SUMMARY The objective of the meeting was to evaluate fishery indicators relative to establishing the state of the stocks, with special attention given to yellowfin and skipjack. Within these indicators, analyses were conducted on catch series, CPUEs and weight and/or size measures for the different stocks. Given the multi-species character of the majority of the tropical tuna fisheries, the Group reiterated the need to develop indicators capable of reflecting the overall state of the fishery and its impact on the ecosystem. RÉSUMÉ L’objectif de la réunion était d’évaluer les indicateurs pertinents de la pêcherie afin d’établir l’état des stocks, l’accent étant mis sur l’albacore et le listao. Parmi ces indicateurs, le Groupe a analysé les séries de captures, les CPUE et poids et/ou tailles moyennes pour les différents stocks. Compte tenu du caractère plurispécifique de la majorité des pêcheries de thonidés tropicaux, le Groupe a réitéré la nécessité d’élaborer des indicateurs capables de refléter l’état de la pêcherie dans son ensemble et l’incidence de celle-ci sur l’écosystème. RESUMEN El objetivo de la reunión era evaluar indicadores de la pesquería relevantes para establecer el estado de los stocks, siendo el rabil y listado las especies a las que se presto una especial atención. Dentro de estos indicadores, se analizaron series de capturas, CPUE y pesos y/o tallas medias para los distintos stocks. Dado el carácter multiespecífico de gran parte de las pesquerías de túnidos tropicales, el Grupo reiteró la necesidad de desarrollar indicadores capaces de reflejar el estado de la pesquería en su conjunto y la incidencia de la misma en el ecosistema. KEYWORDS Yellowfin, skipjack, fishery indicators, tropical tunas, ecosystem
1. Opening, adoption of agenda and meeting arrangements The meeting was opened by Dr. Sergio Gomes de Mattos of the Brazilian Special Secretariat for Aquaculture and Fisheries who presented a welcoming statement (Appendix 1) on behalf of Special Secretary, Mr. Altemir Gregolin. Dr. Victor Restrepo, on behalf of Mr. Driss Meski, ICCAT Executive Secretary, thanked the Brazilian government for hosting the meeting and providing all the logistical arrangements.
The List of Participants is included in Appendix 2. The Agenda (Appendix 3) was adopted after modifications. The List of Documents presented at the meeting is attached as Appendix 4. The meeting was chaired by Dr. Renaud Pianet, and the following served as rapporteurs:
2. Update of basic statistics and catch at size 2.1 Task I The Secretariat presented the catch table (Table 1) for the three tropical species (yellowfin, bigeye and skipjack) for the years 1950-2005. Since the last SCRS meeting held in 2006, some improvement was made to update the ICCAT Task I database according to the changes approved by the Species Group and the SCRS. The modifications carried out to this data set were:
1) the reclassification of Spanish (Canary Islands) and Portugal (Azores) baitboat catches,
2) the adjustment of Venezuelan catches according to a paper by Nova & Ramos and approved by the Tropical Species Group, and
3) the corrections of Guatemalan PS catches that had been reported twice.
The Group reiterated the need, for scientific purposes, to receive from Contracting and Cooperating Parties the fleet characteristics information (Form-1) requested by the Secretariat, together with the Task I data. This information is very important for the historical evaluation of fishing capacity and for characterizing the effort deployed to harvest tunas in ICCAT Convention area.
Discussions on data for Angola and Cape Verde for the years prior to 1960, when national tuna fisheries were very active for these countries, concluded that collaboration with Portuguese scientists and institutes should be encouraged to recover the missing information.
The Group decided that the Task I catch reported under FIS (France, Cote d’Ivoire, Senegal) will be broken down by country before September 2007 by scientists from IRD. It would be helpful to carry out the same exercise for the Task II data in order to link the revised Task I with Task II. Unfortunately, as the basic vessel data were lost, this seems to be nearly impossible.
It was noted that catch for some European-operated vessels with Ghanaian flag was recorded as NEI. It was recommended to extract these from NEI and assign them to Ghanaian flag.
2.2 Task II
The Secretariat reported the availability of Task II data (catch & effort and size sampling) but no thorough analyses of these data set were carried out. The Group reviewed the recommendations made by the previous tropical group meeting held in Sete, France (Anon., 2007), mainly on the need to correct some apparent discrepancies for several fisheries. During this discussion, it was announced that new Task II information from Brazil will be available soon.
The Secretariat recalled the need to have size sampling in the finest scale possible but for some countries it seems difficult to fulfill this request. It was noted that for the European fleet, the size sampling submitted to Secretariat was extrapolated (raised) to some level and, if necessary, the actual size samples could be submitted. It was also noted that it could be useful to have the longline fishery data reported in the same scale.
2.3 Catch at size 2.3.1 Skipjack. The Secretariat presented the new version of catch at size (Table 2) for the years 1969 to 2005. Given the multiple data received at the Secretariat after the first (1999) skipjack assessment, the Secretariat decided to recreate the entire data set. The substitution rules used were tabulated in Table 3. The Group congratulated the Secretariat for its excellent work in this important improvement made by rebuilding the skipjack catch at size. In order to review in detail the available information, the Group conducted more analyses, with the data split by flag, gear and stock.
In order to further quality assure these data, a series of graphical diagnostics were developed to provide a visual method for evaluating the patterns in estimated catch at size at a broad grouping of gear and area definitions. The Species Group decided to disaggregate the catch at size by fisheries as defined in Table 4.
Graphics showing the annual skipjack catch at size profiles for each of these fisheries are shown in Appendix 6 and should be used to guide future analysis of these data.
