. Dr Meyer was in error in recounting his own evidence. He stated that the 9.1ML figure was the crop water usage requirement not the irrigation water requirement. However, this is incorrect - the 9.1ML is derived from 822mm (crop water usage) minus 185mm (rain) and then applying a WFIE of 70% to give the total water needed ((637 / 70) x 100 = 910mm). This gives 9.1ML/ha, which is not the crop water usage requirement but Dr Meyer's calculated irrigation water post rainfall and after WFIE requirement; and 4. Dr Meyer (and the Prosecutor) is incorrect in submitting that the irrigation water figure in the Roth Paper is taken before WFIE is factored in. It certainly is not source water, or the crop water usage figure, but it is the amount after rainfall required from the irrigation source. The fact that Dr Meyer's calculation is significantly more than the amount of water that would usually be required for irrigated cotton crops is apparent from an assessment of the Roth Paper. Dr Meyer himself identified at par 122 of his report that his irrigation water requirement figures are above what would usually be expected. This alone would raise issues as to the reliability of the calculation of Dr Meyer and some justification for this significant departure from the norm.
Mr Purcell's evidence should be accepted where there is a departure from Dr Meyer. He is a highly experienced irrigation engineer who has worked around the world and extensively in Australia in this specialised field and is now based locally. He explained, actual measurement of the variables going into crop water usage must be known before an estimate as to the amount required to irrigate a particular crop on that farm in a particular season can be known. If this is not done, the result is not certain and entirely speculative. There is no criticism made of Dr Meyer's expertise, but he has been instructed by the Prosecutor to be the essential witness in a criminal case as to matters requiring certainty that his methodology cannot and does not purport to provide. Inexplicably, he was not even instructed to inspect Havana North or surrounds and was left to assume that the cotton crop was to be considered a "tall crop" from the moment green shoots appeared.

[15]

Crop growing period

There is some importance to the identification of the crop growing period as the time dictates the period of water demand, and therefore the average amount of water required by that crop. Of particular importance are the start and finish dates as these set the parameters for the determination of the water usage.
Dr Meyer determined the relevant start and end dates from an examination of satellite imagery. The end date was determined as the date where the crop was seen to have "browned off" in that it was either harvested or no longer having a water demand. The evidence of the finish date was both a visual representation together with the representation made to the gin (where the finished cotton product was delivered after harvest) of the date of harvest. There was little criticism by the Defendant (including Mr Purcell) of Dr Meyer adopting the crop finish date, based upon the availability of a discernible visual representation of the crop end date being ascertainable from the examination of the aerial photographs together with the representation made to the gin. For those reasons, I accept the end dates for each of the water years as determined by Dr Meyer.
The start date for the crop, however, has a need for precision. Dr Meyer accepted in evidence that the date a crop is planted as seed can affect water demand depending upon the existing (antecedent) soil moisture conditions. If a crop was planted at a point in time where rainfall had recently occurred, the crop would utilise the water stored in the soil profile rather than relying on the application of irrigation water. Therefore, the identification of the date the crop was planted, and an examination of soil water availability could affect the water demand calculation.
Dr Meyer accepted in cross-examination that the date of planting could have been ascertained from the Defendant, however, such information was not available to him, nor did he request it. Dr Meyer accepted that if he had that information from the Defendant, it would reduce the assumptions he has made as to the growing period: Tcpt, 11 October 2022, p 310(46)-311(13).
Dr Meyer determined crop start date from the point in which he could ascertain some development of a green colour of the crop from an examination of the satellite imagery. The satellite imagery he used was a weekly image rather than a daily image. He visually interrogated the images, and he determined the nominal start date to be the date on which the seeds sprouted a crop that had grown to a height that had broken through the surface level of the soil. The term "nominal" start date was the nomination provided by Dr Meyer, as he attested that he could not ascertain from the satellite images the sowing date but the date on which the green shoots become apparent is used to nominate a start date for the crop.
Dr Meyer accepted that the nomination of a start date could have a material effect on the amount of water required. The consequence of the start date was demonstrated in the example put to Dr Meyer at Tcpt, 11 October 2022, p 329(37)-330(14) where he stated: Q. --year, you identify for 2017/2018 a cotton crop start date on Havana North of 5 December 2017. That's correct. A. Yes. Q. Could I ask you to go to the court book at page 965. In the same volume of the court book, I just ask you to turn to page 965. You'll see some weather records. You'll see on page court book 965 the daily rainfall records for 2017 for the Wee Waa Pendennis station, the station that's 11 km-- A. Yes. Q. --to the south. I'm not suggesting this was in any way intentional, but you'd agree with me that one consequence of the 5 December 2017 start date for the growing season for the 2017/2018 year is that in the amount of rainfall that you factor in over that growing season, it so happens that you miss out the 46.8 ml that occurred on 3 December. That's just a consequence of selecting a particular nominal start date, isn't it? A. That start date was determined from the satellite images that I could see, not determined by trying to bias it one way or the other with any rainfall event. Q. I understand that and I'm not suggesting otherwise, but of course, it's plain to demonstration that as a matter of objective fact, or if another analysis of the satellite data slight variance to the assessment you've made were taken, and the growing season was said to start on 2 or 3 December, then that would increase the amount of rainfall over that particular growing season by proportionally, I suggest, quite a significant amount. A 46-milimetre (sic) figure when we're talking about overall amount of around 185 ml over the growing season. You'd agree with that. A. That would be true.
The evidence discloses that a "nominal" start date was only required to be determined as no direct evidence was sought to determine the actual start date. In circumstances where both the length of the growing period and the actual start date could have the consequence of materially affecting the water demand for the particular growing year the absence of enquiries being made to the Defendant is unexplained. The accuracy of the nominated start date and growing period is uncertain. Whilst Dr Meyer sought to use the evidence available to him, the question to be determined in these proceedings from his evidence is the water demand of the particular crop. Absent a more reliable determination of the start date of the crops and as a consequence the length of the growing period, this input into Dr Meyer's calculation is at best an educated guess.

[17]

Rainfall on Havana North

There is no direct evidence of the rainfall on Havana North in any of the relevant years. Dr Meyer has adopted the rainfall data recorded at the Myall Vale Weather Station. In addition to the rainfall data Dr Meyer also used the relevant climate variables data including temperature, solar radiant energy, humidity and wind from the Myall Vale Weather Station.
Dr Meyer noted in his statement of evidence (Exhibit Z) that "the weather data used needs to be obtained from locations as close as possible to the site of interest". During the course of cross-examination of Dr Meyer, the Defendant provided details of other weather stations in proximity to Havana North (Exhibit 2) which disclosed:
Dr Meyer accepted from considering the rainfall data for each of those weather stations that there was variability across the landscape. He stated in his evidence at Tcpt, 11 October 2022, p 307(4)-308(43): Q. So I suggest to you that that kind of data allows us to arrive at a number of propositions. The first is that evidently the rainfall is not the same at the gauges in these stations over the three or four month periods, December to April in any given year that we are considering, even though the stations are physically proximate to each other on the ground. There's A. I'd like to disagree, with respect, to the figures are here, as in Narrabri Airport, I suggest, is - is - is very different from the other one, two, three, four stations that you've - you've identified there. And you see, within those four stations, the variability is on an annual basis, is actually quite small. Q. But just asking you to go to - and I suppose you would say that the Narrabri Airport station, it's the furthest east, it's 61 kilometres away. A. Yes, and a lot closer to the mountains. Q. But asking you to consider the matter on a month by month basis-- A. Yes. Q. --and even, if you like, you know, disregarding Narrabri Airport, but considering the other stations, Moree, Bellata, Pendennis and Wee Waa, George Street, I suggest to you it's very evident from the figures in the table that there's a high degree of variability, even amongst those stations that are closer together, excluding Narrabri Airport, considered on a month by month basis, over the period December to April in each year. That's correct, isn't it? A. There is variability, as one would expect. Yes. But as I've indicated, I think, in my report, over time and over a - particularly over a five or six month period that we're talking about, a lot of that variability between stations tends to even out. Q. Well, certainly, as we've established by look at - it's some of those figures, it does not exactly even out over a four or five month period, although you may say, on average, the values of the stations converge. Is that what you're saying? A. They tend to be similar, that's correct, when you total them up. That-- Q. On a statistical average, over many, many years. A. A statistical basis. That's correct. Q. But in any given month, of course, as we can see from the table, even looking at those four stations, disregarding Narrabri Airport, in any given month, self-evidently, the rainfall at any one of those stations may exhibit a significant difference from the rainfall at any one of the others. That's something one can see in the table, isn't it? A. There is variability, no doubt about it, and they vary between months, but again, take, for example, in 2018, Bellata, and compare it with the other three stations, it has a high rainfall in February, but then low rain - a lower rainfall in March. And so you see that - that trend happens, it's the vary nature of rainfall in that kind - part of landscape. Q. But sometimes, of course, for example, in the 2019 year, we have one of Bellata's high rainfall months occurring towards the end of 2019 at the 55.6 in November. So, you know, the station may make up an amount, if I could put it that way, at a time it's outside the December to April growing period that we're dealing with. That's correct, isn't it? A. It can be - that can be the case. Yes. Q. Now, look, what one can certainly infer from this information, I suggest to you, as a scientist, is that across the landscape, this landscape within which Havana North is located, it's quite clear that the rainfall in this area, on any given day, in any given month, certainly does not fall uniformly across the landscape. A. That's correct. Q. For example, as there is no gauge - no rainfall gauge information that you've been provided with for gauges on or located on Havana North itself, if one were to go, for example, to the period of February 2018 in the table, even though we get, for example, in that February 2018 table, a figure of 39.2 for Pendennis, a station, you know, the closest station to Havana North, and in a north, south line, 11 kilometres from Havana North. We just don't know, at that time, February 2018, what the rainfall on Havana North was from these figures, do we? A. We don't know, but one - one thing we can say, it's not unlikely to be very significantly different from that-- Q. It could be higher-- A. --because of the very nature of rainfall across that landscape. Q. Exactly. It's highly variable. It would be surprising-- A. It is variable. Q. Surprising, very surprising, if it was exactly same. A. It might be exactly the same. Q. But it could be-- A. It could-- Q. --higher, lower, much higher, much lower. A. It could be-- Q. You just don't know, do you? A. It could be higher or lower.
In circumstances where the water demand for the crop is calculated and thereafter determined as a demand for irrigation water being the amount of water calculated by subtracting rainfall from the crop water demand and the remaining amount is assumed to be met by irrigation water a relative degree of precision is required. That degree of precision is either direct evidence of the rainfall on Havana North (which was not available) or a basis for inferring that the weather data relied upon by Dr Meyer was sufficiently akin to the rainfall expected on Havana North to justify the calculation undertaken. In this case the evidence discloses that the Myall Vale Weather Station may not be an appropriate proxy for Havana North as: 1. It is not the closest weather station to Havana North. The closest weather station is Pendennis weather station 11km from Havana North; 2. The variability on a month-to-month basis between relatively proximate weather stations indicates a variability. Such variability (with the exclusion of the Narrabri Airport Weather Station data) is not explained by landscape features. Therefore, it appears that it is merely a climatic feature of this particular landscape that the rainfall is variable even in locations in relatively close proximity to each other; and 3. Even if another weather station apart from the Myall Vale Weather Station was adopted for the purposes of Dr Meyer's calculation his acceptance that the variability could mean that the rain Havana North received could be higher or lower renders the reliability of the calculation of irrigation water unreliable to the extent of the unknown extent of any such variation in rainfall.
In light of the evidence as to variability, which was accepted by Dr Meyer, and the fact that there are weather stations located closer to Havana North than the Myall Vale Weather Station, the use of the weather data by Dr Meyer is unreliable. This is particularly so where no analysis has been undertaken to permit a finding that a particular weather station has comparability to Havana North in the context of the variable rainfall landscape. The variability of rainfall across the landscape renders the use by Dr Meyer of a single weather station (albeit high level within the range of measured rainfall) ascertaining, at best, what rainfall may be experienced at Havana North within the realm of probability. However, as the rainfall is crucial to the determination of water usage there must be a higher degree of satisfaction that the selection of this particular rainfall data is sufficiently representative of the rainfall at Havana North such that this figure can be adopted as a proxy to actual measurements at Havana North.

[18]

