Solicitors:
Department of Planning, Industry and Environment (Prosecutor)
Mills Oakley (Defendant)
File Number(s): 2017/186631, 2017/186632, 2017/186634, 2017/186635
[2]
The Admissibility of the Expert Reports of Aleksandar Todoroski
The defendant, Leda Manorstead Pty Ltd ("Leda"), was found guilty of three charges relating to breaches of s 125(1) of the Environmental Planning and Assessment Act 1979 ("EPAA") (Secretary, Department of Planning and Environment v Leda Manorstead Pty Ltd (No 4) [2019] NSWLEC 58).
On 25 May 2018 Leda pleaded guilty to a fourth charge, namely, that it committed an offence against s 125(1) of the EPAA insofar as it commenced subdivision work (landfilling and associated earthworks in Precinct 2 of the Cobaki Estate) in accordance with a development consent without a construction certificate ("CC") having been issued by the consent authority, the Tweed Shire Council ("the Council"), or an accredited certifier, contrary to s 81A(4)(a) of the EPAA ("the fourth summons").
It is uncontroversial that the four charges related to the same course of conduct by Leda, that is, the development of land owned by it known as the Cobaki Estate, a major residential development on land located in the Tweed Shire.
The Court must now determine the appropriate sentence to be imposed on the defendant, Leda Manorstead Pty Ltd ("Leda"), for the commission by it of the four offences.
This judgment must be read in conjunction with Leda (No 4). The same abbreviations used in Leda (No 4) are used below, unless otherwise specified.
At issue is the admissibility of three affidavits and two expert reports of Mr Aleksandar Todoroski ("the Todoroski evidence"), an air quality engineer engaged by the prosecutor in these proceedings, the Secretary of the Department of Planning and Environment. The Todoroski evidence comprises:
1. an affidavit sworn 16 March 2020, appending an expert report ("the first Todoroski affidavit" and "the first Todoroski report", respectively);
2. an affidavit sworn 15 May 2020, appending an expert report ("the second Todoroski report"); and
3. an affidavit sworn 1 June 2020 ("the third Todoroski affidavit").
In the Todoroski evidence Mr Todoroski opines as to the air quality at the Cobaki Estate development site where the offending occurred and the actual and likely environmental impacts alleged to have occurred as a result of the commission of the offences by Leda as charged in the first and second summons.
Leda engaged Mr Damon Roddis, an air quality specialist, who has provided the following expert reports:
1. an affidavit affirmed 5 June 2018, attaching a report ("the first Roddis report");
2. an affidavit affirmed 29 April 2020, attaching a report ("the second Roddis report"); and
3. an affidavit affirmed 3 June 2020, attaching a report in reply to the second Todoroski report ("the third Roddis report").
The parties accepted that s 79 of the Evidence Act 1995 does not apply to the sentence hearing in the absence of a direction made to this effect under s 4(2) of that Act. No such direction has been sought nor made. Accordingly, the issue of the admissibility of the Todoroski evidence falls to be determined by recourse to common law principles. Having said this, the cases concerning the admissibility of expert evidence pursuant to s 79 of the Evidence Act continue to provide useful guidance to the Court.
[3]
Objections to the Todoroski Evidence
Leda's objections to the Todoroski evidence were essentially four-fold:
1. first, Mr Todoroski failed to disclose the scientific methodology or modelling upon which his expert opinion is based;
2. second, there is an absence of clear identification of the assumed or accepted facts upon which he built his modelling and/or applied his scientific methodology from which his opinions are derived;
3. third, there is no evidence available which, if accepted by the Court, is capable of proving the assumed facts upon which his opinion is based insofar as some of this evidence is based on facts that he observed on a site visit on 19 September 2017 and, as accepted by the prosecutor, Mr Todoroski has misplaced his field notes for the site visit; and
4. fourth, there can be no reliance on dust plume photos (Figures 6 and 7) contained in the second Todoroski report.
The prosecutor readily acknowledged that the first Todoroski report is "not the clearest of reports and could have been better written" (T91:18-19). This is a courageous understatement. It is difficult to reconcile the confusing nature and generally poor quality of the first Todoroski report with his stated experience as an expert.
It was perhaps for this reason that the prosecutor submitted that the first Todoroski report, when read together with the second Todoroski report and the third Todoroski affidavit, renders all of the Todoroski evidence admissible.
In the alternative, the prosecutor contended that if the objections are upheld, they should be limited to exclude some, but not all, of the material contained in the Todoroski evidence, especially the evidence contained in the first Todoroski report.
[4]
The Unfortunate History of Mr Todoroski's Expert Evidence
On 14 August 2019 Mr Todoroski swore an affidavit in the proceedings annexing an undated report ("the original Todoroski report"). In the original Todoroski report, reference was variously made to the modelling utilised by him and to soil sample data and analytical results.
However, while the results of his modelling were produced, there was no identification or explanation of what model he used or how it was applied. Nor was any sampling data and analysis disclosed in the original report.
Accordingly, as deposed to in the affidavit of Ms Erin Gavin affirmed 16 March 2020 (Ms Gavin is the Acting Director of Planning and Resources Litigation employed by the Department of Planning, Industry and Environment ("the Department")), in November and December 2019 Leda requested that the prosecutor provide it with the modelling and relevant inputs so that its air quality expert, Mr Roddis, could properly analyse and consider the opinions set out in the original Todoroski report.