2.3.2 Yellowfin. A new catch at size dataset for the years 1970-2005 was presented by the Secretariat (Table 5), including the newly complete revision made by the Japanese scientists. The file was split by fleets and area in order to conduct some other analyses in a finer scale.
For both yellowfin and skipjack, the Secretariat presented the mean weight distribution by gear group and all gears combined.
2.3.3 Bigeye. No new information for this species was presented by the Secretariat, as the complete assessment on bigeye will be held in June 2007; the bigeye catch at size will be updated and presented before then.
3. Estimation of landings that are not discarded nor entering in the canning factories (“faux-poisson”) Available information was presented by the Chairman, according to two documents and new data he received from Côte d’Ivoire.
Since numerous years ago in Abidjan, small size tunas, minor tunas and by-catches refused by the canneries are sold on the local market fish as “faux-poisson”. This possibility incited tuna purse seiners to keep onboard part of their catchwhich previously would have been discarded. Consequently, the quantities of landed “faux-poisson” increased rapidly, especially since the 1990s with the development of FAD fishing, catching more small tunas and by-catches. However, quantities were reduced since 1994, because of the implementation of controls carried out by French fishing companies, and the limited carrying capacity of the remaining purse seiners. One can define “faux-poisson” as the part of the industrial tuna fishing catches consumed locally.
The estimation of these “faux-poisson” catches was implemented in 1981, following an initiative from F.X. Bard. First, the unloaded quantities were estimated from information collected from the customs officers at the various exit points of the fishing port, but it appears that this information was widely underestimated. Consequently ORSTOM (formerly IRD) progressively put in place its own system to collect this information, by an exhaustive counting of all vans and trucks carrying tuna outside the port. The system is supposed to be fully operational since 1994.The annual estimated quantities of “faux-poisson” as related to the total tunas unloading are presented in Table 6 and Figure 1.
Within the framework of the EU DG-Fish “Bigeye Project”, complementary to the ICCAT Bigeye Year Program (BETYP), a special sampling (species composition and sizes) scheme was implemented (1997-1999). Then, the sampling was improved in order to estimate the number of vans and/or trucks per day and tuna boats, as well as a visual determination of the species composition of these “faux-poisson” unloading (see form established by R. Dédo, Table 7). Boat information is available since august 1994, and is now reported as a special category.
Results of these samplings for the two periods, 1997-1998 and 2004-2005, are presented in Table 8 and Figure 1. They show quite large variations that may be attributed to the flag origin of the unloadings (increased quantities from Ghanaian carriers). Species composition and size distribution of these catches are also reported in Figures 2 and 3 more detailed analysis of all available information is necessary to improve these estimates, and specially the amount of yellowfin, bigeye and skipjack tunas concerned.
The Group considered that these data were important as the quantities involved are quite significant, especially for skipjack. The fact that it is essentially made of small fish (35-45 cm) is also of concern. It is estimated that even if mixing among species of major tunas is difficult to assess, a more comprehensive study using all available information collected should be attempted and results used for assessments.
It was suggested that the best way to officially take these changes into consideration is to consider them as discards. The Tropical Species Group is asked to study the best solution at its October 2007 meeting.
4. Review of new biological information Two documents were presented to the Group (SCRS/2007/057 and Shuford et al. 2007). Document SCRS/2007/057 described the latitudinal variability in growth rates of eastern Atlantic skipjack tuna (Katsuwonus pelamis). Conventional tagging data collected by ICCAT since the 1960s were reanalyzed using the von Bertalanffy-Fabens growth model. The results of this study suggest that the growth parameters of skipjack tuna vary with latitude. The estimated Lfor skipjack tagged and recovered north of 10°N was very similar to the Lestimated for skipjack tagged and recovered south of 10°N (94.1 cm vs. 92.9 cm, respectively). However, the estimated growth rate coefficients, K, differed in the two regions (0.353 versus 0.195), suggesting that skipjack tuna grow faster in the northern region of the eastern Atlantic Ocean than in the equatorial areas. The growth parameters estimated during this study were consistent with those obtained from skipjack tagged during the Mattes Associées aux Canneurs (MAC) activities in the Senegalese area (L= 97.26 and K = 0.251). In contrast, the estimated growth parameters reported by Cayré et al. (1986) for the same region (L= 62.0 and K = 2.08) were quite different, and were not supported by this analysis.
The Group considered the implications of these results. It is evident that latitudinal variability in the growth rates would complicate age-structured assessment techniques because the size-at-age would be dependant on geographic location and movement patterns. The authors suggested possible alternatives to standard age-structured models including the use of catch-at-size models and growth-transition matrices by large geographic areas.
The Group made several recommendations that could improve future estimation of growth rates for skipjack tuna. The Group recognized that measurement error is a problem common to tagging data. Some members were concerned that excluding fish that were >2 cm smaller upon recapture could bias the estimated growth parameters (since measured size is equally likely to be greater or less than the true size). A possible alternative is to exclude fish if the apparent growth was 3SD from the expected growth. The Group also suggested that an additional bias in region-specific growth rates could occur if the proportion of excluded fish differed between regions.
The Group suggested that the effects of seasonality and possible sensitivity to the minimum acceptable time-at-large could be important to consider.
The Group also recognized that changes in selectivity can cause the perception of a change in growth parameters. This problem has been demonstrated in the U.S. Gulf of Mexico after changes in the minimum legal size. The group discussed possible differences in selectivity north and south of 10°N (i.e. different proportions of FAD versus free-school samples). A possible technique to estimate population growth parameters using fisheries-dependent data (subject to selection) is to use a “truncated likelihood” technique. This technique has been used recently during assessments conducted in the U.S. Gulf of Mexico (pers. comm. M. Ortiz1).