WFIE

Dr Meyer adopted an assumed WFIE of 70%. He did so based upon what he identified as his experience and in reliance upon the Roth Paper.
WFIE represents the amount of irrigation water applied that is actually available to the cotton crop and not lost by means other than crop uptake, such as by infiltration and evaporation. The percentage applied for WFIE has a significant impact on the total amount of water required to be applied to produce an annual crop.
Whilst Dr Meyer indicated that he relied upon his experience he accepted that the WFIE varied depending upon the particular farming practices at each farm. In that regard he stated at Tcpt, 11 October 2022, p 324(50)-325(11): Q. And so, you of course have not measured, and you've not been asked to, so not being critical of you, the actual on the ground irrigation efficiency at Havana North, have you? A. Not at all. Q. Looking at the level of increase that's documented there in this prime facts sheet showing whole of farm irrigation efficiency of 83 % in 2012/13, plateauing somewhat at 81%, still at 81% in 2017/18, you'd agree there's no reason to categorically rule out a whole of farm irrigation efficiency figure of 80% or above in this particular case. And you couldn't do that because you haven't undertaken the empirical exercise on the ground, I suggest. A. That's correct. I can't do that. I have no firsthand experience there.
Having not inspected the Property nor spoken to the farmer he was unable to ascertain what the particular WFIE for the Property was. Nor did Dr Meyer indicate the particular expertise he had in determining WFIE. Whilst he had experience in crop water utilisation his experience did not disclose that he was particularly experienced in the determination of WFIE or any examples where such calculations were undertaken to permit an acceptance of the expression of opinion based upon experience.
Apart from his statement that he relied upon his experience, his evidence relating to WFIE relied exclusively upon his reading of the Roth Paper. Therefore, I find that Dr Meyer's adoption of the 70% WFIE was primarily, if not exclusively, based upon an adoption of the figure identified in the Roth Paper.
The Roth Paper was published in 2013 but was based upon research undertaken in the 23 years from 1988 to 2011. The reference in the Roth Paper to a 70% WFIE related to the observations at p 1040 that: THE WFIE values show a wide range in the data. However, the yearly averages indicate a significant improvement over time. During the late 1990's, the WFIE was ~57%, whereas in the latest industry wide data, collected 10 yeas later, it has risen to ~70%. This indicates that there was less on-farm water loss and that more of the water used on-farm was used productively by the crop. Differences in seasonal conditions can also influence this performance indicator. For example, the highest WFIE was achieved in 2006-7, which was a very dry season. Soil profiles were dry and few irrigation storages were used. There was little, if any, in-crop rainfall across all regions, and surface water allocations were either very low or non-existent. As a consequence, the area planted to cotton on any farm was significantly reduced. This meant that the available water was used carefully and management would have been precise with only small areas to water. Irrigators would have planted their best fields closest to on-farm storages or water extraction points to reduce conveyance losses. The WFIE performance indicator provides an on-farm-irrigation-efficiency benchmark, but does not indicate specifically where water losses and inefficiencies are occurring. Further investigations are required to determination the potential water losses within farms. And to the conclusion at p 1044 that: The majority (at least 80%) of the Australian cotton-growing area is irrigated using gravity surface-irrigation systems. This review found that over 23 years, cotton crops utilise 6-7ML/ha of irrigation water depending upon the amount of seasonal rain received. The seasonal evapotranspiration of surface irrigated crops averaged 729mm over this period. Over the past decade, water-use productivity by Australian cotton growers has improved by 40%. This has been achieved with both yield increases and more efficient water-management systems. The whole-farm irrigation efficiency index improved from 57% to 70%, while the crop water use index in >3kg/mm.ha, high by international standards
The Roth Paper also acknowledged at p 1039 that: The average amount of irrigation water used for all of the studies in Table 1 was 6.97 ML/ha, with a range of 5.37-8.90 ML/ha. There was also a large range between the farms in any given year. The amount of irrigation water used depends on the seasonal rainfall received and the efficiency of farm irrigation systems. Seasonal variability between seasonal average results is evident and expected. For example, 2009-10 and 2010-11 were wet seasons, whereas 2006-7 was a hot dry year. These results have led to the farmers' rule of thumb that, typically, 6-7 ML/ha of irrigation water is required to maximise cotton production.
I do not accept that the adoption of a 70% WFIE is warranted either in the circumstances of this case or on the basis of the findings and proper application of the research in the Roth Paper.
First, the Roth Paper provides an average benchmark of an efficiency which will vary from farm to farm. In order to ascertain whether the "average" identified by the Roth Paper is applicable it is, as was observed at p 1040 extracted above, necessary to understand the irrigation and management systems adopted on each particular farm. By not undertaking the simple and available approach of inspecting the Property to ascertain the implementation (or lack thereof) of irrigation and farm efficiencies the application of the average benchmark identified in the Roth Paper and adopted by Dr Meyer is unreliable at best.
Secondly, the Roth Paper acknowledges that significant changes to average efficiencies can occur in relatively short periods of time. The move from the average of 57% to 70% occurred within a decade. Since the publication of the Roth Paper (and since the dates of the studies upon which it relied) there is evidence of improved efficiencies being adopted in the cotton growing industry. The publications in the 2016 Fact Sheet and the 2019 Fact Sheet indicate that there have been continued improvements in water use efficiency in cotton farming. The 2019 Fact Sheet indicated an increase in WFIE from 70% in 2006/2007 and 2008/2009 to around 83% and 81% in 2012/2013 and 2017/2018. In cross-examination Dr Meyer acknowledged such trend and stated at Tcpt, 11 October 2022, p 324(35)-325(11): Q. You can see the document I provided you is a document I think ultimately produced from yourself via the prosecutor. It's a Department of Primary Industries prime fact document dated October 2019. I'm asking you to turn to page 3 of the document, using the numbers at the bottom of the page. You can see there that one of the points, looking at the middle of that page 3, made in this Department of Primary Industries paper is that in Australian cotton farming whole farm irrigation efficiency is increasing significantly year on year. You'd agree with that. A. That - there is evidence of that. That's true. Q. And so, the figures that this paper sets out, or this fact sheet sets out there, is that there was a 70% figure that they identify in 2006 and 2008, to around 83 and 81% in 2012/13 and 2017/18. Do you see that? A. Yes. Q. And so, you of course have not measured, and you've not been asked to, so not being critical of you, the actual on the ground irrigation efficiency at Havana North, have you? A. Not at all. Q. Looking at the level of increase that's documented there in this prime facts sheet showing whole of farm irrigation efficiency of 83 % in 2012/13, plateauing somewhat at 81%, still at 81% in 2017/18, you'd agree there's no reason to categorically rule out a whole of farm irrigation efficiency figure of 80% or above in this particular case. And you couldn't do that because you haven't undertaken the empirical exercise on the ground, I suggest. A. That's correct. I can't do that. I have no firsthand experience there.
Third, the application of the 70% WFIE to Havana North produced an irrigation amount that exceeded general industry accepted water usage of 6-7ML/ha (Roth Paper at p 1039). Dr Meyer accepted that his water requirements as calculated significantly exceeded the expected water usage: Tcpt, 11 October 2022, p 326(44)-327(10). He sought to distinguish this on the basis that he considered the sums in the Roth Paper to be water "used by the crop" rather than the application of water. I do not accept this distinction. The Roth Paper, as is made plain at p 1039, relates to water amounts referenced in Table 1 of the Paper. The identification of a much higher irrigation water indicates a problem with the calculation which is not explained by Dr Meyer in any acceptable manner and is sufficient to indicate a high degree of overstatement of water demand in his calculation which is not explained or justified such that the figure should be treated with caution.

[19]

The IrriSAT computer programme

The IrriSAT computer programme is the foundational tool utilised by Dr Meyer to determine crop water demand. The IrriSAT Technical Reference identifies in the executive summary at p 1: IrriSAT is a decision support tool to assist irrigators with irrigation water management. The IrriSAT methodology aims to be as simple as possible (in order to limit the number of inputs and parameters required), yet sufficiently complex to accurately estimate irrigation requirements. The IrriSAT methodology described in this document can be summarised as follows: • The water balance approach to irrigation scheduling keeps track of the soil water deficit by accounting for all water additions and subtractions from the soil root zone on a daily basis. • The IrriSAT methodology uses a simplified approach whereby: ∆.S = P + I − ETc. This approach assumes crop water consumption (or evapotranspiration) accounts for the biggest subtraction of water from the soil root zone while precipitation and irrigation provide the major additions. • The soil in the root zone has an upper limit of storing water that can be used by crops. This upper limit is constrained to the field capacity. • As the crop grows and extracts water from the soil to satisfy its water use requirement (ETc), the stored soil water is gradually depleted. • ETc is estimated using field observations and nearby climate observations. • Crop reference evapotranspiration (ETo) is estimated using the FAO Penmen Monteith Tall-Crop (alfalfa) reference using observations obtained from weather stations. • Crop coefficients (Kc) are estimated by directly observing the crop growth on fields using remote sensing techniques. Strong relationships between NDVI and Kc can be used to achieve this. • Tracking the water balance deficit in the root zone allows for a refill point to be defined which indicates when irrigation is required. • Common metrics which can be used to measure the agronomic performance over a growing season include: the irrigation water use index (IWUI); crop water use index (CWUI); and gross production water use index (GPWUI).
As is to be observed from the executive summary the IrriSAT tool is not designed as a tool to determine the past water demand for a particular crop, IrriSAT is a tool to assist in decision-making during crop growth. It is designed to be both simple and complex.
It is a software programme that relies upon the determination of statistical outcome. To that extent it is utilising mathematical algorithms to determine the inputs into the ultimate calculation for water demand. In this process the programme has inbuilt a number of assumptions - some to be the subject of input by the operator and some inherent in the programme.
On page 2 of the IrriSAT Technical Reference it states: Simplifying the Water Balance The IrriSAT water balance model is a simple approach to tracking water deficit in the soil root zone. It further simplifies the water balance described in eq (2) by making the assumption that subsurface flow in and out are generally negligible (consistent with flat irrigated fields found extensively across the major irrigated areas). Additionally, capillary rise is assumed negligible, which would be the case with a water table deep below the soil surface and efficient irrigation practices. In certain irrigation situations these assumptions may not be met so it is important to understand the particular circumstance in terms of how this approach represents these situations. If for instance you have significant deep drainage or capillary rise it is possible to reduce/increase your indicated irrigation inputs to take these elements into consideration. However, in the majority of situations these components are generally small…
The text makes clear that the assumption as to drainage depth and capillary rise may vary and should be considered prior to adoption. Dr Meyer adopted the default assumptions. In cross-examination he stated at Tcpt, 11 October 2022, p 332(4-18): Q. In terms of - could I ask you to go to the IrriSAT technical reference, that I handed you a number of documents ago, and looking at the top of page 2, you'll see there that one of the assumptions made by the water balance model, looking at the first three or four lines at the top of that page is that, the assumption is made that subsurface flows in and out are generally negligible. Do you see that? A. That's correct. Q. Another assumption made by the IrriSAT methodology is that capillary rise so the supply of the water from the subsurface is negligible. Do you see that? A. Yes. Q. And you, of course, haven't been asked to, and haven't had an opportunity to test either of those variables on Havana North, have you? A. Not at all.
Whether the default assumptions are an appropriate reflection of the water balance on Havana North is unknown. As a consequence, there is a degree of uncertainty in the modelling undertaken by Dr Meyer based upon those assumptions.
The Defendant took issue with the use by Dr Meyer of the assumption inputted into the algorithm to enable the estimation reference evapotranspiration that the cotton crop was to be treated as a "tall" crop from the point in time it appeared as green on the satellite images rather than when it reached a height that qualified as "tall" under the IrriSAT definition.
The IrriSAT Technical Reference notes at p 5 that the determination of ETo adopted two standardised crop heights that include a "'short crop" with an average height of 12cm high and a "tall crop" with an average height of 50cm high. Dr Meyer used the tall crop as the reference for the total period of the crop growing period and did not adjust it from short to tall with the growing of the crop. Dr Meyer gave evidence at Tcpt, 11 October 2022, p 317(28)-318(45) that: Q. One of the variables that one can see in the equation in the numerator of that equation is the variable Cn. Do you see that? A. Yes. Q. And that, of course, as set out in the table on the middle of page 5, that's a value that changes according to whether the crop is assumed to be short or tall. That's correct? A. That's correct. Q. As set out in page 5 in this IrriSAT technical reference, the reference to short crop, it's said there, is a crop of around, such as grass, around 12 centimetres high. Whereas a tall crop may be something like alfalfa, around 50 centimetres high. That's correct? A. That's correct. Q. Of course, having regard to the fact that the exercise that you've been asked to undertake in the database is these satellite images of Havana North. And just reading the four corners of your expert report, you weren't provided with any figure for the height of the cotton crop at any particular time over the relevant growing season. Were you? A. No. Q. And you'd agree with me that looking at the difference that arises, depending on whether one goes from a short crop value with a figure of a CN value of 900 in the numerator of the ETo equation, to a tall crop, with a value for CN of 1600 in the numerator, that is a factor that could introduce, I suggest, material variability into the daily ETo value. That's correct? A. No, it doesn't, because you either choose one or the other, depending on the crop that you're dealing with, you don't change the CN value for a particular crop, in this case cotton, which is a - would be regarded - you would use, always use a - an ETo value, which is, uses the tall crop coefficient, because the evaporated characteristics of cotton, and particularly in the environment we are talking about, the tall crop reference, and ETo is the - is the one, the standard one which would be used at all times. Q. You understand, of course, that for example in the - or you'd be prepared to accept from me that in the first month of the cotton growing season on Havana North the cotton crop is generally around 10 centimetres to 25 centimetres in the height. That would seem to be about right? A. That's correct. But the - the way in which that is taken into account is that the Kc value which is applied is appropriate for that particular size crop, with that particular area - green leaf area, which is derived from the satellite in the image. Q. But even in that, just taking that first month period and a cotton crop of, I'm asking you to assume, no higher than 25 centimetres in height. So taking the period from mid-December 2016 to, say mid-January 2017, you agree that if you were calculating crop water usage for that particular month, the relevant type of CN variable that you should be including in the ETo equation for that period, when the cotton crop is less than 25 centimetres, would be the short crop CN figure of 900, rather than the tall crop CN figure of 1600? A. No, that's not correct. You would use - you would always use the tall crop calibration of the Penman-Monteith, because the - and again, from my 20 years of experience of measuring evapotranspiration ETc of crops in these kind of environments indicates that in Australian conditions, with these kind of crops, the reference evapotranspiration value calculated for the tall crop is the appropriate one to use. And the adjustment, in terms of the height of the crop, is accounted for by the amount of leaf area which is associated with that shorter crop. In other words, it would be a smaller leaf area, which is then calibrated through the Kc value, derived from the crop canopy, which is evident. Q. So you suggest that this Kc value varies according to the greenness of the crop, evidently, but you are prepared to accept for the purposes of your calculation an ETo value over that period of time during the growing season, that first month, that does not in fact reflect the actual height of the cotton crop? A. It's irrelevant.
The IrriSAT Technical Reference identifies the coefficients as utilised in daily calculations. I do not take this reference to mean that the height of the crop must be determined daily - it is merely a reference to the fact that the IrriSAT calculation is to be done on a daily basis and will include the coefficient for the type of crop planted as either short or tall and not how it presents as a height on any given day. I also note that the assessment of Dr Meyer was consistent with the evidence of Mr Purcell who stated in cross-examination at Tcpt, 13 October 2022, p 497(39)-498(6) that: Q. Because when you are using IrriSAT at some point you have to make a selection as to whether or not it's a small crop or a short crop. Correct? A. No. It's you stick with the one and there's factors in the - when it's very short, I mean, the cotton plant might with that high. There's a very good chance IrriSAT can't look at anything because it's such a small area of the total field is leaf that it's almost useless, so when it gets further - more developed, then you can use it. Q. I had understood IrriSAT enabling you to make a selection at the start of the process, as it were, of doing the calculation between a short crop and a tall crop - right - depending on the particular crop in respect of which you are seeking to work out water usage, and then that is one of the factors that goes into the IrriSAT algorithm to give you the water usage requirements over time. A. I think if - if somebody was going to that extreme, frankly, I would say they're overdoing the basis of IrriSAT. IrriSAT is based on the tall crop, and I've talked to John Hornbuckle recently about this, because there was a lot of academic toing and froing about whether it should be tall crop or short crop. He has now set it up so it's ready for tall crop, trying to make it easier to use.
I therefore accept that the use by Dr Meyer of the tall crop coefficient for cotton to be appropriate in this circumstance.

[20]

The ±15% rate of uncertainty

Dr Meyer indicated that his figure, as calculated, had an error rate of ±15%. Dr Meyer attested to the basis on which he adopted his rate of uncertainty at Tcpt, 11 October 2022, p 320(10-35) as: Q. You've indicated in your report that you are prepared to attach to your crop water usage figure an uncertainty value of around about 15%. That's correct? A. That's correct. Q. There is, of course, no reason why that approximate uncertainty figure couldn't be slightly higher or slightly less than the 15%. It could have been selected by you, for example, at 20% quite reasonably, I suggest? A. No. The - the figure I gave - I gave you that figure, the 15% is on my experience through, again, 20 years of measuring ETc and ETo, and looking at how well we do in terms of using that relationship between ETo and ETc using the Kc value. And as it indicates that variability is plus or minus 15%. So, in general terms, it could be slightly greater or slightly less of that sort of order. Q. Yes, that's the nature of an uncertain figure. But what I'm suggesting to you is that there is no process of calculation or formula from which that uncertainty figure has been derived as you, I think, have suggested, it's a matter for judgment, and that's the figure that arrived at by you applying your judgment and experience. A. Yes, from my experience. Not my judgment. From my experience of the calculations and the measurements that I've made. Q. But there is no formula in this particular case for example, derived from the IrriSAT programme or any other formula in the IrriSAT reference material that you can take the Court to that explains the calculation of that particular figure is there? A. No.
I take it from this evidence that Dr Meyer is suggesting that the uncertainty rate has been developed from his experience in measuring in real time ETc and ETo and/or using IrriSAT for the purpose it was designed, namely as a daily decision-making tool utilising inputs in effective real time, rather than an after the event tool to estimate water usage of a crop sown and harvested prior to the exercise being undertaken. In such circumstances, Dr Meyer accepts that an error rate of ±15% may be slightly greater or slightly lower.
The calculation being undertaken by Dr Meyer in this case, however, is based upon a number of unknown and untested assumptions and without a familiarity with the farm to which he is applying his assumptions. In such circumstances the degree of precision to which the ±15% error rate was derived is absent. In such circumstances where assumptions must be applied without knowledge of the applicability of those assumptions including particularly his unfamiliarity with the farm in question, the limitation of the uncertainty to ±15% cannot be accepted. Whilst I accept that there is a degree of uncertainty, I do not accept that it is, or is limited to the range of ±15%.

