Christofer H. Boggs
Southwest Fisheries Science Center
National Marine Fisheries Service, NOAA
Honolulu, Hawaii 96822 USA
A review of the Proceedings of the First FAO Expert Consultation on Interactions of Pacific Tuna Fisheries reveals many unresolved issues. Important problems fall into several broad categories: 1) inadequate collection or analysis of fisheries data, 2) inadequate biological and ecosystem data to accurately model fishery interactions, 3) problems with tag and recapture experiments, 4) lack of alternative methods for quantifying fish movements, and 5) a narrow view of fishery interactions. Potential suggestions for action include: 1) improve basic fisheries data collection and stock assessment, 2) expand and improve tag and release programs, 3) develop new methods for tracing fish movements, and 4) simulate best-case and worst-case fishery interaction scenarios and economic impacts for the purposes of risk assessment and discussions of resource allocation.
A goal of the first two fishery interaction consultations was to assess the extent and severity of impacts resulting from sequential fishing by several fisheries on the same (or dependent) resources. The purpose of this paper was to present outstanding issues and problems remaining from the previous consultation (see Shomura et al., 1994). These problems and shortfalls hinder the estimation of fisheries interactions. Questions were posed to trigger discussion. Section 7 of the summary report from the second consultation (Shomura et al., 1995) covers the responses of the consultation participants.
The perspective set by the first consultation was that important issues are those with practical relevance to fisheries management. This perspective was maintained here. As noted in Section 7.1 of the summary report (Shomura et al., 1995), many critical and recurrent problems (i.e., inadequate fishery statistics, inadequate biological and ecosystem data, problems with tagging) are very similar to those in stock assessment research. If such problems prevent the estimation of fishery impacts on a resource (i.e., stock assessment), those same problems can prevent the estimation of fishery interactions among users of the resource. Most of the problems and information gaps have been clearly and thoroughly described in the proceedings of the first consultation and in numerous workshops and symposia on fisheries management. In the interim between the first and second consultations, researchers have done an excellent job of focusing on relevant issues, but many problems and shortfalls remain.
2. PROBLEMS AND QUESTIONS
Many of the problems and issues were perennial and very familiar to participants of the second consultation, which may have inhibited lively discussion. Prior to the consultation the participants were asked to identify issues that had been adequately addressed or problems that had been truly solved, but none did so. It seemed to be unanimous that better information is needed in almost every situation. Realistically, choices must be made regarding priorities for further work. Each problem described below is followed by a trigger question regarding that problem. To provoke discussion and prioritization, the questions were often phrased as choices between alternatives.
2.1 Inadequate Collection or Analysis of Fishery Data
Problem 1: Species targeting, gear characteristics, and gear efficiency are often inadequately described or quantified. The solution to this problem involves a much greater emphasis on the collection of more and better fisheries statistics. However, recent attempts to standardize purse seine catch-per-unit-effort (CPUE) data in the western Pacific have proven to be discouraging and it has been suggested that such attempts may be futile (Anonymous, 1994).
Question 1: Are meaningful measures of relative abundance possible through the collection and analysis of catch and effort data? For some or all fishing methods, are CPUE indices hopelessly confounded by changes in fishing strategy, technique, and other fishery performance factors so that they cannot be corrected to represent fish abundance?
Problem 2: Small scale commercial, artisanal, and subsistence data for many fisheries do not exist or have only recently been collected. Projects have been initiated to retrieve and analyse historical data from such fisheries (He and Boggs, 1996), but useful results have sometimes been obtained quickly and economically through analysis of existing tag recapture results and simulation (SPC, 1994).
Question 2: How much effort should be expended to recover useful historical data from localized, small scale or artisanal fisheries? Is it more practical to focus on improving current statistics and to model interactions in such fisheries based on tagging experiments?
Problem 3: Better species identification, and more (and better) size data are needed from fishery data collection programs.
Question 3: What are the most important categories of fishery data that are not being collected adequately? Assuming similar levels of research funding, what activities could be viewed as less important?
Problem 4: Many Pacific tuna stock assessments are tentative and inconclusive. Spatial distribution, fish movements, and fish size distribution need to be incorporated in assessments.
Question 4: What are good examples of adequate tuna stock assessment, and what additional data are most needed to conduct such assessments for tuna stocks in the Pacific? Are we limited mostly by insufficient information or are access to the data and a lack of human resources to do the work equally important?
