Implications of the Japanese Overcatch of Southern Bluefin Tuna for Data Collection and Assessments of Tropical Tuna

IOTC-2007-WPTT-INF07 Implications of the Japanese Overcatch of Southern Bluefin Tuna for Data Collection and Assessments of Tropical Tuna T Polacheck C Davies

Implications of the Japanese Overcatch of Southern Bluefin Tuna for Data Collection and Assessments of Tropical Tuna Tom Polacheck 1 and Campbell Davies CSIRO Marine and Atmospheric Research GP Box 1538 Hobart, Australia Abstract Japanese longline data collected from vessel reported log book data constitute a central component in most of the stock assessments for the world s major tropical tuna and billfish fisheries (e.g. yellowfin tuna, bigeye tuna and swordfish). A review of Japanese market statistics was undertaken in 2006 by Australia and Japan in relationship to catches of southern bluefin tuna (SBT). The Market Review found that very substantial and continuous IUU SBT catches had been occurring by Japanese longline vessels since at least the early 1990s. Over the period 1990-2005, Japanese catches were estimated to have exceeded officially reported catches by at least a factor of 2. This paper explores the potential implications of these large IUU catches in terms of other tuna and billfish catch and effort data from Japanese longline vessels and for stock assessments that are dependent upon them. The large SBT IUU catches may have resulted in the misreporting of catches as other species and misreporting of the location of fishing effort both of which could bias CPUE indices and the stock assessments for other tuna. The large SBT overcatches and IUU fishing highlights the problems and risks of relying predominately on data from vessel supplied logbooks for conducting stock assessments. Concerns about the lack of verification of logbook data need to be raised as a matter of priority. There is an urgent need for the scientific fishery community to be more pro-active in the development and implementation of independent ways to monitor and verify catches and fishing effort (e.g. scientific observers, video monitoring, port sampling, etc,). Introduction Japanese longline data collected from vessel reported log books constitute a central component in most of the stock assessment for the world s major tropical tuna and billfish fisheries (e.g. yellowfin tuna, bigeye tuna and swordfish) conducted by regional fishery management organizations (RFMO) (i.e. ICCAT, IATTC, WPFC, IOTC). In particular, the catch per unit effort (CPUE) indices generated from these Japanese logbook data are either the sole or one of the principle measures of relative abundance utilized in these assessments. The large reliance on of these logbook data stems from: 1. The logbook data provide a long time series of consistently recorded catch and effort data often stemming back to the start of any significant commercial catches; 2. The logbook data provide both wide-spread geographic and seasonal coverage throughout the major ocean basins; 3. The Japanese fleet has generally been one of the major harvester of the species being assessed at least for the longline component of the fishery; 1 Corresponding author tom.polacheck@csiro.au 1

4. The difficulties of obtaining fishery independent relative abundance indices for these large wide-ranging pelagic species has necessitated a reliance on fishery dependent CPUE in the stock assessments; and 5. A perception has existed that there is a high degree of accuracy and reliability in the reporting of the catches and fishing effort by Japanese fishermen. These five factors have also been important in the use of these data in recent debates on the worldwide status of large pelagic fish resources (Myers and Worms 2003; Walters 2003, Hampton et al 2005, Polacheck 2006, Sibert et al 2006). While a perception of accuracy has surrounded these Japanese data, little or no information actually exists to evaluate whether in fact this is the case. Little actual verification (e.g. reliable monitoring and cross checking of actual landing) appears to have actually been undertaken. In 2006, a review of Japanese market statistics was undertaken by Australia and Japan in relationship to catches of southern bluefin tuna (SBT). This review revealed that there has been very substantial and continuous unreported longline overcatches of SBT since at least the early 1990s both in the officially reported Japanese catch statistics and in the associated log book data used for stock assessment purposes (Anon. 2006a, b, c, d; Polacheck et al 2006) 2.While the Market Review was confined to an examination of the amount of SBT being sold in Japan, the existence of these large overcatches of SBT has potentially wider and larger implications for the reliability of Japanese tuna longline catch and effort data and for stock assessments that are dependent upon these them. The purpose of the current paper is to examine some of these implications. Background on Overcatch SBT is a long lived, mobile, highly valued fish found throughout most of the southern temperate oceans except in the more easterly regions of the South Pacific. Surface and longline commercial fisheries for SBT began in the 1950s and the stock has been very heavily fished (Caton 1991). A major component of the surface fishery (that off the southeast coast of Australia) collapsed in the late 1970s and tagging studies indicated very high exploitation rates on juveniles in the 1980s. In response, beginning in 1984, Australia markedly reduced its catch of juveniles in the surface fishery. Informal international management arrangements involving Australia, Japan and New Zealand were initiated in the early 1980s and were subsequently formalised with the establishment of the Commission for the Conservation of Southern Bluefin Tuna (CCSBT) in 1993. Total allowable catch limits (TAC) were introduced in 1985 and were divided into national allocations. The TACs were progressively lowered with a major reduction of around 50% for the 1989 fishing year. The early limits resulted in reductions in catches from the surface fisheries. However, it was not until the 2 The actual review report of the Japanese market has not been placed in the public domain by the Commission for the Conservation of Southern Bluefin Tuna (CCSBT) which is the international regional fisheries management organization for SBT and for which the review was presented. As such the actual Market Review report is considered to be confidential at this time and can not be directly referenced. However, numerous reports and papers which document and summarize the results of the review are publicly available from the CCSBT. This paper is based on the information in these publicly available documents. 2