Mr Todoroski refused to provide the modelling on the basis that it was sensitive, confidential and "commercial-in-confidence" to his business. Leda therefore issued a subpoena to produce to the prosecutor and to Mr Todoroski for his modelling data, sampling and field notes for the soil sampling, laboratory results for the soil analysis, and emission calculations.
The subpoena was stood over several times pending the receipt by Mr Todoroski of legal advice.
Ultimately, Mr Todoroski prepared an updated expert report including modelling that did not use sensitive, confidential or commercial-in-confidence information (that is, the first Todoroski report). The first Torodorski report wholly replaced the original Todoroski report.
Meanwhile, an email exchange between the solicitors for the Department and the solicitors for Leda commencing on 16 April 2020, established that, according to Mr Roddis, most of the material previously requested under the subpoena (which was no longer pressed) has been provided but that the following information remained outstanding:
∙ Laboratory results for soil analysis, including any email, written note or document containing sample results;
∙ Sampling and field notes for the soil sampling; and
∙ Emissions calculations (including spreadsheets and written notes) used to generate the emission rates that have been used in the dispersion modelling.
By way of reply dated 20 April 2020, the solicitors for the Department provided the laboratory results and the emissions calculations, but advised that Mr Todoroski has either "lost or misplaced any field notes he may have had, and is unable to locate those so cannot provide them."
It is not therefore in dispute that modelling data relied upon by Mr Todoroski was provided to Mr Roddis on 2 and 21 April 2020, and has been reviewed and responded to by Mr Roddis in his first report.
In response to the first and second Roddis reports, Mr Todoroski prepared his second affidavit and report. In the second Todoroski report, Mr Todoroski clarified and corrected errors in the modelling and results contained in his first report. The errors were not insignificant and produced a different set of figures and findings in the second report.
It is within this context that determination of the voir dire takes place.
[5]
What is Required for Expert Reports to be Admissible
In the seminal decision of Makita (Australia) Pty Ltd v Spowles [2001] NSWCA 305; (2001) 52 NSWLR 705 Heydon JA (as his Honour then was) said the following (at [85] and [87]):
85 In short, if evidence tendered as expert opinion evidence is to be admissible, it must be agreed or demonstrated that there is a field of "specialised knowledge"; there must be an identified aspect of that field in which the witness demonstrates that by reason of specified training, study or experience, the witness has become an expert; the opinion proffered must be "wholly or substantially based on the witness's expert knowledge"; so far as the opinion is based on facts "observed" by the expert, they must be identified and admissibly proved by the expert, and so far as the opinion is based on "assumed" or "accepted" facts, they must be identified and proved in some other way; it must be established that the facts on which the opinion is based form a proper foundation for it; and the opinion of an expert requires demonstration or examination of the scientific or other intellectual basis of the conclusions reached: that is, the expert's evidence must explain how the field of "specialised knowledge" in which the witness is expert by reason of "training, study or experience", and on which the opinion is "wholly or substantially based", applies to the facts assumed or observed so as to produce the opinion propounded. If all these matters are not made explicit, it is not possible to be sure whether the opinion is based wholly or substantially on the expert's specialised knowledge. If the court cannot be sure of that, the evidence is strictly speaking not admissible, and, so far as it is admissible, of diminished weight. And an attempt to make the basis of the opinion explicit may reveal that it is not based on specialised expert knowledge, but, to use Gleeson CJ's characterisation of the evidence in HG v The Queen (at 428 [41]), on "a combination of speculation, inference, personal and second-hand views as to the credibility of the complainant, and a process of reasoning which went well beyond the field of expertise".
…
87 There is no doubt about Professor Morton's authority, experience, qualifications and skill. It is also the case that Professor Morton's report is quite lengthy and detailed. But, given that the court is not obliged to take the opinion of an expert as conclusive even though no other expert is called to contradict it, can it be said that Professor Morton's report goes beyond a series of oracular pronouncements? Does it usurp the function of the trier of fact? More vitally, did it furnish the trial judge with the necessary scientific criteria for testing the accuracy of its conclusions? Did it enable him to form his own independent judgment by applying the criteria furnished to the facts proved? Was it intelligible, convincing and tested? Did it go beyond a bare ipse dixit? Did it contain within itself materials which could have convinced the trial judge of its fundamental soundness?
At common law the admissibility of expert opinion evidence depends on the proper disclosure and evidence of the factual basis of the opinion, or the so-called 'basis rule'. The expert must disclose the facts (usually by way of assumptions) upon which the opinion is based, those facts must be capable of proof by admissible evidence, and evidence must be admitted to prove those assumed facts (Dasreef Pty Ltd v Hawchar [2011] HCA 21; (2011) 243 CLR 588 at [66]-[90] per Heydon J).
