AUSTRALIA - MEASURES AFFECTING IMPORTATION OF SALMON
- RECOURSE TO ARTICLE 21.5 BY CANADA -

D. DUE PROCESS

Australia

4.51 Australia also raised some due process concerns. These relate to two matters: (1) Canada's supplementary submission of 30 September 1999; and (2) the requirement of Canada to demonstrate a prima facie case and Australia's opportunity to respond to claims made against it.

4.52 The Panel granted Canada an additional period of time to provide arguments relevant to AQPM 1999/64. However, Canada's supplementary submission is almost entirely concerned with a disease of non-salmonids endemic to the whole of Australia, the factual matters of which was publicly available information.

4.53 In its questions to parties, the Panel invited Canada to correct deficiencies in its claims and arguments (Questions 12 and 13). These included Canada's failure to identify specific measures as alternative measures under Article 5.6, and to identify specific SPS provisions claimed against in its supplementary submission. Australia's concerns relate to Canada's burden of demonstrating a prima facie case, and Australia's opportunity to respond to claims made against it.

E. MEASURES TO COMPLY WITH THE RECOMMENDATIONS AND RULINGS OF THE DSB

Canada

4.54 Canada argues that the relevant AQPM¹⁷ sets out new policies to be applied by the Director of Quarantine (or his/her delegates) when considering, under section 70 of QP 1998, whether to grant an import permit under Section 43 of QP 1998. Unless and until these policies are fully implemented, there is no basis for Australia's contention that measures exist to implement the recommendations and rulings of the DSB.

4.55 Australia did not implement its new measures for non-salmonid finfish until 1 December 1999. It will not fully implement its new policies for live ornamental finfish until 2002.¹⁸ It has also become clear in discussions with Australian officials, that AQIS has not implemented a number of the steps required to give effect to its new policies for salmonids. For example, the new policy purports to currently allow the importation of non-"consumer-ready" product for processing at an approved Australian facility. In fact, however, AQIS has yet to finalize its criteria for granting the facility approvals.

4.56 Issuing policy statements does not constitute the implementation of new measures or compliance with the recommendations and rulings of the DSB. There is a fundamental difference between a Member stating its intention to implement, as it is required to do within 30 days after the adoption of the panel or Appellate Body Report and actually fulfilling that intention within the reasonable period of time as determined by Article 21.3 of the DSU.

4.57 Australia's non-compliance, or lack of measures, is particularly glaring in the case of live ornamental finfish. Australia's lack of measures with respect to live ornamental finfish played an important part in the DSB's rulings that Australia is acting inconsistently with Article 5.5 and by extension Article 2.3 of the SPS Agreement. Even next year, not all of Australia's requirements for live ornamental finfish will be in force. Disease testing requirements for goldfish will not fully take effect until the beginning of 2002.

4.58 If, as Australia claims, all of its requirements are necessary to achieve its appropriate level of protection, then it will not achieve that level of protection until 2002. Until then, it will not have complied with the recommendations and rulings of the DSB. Australia knew this on 19 July 1999. Australia seems to believe that it may unilaterally grant itself months or even years of additional time for compliance, regardless of the 6 July 1999 deadline imposed by the arbitrator.¹⁹ Australia can determine how it will implement, but it cannot determine when.

4.59 Furthermore, not all of the requirements Australia imposes are necessarily even listed in the AQPMs. Thus, when Australia approved a health certificate for the importation of Canadian farmed salmonids on 26 November, it required a declaration for Atlantic salmon and rainbow trout that the fish do not come from a farm infected with ISAV. This was specified in AQPM 1999/69 and was addressed in the 1999 Report.²⁰ However, Australia has also imposed a requirement that the fish do not come from waters within 10 kilometers or one so-called tidal interchange of an infected farm, whichever is greater. This is specified neither in the AQPMs nor is it mentioned in the 1999 Report.

4.60 Australia appears to ascribe sinister motives to Canada's delay in even applying for certification for farmed salmonids. However, there was no bad faith on Canada's part. Under Canadian law, wild salmon fall within federal jurisdiction whereas farmed salmon fall jointly within federal and provincial jurisdiction. Thus, Canada could seek certification for farmed salmonids only after consulting with its provincial governments.

Australia

4.61 Australia submitted that it had taken the following actions in response to the recommendations and rulings of the DSB in relation to the provisions of the SPS Agreement.