Problem 5: In many areas the fisheries statistics available to any one nation or agency are incomplete due to confidentiality rules that prevent the release of data or that require aggregation of data into a summary form that hinders useful analysis. Confidential data may often be exchanged through the establishment of specific collaborative research projects between nations or agencies. However, no comprehensive program exists to enable mutual exchange or centralized compilation of detailed fishery statistics from all major Pacific tuna fishing nations.
Question 5: Would the establishment of a regional monitoring and management authority to exchange data on all tuna fisheries throughout their range speed up the assessment of stocks and the quantification of fishery interactions? Alternatively, is the current practice of arranging collaborative data exchanges to accomplish specific research goals adequate and appropriate given the very limited human resources available to conduct the needed analyses?
2.2 Inadequate Biological and Ecosystem Data
Problem 6: Although the importance of discriminating between stocks is obvious, the practical meaning of genetic stock structure studies conducted to date is unclear. Lack of genetic differences between groups does not necessarily imply that a single stock should be assumed for management purposes. For example, a lack of genetic heterogeneity between areas (i.e., between fish from the Atlantic and Pacific Oceans) implies that there is a single genetic population. However, movement of a few fish per generation between these areas can produce genetic homogeneity. When gene flow is substantial but few fish actually move between areas, separate management is justified. Conversely, observation of genetic differences between groups does not necessarily imply that separate stocks should be assumed for management purposes. A number of mechanisms (e.g., spawning site fidelity) may contribute to geographic heterogeneity in genotypes even when substantial movement of fish between areas is documented. Although genetic researchers do find geographic differences, few recommend actual population boundaries for stocks because the geographic resolution of their work is usually limited to a handful of locations and because the practical implications of their results are vague. Before much larger scale genetic studies are conducted researchers should clarify the meaning of results of such studies.
Question 6: What is the relevance of genetic studies to the question of whether or not separate fisheries exploit a common resource? What testable hypotheses relevant to fishery management can be generated from existing data? Could hypothesis testing be conducted on more tractable populations as a proxy for tuna?
Problem 7: Various growth rate estimates exist for tunas from different parts of the Pacific. Differences are found between growth rates estimated by analysis of modes in size-frequency distributions and growth rates based on otoliths or other hard parts. Differences are also found using similar methods in the same area. Still, growth rate estimates are probably much less variable than many other parameters used in stock assessments. Much emphasis is placed on the importance of size-frequency data in the more sophisticated stock assessments and analyses of fishery interaction, and collection of more and better size data has been recommended. These data are important and useful without independent estimates of growth rate. However, resolving the differences among independent growth rate estimates would allow stock assessment and fishery interaction models based on size-frequency data to be constrained in useful ways (e.g., by ruling out faster or slower growth rates than those resulting from hard part studies).
Question 7: Should growth rate studies receive much further attention, or are the differences between methods and areas acceptable given the poorer quality of information on other vital rate statistics?
Problem 8: Natural mortality is wrongly assumed to be independent of fish size. In tagging studies natural mortality is estimated from the decline in tag recoveries over time (i.e., total mortality) under different levels of exploitation (i.e., fishing mortality). However, natural mortality is difficult to distinguish from fish movement away from fished areas. Moreover, differences in natural mortality with fish size are difficult to estimate because in comparison with small fish, large fish are more difficult to tag. Also, large fish may have different vulnerabilities to capture and different movement patterns from small fish. Size-specific differences in natural mortality could be especially important when fishery interaction models based mostly on tag results from small fish are extrapolated to whole stocks. Some fisheries are highly dependent on larger fish, and the modelled effects of fishery interactions on such fisheries could be erroneous.
Question 8: How can natural mortality and total mortality of older tunas be estimated? How sensitive are fishery interaction models to size-specific differences in natural mortality? Can we assume that the differences are so small, or that the models are so robust, that the problem may be ignored?
Problem 9: For most tuna stocks, little is known regarding the ecological causes of variation in abundance among age classes. Cannibalism, interspecific predation, competition, and trophic dynamics probably affect recruitment, growth, reproduction, and mortality of tunas. For some stocks it has been hypothesized that strong year-classes occur in alternating years or that strong year classes accompany or follow major environmental perturbations, but the biological mechanisms for such relationships are not well known. Generally, single-species models are used for stock assessment and interactions analyses. The number of unknown ecosystem variables is too great for ecosystems models to have had much practical application to fishery interaction problems. A definitive study of the food web dynamics of the pelagic ecosystem would be a massive undertaking.
Question 9: Do priorities dictate continued deferment of multispecies ecosystems approaches to explaining the dynamics of the exploited resources, or are there particular fisheries and stocks where such methods could be practically applied?