In the leading authority of Dasreef, the plurality (French CJ, Gummow, Hayne, Crennan, Kiefel and Bell JJ) endorsed the remarks of Heydon JA in Makita and said that (at [37] and [41]-[42], footnotes omitted):
37 It should be unnecessary, but it is nonetheless important, to emphasise that what was said by Gleeson CJ in HG (and later by Heydon JA in the Court of Appeal in Makita (Australia) Pty Ltd v Sprowles (51)) is to be read with one basic proposition at the forefront of consideration. The admissibility of opinion evidence is to be determined by application of the requirements of the Evidence Act rather than by any attempt to parse and analyse particular statements in decided cases divorced from the context in which those statements were made. Accepting that to be so, it remains useful to record that it is ordinarily the case, as Heydon JA said in Makita, that "the expert's evidence must explain how the field of 'specialised knowledge' in which the witness is expert by reason of 'training, study or experience', and on which the opinion is 'wholly or substantially based', applies to the facts assumed or observed so as to produce the opinion propounded". The way in which s 79(1) is drafted necessarily makes the description of these requirements very long. But that is not to say that the requirements cannot be met in many, perhaps most, cases very quickly and easily. That a specialist medical practitioner expressing a diagnostic opinion in his or her relevant field of specialisation is applying "specialised knowledge" based on his or her "training, study or experience", being an opinion "wholly or substantially based" on that "specialised knowledge", will require little explicit articulation or amplification once the witness has described his or her qualifications and experience, and has identified the subject matter about which the opinion is proffered.
…
41 Contrary to submissions on behalf of Mr Hawchar, this analysis does not seek to introduce what has been called "the basis rule": a rule by which opinion evidence is to be excluded unless the factual bases upon which the opinion is proffered are established by other evidence. Whether that rule formed part of the common law of evidence need not be examined. It may be accepted that the Law Reform Commission's interim report on evidence denied the existence of such a common law rule and expressed the intention to refrain from including a basis rule in the legislation the Commission proposed and which was later enacted as the Evidence Act 1995 (Cth) and the Evidence Act 1995 (NSW). What has been called the basis rule is a rule directed to the facts of the particular case about which an expert is asked to proffer an opinion and the facts upon which the expert relies to form the opinion expressed. The point which is now made is a point about connecting the opinion expressed by a witness with the witness's specialised knowledge based on training, study or experience.
42 A failure to demonstrate that an opinion expressed by a witness is based on the witness's specialised knowledge based on training, study or experience is a matter that goes to the admissibility of the evidence, not its weight. To observe, as the Court of Appeal did, that what Dr Basden said about the volume of respirable dust to which Mr Hawchar was exposed over time was "an estimate" that was "contestable and inexact" no doubt did direct attention to its worth and its weight. But more importantly, it directed attention to what exactly Dr Basden was saying in his evidence and to whether any numerical or quantitative assessment he proffered was admissible. And if, as the Court of Appeal observed, his opinion on that matter lacked reasoning, the absence of reasoning pointed (in this case, inexorably) to the lack of any sufficient connection between a numerical or quantitative assessment or estimate and relevant specialised knowledge.
In HG v The Queen [1999] HCA 2; (1999) 197 CLR 414 Gleeson CJ had described the position thus (at [39], footnotes omitted):
39 The opinions of Mr McCombie were never expressed in admissible form. An expert whose opinion is sought to be tendered should differentiate between the assumed facts upon which the opinion is based, and the opinion in question. Argument in this Court proceeded upon the basis that it was possible to identify from Mr McCombie's written report some facts which he either observed or accepted, and which could be distinguished from his expressions of expert opinion. Even so, the provisions of s 79 will often have the practical effect of emphasising the need for attention to requirements of form. By directing attention to whether an opinion is wholly or substantially based on specialised knowledge based on training, study or experience, the section requires that the opinion is presented in a form which makes it possible to answer that question.
The trinity of authority referred to above have been applied in a plethora of subsequent cases (one example relied upon by Leda is Nicholls v Michael Wilson & Partners Ltd [2012] NSWCA 383 at [234]). In Kyluk Pty Ltd v Chief Executive, Office of Environment and Heritage [2013] NSWCCA 114, albeit in the context of the application of s 79 of the Evidence Act, Schmidt J succinctly summarised the principles as follows (at [176]-[177], with whom McCallum J, as she then was, agreed at [138]):
176 In the result, the current position appears to be that for expert opinion evidence to be admissible under s 79 of the Evidence Act, it must satisfy the two criteria identified by the plurality in Dasreef. That is, it must establish that the expert "has specialised knowledge based on the person's training, study or experience" and that the opinion expressed by the expert "is wholly or substantially based on that knowledge". The evidence must also be presented in a form which reveals the facts and reasoning on which the opinion rests.
177 An expert opinion which meets those requirements need not be excluded if all of the factual bases upon which the opinion is proffered are not established by the expert's own evidence. Even if facts which the expert "assumes" or "accepts" in reaching the opinion expressed are not proved in some other way, then the opinion may still be admissible. That will depend on the nature of those facts and what bearing they have on the opinion. If they provide but a small part of the basis upon which the opinion rests, then the failure to prove those facts may have but little impact, and not render the opinion inadmissible. The failure to prove facts which provide a significant basis for the opinion might, by way of contrast, be such as to render the opinion no longer relevant to a fact in issue, no foundation for the opinion having been established. Such an opinion, even if it were admitted, would be of no value. Where an opinion is admitted, the failure to establish a fact which is not of such significance, may nevertheless have an impact on the weight given to the opinion.
In summary, for expert evidence to be admissible, both the facts upon which the opinion is based and the reasoning process leading to the formation of the opinion must be exposed. If not, it is questionable whether the opinion is based on particular specialised knowledge. A bare ipse dixit (or, 'I say so') will not suffice and is likely to render an expert report inadmissible on the basis that it will neither be intelligible and convincing nor capable of testing.