4.62 With respect to the finding that the quarantine import prohibition on fresh chilled or frozen salmon was being maintained without a proper risk assessment (Article 5.1 and by implication Article 2.2), a risk assessment was undertaken on fresh chilled or frozen salmon from Canada as part of a generic Import Risk Analysis (IRA) on non-viable salmonid products and other non-viable marine finfish. The 1999 IRA process identified certain diseases in salmon of Canadian origin. These diseases were considered in the context of Australia's appropriate level of protection (ALOP), with the list of diseases that warranted consideration of risk management measures narrowed, at a baseline of evisceration (the level at which the product is internationally traded), to six diseases.

4.63 Risk management measures were evaluated on a disease-by-disease basis and measures adopted commensurate with risk. Based on this risk analysis, the Director of Quarantine decided that quarantine prohibition on fresh chilled or frozen salmon should be removed. Broadly, fresh chilled or frozen Canadian salmon may now be imported in three ways: (1) product in consumer-ready form; (2) product for processing; and (3) product which meet equivalent approaches to managing risk.

4.64 The relevant seafood importers association was closely consulted with a view to ensuring the least trade restrictive approach.

4.65 Entry into Australia is subject to permit, in accordance with QP1998 as amended. The cost of an import permit is A$60 valid for multiple consignments for up to 2 years. Legal action was not required to implement the decision, which entered into effect on 19 July 1999.

4.66 In order for commercial trade to flow, exports of consumer-ready product must be accompanied by certification in a form approved by AQIS. An import permit will not be granted absent such certification. AQIS has approved certification from Canada, New Zealand and the US.

4.67 An ABARE report-in-progress assessed that Canada would be more competitive against domestic product and other importers in the supply of frozen salmon, than for fresh or chilled salmon.

4.68 With respect to the finding that there were arbitrary or unjustifiable distinctions in the levels of protection considered to be appropriate in different situations (between fresh chilled or frozen salmon on the one hand and on the other hand whole frozen herring for use as bait and live ornamental finfish) which resulted in a disguised restriction on international trade (Article 5.5 and second sentence Article 2.3), in addition to the measures applying to the salmon product based on a risk assessment, risk assessments were undertaken, inter alia , on the disease risks associated with whole frozen herring for use as bait and on the disease risks associated with live ornamental finfish. The risk assessments formed part of IRA's on non-viable salmonids and other non-viable marine finfish, and on live ornamental finfish that were undertaken in parallel. The risk assessment on live ornamental finfish was limited to those imported fish that are not prohibited entry by Australia's environmental laws.

4.69 The measures applying to salmonids entered into effect on 19 July 1999. The measures applying to non-viable marine finfish other than salmonids have been in effect from 1 December 1999. The measures on ornamental finfish are being progressively introduced from 1 December 1999.

4.70 As a result of the IRA process, restrictions on imports of whole, frozen herring for use as bait were significantly tightened. The Director of Quarantine decided that non-viable herring of the genus Clupea will generally be prohibited entry.

4.71 AQPM 1999/79 sets out the measures applying to non-viable non-salmonid finfish. Different measures apply to: consumer-ready product (Part A); all fish product of New Zealand origin (Part B); head-off eviscerated fish or fish further processed products not meeting criteria in Part A (Part C); and all other non-viable non-salmonid finfish (Part D). For herring bait and feed fish, end users have the option of importing against the standard conditions set out in Parts A, B or C, or of importing under permit subject to the conditions set out in Part D.

4.72 Whole round herring for use as bait may be allowed on the basis of importers making a scientific submission to AQIS. AQIS will then make an informed assessment of the quarantine risk presented by the proposal. Information in the submission must include:

• details of the product to be imported;
• waters in which the fish were farmed (if appropriate) and caught;
• intended end-use of the fish; and
• details of pre-export and post-arrival management that would mitigate quarantine risks associated with importation.

4.73 Legal action was required for the application of these measures. This is embodied in the Quarantine Proclamation 1998 as amended. The date of effect of the amendments was 1 December 1999.

4.74 With regard to live ornamental finfish, as many species of live ornamental finfish are prohibited entry to Australia by virtue of the Wildlife Protection (Regulation of Exports and Imports) Act 1982, the risk assessment on live ornamental finfish was limited to those species which are specifically permitted entry under Schedule 6 of that Act.

