Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 131
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA 4 Findings and Recommendations Significant advances have been made in the United States in recent years in the collection, management, and use of data related to marine fisheries. The National Marine Fisheries Service (NMFS) and the states have experimented with new technologies, such as vessel monitoring systems, electronic logbooks, acoustic fish detection, the Global Positioning System (GPS), electronic zebra code scanners, and modern tagging methods. Despite these advances, traditional fisheryindependent surveys seem to be underfunded in some areas, commercial data are underused, and both may sometimes not be of high enough quality for their intended uses. Congress has responded to the need for surveys by providing funds for new and more capable fishery research vessels, but budgets for using survey vessels are under continual pressure. Fisheries data collection, management, and use will suffer without provision of adequate, sustained funding from Congress that is applied appropriately by the Department of Commerce (DOC), the National Oceanic and Atmospheric Administration (NOAA), and NMFS. The committee did not analyze the fishery data collection budgets of the states or NMFS, so it cannot determine whether they need budget increases or could use their existing resources more efficiently. It does seem clear, however, that NMFS is being asked to provide more data to diverse groups for a variety of purposes. Major advances in fisheries data management could be achieved by continuing the use of new computing and communication capabilities and increasing integration and standardization of data management on regional and national bases. NMFS and its state and regional partners on the U.S. Atlantic coast are implementing the Atlantic Cooperative Coastal Statistics Program (ACCSP), which seems to be a good model for regional data management systems because of its emphasis on data standardization and full information access by the program partners, including federal and state agencies, and participation by interstate commissions, commercial and recreational fishermen, and environmental advocates. At the direction of Congress, NMFS developed a plan for a nationwide Fisheries Information System, an umbrella data management system that will incorporate existing and planned regional systems and help them coordinate data standardization and access. Such new data management systems have become even more impor-
OCR for page 132
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA tant as the most recent reauthorization of the Magnuson-Stevens Act added new requirements for identification of essential fish habitat. The committee drew on input from its public meetings, published documents, and its own analyses and discussions to develop a series of findings and recommendations that (1) analyze the 1999 summer flounder stock assessment and provide advice to NMFS about ways to improve these assessments (see Appendix D) and (2) identify more general data issues needing attention. Many of the recommendations related to the summer flounder assessments should be applied more broadly in other U.S. fisheries; relevant examples will be provided later in this chapter. Improvements in the credibility of NMFS data collection and stock assessment procedures could be achieved if the following recommendations were implemented. Some of the major issues that emerged during the committee's work were the timing of data availability to managers and fishermen and the waste of the great potential of commercial and recreational fishermen as data collection partners. Delay in the availability of data was a major theme of the committee 's discussions. In particular, fishery-dependent data, especially commercial and recreational landings data, must be available on time scales that match needed management actions. The degree of timeliness needed varies depending on the type of data and the management system used. For example, for management based on total allowable catch, total catch data and data on discards in other fisheries should be available soon enough to allow closing a season early. Presently, most commercial data meet this criterion, but recreational catch data generally are not available on this time scale. Management with closed areas is best accomplished with real-time information about vessel location. The expansion of the recreational share of many fisheries has exacerbated the data timeliness problem because data collection systems for recreational fisheries are not designed to make data available in a usable form quickly, so that in-season1 management of most recreational fisheries is presently not possible. State and federal budget limitations will probably always constrain fishery-independent surveys. The portion of the NMFS budget for research and data collection has been diminished in recent years from budget cuts and “earmarks,” without positive adjustments to base budget that would offset personnel costs and other costs increased by inflation. Commercial and recreational fishermen are a large potential source of data about the fish stocks they exploit. The committee believes, therefore, that it is imperative for NMFS and the councils to improve the quality of data available from commercial and recreational fisheries. NMFS finds it difficult to use certain forms of commercial data, especially measures of fishing effort and CPUE, because of valid concerns about the data's usefulness; the committee believes these concerns could be better addressed through constructive engagement with the commercial sector. Improvements in data quality will occur only if NMFS and industry work cooperatively to create more effective and efficient data collection and management systems and create an environment that fosters the availability of accurate, precise data with adequate protection of privacy and confidentiality. Other stakeholders, such as environmental advocates and ceremonial and subsistence users should also participate in data collection and management so that their knowledge and interests are considered. Responsibility for the current failure to use certain kinds of fishery-dependent data (e.g., commercial catch rates and logbook data such as the landings by species and locations of catch) for stock assessments can be ascribed to both NMFS and the industry, and both need to make good-faith efforts to work together to improve data availability, thereby fostering improvements in management and sustained fisheries yields. 1 In-season management refers to changes in the season length based on a sector's actual catch in relation to the total allowable catch allotted to that sector.
OCR for page 133
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA IMPROVING DATA COLLECTION Matching Data Collection Costs to Benefits from Fisheries Findings: In an ideal world, sufficient resources would always be available for data collection, but resources have never been adequate to meet this ideal and NMFS has not been able to obtain satisfactory increases in its data collection budget. The committee could find no existing analyses of the costs and benefits of data collection and management for specific fisheries, particularly of the ratio of marginal costs and marginal benefits for each additional dollar spent on data collection. The potential value of management advice (e.g., in terms of decreasing costs or increasing the long-term potential yield of fisheries) may not be related to the frequency and cost of surveys and other data collection and management systems (e.g., logbook programs, port sampling, observer programs). As a consequence, it is not clear whether allocation of data collection resources is based on objective analysis, or instead reflects regional tradition, bureaucratic inertia, or political incentives. Mismatches between costs and benefits may exist because social and economic factors are not accounted for properly. A comparison of data collection and management costs and the marginal benefit of such expenditures for fisheries management would allow NMFS to allocate its data collection effort more effectively within and across fisheries. Data collection could be made more costeffective through changes in the allocation of survey effort, collection of commercial and recreational catch and effort statistics, and optimization of regional monitoring through observer programs and vessel monitoring systems. NOAA recognized the need for such a costbenefit approach in its 1998 Fisheries Data Acquisition Plan: Because fiscal resources for fisheries management are not unlimited, a hierarchical system of priorities must be set to ensure that the most important data needs are met. First priority must go to endangered or threatened species, [examples given include marine mammals and overfished stocks] Another consideration when setting priorities is the value of the stock or species to the U.S. economy. However, the ecological importance of a species is also considered. Surveys may be conducted on a species of little economic value, but is an important forage fish for, or predator on other stocks, or is a bycatch species (NOAA, 1998). The status of fisheries in the U.S. Exclusive Economic Zone as national public-trust resources suggests that a nationwide prioritization of expenditures would be desirable. Such an exercise would make costs and benefits more explicit and would give NMFS a more objective basis for changing data collection intensity or requesting additional resources. Implementation of a costbenefit approach to data collection and management will be hindered if sufficient resources are not appropriated by Congress for necessary analysis and implementation. Recommendations: NMFS should allocate its data collection resources by some rational plan between its own data collection efforts and the efforts of others. Data from non-NMFS sources will require special attention to such matters as data quality and coverage (by fishery, species, times of year, and location). Reallocation of data collection resources may not result in a net cost savings. Congress should encourage NMFS to conduct a nationwide analysis of the costs and benefits of optimizing data collected for each fishery, including the value of fish stocks for commercial, recreational, and non-consumptive uses. Analyses should include appropriate multipliers to capture benefits of recreational and commercial fisheries to the broader economy (e.g., bait and tackle purchases, boat rentals, sales by fish dealers and retailers), as well as ecosystem benefits. For example, a table such as the following might be constructed:
OCR for page 134
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA Species or Species Complex Potential Value of Harvest and Other Benefits Spawning Stock Biomass Long-Term Potential Yield Importance of Species in Food Web/ Endangered Status Precision Needed Estimated Data Collection Cost (itemize by data source) Such a table could be an extension of the information on recent average yield, current potential yield, and long-term potential yield given in the NMFS Our Living Oceans reports (e.g., NMFS, 1999). The primary intent of such an analysis would be to guide the federal investment in fisheries data collection and management. States could use a similar approach to evaluate the costs and benefits for individual fisheries at the state level, which may be only partially related to federal costs and benefits. An interim measure could be to try the approach for one major fishery on each coast or within each council region. The key point is that benefits and costs of fisheries data collection and management need to be measured in common scales across fisheries, whether in dollars or in other scales that can be used to quantify noneconomic costs and benefits. Quantifying environmental benefits of data collection may be difficult and proxy measures may be needed so that environmental benefits can be included in the total benefit of data collection. Examples of proxies might include the standing stock biomass and information about the importance of a species in the food web. Such an approach should consider the potential value of recovery of an overfished stock. This type of analysis would allow more informed decisions to be made about which fisheries merit increased funding for data collection and management. As part of the recommended cost-benefit analysis, the precision of each data source should be determined, because it will not be cost-effective to measure one input (e.g., commercial or survey CPUE) with great precision, while only measuring another input approximately, because the precision of the output (e.g., estimated fishing mortality) and the management that is possible will often reflect the precision of the most imprecise input. An analysis of the costs and benefits of data collection and management should also identify areas of research needed to make better cost-benefit decisions. The federal government may be subsidizing some fisheries by spending more on data collection and management than the fisheries are worth to the nation. When this is the case, the government could decide that such expenditures are not cost-effective. If this led to cut-backs in data collection and required more conservative management, industry should have the option to pay for data collection that might (or might not) allow higher TACs. Otherwise, where data collection and management costs exceed benefits to a specific fishery, less data-intensive management (reflecting perhaps more biologically conservative actions) should be considered by the regional fishery management councils. This review should be updated every 5 to 10 years to account for changes in the value of the fisheries, the development of more cost-effective techniques for data collection and management, and improved understanding of specific fisheries. An important contribution to this cost-benefit analysis would be quantification of the costs and benefits of different data sources for different species. This information would allow determination of the appropriate mix of data types and sources, for example, sampling of landings, observers, vessel monitoring systems (VMSs), logbooks, and dealer reports. The appropriate mix
OCR for page 135
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA may be different for each fishery, but VMSs may be particularly useful in fisheries managed with closed areas, observers are useful for all fisheries in which bycatch is a problem, and logbooks are useful in all areas. The committee's recommendations assume that funding will constrain most fishery-independent surveys to groups of species, although additional funding would allow NMFS to examine how well its surveys characterize individual species within multispecies complexes. Surveys almost always focus on groups of species because NMFS does not have the resources (financial, personnel, ships) to optimize surveys for individual species, and the value of most individual species does not merit such an approach. This creates a dilemma: each survey will be sub-optimal for any particular species. How can this situation be remedied in the context of limited resources? If Congress, NMFS, the councils, industry, or environmental advocacy groups believe that better data are needed for any single species, (1) funding could be shifted from other sources or new funding appropriated, (2) fishermen who target the species could be asked to contribute to directed surveys of it (through taxes or in-kind contributions of services in helping with the surveys), (3) NMFS could find new ways to improve commercial and recreational data and use these data more extensively in stock assessments, and/or (4) new methods and technologies, such as acoustic surveys or adaptive sampling, could be implemented. Greater Use of Fishery-Dependent Data Findings: Data collected from scientifically designed surveys are often contrasted with that collected from fishery operations. In fact, both have benefits and drawbacks as data gathering mechanisms. Generally, greater control is exercised over the gathering of data from scientific surveys. For a scientific survey, a full statistical design can be implemented, accounting for potential sampling biases. Unfortunately, the cost per sample is large, often resulting in a low sample size, thereby reducing the precision. Loss of precision leads to uncertainty through greater variation in the estimates. The Marine Recreational Fisheries Statistics Survey (MRFSS) is an example of a scientifically-designed survey that can provide sufficient precision for most modes at the regional level for the recreational component of many fisheries. However, it estimates only what is caught by anglers and does not monitor the fish population directly. Preferences of anglers will motivate them to fish in a manner that biases the catch toward their desired species and fish size. Data from commercial fishery operations (e.g., from logbooks, observers, port sampling) do not represent random samples of the fish population. Rather, such data reflect the characteristics of those portions of the fish population that are subject to harvest. Cost per fish sampled is low relative to the cost of a survey and data available from intensive fisheries resemble a census more than a survey sample. Thus, the effective sample size for commercial data may be orders of magnitude larger than what may be available via surveys. The fact that it is a fishery-directed harvest in which fishermen are pursuing their own goals, however, means that only a portion of the fish population may be targeted and that targeting may change with time. As a consequence, data from fishery operations present a biased perspective of the population (e.g., fisheries target high concentrations of adult fish), and perhaps more importantly, this bias may change over time and not correlate well with actual fish abundance. Part of the bias arises from not understanding how social and economic factors affect when, where, and how fish are harvested. The lack of control over fishery targeting practices (i.e., lack of structured sampling design applied to the population) makes it necessary to account for potential biases when such data are used in stock assessment procedures. The committee believes that it would be more cost-effective to find ways to improve the collection and use of data from commercial and recreational fisheries in stock assessments than for the
OCR for page 136
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA government to conduct vastly increased surveys, although there are areas in which surveys need to be increased. Data obtained from commercial and recreational fisheries can be a valuable resource if their inherent biases can be understood and accounted for in the assessment process. The value of commercial data is not limited to population assessment alone. These data may also prove useful for tracking ecosystem changes or to monitor the effects of fishing regulations over time. Such effects are not always apparent from fishery-independent surveys, due to their directed objectives and cost restrictions. Recommendations: NMFS and the councils should invest in finding ways to improve data from commercial and recreational fisheries to make these data more useful in stock assessments, rather than establishing new fishery-independent surveys. Existing surveys should be made more cost-effective by incorporating new technologies and management methods. In implementing this recommendation, NMFS will need to understand, account for, and reduce (if possible) the inherent biases in fishery-dependent data of different types. Use of fishery-dependent data also should be guided by the evaluation of costs and benefits recommended earlier. Minimizing and Accounting for “Data Fouling” Findings: Data fouling is a serious problem that can result from specific types of management. Also, any change in management regime can cause data fouling by changing the spatial and temporal extent of fishing discard rates, misreporting, and other factors. For example, trip limits established in the summer flounder fishery changed the areas targeted by commercial fishermen. Recommendations: Assessments should take into account the effect of regulations on how