These principles are in conformity with an expert witness's obligation to comply with the Expert Witness Code of Conduct contained in Sch 7 to the Uniform Civil Procedure Rules 2005 ("UCPR") ("the Code") (see r 3J in Pt 75 of the Supreme Court Rules 1970 ("SCR"). Rule 5.2(1) of the Land and Environment Court Rules 2007 applies Pt 75 of the SCR to criminal proceedings in Class 5 of this Court's jurisdiction). Clause 3 of Sch 7 to the UCPR relevantly provides:
3 Content of report
Every report prepared by an expert witness for use in court must clearly state the opinion or opinions of the expert and must state, specify or provide -
…
(d) the assumptions and material facts on which each opinion expressed in the report is based (a letter of instructions may be annexed), and
(e) the reasons for and any literature or other materials utilised in support of each such opinion, and
…
(g) any examinations, tests or other investigations on which the expert has relied, identifying the person who carried them out and that person's qualifications, and…
The requirement to comply with the Code and with Div 2 of Pt 31 of the UCPR (which concerns provisions applicable to expert evidence) is also stipulated in r 31.23 of the UCPR and in the relevant practice note of the Court, namely, Practice Note - Class 5 Proceedings (at [36]).
[6]
Not All of the Todoroski Evidence is Admissible
It is convenient to deal with the second, third and fourth criticisms of the Todoroski evidence before turning to the much more vexed principal objection raised by Leda in its first complaint.
[7]
The Absence of Clear Identification of Assumed Facts
In his first report, Mr Todoroski set out the limitations in developing an air dispersion model quantifying the likely effects of the excess dust emissions from the site on residents in the absence of specific and detailed information about the size of the exposed disturbed area, the quantities of material moved, how the material was moved, and how it was handled and processed (at paragraphs 17 and 18). However, he went on to state that he relied upon the information in a report by Mr Leo Watts to quantify the exposed disturbed areas (at paragraph 19). He noted that he did not model dust from material movement but limited his modelling to the impacts of dust generated by wind erosion from the surface of the exposed disturbed areas, thereby providing an underestimate of the likely dust levels (at paragraphs 20 and 21).
Mr Torodorski goes on to describe his inspection of the site, the soil samples taken by him during the site inspection and the observed site silt levels and their distribution (see further below).
In his first report it is clear that Mr Todorski utilises the default wind erosion emission factor used in dust modelling assessments in NSW as set out in the National Pollution Inventory (while he does not state this explicitly, this is the plain inference to be drawn from a fair reading of paragraphs 25 and 26 of the first Todoroski report).
He then expressly sets out "some key assumptions made" by him in providing his opinion (at paragraphs 28 and 29, footnote omitted):
28 I chose to prepare a generally conservative (likely to underestimate effects) modelling assessment. Some key assumptions were made, including;
(a) Only wind erosion was modelled, despite other material handling activity on the Site. Such other activity would generate significant dust emissions, including during times of poor wind dispersion, when impacts per unit of emissions may be larger than on windy, high dispersion days.
i. This is likely to significantly underestimate the actual impacts. I considered modelling the potential excess dust generated by excess plant and equipment movements and material handling but was unable to do so accurately in the absence of information about the material handling that occurred.
(b) The modelled emissions are proportional to the wind speed. Under low wind speed nil or negligible emissions occur in the modelling and high emissions occur, per the emission factor, under higher wind speeds (as would occur in reality for wind erosion).
(c) The default Australian (NPI) emission factor of 0.4 kg/ha/hr was applied, despite my observations as set out in Point 24 and 25 indicating that a higher value would be expected for the erodible material on the Site.
(d) A default particle size distribution was used in the modelling, despite the surface soil samples indicating a much larger fraction of fine material. The site specific data reveal approximately five times more of the PM2.5 material that would have been generated.
(e) The 5.59 ha of the exposed disturbed surfaces generally nearest to the most likely impacted receptors were excluded from the modelling of excess dust impacts in every instance.
(f) It was assumed that all undisturbed surfaces and roads did not have any emissions.
29 Other aspects I considered include that watering of a main road on the Site is evident in one Google Earth image of the Site, and watering would reduce wheel generated dust. Whilst I did not include wheel generated dust emissions in the model, it is reasonable to assume that the watercart may have also sprayed water at other times, or on other surfaces. However, many watercarts would be needed to control wind erosion from areas this large on a warm and/or windy day, and I did not find any aerial images which show many carts on the Site. I did not find any images that show fixed water sprays or irrigation sprays either. In this regard, I also considered the information from Site representatives in their response to dust complainants which indicate the Site was vacant on weekends and after hours. Thus staff could operate any water carts were not present out of hours to control dust from wind erosion.
(a) I also considered whether all of the modelled exposed disturbed area would emit dust, noting that for example the nearest 5.59 ha to residents cannot actually be in the same place in every scenario, given that the Site could not be developed if this were the case. Whenever the allowed 5.59 ha area is further from the receptors than modelled, the impacts due to the excess exposed disturbed area increase. The consideration in this regard was to assume, unrealistically but conservatively, that the allowed 5.59 ha was always nearest to the receptors as this provides the largest allowed impact and least excess impact. It was also taken as fact that the areas in Leo Watt's affidavit represent the total exposed and disturbed land that was not stabilised in terms of wind erosion.
Additional assumptions were contained in the second Todoroski report (for example, the basis for his emission factor selection set out at paragraph 74).