The Group also discussed the importance of collecting direct observations of age for tropical tunas from hard part analysis using modern techniques. If direct observations of age composition were available for tropical tunas, it would be unnecessary to estimate the age composition using techniques (age-slicing, age-length keys) that are known to suffer if size-at-age is highly variable, or if spawning occurs throughout the year.
Shuford et al (2007) was presented to the Group. This document addresses the 2003 ICCAT yellowfin stock assessment recommendation that “research should be conducted to validate the growth curve used for yellowfin tuna.” In this study, an age and growth curve is derived using directly determined age estimates from the daily microstructure of sagittal otoliths from 132 fish (5.2-179 cm FL) caught in the Gulf of Guinea and North Carolina. Using the age estimates and corresponding fork lengths, the growth curve and parameters were derived with the von Bertalanffy growth function. The resulting growth coefficient (k) is 0.281; theoretical maximum FL, (L) 245.5cm; and theoretical FL at age zero (t0), 0.042.
In order to evaluate the relative contribution of the various spawning periods throughout the year (which are correlated to the various spawning areas), it is recommended that further research be conducted using direct ageing (or other appropriate methods) capable of estimating birth month since it appears that there may be less seasonality in birth month for some tropical species (e.g., see Appendix 5). Sampling should be designed to be representative of the catch and/or stock.
The data supporting this von Bertalanffy growth function were compelling. However, there was discussion that an early period of slow growth has been demonstrated using tagging data. The Group did not conclude that there was sufficient evidence at this time to reject the two-stanza growth function currently assumed for yellowfin, although further investigation of the matter (including review of relevant tagging studies) was encouraged. The Group also recommended that analyses be conducted to determine the effect of the different growth functions on the estimated age-composition, and the implications of those differences on estimates of total mortality derived from catch-curve analyses. This analysis was conducted during the meeting, and is summarized in Appendix 5. The results indicate that the age-composition derived from age-slicing is sensitive to the growth curve, and that this difference may be sufficient to cause changes in estimations of total mortality.
The Group concluded that changing the growth model could alter estimates of stock status. Therefore, the Group recommended that the original data be recovered. These data should be used to determine how the growth functions were developed, to test alternative model fits and structures and to investigate the potential to develop a model that incorporates both data sets.
5. Review of fishery indicators General indicators are in Section 5.2; Section 5.1, which follows, has species-specific indicators.
Total catches of Atlantic tropical tunas (yellowfin, skipjack and bigeye) by gear are shown in Figure 4. Overall catch shows an increasing trend up to reach 474, 084 t in 1994 followed by a decreasing trend in the most recent period. Trends of catches are mainly due to the development of the purse seine fishery.
5.1 Indicators relevant to stock status
General trend in skipjack catch
As shown in Figure 5, the total catches of skipjack in the Atlantic Ocean depict a continuous decline since their maximum reached in 1991 (175,000 t) but with a slow increase in the recent years (138,000 t in 2005).
As concerns the East Atlantic, the skipjack fishery underwent important changes in the early 1990s, following the introduction of artificial floating objects (FADs) by the European Union purse seiners and in the Ghanaian purse seine fisheries and baitboat fisheries. In addition, a fishing method was developed in which the baitboat is used at the floating object in the waters off Senegal, Mauritania and the Canary Islands (1992). All these changes have caused an increase in the exploitable biomass of the skipjack stock (in particular, due to the expansion of the fishing area to the west), and in skipjack catchability. Despite these changes, the catches for purse seiners decreased, as previously described for the total catch. This decrease is likely a consequence of the implementation of the moratorium on FAD fishing as well as the decrease in the nominal fishing effort. However, the catch increased in the recent years. For the pole and line fisheries the catch remained stable at about 40,000 t since the late 1970s.
In the West Atlantic, after fluctuating between 2,000-3,000 t/year, the catch of skipjack rocketed up to about 40,000 t in 1985. Since this maximum, the western catches (mainly due to the Brazilian baitboat fishery) were maintained at a level close to 30,000 t (Figure 6).
Apparent abundance indices The Group stressed the importance of updating the catch rates of the main fisheries reporting catch of skipjack. It must be stressed that skipjack tuna is often a secondary species, depending on the price differential and on the catchability of other target species. Consequently, estimation of the effective effort exerted on skipjack (e.g. effort proportional to fishing mortality) remains problematic and catch rate may sometimes depict a different trend than abundance.
For purse seiners fishing alternatively on free schools and on FADs, it was considered that search time may be the best measure of basic effort on free schools. It was also suggested that the analysis data set might be further restricted to effort associated with free school sets by assuming that vessels which travel longer distances overnight are moving between FADs, as they can’t be searching for free schools at night. However, this approach would likely require further study, including the incorporation of VMS data, to determine if it is both feasible and appropriate. A new EU funded CEDER Project (Catch, Effort and Discards Estimates in Real Time), which started in 2006, will be partially dealing with this question. The basic idea developed within the framework of this project is to analyze the individual trajectory of purse seiners in order to characterize fishing behaviors reflecting searching time for un-associated school or running toward a FADs previously detected by radio range beacon (bearing in mind, however, that whatever the fishing mode researched, every tuna school detected by chance can be set on). Other factors which might be considered include the changes over time that have resulted in reduced time necessary to make sets and to offload catch (increasing efficiency of fishing effort over time).