[21]

Professor Albert Van Dijk

Professor Albert Van Dijk holds qualifications in Earth Sciences. He is presently the Director of the Centre for Water and Landscape Dynamics and holds a chair in Water Science and Management at the Fenner School of Environment & Society, Australian National University. He has undertaken research with respect to the interaction between vegetation and the hydrological cycle and related processes.
Prof Van Dijk prepared a written expert report (Exhibit N). In his report Prof Van Dijk indicated that he was requested to provide an estimate of the annual level of rainfall at the Property over the relevant water years. He was also requested to estimate the volume of runoff that would be available for capture by the Property during those relevant water years.
Prof Van Dijk concluded: 1. The best estimate for total rainfall: 1. For the period 1 July 2016 to 30 June 2017 was 544mm; 2. For the period 1 July 2017 to 30 June 2018 was 329mm; and 3. For the period 1 July 2018 to 30 June 2019 was 285mm. 1. Of the amount of total rainfall for each year 3.3% of that rainfall would have fallen directly into the water storage dams; 2. Of the amount of total rainfall each year an additional amount of runoff was captured as tailwater. This calculation included water captured flowing onto the site from areas outside the Property boundary to the south-east. The amounts of overland flow that would have been captured into the water storage dams in the following years: 1. For the period 1 July 2016 to 30 June 2017 was 207ML; 2. For the period 1 July 2017 to 30 June 2018 was 63ML; and 3. For the period 1 July 2018 to 30 June 2019 was 16ML; 1. Each of these figures have an element of uncertainty.
Prof Van Dijk observed that there was no rainfall data available for the Property as the Defendant did not maintain rainwater gauge records.
Prof Van Dijk estimated the annual rainfall using a statistical method. He stated: Daily rainfall for any location Australia can be estimated using interpolation, a statistical method to estimate rainfall at all points in between individual rainfall measuring stations. These interpolated estimates are available for any location in Australia from two sources: (1) the Bureau of Meteorology ("BOM" from here on) and (2) Queensland Government ("SILO" from here on). Both estimates are obtained using very similar methods and a very similar set of gauge observations, and are also quantitatively very similar. Both estimates were downloaded for property coordinates.
Prof Van Dijk utilised the data obtained from seven weather stations in the locality. From that data he determined both the rainfall and the runoff as calculated by him.
The weather stations he selected were all within the range of 155m to 190m above sea level. Given the very flat terrain of the landscape he did not expect strong topographic variation, although he expected average rainfall would decrease towards the west.
The statistical interpolation estimates obtained by BOM and SILO are based on surrounding gauges and can be considered an average of them, weighted by their proximity. As a general rule, rainfall differences for locations closer to each other are smaller than differences for locations that are further apart, although for annual total rainfall Prof Van Dijk did not detect such a distance effect for the stations he considered, and therefore considered rainfall at each of the stations to be equally reasonable estimates of rainfall at the Property. As his best estimate (drawing on a broad array of available data), he adopted the average of the BOM and SILO estimates.
As to his method of calculation of the available runoff, Prof Van Dijk drew on data from surrounding weather stations and sought to identify events where rainfall exceeded a level required to generate runoff (the initial loss or IL level). Prof Van Dijk considered that a range of 25mm to 67mm IL was appropriate given the nature of the assumed soils (vertosols), the circumstances in which runoff generation was being considered (on levelled fields) and drawing upon research that had been conducted of runoff generation from irrigated cotton on vertosols (which found that runoff generation occurred after 25mm to 65mm of rain).
Prof Van Dijk noted that vertosols are clay-rich soils that have strong shrinking and swelling properties and crack when dry. These cracks then increase the amount of rainfall required to fill the cracks and consequently the amount of rainfall required for the soil to produce runoff.
Prof Van Dijk operated on the basis of a range of estimates and, due to the uncertainties involved, he recommended considering the full range of estimates. His best estimate was the middle of the range of his IL estimates (46mm).
In terms of potential runoff from off-property, Prof Van Dijk considered that an additional area of 200ha may have been able to generate runoff capable of being captured on the Property. This area of 200ha fell to the south-east of the Property and was bounded to the east by a drainage channel which Prof Van Dijk identified in LiDAR data as being about 80cm or 90cm deep.
The area of land available for runoff was identified by representation in Figure 1 of Exhibit N:
Prof Van Dijk was cross-examined upon his evidence to the following general effect: 1. He was not an irrigation engineer; 2. He had not been to Havana North and had made his assessment from a consideration of satellite imagery; 3. He had no data to determine the rainfall or runoff that actually occurred on Havana North. He derived an estimate of rainfall based upon the derivation of a statistical probability having regard to the data from the weather stations he considered. He derived a statistical probability that there was a one in six chance that the rainfall at the Property was higher than the highest value derived from the six weather stations. The probability would change depending upon the number of weather stations examined as part of the exercise; 4. As to the area of offsite runoff harvested, he had only considered an area to the south-east as his letter of instructions had indicated that the Defendant had identified that area as an area for runoff collection. He did not consider any other locations for non-flood but usual rainfall runoff harvesting; and 5. In the event of flood, the Property was within a flood plain and flooding of the Namoi River could reach the Property depending upon the height of the flood.

[24]

Mr James Purcell

As previously observed, Mr James Purcell had attended at the Property. He observed that: 3.2 Floodplain Harvesting Potential 3.2.1 Havana North has floodways on all four sides of the farm. These floodways are part of the lower Namoi Floodplain Management Plan (see Attachment 6 and 7). 3.2.2 Havana North has five (5) Water Harvesting Points (see Attachment 4 and 5) where any water in these floodways can be captured by below ground drains and pumped into any one of the four (4) dams or ring tanks. Water can be captured at these Water Harvesting Points whenever water is in the floodways. Water can be flowing in the floodways from local storms upstream to the south-east of Havana North or from Namoi River floods like November 2021 and September 2022 where flood water breaks out of the northern banks of the Namoi River and flows through the designated floodways. 3.2.3 The local catchment area in the floodways upstream of Havana North has been determined using publicly available Digital Elevation models. The catchment area is approximately 2,932ha as shown in Attachment 8.
Mr Purcell identified the location of the five water harvesting points in Attachment 5 of his report (Exhibit 1) reproduced as follows:
He considered the appropriate calculation of rainfall and runoff to be: 6.4 Effective Rainfall 6.4.1 At any time during the crop season when it rains, some of the rain infiltrates into the soil (initial loss) and adds to the soil moisture in the plant root zone. This is effective rainfall. 6.4.2 There are a number of rainfall runoff computer models available that calculate the component of effective rainfall and the remaining rainfall runoff from a catchment or field during a rainfall event. 6.4.3 In my opinion, one of the most appropriate rainfall runoff models for farms is the daily rainfall runoff model developed by the Soil Conservation Service of the United States Department of Agriculture called "USDA", 19713. 6.4.4 I have been using this model for rainfall runoff and infiltration since 1972 and while working for the Queensland Water Resources Commission in 1977, I worked with a project comparing gauged catchment runoff volumes versus USDA modelled runoff. While the runoff naturally varied with rainfall intensity, the seasonal volume of runoff measured was consistent with the modelled volume. 6.4.5 The other advantage of the USDA model is the runoff coefficient depends on the: • Land use or cover; • Farming treatment; • Hydrologic condition; and • Soil type. 6.4.6 The initial infiltration and runoff component of a given rainfall event is also varied in the USDA model according to the antecedent moisture condition (AMC) which is determined by the magnitude of rainfall recorded in the previous five days before the rainfall event or if the irrigation field has been irrigated. 6.4.6 The USDA rainfall runoff model is designed to work in agricultural systems. 6.4.7 Once the runoff component of a rainfall event is calculated the infiltrated component is also known and adds water to the soil moisture in the crop root zone. 6.4.8 Given that the irrigation of a field changes the AMC value used to calculate the runoff from that field, it is necessary to have records of when each field starts being irrigated and when each field finishes being irrigated. For instance, if it took 2 days to irrigate a field in 4 shifts, then a rainfall event on day 2 would yield more runoff from the first half of the total field. 6.4.9 Obviously, good rainfall records over the full area of the farm are required to use any model. Unfortunately, Havana North do not use rainfall gauges at all on the farm. 6.4.10 When asked why, Mr Garry Phelps, owner of Havana North, told me that he used to have three rainfall gauges but found that in most storms the measured rainfall varied significantly across his farm, so he stopped using them. 6.4.11 Modelling rainfall runoff using any model without accurate and reliable rainfall measurement is problematic. As a result, in my opinion, it is not possible to calculate reliable estimates of available rainfall runoff or effective rainfall at Havana North from the currently available data. 6.5 Floodplain Harvesting 6.5.1 While it is obvious from an inspection of the farm infrastructure that significant floodplain harvesting is possible at Havana North, again no measurement or recording of floodplain harvesting take has been kept by Havana North. 6.5.2 Attachment 8 shows the ground contours of the floodway catchment upstream to the south-east of Havana North. These contours show that the catchment is approximately 2,900ha. I note that this catchment area was estimated by Professor Van Dijk to be "close to 200ha" in Line 411 of his Expert Report. 6.5.3 There are two below ground drains shown in blue on Attachment 8 on the eastern side of Doreen Lane which will allow capture some of the floodplain flow but that still leaves a significant volume of runoff passing Havana North's water harvesting infrastructure. 6.5.4 Again, without reliable and accurate rainfall measurements, calculation of runoff from this catchment is problematic. 6.5.5 I note the calculated estimate of rainfall runoff from Havana North and the adjacent floodplain in the Expert Report of Professor Van Dijk. In my opinion, an initial loss for any rainfall event of 67mm with a continuing loss of Nil, as proposed by Professor Albert Van Dijk in Line 185 to 197 in his Expert Report is not suitable for the soils in the vicinity of or on Havana North. A comparison with the USDA rainfall runoff model indicates that for the soils at Havana North (soil group C or D) the antecedent moisture conditions (AMC) values of: a. KII value = 84 b. KI value = 70 c. KIII value = 95 The initial losses for theses AMC values are: KI = 21 mm KII = 9 mm KIII = 2.5 mm The initial loss of 67mm would be for a deep sandy soil like a Cypress Pine Forest not an irrigation farm on cracking clay soils. I note that Professor Van Dijk acknowledges in line 195 of his Expert Report that the AWRA-L model used to calculate the 67mm initial loss "was designed and calibrated to estimate river inflows in freely draining catchments, not to estimate tailwater generation on irrigation farms". 6.5.6 In my opinion, the estimated overland flow and rainfall runoff from Havana North of 207 ML by Professor Van Dijk in line 430 of his Expert Report is incorrect and would significantly under estimate the actual rainfall runoff because the catchment area is incorrect and the likely initial loss of 46mm, in my opinion, is significantly too high when compared to the values from the USDA model above. The modelling by Professor Van Dijk also does not consider the Antecedent Moister (sic) Condition of the external catchment or the fields on Havana North to calculate the runoff.
Mr Purcell was cross-examined upon his evidence. He was criticised for his use of a model developed in the United States for application in Australia. Mr Purcell attested that he had used the model for 30 years in NSW and was aware that it was utilised in other states of Australia. He maintained his position that the model was an appropriate and accurate measure of predicting runoff. Mr Purcell accepted that matters of judgment and experience were required to implement the model, and that the accuracy of the outcome depended upon such judgments.
As to the use of rainfall data he was concerned that where farm specific measurements were not available that the use of a proximate weather station data was examined for application to a nearby farm. He stated at Tcpt, 14 October 2022, p 525(29)-526(19): Q. In other words, you've carried out the exercise of predicting rainfall on a farm by reference to rainfall data from surrounding weather stations, having satisfied yourself that they are representatives - so the data is representative of the likely fall on the farm. Correct? A. No, sometimes I'll say, "No, I'm sorry. I've got no rainfall data for your farm. I can't help you." Q. All right. But on other occasions, as I take it from one of your earlier answers, you have come to the view that yes, I can predict the runoff on your farm based on the rainfall data from the surrounding weather stations. Correct? A. It would have to be very close to the farm for me to do that. Q. Do I take it from what you said earlier about topography that certainly in locations where the land is relatively flat as it were, you would be less concerned about variability between or variance I should say, between the recorded rainfall data at a nearby weather station and that on the farm? Do you accept that? A. No, I don't. You - you're missing the point. Rainfall - spatial variation with rainfall is unpredictable and quite bad. So, what I'm looking for is - is a rainfall station at Havana North because that's where I'm calculating the - the runoff and the infiltration. If I can't get that, I've got to make a decision whether I'll even do the exercise because the rainfall data is key information for the model. It is all based on daily rainfall figures. If those figures aren't accurate and representative to that farm, I can't calculate the runoff for that farm. Q. When you say representative for that farm - I think you given the earlier answer already - that rainfall data from, you accept, rainfall data from surrounding weather stations can sometimes in your own assessment be representative of rainfall on the farm. Correct? A. No, you've - you - you take - you've taken it further than what I - what I explained. You've got to understand that even within the farm of Havana North, there's three rainfall gauges we used previously and found it a very - that's the problem you have when you don't have local rainfall data for that farm. So, for me to give you a reasonable runoff figure, I need to have rainfall at that farm or from a BOM station which is very close to that farm. Q. I had thought from your earlier answer that you would not only look at the one nearby weather station, but you would also look at various nearby weather stations. A. Well, it depends if there's - if there's various stations close to the farm.
Mr Purcell was also cross-examined upon his designation of soil types as it affects runoff. Mr Purcell was critical of Prof Van Dijk utilising a standard amount of rainfall that would be lost to the soil profile prior to runoff (the IL value). Mr Purcell had adopted soil classifications that varied depending upon the treatment such soil had undergone at the time of rainfall, for example, a cultivated field compared to a fallow field. Depending upon the nature of the use and treatment of the land the runoff will vary.
Further Mr Purcell indicated that in order to determine likely runoff volumes it is necessary to know the amount of rainfall in the preceding 5-day period. He explained this evidence at Tcpt, 13 October 2022, p 503(25-50): Q. So, answer that question. A. Okay. What causes run-off from any catchment is the soil type, the cover, the antecedent, moisture condition, and the amount of rainfall on that day. All of those effect how much run-off comes from that catchment to Havana North. Q. The antecedent moisture condition, is that the same thing as how wet the soil before the rainfall event? A. No. Q. What do you say-- A. How wet the soil is when - we had a storm today, if there has been virtually no rain for the previous five days, there will be less run-off that if it was raining like hell for the last five days. Q. Precisely. I accept that. And vice versa-- A. Yeah. Q. --if, for example, we had a rainfall event on this particular day and it hasn't rained for two weeks, three weeks, for example, then one might say, logically, that it would take a whole great more of rainfall amount for the rain to run. A. It's actually more sensitive in the five days, because once it gets past five days, the effect of the dryness reduces. The peak effect is in the - in the previous five days. It could go for months, but it doesn't - it doesn't continue to degrade at the same rate.
Absent information of the type identified by Mr Purcell he was of the opinion that it would not be possible to calculate an amount of runoff on Havana North.