Further, as was acknowledged in the email exchange referred to above (at [20]-[21]), Mr Todoroski provided the laboratory results for the soil analysis (including all notes) and the emissions calculations (including spreadsheets and written notes) used to generate the emission rates employed by him in his dispersion modelling. In the second Roddis report, Mr Roddis noted that this information had been provided by Mr Todoroski and that "the information provided is now broadly what I would expect to allow me to make a more informed evaluation."
In my opinion, the factual bases upon which the opinions proffered in the Todoroski evidence are based have been sufficiently set out in the Todoroski evidence and, where necessary, supplemented by the provision of the additional information. Whether they can be proven or established is not relevant to the question of admissibility, but goes to the weight to be afforded to the evidence.
[8]
Failure to Provide Field Notes of Site Visit and Soil Sample Collection
In the first Todoroski report, Mr Todoroski described his site visit and the conditions he observed at the time as follows (at paragraphs 6-11):
6 I visited the Site on 19 September 2017 between approximately 9:00 am and 12:30pm. I took four surface soil samples, some photographs and made various observations.
7 The conditions during my site visit were relatively dry, hot and windy.
8 Weather conditions as measured by the Bureau of Meteorology weather station at Coolangatta during the site visit period indicate that temperatures ranged from 21.7 to 25.3 degrees Celsius with winds from the northwest at speeds ranging up to 6.5 to 8.9 metres per second.
9 Just prior to entering the gate to the south west of the Site, I observed significant plumes of dust and willy-willies or dust devils in the southern exposed areas of the Site. The visible plumes of dust were several hundred metres in length and approximately 50 m in height, however the visible dust dissipated quickly. The dust was pale, off-white in colour and was being generated by wind acting on the surface of an area of similarly coloured exposed material.
10 During the entire time of the site visit no plant or machinery was operating. Several plant items were parked near a shed in the south eastern part of the Site, and no plant items were observed elsewhere on the Site. Only the light vehicle I travelled in during the site visit was moving on the Site.
11 The site visit essentially entailed a clockwise inspection of the Site, starting in the south west corner. I collected four samples of surface material during the site visit, at the locations shown Figure 1 [sic]).
A map was included indicating the locations of the surface soil samples taken by him on the visit (Figure 1).
His description of his site visit and surface soil sample collection was elaborated upon later in his first report (at paragraphs 22 to 24):
22 As outlined previously, I conducted a site inspection and obtained four surface soil samples. I was only able to identify some limited locations at which to take these samples, i.e. where the surface materials would have been similar to those prior to establishing a grass cover, or when complaints were made and activities were occurring near residences in the north west of the Site.
23 The sample data are consistent with my site observations. Sample 1, being generally firm, undisturbed in-situ material showed the smaller fraction of silt material (although at the time of the complaints may have been disturbed by various activities at the Site), Sample 2 and Sample 4, have a larger fraction of silt, and Sample 3, taken on pale, off-white material emplaced on the Site has a high fraction of silt material, consistent with my observations of it being very dusty.
24 I considered the observed site silt levels, the size distribution of silt, my observations of the nature of the surface material and any crusting and surface silt layers, the dust from wind erosion evident in complainant's videos and photos, that I observed on the Site, and from the action of a light vehicle on the Site on the day I inspected the Site in determining the emissions factor to use to represent the likely rate of dust lift-off at this location due to wind erosion.
Further detail concerning his site visit and the soil sample collection was provided in the third Todoroski affidavit, which formally appended the certificates and the results of the analysis of the four soil samples taken, including an analysis of the particle size distribution.
As was discussed above, these analyses were provided to Mr Roddis by 21 April 2020.
Mr Roddis reviewed the laboratory results for Mr Todoroski's soil samples and noted that they were broadly consistent with his findings (p 12 of the second Roddis report. As detailed in his second report, Mr Roddis completed his own site inspection and soil sampling on 10 October 2019). Using his results, Mr Roddis was able to conclude that the material on site would not become readily wind-erodible fine material, which he noted was contrary to the conclusion stated in the second Todoroski report (p 11).
Plainly the results from Mr Todoroski's soil sampling should have been appended to the first Todoroski report, but their omission does not, in my view, render any of the Todoroski reports inadmissible in light of their provision to Mr Roddis in April 2020 and their attachment to the third Todoroski affidavit.
Similarly, the fact that Mr Todoroski has lost his field notes does not render the Todoroski evidence inadmissible. While there is no information on how the samples were collected, there is sufficient detail in the first Todoroski report and third Todoroski affidavit as to the conditions under which the site visit occurred and where on site the samples were collected, to render the Todoroski evidence admissible. The lack of field notes is, in this instance, a matter going only to the weight to be attributed to the Todoroski evidence in this regard.
[9]
Reliance on Figures 6 and 7 in the Second Todoroski Report
Leda separately objected to paragraph 24 of the second Todoroski report where Mr Todoroski relies on Figure 6 (a photo of dust blowing in Hagavatn, Iceland) and Figure 7 (an image of dust concentrations he has "measured for visible dust plumes at another location").
Assuming Figure 6 is what is also referred to as "Figure 4" in paragraph 24(d) of the second Todoroski report, leaving aside issues of weight, the evidence is admissible on the basis that it depicts what a dust plume "that is many kilometres in width…at concentrations well above 50 μg/m3" looks like, an observation that Mr Todoroski opines that he is familiar with.