fishermen conduct their operations and how this could change the composition of fish caught and data available for stock assessments and other purposes. More fundamentally, councils should explicitly consider the effects that proposed regulations and management regimes would have on data quality and should attempt to design systems that will achieve their management goals with as little data fouling as possible, for example, by implementing or expanding observer programs. NOAA (e.g., NMFS, Sea Grant) should support both internal and external research to identify and evaluate incentives for accurate reporting and disincentives for misreporting and to study the effects of regulations on the industry. In particular, NMFS should engage more social scientists to help build the knowledge base needed to move management beyond trial-and-error to a more predictive capability. Emphasis should be on the relationship between different types of regulatory approaches and fishermen's attitudes and behaviors toward fish harvesting and data reporting. Fishery-Independent Surveys Survey Design Precision of Survey Data Findings: Once in place, long-term fisheries survey programs are rarely evaluated to determine whether the survey design provides accurate and precise estimates of abundance or relative changes in abundance. Precision is a function of survey design and sampling intensity and reflects the uncertainty associated with the survey indices. Precision can be evaluated for the NMFS surveys by examining the current allocations of survey sampling stations among statistically designed strata to determine whether they correspond to the current spatial distribution of the target species and to judge if the allocation of samples to strata increases the precision of the overall abundance estimates. Methods for evaluating the precision of stratified random surveys are available in the statistics literature and have been used to evaluate similar fishery surveys (Smith and Gavaris,
OCR for page 137
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA 1993; Smith, 1996; see Appendix C for summer flounder example). Although this kind of evaluation may be complicated by the multispecies aspect of surveys such as those conducted for New England groundfish, it is useful to evaluate the survey design for each species to determine whether the existing design is a reasonable compromise for the mix of species. Reallocation of survey effort among sampling strata is a useful method for refining surveys as priorities change. NMFS has employed this approach well in the past to improve the level of information gained from its surveys. However, many of the strata in the 1995 fall New England groundfish survey and the deep strata for the 1995 spring groundfish survey contain only one sample per stratum. When strata contain only one sample, variance is inestimable, except by extrapolation from other areas. Recommendations: The statistical precision of each NMFS survey should be evaluated. Estimates of the maximum expected precision given the current number of tows should be calculated and used to evaluate the statistical power of the survey to detect changes in abundance over time. The gains, or potential gains, from better survey designs, or an increase in sampling intensity, should be evaluated in relation to the assessments in which the data are used. An increase in the precision of state surveys might also be achieved if similar analyses were applied to them. NMFS should periodically review whether sampling effort should be reallocated, even when the overall objectives of the survey stay the same. For example, if the range of the species expands to offshore areas, NMFS may need to sample these areas more heavily. In the case of the summer flounder, this type of change was made, but user groups, particularly fishermen, did not seem to know the extent of the survey expansion and the results of the changes, both in terms of raw data and assessment results. In relative abundance plots, the symbols denoting abundance at each survey station should identify the location of survey stations where there was zero catch. Small changes like this could greatly enhance communication. Appropriate statistical design should always include more than one tow per stratum. If this is not a feature of an existing sampling design or extra stations cannot be occupied during the execution of a survey, strata should be redefined before the survey or combined for analysis if the number of tows becomes limited during the survey. 2 In general, surveys should be cost effective, statistically well designed, conducted frequently enough to detect significant trends in abundance, and take into consideration the biology of the species. Standardization of survey methods within management units to meet these needs should increase economic efficiency, quality control, and comparability of data. Standardizing within management units may be necessary, but flexibility across species and management regions is needed to address the diversity of issues and environments observed in marine systems. State surveys apply a great deal of sampling effort, which is often allocated in relatively few tows per stratum each month, several different months each year. For many states the monthly surveys seem individually inadequate to develop precise estimates of abundance. Combining these surveys across months may also be inappropriate if a species is seasonally migratory and will change distribution among strata over the several months of survey effort. States should determine whether consolidating all their survey effort into a single annual survey might yield much more information than monthly surveys. Such analysis and coordination could be undertaken under the auspices of the interstate marine fisheries commissions. Accuracy of Survey Data: Frequency and Spatial Extent of Surveys Findings: The frequency of NMFS surveys varies from stock to stock and region to region. For 2 Strata may differ by species and year to year, so post hoc combination of strata is more appropriate than combining strata in the sampling phase.
OCR for page 138
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA example, there are two general trawl surveys per year, plus the winter flatfish survey, on the U.S. East Coast. Other surveys (e.g., West Coast Continental Shelf groundfish) are conducted once every three years (see Table 3-3). Many federal surveys (the Atlantic Coast is an exception) do not provide information about seasonal distribution because they are conducted only annually or less frequently. The seasonal distribution of a species may be important to monitor in terms of how it influences catch rate for both survey and fishing vessels and in terms of how vulnerable different components of the stock are at different times of the year. Infrequency of surveys may hinder management, especially for fully exploited fisheries and for fish stocks whose dynamics change significantly from year to year. NRC (1998a) found that the single most important factor in achieving accurate results using any stock assessment model is an accurate indicator of relative abundance over time. Where vessel time is limited, surveys with twice the effort every second year—or thrice the effort every third year—may lead to more reliable assessments than annual surveys. Sometimes annual surveys may be the best option, but not universally. It is possible that the biology of a species complex, the health of its stocks, or the management approaches being used would not require an annual survey, particularly if good fishery-dependent data were available and the fishery were not in a situation in which recruits make up a significant portion of the biomass. Recommendations: The NMFS examination of the costs and benefits of data collection should include the frequency and timing of surveys in each region, with consideration of factors such as the biology of the managed species, state of the stocks, the current and potential economic value of the species, and the availability of other accurate indices of trend (e.g., commercial CPUE). NMFS should report their findings to Congress to help members of Congress understand subsequent realignments of survey activity that may be worked out between NMFS and the regional councils. In addition, NMFS should attempt to improve the quality of commercial and recreational data to the point that they could substitute for some survey data. Findings: The accuracy of survey data is mainly a function of sampling over a stock's entire geographic range and adequate sampling of all age and size classes. Distributions of fish over time and space and at different life stages are important considerations in designing the spatial extent of surveys. As noted by NRC (1998a), surveys should consider that stocks may shift over time due to regime shifts and other environmental changes. For some stocks, such as summer flounder, it is not clear in available documentation that the surveys cover the entire range of the stock or adequately account for seasonal movement. For example, catch numbers of summer flounder in the winter surveys increased all the way to the edge of the survey (at the continental shelf edge) in some areas. NMFS added stations along the shelf edge for the winter survey (as weather permits), in response to industry concerns. The assumption in the subsequent analysis was that there were no significant concentrations of fish beyond the survey zone. Another example of a survey that may not sample a stock's entire range is for Greenland halibut (turbot) in the Bering Sea. Turbot is primarily a deep-slope species, yet no information is available regarding its abundance in its prime habitat because the triennial slope survey does not sample deep enough. The range of a stock can be monitored through the spatial distribution of abundance indices in the surveys and the locations of commercial or recreational catches. Using fishery activity to detect changes in a species range may not be effective, however, if management is changed in such a way that fishing time or place are restricted (e.g., trip limits for summer flounder reduce fishing activities far from port).