[25]

Prosecutor's submissions

Given how dry the 2018 and 2019 years were in particular, the Court may consider that it is likely that the soil experienced cracking of the type described by Prof Van Dijk. Potential runoff events onto soils in such a state might be expected to attract an IL value towards the higher end of the range endorsed by Prof Van Dijk, making the adoption of a lower IL value (and an assumption of no continuing loss) a particularly conservative approach.
The Prosecutor submitted that the approach of Prof Van Dijk to the estimation of rainfall constituted sound scientific practice, drawing on the best available data. The Defendant did not maintain rainfall gauges on its Property and, in any event, gauges that are not maintained by weather stations may be subject to unreliability due to factors such as: substandard equipment; improper placement; and the absence of a system of routine checking (or adherence to any such system). What Dr Van Dijk relied upon was accepted scientific practice.
The calculation of runoff by Prof Van Dijk again used the appropriate and available data from surrounding weather stations. His determination of the IL value was appropriate having regard to the nature of the vertosol soils and was supported by research
In terms of potential runoff from off-property, Dr Van Dijk considered that an additional area of 200ha may have been able to generate runoff capable of being captured on the Property. This area of 200ha fell to the south-east of the Property and was bounded to the east by a drainage channel which Dr Van Dijk identified in LiDAR data as being about 80 or 90cm deep.
It is apparent from the evidence of Mr Purcell that there are two such drains to the east of the Property and, from the floodway map, it can be seen that those two drains cover the entire width of the floodway. While significant enough events (such as flooding of the Namoi River) might generate sufficient flow volume and velocity to overwhelm those two drains, in the absence of any evidence of such events, the basis upon which Prof Van Dijk operated (contemplating an area of 200ha) was both rational and reasonable.
Outside of the September 2016 flood event, there is no evidence of any such significant events occurring over the relevant years. The maximum daily rainfall at any of the surrounding weather stations between July 2016 and December 2019 was 62.1mm at the Wee Waa (George Street) Weather Station on 30 March 2019. The next highest separate event recorded was 52.4mm at the Bellata (Aberfeldie) Weather Station on 26 January 2017. The Court may find the rainfall records entirely unsurprising given that the region (and much of NSW) was experiencing significant drought conditions in 2017, 2018 and 2019.
Insofar as the Defendant may suggest that it is reasonably possible that the Property experienced significantly higher rainfall than that recorded at other areas within the region, the Court would apply its common sense. Much of the charge period contemplates drought conditions (including, in some instances, weeks where no rainfall was recorded at any of the weather stations proximate to the Property). The dryland crops (dependent upon rainfall) performed poorly. The speculative proposition, unsupported by any data and contrary to evidence of the performance of the dryland crops in 2016/2017 and 2018/2019, that the Property may have experienced significantly higher rainfall, would not be entertained as a reasonable possibility.
None of Mr Purcell's evidence undermined or put in doubt the reliability of the expert evidence of Prof Van Dijk.

[26]

Defendant's submissions

There were large, acknowledged, uncertainties in the estimation of rainfall on Havana North. By way of example, the statistical probability derived by Prof Van Dijk would vary depending on the number of weather stations used in the analysis. If only two weather stations were used, then there was a 50% chance of a higher rainfall figure than the highest taken at those stations over one year and a 1/8 or 12.5% chance that occurred every year over three years.
Prof Van Dijk's overland flow calculations were limited by his instructions to only that area of 200ha to the south-east and had not investigated other harvesting points in the north. In his oral evidence, he acknowledged the amount he calculated could have been greater than he had identified, as a possibility. The omission of a consideration of runoff of the four other identified harvest locations and having regard to the catchment area identified by Mr Purcell places significant unreliability on these calculations.
There was acceptance by Prof Van Dijk that there was variability of rainfall in this landscape. The assumption that a determination of statistical probability of rainfall from a number of weather stations does not overcome the issue of variability - no witness including Prof Van Dijk was able to attest as to what the rainfall in fact was at Havana North or on what basis the analysis done by him would be a sufficient proxy for that determination of fact. This is a significant deficiency in the Prosecution investigation and evidence. This figure is obviously crucial and an essential fact in proving how much came up from the bores in each year. In summary, a selection of the evidence as to the unreliability and uncertainty as to the amount of rainfall to be attributed to Havana North is as follows: 1. Prof Van Dijk says the best estimate of rainfall in 2016/2017 on Havana North is 544mm but with a 16.6% chance it was equal to or more than 606mm; 2. Prof Van Dijk says the best estimate of rainfall in 2017/2018 on Havana North is 329mm with a range of 240mm-434mm and either the same 16.6% chance of exceeding 434mm in this year considered alone or 3% chance if the year before had exceeded 606mm; 3. Prof Van Dijk says the best estimate of rainfall in 2018/2019 on Havana North is 285mm with a range of 223mm-346mm and a 16.6% chance of exceeding this upper figure in this year considered alone but less than that if each of the other two previous years had also considered their upper end of the range figure; 4. Prof Van Dijk opined that the variation in rainfall across this landscape is "largely random"; 5. Dr Meyer conceded rainfall across this landscape was variable; and 6. Mr Purcell's rainfall spatial variation is that it was not possible to calculate reliable estimates of non-bore water sources due to absence of required data (including rainfall data).
Prof Van Dijk acknowledged that he could not rule out overland flow when his Sentinel satellite images had not shown the land, which was the case when the land was obscured by cloud or in the gaps between the satellite images taken every fifth day. This creates a reasonable doubt, in itself, as regards to the frequency of overland flow events.
When Mr Purcell's evidence as to the larger catchment allowing rainfall to flow to Havana North and the unknown rainfall figures for either anywhere in the catchment or Havana North are considered, the existence of a reasonable doubt as to one source of water able to supply the alleged deficit asserted by the Prosecutor is unavoidable. As was explained by Mr Purcell, the drains seen in his catchment diagram would not stop all overland flow, and Prof Van Dijk has not factored in overland flow from the north of the property or considered the input from off farm overland flow from the other farm collection points identified by Mr Purcell and overland flow from rainfall to the east from Mr Purcell's 2,000ha catchment cannot be ruled out even given the drains identified by Prof Van Dijk and on Mr Purcell's map. Prof Van Dijk has assumed that no runoff can reach Havana North from the east of these, but as Mr Purcell has explained the movement of runoff past those drains to Havana North is dependent on rainfall.
There is no doubt that during the September 2016 floods, having regard to Mr Falkenmire and Mr Campbell's evidence flood flows would have reached Havana North, but crucially there is a reasonable doubt not disproved by the Prosecution case as to the amount of ordinary rainfall runoff (that is, non-flood runoff) reaching Havana North past these drains from time to time. As Prof Van Dijk concedes the area to the east is "one big floodplain effectively", the drains may not even be there anymore, and he has not inspected them on the ground although he "believes" their depth was about 80cm or 90cm at some unspecified date at some unknown point in time from LiDAR.
Further, there is no evidence before the Court as to the significance of the multiple rises in the Namoi River revealed in Annexure B of the affidavit of Andrew Falkenmire, although below the peak of September 2016, during the subsequent years. Whether or not the fluctuations shown in Annexure B lead to additional overland flows from the Namoi River into the flood catchment and past Havana North has not been ruled out on the evidence and is a question raised by Mr Falkenmire's evidence itself.
In addition, Prof Van Dijk accepted that it was a "distinct possibility" that things could have changed as regards to the drains since he saw them in the LiDAR image of an unspecified date Further, there are no calculations or images supporting that opinion and no evidence as to the slope of the sides or shape of the channels, or whether they are permanently moist, all matters that could affect the ability of overland flows to move past the drains from west to east. The significance of that depth is unknown from the evidence and cannot be known as Mr Purcell explains, without rainfall figures for the catchment, and it would follow from Mr Purcell's evidence, information as to the soils and treatment (for example, crop growth) within the catchment at the relevant times. Mr Purcell did not agree that overland flows could not move past these drains and gave evidence that this further information would be required to determine which flows would, and which would not.
Mr Purcell's evidence creates doubt as to the reliability of Prof Van Dijk's opinions in relation to both the extent of the catchment to the east, whether he has considered all points of overland flow into Havana North, including from the north, and the reliability of his conclusions as to the amount of runoff from the land surface of the farm and off the farm as he has not considered the surface treatment (for example, crops grown, whether there is mulch on the ground, type of crop and stage of growth) or the soils of the larger catchment area identified by Mr Purcell or in the northern areas.
As Prof Van Dijk conceded he is not a farm irrigation engineer. Mr Purcell is, and the Court would accept his evidence over Prof Van Dijk's on this matter or at the very least accept that it identifies an unresolved contest between experts and hence a reasonable doubt. Professor Van Dijk's identification of 46mm or 67mm as the IL value on Havana North is highly questionable, and likely wrong, having regard to both Mr Purcell's evidence and the Roth Paper.

[27]

Findings on rainfall

The amount of available rainfall has the potential to affect the amount of irrigation water required, and therefore the amount of irrigation water to be inferred from the evidence to have been taken from the bores. Rainfall has a direct contribution being the amount of water that falls on the crop and into water storages. Rainfall also has an indirect contribution being the harvesting or collection of runoffs from surrounding land by the Property for use by it in the cultivation of the crop in each water year.
There was no direct evidence of the amount of rainfall that in fact fell on the Property on the relevant dates. Rainwater was not measured at the Property. The Prosecutor has sought to rely upon rainfall data collected at various measuring locations in relative proximity to the Property. Prof Van Dijk considered seven such locations.
Prof Van Dijk accepted that there were differences in rainfall across the region, as evidenced by the wide range of rainfall data collected on each day from the numerous weather stations. His calculation of rainfall as a statistical probability was the only calculation available to him on the data available and in those circumstances, he was doing the best he could with the evidence available to him to determine a "likely" estimate of rainfall. That statistical probability, however, did not account for the evident variability as between the seven weather stations, but assumed a degree of consistency. On that basis, absent some evidence to establish that Havana North would have the amount of rain determined as a statistical probability from the consideration of the seven weather stations, the actual amount of rain that fell on Havana North remains a statistical probability but not evidence of a fact that such amount fell on Havana North. In light of there being analysis in the evidence of the comparability of rainfall on Havana North to the range of weather stations considered, at best the full range of the probability must be examined as available in any calculations in which Prof Van Dijk's estimates are utilised, and they are to be understood as statistical probability of likely rainfall in the locality generally.
As to his calculation of runoff, Prof Van Dijk was limited in his instructions to merely consider the harvest point near Dam 1. Having regard to the evidence of Mr Purcell that was not challenged in cross-examination, there are four other locations from which runoff can be harvested. Prof Van Dijk has not considered these areas, and therefore his calculation of runoff cannot be accepted as the total amount of available runoff in any of the charge periods.
As to the area of runoff that Prof Van Dijk did calculate he did not have the benefit of physically viewing the area, rather, he relied upon satellite imagery that he accepted may have changed between the capture of the image and the charge periods. Mr Purcell had visited and physically viewed the area. In the circumstances I find his evidence, to the extent that it differs from Prof Van Dijk with respect to the physical characteristics of this area must be preferred.
The calculation of the IL is capable of having a significant impact on the quantity of runoff calculated. Prof Van Dijk and Mr Purcell utilised different methods to determine the IL. In substance, however, the difference appeared to relate primarily to the effect of the classification of soil type particularly having regard to the use to which the land was being put, and that Prof Van Dijk considered each day rain fell as a separate "rain event" and as such deducted the IL on each occasion. The capacity of soil to recharge (the IL rate) between days of rain will depend upon the volume of rain and the period over which it fell. I consider that recommencing the IL on each day is artificial and may render the calculation unreliable. The ability of the water to runoff and the extent to which it will be absorbed into the soil profile will depend upon its treatment. As Prof Van Dijk accepted compacted soil will produce greater runoff than uncompacted soil. To that extent, Prof Van Dijk's analysis from satellite imagery does not permit the type of analysis at the level of detail that would produce a confident calculation of runoff. Even on the area of runoff calculated by Prof Van Dijk his calculation is to be treated with caution.

[28]

Mr Robert Day

Mr Robert Day is a special analyst for the Prosecutor, he prepared a written expert report (Exhibit M). In that report he attested to the process he undertook to calculate the volume of water in each of the four relevant water storage dams at nominated dates within each water year.
Mr Day provided in tabular form the water volume calculated by him by reference to what he described as the: high plausible volume; digital estimate volume; and over conservative volume. It was contended by him that for the water storages excluding Dam 2 the digital estimate volume was, in his opinion, the most likely volume of water at each date, however, it was plausible that the actual volume in each Dam would be within the range expressed by the high plausible to low plausible (over conservative) volumes.
As to Dam 2, it had been altered since the base date data (as explained below) and accordingly, he considered that the full range of calculations from high plausible to over conservative should be used to accommodate for this variation.
Mr Day utilised satellite imagery together with Geographic Information System (GIS) software to calculate the surface area of the water in each storage area at identified points in time. Utilising that surface area Mr Day mapped those figures against what was identified as "storage volume curve" (SVC). The SVCs were established as part of the NSW Healthy Floodplains Project.
Mr Day attested that the SVCs were specific to the dams themselves and were not generic approximations. The SVCs were developed from LiDAR data to which a Standard Bathymetry Model (SBM) was applied in order to account for water in the base of the dams at the time of the LiDAR data-capture.
In cross-examination Mr Day accepted: 1. He had not visited the site; 2. The LiDAR information he relied upon was captured in 2007; 3. The SVC that was applied to the subject water storage dams on the Property was not developed by him but taken from a "library" of SVCs developed by someone other than him, but those SVCs were applied by him in his calculations. Whilst not a "generic" SVC in that they could be applied to any water storage dam they were SVCs "specifically" designed for the subject water storage dams; 4. He had not seen the calculations underlying the SVC; 5. The calculation of the SVC relies upon assumptions that: 1. all of the dams are built in a similar way; 2. the height of the dam wall corresponds to the width of the wall; 3. the wall batters are similar in that the volume of the wall corresponds to the volume of the borrow pit excavation of the base of the wall; and 1. To the extent that the construction water storage dam does not correspond with the assumptions above the assumptions that underly the calculation would be inapplicable.