Problem 10: To detect and quantify the impacts of fishery interaction against the background of natural variability, we will need a better understanding of the effects of oceanographic variation on fisheries. Oceanographic studies, especially those utilizing remote sensing technologies combined with ocean circulation models, are needed to explain variation in fishery performance, allow closer estimation of trends in fish abundance, and permit better analyses of fishery interaction. However, the mechanisms linking oceanographic variables to tuna abundance, distribution, and catchability are probably trophic, complex, and potentially chaotic. It may be that substantial natural variation in abundance, distribution, and catchability will remain unexplained and fishery interactions will remain difficult to detect or quantify. Many efforts to quantify the effects of fishing intensity on fish abundance may continue to treat natural variation as noise.
Question 10: If fishery impacts on stocks are so small relative to natural variations that the fishery impacts cannot be detected, then are the fishery impacts very important? Natural variations in abundance may exhibit more noise and less trend than fishing power or species targeting. The latter tend to exhibit coherent trends that oppose and mask declines in abundance. Would it be more practical to direct our efforts to ensure that fishery data adequately quantify a useful standard of fishing effort than it would be to try to account for natural variability in fish distribution and abundance?
2.3 Problems With Tag and Recapture Experiments
Problem 11: The low rate of yellowfin tuna tag recoveries from longline gear compared with surface gear suggests that there could be a segregation of fish between deep and shallow habitats at some stage of life. Analysis of the sparse returns of yellowfin in the larger size groups important in the longline fishery suggests that the returns from this fishery are lower than expected. Biological studies may shed some light on this issue by indicating other differences between surface- and subsurface-caught yellowfin tuna. However, more tagging data on large tuna of all species would be useful for many reasons, including the estimation of size-specific mortality. Research longline fishing indicates that large tuna caught by longline gear can be tagged and recovered, as can large tuna caught by recreational fishermen.
Question 11: Could the tagging of thousands of large tuna ever be practical? Would a cooperative program involving the agencies conducting research longline operations together with recreational fishermen be able to tag and release useful numbers of large tuna in the Pacific?
Problem 12: The trivial rewards given to those who recapture and report tagged fish probably contribute to a low reporting rate of recaptures. Although assessments are robust to non-reporting as long as the rate is somewhat stable, the effectiveness of tagging programs is diminished in direct proportion to non-reporting. A clear outcome of the previous consultation was that tagging programs were among the most effective means of studying fishery interaction. Such programs are so costly that it would seem to be relatively cost-effective to increase the rewards offered for tags, or to make greater use of lotteries for those who report tag recoveries.
Question 12: Should the consultation recommend that a fund and/or a cooperative program be established to increase the rewards offered for tag returns?
Problem 13: More work is needed to characterize the statistical properties of estimates of fish movements based on tag return data using the various models that have been devised for that purpose. The previous consultation strongly recommended that the same tag return data be analysed using several of the existing models to explore differences due to technique. Additional analyses are needed to examine how stable apparent patterns of directed movement may be from one time period to another. Although most tagging programs are too expensive to continue indefinitely, existing results should be closely examined with regard to time trends in directed components of movement. It may be that diffusive components of movement are more consistent than directed movements, and that movements that appear to be directed in individual years are actually more random when analysed in aggregate over many years. Perhaps only directed movements which are demonstrably repetitive should be used in simulations of interaction between geographically separate (or partially overlapping) fisheries.
Question 13: Are the movement patterns determined from tag-recapture experiments reproducible when analysed using different models? And are the patterns repetitive over time? In fishery interaction simulations, might it be useful to use both the observed pattern of tuna movements as well as an alternative pattern in which the directed components are made random (i.e., the directed components are not repetitive over time)?
2.4 Lack of Alternative Methods for Quantifying Fish Movements
Problem 14: Alternative methods are needed to survey the movement patterns of tuna that are difficult to tag in large numbers, and to document movements in areas where tuna are seldom caught. Tag and recapture experiments have proven to be invaluable in the study of fishery interaction, but there remain aspects that have been difficult to address using this technique, such as the movements of larger tuna, the possible segregation of yellowfin tuna by depth, whether or not observed directed movements are repetitive, whether they are segments of true migrations, and what sort of movements may be under-represented as tuna enter areas where fishing is not intense or vulnerability to fishing gear is reduced.
More tagging experiments are needed in a host of locations and fisheries, yet the cost is daunting. The largest and most useful tagging programs have been directed towards the most easily tagged species and sizes of tuna but have nonetheless been quite expensive. Cost-effective alternatives to large-scale tagging experiments would be very useful. However, tagging experiments have the dividend that they also allow the estimation of mortality, whereas the existing alternative approaches to studying movements do not.
Alternative methods of studying fish movements include surveys that map changes in observed abundance (e.g., visual, video, acoustic, laser, and radar surveys) and methods that track individual fish (e.g., sonic tracking and archival tags, Problem 15 below). Some of the survey methods might be used to follow individual tuna aggregations or to measure instantaneous movement vectors. More often, the survey methods rely on the assumption that fish move as the patterns of high abundance move, whereas the patterns may simply reflect aggregation and dissipation of fish due to other processes.
Question 14: Are there any promising new technological developments that suggest feasible survey methods for documenting tuna movements? Considering the benefits of conventional tagging studies as well as their high cost and limitations, could any of these survey methods be more cost-effective than tagging programs?
Problem 15: Alternative methods for documenting tuna movements through the study of individual fish are needed. Sonic tracking of individual fish is generally too short in duration to provide useful data on long-term movements. This method could be useful for observing behavioural segregation of large yellowfin tuna into either the deep or the shallow zones exploited by longlines or purse seines (respectively), but would not indicate when the habitat segregation originated or how vulnerable each type of yellowfin might have been to different fisheries at an earlier stage in life (i.e., type B interaction of Hampton, 1994). Moored arrays of hydrophones could be useful in monitoring longer-term migrations of tuna tagged with individually recognizable transponders, but the arrays would have to be huge. Alternatively, tuna could carry archival tags that would store movement data while out of the range of moored hydrophones. Archived data would be downloaded when the fish encountered one of the widely scattered moorings. Archival tags appear the most promising as a method of recording long-term movements, but because tag recapture rates are very low for most species, a fishery-independent archival tag data recovery system may be required.
Genetic and chemical studies of individual fish might also provide valuable information on past movements or origins of tuna. If spawning site fidelity exists, genetic tests might indicate the origins of fish. Alternatively, the absence of movement between broad areas might be established by reliable classification of individuals into geographically distinct genotypes (but see Problem 6). Chemical microconstituents concentrically deposited in otoliths could be so site-specific that their sequence might reveal an origin and history of movement. Alternatively the microconstituents may be so dependent on a multitude of variables (i.e., food and temperature) as to be undecipherable.
Question 15: Are there promising new technological developments that may allow tracking (or backtracking) of individual fish movements over the long term? Again, considering the benefits of conventional tagging studies as well as their high cost and limitations, could any of these methods be more cost-effective than tagging programs?
2.5 Narrow View of Fishery Interactions
Problem 16: Fishery interactions studies have mostly focused on marginal interactions rather than on absolute interactions (Kleiber, 1994), which may have minimized the apparent magnitude of fishery interactions. Marginal interaction is the effect of variations in the intensity of one fishery on the performance of another, whereas absolute interaction is the sum effect of the existence of one fishery on the performance of another (Kleiber, 1994). Because variations in fishing effort are generally much less than the total fishing effort by an entire fishery, marginal interactions are likely to be less intense than absolute interactions. For many fisheries reliable data may exist only for recent times (or not at all). In other cases traditional fisheries no longer exist, and it is not known whether or not fishery interaction contributed to their demise.
Data on the performance of subsistence or artisanal fisheries before the large, distant water fisheries became established almost never exist, so that empirical detection of absolute interaction is unlikely. Declines in the apparent abundance of tuna characterized the development of longline fishing in the Pacific, as was to be expected as the level of exploitation increased. One way to address absolute interaction in such cases is through simulation, as was recently done for the fisheries of Kiribati (SPC, 1994), where the effect on local fishery performance of eliminating the entire Pacific purse seine fishery was explored.
Question 16: Should greater attention be given to estimating absolute, rather than marginal interaction, especially in the case of interactions between very large fisheries and artisanal or subsistence fisheries?
Problem 17: The importance of detecting fishery interactions even when the interactions are small was discussed at the previous consultation. It was suggested that finding and quantifying such interactions were important in order to be able to predict larger interactions that could occur as fisheries expand. However, predicting larger interactions at increased levels of effort based on poorly quantified relationships at lower levels of effort may be misleading.
Question 17: When a fishery interaction is not clearly evident and likely to be weak, how appropriate or valuable is it to search harder for empirical evidence of the hypothesized interaction? How useful is a projection based on empirical evidence of a small interaction? If the concern is over a larger interaction that may occur at higher effort levels, why not allow the fishery to grow until such an interaction becomes clearly evident?