1989 fishing year that the catch limits became restrictive for the Japanese longline fleet (i.e. the Japanese longline fishery reported that it was not able to catch its limit prior to this year). From 1989, catch limits essentially remained fixed until 2006 3. Concerns about potential unreported catches for SBT began to be raised as early as 1990 i.e. shortly after the agreed catch limits for SBT became restrictive on the Japanese longline fleet (Anon., 1990, 1991; Polacheck and Klaer, 1991). While a primary focus of these early concerns was on potential increase catches by non-parties to the informal tri-lateral and subsequent CCSBT convention, there was also concerns about the financial incentive that restrictive quotas generate for under-reporting of catches. Direct evidence became available in 1996 which demonstrated that IUU fishing was in fact occurring by Japanese vessels. In December of that year at least 50 longline vessels were detected by an Australian surveillance flight fishing in a prime SBT fishing area in the southeast Indian Ocean. The area had been officially closed to Japanese fishing by the Government of Japan because their SBT quota had been caught. Forty vessels were identified as Japanese, several as Korean 4 and the rest were unidentified (Anon. 1997a). This surveillance flight appears to have been a once off-event and in any case had only a limited spatial coverage. As such, it did not provide any direct measure of the extent of illegal fishing, the magnitude of the SBT overcatch and the number of years it had been taking place. The CCSBT did not undertake any follow-up procedures to improve monitoring or to improve compliance, although Japan did institute some additional regulations on its fleet (Anon. 1997a). The CCSBT relied on Japan s assurance that this was not symptomatic of larger scale IUU fishing by its fleet and that they would instituted sufficient measures to prevent further out of season fishing. One consequence of data collected by the surveillance flight was recognition and agreement within the CCSBT Scientific Committee that uncertainty existed about the magnitude of the actual catches. In fact, the CCSBT Scientific Committee agreed that the actual level of SBT catches was a major source of uncertainty that needed to be addressed when conducting SBT stock assessments (Anon. 1997b). However, consensus was never able to be reached on any specific overcatch scenarios to include in the stock assessment process 5 and this lack of consensus was one of the reasons that prevented agreements within the Scientific Committee on the stock assessments in the late 1990s. 3 The catch levels set for Australia, New Zealand and Japan by CCSBT remained at their 1989 levels except for years 1998-2003 in which the CCSBT was unable to agree on a global TAC. In these years, the three parties voluntary agreed to keep their catches at their past level except in 1998 and 1999 when Japan unilateral increased its catches to undertake an experimental fishing program (Polacheck, 2002). Also, the CCSBT increased its global TAC in 2003 to accommodate new member catches and to acknowledge the existence of non-member catches (Anon. 2003). 4 Note that Korea was not a member of the CCSBT at this time. 5 Beginning in 1996, stock assessments conducted by Australia contained a range of catch scenarios including ones which allowed for catches above those officially reported to provide an indication how uncertainty in catches affected the results (e.g. Polacheck et al 1996, 1998, 2001; Polacheck and Preece, 2001). Stock assessments conducted by Japanese scientists have only used officially reported catches. 3