4.75 Live ornamental finfish represent a special case in terms of disease prevalence and risk management. There is only limited scope for generic risk management comparable to arrangements applying to non-viable finfish. Diseases of live ornamental finfish may be localised, in many instances at premises level and disease status may alter rapidly. Diseases are also often species-specific. Risk management options for live fish are different from that for non-viable fish (e.g. evisceration is not an option for a product whose commercial value is solely as a live fish). The effectiveness of measures is an important consideration. For live ornamentals, certification and permit systems make a significant contribution to risk management as well as other measures such as quarantine withholding periods and visual inspection.

4.76 The pre-arrival and post-arrival conditions presently attached to freshwater ornamental finfish are already rigorous and certification/permit conditions are far more onerous than those attached to the import of non-viable finfish. Simplistic and crude product comparisons of certification or product specification requirements, including on a numerical basis, are not scientifically sound and are not an appropriate formula for assessing the relative effectiveness of risk management between different products.

4.77 The IRA on live ornamental finfish identified four diseases in common between Canadian salmon and live ornamental finfish:

• Aeromonas salmonicida (typical) - goldfish only;
• Aeromonas salmonicida (atypical) - goldfish only;
• Yersinia ruckeri (Hagerman strain) - single finding in goldfish; and
• ENV - marine ornamental species only.

4.78 On 19 July 1999, the Director of Quarantine decided that additional quarantine measures should apply to the import of live ornamental finfish. The measures are generic in policy terms, but will have specific application in regard to species and disease prevalence (including at premises level). The additional measures, together with the necessary administrative arrangements, are being progressively introduced.

4.79 These arrangements, which reach down to the level of agent-specific certification and sampling, require a detailed sub-set of administrative procedures and practices, including in relation to approvals for private quarantine facilities and quarantine security. All administrative arrangements must be in accordance with Australia's domestic legal framework, including in regard to the delegated authority of individual AQIS officers. Other than in regard to live marine finfish (for which legislative amendment was required), no legal action was required to implement these measures. Legal authority for the delegation of quarantine decision-making is contained in QP 1998.

4.80 Therefore, Australia argues, it is clear that Australia has implemented. The measures applying to salmon and other non-viable marine finfish are in force. A certificate for the import of Canadian salmon has been approved and an import permit granted. This certificate is irrefutable evidence that Australia has removed the import prohibition on fresh chilled or frozen salmon from Canada and that the measures as described are being applied to fresh chilled or frozen salmon from Canada. The additional measures applying to live ornamental finfish were progressively introduced from 1 December 1999.

4.81 In regard to approvals for processing plants, the applications must come from the operators of the plants concerned. AQIS will respond to each application as received and will decide whether to authorize the plant against the risk management policies decided by the Director of Quarantine. To date, no applications have been received.

4.82 From a trade effects perspective, it is noted that the different dates of application for the measures on salmonids and for the other two product groups involve immediate trade liberalisation for salmonids and transitional periods for a movement to quarantine measures with potentially more trade-restrictive effect applying to the other two product groups. Consistent with SPS obligations, AQIS has taken into account factors such as shipping times, commercial contractual arrangements and, in the case of live ornamental finfish, the need for sufficient lead times to adjust to more complex administrative arrangements that will apply, including the need for approved post-arrival quarantine premises. In this context, attention is drawn to paragraph 2 of Annex B of the SPS Agreement:

"Except in urgent circumstances, Members shall allow a reasonable interval between the publication of a sanitary or phytosanitary regulation and its entry into force in order to allow time for producers in exporting Members, and particularly in developing country Members, to adapt their products and methods of production to the requirements of the importing Member."

F. ARTICLE 5.1 OF THE SPS AGREEMENT

Canada

4.83 Canada argues that even if Australia has measures in existence, those measures are inconsistent with the SPS Agreement. The new policies that Australia has implemented or may implement under its various AQPMs are themselves inconsistent with the SPS Agreement. Australia's measures are not based on a risk assessment, contrary to Article 5.1

4.84 There are two elements to the obligation under Article 5.1 of the SPS Agreement. First, there must be a proper risk assessment or risk assessments on which it relies, within the meaning of Article 5.1 and Annex A, paragraph 4. Second, Australia's measures must be based on that risk assessment or those risk assessments.²¹ In the present case, a proper risk assessment does not exist to support Australia's new measures. Even if it did, Australia's measures are not based on it.

4.85 Australia's 1999 Report is an impressive survey of the scientific knowledge on certain fish diseases but it is not a risk assessment within the meaning of Article 5.1 and Annex A:4 of the SPS Agreement. As Dr. Wooldridge confirms, the 1999 Report neither evaluates the likelihood of the establishment or spread of the disease or disease agents of concern to Australia, nor does it evaluate the likelihood of entry, establishment or spread according to the measures which Australia might apply.²² Nothing in Australia's submissions refute this. Either failure is sufficient for the Panel to find that Australia maintains its new policies without a risk assessment, contrary to Article 5.1 and, by implication, Article 2.2 of the SPS Agreement.

Australia

4.86 Australia welcomes the clarifications provided by the experts of their views on the adequacy and appropriateness of the 1999 IRAs. The experts agreed on the integrity and transparency of the process. Australia engaged a large number of highly qualified aquatic science and risk management experts to ensure the integrity of the 1999 IRAs.

4.87 Dr. Br�ckner advised that Australia's import risk analysis on non-viable salmonids and non-salmonid marine finfish was "acceptable and scientifically justifiable". Dr. McVicar confirmed Australia's view about the limited applicability of quantitative risk analysis to regulatory decision-making on fish diseases generally and, in particular, the present case. Dr. McVicar concluded that he was satisfied with Australia's risk assessment and that the minor improvements that could be made would be unlikely to warrant any changes to Australia's conclusions.

4.88 In the original dispute, the Appellate Body considered that:

"A proper risk assessment of this type must evaluate the likelihood, i.e. the probability, of entry, establishment or spread of diseases and associated biological and economic consequences as well as the likelihood, i.e. probability, of entry, establishment or spread of diseases according to the SPS measures which might be applied".²³

This is the legal test before the Panel. It is not one of the individual views of experts according to a preferred technique or standard. The SPS Agreement is not prescriptive about detailed techniques and methodology of a risk assessment. It would not function as a legal instrument if its legal obligations were to be hostage to fashions in risk assessment. The legal test is not one of whether Australia should or could have undertaken a quantitative risk assessment. Indeed as Dr. Wooldridge advised, it was unknown whether there would have been a different outcome if the risk assessment had been conducted in the way that she prefers.

4.89 The SPS obligation in regard to "techniques" is that a WTO Member should take into account those techniques developed by the relevant international organisations. The structure of the 1999 IRA comprises three components: hazard identification, risk assessment and risk management. This conforms with the OIE guidelines on import risk analysis as stated in the International Aquatic Animal Health Code (1997) (the "OIE Code") and the OIE International Animal Health Code.

4.90 The following diseases of Canadian salmon, for which an OIE standard of evisceration exists, involve risk management based on that standard:

• Piscirickettsia salmonis: all fish;
• IPNV: all fish, except juveniles; and
• VHS: all fish.

Canada

4.91 Canada contends that the 1999 Report does not meet OIE standards. According to Dr. Wooldridge, who helped to draft the standards in the 1999 OIE Animal Health Code upon which the Aquatic Animal Health Code is based, the 1999 Report does not meet those standards. Canada agrees. In Canada's view, the 1999 Report does not meet the requirements of either the 1999 draft OIE Code or the 1997 OIE²⁴ Code because it does not assess risk.

4.92 As Dr. Wooldridge made clear, the 1999 Report appears superficially to be a risk assessment but closer scrutiny reveals its inadequacies. Thus, the 1999 Report appears to contain information or elements that are required by the OIE Code, such as exposure assessments and release assessments, but this information has not been used in a full and transparent manner that enables an outsider to see how Australia arrived at risk conclusions. Among other things, significant information has been "systematically left out" in the exposure assessment section, and the exposure and release assessments were not integrated.²⁵ Thus, the 1999 Report does not satisfy, among other things, the fourth step of risk assessment (risk estimation) in the draft 1999 OIE Code.

4.93 The 1999 Report also does not consider the probability that the disease agents of concern will complete the full range of steps to become established or spread in Australia. The steps are specifically identified in the risk assessment section of the OIE Code.²⁶

4.94 Canada notes that Australia argues that the proposed 1999 OIE Code makes clear that as a general principle, qualitative risk assessments are valid. This has never been in issue. The issue is whether the 1999 Report is a valid risk assessment. The problems that Canada and Dr. Wooldridge have identified in the 1999 Report relate not to whether Australia took a qualitative approach but the particular "methodology" Australia used. Thus, Canada's position that Australia could have relied more on the New Zealand risk assessments or the Vose Report was not intended as an argument that Australia necessarily needed to do a quantitative risk assessment. Instead, it goes to the issue of whether Australia used the best available information. The use of the best available information is also a principle of risk assessment identified in the 1999 OIE Code.²⁷ At least one of the peer reviewers of the 1999 Report repeatedly advised that it would be highly relevant for AQIS to consider the results of its qualitative analysis and the proposed risk reduction measures in the light of the Vose and New Zealand risk assessments.

4.95 Dr. Wooldridge also stated that, like Canada, she too does not understand why Australia has not attempted a quantitative assessment even though it is not a requirement of the SPS Agreement. Dr. McVicar states that quantitative risk analysis is severely constrained with fish diseases due to the lack of adequate data in key areas. However, Dr. Wooldridge has noted that, "there was generally much more data in existence for almost any quantitative risk assessment than at first sight seemed likely to be available".²⁸

4.96 Moreover, where data are unavailable for any given step in the risk pathway, a quantitative risk assessment can adopt the highly conservative approach of assuming a 100 per cent probability that that step will be completed. The Vose Report took precisely this highly conservative approach yet still concluded that the likelihood of the entry, establishment or spread of A. salmonicida and Renibacterium salmoninarum into Australia through the importation of Canadian, wild, ocean-caught Pacific salmon for human consumption was negligible.

4.97 Canada also notes that both the 1997 OIE Code and the proposed 1999 version of the OIE Code clearly emphasize the importance of commodity volume in assessing risk. Thus, the 1999 version considers it a principle of risk assessment that: "risk increases with increasing volume of commodity imported".²⁹ The 1999 Report ignores the effects of volume on risk, particularly in the case of bait and feed fish. Australia therefore cannot claim that the 1999 Report meets the standards of the OIE Code.

1. Evaluation of the likelihood of entry, establishment or spread of diseases and the associated consequences

Canada

4.98 The 1999 Report has accorded unjustifiable subjective weightings to the scientific evidence before it and has seemingly ignored its own conclusions as to the likelihood that specific sub-events would occur. The result is that its unrestricted risk estimates are inexplicably weighted against eviscerated salmon product. The 1999 Report cannot therefore be said to have adequately evaluated the likelihood of entry, establishment or spread of disease and the associated consequences as it is required to do under Article 5.1.

4.99 Although the probability terms used are defined in the 1999 Report, it is difficult if not impossible to understand the basis on which Australia applies these terms to event probabilities. This subjectivity problem is apparent in the absolute probabilities (e.g. "low", "moderate", "very low", "negligible") that Australia has chosen to assign to the likelihood of events occurring in respect of individual disease agents. It is also a problem in the relative probabilities Australia has assigned among disease agents or products. These failings are endemic to the document.

4.100 For example, Australia has found that (a) the incidence of pathogens entering the aquatic environment via human consumption of imported salmonids would be "extremely low"; (b) the titre of A. salmonicida in eviscerated imported product would be "extremely low"; (c) solid-waste and liquid waste disposal pathways would reduce these already extremely low titres by "orders of magnitude"; (d) other pathways are essentially irrelevant; (e) exposure even to a "low" titre of A. salmonicida would need to be maintained for a prolonged period for infection to result; and (f) there is an absence of significant salmonid populations in most of Australia. It is therefore inexplicable that Australia would conclude, as it does, that the probability of susceptible fish being exposed to an infectious dose of typical A. salmonicida would be merely "low".³⁰

4.101 Dr. McVicar's subsequent comments on A. salmonicida reflect his views on the likelihood of the entry of the disease agent into Australia. However, they do not address Canada's specific concern, set out in paragraphs 58 to 61 of Canada's First Submission, regarding Australia's estimate of the likelihood of disease establishment. In arriving at an estimate of the likelihood of disease establishment, Australia appears to have ignored its own conclusions regarding other steps in the pathway that are necessary preconditions for disease establishment.

4.102 There are similar discrepancies in the conclusions the 1999 Report reaches regarding the consequences of disease establishment. For example, Australia acknowledges that Australian salmonids are "routinely vaccinated to prevent disease due to Vibrio anguillarum" and that the consequences of the establishment of Vibrio salmonicida (V. salmonicida) could be mitigated by similar means. In the case of A. salmonicida , the Report notes the advice of Dr. McVicar that successful control methods, including vaccination, have significantly reduced the importance of A. salmonicida in Scotland.³¹ It finds however, that in addition to increased costs, Australia's "''disease and chemical residue free' image could also be harmed". The Draft fails to reconcile this "image" with the reality of its control programme for Vibrio anguillarum (V. anguillarum). The result is that it overstates the consequences of A. salmonicida introduction.

4.103 All risk assessments are to some degree subjective. Whether subjectivity will be an actual problem in a qualitative risk assessment will depend on how the qualitative terms are used. A qualitative risk assessment will not be too subjective when, at a minimum, it is sufficiently transparent that each conclusion follows logically from those that came before it and when one can have reasonable confidence in the levels of risk assigned. The 1999 Report fails on both counts. Dr. Wooldridge agrees. The 1999 Report failed to develop conclusions that follow logically from the sequence of previous conclusions, and systematically excluded the information regarding the pathways for disease release into the aquatic environment from its exposure assessment for each disease agent. It therefore did not evaluate likelihood of the entry, establishment or spread of the disease agents of concern.

4.104 Nothing in Australia's rebuttal identifies where the 1999 Report has evaluated the likelihood of the establishment or spread of the diseases of concern, rather than simply stating conclusions. The best that Australia can show in its rebuttal is that the 1999 Report determined that wastewater from processing plants may contain a higher concentration of pathogens and that 3. Diseases of imported salmonids versus domestic fish that wastewater bypassed sewerage systems or were directly discharged into waterways without treatment such pathogens could perhaps enter the aquatic environment in significant quantity.³² Or it is speculated that significant high level exposure of susceptible fish to a significant titre of IHNV from, for example, regular discharge of untreated effluent from a salmon processing plant, could, possibly, result in the establishment of infection.³³

4.105 One will search in vain for an evaluation of the likelihood or probability that any of these events will occur. As with the 1996 Final Report, the 1999 Report focuses in this case on possibility, not probability. For example, the 1999 Report does not say what is the likelihood that processing plants will contain a higher concentration of the pathogens of concern, or that these pathogens will then regularly be discharged directly into the aquatic environment without treatment, or that a susceptible host will be present.

4.106 Nor does the 1999 Report consider the probability of the full range of steps that must be completed in order for a pathogen of concern to become established or spread in Australia. The steps are specifically identified in the risk assessment section of the OIE Code.³⁴ Yet the 1999 Report does not evaluate these probabilities for the disease agents of concern.

4.107 By contrast, the Introduction to the 1999 Report does reach conclusions on the establishment and spread of aquatic disease agents generally. It concludes that the incidence of pathogens even entering the aquatic environment via the human consumption of imported salmonids or marine finfish would be extremely low³⁵; that wastewater would greatly dilute and reduce the loads of any pathogens present; and that the probability of pathogens even entering the aquatic environment via a solid waste pathway would be very low. The Introduction dismisses other pathways as not significantly increasing the probability in the risk analysis overall.³⁶

4.108 As Dr. Wooldridge notes, the foregoing information and the conclusions drawn from it do not appear to have been taken into account in the individual exposure assessments for specific diseases in salmonids. Had Australia done so, it might well have concluded in the 1999 Report that for each disease, the overall probability of even aquatic exposure to salmonid product was exceptionally low at the highest.

4.109 Dr. McVicar states that the 1999 Report has made a valid assessment of the probability of the establishment or spread of each disease of concern in Australia. However, he does not explain how or why he reaches this conclusion. He states, for example, that "the level of viable infectious agents likely to be remaining in gutted carcasses in the parts usually removed and disposed of before human consumption was considered by Australia to warrant additional safeguards". However, he does not address Dr. Wooldridge's concern that Australia has not attempted to specify at what levels these infectious agents would remain. In addition, as Dr. Wooldridge notes, even if imported salmonid product is infected, the Report identifies the very low probabilities such product will enter the aquatic environment but does not consider this in reaching its conclusions on establishment or spread. Nothing in Dr. McVicar's response contradicts Dr. Wooldridge's finding.

Australia

4.110 The 1999 IRA evaluated the quarantine risk with respect to those specific disease agents identified in hazard identification as warranting further consideration. Chapter 4 assessed the unrestricted risk posed by each disease agent in accordance with the OIE standard: (1) release assessment, (2) exposure assessment, and (3) consequences assessment. A risk evaluation matrix was used to determine whether the quarantine risk posed for each disease agent met Australia's ALOP.

4.111 The 1999 IRA considers the factors relevant to the release assessment, including disease prevalence and titre of disease agent. Relevant factors considered were in accordance with the OIE Code:

• the probability of fish being infected with a disease agent in the exporting country;
• the probability of the infective agent being present in particular tissues imported; and
• the probability of the infective agent surviving pre-import conditions or treatment.

The 1999 IRA considered a broad range of salmonid species, countries of origin, production systems, disease agents, types of salmonid product and methods of treatment. Quantitative data on many of these factors is lacking. The risk analysis was based largely (but not exclusively) on qualitative information.

4.112 The 1999 IRA recognises the importance and influence of the many variables to the release assessment in the general discussion (pages 14-23) and in the risk assessment of specific disease agents in Chapter 4. For each disease agent, the key findings include information on prevalence and the distribution of disease agents in various tissues. This influences the extent to which risk management measures would reduce the probability of an agent occurring in imported product.

4.113 Canada is seeking to imply that the validity of a qualitative risk assessment depends on whether it was possible to conduct a quantitative risk assessment. Canada claims that the 1999 IRA's evaluation was "highly subjective" and the "reasonableness" of Australia's probability assignment goes to the validity of qualitative vs quantitative risk assessments.

4.114 In her responses to Questions 1 and 2, Dr. Wooldridge appears to be suggesting that a qualitative risk assessment, which compares risk in different diseases, cannot be valid unless accompanied by quantitative analysis. By definition, a qualitative risk analysis will use qualitative terms in assessing risk. Dr. Wooldridge's comments go towards the validity of qualitative, as compared to quantitative, risk assessments.

4.115 Canada's and Dr. Wooldridge's contentions are not supported by international and WTO practice. International practice, as reflected in the OIE International Animal Health Code (1999), affirms the equal validity of qualitative and quantitative methods. In addition, the Panel and Appellate Body in the current dispute stated that likelihood may be expressed either quantitatively or qualitatively. Neither are Canada's and Dr. Wooldridge's contentions supported by Dr. Wooldridge's own evidence to the original Panel and the advice of Dr. McVicar and Dr. Br�ckner.

4.116 Consistent with the OIE Code, Australia conducted the 1999 IRA on a qualitative basis. In the absence of definitive, quantitative data on factors relevant to quarantine risk, AQIS applied appropriately conservative professional judgment based on available scientific information and the advice of experts in relevant fields. A quantitative approach would not have provided a more objective evaluation of risk as expert judgment would still be required to address data gaps.

4.117 Canada's assertions that prevalence is not translated into risk for A. salmonicida (typical) and of anomalous relative probabilities in regard to IPN and ISA, is rejected by the scientific basis of the 1999 IRA process. This is set out in the hazard identification, risk assessment and risk management Chapters of the 1999 IRA. The risk assessment covering Canadian salmon does not address viral encephalopathy and retinopathy virus (VERV) which is not a disease agent reported in salmonids; nor is it an exotic disease to Australia.

4.118 The risk from V. salmonicida was not considered to warrant specific risk management measures. V. anguillarum is not an exotic disease and therefore outside the scope of the risk assessment. In addition, both V. anguillarum and V. salmonicida may be managed through the use of bath vaccinations. The efficacy of vaccination is high and protection enduring. A. salmonicida is managed with oil adjuvent vaccines which must be individually injected into the body cavity of each fish. This vaccine has adverse effects including the development of lesions in the carcass, increased cost-of-production and reduced growth rates. While the efficacy is good, it is not as high as for V. anguillarum and V. salmonicida vaccines, nor is protection as enduring. Recognised differences in the administration, efficacy and adverse effects of vaccination between V. anguillarum and A. salmonicida were taken into account in the risk assessment.

4.119 Consideration of previous import volumes and time periods is not a requirement of a qualitative risk assessment. The absence of disease incidents associated with frequent large volume commodity imports does not indicate that the material can pose no risk. It only indicates that the import has had a low risk in relation to the specific materials and conditions of the previous importations.

4.120 In summary, the 1999 IRA evaluates the likelihood i.e. probability, of entry, establishment or spread of the diseases of concern, and the associated consequences. The use of qualitative terms in assessing likelihood is consistent with a qualitative risk assessment and is valid with regard to the scientific analysis.