Eastern Atlantic The analysis of the changes in nominal catch rates and in several fishery indicators of the EU tropical surface fishery (baitboat and purse seine) operating in the eastern Atlantic Ocean was presented in document SCRS/2007/058. Figure 7 depicts the change over the period 1991-2006 of the CPUEs for the EU purse seiners for both fishing modes (i.e., non-associated schools and FADs set). From this figure, it can be seen that the catch rate for skipjack caught in free school fluctuates over the years (about 1 to 4 t/searching day) but without evidence of any trend. The increase in CPUEs for skipjack associated to FADs since 1999 may be partially due to a larger catch by positive set (Figure 7; lower panel).
Regarding the European baitboats based in Dakar (Senegal), the nominal catch rates of skipjack have increased regularly over the entire time series (Figure 8). When analyzing these data it must be kept in mind that since the beginning of the 1990s these baitboats have developed a fishing technique (mainly for targeting bigeye) in which the baitboat is used as the floating object, fixing the school (comprised of bigeye, yellowfin and skipjack) during the entire fishing season in waters off Senegal and Mauritania. As a consequence, it makes sense to assume that the adoption of this fishing technique has increased the overall catchability of tunas. Notice however that the pattern described for skipjack contrasts with the decreasing trends in CPUEs observed for the other two tropical tuna species.
During the meeting the Group held in Séte (France) in 2006, it was recommended that analyses of the CPUE trends for fisheries along the fringes of species distribution be conducted by scientists from various Contracting Parties. The results of the standardization of the CPUEs for the Azorean baitboats was presented and discussed during the Species Group. As expected, due to the location of this fishing area with respect to the range distribution of skipjack, the standardized index showed a high variability, but without a significant tendency (Figure 9).
Western Atlantic In contrast with the large fishing areas observed in the eastern part of the Atlantic Ocean, the fishing grounds in the western Atlantic are generally more coastal (one explanation is that the deepening of the thermocline and of the oxycline to the central Atlantic limits the areas with suitable conditions for surface tuna fishery). The Brazilian baitboat fishery represents an interesting study case as skipjack is their main target species. The nominal CPUE was not updated since the last SCRS meeting but there is no evidence of any trend, even if a large variability is observed (Figure 10). Note that the catch rates reported for this fishery are higher than the CPUEs observed in all the eastern baitboat fisheries. For the Venezuelan purse seiners, fishing mainly in the Caribbean Sea, no new information has been provided since the last SCRS meeting. The reason for the decrease in nominal CPUEs observed since 2002 (Figure 10) is not known (possible causes include changes in local environmental conditions, lower abundance of skipjack or its prey, lower assistance of the baitboats to hold tuna schools stationary near the surface during the encircling operation, etc.).
Preliminary results concerning the standardization of the catch rate of skipjack caught by the U.S. recreational fishery with a delta-lognormal model were presented during the meeting. Following the recommendations of the Species Group, a second analysis accounting for a better categorization of some explanatory variables (e.g., larger geographical areas, omission of strata where the presence of skipjack is unlikely, etc) was tentatively done. It was shown that the proportion of positive catch increased slowly, whereas the catch for positive trips remained relatively constant. However, since the confidence intervals remain very large and the standardized CPUEs fluctuate similarly than the nominal index (Figure 11), further studies are needed.
Mean weight The Group also discussed a number of other analyses that may provide some indication of stock status in the absence of a full assessment. Among these auxiliary indicators, changes in average weight of fish in the catch over time may be useful. A reduction in average weight may be reflective of increasing or sustained high fishing mortality, although an initial reduction in average size is expected in a fishery and changes may also reflect changes in selectivity or recruitment pulses (at smaller sizes).
Size samples for longline-caught fish throughout the Atlantic are too sparse to provide reliable indication of pattern in average size of the overall longline catch (see Appendix 6).
Eastern Atlantic The mean weight of skipjack in the eastern Atlantic varies between 2 and 3 kg (Figure 12). A slow decrease can be observed from 1969 to the late 1980s for both purse seiners and baitboats. This indicator stays stable in the last decade. The Group stressed the need to further examine information for the fisheries identified in Table 4 in order to better evaluate the impact of the fishing pressure on the different components of this stock.
Western Atlantic The average weight of skipjack is larger in the western Atlantic than it has been showed for the East (3-4.5 kg vs 2.5-3 kg, respectively; see Figure 13). This is specifically the case for the directed Brazilian baitboat fishery during the last 20 years (with exception of the last two years for which this indicator decreased). The mean weight for skipjack landed by purse seiners (mainly from Venezuela) is lower than for baitboats, and decreases continuously until approximately 2 kg.
Information on the change over time of the average length of skipjack caught by the U.S. longline were presented in document SCRS/2007/054. Even if skipjack is an incidental catch for longline and even if there is some variability in mean length between the different areas fished by this fishery the global pattern suggests a slow increase in length for the last 15 years (approximately from 62 to 70 cm). This pattern corresponds to a change in average weight approximately from 4 to 5 kg (Figure 14).
5.1.2 Yellowfin Catch trends In contrast to the increasing catches of yellowfin tuna in other oceans worldwide, there has been a steady decline in overall Atlantic catches since 2001 (Table 1; Figure 15). Atlantic surface fishery catches have shown a declining trend from 2001 to 2005, whereas longline catches have generally fluctuated without clear trend. Purse seine catches have declined over 55% from the peak catch level in 1990.
It is not immediately clear to what extent declines in catch might be the result of reductions in stock size or are due to reductions in targeted effort levels. It is expected that the reduction in purse seine carrying capacity in the Atlantic (a result of the aging and reduction of the existing purse seine fleet as newer vessels are introduced instead in the Indian Ocean) has had a profound impact on catch levels (Figure 16). The shift since 1990 toward a greater proportion of sets on FADs (Figure 17) has increased the proportion of skipjack in each set as well as decreased the average size of yellowfin on a set. It is possible that reductions in longline catches may be due, at least in part, to shifts toward targeting bigeye tuna.
Abundance indices Purse seine and baitboat in the east Atlantic The largest catch and effort database, which covers the largest and most widespread fishery for yellowfin tuna, is that for the eastern tropical Atlantic baitboat and purse seine fisheries operated by European and associated fleets. Document SCRS/2007/58 presents the main statistics and characteristics of these fisheries from 1991 up to 2006.
The purse seine catches have generally shown a declining trend, particularly since 2001, although a slight recovery is observed in the preliminary 2006 catches due to increases in free school catches (Figure 19). The nominal catch rates show increasing or stable trends (Figures 20 and 21). However, these trends have not been adjusted to account for technological and methodological changes in the fishery (i.e.use of sonar on FADs, bird radar) which are expected to have increased fishing power.
The yellowfin tuna catch rates have shown in general a decline for the Dakar/based baitboats (Figure 8) while the carrying capacity steadily increased (Figure 18). It is not clear how the catch rates trend has been affected by the adoption in the 1990s of techniques using the baitboat as a fish aggregating device. Catch rates since the late 1990s have fluctuated without a clear trend.
Purse seine in the west Atlantic No new information was available from the purse seine fisheries in the west Atlantic.
Longline and rod-and-reel indices Several CPUE indices were presented at the meeting from fisheries other than purse seine. All the indices were standardized using GLM, differing in the assumption of the error distribution (log-normal or Poisson). They had the same basic factors in common, such as year, season and area, along with other factors particular to each case.
Two indices were presented for the U.S. pelagic longline fishery from 1987 to 2006 (SCRS/2007/055) derived using fisherman logbooks, in number of fish and in biomass (adjusting logbook numbers by time-area average weights from observer data). These indices show a clear declining trend until 2003, followed by an increase in the last three years (Figure 22). Indices in number of fish were also developed for the U.S. pelagic longline fishery using observer data and are presented in document SCRS/2007/056. The observer based index covering the entire U.S. pelagic longline fishery is shown compared to the logbook-based index in (Figure 23); the relatively high variability of the index may be influenced by the sample size of observer data. A separate index was also calculated for the U.S. pelagic longline in the Gulf of Mexico using the U.S. observer data (Figure 24), showing high variability with no clear trend. The completion of the planned incorporation of Mexican observed data in order to calculate a combined index may improve these results.
An index in numbers for the Brazilian longline fishery (1978-2005) was presented in document SCRS/2007/059. No clear trend was evident, and values were highly variable during the 1990s (Figure 25). Discussion centered around the importance of the targeting factor, which was calculated using cluster analysis. It was considered that actual changes in targeting behavior may not be completely captured by this approach, and that this may be at least a partial explanation for the variability.
Okamoto (2007) presented an index for the Japanese longline fleet (1965-2005) in number of fish using, in addition to the common factors, environmental factors such as sea surface temperature. The catch and effort data set was aggregated by month, 5-degree square and the number of hooks between floats. Data for 2005 were considered preliminary. The index showed a clear overall declining trend (Figure 26), although point estimates for 2003 and 2004 had shown some recovery (prior to the preliminary 2005 estimate). This index has the broadest coverage in time-area of any of the yellowfin standardized abundance indices.
An index, in number of fish, for the U.S. rod and reel fishery (1986-2006) was presented in document SCRS/2007/052. This analysis used intercept survey data. Trends were quite variable, with a declining trend from 1986 to 1992, followed by an increasing trend (with large fluctuations) until 1999, and finally a generally decreasing trend. Concerns were raised that that successful trips within the analysis data set were rare, which may contribute to the variability of the index. There was discussion of alternatives which might potentially restrict the analysis data set to focus more on yellowfin directed effort; this alternative treatment was performed during the meeting and the updated indices (slightly changed as a result) are shown in Figure 27. Additional improvements to the analysis are being considered for calculations to be presented at future meetings.
Average length/weight trends The overall average weight trend, calculated from available catch-at-size data, has been declining fairly steadily since 1973 (Figure 28). The average weight trends by gear group are shown in Figure 29. The average weight in purse seine catches, which represent the majority of the landings, has been declining since 1992. This decline is at least in part due to changes in selectivity associated with fishing on floating objects. Less dramatic declines in average weight have also occurred for baitboat fisheries. Longline fleets, which have not undergone such a dramatic shift in selectivity and which target larger fish, have experienced and general decline in average weight since 1975, although calculated averages have varied widely beginning in 2002. The causes of the variability in longline average weight in the most recent years are not clear, although much of this data is still regarded as preliminary. The value for 2005 is preliminary.
In order to determine changes in average weight for the purse seine fleet independent of the change in selectivity associated with FAD (also know as “log”) fishing, the trends for the European and associated purse seine fleet were examined separately by set type (Figure 30). Average weights have generally declined for both FAD and free school sets, although there was a slight recovery in 2006. This suggests that the decline in purse seine average weights cannot be entirely attributed to an increase in the proportion of sets on FADs. The complete size distributions for the two set types, comparing the period 2001-2005 to 2006, during the most recent six years are shown in Figure 31. The size distribution for the same period in the European and associated baitboat fleets is shown in Figure 32. Information on the average length of yellowfin tuna, by year, caught by the U.S. longline fleet was presented (SCRS/2007/054) during the meeting and is shown in Figure 33. There was no substantial trend in mean length for the time period 1992-2006.
An examination of these various signals in the fishery is no substitute for a full stock assessment. However, the declines observed in the standardized abundance trends from major longline fisheries, combined with declines in average weight which cannot be entirely explained by increased targeting of FAD purse seine sets, reinforce the need to conduct a full stock assessment for yellowfin tuna. It is important that such an assessment take these and other indicators into account.
5.1.3 Bigeye General trend in bigeye catch by gear
Total annual catch (Figure 34) steadily increased and reached to 130,000 t in 1994, and has rapidly decreased to less than half, about 60,000 t in 2005. This large decline was observed not only in longline but also purse seine and bait boat fisheries. Longline catch account for about 60-70% of the total catch. In 2005, the catch of the major fisheries was 35,400 t, 12,500 t and 11,000 t for longline, purse seine and baitboat fisheries, respectively.
Fishery indicators presented in this meeting
Standardized bigeye CPUEs for one longline fishery, two baitboat fisheries, and one purse seine fishery were presented during meeting. CPUE for Chinese Taipei and Japan, which are the main fleets catching bigeye, will be updated until the next bigeye assessment in June, 2007 using the latest available data.
EU purse seine (SCRS/2007/058): The bigeye catch by the EU purse seine fishery, which was about 30,000 t in 1993, decreased to 13,000 t in 1998 and has remained at the same level thereafter. Although 80-90% of bigeye has been caught by operations on log-associated schools, the free swimming school catch was exceptionally high, about 30% in 2006. CPUE for each type of operation (free and log) has been kept at the same level, and does not show any marked trend except for relatively high CPUE of free school set in 2006 (Figure 35).
The Group discussed how purse seine CPUE should be standardized for the next bigeye stock assessment in June 2007. In the two CPUE indices, ‘catch per positive set’ and ‘catch per searching day’, the former one would be appropriate for small bigeye, and the CPUEs of two types of operation, free and log school sets should be treated separately (see also Section 8).
EU baitboat, Dakar-based (SCRS/2007/058): The bigeye catch of EU baitboats has fluctuated around 2,000-4,000 t except for high catches of 5,000-7,000 t in 1999 through 2001 (Figure 36). Total effort has constantly increased from 1,500 days in 1991 to around 3,500 days in the latest years. CPUE (tons/day) which fluctuated between 1.5 and 2.5, showed a sudden drop to less than 1.0 and has remained at a low level up to now (Figure 37).
Brazilian longline (SCRS/2007/059): The bigeye CPUE of the Brazilian longline fishery was standardized for 1978-2005 using the same data set and method as that used for yellowfin. Standardized bigeye CPUE (fish/100 hooks) was stable around 0.3-0.5 until 2000, after which it has remained at a low level of about 0.2 (Figure 38). It was mentioned that the information on targeting or gear configuration is not available, although they should be important factors to standardize yearly and seasonal changes in targeting strategy.
Azores baitboat: The Azores baitboat catch, which has fluctuated between 1,000 and 6,000 t for the period from 1950 to 1999, decreased to less than 1000 t, thereafter. Bigeye tuna are mainly caught in the second quarter, from May to July. CPUE standardized by Delta-lognormal model showed constant declining trend until 2003, and an increasing trend was observed in 2004 and 2005 (Figure 39). The Group considered that the CPUE of Canary Islands fishery should be different from that of Azores baitboat.
Mean weight During this meeting, size and mean weight analyses for the latest years were presented for the EU purse seine and baitboat fisheries. The size information for this species will be summarized including other fisheries in the next bigeye assessment meeting which will be held in June, 2007.
EU purse seine: The mean weight of the catch was constantly around 4-5 kg in the catch of log school set while that of free school set has been fluctuating between 5 and 14 kg (Figure 40). It is noteworthy that large bigeye around 140-170 cm were predominant in free swimming school set in 2006, while the size composition of log school in 2006 is similar to that in the past years (Figure 41).
EU Dakar-based bait boat: The mean weight of the EU baitboat (Figure 42) was high, at about 10kg from 1991 to 1997 and suddenly declined to 7 kg in 1998. This low mean weight lasted until 2004 and returned to high level about 10 kg in 2005 and 2006.
5.2Fishery indicators relevant to multi-species fisheries and ecosystems The Group stressed that the tropical tuna fisheries are multi-species fisheries, with strong interactions between the selectivities and fishing mortalities among yellowfin, skipjack and bigeye tunas. Considering this, it may be useful to develop new indicators reflecting the status of the fishery as a whole. Along these same lines, it was also recognized that the tropical tuna fisheries influence the pelagic ecosystem.
No document was presented this year on this matter. Nevertheless the Group considered that the points of the meeting of Sete (2006) continued being valid:
− Indicators for Sustainable Development of Marine Capture Fisheries (Code of Conduct for Responsible Fisheries).
− Current usage of fisheries indicators and reference points, and their potential application to management of marine fisheries.
− Community indicators can measure the impact of fishing.
− The FAO framework for use of quantitative indicators and performance measures to manage target species and ecosystem impacts.
In this sense the EU plans to integrate information of the impact of fishing on the ecosystems in its new Data Collection Regulation, starting in 2008.