In my view, the photos at Figure 7 fall into the same category. This is reinforced by the fact that they are images of dust concentrations that he has measured for visible dust plumes. The undisclosed location of the photos, while a matter going to the weight to be placed on this evidence, does not render them inadmissible. This is because the reliance placed on them by Mr Todoroski is not location specific. Rather, it is to depict what a dust plume at those concentrations looks like.
[10]
Failure to Disclose Modelling Used
More problematic is, as Leda identifies, the fact that nowhere in the first Todoroski report does Mr Todoroski state what model he has used to formulate his results and provide his opinions. This was so notwithstanding that one of the principal reasons for providing the first Todoroski report was "to present the modelling results on the basis of generally used model settings in the air assessment industry (i.e. the suite of various settings I've encountered and used in my previous places of employment)".
After setting out the assumptions upon which he relied, including the default wind erosion emission factor that he utilised, Mr Todoroski explained his overall methodology by stating that he focused on excess amounts of dust rather than on compliance with total dust levels referable to the NSW EPA criteria. While he does not expressly state what this criteria is, at the back of his first report there is a reference to "NSW EPA (2017) 'Approved Methods for the Modelling and Assessment of Air Pollutants in NSW', January 2017", which, it may be inferred, is the relevant criteria. Similar references at paragraphs 34 to 36 and 42 of his first report support this conclusion.
As a guide to potential background dust levels at the site, Mr Todoroski set out the levels recorded at the nearest publicly available monitoring station in Springwood, Queensland, in Table 1 and Figure 2 of his first report (see paragraph 32). He acknowledged that "due to the large separation distance, there is significant uncertainty as to how well these data may represent the Site background dust levels."
The background dust levels were then "added to the modelling results (shown later) to indicate the total dust levels" (paragraph 33).
Under the heading "RESULTS", Mr Todoroski referred to the "modelled PM10 and PM2.5 results at three nearby receptor locations" for three categories of exposed areas. These results were set out in Tables 2 to 4. While the significance of the results and their likely impact was explained by Mr Todoroski (at paragraphs 40 to 43), how these predicted results were calculated and by what method was not revealed.
Under the heading "SUMMARY" Mr Todoroski opined as follows (at paragraphs 44 to 49):
44 In my opinion, I consider the operation of the Site, with many times larger exposed disturbed areas than I understand were allowed, and operation for many years without any apparent stabilisation of the exposed surfaces (such as establishing grass), is likely to have led to unacceptable levels of impact at nearby receptors.
45 I cannot quantify the exact degree, duration or extent of such impacts due to a lack of exact quantitative information, and instead I have conducted modelling to estimate the potential wind erosion dust impacts with the available data.
46 Wind erosion is only a subpart of the total emissions that would affect resident exposure, and the modelling is based on generally conservative assumptions. Thus I consider that the predicted results in this report would significantly underestimate the likely actual effects.
47 My site observations, the consistency of the complaints data with my observations, and the generally high modelled impact results, especially for annual average PM2.5 and 24-hour PM10, despite the conservative assumptions are key factors for my opinion in this situation.
48 In my opinion, the most significant finding from my work is that the impacts from the Site would have been 3 to 4 times higher at receptors for the key annual average PM2.5 metric most closely associated with health impacts. The absolute PM2.5 levels are likely to have been significantly higher than modelled for a range of reasons, including the modelling conservatively underestimating the fraction of PM2.5 in material on the Site, but the relative difference between the impacts of what was allowed and occurred would generally remain in the range of 3 to 4 fold or greater whilstsoever the allowed 5.59 ha remained nearest to receptors. This can reasonably be interpreted to mean there would have been approximately 3 to 4 times the health impact than allowed.
49 The other relevant findings are that due to excess dust from the excess disturbed areas in general the EPA criteria would have been exceeded for 24-hour PM10 by a large margin, and are likely to have been exceeded for the other PM10 and PM2.5 metrics also.
Again, the modelling that Mr Todoroski referred to in his "SUMMARY", and throughout the body of his first report (for example, the reference to "predicted results" at paragraph 31 or the "modelling results" at paragraph 33), remained undisclosed.
There was reference to a model in the "References" section at the end of his first report, namely, the "'Generic Guidance and Optimum Model Setting for the CALPUFF Modelling System for Inclusion into the Approved Methods for the Modeling [sic] and Assessments of Air Pollutants in NSW, Australia'. Prepared for the NSW Office of Environment and Heritage by TRC Environmental Corporation" ("CALPUFF model"). However, there was no explicit discussion of the model in the first Todoroski report, whether it was in fact used by Mr Todoroski, and if it was, how it was used by him. In my opinion, having regard to the legal principles referred to above, this omission is fatal to the admissibility of most of the first Todoroski affidavit and report.
The prosecutor submitted that an inference ought to be drawn that the CALPUFF model referenced at the end of the first report was the same model referred to by Mr Todoroski throughout that report.
The difficulty with this submission is the fact that the refusal by Mr Todoroski to provide the modelling used by him to produce his original report, together with the subsequent withdrawal by him of that report and its replacement with another report (the first Todoroski report) that appeared to use different modelling, has a very real tendency to create considerable uncertainty as to what model was used. Adding to the confusion was the suggestion in the original report, not displaced by the circumstances of the provision of the first Todoroski report, that the model used by Mr Todoroski was his own model, not that of a third party. The reference to the CALPUFF model at the end of the first Todoroski report appears, absent further elucidation or clarification, to be a reference to a generic model, not a model created by Mr Todoroski.