[29]

Mr James Purcell

Mr James Purcell provided evidence as to the assessment of storage capacity of each of the water storage dams. In doing so he relied upon survey data prepared for each dam: 7.0 Dam Storage Curves - Havana North 7.1 A storage curve for any dam or ring tank is the relationship which links the water depth, the volume stored at any depth and the surface area of the water surface at any depth. 7.2 These relationships are determined by survey using ground, GPS, LiDAR (aerial) and/or bathymetric (below water) survey methods. 7.3 As part of the Floodplain Harvesting Project run by the NSW Department of Planning and Environment, all storages or dams shall have to have a storage curve survey completed which is certified by a registered surveyor. 7.4 Attachment 10 contains the storage curve data for: • Dam 1; • Dam 2; • Dam 3; and • House Dam (Dam 4). These storage curves are based on detailed surveys completed by SKM Consultants of Moree on 29th August 2022 and certified by Registered Surveyor, Paul Covell. 7.5 For each dam the volume in megalitres (ML) and surface area in square metres is given for every 0.1m in water depth from empty to top of bank. 7.6 It is not possible to store water at the top of bank level because of wave action caused by wind and inconsistent top of bank levels. The top water level is therefore some distance below the top of bank to prevent waves in a storm breaking over the top of the bank and to provide and a margin to allow for slight low spots in the finished bank. 7.7 The vertical distance from the average top of bank level to the top water level is called "freeboard". 7.8 For large storages a freeboard of 1.0 metre is normally allowed. For smaller dams like the Havana North dams a freeboard of 0.6 to 0.8 metres is recommended. Depending on the farmers risk management, higher water levels than the normal top water level can be used for short periods in say a dry season where water capture is paramount. 7.9 In my opinion, based on experience and observation, I would recommend a freeboard for Dams 1, 2 & 3 of 0.7 metres and 0.5 metres for the House Dam (Dam 4). 7.10 I would also be aware that at times Havana North will run with 0.3 metres freeboard for short periods. 7.11 Table 7.1 shows the volume for each dam at Havana North with the recommended 0.7 metre and the short-term 0.3 metre freeboard. Table 7.1 Havana North Dam Volumes Dam Freeboard (m) Top Water Level (m) Volume (ML) 1 0.7 182.1 477.2 0.3 182.5 547.5 2 0.7 181.95 322.3 0.3 182.35 362.2 3 0.7 179.65 286.2 0.3 180.05 342.2 4 0.5 180.75 23.4 (House Dam) 0.3 180.95 26.4

[33]

Findings on stored water availability

The volumes of storage capacity in Dams 1 and 3 are largely in agreement between the parties and the volumes as identified by survey in Mr Purcell's report should be accepted as they relate to actual measurement rather than estimates based upon the application of mathematical assumptions. Dam 2 had its maximum capacity using the bathometric curve calculated prior to the enlargement of the Dam. As such, the evidence of survey in Mr Purcell's report is the more accurate determination of maximum volume and should be accepted.
Dam 2 was altered in size. On that basis, Mr Day accepted that his estimates may be unreliable such that he suggested that the high probable figures be utilised in any analysis. The analysis by Mr Day of Dam 2, whilst recognising a degree of unreliability cannot be overcome by adopting another figure that has been calculated absent a proper determination of the capacity of Dam 2. For the purposes of utilising Mr Day's calculations real caution must be exercised in adopting his figures for Dam 2. To that extent the high probable figures are not a reliable proxy for an inaccurate calculation of the actual size of Dam 2 and are at best an educated guess that adopting such a figure may overcome the inadequacy in the original calculation.
Applying the maximum figures has a degree of certainty as they have been calculated by survey and to the extent there is some degree of agreement between the survey and the bathometric analysis, I am able to derive some confidence from the bathometric calculations for the maximum figures. However, the survey also discloses that the features of the water storage dams below the surface level are not uniform. The bathometric analysis assumes a smooth surface and bank construction techniques neither of which are present in the dams at Havana North. Therefore, the digital estimate volume fails to take into account those anomalies. Where the estimates of Mr Day to be utilised in any calculation the high plausible value would be the best measure of an estimate as the anomalies would be, at least to some extent, subsumed in the calculation. However, even on the high estimate level analysis for point in time analysis the calculation is no more than an estimate.

[35]

Prosecutor's submissions

The Prosecutor submitted that I should give a number of directions relating to the determination of this case, which directions (excluding the honest and reasonable mistake of fact direction) were: Onus and standard of proof 1. As this is a criminal trial the burden of proof of the guilt of the defendant is placed squarely on the prosecutor. That burden rests upon the prosecutor in respect of every element or essential fact that makes up the offence charged. That burden never shifts to the defendant. There is no obligation on the defendant to prove any fact or issue that is in dispute. It is not for the defendant to prove its innocence but for the prosecutor to prove its guilt. 2. Proving the defendant's guilt beyond reasonable doubt is the standard of proof the prosecutor must achieve before the defendant can be convicted and the words mean exactly what they say - proof beyond reasonable doubt. 3. The burden of proof on the prosecutor does not mean that the prosecutor must prove beyond reasonable doubt every single fact that is in dispute but the prosecutor must prove the elements of the charge and must prove those elements beyond reasonable doubt. 4. In a criminal trial there is only one ultimate issue that a tribunal of fact has to decide. Has the prosecutor proved the guilt of the defendant beyond reasonable doubt? If the answer is "yes", the appropriate verdict is "guilty". If the answer is "no", the verdict must be "not guilty". Liberato direction 5. The defendant relies on an account of events given to NRAR investigators by its directing mind, Mr Phelps. His account is to the effect that he had no reason to believe that more than 988ML was being taken in any given water year and no reason to believe that the meters were not recording accurately. He received most of his irrigation water from rainfall and associated runoff, was not a record keeper and farmed by "gut feel". He denied that he would have used in the order of 10ML/ha of irrigation water. He denied having ever interfered with the meters. 6. Insofar as the various aspects of Mr Phelps' account involve a denial of elements of the prosecution case then if they are accepted as true, an acquittal must follow. 7. If they are difficult to accept but it is considered that any such aspects might be true, then an acquittal must follow. 8. If the evidence is disbelieved then it should be put to one side and the question will remain: has the prosecutor, upon the basis of evidence that you do accept, proved the accused's guilt beyond reasonable doubt? Circumstantial case 12. Where a prosecution case rests substantially on circumstantial evidence, a tribunal of fact cannot return a guilty verdict unless the prosecutor has excluded all reasonable hypotheses consistent with innocence: The Queen v Baden-Clay (2016) 258 CLR 308 at [46], [50]. For an inference to be reasonable it must rest upon something more than mere conjecture: The Queen v Baden-Clay at [47]. 13. In considering a circumstantial case, all of the circumstances established by the evidence are to be considered and weighed in deciding whether there is an inference consistent with innocence reasonably open on the evidence: The Queen v Hillier (2007) 228 CLR 618 at [46]. The evidence must be considered as a whole and not by a piecemeal approach to each particular circumstance: The Queen v Hillier at [46] Inferences 14. Inferences are conclusions of fact rationally drawn from a combination of proved facts. If A, B and C are established as facts then one might rationally conclude that D is also a fact, even though there might be no direct evidence that D is indeed a fact. Inferences may be valid or invalid, justified or unjustified, correct or incorrect. 15. In a criminal trial where the tribunal of fact must be satisfied of the guilt of the defendant beyond reasonable doubt, the tribunal of fact should be extremely careful about drawing any inference and should examine any possible inference to ensure that it is a justifiable inference. 16. In the context of a criminal trial, an inference should not be drawn from the direct evidence unless it is a rational inference in the circumstances.
Having regard to the nature of the Prosecutor's case, I consider that these directions are appropriate.
The Prosecutor's submissions were that the evidence of Dr Meyer's estimation of water demand should be accepted as a rational inference such that it could be used to establish the essential element of the charges that the Defendant had taken more water than authorised. Once Dr Meyer's calculation is accepted the evidence as to water sources, other than the bores, is applied to the calculation and the deficit determines the amount of bore water that must have been used - there being no other water source to meet that deficit.
To the extent that Dr Meyer's evidence was criticised "the Court may wish to proceed on the conservative approach of adopting the lower bound of Dr Meyer's estimate (that is, the -15% figure)". As to the figures of Prof Van Dijk it was suggested that the higher end of his estimates of rainfall and runoff could also be adopted, and Mr Day's high plausible volume be applied. It was considered that such an approach would overcome any criticism of the evidence and permit a finding most favourable to the Defendant. It was submitted that even on that basis the evidence would disclose a deficit in water that would need to be met by the taking of bore water in excess of the permitted amount.

[36]

Defendant's submissions

Amongst multiple other alternative rational hypotheses consistent with the Defendant's innocence, the following alternative analysis of the water take is a rational hypothesis that has not been disproved by the Prosecution (even without factoring in: 1. The potential for sui generis storms on Havana North or parts of it, or the adjoining catchment arising due to rainfall variability; 2. The potential unknown and unmeasured soil moisture amounts at the beginning of each season; 3. The unknown and unmeasured additional overland flow amounts based on Mr Purcell's 2,000ha catchment, 10 times larger than Prof Van Dijk's 200ha; and 4. The unknown and unmeasured amounts from other water collection points including the north of the Property, and without factoring in the unreliability of Mr Day's figures about which the Prosecutor was given express notice in the ROI but failed to investigate.
Whilst not accepting the appropriateness of Dr Meyer's calculations or the assumptions therein, when the evidence, taken as a whole is considered, there are clear, reasonably plausible figures that may be rationally inputted into Dr Meyer's calculations. These figures are supported on the evidence. Even on this basis, the Prosecutor has not excluded a reasonable hypothesis of innocence. For example, the following calculations demonstrate such a hypothesis is available: 2016/17

320ha
Only 1186ML needed (starting with Dr Meyer's 822mm crop water requirement, based on 20% error rate in crop water usage, Wee Waa George St BOM gauge rainfall (185mm), WFIE of 90%).
Subtract 406ML bore usage gives 780ML needed.
Van Dijk says 242ML in dam rain and 798ML in farm runoff.
Zero deficit even on the amount recorded on the meters, and a surplus if the lawful potential take as pleaded of 988ML is considered. 2017/18
342ha
Only 1989ML needed (starting with Dr Meyer's 857mm crop water requirement, using 20% margin of error, 162mm Wee Waa BOM rainfall, see T330.29-34 and a 90% WFIE figure, giving 5.81ML/ha).
Minus 546ML bore usage.
Van Dijk says 173ML in dam rain and 596ML runoff.
Day says at start of July 2017 around 675ML was in the dams on his high plausible figures.
Zero deficit (a 1ML surplus) even on the amount recorded on the meters, and a surplus if the lawful potential take as pleaded of 988ML is considered. 2018/19
179ha eastern fields
Only 1102ML needed (starting with Dr Meyer's 873mm, using a 20% reduction for uncertainty, minus 144mm at the Wee Waa BOM for this year December - April, and applying a 90% WFIE, gives 6.16ML/ha).
157ha western fields
Only 937ML needed (starting with Dr Meyer's 852mm, applying a 20% uncertainty, minus 144mm for Wee Waa BOM December - April, and applying a 90% WFIE, gives 5.97ML)
Total both eastern and western fields: 2039ML. Minus 573ML bore usage.
Van Dijk says 138ML (CB 84, [36]) on dam rain and 1126ML in farm and off farm runoff (CB 1384 line 484, CB 84, [37], based on a plausible 346 mm in this year, CB 1375).
Day says 149.3ML in his high plausible as at 2 July 2018, CB 1268. Deficit of 52.7ML on the amount recorded on the meter, but a surplus if the lawful potential take as pleaded of 988ML is considered. The 52.7ML is only an 9.1% exceedance of the metered amount and so is explicable by the inaccuracy of the Davies Shephard meters as measured in the Prosecution's Judge report).

Having regard to the evidence as a whole, the Court must entertain a reasonable doubt as to the Defendant's guilt in relation to the Water Take Charges, due to the multiple successive assumptions, each attended by significant uncertainty and margin of error, on which the Prosecution case relies.
Should the Court accept Mr Phelps as a witness of truth and accept his evidence that the meters were not tampered with, then that is a finding of fact directly inconsistent with the Defendant's guilt, as it entails that no more than the amounts as recorded on the meters were taken.
Another route to the same conclusion is a finding that WFIE may have been greater than Dr Meyer's 70%, that the extreme rainfall variability produced more rainfall on Havana North than postulated by Dr Meyer and Prof Van Dijk, and that a contribution by overland flow from an area greater than the 200ha identified by Prof Van Dijk could have supplied any remaining deficiency between Dr Meyer's crop water usage and the metered amounts. There is also uncertainty as regards to the crop water usage figures of Dr Meyer and the high plausible dam capacities of Mr Day based on the 2007 LiDAR data and absent on the ground survey that adds to what must be the reasonable doubt as to guilt.
There are two mandatory requirements for the success of a prosecution based on circumstantial evidence: see Chamberlain: 1. While an inference of guilt can be drawn from a combination of facts none of which viewed alone would support the inference, the tribunal of fact cannot view a fact as the basis of an inference of guilt unless satisfied beyond reasonable doubt as to the existence of that fact; and 2. The inference of guilt must be the only inference that is reasonably open on all the primary facts as found.
The Prosecution case here fails both necessary requirements for the success of a circumstantial evidence prosecution as identified in Chamberlain.
Dr Meyer's methodology is no doubt very useful for agricultural or catchment management purposes at a broad scale. However, considered in the context of the evidence as a whole it lacks the precision or certainty required to prove the present charges to the criminal standard.