OCR for page 139
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA Recommendations: NMFS should ensure that the geographic ranges of its surveys cover the geographic ranges of each species managed. In particular, the spatial extent of surveys should be based on good evidence that the ranges of the target species included in the survey have been reached, using objective threshold criteria for deciding when the range has been covered sufficiently. Information about the geographic extent of surveys and commercial catches should be shared with managers and stakeholders in an appropriate forum. The historic locations of summer flounder, as determined by catches from fishing activities or exploratory surveys in winter, need to be analyzed to evaluate the extent of the stock area with respect to existing strata boundaries. Although this information may already be available in various forms, it needs to be drawn together and presented in a form accessible to the public, to either correct misperceptions of the industry or initiate action to improve the sampling if industry concerns are warranted. NMFS should use information provided by commercial (e.g., logbooks) and recreational fishermen regarding geographic locations of stocks (e.g., summary plots of commercial or recreational harvest locations), keeping in mind that if regulations limit the places in which it is legal or cost effective for fishermen to fish, the spatial distribution of catches may not indicate where fish are actually to be found. NMFS can consult with fishermen by presenting planned survey locations to them. All user groups should watch for patterns in fisheries data that may appear to be associated with poor spatial coverage (i.e., seeing few large fish in the catch) and could indicate other issues of concern, such as underreporting of discards or unaccounted-for migrations. The existence of separate stocks should be examined and, if found, their significance should be determined. Data to differentiate among stocks should include information collected from tagging studies, characterizations of parasitic fauna, genetic studies, recruitment patterns, and growth rates (NRC, 1994a). Stock identifications should recognize that unit stock concepts are defined not only by genetic uniqueness, but also by the degree to which stocks differ in local recruitment characteristics, feeding and spawning ground fidelity, growth patterns, and other lifehistory characteristics. Survey data, as well as fishery-dependent data, may need to be collected at a finer scale than presently to manage such stocks separately, but the added costs of more detailed data should be evaluated against the additional benefits that would result. Essential Fish Habitat Data Findings: A dramatic increase in information on essential fish habitats is likely to result from the emphasis on this matter in the Magnuson-Stevens Act. Although this information is being gathered principally to identify habitats critical to various life history stages of important marine species, the result will be a compendium of information that can be used more broadly to serve the needs of fisheries assessment and management. For example, knowledge of the distribution of favorable habitats for adult fish can be used to design more appropriate sampling strata, leading to increases in precision in population estimates. As another example, key habitats for eggs, larvae, and juveniles can be used to identify areas for more effectively monitoring year-class strength, leading to better short- and long-term projections of population trends. Recommendations: Information on essential fish habitat should be collected and managed in the context of the broader needs of fisheries science, assessment, and management. Stock assessment scientists, fishery managers, and fishery ecologists should work together in setting objectives for collecting data related to essential fish habitat, and in establishing a means of managing and accessing these data. This latter goal should be developed in parallel with ongoing efforts that focus on more traditional data management. An important aspect of managing data related to essential fish habitat is to ensure that the format
OCR for page 140
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA and content of the data conforms to other federal oceanographic data in a spatial and temporal framework. Sampling Gear and Methods Findings: The use of gear standardized over time is important for maintaining consistent measures of relative abundance. Surveys are generally designed to yield estimates of abundance that can be related to estimates based on catches with the same gear in previous years. Commercial fishermen are constantly updating their gear and methods and try to fish in areas where they can maximize their profits. A common complaint by fishermen about surveys is that NMFS uses outdated gear, inefficient fishing methods, and random or set stations, which fishermen believe biases the species and size compositions of survey catches in ways that reduce the TAC. In some cases, commercial fishermen have been able to demonstrate that survey gear has not been used correctly, preventing the determination of even relative abundance, for example, in a West Coast fishery (Lauth et al., 1998). Fishery-independent surveys for summer flounder and other groundfish in the Northeast region use 30-minute tows to increase the tows per station and increase precision, whereas commercial fishermen often tow their trawl nets for 3 hours or longer. Fishermen report that longer tows allow them to catch large fish that are hard to tire with shorter survey tows. Fishermen who attended the committee's meetings were unsure about what gear was used in Northeast Fishery Science Center's summer flounder surveys, but doubted whether the gear would sample summer flounder in an unbiased manner, given the tow duration. Most U.S. surveys are conducted using bottom trawls, so mid-water and surface-dwelling fish species are undersampled. Hydroacoustic and other remote sensing methods are not commonly used for U.S. fishery assessments, with a prominent exception being the hydroacoustic surveys used to assess spawning aggregations of Alaskan pollock and Pacific whiting (e.g., Traynor, 1997). Acoustic methods do not work for flatfish because these methods have difficulty distinguishing fish from seafloor, but these methods can greatly extend survey capabilities for midwater fish. Recommendations: When survey gear is outdated, has unstable performance or is hard to set up correctly, effort should be directed at improving the gear and providing some level of cross-calibration so the value of historic data is maintained. Key issues in gear selection include credibility of results as well as bias in sampling and sample-to-sample variability. If use of modern gear and methods could decrease bias and variability without seriously compromising survey data time series, NMFS should consider updating its gear and changing survey methods to make them more similar to current practice in the fishing industry. If changes are made, however, they should be done with appropriate parallel use of old and new gear for as long as necessary to provide adjustment factors for historic data. NMFS should consider hiring commercial fishermen to participate in survey cruises to see how sampling gear is used and where surveys are conducted. Conversely, fishermen should acknowledge that even if survey gear is somewhat antiquated, as long as it operates as it was intended, it may still provide an adequate index of relative abundance. NMFS should document and communicate information to stakeholders about the gear and methods (such as tow duration) used in its surveys, as well as an evaluation of the gear's efficiency at capturing different species. This is particularly important for summer flounder because the committee is unsure if roller gear, whether 6 inches or 36 inches, is appropriate for surveying this species. NMFS should endeavor to communicate to stakeholders why survey gear may be operating differently than commercial gear and how such differences may or may not affect assessments. In the case of summer flounder, the effect of tow duration on the size distribution of
OCR for page 141
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA fish caught should be studied to understand differences in gear selectivity that may exist between surveys and commercial catch. Gear and operating procedures used in all surveys should be evaluated on a regular basis (e.g., every 5-10 years and particularly at the time research vessels are changed) to ensure that they are still reasonable by industry standards. Modernizing gear and procedures is one way to increase the credibility of survey results to fishermen. Another means of boosting NMFS' credibility would be to run parallel surveys with commercial fishermen. This will help determine how catchabilities compare between the survey and the fishery and help improve communication between NMFS and commercial fishermen. It also can help in the interpretation of fishery-dependent data, for example, how commercial CPUE is changing over time relative to survey CPUE. Incorporating fishermen's knowledge into sampling design and analysis, especially about where and how long to fish, could improve sampling efficiency and NMFS' credibility with fishermen. The interstate commissions should initiate or continue efforts to get states to standardize their surveys to improve the comparability of their survey data, decrease the frequency of their surveys to increase the stations on any specific survey, or in some cases, shift from doing surveys to other activities more useful for coastwide stock assessments. New survey methods should be pursued. NMFS should increase its use of hydroacoustic and other developing methods to estimate the stocks of surface-dwelling, mid-water, and other vertically-oriented fish species (e.g., rockfish) that are not susceptible to bottom trawls usually used in NMFS surveys, where this would lead to a cost-effective gain in precision. Although acoustic methods don't work for flatfish, they may be useful for other species. Survey Vessels Findings: Fishery-independent surveys using calibrated vessels owned by NOAA provide vital and irreplaceable data for stock assessments and ecological monitoring and other fishery assessment purposes. Trawl surveys are especially susceptible to variability caused by differences among vessels, so calibrated vessels are a necessity. Many kinds of data can be collected only by using survey vessels, although NMFS charters 40 percent of its survey and research days at sea on commercial and university vessels annually for other purposes. NOAA survey vessels are aging and in need of replacement and it appears that the number of ships and ship time available for surveys is not adequate to meet all critical needs. Recommendations: The committee endorses the efforts of Congress and NMFS to maintain a strong fleet of NOAA survey vessels, particularly for trawl and acoustic surveys, by replacing aging vessels with new, more capable, and more quiet ones. NMFS should continue to use charter and lease-back arrangements, where appropriate, even as the agency acquires new survey vessels. Congress should not only fund the construction of new vessels, but also should provide adequate funding for survey and research work performed by these vessels. Adaptive Sampling Findings: Adaptive sampling provides a means of increasing the precision of certain kinds of survey estimates, especially total biomass of some species, by using information obtained during the survey to determine where additional sampling should be done. In a sense, fishermen act as adaptive samplers when they locate and fish intensively in areas where the fish are present, but they do not sample in a way that is statistically valid. Although adaptive sampling is generally a single-species approach, sampling rules may be defined so that precision would be increased for groups of species that have similar distributions. Knowledge of habitat preferences can be used to design adaptive sampling schemes. Adaptive sampling is not appropriate for other purposes, such as mapping species distributions,