In these circumstances, to draw the inference invited by the prosecutor as to which model was used by Mr Todoroski in his first report would be unsafe.
Neither party could find definitive authority on whether it was permissible for the Court to read the two Todoroski reports together, so that the second report cured the defects of the first report (see the discussion below). In this context it is worth observing that in Dasreef the joint judgment noted that the requirements in Makita can "be met in many, perhaps most, cases very quickly and easily" (at [37], quoted above).
In X v Sydney Children's Hospital Speciality Network (No 5) [2011] NSWSC 1351, faced with a Makita objection to a joint expert report that did no more than record one word answers and expressed no reasons, Adamson J found that the submission that when the single expert reports and the joint report were read together the requirement for reasons was satisfied, had "significant force" (at [56]). Ultimately her Honour did not decide the issue, disposing of the objection for other reasons.
While such an approach is superficially attractive, even reading the two Todoroski reports together, as urged by the prosecutor, does not assist. Not only does Mr Todoroski not state what model he used in his first report anywhere in his second report, his second report corrects fundamental errors made by him in his first report that largely render his critical results and findings in the first report otiose (for example, the results in Tables 2 to 4: see [75] below). Given the constantly evolving nature of the modelling used by Mr Todoroski across his original, first and second reports, to draw appropriate inferences as to the scientific methodology used in the first report by recourse to the second report would, to reiterate, not be safe.
Finally, that Mr Roddis was able to engage with the first Todoroski report in the second Roddis report and make a "more informed evaluation" following the provision of the requested information, including the dispersion modelling information, does not, of itself, ameliorate the deficiencies inherent in the first report thereby rendering it admissible.
Having said this, I accept the submission of the prosecutor that there are some parts of the 16 March 2020 affidavit and the first Todoroski report that are nevertheless admissible. That is, evidence given by Mr Todoroski as to what he observed on site, what surface soil samples he collected, and his evidence concerning the NSW EPA criteria and the NPI emission factor. What weight is to be given to this remaining evidence will be another matter. I set out the evidence that is admissible in the first Todoroski affidavit and report at the end of this judgment.
In the second Todoroski report, Mr Todorsoki examined the first and second Roddis report and tested the veracity of the Roddis modelling. He also corrected errors contained in his first report.
Many of the defects contained in the first Todoroski report were not present in the second Todoroski report. Specifically, unlike the first Todoroski report, there is express reference to the use of the CALPUFF model in the second Todoroski report (at paragraph 30):
30 In terms of a direct comparison between the Roddis and my results, there is an obvious four fold difference that must arise due to just the adopted emissions factor (assuming all else being equal, which isn't the case). I use the Australian default emissions factor, and Roddis uses the default US factor. Roddis underestimates the size of the exposed area and uses spatial scaling or spatial transfiguration to correct this. This would cause a further multiplicative factor of difference between the Roddis results and mine, perhaps several fold times less for Roddis relative to mine (but just how large this factor is cannot be readily determined). In general I also find that the air dispersion model Roddis used (AERMOD), will result in short term 24-hour average dust levels predictions approximately half those predicted in my chosen model CALPUFF. Compared to CALPUFF, AERMOD shows significantly poorer correlation between the modelled results and the actual observations when modelling sources near the ground, but similar, good results when modelling tall industrial stacks. However, this difference can vary greatly depending on the specific case, and there is a large sensitivity in this case (in both models) when the receptors are so close to such large sources of dust. In this case I would not be willing to venture so far as to estimate what factor of difference there may be between the models alone (all else being equal), other than to comment that in general, the CALPUFF model results are usually higher.
Reading the report as a whole, it is clear that the modelling used by Mr Todoroski in his second report to provide his opinions is the CALPUFF model.
In addition, many of the calculations, explanations and elucidations of the scientific methodology employed by him that were absent from the first Torodorski report could be located in the second Todoroski report. While the second report was not as pellucidly written as could be expected from an expert giving evidence in legal proceedings, it did not impermissibly infringe the legal principles governing the adducing of expert evidence referred to above.
Once it is understood that:
1. the model used by Mr Todoroski in the second report was the CALPUFF model;
2. most of the salient assumptions, facts and reasoning processes leading to the formation of the opinions expressed in the second Todoroski report (albeit some of which were found in the first Todorski report) were set out; and
3. additional material was provided to Mr Roddis in April 2020, including information that was "broadly consistent with a conventional atmospheric dispersion modelling set up",
the second Todoroski report is, on balance, admissible.
In his second report, Mr Todoroski used his modelling to both critique the modelling used by Mr Roddis and, as stated above, to correct errors in the first Todoroski report. Insofar as the former use is concerned, the second Todoroski report is analogous to a peer review of the first and second Roddis reports. Insofar as the latter use is concerned, there is, as the prosecutor submitted, nothing impermissible in an expert witness clarifying and correcting in a subsequent report mistakes made in an earlier report. By conducting a more thorough analysis of his modelled material and in revising his calculations, this is exactly what occurred in the second Todoroski report (see paragraph 36(a)).
It is not correct to say, as Leda does, that no explanation is given by Mr Todoroski for the revised calculations at paragraph 36(a) of his second report. On the contrary, Mr Todoroski states that he found a "significant overestimation" in his "result for the maximum 24 hour average PM10 concentration" due to "using a grid residual method to interpolate between contour lines to obtain a result at a specific intermediate point", which would usually not be problematic but was in this instance because of the "very steep and non-liner [sic] change in the dust levels near the receptors". While somewhat opaque, the explanation is nevertheless sufficient.