Beginning around 1997, Australian Industry informally suggested that the Japanese market statistics indicated substantial overcatches of SBT. CCSBT Scientists encouraged the Australian industry to table actual analyses of demonstrating this (e.g. Anon. 1997c). However, it was only in 2000 that a written document was provided in which it was asserted that analyses of market statistic showed about 100% overcatch by Japan (Jeffries, 2000). However, the document provides little details and its author subsequently repudiated the results stating upon further consideration that he considered the market data unreliable (Anon 2000). Nevertheless, Australian industry continued to verbally express concerns about large SBT overcatch based on their knowledge and perception of the Japanese markets and statistics from them. In 2005, these concerns along with actual documented sources and analyses provided by industry, resulted in Australia undertaking independent analyses of market data. Based on these analyse, Australia tabled documents at the CCSBT Scientific Committee and Commission meetings which suggested that the Japanese market data demonstrated that there had been substantially more SBT sold in the Japanese wholesale fish markets than could be accounted for in official catch statistics for an extended number of years (Australia, 2005a 6 ). The papers tabled by Australia resulted in Australia and Japan agreeing to undertake an independent review to be reported back to the Commission of the Japanese SBT market data anomalies. The purpose of the review was to determine whether or not over-catching had been occurring relative to the total allowable catch, and if so, over what period and what was it the source (Anon 2005a) 7. Overcatch Based on Market Statistics Japan is the primary market for SBT with a very high percentage of the world catches consumed as sashimi within Japan. A large fraction of the SBT in Japan is sold as whole fish through auctions at more than 16 wholesale markets, with the Tokyo Tsukiji market being the largest. Statistical data on the amount of SBT sold at auction are available from a large number of these. However, access and interpretation of the data are not straightforward. The marketing and distributional system is complex (Williams 1986). Interpretation of these market statistics in terms of catches by Japanese longliners is confounded by a number of factors including: 6 The paper tabled by Australia at the scientific committee was produced and presented during the meeting (Anon. 2005b). As such, under the rules of procedures that the Scientific Committee operates, it was not a formal meeting documented and has no official status (i.e. it is not publicly available from the CCSBT Secretariat). However, the paper was formally presented to the Scientific Committee, there were extensive discussions about it and the results within it were used to form part of the agreed conclusions from the meeting (see paragraphs 40 and 49 to 53 in the report). In spite of reference to the working paper in the report, at the time of adoption of the final report, Japan refused for the working paper and/or that portion of it that was extensively discussed to be included as an attachment to the report. As such, it was not included and is a phantom document (see Polacheck et al 2006 for further details and which also contains a copy of a portion of this document). 7 Concurrent with the review of the Japanese market data, it was also agreed to undertake an independent review of possible Australian SBT farming operation anomalies at Port Lincoln to determine whether or not over-catching was occurring in this fishery relative to the total allowable catch (Anon. 2005b, JFA 2005). As the anomalies in SBT farming operations have no direct implications for tropical tuna assessment, this review is not discussed here (see Anon 2006a for more details on this review) 4

1. Only a fraction of the frozen wild whole tuna sold actually goes through these wholesale markets. There are direct sales of tuna that by-pass the wholesale market system (e.g. bulk purchases by supermarket chains); 2. The complex distribution and market systems results in some tuna being included in the sales statistics for more then one wholesale market (i.e. double counting exists in the simple sum of the total SBT sold across all markets); 3. A fraction of the wild caught frozen imported longline catches from other countries (e.g. Korea and Taiwan) are also sold in the wholesale markets; 4. A large fraction of the farmed SBT caught by the Australian surface fishery are frozen and a portion of these are sold in the wholesale market auctions; 5. There can be a substantial lag between the time when an SBT was caught and when it is actually sold as a result of the long length of longline cruises and because tuna can be held in frozen storage for extended periods (several years) depending on current prices and overall availability of competing product. The Market Review compiled statistical sales data for SBT and collected information with respect to the first four of the confounding factors listed above. Based on their investigation, the Market Review concluded that there had been substantial unreported catches of longline caught SBT relative to the officially reported catches to the CCSBT. The review provided two specific scenarios for their best estimate of the annual sale of SBT from 1985 through 2005. The two scenarios represented alternative views within the Market Review team on the level of double counting that was represented in the wholesale sales statistics. The Market Review did not attempt to account for the lag between when the catches were caught and when they were sold. Nevertheless, when comparing the cumulative total reported Japanese longline catches over this period and the estimates of the told longline SBT sold at market discounted for non-japanese catches from the review there is clearly a very large discrepancy (e.g ~178,000 tonnes above Japan s national allocation of ~133,000 tonnes for the period 1985-2005 or an estimated total catch of ~2.4 times above their national allocations, Hurry, 2006). The Market Review identified four potential sources for the overcatches (Polacheck et al 2006): 1. Non-SBT registered Japanese vessels illegally catching SBT and unloading as different tuna without being inspected 2. SBT registered Japanese vessels under report SBT and unloading undeclared SBT as different tuna without being inspected 3. Import wild SBT caught by foreign vessels and imported to Japan via freezer boat or container 4. Transshipment of SBT caught by foreign vessels and delivered via freezer boat as domestic SBT and then unloaded as different tuna. These four sources are not mutually exclusive. All four of them could have been simultaneous sources contributing to the overcatch and their relative 5