OCR for page 155
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA An external report assessing NOAA fisheries research vessels also recommended that NMFS, nationwide, put more emphasis on developing “best practices” for interactions with fishermen, indicating that “NMFS as a whole needs to devote very serious effort and experimentation to the sociological aspects of its operations. They could use professional assistance in this area, as part of the legally-mandated increased emphasis on socioeconomic considerations in ecosystem-based management. NMFS must partner with a much broader segment of the interested community— fishermen, academics, industrialists—to a much greater degree than it ever has, in all aspects of its fisheries oceanography and fisheries monitoring and survey efforts” (Dorman, 1998). Recommendations: NMFS should identify approaches that maintain the statistical rigor needed for long-term fishery assessments, while making the best use of local knowledge among commercial fishermen with expertise about specific stocks and gear types that are efficient at catching targeted stocks. Canada, New Zealand, and other nations provide models for positive interactions with industry. NMFS should consider hiring commercial fisherman to participate in surveys (in addition to opportunities for unpaid participation) to see how sampling gear is used and where the surveys are conducted. NMFS also should carry out some joint sampling cruises using NMFS and commercial vessels, with exchanges of crew. Harms and Sylvia (1999) suggested that collaborative research between fishermen and scientists should be “undertaken only if there is (1) equal partnership in planning and implementation, (2) adequate funding, (3) competent management, and (4) commitment to begin small and build on success.” If cooperative research is to be adopted on a broader basis, institutional changes in both industry and government agencies will be needed, and the recommendations of Harms and Sylvia should be considered. Cooperation with Recreational Fishermen Findings: The lack of a national program for saltwater fishing licenses greatly complicates estimation of recreational catch and effort. Such a requirement is controversial because many states do not presently require licenses, and anglers in those states do not want to face additional regulations and a perceived intrusion of government into their private lives. However, requiring licenses for marine recreational fishing (even free ones) could improve data collection efforts by providing a comprehensive sampling frame and eliminating the inefficient random-digit dialing surveys (but not the expensive intercept surveys). License frames are of greatest value when they obtain uniform information and are coordinated among the states. In theory, recreational effort assessments could be less costly (in terms of time, money, and staff) if based on license sample frames, because anglers would be identifiable and sampled more easily than with current methods. However, some states ' requirements for saltwater licenses exempt certain classes of anglers (see Table 3-9), which would complicate attempts to use saltwater recreational fishing licenses. Recommendations: NFMS should increase its dialogue with recreational fishermen to jointly develop and implement improved data collection for recreational fisheries, through MRFSS, specially designed and coordinated tagging studies, and alternatives to MRFSS that allow in-season adjustments to recreational quotas. MRFSS should continue to evaluate whether saltwater fishing licenses and longitudinal sampling would provide cost-effective alternatives to random-digit dialing, keeping in mind that merely using license frames may not make MRFSS more timely or cost-efficient if different states use vastly different licensing programs. Review Findings: The review of data collection procedures is usually handled tangentially in the course
OCR for page 156
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA of reviewing stock assessments, but few reviews focus on data collection. Recommendations: NMFS, in conjunction with the regional councils, should review all aspects of its data collection activities, on a fixed, publicly-announced schedule including all types of fishery-dependent and fishery-independent data. Such reviews should include both a scientific peer review and a stakeholder review. As part of this process, commercial fishermen and other stakeholders should participate in actual data collection exercises. IMPROVING DATA MANAGEMENT Defining User Groups and User Needs Findings: Many types of stakeholders could be users of new fishery databases. Different user groups (federal and state agencies, fishermen, scientists, managers, environmental advocacy groups, consumers, local communities) have different needs and concerns about fish stocks and marine ecosystems. The primary focus of most existing data collection and management activities is on collecting sufficient data to conduct an accurate and precise scientific assessment of biomass relative to previous years. However, other potential users have different questions. For example, fishermen may want to know how the catch of different species correlates with environmental variables and how their business compares against port and fleet average performance. Community leaders may want to know how fishing is expected to fare in future years, whether fishing activity is shifting to or from their port, and the value of the local multiplier effect for different kinds of fishing-related businesses. Environmental advocates might want to know how fishing is affecting marine ecosystems and trends in bycatch. NMFS seems to be giving a low priority to putting fisheries data into the same geospatial format used for data from other agencies and working with other agencies to conform to the same data standards. Recommendations: Research on stakeholder and user concerns and needs should be conducted and used to improve the outcome of any new fisheries data management system. NMFS and others responsible for data collection should recognize that the data they collect may be used in a broader context. Stakeholders should be enlisted to identify broader uses for both traditional and new data. NMFS, interstate commissions, and states should involve expected and potential data users and providers in designing their fishery data management systems. Mechanisms for involvement should identify and address the concerns and needs of data users and providers in relation to fishery databases. The kind of involvement used in design of the Atlantic Cooperative Coastal Statistics Program (ACCSP) may serve as a useful model. Databases and Data Management Systems Findings: ACCSP has been a good model of a regional data management system up to this point in its development. For example, it has identified a core data set and requires inclusion of all necessary data in a single database available to all partners. Less desirable are systems that allow one partner—either states, commissions, or the federal government—to have absolute control of data. Recommendations: Regional databases should standardize their data collection, management, and quality control activities. The committee agrees with the directive of Congress in requesting a plan for a nationwide Fisheries Information System (FIS). The FIS design (based on coordinated regional systems) is good and its reliance on national standards is a positive feature. The FIS is ambitious, however, and for it to be successful, (1) Congress must provide adequate funding and (2) cooperation and balance among regions must be ensured. ACCSP, other regional databases, and the FIS should specify the national and international standards to which the system will comply, NMFS should:
OCR for page 157
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA continue its plans to implement national data standards that are designed to promote data exchange, include metadata with all data sets to ensure that data can be understood properly, promote the wide distribution of shareable data and metadata, create the ability to cross-validate logbook, observer, dealer, and VMS records to assess and improve the quality of fishery-dependent data, and work with the U.S. Fish and Wildlife Service (FWS) to prevent confusion between the NMFS Fisheries Information System and the FWS Fishery Information System, a data system for freshwater migratory fish. Findings: Estimating the costs and benefits of existing and planned fisheries data management systems is an important task. Some analyses of costs and benefits of regulatory impact are required as part of fishery management plans (FMPs; MSFCMA Sec. 303[a]), although FMPs do not require analysis of data management as a separate phase. Analysis of the actual costs and benefits of an operational FMP often include factors that are difficult to quantify in economic terms, such as better information availability and streamlining an activity. The committee agrees with NMFS that developing an FIS is critical for fulfilling its responsibilities pursuant to the Government Performance and Results Act (GPRA) of 1993 in terms of collecting relevant data, tracking progress in achieving performance goals, and ensuring integrity and accountability. Recommendations: NMFS should continue to attempt to find ways to contain costs and increase benefits from its fisheries data management activities. In part, this can be accomplished by continued cooperation with states and regions in data management and looking for opportunities to build on existing efforts (e.g., through FIS), rather than duplicate them. Another means to achieve this goal would be increased use of commercial data and better recreational data that could be available by following recommendations given earlier in this chapter. Institutional Arrangements Findings: Several different paths have been taken in different regions of the United States to manage fisheries data. ACCSP and FIS are providing (or plan to provide) centralized regional and national databases, respectively. In contrast, several commercial firms are providing value-added products to fishermen for operational purposes, including data for smaller areas and for specific fleets or fleet segments. Government data systems provide the potential benefits of access to all users, long-term storage, and nationwide standards for data collection, but are presently limited by the confidentiality of data and the often significant time lag between data collection and availability. Commercial systems can quickly provide very specific products and can establish whatever level of confidentiality is specified by a consortium of data contributors and financial supporters. Recommendations: NMFS should encourage both centralized governmental and decentralized commercial data management, depending on the characteristics of the product needed and the capabilities of different sources to achieve the identified product characteristics. Centralized governmental data management can be achieved by continuing the development of the planned umbrella Fisheries Information System. At the same time, NMFS should identify which sources of data and information might best be managed by commercial firms and find ways to encourage such commercial development through Small Business Innovation Research grants, Cooperative Research And Development Agreements, and other programs. Findings: Institutional arrangements for management of data are evolving toward systems of regional and national coordination and access. The situation in the Gulf of Mexico region, where funding was appropriated directly to the Gulf States Marine Fisheries Commission, with no assurances of partnerships in system development and data access, could hinder the implementation of the FIS and degrade data quality.