[37]

Findings on evidence of guilt on amount of water taken

The Prosecutor's submissions relied upon an acceptance of the amount of crop water and irrigation water demand as determined by Dr Meyer. Whilst I have considered the evidence of Dr Meyer above it now comes to put that evidence into the context of the whole evidence in order to determine whether, in this circumstantial case, an inference can be drawn from the totality of the evidence that establishes beyond reasonable doubt that the Defendant took more water than permitted. In that respect, the Prosecutor must exclude all reasonable hypotheses consistent with innocence.
In light of my findings as to Dr Meyer's evidence, I consider that the Prosecutor has not, on that evidence alone, established a rational inference that the amount of crop water and irrigation water demand as identified by Dr Meyer is a rational or reasonable determination of demand such that the necessary inference can be drawn or that it is sufficient to exclude all reasonable hypotheses consistent with innocence. In taking all other evidence into account the deficiencies in Dr Meyer's evidence are not overcome to the extent that either a rational inference can be drawn as to the amount of crop water or irrigation water demand notwithstanding his evidence or in supplement to his evidence.
The evidence of Prof Van Dijk with its deficiencies in the calculation of stormwater runoff alone renders the assessment of water sources unreliable. His use of seven weather stations, too, does not overcome the inadequacy in the evidence that such an approach is a sufficient proxy for actual evidence of rainfall in circumstances of evidence of rain variability over this landscape. Therefore, in combination with the evidence of Dr Meyer, the deficiencies in Dr Meyer's evidence identified by me are not overcome and may in fact be compounded if Prof Van Dijk's figures are factored into Dr Meyer's analysis.
The evidence of Mr Day, absent a necessary determination of crop water demand cannot stand alone, this evidence only operates in a framework provided by Dr Meyer's evidence. If Dr Meyer's evidence is insufficient to permit an inference of water demand (as I have found) then Mr Day's evidence cannot add any weight or force to the drawing of the necessary inference.
Further, the examples formulated by the Defendant above at [174], demonstrate that even on an acceptance of the fundamental underlying evidence of Dr Meyer's and Prof Van Dijk's adopting variations to their figures which were accepted by them as being open (albeit not adopted by them) indicates hypotheses consistent with innocence. I do not consider that the evidence of Dr Meyer or Prof Van Dijk is sufficiently reliable to permit such an exercise to be rationally undertaken as such evidence still retains the unfounded assumptions that I have determined should not be made, this exercise demonstrates the range of hypotheses on the evidence that remain unexcluded by the Prosecutor.
The Prosecution case relies upon the accuracy of Dr Meyer's estimate of the water needs of each cotton crop in each water year. Whilst the Prosecutor was prepared to accept that "the Court may wish to proceed on the conservative approach of adopting the lower bound of Dr Meyer's estimate (that is, the -15% figure)" it does not submit that any other variation to Dr Meyer's estimates are appropriate. For the reasons outlined above, I do not accept this approach on the evidence. Rather than Dr Meyer's evidence providing a reliable figure of the water needs of the crop his evidence disclosed that: 1. His estimate of 15% error rate was based upon no statistical or other basis, merely an estimation based on "experience" which experience was not identified in any meaningful sense when it came to determining the accuracy of his capacity to estimate crop water usage; 2. The WFIE derived by him was based solely on the adoption of the figure from the Roth Paper which estimate was not intended to be used as such a benchmark. Absent having an understanding of the nature and manner of the farm in practice, such an adoption was unreliable and was likely to be higher if trends experienced in the cotton industry generally had also been adopted on the Property; and 3. The use of rainfall data from weather station data where the evidence is that there is localised variability, and no evidence was adduced to support the selection or adoption of a single or multiple weather stations as an appropriate proxy for actual weather data from Havana North.
If Dr Meyer's estimate of water usage does not establish water demand beyond reasonable doubt the Prosecutor accepted that on its case such would be sufficient to dismiss the charge and that the evidence of Prof Van Dijk and Mr Day, and any of the other evidence would not be sufficient to prove the charge absent the determination of a water usage requirement.
Even if Dr Meyer's figures could be utilised in some way to determine a calculation of water usage not undertaken by him but mathematically available such would be insufficient to establish that the water taken exceeded the licenced amount. Prof Van Dijk, in accordance with the instructions given to him only considered rainwater harvesting to the south-east of the Property. Both the statement of Mr Phelps in his ROI and the evidence of Mr Purcell is that the Property had been developed to utilise rainwater harvesting from all of the property boundaries and contrary to the instructions given to Prof Van Dijk there are five formalised rainwater harvesting points along the boundaries. These areas have not been considered in the Prosecutor's evidence nor the impact on water availability to the Property determined to any meaningful extent. Absent such water availability being considered and determined the available rainwater to the crop cannot be established beyond reasonable doubt. Absent such satisfaction a figure for water from the bore that is required to meet the crop demand cannot be determined to the relevant satisfaction, and therefore no determination, by inference or otherwise, can be made as to the amount that would be required to be taken from the bore.
Moreover, the evidence of Mr Phelps as to his water usage and harvesting practices have not been demonstrated to be wrong or not to be accepted. The reference to water harvesting is consistent with the evidence of Mr Purcell of water harvesting locations. Further, his estimate of water usage is consistent with industry averages and is, as demonstrated by the Defendant's calculations above broadly consistent with the alternative water demand and usage calculations. I find his evidence plausible.
As identified in the directions formulated by the Prosecutor, if Mr Phelps evidence or some aspects of his evidence as to his water usage may be true then the charges must be dismissed.
For those reasons, I find that the Prosecutor has not established beyond reasonable doubt that the Defendant took water in excess of its licence limit in any of the relevant water years. Accordingly, those charges must be dismissed. As the essential element is necessary to be established in both the primary and the alternate charges, the offence on each basis cannot succeed.
As the Prosecutor has failed to establish a fundamental element of the charge and I have determined that the charges are to be dismissed it is not necessary for me to determine whether the Defendant held the necessary intention to warrant a finding of guilt, nor is it necessary for me to determine the defence raised by the Defendant pursuant to s 60F of the WM Act.

[40]

Mr Andrew Judge

Mr Andrew Judge is an approved signatory for the National Association of Testing Authorities and National Measurement Institute. He is experienced in hydraulic and metering testing.
Mr Judge prepared a written statement of evidence (Exhibit L) relating to his opinions.
The Diesel Meter, the Electric Meter and the House Meter were delivered to Mr Judge's laboratory for testing. Mr Judge undertook a visual examination of the three meters and reported: 1. All three units felt rough when turning the impeller implying some of the bearings in the units may be worn or used; 2. The meter seals on MPA 9953/250 (the Diesel Meter) were in place but were "damaged and easily removed"; 3. The meter seals on MPA 11274/250 (the Electric Meter) appeared undamaged; 4. The meter seals on MPA 11356/250 (the House Meter) were missing; and 5. All three meters had tapings fitted with bolts directly over each meter's impeller. The bolts used in each of the units appeared to be different and could be easily removed. There were rust marks on the lower part of the meter flange, indicating exposure to water through the tapings. These tapings and bolts do not appear in the supplier's standard documentation found for this type of meter.
Mr Judge noted that the meter body, stem and propeller, with base plate of each unit was delivered to him. He was advised that field officers were unable to remove the installed meters pipes as the meters were welded to long steel pipes. His facility manufactured a replicated pipe (PVC) section based on available specifications to enable testing.
The testing of each meter was conducted using a "flying start" method which involved setting the flow rate when the rig and the equipment under test (EUT) are stable, at which point the readings for the test are started. Test flow rates were chosen by him to cover the potential flow range nominated for the EUT.
The counter on each meter was read and photographed via a time piece (a digital watch) synchronized to the logging system. The video and photographs of each testing of each meter was tendered in evidence.
The accuracy testing undertaken by him provided summary results as depicted in Table 2 (Exhibit L) as follows: 4.1 Accuracy testing The accuracy testing results are present in Appendix A. A summary of these results are presented in Table 2. Meter Time Reference Volume (I) Flow Rate (l/s) EUT Volume (l) Error Uncertainty K MPA 9953/250 No Flow recorded on EUT 00:38:42 179207 77.2 159500 -11.00% 0.34% 2.0 00:31:11 323991 173.2 288000 -11.11% 0.39% 2.0 MPA 11274/250 00:49:10 59070 20.0 52000 -11.97% 0.38% 2.0 00:36:40 164435 74.7 149500 -9.08% 0.34% 2.0 00:37:22 407830 181.9 367000 -10.01% 0.33% 2.0 MPA 11356/250 00:31:19 50024 26.6 37000 -26.04% 0.38% 2.0 00:30.21 159070 87.4 122000 -23.30% 0.35% 2.0 00:19:09 208655 181.6 192000 -7.98% 0.40% 2.0

[41]

In response to specific questions posed to him by the Prosecutor, Mr Judge's responses were: 4 Further clarifications Further clarifications regarding the Crown Solicitors correspondence (Annexure 2) is provided below. Q1 What observations did you make of the state of repair of each of the 3 Davies Shepard meters? All three units where showing their age, but they appeared to be sealed with no leakage. Q2 Did you observe any signs of recent maintenance on any of the 3 Davies Shephard meters? The meters did not show evidence of maintenance, but with this unit it would be hard to be sure. Q2a Would you expect signs of maintenance to be observable generally? It would be hard to see evidence of maintenance on these meters. Q3 Based on your experience, is the observed state of each of the meters consistent with ordinary wear and tear over time? The ware (sic) on these meters appears consistent with their age. Q4 What might the damaged seals on meter MPA 9953/250 indicate? Appears as though this maybe corrosion and may be a function of age. Q5 What might the undamaged seals on MPA 1t274 indicate? The meter has not been opened, or if it has been opened the seal re-fitted Q6 What might the missing seal's on MPA 11356 indicate? Either the seal has corroded or, it may have been removed to open the meter for maintenance such as replacing the gasket between the body of the meter and counter unit. As you noted in your previous Report, each of the 3 Davies Shephard meters had a bolt located beneath the counter (see photos at Appendix B to that report). Based on your experience and expertise: a) Would that bolt serve any functional purpose? Q7 b) Could the bolt and/or the opening be used to impede the operation of the meter? c) If the answer to 7(b) is yes: I. How would this be done? II. Would you expect this to leave any observable signs (for example damage to the impeller)? a) Yes. They do not appear in the meter documentation we have for these units, and the only practical function would be for inspection inside the pipeline, though the technology to inspect through such as small tapping was not readily available at the time of the meter manufacture. b) Yes. c) Remove the current bolt and install a longer bolt which would prevent the impeller from spinning. If this is done when the meter is at rest, there should be no mark left on the impeller. Q8 Could the opening on each of the 3 meters have been used for the insertion of a grease nipple connected (internal to the pipe) to a copper tube in order to lubricate the bearings? a) If the answer to that question is yes, would you expect that installation (assuming it has since been removed) to have left any observable sign? No, only one meter had a grease nipple in the unit and inserting grease in the current tapping in the unit would just send grease into the water inside the pipe.

[43]

Mr Judge was cross-examined upon his evidence. During the course of cross-examination, the video evidence of the testing procedure was played to the Court.
Mr Judge agreed that contrary to his report that identified that the Diesel Meter showed a zero-flow rate the video disclosed that the meter in fact registered flow during testing. Mr Judge was cross-examined as follows (Tcpt, 7 October 2022, p 242(10)-(24)): Q. Summing up the amounts - just staying with court book 522 - although there was a flow of 5,021 that you record for that first test of the meter on the diesel pump, you don't have anything in the EUT volume column; the reason why you haven't inserted anything there is based on what you are saying to be the case is that there was no reading or turnover on the dial. A. It was very slow. We weren't considering it. Q. Mr Judge, you accept then there was some turnover on the dial? A. There may have - there may have been a slight amount. Q. So the figure there shouldn't be zero, should it? A. Probably not zero. Be close to. I'm not exactly-- Q. So you accept your report in that respect isn't entirely accurate? A. Not 100% on that one either.
The reference in Table 2 above includes a range of error in the metering. That error range for each pump could be caused by mechanical features such as a lack of lubrication or the age of the bearings surrounding the impeller or fair wear and tear from the old age of the pumps.
The pumps were tendered in evidence and physically examined by Mr Judge in cross-examination. He accepted that all three pumps identified the presence of a grease nipple, two remained in existence and one had been removed or damaged. He accepted that his response to question 8 above was incorrect.
The Prosecutor adduced affidavit evidence from Mr Wayne Nott, Customer Field Officer for WaterNSW. He gave evidence upon which he was cross-examined relating to his reading of the meters at the Property and the evidence of the volume of water that the meters indicated at each relevant date. He also gave evidence of the occasions he attended with investigators relating to the investigation of the matters for which the Defendant had been charged.
The Prosecutor adduced affidavit evidence from Ms Leann Davidson, Team Leader Investigations of the Prosecutor. She gave evidence generally in connection with the investigation of the Defendant together with evidence, relating to inspections she had undertaken of the Property and the pumps.
The Prosecutor adduced affidavit evidence from Mr Charles Moss, Manager-Water Information at the Department of Planning Industry and Environment. He gave evidence relating to the accounting of water usage and an explanation of water account documentation.
The Prosecutor adduced affidavit evidence from Ms Helen Singelton, who was employed as a Senior Investigator contracted to the Prosecutor for a period concurrent with the investigation of the Defendant. She gave evidence relating to her attendance upon the Property in connection with the investigation of the matters for which the Defendant has been charged.
The Prosecutor adduced affidavit evidence from Mr Scott Mathieson, Senior Investigator with the Water Enforcement Taskforce. He gave evidence, upon which he was cross-examined relating to his investigation of the offences with which the Defendant was charged. On 25 June 2019, Mr Mathieson was in attendance at Manly Hydraulics Laboratory (MHL) when the testing of the pumps was undertaken by Mr Judge. Mr Mathieson produced a large number of documentary materials including a transcript of the ROI with Mr Phelps who attended as a representative of the Defendant.
The Prosecutor adduced affidavit evidence from Mr Andrew Mannall, Lead Investigator for the Prosecutor. He gave evidence, upon which he was cross-examined relating to inspections he had undertaken of the Property and the pumps. He also assisted in the removal of the pumps from the Property, stored and later transported the pumps to MHL. On 25 June 2019, Mr Mannall was in attendance at MHL when the testing of the pumps was undertaken by Mr Judge. Mr Mannall also conducted the ROI with Mr Phelps who attended on behalf of the Defendant.
As part of the evidence of the investigators video evidence of the removal of the pumps from Havana North and the installation of the pumps at Mr Judge's laboratory were tendered and viewed. The video evidence disclosed that there was a degree of physical force used as disclosed at: 1. On-site removal: 1. Diesel Pump - Tab 20 of Exhibit SM-1 at 00:00 (hammer and screwdriver used to remove pump); 2. Electric Pump - Tab 20A of Exhibit SM-1 at 02:41 (hammer and screwdriver used to remove pump); and 3. House Pump - Tab 23 of Exhibit SM-1 at 02:08 (hammer and screwdriver used to remove pump); 1. In laboratory: 1. Electric Pump - Tab 25 of Exhibit SM-1 video MAH00022 at 02:19-02:36 (removing locator screw with metal object); and 2. House Pump - Tab 25 of Exhibit SM-1 video MAH00022 at 02:30:09-02:30:18 (removing locator screw with metal object).
In addition, the Prosecutor relied upon the circumstantial evidence referred to in connection with the Water Take Charges for an inference that the amount of water taken so far exceeded the allocation of water that the pumps must not have been operating properly and a failure to ensure proper operation.

[45]

Prosecutor's submissions on the Metering Charges

The metering charges are pleaded in the alternative. Should the Court find the charge under s 91l(2) of the WM Act proven then the alternative charge pursuant to s 91H(2) need not be considered.
Insofar as the result for the Diesel Pump at low-flow (26L/s) is concerned, the Prosecutor does not invite the Court to accept it as an accurate representation of the performance of that meter under testing. The time under testing can be seen to have only been 3 minutes and 10 seconds. By contrast, the remaining eight tests were conducted over periods ranging from between 19:09 minutes and 49:10 minutes. Given the low-flow rate of 26L/s, the total volume that would have passed through the meter over 3 minutes and 10 seconds would only have been in the order of 4,940 litres. Noting the scale of the meter, that would only have been expected to register a small increment.
It is apparent from viewing video footage of the test that there was at least some level of movement on the final digit of the meter. Mr Judge conceded as much. In those circumstances, the low-flow result for the Diesel Pump cannot be accepted as reliable and should be put to one side. While Mr Judge stated that it would "probably" have registered a similar error margin to the medium and high-flow rate testing, that opinion did not derive from any testing done that Mr Judge actually did and was simply an assumption on his part that may or may not be accurate.
The balance of the tests results may, however, be accepted as accurate. It was undertaken at an accredited laboratory with carefully calibrated testing conditions and performed over a sufficient period of time as to produce reliable results. In the Prosecutor's submission the Court would act on those figures as representing the extent of inaccuracy in the meters as at the date of testing and, at least in proximate terms, over the period alleged for the purposes of the ss 91H and 91I charges of the WM Act.
The discrepancy between the overall increment on the meters whilst at the laboratory and the figures recorded as incremented on the meters whilst under test conditions (that is, when the flow was stabilised at a known rate) is readily explained by the fact that additional water was pumped through the meters whilst they were not under test conditions. The volume of that water was, according to Mr Judge, not able to be accurately quantified. Any exercise of attempting to compare the test results to the overall increment of the meter is therefore not a valid one.
Whilst the Court would not act on the low-flow test result for the Diesel Pump at a low-flow rate, that test is clearly severable from the balance of the testing which was conducted under appropriate conditions.
The specifications for Davies Shephard meters indicate an accuracy of ±3% across varying flow rates. The tests conducted on these meters indicate that they fell well outside the factory accuracy rate of ±3%, up to: 11.1% (Diesel Pump), 11.97% (Electric Pump) and 26.04% (House Pump).
The element of s 91I(2) of the WM Act that the Prosecution must satisfy beyond reasonable doubt is that the metering equipment, the subject of each of the charges, was "not operating or not operating properly". The fact that "not operating" and "not operating properly" are both avenues of liability indicates that "not operating properly" must have work to do: the section does not only contemplate a meter that is not recording water take at all.
The section says nothing about "fair wear and tear" being an acceptable justification for a metering device "not operating properly", it simply provides for liability in circumstances where water is taken whilst metering equipment is "not operating properly". The question that arises for the Court is: when is a water meter "not operating properly"? In the Prosecutor's submission the answer to that question is: when it is not fulfilling its intended function of accurately measuring flow.
It can reasonably be expected that there will be a level of tolerance in a meter. A legislative intent to provide for liability in circumstances where a meter is operating within its factory tolerance would not, absent a contrary indication, be inferred. Here the factory figure was ±3%. These meters were well outside that tolerance. Wherever the precise point arises, in the Prosecutor's submission an error rate of greater than 10% clearly leads to a conclusion that a meter is "not operating properly".
It is necessary to consider some context. These are meters installed in connection with "metered works" under the WM Act. "Metering equipment" is defined to include "any device used for or in connection with measuring the flow of water and any ancillary wiring, pipework, telemetry equipment or apparatus and any supporting structure". These are meters used for the purpose of measuring the flow of water taken from a carefully calibrated system under the WM Act, which involves available water determinations under Pt 2 of the WM Act and, in this case, the provisions of the Water Sharing Plan for the Upper and Lower Namoi River Groundwater Sources 2003 in each water year. "Operating properly" needs to be understood in that context.
Any suggestion that 26% could be regarded as falling within the bounds of proper operation simply could not be accepted. In the Prosecutor's submission, error rates in excess of 10% fall outside the bounds of the "proper operation" of a meter.
In terms of s 91H(2) of the WM Act, conducting maintenance generally is not an answer. There would be no utility, to use a crude example, in changing the tyre on a car for example if the air filter was blocked. It would do nothing to address the problem. The relevant element of the charges is "failing to ensure the proper operation" of the meters. The Prosecutor submits that the Court would be satisfied to the requisite standard that, whatever may have done to inspect any of the meters (and in the case one meter, take the meter for repairs in 2019), that did not ensure the proper operation of the meters. Indeed, various statements made by Mr Phelps himself establish that the Defendant did not carry out regular maintenance on the meters: 1. Mr Phelps' comments as to his lack of a maintenance regime for the meters; 2. Mr Phelps' statement to the effect that maintenance is only performed when it's brought to his attention; and 3. Mr Phelps' comments to the effect that he has never done anything to maintain the meters unless directed to.

[46]

Defendant's submissions on the Metering Charges

It is an offence against s 91H(2) of the WM Act that the Defendant failed to ensure the proper operation of the relevant metering equipment.
The Prosecution case is in part based on the evidence of Mr Judge, that there is an error rate associated with each meter, and also a general suggestion that because more than 988ML were taken in each water year from 2016 to 2019, the meters must not have been properly operational and the Defendant must not have ensured they were operating properly from January to April 2019 and January to June 2019. Mr Judge's evidence is the only evidence before the Court showing an error rate in the meters.
The Defendant says that such an error rate arose from fair wear and tear and the nature and operation of these old meters and meters of this type, that the meters were operating during the relevant periods and that there was no failure to maintain. This was also Mr Judge's evidence, that the error rates he identified could be due to fair wear and tear.
The Prosecution needs to establish that a standard was breached which means that the meters were not operating properly, as they were clearly operating in the charge period (water was measured as being taken).
The Defendant says s 91I of the WM Act charge is inherently uncertain and inadequately particularised, as what amounts to proper operation is unclear on the face of the charge and has not otherwise been clarified in the Prosecution case at trial.
The Prosecution has not proved its case even if the percentage error rate during certain tests at certain selected flow rates in Mr Judge's written evidence are accepted - why a 7% or 11% error constitutes a failure to properly operate is not explained. This is not perfect operation, but the offence provision does not use that language. The Defendant says that even if those error rates are accepted, the charge has not been proven as what is shown is substantially effective operation, with a margin of error or accuracy.
There is no evidence capable of proving that the meters were not in situ during the charge periods January to April 2019 and January to June 2019. Mr Nott confirms this by inspecting the meters in January and April 2019. This matter can be accepted by the Court.
To the extent that the Prosecutor relies on Dr Meyer's crop water estimates for the Metering Charges, there is unresolved internal inconsistency in its case between the amounts posited by that expert evidence and the recorded error rates of Mr Judge.
Due to this internal inconsistency and conflict of evidence, the Prosecutor has not proved to the criminal standard that the error rates identified by Mr Judge at various flow rates after the removal of the meters applied for the take that occurred between January to June 2019, and nor has any greater error rate been proved to the criminal standard of proof.
In particular, Dr Meyer does not provide evidence directed towards the charge periods of these offences. His expert evidence does not engage with the matter that the Prosecution is required to prove during the charge period. Any assertion by the Prosecutor that the meters during these identified charge periods were recording at error rates above those in the Judge report is the subject of a reasonable doubt by reason of the Prosecution's own evidence in that report. Even if the Court were to accept Dr Meyer's evidence as to the irrigation water requirement on Havana North over the various nominal growing seasons selected by him, that is not capable of establishing that these meters under recorded in the January to April or June 2019 charge periods, particularly given the significant variability in error rates identified by Mr Judge.

[49]

Evidence does not prove meters were not operating properly or a failure to ensure proper operation

The specifications of the Davies Shephard meters do not provide an assurance of 100% accuracy. To the contrary, they specifically address various causes of inaccuracy and identify a 3% level of inaccuracy by design and only within certain flow velocity conditions. Further, the specifications specifically advise that meter accuracy is related to water velocity being dependent on the ratio of bearing and gear friction at lower velocities and at high velocities by wear of bearings and maintenance expenditure, with velocities over 2.5mps not being tolerated. However, in the absence of any evidence as to the flow velocities at the time water was extracted in the charge period, or as to the ratio of bearing and gear friction, the level of accuracy of the meters in actual use during the charge period is unknown and as a result not established beyond a reasonable doubt on the evidence before the Court. There is in fact evidence in the ROI that flow rates were extremely variable on the farm, which are unknown for the January to April 2019 and January to June 2019 charge period.
Mr Judge confirmed that it would have been preferable to test the meters in situ. He also confirmed that he did not know what the water pressure during use was at Havana North.
On Mr Judge's figures, disregarding the flawed first diesel pump meter test (that Mr Judge conceded probably had an error rate the same as the other two tests), the average error rates in his Table 2 are 11.05%, 10.35%, and 19.1%.
The Defendant's technical expert Mr Mann gives evidence that the Davies Shephard meters were subject to fair wear and tear issues due to grit chewing out the nylon impeller bearings.
Further, the House Pump meter, site 17636 Lic No 142154 (proceedings 2020/88368) was also interfered with and potentially damaged by the Prosecutor's investigations prior to the charge period, during the inspection of August 2018: see the second affidavit of Wayne Nott and the affidavit of Leann Davidson documenting the force used to dismantle the meter and its reassembly, all prior to the charge period.
Mr Mann gave evidence that the magnets in the meters may be damaged by being struck, and the Davies Shephard leaflets show the complexity of the mechanism inside the capsule.
In relation to the s 91H of the WM Act charge, there is also objective evidence of maintenance by drilling grease holes, installation of a new impeller in one meter with the purchase price still marked on it and the oral evidence of Mr Mann that one of the meters had undergone a complete element change.
Reasonable doubt has not been eliminated in regard to the Metering Charges. There is no evidence of the level of inaccuracy of the meters when in situ, the flow rates used when the meters were in situ during the charge periods and evidence that the flow rates may affect accuracy.
There exists a reasonable doubt as to whether the meters were "not operating" or "not operating properly". Having regard to Mr Judge's evidence and expert report that at its highest the meter error rate is 26% on one reading, and otherwise below a 20% error rate taking or an average of the three tests, for the other meters, around 10%. The meters have been in place for a significant amount of time and Mr Judge's evidence is that this arose from fair wear and tear. Mr Mann's evidence strongly supported this opinion and explained that the Davies Shephard meters although robust in some respects were affected by wear of the nylon bushes on the impeller and could be compromised with the farmer having no way of detecting this without dismantling the meter entirely. They became less accurate over time, and this had nothing to do with poor maintenance.
The term "proper operation" requires examination of the purpose of the machine's operation: Ozbinay v Crowley (1993) 17 MVR 176 at 5 (per Byrne J). The purpose is to measure the volume of water taken from the bores for dual reasons. For the Defendant it was to give a measure of the water taken for farm management purposes. They perform this function still. The meters were not put in pursuant to the WM Act, although it is now agreed that they were installed at all relevant times within the meaning of that term in the WM Act. They also assist in enabling the new regulator, WaterNSW assess the amount of water taken and whether a farmer is within their water allocation. So, the meters provide information as to water volume to both the farmer and the regulator, and always have done.
There is no suggestion during the charge period from January to June 2019 that the amount of water taken (even at Mr Judge's error rates of a maximum of 10%-20% minus 3.3% for uncertainty and the "guaranteed" error rate during certain flow velocity ranges) prevented the achievement of the above purpose or precluded the meters performing these functions. The evidence in the ROI is that Mr Phelps was able to rely on the meters, and so did WaterNSW.
The Macquarie Dictionary definition cited by the Prosecutor at par 31 of the Prosecutor's closing submissions (PCS) assists in that inherent in the ordinary meaning of "properly" and "proper" is a reference to appropriate to the circumstances or conforming to established standards. The circumstances are that these are meters that had been installed for many years to measure the volume drawn up from each bore, but never with any particular standard of accuracy being required or imposed. The Prosecutor's case now depends on the Court finding than an occasional 8-23% error rate in addition to that guaranteed by the manufacturer puts the meters outside the definition during a four month period. There is no evidence of any issue having been taken with the accuracy of the meters by the expert Customer Field Officers who read the meters (such as Mr Nott), and no test was demanded or standard of accuracy specified in the decades before the current charge period of January to June 2019.

[50]

The error rates vary as between the meters

From 29 January to 29 April 2019 the records in the affidavit of Charles Moss suggest that the amount of water used was 55ML for the House Pump, 32ML for the Diesel Pump, and 68ML for the Electric Pump. From 29 January to 4 June 2019, the records in the affidavit of Charles Moss show that the amount of water used was the same, the meters being removed on the last day of the charge period. At a 19.1% error rate (the average error rate found for the House Pump meter), this amounts to a maximum discrepancy of 10.45ML on the 55ML taken through the House Pump during the charge period. However, as the flow rates during the charge period is unknown, even on Mr Judge's figures the applicable error rate could have been as low as 7.98% or factoring in the 3% error rate and 0.4% uncertainty, an unexplained discrepancy of only 4.58% for the House Pump. Over the 55ML drawn from the House Pump during the charge period, this amounts to only 2.5ML out of 55ML. This is 0.25% of the yearly water allocation of 988ML. In relation to the meter on the Diesel Pump, Mr Judge's average error rate is 11.055%, which minus 3% and 0.38% uncertainty gives 7.6%. Over 32ML taken from the Diesel Pump, this gives 2.4ML out of 32ML, which is 0.24% of the yearly water entitlement. For the Electric Pump, the lowest error rate is 9.08%, which minus 3% and 0.34% uncertainty gives 5.74%. Proportionately to the 68ML taken by the Electric Pump in the charge period, this gives 3.9ML which is 0.39% of the yearly water entitlement.
Proportionately minor discrepancies like this have not been proved to be material by the Prosecutor, and the Prosecutor's case in relation to the Metering Charges is put on the basis that the discrepancies are material: see par 31 of the PCS.
These proportionately error rates do not fall within the ordinary meaning of a deficiency that deprives the meters of "proper operation" as the function or purpose of keeping an adequate record of compliance or not with the water allocation over the charge period is not nullified. They are not material discrepancies and have not been shown to be so by any test, standard or other criteria. The mere assertion of materiality in the Prosecution case exposes this gap. No evidence was called by any expert or officer explaining why these discrepancies are material and no standard was tendered establishing this.
Further, there are doubts as to whether Mr Judge's methodology calculated the actual error rates or whether they were lower because of unquantified additional amounts of water that are shown on the meter dials in the photographs but not in Mr Judge's Table 2. He provided an explanation for this, that the additional amounts were somehow outside the actual start and finish of the test. In the absence of evidence as to how much water was not recorded in the tests which passed through these periods this is only an explanation in concept or aspiration as Mr Judge explained, there is no record of those amounts that is satisfactory, and so a reasonable doubt arises as to whether or not, in addition to the 3% error rate of the MHL testing meter and Mr Judge's 0.3% or thereabout uncertainty figure in Table 2, his stated error rates are in fact higher than the actual error in any of the meters. This reasonable doubt has not been dispelled in the evidence.
The affidavit of Stuart Mann provides evidence at par 51 in relation to the meter on the Diesel Pump that that on physical inspection, even when totally removed, it showed no external sign that it would under record water flow. The same evidence is given in relation to the meter on the Electric Pump (at par 47), and the meter on the House Pump (at par 39).
A reasonable doubt as to whether the error rates identified in the Judge report were those of the meters when in situ is created by the second affidavit of Stuart Mann dated 10 October 2022 which gives evidence that when the meters or flanges are "impacted" the internal magnet can be affected: see par 10. Mr Mann explained that he would be speculating and did not know whether there was any adverse affectation as regards to the removal of these meters. However, it is plain from his affidavits that that was not the task he was instructed to undertake. His opinion evidence as to the sensitivity of the meters to "impact" is unaffected by this. It survives and gives rise to the reasonable possibility of such adverse affectation on the meters has not been excluded to the criminal standard by the Prosecutor. Clearly, Mr Mann could not have gone any further than he did with his expression of opinion as it must forever remain unknown whether or not the potential partial disengagement of the spindle from the magnet has occurred to affect the operation of these meters as they were (a) not opened up to expose the magnet at MHL after removal; and (b) not opened up and examined before their removal and impact on site. Unless both things had occurred, it cannot be known whether the impacts of 4 June 2019 caused a physical disconnection of the magnet from the spindle. Mr Mann's evidence identifies a reasonable doubt as to that having occurred and that has not been closed off on the Prosecution evidence, as there is no evidence of such before and after examination.
Further, for the ss 91I and 91H of the WM Act charges the absence of proper operation needs to be proved within the specified charge period in January to April or June 2019. That has not been done. There is no evidence as to the error rate that applied at that time and during the extraction that occurred over that period. If the Judge error rate is applied it is not capable of proving the offence, and there is a reasonable doubt as to whether that measured error rate did apply due to the violent process involved in removing the meters.