However, it was underlined that at this time, the Group is working with several indicators that could be considered from the multi-species point of view (catch of all species/set, CPUE). This discussion will have to be transferred to the Sub-Committee on Eco-Systems for a wider vision of the problem. The Group examined several general indicators of the purse seine and baitboat fisheries, such as: change over time in carrying capacity (Figure 16 and 18), the total number of sets and the % of successful sets by fishing mode (Figure 17), and total area visited and fished (Figure 43 and 44).
For the purpose of examining the pattern in the overall longline fishery production and catch rates in the eastern tropical Atlantic to compare with those available for the surface fisheries, catch and effort by 5x5 in the region ranging from the African coast seaward to 30oW longitude and between 10oS and 25oN latitude was examined. These data were as recorded in the CATDIS file and the estimates of longline fishing effort produced by the Sub-Committee on Ecosystems at their March 2007 meeting. Figure 45 shows the pattern in effort, catch and nominal CPUE by year for the entire period and for the period since 1991 to compare with similar views of the tropical purse seine fleet; Figure 46 presents the total area visited and fished (1ºx1º) by the Japanese longline fleet from 1977.
With the aim of favoring multi-species approaches in tropical tuna fisheries, comparisons of changes over the years in the total mortality Z for yellowfin and skipjack were conducted during the meeting with the use of the catch-at-size data provided by the Secretariat. An apparent Z estimate for each species was performed on the basis of the following equation:
Z = K*( L∞ -L moy) / (L moy - L c), where L∞ and K represent the conventional parameters of the von Bertalanffy’s growth curve, Lc = the length at which fishes are fully recruited, and Lmoy the average length for fishes fully recruited (Beverton and Holt, 1956) (Figure 47). In this figure it can observe that skipjack total mortality decreased from final of the 1990s, possibly due to the moratorium on floating objects, since skipjack has been the main affected species.
During the present Species Group meeting another indicator was updated. This indicator, termed “Skew” hereafter, is based on the reasonable assumption that an increase in fishing effort reduces the proportion of older individuals in the population and as a result the age/size composition becomes more skewed (Rosenberg and Brault (1991) in Caddy, 2004). This indicator is defined as:
Skew = N -1 Σ [ (X i – X moy ) / s ] 3, where N is the number of age/size classes, X i is the relative abundance of the ith class, X moy and s are the average abundance and the standard deviation respectively. The changes over the years of this indicator for the three species are shown in Figure 48. The general trend depicted in this figure (i.e., assumed to reflect an overall increase in fishing mortality) reinforces that was observed in the previous analysis of Z (even though some discrepancies may exist between both indicators).
The Group recognizes the interest of this type of indicators and suggests that this type of indicators be compared with the changes over time of the status of the stocks as estimated by the conventional assessment models in futures analyses. Furthermore, the Group pointed out the need to assess the accuracy of such indicators by simulation studies, especially considering the possibility of highly variable size at age for certain species.
5.3Other indicators Along this line and consistent with ecosystem approaches, the Group also emphasized the need to further evaluate and understand the relationships between environmental conditions and indicators and tuna fishery success. The Group recommended working toward identifying reliable environmental indicators for explaining tuna availability and abundance. It would also be desirable to better understand the effects of relative price paid for yellowfin, bigeye and skipjack on changes of targeting between these species and other economic factors.
6. Review of progress made on improving the tagging database No specific papers related to analyses of the tagging database for either yellowfin or skipjack tuna were presented. The Secretariat provided the updated status of the tagging database, tables of the current tagging data for skipjack and yellowfin and maps of the tag and recapture locations (ICCAT 2006). The Secretariat has devoted substantial effort to incorporate the suggestions in SCRS/2007/057 to correct problems in the database. However, it was felt that several more months of devoted work on the database is needed to make the database fully useable for yellowfin and skipjack tuna assessments. Nevertheless, the tagging database includes records for 1,711 recaptures of 19,361 tagged yellowfin (0.09%), 6,538 recaptures out of 42,483 tagged skipjack (15%) and 3021 recaptures out of 14,170 tagged bigeye tuna.
There are many applications of this data for estimating mortality rates (Ortiz de Zárate and Bertignac, 2002), improving growth estimates and determining movement rates and stock structure that could be applied to yellowfin and skipjack tuna assessments. In particular the latitudinal variability in growth rates (SCRS/2007/057) may require estimating movement and or/mixing rates between regions. As of now, however, no new information has been obtained from the tagging database to change the current single stock status of yellowfin tuna assumed in the 2003 stock assessment (Anon, 2004).
It was noted that the tagging data are highly desirable to apply the Multifan-CL assessment models (Miyabe et al. 2005). In particular it is necessary to have records of the time and fishery released from for every tagged animal and the time and fishery of the recapture in order to estimate mixing rates and fishing mortalities by fishery. Information on all tagged fish and their fishery of capture is lacking for skipjack and yellowfin, however, this data is available for bigeye tuna, and will likely be used in the upcoming assessment. Further, it is recommended that individual users of the data and individual tagging entities participate in the process of error checking and ground-truthing missing records.
7. General recommendations Recommendations directed to the tropical species groups are contained elsewhere in this report. Recommendations in this section are of a more general nature and addressed to the SCRS or CPCs.
Fleet information. The Group noted that relatively few CPCs are submitting complete fleet size and fleet characteristics information together with Task I data, as required. Fleet information is extremely valuable to species groups and the SCRS because it allows for an adequate characterization of fisheries. It is especially important to have complete historical fleet information in order to understand changes in exploitation over time. The Group recommends that the SCRS consider ways in which existing gaps in fleet information may be filled, and ways in which future reporting may be improved.