OCR for page 158
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA Recommendations: The Gulf of Mexico, Pacific, and other regions should follow the ACCSP lead in identifying needs and recognizing where efficiencies could be gained and quality could be improved by standardizing protocols and identifying regional concerns. Congress should ensure that systems are compatible and that the proper level of centralization or decentralization is achieved. The role of commissions in data management needs special attention. The overriding issue is whether a centralized, decentralized, or mixed approach to data management is most efficient and meets the needs of different levels of government. In relation to the commissions, NMFS should determine: What are the benefits and costs to federal management of having the commissions involved in collecting and managing data from the states? Would it be more cost-effective for the commissions, NMFS, or private contractors to manage state data? Does the imposition of an additional administrative layer between the states and NMFS hinder data flow and imposition of standards for data collection and management? Implementing Standards and Improving Quality Control Findings: Standards of potential importance to fisheries data management range from commonly accepted data and information standards such as SQL, TCP/IP, Z39.50 searchable indices, CORBA, and Government FIPS standards for federal systems, to Internet standards (TCP/IP, HTTP). Metadata standards of potential importance include the Federal Geographic Data Committee (FGDC) Content Standard4 for Digital Geospatial Metadata and the Biological Data Profile of the Content Standard of Digital Geospatial Metadata. The U.S. Fish and Wildlife Service (FWS) has established and promoted data standards to ensure compatibility of data management and use throughout the agency. Recommendations: NMFS should follow standards set by the Federal Geographic Data Committee, so that its data are compatible with data from other agencies. This is of more than bureaucratic interest because several agencies collect data that could be useful to fisheries management if NMFS data and data from other agencies were in compatible formats. NMFS should follow the FWS example in the attention given to stewardship, promotion of, and accessibility to data standards through Web-based information sources. Any standards specific to fisheries should be set cooperatively among stakeholders, as practiced in other industries such as electronics and computers (e.g., by the American National Standards Institute). Computer hardware to be used onboard fishing vessels should comply with the NMEA-0183 standard for marine interfaces. NMFS should attempt to implement nationwide standard error-checking procedures. Improving Technologies Findings: Data entry without verification and manual transfer of logbook data to electronic systems are opportunities for errors to enter databases. Electronic logbooks are a promising new development being pursued in a NMFS project, jointly with the industry, in the Pacific Northwest region. VMSs are either being used or will be used soon in many U.S. fisheries. Recommendations: NMFS and the regional councils should require double-entry and other verification techniques and continue to pursue the possibilities for electronic submission of logbook information. NMFS should implement VMSs in such a way that their value is much greater than the added financial and bureaucratic costs to fishermen. 4 Data content standards provide semantic definitions of a set of objects and the relationships among them (see http://www.fgdc.gov).
OCR for page 159
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA Review Findings: Many of the regional data management systems seem to have arisen and accreted from within states or regions without much external review or even periodic internal review. Recommendations: NMFS and other agencies with data management responsibilities should have their systems reviewed by outside experts, considering both how well they meet their intended purpose and how well they adhere to relevant national and international standards. IMPROVING DATA USE Data in Stock Assessments Findings: Fishermen often believe that data are mishandled or misrepresented to support hidden agendas. This perception may arise because fishermen do not understand or agree with the significance or scientific merit of a particular data collection method. However, some suspicion may be understandable because it is not always clear why some data are used and other data are ignored. Changes in assessment methods appear to occur frequently and such changes often come as a surprise to fishermen. In the case of summer flounder, commercial fishermen believe that NMFS surveys do not cover the full range of the species, information from commercial logbooks is ignored, sportfishing information is lacking in timeliness and quality, and changes are made in management actions before the consequences of previous actions are fully realized. The Food and Agriculture Organization's report Precautionary Approach to Fisheries recommended that stock assessment processes should include “a process for assessment analysis that is transparent....” (FAO, 1995, p. 14), to improve the understanding and trust of stakeholders in assessments. Improving Fish Stock Assessments (NRC, 1998a) gave extensive advice for improving data use in terms of stock assessments: A variety of assessment models should be applied to the same data. Greater attention should be devoted to including independent estimates of natural mortality in assessment models. Fish stock assessments should include realistic measures of the uncertainty in the output variables whenever possible. New stock assessment techniques should be developed that can yield accurate and precise estimates, even though some of the data are incomplete, ambiguous, and variable. New techniques should also take into account the effects of environmental fluctuations on fish populations. Recommendations: NMFS should make its stock assessment process more transparent and accessible to stakeholders. Prior to a major modification in any aspect of a stock assessment procedure, NMFS scientists should discuss proposed changes, and the reasons for them, with key user groups (i.e., fishermen, managers, and environmental advocacy groups). In general, NMFS should make the objectives of each survey clear, based on existing and anticipated management programs. NMFS should present assessment results with and without the proposed changes in stock assessment procedures, to show how the results differ and why the modifications make the assessment more realistic. When a survey design is modified, stakeholders should be informed of the changes and what they are intended to accomplish. Feedback from stakeholders should be acknowledged and addressed either by incorporating suggestions or by providing a reasonable rationale as to why the existing approach is most appropriate. Objectives of stock assessments should be communicated to stakeholders in a way that can be understood. NMFS should review the objectives of surveys periodically and publicly to ensure that surveys are designed correctly to meet these objectives. The committee supports the recommendations of NRC (1998a) listed above. Institution of these measures would help fishery scientists and
OCR for page 160
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA managers understand fisheries better and understand the level of uncertainty more fully. Access to Fisheries Data Findings: It is difficult for individuals outside NMFS to access many types of fisheries data held by the agency, although NMFS does provide some basic queries for both commercial and recreational data on its Web site. Even data that do not need to be confidential can be difficult to find and use because they are kept on many different servers, do not feature user-friendly interfaces, and/or require special permission to access. Furthermore, data that are available are not always in a user-friendly or even a computer-friendly format. The Open GIS Consortium is well advanced in developing Web-based access to geospatial data and provides a ready standard for government data systems. Recommendations: NMFS should develop and publicize (including on their Web site) a data access policy and instructions for accessing data, which may need to be different for different users because of confidentiality concerns. Standardized formats or access through standardized query programs should be enhanced so that any qualified user can access and use the data readily to answer questions that arise across data sets, and over a variety of scales within data sets. NMFS should participate actively in the interagency Open GIS Consortium and adhere to consortium guidelines in creating its systems for data management and access, so that its data are compatible with data from other federal agencies and accessible through the same means. Fishery scientists in NMFS and academia should continue to have access to data from logbooks and observers without aggregation, if they agree to respect confidentiality of the data when reporting the results of their research. Access to data and metadata both within and outside NMFS should be encouraged because involving new people in data analysis can provide new perspectives about the data and a greater number and variety of users is more likely to reveal problems and errors in data sets, improving quality control. Confidentiality Findings: Confidentiality of fisheries data is restrictive to the point of hindering both research and management. State and federal restrictions to free access to data can hinder development of the kind of socioeconomic models that enable scientists and managers to determine whether regulations and management measures have been effective or to predict whether potential measures are likely to be effective. These policies neglect the rights of the public to have greater information about the use of public-trust resources. The privilege to exploit marine fish resources should carry some obligation on the part of fishermen, balancing reasonable protection of proprietary information against the large need of managers to be well informed and able to manage the fishery. For example, fishing in some areas may be detrimental to the environment. If information on fishing areas is confidential, interested stakeholders would have a difficult time determining how much fishing is being conducted in sensitive areas. Conversely, some level of confidentiality may be necessary to allow fishermen to maintain their businesses and to promote reporting of high-quality information about location, landings, and bycatch in some fisheries (e.g., reporting of halibut fishing information to the International Pacific Halibut Commission), information that might not be as accurate if it were not confidential. NMFS and its partners in development of new fishery data management systems (e.g., the ACCSP) have taken the approach of creating cooperative systems managed independently of NMFS and with access from all contributing partners. Because of such independence and broad participation, such systems may be more credible with all stakeholders than the current generation of systems that are less accessible to all stakeholders. Coastal states (except Alaska) share the same data confidentiality standards, with the rule of thumb being that there must be three or more