contribution could have varied over time. However, subsequent discussion and decisions within the CCSBT indicate that the third and fourth sources are not seen as likely (e.g. Japan has acknowledged that substantial overcatch occurred in at least the most recent year. Additionally, only Japan has been asked and agreed to reduce it catches below its previous allocations in response to the Market Review (Anon. 2006a). The working assumption both in the CCSBT and in the current paper is that most if not all of the overcatch is the result of IUU fishing by Japanese longliners. Polacheck el al (2006) and the CCSBT Stock Assessment Group (SAG, Anon 2006b) 8 developed estimates of the annual longline catch based on the two market sales scenarios in the Market Review. The estimates are based on the assumption that the catch caught in any given year was sold over the subsequent next two year in an approximate 70/30% split 9. Based on this assumption for the lag between catch and time of sales, the market statistic show that there has been substantial overcatches since 1990 relative to Japan s reported catches (i.e. >100% in total) and over 200% in some years irrespective of which of the Reviews scenario is used (Figures 1-2). Prior to 1989 and the time when catch limits were not restrictive in terms of reported catches, the market statistics suggest small or no net ovcrcatches from 1984-1988 (i.e. -3% or 7% depending upon the scenario considered). Direct Implications for Tropical Tuna Catch Statistics and CPUE Analyses At first consideration, a substantive direct link between SBT overcatches and reported Japanese longline catches or CPUE for tropical tuna might seem unlikely given the spatial separation between the fisheries. However, 30% or more of the Japanese longline fleet is reported to have participated in the SBT fishery in every year since 1983 (Table 1). These vessels report that they are fishing for SBT for only part of the year. During the rest of the year, they report changing fishing grounds. As such, depending upon how the actual overcatches were taken and how the catches and effort associated with these were reported in the log books from the vessels involved, the catch and catch rate data for other tuna species could be greatly affected. The large overcatches could have stemmed from an underreporting of the actual SBT catches while vessels were legally fishing for SBT (i.e. by longliners authorized to fish for SBT during the Japanese official season). In this case, the vessels may simply have decided not to report anything in their logbooks. However, given the magnitude of the overcatch, they may have decided to report at least a fraction of the overcatch as other species (e.g. bigeye) so as to avoid a large discrepancy between the quantity of fish being unloaded and the 8 The SAG is a technical group under the CCSBT Scientific Committee that undertakes and reviews the technical and analytical aspects of the SBT stock assessments. 9 The overcatch estimates in Anon (2006b) and Polacheck et al (2006) are based on different assumptions about how to account for the assumed 70/30% lag in the sales relative to the time of catch. Anon. (2006b) assumed that 70% of the catch caught in a year was sold in the following year and the remaining 30% in the following year. This approach requires assuming that no overcatch existed in a known year in the past to be solvable. Polacheck et al (2006) assumed that the total sales in the market in a given year were comprised of 70% of fish caught the previous year and 30% caught two years prior. The approaches yield similar results and results from the second approach are presented here. 6

quantity reported in the log book (mis-reporting of total catch would be easier to detect then the mis-reporting of the species composition). Alternatively, large illegal catches of SBT could have occurred as the result of vessels fishing in areas and time period when fishing for SBT was closed. In this case, both locations and actual catches are likely to have been misreported (e.g. it would be highly suspicious if a vessel were to report long periods of non-fishing or long periods of fishing with no catch). There is also the possibility that a substantial fraction of the SBT overcatch was taken by vessels fishing legally outside of the SBT regulated fishing areas but mis-reporting SBT catches as other species. This situation can arise because there are large areas where SBT are known to occur but not defined to be within the seasonally regulated areas used by Japan for controlling their SBT catches. Some of these areas (particularly in the Indian Ocean) historically have had large catches of SBT and have never been closed by Japan to fishing by its longline vessels - both those with authorization to fish for SBT and the remainder of the fleet (see section below on Area 2 for a specific example and also Polacheck et al 2006 for more detail). Observer Data The only way that the overcatch of SBT is unlikely to have had no affect on the reported catch and/or effort for other tuna species, if all of the overcatch was the result of under-reporting of SBT catches during targeted fishing for SBT or nonreporting of SBT by-catch while fishing for other species while accurately reporting of effort and the catches of all other species. The latter seems highly unlikely given the magnitude of the overcatch. In terms of the former, this would require that vessels were achieving catch rates 2-3 times greater then those reported in some years. Polacheck et al (2006) examined available observer data to see if there was evidence for large discrepancies between reported log book SBT catch rates and those directly observed. There are three sources of available observer data for Japanese longline vessels fishing for SBT: (1) Australian observers on vessels fishing with in the Australian Fishing Zone; (2) New Zealand observers on vessels fishing in the New Zealand exclusive economic zone (EEZ) and (3) a combination of Australian, Japanese and New Zealand observers on vessels fishing on the high seas collected under a collaborative project among these three countries known as the Real Time Monitoring Program (RTMP). The first two of these are limited in their spatial extent but cover long time spans (i.e. for Australia from 1979-1997 with detail data from 1991 and for New Zealand from 1990-2006). The RTMP data provide the broadest spatial coverage but are limited to the years 1991-1994. For both the RTMP and Australian observer data, direct comparison of the observed catch rates with either those logbook reported by the vessel when they had observers or in the logbook data for vessels without observers fishing in the same area and time period suggest that vessel reported catch rates tend to be somewhat larger than the rates observed by observers, while there is little difference in the reported catch rates between vessels with and without observers (Figure 3 and see Polacheck et al 2006 for more detail). Within the New Zealand EEZ, observer coverage has generally been high essentially 100% in recent years. As such, direct comparison of catch rates from observed and unobserved vessels within the New Zealand EEZ would not be expected to be particularly revealing. Nevertheless, the CPUE levels observed by New 7

Zealand observers are similar or less than those reported in general by Japanese fleet for the Tasman Sea or the general SBT fishing grounds (i.e. Areas 4-9) (e.g. Polacheck et al, 2004; Basson et al, 2005). As such, none of the available observer data suggest any substantive under reporting of the SBT catch rates in the vessel reported logbook data. It has been suggested that the lack of any substantial discrepancy between the observed and logbook reported catch rates is because vessels with observers deliberately fished ineffectively (e.g. by choice of set location, depth fished, baiting practices, etc.) when observers were aboard to prevent suspicions that catches and catch rates were being substantially underreported. In the case of the fishing within the Australian and New Zealand waters, it seems unlikely that this was occurring. In Australian waters, observer coverage was relative high (as great as 78% in some years). In addition, all vessels were required to under go pre and post inspection of the catches in their freezers. Any large discrepancies between observed and reported catches on the order of the 200% or more that would have been required too produce the overcatches should have been apparent. In the case of New Zealand, the 100% observer coverage would mean that deliberately poor fishing can not explain why catch rates in New Zealand waters have been similar or less than those reported in nearby waters. In addition, if substantial underreporting of catch rates was not occurring within Australian and New Zealand waters, there is the question of why vessels would have been willing to both pay and spend considerable amounts of time to fish there if could have achieve substantial higher catch rates on the high seas and not had to have reported them. Latent Effort and Effort Required to Catch the Overcatch As noted above one potential source of SBT overcatch is from Japanese SBT registered vessels fishing outside of the official Japanese SBT season and failing to report their catch or mis-reporting the species code. This could result in substantial catches of species such as bigeye and/or yellowfin being over reported and considerable amounts of effort incorrectly being judged as targeted at these species. One measure of the potential for this to have occurred is to determine how much potential latent effort existed in the fishery. The number of available fishing days that existed each year outside of the official SBT season combined with the number of SBT vessels provides an indication of the extent that fishing outside of the official season could have been a potential source of significant overcatch. Polacheck et al (2006) calculated the number of available fishing days for authorized SBT vessels that fell with the official SBT fishing season during a given year based on the number of vessels and fishing days allocated to different areas. Based on these calculations 10, no more than 36% of the available fishing days for SBT vessels would have fallen within official SBT seasons (Figure 4). As such, substantial latent effort existed within the SBT fleet, itself, for conducting fishing outside of the official season. Given that this type of IUU was documented to have occurred in 1996 (Anon 1997ba and see above), it can 10 Note that for the years 1990-92 and 1995-98 detailed information was not available on the number of vessels by fishing area. For these years, the numbers were imputed based on the pattern seen in other years. See Polacheck et al (2006) for details. 8

not be considered implausible that such fishing activity was at least a substantial part of the source of the SBT overcatch. To the extent that this was a source of the overcatch, it is likely to have resulted in over reporting of both catch and effort for other speices. Reported Catch and Effort in Area 2 SBT statistical Area 2 (Figure 5) is one such area that is never closed to longline fishing by Japan and is an area in which substantial quantities of SBT have been caught in the past. This statistical area is just to the north of Area 8 in the southeast Indian Ocean, which is one of primary fishing grounds for high quality SBT. Thus, Area 2 would be one logical place for either conducting illegal fishing operations for SBT or for vessels to report their location when fishing for SBT in a nearby closed area. Area 2 is recognized as a staging ground for SBT spawners ( the Oki grounds). The largest catches of SBT ever occurred during the early 1960s and Area 2 was the dominant source of the SBT catch in these years with over 75% of the annual reported Japanese SBT catch in some years. In 1971, much of this area was voluntarily closed by the Japanese industry to SBT fishing between December to March as a measure to help protect migrating spawning fish. However, by this time reported effort in this area had already declined to low levels (Figure 6) in preference for targeting the main more southern feeding grounds. After this period, reported SBT catch rates in Area 2 have always been low except for an increase reported in 2005 (Figure 6). While reported catch rates of SBT remain low in Area 2, the amount of reported effort increased dramatically after catch quotas became restrictive on the Japanese fleet in 1989 (Figures 6). This could reflect a displacement of effort towards bigeye (which are also found in this area) when the SBT fishery was closed or an area where catches (given the large historical catches of SBT from this area) and possibly location of effort was mis-reported. The available information does not allow evaluation if this was occurring and if so to what extent. Nevertheless, it is informative to explore the extent to which the large increases in reported effort in Area 2 could have accounted for the calculated overcatches from the Market Review. Figure 7 compares the amount of reported effort in Area 2 and the amount of effort required to have caught the overcatch using calculations described in Polacheck et al (2006). The comparisons in this figure suggest that the increasing effort in Area 2 between 1989 and 1995 would have been sufficient to explain a large fraction of the overcatch in those years. Subsequently, the extent to which the reported effort in Area 2 would not have been sufficient depends in part on the assumption made for what cath rate is assumed to apply for the overcatch. In any case, the reported effort in Area 2 was still sufficient to potentially have been an important component contributing to the overcatch. As such, it may also be a factor affecting the catch and catch rate data for other species. In this regard, it would be worthwhile to evaluate the extent to which the catch and catch rates from Area 2 contribute to the overall Japanese catches and CPUE statistics for other tuna in the Indian Ocean especially bigeye. 9

Implication for Catch and Effort Monitoring The large SBT overcatches and IUU fishing highlights the problem and risks of relying predominately on data from vessel supplied logbook for conducting stock assessments. That large incentive for fishermen to mis-report when restrictive quotas exist is obvious. Without effective independent monitoring of vessels activities and landings, it really is not surprising that large overcatches occurred in the SBT situation particularly given the high value associated with individual fish. One question about the reliability of the Japanese log book data is whether issues of reliability should only be seen as a concern in situations and time periods for which restrictive quota have been in place (e.g. for SBT, NBT and bigeye in the Atlantic). The estimates produced by the Market Review suggest that the beginning of the overcatches were associated with the catch quota becoming restrictive on the Japanese fleet. However, the estimates of the actual magnitude of the SBT overcatch particularly pre-2000 depend upon a number of assumptions for which there is little direct data. One of the most critical is the assumed value for the proportion of SBT that by-pass the wholesale market system. Australian industry has suggested that the figures used in the Market Review are too low. If this were the case then the market statistics would suggest substantial overcatches of SBT were occurring during most of the 1980s and prior to when catch quotas were restrictive. The motivation for misreporting is not clear. Australian industry has suggested that concerns about taxes would have existed (Brian Jeffries, personal communications). Such concerns raise questions about the general reliability of the Japanese logbook data and not simply for SBT. Given the apparent lack of effective independent monitoring of the Japanese fleet in the past and the large SBT overcatches, independent assessment of the reliability of these data for other species and time periods (e.g. additional market reviews) would seem warranted particularly given the central role of Japanese CPUE time series in most tuna and billfish assessments. The large overcatches and reliance on unverified logbook data raises general questions about the lack of rigour and standards for data that have been used in fishery stock assessment and in the provision of scientific advice. Accepting logbook data for use in stock assessments without independent monitoring and verification is standard practice in many fora - particularly at the international level where it is difficult to question official statistics and where any independent monitoring requires consensus. When there are conflicts of interest or at least perceived conflicts of interest combined with the mistrust that often exists between those supplying the data and the potential use of the data for compliance, management and/or science; standard scientific practice would require independent review and evaluation of the objectivity and accuracy of such data before the data were used in the provision of scientific advice (e.g. compare the lack of standards here with the requirement for double blind administration of drugs in medical trials). Of course the problem in many fisheries (particularly tuna fisheries) is that with out the logbook data, there would be little or no basis for conducting a scientific assessment. Nevertheless, any analyses or assessment is conditional on the assumed accuracy and reliability of the input data. At what point, is the appropriate response for requests for scientific advice to respond that it is not possible to provide 10

meaningful advice based on unverified and potentially unreliable and biased data? At the very least, it would seem that there is a need to ensure that concerns about the lack of verification of logbook data are clearly raised as a matter of priority. There is a need for the scientific fishery community to be more proactive in the development and implementation of independent ways to monitor and verify catches and fishing effort (e.g. scientific observers, video monitoring, port sampling, etc,) Literature Cited Anon. 1990. Report of the ninth meeting of Australian, Japanes and New Zealand scientists on southern bluefin tuna. 17-22 September 1990. Hobart, Australia. Anon. 1991. Report of the tenth meeting of Australian, Japanes and New Zealand scientists on southern bluefin tuna. 17-22 September 1991. Wellington, New Zealand. Anon. 1997a. CCSBT Report of the Resumed Third Annual Meeting (Revised). 18 22 February 1997. Canberra, Australia. Anon. 1997b. Report of the Third Scientific Committee Meeting of the CCSBT. 28 July 8 August 1997. Canberra, Australia. Anon, 1997c. Outcomes of discussions of the informal sceientist workshop 13 May 1997. Coogee Bay Hotel Sydney, Australia. Anon. 2000. AFMA SBTMAC FAG Meeting Report.:Report of the Special 2000 SBT FAG Meeting. 7-8 September 2000. Hobart, Tasmania. Anon. 2003. CCSBT Report of the TenthAnnual Meeting of the Commission. 7-10 October 2003. Christchurch, New Zealand. Anon. 2005a. CCSBT Report of the Twelfth Annual Meeting of the Commission.15 October 2005. Narita, Japan. Anon. 2005b. CCSBT Report of the Tenth Meeting of the Scientific Committee. 9 September 2005. Narita, Japan. Anon. 2006a. CCSBT Report of the Thirteenth Annual Meeting of the Commission. 10 13. October 2006. Miyazaki, Japan. Anon. 2006b. CCSBT Report of the Special Meeting of the Commission. 18-19 July. Canberra, Australia. Anon. 2006c. CCSBT Report of the Seventh Meeting of the Stock Assessment Group. 4-11 September 2006. Tokyo, Japan. Anon. 2006d. CCSBT Report of the Eleventh Meeting of the Scientific Committee (September 2006) 11

Australia. 2005. Comparison of CCSBT catch data with Japanese auction sales of frozen SBT. CCSBT-EC/0510/25. Basson, M,, J. Hartog, D. Kolody, T. Polacheck. 2005. Fishery indicators for the SBT stock 2004/05. CCSBT-SC/0509/25. Caton, A.E.. 1991. Review of aspects of southern bluefin tuna biology, population and fisheries. pp181-357. In: World Meeting on stock assessment of bluefin tunas: strengths and weaknesses. Special report. Edited by R.B. Deriso and W.H. Bayliff. La Jolla, California: Inter- American Tropical Tuna Commission. Jeffriess, B. 2000. Catches by other countries. AFMA SBT-FAG/2000/20. JFA. 2005. Preliminary analysis on growth rates of farmed SBT through trade data and other related information. CCSBT-EC/0510/29. Hampton, J., J. R. Sibert, P. Kleiber, M. N. Maunder and S. J. Harley. 2005. Fisheries: Decline of Pacific tuna populations exaggerated? Nature 434 (E1-E2) doi: 10.1038/nature03581. Hurry, G. 2006. Opening Statement by Australia. CCSBT Report of the Thirteenth Annual Meeting of the Commission. 10 13. October 2006. Miyazaki, Japan. Itoh, T. and K. Miyauchi. 2005. Review of Japanese SBT Fisheries in 2004. CCSBT-ESC/0509/SBT Fisheries/Japan. Myers, R.A. and Worm, B. 2003. Rapid worldwide depletion of predatory fish communities. Nature 423: 280-283. Myers, R.A. and B. Worm. 2005. Fisheries: Decline of Pacific tuna populations exaggerated? Myers and Worm reply. Nature 434 (E2) doi:10.1038/nature03582 Polacheck, T. 2002. Experimental catches and the precautionary approach: The Southern Bluefin Tuna Dispute. Marine Policy.26:283-294. Polacheck, T. 2006. Tuna longline catch rates in the Indian Ocean: Did industrial fishing result in a 90% rapid decline in the abundance of large predatory species? Marine Policy 30:470-482. Polacheck, T. and N. Klaer. 1991. Assessment of the status of the southern bluefin Tuna Stock using virtual population analysis - 1991. 10th Trilateral Scientific Meeting on SBT SBFWS/91/6. Polacheck, T., A. Preece, A. Betlehem and K. Sainsbury. 1996. Assessment of the status of the southern bluefin Tuna Stock using virtual population analysis - 1996. CCSBT/SC/96/26. 12

Polacheck, T., Preece, A. and Klaer, N. 1998. Assessment of the Status of the Southern Bluefin Tuna Stock using Virtual Population Analysis - 1998. CCSBT-SC/9807/17. Polacheck,T., A. Preece and D. Ricard. 2001. Assessment of the the Status of the Southern Bluefin Tuna Stock Using Virtual Population Analysis 2001. CCSBT-SC/0108/20. Polacheck, T. and A. Preece. 2001. An Integrated Statistical Time Series Assessment of the Southern Bluefin Tuna Stock based on Catch at Age Data. CCSBT-SC/0108/20. Polacheck,, T., D. Kolody, M. Basson and J. Gunn. 2004. Fishery indicators for the SBT stock 2003/04. CCSBT-SC/0409/21. Sibert, J., J. Hampton, P. Klieber and M. Maunder. 2006. Fishery induced changes in biomass, size structure and tropic status of top-level predators in the Pacific Ocean. Science 314: 1773 1776. Walters, C. 2003. Folly and fantasy in the analysis of spatial catch rate data. Canadian Journal of Fisheries and Aquatic Science 60: 1433-1436. Williams, S.C. Marketing tuna in Japan. Quensland Fishing Industry Training Committee. Brisbane, Australia. 13

Table 1: Number of Japanese longline vessels, the number that reported catching some SBT, and the number that caught more than 100t. The data in recent years are preliminary (modified from Itoh and Miyauchi, 2005). Percent of Global fleet Year All longline 1 SBT>0 2 SBT>100 3 SBT>0 2 SBT>100 2 1983 770 270 265 35 34 1984 761 287 276 38 36 1985 773 293 275 38 36 1986 771 271 253 35 33 1987 770 276 248 36 32 1988 759 255 223 34 29 1989 764 256 229 34 30 1990 758 250 240 33 32 1991 737 196 187 27 25 1992 723 205 192 28 27 1993 722 209 186 29 26 1994 716 201 193 28 27 1995 703 210 201 30 29 1996 674 230 218 34 32 1997 661 213 205 32 31 1998 663 220 205 33 31 1999 528 188 183 36 35 2000 529 180 168 34 32 2001 529 196 187 37 35 2002 523 176 168 34 32 2003 517 173 162 33 31 2004 506 169 165 33 33 1: The total number of Japanese high sea longline vessels. 2: The total number of Japanese high sea longline vessels which operated in the statistical areas 4-9. 3: The total number of Japanese high sea longline vessels which operated in the statistical areas 4-9 and caught more then 100 tonnes of SBT. 14

2004 2000 1996 1992 1988 1984 16000 14000 12000 10000 OVercatch (t) 8000 6000 4000 2000 0-2000 -4000 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 Year Case 1 Case 2 Figure 1. Estimated amount of overcatch in a year. The overcatch is calculated as the difference between the Market Review estimates and official Japanese catch statistics and assuming that 70% of the estimated market sales in a given year were caught in the previous year and 30% were caught two years previously. The lines labelled Case 1 and Case 2 are based on the Market Review s Case 1 and 2 scenarios for the amount of double counting that exists in the market statistics. (Adapted from Polacheck et al 2006). 280 25000 230 20000 Percent 180 130 80 30 15000 10000 5000 Catch (t) -20 0 Year Case 1 Case 2 Official Catches Figure 2: Japanese official catches of SBT as reported to the CCSBT for years 1984-2004 including RTMP and EFP catches (see Polacheck et al 2006 for details) and the percentage overcatch for the Case 1 and Case 2 scenarios from the Market Review lagged as in Figure 1. 15

Log Book 2 4 6 8 10 12 RTMP Vessel w/observer 0 5 10 15 20 0 5 10 15 20 25 RTMP observer 0 5 10 15 20 25 RTMP observer RTMP Vessel no observer 0 5 10 15 20 25 30 RTMP Vessel no observer 0 5 10 15 20 25 30 0 5 10 15 20 25 RTMP observer 0 5 10 15 20 RTMP vessel w/observer Figure 3: Comparison of the catch rates (number of SBT per thousand hooks) in a 5 square/month stratum based on different sources for the catch and effort data. Upper left: Japanese vessel reported log book data compared to RTMP observer data; upper right: RTMP vessel reported data when observers were present compared to RTMP observer data; lower left: RTMP vessel reported data when no observers were present compared to RTMP observer data and lower right: RTMP vessel reported data when no observers were present compared to RTMP vessel data when observers were present. (From Polacheck et al 2006; see this reference for details and other comparisons) Percent 36 30 24 18 12 6 0 1990 1995 2000 2005 Year Figure 4: The percentage of available days within a year that Japanese SBT vessels could officially have fished for SBT within regulated SBT fishing grounds (see Polacheck et al 2006a for details). 16

SBT Statistical areas Latitude -60-40 -20 0 15 14 10 9 8 2 1 3 4 7 6 5-50 0 50 100 150 200 Longitude Figure 5: SBT statistical areas defined by the CCSBT. Note that areas 1-10 have been those traditionally used for Japanese longline data. Areas 14 and 15 are recently defined areas to encompass areas where there has been significant Taiwanese reported catch and effort. Japan does not provide complete catch and effort data to the CCSBT for these latter two areas. Area 2 sets 0 2000 4000 6000 8000 10000 1970 1980 1990 2000 year cpue 0 5 10 15 1970 1980 1990 2000 Figure 6: Annual number of reported longline sets (upper panel) and catch rates (number of SBT per 1000 hooks lower panel) by Japanese vessels in Area 2. year 17

46000 Sets 36000 26000 16000 6000-4000 1984 1989 1994 1999 2004 Year Area 2 Case 1 Case 2 46000 Sets 36000 26000 16000 6000-4000 1984 1989 1994 1999 2004 Year Area 2 Case 1 Case 2 46000 Sets 36000 26000 16000 6000-4000 1984 1989 1994 1999 2004 Year Area 2 Case 1 Case 2 Figure 7: Comparison of the number of sets reported in Area 2 compared to the calculated number of sets required to account for the Case 1 and Case 2 overcatch scenarios assuming that all of the overcatch came from unreported sets. Upper panel assumes the catch rates for the unreported catch equaled the nominal CPUE for SBT; mid panel that they equaled the maximum of the nominal rate in Area 7, 8 or 9 (the primary Japanese fishing areas for SBT) and lower panel that they equaled the average CPUE in the top 20% ranked 5 square/month strata (taken from Polacheck et al 2006). 18