[52]

Honest and reasonable mistake

The evidence of Mr Phelps is relied upon to establish an honest and reasonable mistake of fact as to the operation and proper operation of the meters.
Mr Phelps in the ROI explained that he went on what the meters told him and checked them regularly to see they were working. There was no indication they were not operating properly. This was reinforced by Mr Nott's oral evidence that he never detected a problem with the meters. He has given oral evidence that from time to time he has taken them to be fixed.
The leading authority on the interpretation of statutory offences remains the High Court's decision in He Kaw Teh v R (1985) 157 CLR 523 (He Kaw Teh). The starting point in construing any statutory offence provision is that "[t]here is a presumption that mens rea, an evil intention, or a knowledge of the wrongfulness of the act, is an essential ingredient in every offence; but that presumption is liable to be displaced either by the words of the statute creating the offence or by the subject-matter with which it deals, and both must be considered": Sherras v De Rutzen (1895) 1 QB 918 at 921, cited by Gibbs CJ in He Kaw Teh at 528. There is nothing in the statute that displaces this presumption of mens rea here.
It is not understood this analysis is disputed by the Prosecutor.
The evidence is that Mr Phelps did not know he was doing anything wrong or that the meters were not operating effectively.
The general principles applied to statutory offence provisions were summarised by Brennan J in He Kaw Teh at 582 as follows: There is a presumption that in every statutory offence, it is implied as an element of the offence that the person who commits the actus reus does the physical act defined in the offence voluntarily and with the intention of doing an act of the defined kind. There is a further presumption in relation to the external elements of a statutory offence that are circumstances attendant on the doing of the physical act involved. It is implied as an element of the offence that, at the time when the person who commits the actus reus does the physical act involved, he either - (a) knows the circumstances which make the doing of that act an offence; or (b) does not believe honestly and on reasonable grounds that the circumstances which are attendant on the doing of that act are such as to make the doing of that act innocent. The state of mind to be implied under (2) is the state of mind which is more consonant with the fulfilment of the purpose of the statute. Prima facie, knowledge is that state of mind. The prosecution bears the onus of proving the elements referred to in (1) and (2) beyond reasonable doubt except in the case of insanity and except where statute otherwise provides. Once it is raised by the evidence, honest and reasonable mistake of fact becomes an element of the offence which the prosecution must negative beyond reasonable doubt as set out in CTM v The Queen (2008) 236 CLR 440 per Gleeson CJ, Gummow, Crennan and Kiefel JJ at [8]. See also Poidevin v Semaan (2013) 85 NSWLR 758 at [14] per Leeming JA, with Ward and Emmett JJA agreeing.
The evidence in the ROI demonstrates that Mr Phelps had no belief at all that he was exceeding his water allocation or that the meters were other than generally reliable and recording the water taken whenever he observed them, and it is evident from the explanation of events given in the ROI that it is fanciful to suggest that he did not have an honest and reasonable belief in the proper operation of the meters because an analysis such as that of Dr Meyer was advanced after the charge period.

[54]

Findings on the Metering Charges

The principle charge in these proceedings is a breach of s 91I(2) of the WM Act. If that charge is not proven then the alternative charge against s 91H(2) falls for consideration. As to the essential elements of the s 91H of the WM Act charge there is no dispute that there was at the relevant date metering equipment installed in connection with a water supply work, the dispute is whether the Defendant failed to ensure its proper operation.
The term "proper operation" is to be understood having regard to the principles of statutory construction. In this case, the Defendant contends that such a term excludes fair wear and tear. The Defendant contends that the considerations set out rely upon the considerations of Pain J in Harris at [96]-[103] as supporting such construction. In that case, her Honour held: 96 The starting point for ascertaining the meaning of a statutory provision is the text of the provision considered in light of its context and purpose, as the plurality said in SZTAL v Minister for Immigration and Border Protection (2017) 262 CLR 262; [2017] HCA 34 at [14]; see also Gageler J (who was in dissent but not on the principles of statutory construction) at [37]-[39]. 97 Provisions are to be given their ordinary and natural (or conventional) meaning that is appropriate having regard to the immediately surrounding words and their grammatical usage: Project Blue Sky v Australian Broadcasting Authority (1998) 194 CLR 355; [1998] HCA 28 at [78] per McHugh, Gummow, Kirby and Hayne JJ; CIC Insurance Ltd v Bankstown Football Club Ltd (1997) 187 CLR 384; [1997] HCA 2 at 408 per Brennan CJ, Dawson, Toohey and Gummow JJ; Interpretation Act 1987 (NSW) s 6. Section 33 of the Interpretation Act provides that a construction that promotes the purpose or object of the statute is to be preferred. 98 The purposes of the WM Act are contained in s 3. It must be acknowledged, however, that a statute may, expressly or impliedly, have a number of different, and potentially competing, purposes: Construction Forestry Mining & Energy Union v Mammoet Australia Pty Ltd (2013) 248 CLR 619; [2013] HCA 36. 99 Another important principle of interpretation, especially when considering a statute regarding a criminal offence, is the principle of legality. As such, it is assumed that, absent clear language, Parliament did not intend to abrogate or curtail citizens' rights, including by the imposition of criminal sanction and if so, only to the extent clearly delineated: Coco v The Queen (1994) 179 CLR 427; [1994] HCA 15 (Coco). 100 In construing such provisions, the convenience in carrying out an object authorised by the legislation is not a ground for eroding fundamental common law rights: Coco at 436 (quoting with approval Plenty v Dillon (1991) 171 CLR 635; [1991] HCA 5 at 654). 101 In the present case, where Parliament purports to prohibit certain conduct and to make it an offence (with a nearly unlimited maximum penalty, currently up to $5,005,000.00 and $264,000.00 per day thereafter), the onus is upon the State to clearly and unambiguously draft the terms upon which such punishment may be imposed. 102 If, applying the principles of construction set out above, there is ambiguity in the section, that ambiguity is to be resolved in favour of the Defendants, or in other words, if there are at least two reasonably open meanings, the Court must give effect to the more lenient one: Beckwith v R (1976) 135 CLR 569; [1976] HCA 55 per Gibbs J at 576. 103 I adopt as correct the Defendants' submissions on the proper approach to statutory construction, particularly in light of the need to construe criminal provisions fairly.
The principles set out by her Honour above are not controversial. However, her Honour was not considering s 91I(2) of the WM Act, and therefore made no findings as to its appropriate construction.
In this case, adopting the principles of construction as outlined by her Honour, that I consider apply equally to the task of construction in this case, I cannot accept the construction contended for by the Defendant.
The text of s 91I(2) is: A person who takes water from a water source to which this Part applies by means of a metered work while its metering equipment is not operating properly or is not operating is guilty of an offence. The term "operating" is used twice, once where the meter is performing and once where it is not. The terms "operating" and "operating properly" are not defined terms in the WM Act. The meaning is therefore to be taken as the ordinary and natural meaning as determined having regard to the context and objects of the relevant Act.
Whilst not determinative both parties relied upon the dictionary definition of these terms to assist in ascertaining the ordinary and natural meaning. The relevantly applicable definitions of "operate" and "properly" in the Macquarie Dictionary (Online) are: Operate verb (operated, operating) -verb (i) 1. to work or run, as a machine does. 2. to work or use a machine, apparatus, or the like. Properly adverb 1. in a proper manner. 2. correctly. 3. appropriately. 4. decorously. 5. accurately. 6. justifiably. 7. Colloquial completely.
The Prosecutor submitted that these terms should be construed in the context of the legislation as meaning: appropriate to the purpose or circumstances; fit, suitable or conforming to established standards. Such a construction was not inconsistent with the context of the Act.
The context of the WM Act does dictate some degree of accuracy of metering equipment. The objects of the WM Act in s 3 as they realte to use of water identify the following: (e) to provide for the orderly, efficient and equitable sharing of water from water sources, (f) to integrate the management of water sources with the management of other aspects of the environment, including the land, its soil, its native vegetation and its native fauna, (g) to encourage the sharing of responsibility for the sustainable and efficient use of water between the Government and water users, (h) to encourage best practice in the management and use of water.
The WM Act makes provision for the licensing of water taking and the payment by the water user for such water. The equitable allocation and the efficient use and economic return as provided for in the WM Act would be undermined if the metering devices that are the means by which compliance with a licence are to be measured and payment determined could be left to deteriorate through normal wear and tear such that they no longer serve that function.
That being said, I accept the Defendant's submissions that does not mean that proper operation is perfection. Rather, as submitted by the Prosecutor it would permit a reduction in operation below perfection such that there is no material under recording of the water taken by each metering device.
In that context, "material" means something which is of consequence and not de minimis, having regard to the context and purpose of the functioning of the metering device.
In this case, for the reasons outlined above in connection with the Water Take Charges, I am not satisfied that the circumstantial evidence permits the relevant inference to be drawn that there was a substantial take of water from the bores such that it would establish that an amount of water was, by necessity, taken from the bores in excess of the water allocation.
The evidence, apart from the circumstantial, that remains for consideration is that of Mr Judge. Mr Judge's evidence was not satisfactory as it related to the low-flow measurements of the Diesel Pump. From my examination of the video evidence the meter showed movement during this test. The Prosecutor seems to accept that this evidence is not sufficient to establish the flow rates of that pump at low-flow.
That being said, I do not consider that Mr Judge was acting untruthfully or otherwise inappropriately with respect to the other tests undertaken by him. His process was one that was appropriate for the type of testing undertaken by him. The "running start" process whereby the pipe was charged with water and the flow achieved prior to the recording of volumes through the meter and measurement against the metering device recording the volume was appropriate. That process ensured that the meter was receiving the water at the flow rate being measured. The fact that the water that flowed prior to the test was not recorded by Mr Judge was not inappropriate or otherwise representative of an inconsistency, as that part of the test was not recorded and did not affect the final outcome.
The further criticism of Mr Judge's testing as to the fact that it was undertaken on a system of pipes installed in the laboratory rather than in situ is also to be rejected. The system of pipes was installed having regard to the system that existed in situ and was a replication of that system.
As to the risk of the removal process of the meters undertaken by the Prosecutor somehow damaging the metering device by dislodging the magnet it was apparent that the officers who removed the meters did so with some force. The force was applied by means of hitting the meter surround with tools using some degree of force. The meters were tendered in evidence and were examined in the witness box by Mr Mann. He stated that without taking apart the meter casing it could not be established whether the magnet had dislodged during the removal process. If the magnet had been dislodged, the meter would read inaccurately, rather than cease to function all together. Mr Mann, however, did not suggest in evidence that he had viewed the removal process or laboratory processes as shown on the video tendered in the proceedings. He expressed no opinion as to whether the actions taken by the Prosecution investigators was of the type of force to which he referred. Further, not having examined the state of the magnet in the actual pumps, this evidence is not evidence that the force used on these particular pumps was sufficient to dislodge the magnet or otherwise render the meter reading unreliable. The evidence is mere conjecture and insufficient to permit of any reasonable use in these proceedings.
The specifications for the particular Davies Shephard meters used in this case identified that the guaranteed accuracy at installation was ±3%. Therefore, an error rate of that quantum was an expected variation at installation. Accordingly, in order for the particular metering device to be operating properly at installation an error rate of this degree would be acceptable. At installation such a meter would be considered to be operating properly even if it recorded a water take 3% less than in fact was being taken.
The degree of error identified by Mr Judge must also be considered in light of the accepted error rate of the meter. If at installation the degree of variability could be under recorded by 3% and the meter properly functioning the error rates identified by Mr Judge must be adjusted to reflect this acceptable error rate. Accordingly, I accept the Defendant's submission that on the evidence the error rates identified by Mr Judge must be reduced by 3%. The question then arises as to whether the error rate that remains is material.
The adjusted percentage error rates therefore vary from between: 23.04%-4.98% for the House Pump; 8%-8.11% for the Diesel Pump; and 8.97%-6.03% for the Electric Pump.
The adjusted error rates demonstrate that the pumps were not operating within the guaranteed error rate and in all respects were under recording water take by more than 3% and in all cases were more than double that accepted amount. This operating level was, on the evidence, likely caused by the wear and tear on the component parts of the pumps over time. The evidence is also apparent that apart from when a pump failed no regular checking of the pump metering system, or the pump more generally was undertaken. That is, they were left to operate unchecked until failure. Whilst the pump was generally considered robust the evidence also indicates that it was accepted that the accuracy of the pump would diminish with time.
Whilst I accept that the pump accuracy diminished by fair wear and tear, I do not accept that such diminution is not to be considered in the assessment of whether the pump was operating properly for the reasons outlined above. A diminution of the accepted error rate to double or triple that sum is more than a de minimus change. The allocation of water relies upon a degree of accuracy in recording - absent that degree of accuracy both efficiency and sustainability of water resources is diminished. In this case, even excluding the anticipated 3% inaccuracy on the total water take allocated a 4.98% error would result in an additional 49.2ML being taken. Such sum is not inconsequential or not material in the context of the legislative scheme.
As to the honest and reasonable mistake defence, the only evidence of Mr Phelps' dealings with the pumps is that contained in the ROI. To that extent he gives evidence that he satisfied himself from time to time that the meter dials were moving and when the pump stopped working, he had it fixed. The relevant mistake of fact must be that the meters were operating properly - Mr Phelps, on the available evidence, does not appear to have ever turned his mind to this fact. The movement of the dial of a meter or the fact that a pump is pumping is not a determination or assessment of the proper operation of the pump or meter. Accordingly, I find that the Defendant did not hold a relevant mistake as to fact.
As to the s 91M of the WM Act defences such only relate to Tier 1 offences. The Metering Charges are Tier 2 offences, and therefore are not available to the Defendant in these proceedings.
Accordingly, I am satisfied that the Prosecutor has established beyond reasonable doubt that the three meters were not operating properly during the water year as charged contrary to s 91l(2) of the WM Act. Accordingly, there is no need to move to consider the alternative charges.

Natural Resources Access Regulator v Lidokew Pty Ltd

At a glance

Court

Decision date

Before

Source

Judgment (56 paragraphs)

Parties

Legislation Cited (3)

Cases Cited (24)

Cited By (1)