Size frequencies. The Group commended the Secretariat's quality control efforts to improve the size frequency data. However, there are limitations about how much the Secretariat can do in terms of data validation. For this reason, it is recommended that national scientists review all of the size frequency data maintained in the Secretariat databases for validation purposes (e.g., checking for outliers, duplicate entries, incorrect gear classifications, etc., and generally ensuring that the database contains what was reported originally). This review should be completed by October 2007. After this deadline, the species groups should decide how to proceed.
Faux poissons. The Group considered that these quantities of unreported catch involved as faux poissons are quite significant, especially for skipjack but perhaps less so for yellowfin and bigeye because the sizes of the fish in question are small. It was recommended that the SCRS take these catches officially into consideration as discards.
8. Other matters In order to facilitate some of the work that needs to be carried out for the MULTIFAN-CL analyses that are planned before the June 2007 bigeye assessment, the Species Group reviewed the definition of fisheries in the model and data availability. A condition to the discussions was the objective to make as few changes as possible to the fisheries that were defined in the 2004 assessment. The following is a summary of the conclusions reached:
− Fisheries. The Species Group recommended splitting the recent (since 1991) purse seine fishery (formerly "Fishery 3") into two separate fisheries, depending on set type (free school vs FAD).Other refinements to fishery classifications were made such as: combining Ghanaian BB+PS into the same fishery; ensuring that catches for all fleets are taken into account in the model (for this purpose, Region 3 catches by surface gears would be added to the catches by Other LL in the same region); ensuring that Spanish BB catches in the ETRO area would be included in the corresponding fisheries. A summary of the fisheries is contained in Table 9.
− CPUE series. CPUE series by year and quarter will be prepared by the Secretariat using Task II data, except the following which will be prepared by national scientists:
1) Japanese LL (series 10, 11 and 12)
2) Brazil LL, separate series for Fisheries 14 and 15.
3) Azores BB, for Fishery 9
9. Report adoption and closure The Chairman again thanked the local hosts for the organization of the meeting. The report was adopted and the meeting adjourned.
ANON. 2007. Report of the 2006 ICCAT Inter-sessional Tropical Species Working Group (Séte, France, April 24 to 28, 2006). Collect. Vol. Sci. Pap. ICCAT, 60(1): 1-90
ARIZ J. and D. Gaertner. 1999. A study of the causes of the increase of the catches of bigeye tuna by the European purse seine tuna fleets in the Atlantic Ocean. Program UE DG-Fish (96/028).
BEVERTON, R.J.H. and S.J. Holt. 1956. A review of methods for estimating mortality rates in exploited fish populations, with special reference to sources of bias in catch sampling. Rapp. P.-V. Réun, CIEM, 140: 67-83.
CAYRE, P., T. Diouf, A. Fonteneau et M.H Santa Rita Vieira. 1986. Analyse des données de marquages et recaptures de listao (Katsuwonus pelamis) realizés par le Sénégal et la République du Cap-Vert. In. Proceedings of the ICCAT Conference on the International Skipjack Year Program. P.E. K. Symons, P.M. Miyake & G.T. Sakagawa (eds.) pp. 309-316.
CAYRE, P. and F. Laloe. 1986. Relation poids-longueur de listao (Katsuwonus pelamis) de l’Océan Atlantique. Proc. ICCAT Intl. Skipjack Yr. Prog. 1: 335-340.
GAERTNER, D., J.P. Hallier and M.N. Maunder. 2004. A tag-attrition model as a means to estimate the efficiency of two types of tags used in tropical tuna fisheries. Fish. Res., 69: 171–180
GAERTNER, D., P. Kebe and C. Palma. 2007. Some clues for correcting the tagging database of tropical tunas. Collect. Vol. Sci. Pap. ICCAT, 60(1): 185-189.
HERVE A., R. Dédo, S. Diomande et A. Gobo. 2004. Actualisation des quantités de "faux-poissons" débarquées par les senneurs à Abidjan de 1981 à 2004. Collect. Vol. Sci. Pap. ICCAT, 56(2): 443-527.
MIYABE, N., Y. Takeuchi, H. Okamoto and V.R. Restrepo. 2005. A new attempt of Atlantic bigeye tuna (Thunnus obesus) stock assessment by statistical integrated model (MULTIFAN-CL). Collect. Vol. Sci. Pap. ICCAT, 57(2):177-200.
OKAMOTO, H. 2007. Standardized Japanese longline CPUE for yellowfin tuna in the Atlantic Ocean from 1965 up to 2005. Collect. Vol. Sci. Pap. ICCAT, 60(1): 342-350.
ORTIZ DE ZÁRATE, V. de & M. Bertignac. 2002. Analysis of tagging data from North Atlantic albacore (Thunnus alalunga): attrition rate estimates. Collect. Vol. Sci. Pap. ICCAT, 4(5): 1438-1453.
ROMAGNY B., F. Ménard, P. Dewals, D. Gaertner et N. N'Goran. Le “ faux-poisson ” d’Abidjan et la pêche sous DCP dérivants dans l’Atlantique tropical Est : circuit de commercialisation et rôle socio-économique. In « Pêche thonière et dispositifs de concentration de poissons. Ed. Ifremer, Actes colloq, 28, 634-652.
SHUFORD, R.L., J.M. Dean, B. Stequert, E. Morize. 2007. Age and growth of yellowfin tuna in the Atlantic Ocean. Collect. Vol. Sci. Pap. ICCAT, 60(1): 330-341.