OCR for page 161
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA reporting entities (e.g., vessels or seafood dealers) at any level of data summary before the data are not confidential and can be available to the public. Alaska requires a minimum of four or more reporting entities. On the U.S. West Coast, states have authority to collect logbook information and they set policies on confidentiality. States use a common (NMFS-designed) logbook and store the data with NMFS, but NMFS must respect state data policies. For some states, NMFS is not allowed full access to all state data; for example, Alaska restricts access to confidential data, except to selected NMFS personnel for stock assessment and law enforcement purposes. Two recommendations of NMFS (undated) that are particularly relevant to this report are that the next reauthorization of the Magnuson-Stevens Act should (1) establish a sunset period on confidentiality of fisheries data and (2) eliminate existing prohibitions on collecting economic and financial statistics from marine fisheries. Recommendations: Congress and the states should re-evaluate their existing policies on data confidentiality, while respecting the rights of fishermen as small business owners as much as possible. The assertion that fisheries information has intrinsic value proprietary to the fisherman should be examined in light of the mandate to NMFS and the regional councils to manage fisheries as public-trust resources. Specific consideration should be given to establishing sunset periods for data confidentiality for all fisheries data, when the information no longer retains significant proprietary value. The proprietary periods should be determined for each fishery in a public forum including scientists, managers, fishermen, and environmental advocates, and should be included in the fishery management plan for each fishery, as each fishery is unique. Sunset provisions for data confidentiality should be developed by government (states and federal) working with data providers. The effects of the loss of confidentiality on precision and bias in logbook and other reporting should be considered in setting the proprietary period for each type of data. Matching Management to Data Available Findings: The effectiveness of fisheries management depends on the use of timely data of suitable accuracy and precision to provide answers to questions about a stock's current status, desired future status, and actions needed to achieve the desired status. Greater precision and accuracy often require greater resources for sampling and analysis. Moreover, some sources of inaccuracy and imprecision may be difficult to eradicate even with unlimited expenditures. Thus, the question for management should become: What level of inaccuracy and imprecision in advice and hence in management is tolerable if a particular management regime is to achieve its goals? The answer to this question clearly depends on the management regime chosen and its objectives. If managers are not prepared to pay for the needed precision or if that precision may not be achievable at any price, they may have to modify either their management objectives or management tools. As demonstrated in Chapter 3, management and data needs are closely related. Recommendations: The regional fishery management councils and NMFS should work together to match management to data that are available at a reasonable cost. Such an analysis will depend on completion of the review of the costs and benefits of fisheries and data collection recommended earlier in this chapter. Cooperation and Communication Findings: Part of the image problem shared by NMFS and the regional councils is lack of communication on a level that is informative and appealing to stakeholders. Many stakeholder groups are affected by the collection and use of fisheries data, including commercial and recreational fishermen, and environmental advocacy groups. Few stakeholder groups have a good understanding of why fisheries data are collected and used in certain ways. Greater outreach to
OCR for page 162
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA these audiences to improve their understanding and the perceived credibility of fisheries data is needed. User groups need to know about the information contained in a data set, even though the analysis for which it is used may be complex. It is important for scientists and managers to improve their communication of the data available and to make such data available to stakeholders more readily and in a user-friendly form. When this is not achieved, a lack of trust develops between those who control access to data and those who cannot gain access. In many cases, disagreement of fishermen with the results of stock assessments can be traced to NMFS not explaining the sources of variability in the data and the uncertainty of the models being used. In the current fisheries management system, several activities occur sequentially: Data are collected. Stock assessments are conducted. Management recommendations are made. Fish are allocated among user groups. Fishing regulations are designed and implemented. For individuals, this process determines either their opportunity to make a living (commercial and charter sectors) or their ability to engage in their recreational activities. When their income or their favorite pastime is threatened, people respond by attempting to manipulate the management process at every stage. No amount of communication or transparency of the process will eliminate these conflicts among users. But a more open and innovative assessment process could improve the credibility of the resulting assessments and perhaps move harvester conflicts away from the stock assessments to other stages. Recommendations: NMFS should improve its outreach to specific stakeholder audiences by seeking (1) perspectives from stakeholders regarding what information they would find useful and how best to get it to them and (2) perspectives from data gatherers and stock assessment scientists regarding what is important for user groups to know about the quality and limitations of data. NMFS and councils should ensure that stakeholders feel they are getting all the information they need to make decisions and understand how councils use data to make management decisions. Such outreach could take place in conjunction with meetings of the regional councils and/or at special meetings convened by NMFS in port areas. Outreach events would be one way to communicate information (and uncertainty) to target audiences in person. Another complementary approach would be to make data more accessible through Web-page queries and more sophisticated forms of graphical data presentation. Several efforts by NMFS and other parts of NOAA—providing the ability to query aggregated commercial and recreational data through the NMFS Web site and data visualizations provided through the Pacific Marine Environmental Laboratory Web site—are a first step in the next generation of data access. NMFS could use its new communication experts for this purpose, remembering that communication is a two-way process, not merely a one-way dissemination of information. Increased cooperation and communication between NMFS and industry is crucial for improving the collection and understanding of fisheries data. However, it is important that NMFS conduct all activities with industry in an open manner to avoid the appearance of collusion. NMFS should improve its use of data visualization to communicate scientific information to the public, moving away from display in tabular form to more graphic displays, including plots, maps, and pictures, without lowering the quality of its data or presentations. The accuracy of the information available needs to be conveyed (e.g., maps are not always equally accurate over the entire range displayed). The methods and assumptions used for data summarization or analysis should be included with the data and analyses. In communicating data and results of analyses, NMFS should tailor its approaches to different audiences that may require different levels of detail. It would be useful for the public to see
OCR for page 163