Problem 18: At the start of the first consultation, the participants agreed that fisheries economics issues were beyond the scope of that meeting. Economic studies, however, remain extremely relevant to the study of fishery interactions. Fishery interaction issues often arise well before the population of the exploited resource is threatened, and while the risks and benefits of management alternatives are more economic than biological. When population stability is not the issue, assessment of the risks involved in allowing one fishery to expand at the detriment of another should emphasize the economic costs and benefits of each type of fishing. Furthermore, market interactions between fisheries (as commodity suppliers) may be more easily detected and quantified than resource-mediated (biological) fishery interactions.
Question 18: Should greater attention be given to the economic interactions between tuna fisheries as well as the economic costs and benefits of tuna fisheries?
The participants reactions to the trigger questions (see Section 7 of the summary report, Shomura et al., 1995) may reflect a much broader understanding of the problems and issues than those of this author. Nevertheless, I offer the following preliminary generalizations (in advance of the second consultation).
Fishery data collection activities must be expanded and improved. Such activities are old-fashioned, rarely provide results publishable in peer-reviewed literature, and are not considered to be research by the academic community. These activities are essential, however, to evaluating fisheries interactions. Understanding of oceanographic and ecological factors affecting resource dynamics and fishery performance may be greatly improved, or these factors may remain poorly understood sources of noise. Either way, measures of fishery performance based on fishery data will continue to be essential in evaluating the need for management. The most important requirement for fishery data collection, beyond defining the scale and location of each fishery and the species and sizes of fish caught, should be to provide measures of fishery performance that account for changes in fishing power, gear efficiency, and targeting.
Tag and recapture experiments have proven to be very valuable and complementary to analyses of fishery statistics in identifying and quantifying fishery interactions. No other research method has provided as much useful information or has been the basis of so many practical applications in this field. However, there are very important issues involving fisheries for larger tuna for which tagging experiments have fallen short of success. A major effort to tag larger tuna should be initiated, or alternative methods for measuring the movements and mortality rates of larger tuna must be developed. Simulations may provide the best means of resolving fishery interaction issues. Reasonable upper and lower bounds of possible movement rates, and natural mortality, along with the best available fishery statistics, should be used to simulate each potential interaction. Economic information should be used to assess the risks implied by the best- and worst-case simulations, to explore the consequences of allocation alternatives, and to assess the potential benefits of regulating the fisheries.
4. REFERENCES CITED
Anonymous. 1994. Report of the third meeting of the Western Pacific Yellowfin Tuna Research Group, Pohnpei, Federated States of Micronesia, 21-23 June 1993.
Hampton, J. 1994. A review of tuna fishery-interaction issues in the western and central Pacific Ocean. In: Shomura, R.S., J. Majkowski and S. Langi (eds.). Interactions of Pacific Tuna Fisheries. Proceedings of the First FAO Expert Consultation on Interactions of Pacific Tuna Fisheries, Noumea, New Caledonia, 3-11 December 1991. FAO Fish. Tech. Paper (336/1): 138-157.
He, X., and C.F. Boggs. 1996. Do local catches affect local abundance? Time series analysis on Hawaii's tuna fisheries. In: Shomura, R.S., J. Majkowski and R.F. Harman (eds.). Scientific Papers from the Second FAO Expert Consultation on Interactions of Pacific Tuna Fisheries, 23-31 January 1995, Shimizu, Japan. [This volume]
Kleiber, P. 1994. Types of tuna fishery interaction in the Pacific Ocean and methods of assessing interaction. In: Shomura, R.S., J. Majkowski and S. Langi (eds.). Interactions of Pacific Tuna Fisheries. Proceedings of the First FAO Expert Consultation on Interactions of Pacific Tuna Fisheries, Noumea, New Caledonia, 3-11 December 1991. FAO Fish. Tech. Paper (336/1): 61-73.
Shomura, R.S., J. Majkowski and S. Langi (eds.). 1994. Interactions of Pacific Tuna Fisheries. Proceedings of the First FAO Expert Consultation on Interactions of Pacific Tuna Fisheries, Noumea, New Caledonia, 3-11 December 1991. FAO Fish. Tech. Paper (336/1-2).
Shomura, R.S., J. Majkowski and R.F. Harman (eds.). 1995. Summary Report of the Second FAO Expert Consultation on Interactions of Pacific Tuna Fisheries, 23-31 January 1995, Shimizu, Japan. FAO Fish. Rep. 520: 58 p.
SPC (South Pacific Commission). 1994. Oceanic Fisheries Programme work programme review 1993-94 and work plan 1994-95. Working Paper 5, Seventh Standing Committee on Tuna and Billfish, Koror, Republic of Palau, 5-8 August 1994.