In relation to the new figures in Tables 2 to 4 of his second report, Mr Todoroski conceded at paragraph 36(a) that the data extracted in his first report was not "exact" and that results presented in Tables 2 to 4 of his first report were therefore not correct. While this may raise serious doubts about the efficacy of Mr Todoroski's modelling, it does not render the evidence contained in the second Todoroski report inadmissible. As the prosecutor observed, the model outputs data because it runs simulations each day. The findings of Mr Todoroski in the second report are based on a complete data set and not merely a maximum 24 hour value (as calculated in the first Todoroski report). As Mr Todoroski noted, while this affects specific findings, it did not alter his fundamental conclusions which were based upon results that were what he considered to be "largely unaffected" by the earlier modelling error. Whether this albeit limited explanation will withstand challenge remains an open question, but the evidence is nevertheless admissible.
Leda raised a separate objection to paragraph 86 (and presumably paragraph 87, although it was not expressly referred to by Leda) of the second Todoroski report, which states that:
Testing of Roddis's recommended modelling approach for rainfall and including wind.
86 I have tested the effect of using Roddis's recommended rainfall modelling approach and also the default wind speed threshold. In this regard, I note the following:
(a) In 2015, Roddis is correct that including rainfall does result in generally less impacts.
(b) Including the effects of rainfall per Roddis however does not lead to lower impact during the charge period, where up to 35% more impacts arise.
(c) Roddis does not consider the complete second charge period, or any of the first charge period in his claims. In regard to the relevant charge periods, he is incorrect in a materially significant way.
In paragraph 87 Mr Todoroski set out the results of his "testing". Later in the paragraph he indicated that he used his modelling do to so. Consistent with the reasons given above, the evidence at paragraphs 86 and 87 of the second Todoroski report is admissible.
By way of further objection, Leda submitted that at paragraphs 97 to 99 of the second Todoroski report Mr Todoroski varied the calculations in his first report "without explanation as to why he has done so". This is not correct. The explanation is that provided earlier by him in his second report at paragraph 36(a) (discussed immediately above). As he stated at paragraph 96, the "result [sic] correct overestimations arising from inexact interpolation methods used previously due to time pressures in preparing my earlier 2020 affidavit." The "results" referred to by him were those in Tables 2 to 4, and the error impacted his previous calculation of the "annual average results".
Finally, Leda complained that the second Todoroski report introduced new material (for example, photographs) and new findings (presumably the revised modelling results).
It should be noted that the second Todoroski affidavit (appended to which was the second Todoroski report) was filed and served pursuant to Court order. At no time has Leda asserted that it is prejudiced by this additional material. If it genuinely is, this can be readily accommodated by allowing Leda time to respond to the material in light of the fact that the sentence hearing is listed in two tranches (4-12 June and 24-28 August 2020).
[11]
Summary of Rulings and Orders
Consistent with the reasons given above, my rulings on the objections to the Todoroski evidence are set out below.
[12]
First Todoroski Affidavit and Report
In the affidavit dated 16 March 2020 and the first Todoroski report most of Leda's objections are upheld and the following evidence is not admissible:
1. the second sentence of paragraph 8 of the affidavit to the end of that paragraph;
2. all of subparagraphs 3(a) and (b) of the report;
3. paragraphs 19 and 20 of the report;
4. the second sentence of paragraph 21 of the report;
5. paragraph 29(a) of the report;
6. paragraph 30 of the report;
7. paragraph 38 of the report to the conclusion of the first report, excluding the "References" section.
While at first glance the material contained in paragraph 28 and the remainder of paragraph 29 refer directly to the rejected modelling evidence, because this evidence is no more than a statement of the assumptions used by Mr Todoroski in carrying out his modelling exercise, they are, in my opinion, admissible. As the prosecutor submitted, just because evidence is modelling related does not, of itself, render it inadmissible.
[13]
Second Todoroski Affidavit and Report
The objections raised by Leda in respect of the second Todoroski affidavit and report are rejected. The second Todoroski affidavit and report is therefore admissible in its entirety.
[14]
Third Todoroski Affidavit
As is apparent from the analysis above, most of Leda's objections to the Todoroski evidence were directed to the first and second Todoroski affidavits and appended reports.
As alluded to above, the third Todoroski affidavit was directed to:
1. the formal provision of the results of the analysis of the soil samples taken by Mr Todoroski during his site visit (which had previously been provided to Mr Roddis); and
2. complaints in the form of emails from residents living near the Cobaki Estate and photographs and videos that were attached to those emails, including in particular, emails, photographs and videos sent from Mr Robert Morrison to Mr Stewart McLachlan at the Department. Mr Todoroski had regard to these complaints and the attached material in preparing his first and second reports.
In my opinion, none of the objections raised by Leda discussed above are applicable to the third Todoroski affidavit. That affidavit is therefore also admissible in its entirety.
Exhibits 1 and 2 on the voir dire are to be returned.
[15]
DISCLAIMER - Every effort has been made to comply with suppression orders or statutory provisions prohibiting publication that may apply to this judgment or decision. The onus remains on any person using material in the judgment or decision to ensure that the intended use of that material does not breach any such order or provision. Further enquiries may be directed to the Registry of the Court or Tribunal in which it was generated.
Decision last updated: 12 June 2020