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA what the different data sources (commercial, recreational, and survey) indicate about a fish stock, including overlays on maps to illustrate geographic coverage. Such communication should avoid statistical and other jargon so that non-specialists, fishermen, and the public can fully understand the significance of the information. NMFS should consider using state Sea Grant Marine Advisory Service units more often to help with outreach. One innovative and useful approach is to conduct fishery assessment and management simulations with real fisheries data in a workshop setting to explore with commercial fishermen, recreational fishermen, environmental advocates, and others how assessments are developed and how management decisions are based on assessments. This approach has been used by several fishery scientists to provide opportunities to focus attention of stakeholders on the models and data, rather than on each other (Holling, 1978; Walters, 1986, 1994). In using this approach, it is important that the objectives of the workshops be very specific and that they be conducted outside NMFS, to provide an objective mediator. Simulation workshops should explore issues such as the following: What are the management implications of using random stratified methods for surveys versus going where fishermen know there are concentrations of fish? Why are survey gear and methods kept constant over time? Why is it difficult to use short-term charter vessels for trawl surveys? Does it make a difference in setting TACs whether a decline in recruitment, biomass, or catches are caused by fishing versus climate, habitat loss, pollution, or other environmental factors? How do unreported landings and bycatch affect assessments and subsequent management? How would observations of fishermen change the analyses if included? What kind of observations could and could not be included? In each case, the goal should not be to justify NMFS procedures, but to expose the assumptions and procedures of modeling and TAC setting to the stakeholder communities and explore together what might be the consequences of changing assumptions. Such an activity might bring new insights to NMFS scientists, council members, commercial and recreational fishermen, environmental advocates, and others, and could indicate to NMFS changes needed in its assumptions or procedures. Preparation for simulation workshops should include thoughtful analysis and development of techniques and software. NMFS should seek assistance for such an effort from other parts of NOAA (e.g., the Joint Research Institutes and National Sea Grant College Program) and other sources. NMFS should also create, and make available to the public, manuals describing survey operations that specify survey design, tow time, gear used, and other important factors, such as is done for West Coast fisheries (Anonymous, undated; Munro and Hoff, 1995; AFSC, 1998; Wilkins et al., 1998). NMFS should identify which data sources (including which individual surveys) are included in each stock assessment and what weight is given to each. If the weighting of data sources changes over time, NMFS should communicate what criteria are used to change weightings. Uncertainty in Data, Models, and Model Outputs Findings: Despite the significant effort that goes into designing and implementing fishery surveys to sample individual fish to characterize populations, survey data still usually enter assessments as point values with no corresponding estimate of uncertainty, such as variance. The result is an assessment that gives each observation equal value, whereas in reality some reflect more precise and/or more accurate information than others. This approach decreases incentives for improving data quality. One reason for not including uncertainty in models may be that it is
OCR for page 164
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA difficult to distinguish among the different types of uncertainty entering an assessment model. Uncertainty can enter assessments through observation error, natural variation in the environment (process uncertainty), model misspecification, and uncertainty that arises in management implementation and institutional interactions. Another reason that uncertainty is sometimes not acknowledged is that managers and/or scientists are concerned that fishermen might use uncertainty as justification for less conservative regulations. Recommendations: NMFS should take stronger steps to characterize uncertainty in their assessments and present uncertainty to fishery managers. This may involve an expansion of existing approaches, especially for key species or species complexes, to include a better representation of the level of each kind of uncertainty listed above. Recognizing that the added computational effort may slow the assessment process, the committee recommends that NMFS prioritize the types of uncertainty to address first, moving on to lower priority uncertainties as time and resources permit. Some sources of uncertainty might be determined less frequently and during model development, whereas others should be included with every new assessment. Approximations (e.g., Gaussian approximation using the Hessian approach) might be used for annual updates, whereas better and more complete estimates (e.g., Monte Carlo, Markov Chain, or full bootstrap procedures) might be carried out as assessments are reviewed or when new assessments are developed. NMFS must also decide how it wants to evaluate and present information on variability and uncertainty in a way that helps and does not hinder management. It is important that a case be made for why significant uncertainty should compel managers to be cautious, rather than giving a signal to stakeholders that stock assessment scientists do not know the condition of the stock and thus become an excuse for inaction or for selection of TAC levels that are too high. Presentations should include information about the risk involved in different TAC levels that could be set. Review Findings: The information content of an assessment, namely the data and the model structure and assumptions, should be subject to periodic peer review. This serves two purposes. The first is to have independent scientists scrutinize the process and the product to give public assurance of its scientific integrity, or to identify problems in the analyses. The second is to inject new ideas and research directions into the process that reflect the current state of the art. NMFS and the councils have experimented with different methods of peer review. Currently NMFS, in cooperation with the councils, has several ad hoc committees that review completed assessments prior to submission to a council. Two such committees are the Stock Assessment Review Committee (SARC), which reviews NMFS' East Coast assessments, and the STock Assessment Review (STAR) panel, which reviews NMFS' West Coast assessments. A standing body of advisory scientists called the Scientific and Statistical Committee (SSC) also exists for each council. In addition, several plan teams exist for each council, but their primarily role is to provide management guidance rather than provide a review per se. This peer review system, as it currently operates, has some shortcomings. The SARC and STAR processes, although comprehensive, are not truly independent because the review scientists are typically NMFS employees from other regions and state scientists from the region of concern. The SSC, on the other hand, although generally showing greater independence, at least in terms of composition (although NMFS scientists are on these committees also), does not have the opportunity to examine the issues in as fine a detail as the SARC/STAR panels or as, for example, the committee has in the course of its review. Together, the SSCs and SARC/STAR groups represent the thoroughness and independence needed in a review, but deeper consideration of this process might lead to a more constructive framework. NMFS is experimenting with a Center for Independent Experts to conduct independent reviews, but this approach most often has used single external reviewers rather than teams.
OCR for page 165
IMPROVING THE COLLECTION, MANAGEMENT, AND USE OF MARINE FISHERIES DATA Recommendations: Congress should recognize the need for both a scientific assessment and evaluation process, and a thorough and objective review system. The committee acknowledges that NMFS has taken great effort to provide the means for establishing the scientific integrity of its work, but the committee also recognizes that it is too much to ask NMFS to objectively evaluate that same work. A greater degree of independence in the peer-review process is needed in order to maintain the integrity and scientific credibility of the NMFS assessments. To do this, the scientific products, such as the assessments produced by NMFS and others, should be peer reviewed by scientists who are not directly involved in the assessments or work directly for NMFS (scientists from state fishery agencies could be involved). Because of the limited number of stock assessment scientists outside NMFS and the large number of assessments produced each year, not every assessment should be reviewed each year, but every assessment should be externally reviewed on a regular basis, for example, every three to five years. The SSC could be considered the appropriate independent body for this review, if the SSC is made up of informed but otherwise independent scientists, but time is often a limiting factor for these volunteers in their deliberations leading to council advice. The committee therefore suggests that regional councils consider supporting a stock assessment scientist on staff. Such a scientist could be assigned the task of organizing assessment peer reviews while highlighting issues of scientific and managerial concern under the direction of the SSC and the council executive director. In this way, NMFS could present and defend their work in a public forum and the councils would be able to review this work in an objective fashion. Another vehicle that could be used for such independent reviews is the Center for Independent Experts that is funded by NMFS but which operates independently from NMFS. RESEARCH NEEDS Finding: Stock assessment science and fisheries management are still developing fields. Improvements in each are still needed and will be fueled by continued research and development. Recommendations: Congress should support and NMFS should continue to fund research to improve our ability to characterize fish stocks quantitatively and manage them in the context of the important but sometimes conflicting goals of the Magnuson-Stevens Act and its National Standards. NMFS should fund both internal and external research (biological, economic, and social) relating to: developing methods for evaluating ecological benefits of fish stocks and fisheries; developing new methods for stock assessment; minimizing data fouling and misreporting; testing adaptive sampling for surveys, including both NMFS and industry data collection; testing electronic logbooks and VMSs that offer value-added features to fishermen; linking environmental, economic, and social data, and climate forecasts to stock assessments; studying the feeding habits and the distribution and types of prey and predators of important non-commercial species, to understand the functioning of the marine ecosystems affected by fishing activities; understanding the economic and social motivations of harvesters so that greater use can be made of fishery-dependent data; improving design of recreational fishing surveys; and conducting stock assessments combining recreational and commercial data with very different error and uncertainty structures. Emphasis should be given to research exploring the relationships between different types of regulatory approaches and fishermen's attitudes and behaviors toward fish harvest and data reporting. Research should also identify the most important incentives and disincentives that could be used to promote accurate reporting.
Representative terms from entire chapter: