Why NEFSC Conducts Surveys
Winter trawl survey along the continental shelf (NOTE: Discontinued after 2007 survey)
WHAT TYPES OF GEAR DO WE USE?
To evaluate the status of exploited fishery resources, we must collect and analyze many different kinds of information. Basic landings statistics (including the numbers and weight of each species landed) and demographic data (such as length and age samples) characterize what is brought ashore. At-sea sampling aboard commercial fishing vessels is used to establish the numbers and length/age composition of animals culled overboard. Telephone recall and roving samplers are used to estimate recreational catches. Effort data collected with catches are combined into indices of stock abundance based on catch-per-unit effort (CPUE) ratios. Data from these three types of collection programs are generally referred to as fishery-dependent information: these data are derived directly from the commercial and recreational fisheries.
Fishery-dependent data are vital to our ability to monitor stocks, and for some species are often the only reliable source of data. However, use of fishery-dependent data alone may severely limit our ability to evaluate and make predictions about the status of some stocks. For example, in fisheries heavily dependent on the yearly incoming age group (the new recruits), fishery data alone cannot be used to forecast catches because very small fish are generally not taken with standard fishing gear. Likewise, CPUE may not be a reliable measure of abundance for schooling species, or when the increase in fishing technology cannot be factored into the relationship between catch and fishing effort. Consequently, fishery scientists throughout the world are conducting research vessel sampling programs to gather fishery-independent information (Clark 1981). The scope and use of research vessel surveys in assessing Northeast fishery resources is reviewed here.
Fishery-independent surveys of Northeast fishery resources are conducted for six important reasons:
(1) To monitor recruitment: Research surveying is generally conducted with sampling gear equipped with smaller mesh than is allowed in most fisheries. Small-mesh gear is used in order to estimate the abundance of very small animals that will eventually become large enough to be caught in standard fishing gear. To predict future landings and stock sizes, we must estimate the survival of fish already large enough to be retained by harvesting gear as well as the incoming recruitment to the fishery each year. Depending on the species, research vessel surveys can allow extrapolation of the strength of incoming age groups up to several years before they are allowed to be landed. For example, American plaice reach the minimum size for landing (14 in.) at about age 6. Trawling surveys begin to sample small American plaice during their first year of life, thus allowing five estimates of the relative numbers of small plaice in the population before animals are large enough to land.
For some species, however, growth rates are much faster, and thus the time interval between when fish are detected by the surveying gear and when they are landed by the fisheries is much shorter. In the case of the Atlantic sea scallop, annual dredge surveys are conducted. The relative catch in numbers of sea scallops per haul of the dredge is plotted for two size categories of scallops: those with a shell height smaller than 70 mm (2-3/4 in.) are designated prerecruits; and animals larger than or equal to 70 mm are designated harvestable. The abundance of all sizes of sea scallops has declined significantly since the peak in 1989.
(2) To monitor abundance and survival of harvestable sizes: Research vessel samples generally span the full size and age range of a population on the shelf. Although recruitment prediction is one important element of fishery forecasts, it is equally important to calculate the survival rate of the portion of the stock already subjected to fishing. The catch-at-age data collected from the surveys are one important source of information used to estimate survival rates from one year to the next. A simple estimate of the survival of various age groups in the population can be computed from the indices of abundance in two consecutive years. If the catch-per-trawl-haul of age 4 cod was 100 individuals in 1991 and the catch-per-haul of the same fish, now age 5 in 1992, is 40 individuals, the estimated survival rate is 40/100 = 0.4 = 40 percent. In practice, fishery scientists usually combine catch-at-age data from the surveys with similar data from the fishery catch to improve estimates of fishing mortality and stock sizes. These combined estimates allow calculation of the population that must have existed to yield the catch levels observed during the recent history of the fishery.
Sampling the abundance of harvestable sizes from research vessel surveys may be the only source of data available for species that have never been fished in the past, or are only fished at very low levels. Thus, dredging surveys conducted in the 1960s and 1970s were the only source of information on the abundance of the ocean quahog resource of the Middle Atlantic, Southern New England and Georges Bank areas. Minimum population estimates were made by expanding the average catch-per-square-nautical-mile from the surveys by the number of square nautical miles of sea bottom inhabited by the stock. Similarly, current knowledge of the stock biomass of spiny dogfish and skates is based only on surveys, since catch-at-age based studies have not been undertaken.
(3) To monitor the geographic distribution of species: Some species lead sedentary lives while others are highly migratory. Research vessel surveys are a major source of data on the movement patterns and geographic extent of stocks. Distribution maps can be drawn from reports of fishermen, but these may give a biased picture of the stock, emphasizing only where high density fishable concentrations exist. Distribution data are important not only for fishery management, but also for evaluating the population level effects of pollution and environmental change.
(4) To monitor ecosystem changes: With few exceptions, surveys conducted by the Northeast Fisheries Science Center are designed to be multipurpose. Bottom trawl surveys are not directed at one species, but rather generate data on nearly 200 species of fish and invertebrates in northeastern U.S. continental shelf waters. Many of these species are relatively rare, and have little or no commercial or recreational value. However, when we evaluate the effect of intensive harvesting on selected species, we can observe the response of the entire animal community. The dramatic changes in the system reflect the depletion of several important commercial fishery species, such as, haddock, yellowtail flounder, pollock and plaice and an increase in winter skate, spiny dogfish, and other commercial fishery catches. These data suggest ecosystem-level responses to intensive harvesting, which may have important implications for developing harvesting strategies for the community of species, rather than the individual stocks. A multispecies surveying approach thus provides an important research opportunity in the emerging field of ecosystem-based management.
(5) To monitor biological rates of the stocks: Apart from basic information on the abundance and distribution of species, research vessel survey data are collected on a range of biological rates for stocks. These processes include growth, sexual maturity, and feeding. Changes in growth and maturity parameters directly influence assessment calculations related to spawning stock biomass, yield per recruit and percent of maximum spawning potential. Over the past three decades, these parameters have changed dramatically for some species. Faster growth and earlier onset of maturity have been observed for haddock and cod. It is thus important to monitor these rates continuously if stock status is to be accurately determined. Likewise, diet data collected via examination of stomach contents at sea will be increasingly important as scientists try to evaluate how harvesting affects species linked by predator-prey relationships.
(6) To collect environmental data and support other research: Research vessel surveys are generally conducted 24 hours a day when the vessels are at sea. This presents a superb opportunity to collect environmental information (temperature, salinity, pollution levels, and so on) and to allow other researchers to piggyback on surveys to collect a host of data not directly related to the stock assessment. All research vessel surveys conducted by the Northeast Fisheries Science Center collect and archive an extensive array of environmental measurements and usually have a "shopping list" of samples to be obtained for researchers at academic institutions, other government agencies, and the private sector. On every survey there are scientific berths allocated to cooperating scientists and students in order to foster this cooperative approach to marine science.
The various types of research vessel surveys conducted by the Northeast Fisheries Science Center are described below:
(1) Spring and autumn bottom trawl survey: The spring and autumn bottom trawl surveys conducted by the Northeast Fisheries Science Center are the longest running continuous time series of research vessel sampling in the world. The autumn survey began in 1963; the spring in 1968 (Azarovitz 1981). These surveys cover the ocean environment from 5 to 200 fathoms deep, from Cape Hatteras, North Carolina to well beyond the Canadian boarder. About 300 half-hour trawl sets are made at sites (stations) randomly chosen prior to the beginning of each survey. The objective of each tow is not to catch large numbers of fish, but to capture a representative sample of the various species and relative numbers in a given area. The distribution of trawling locations is allocated according to a statistical method that divides the region into a number of smaller areas (strata) with similar depth characteristics. The method is a stratified-random sampling design, and is commonly used for a wide variety of statistical estimates, including exit polling for elections. In the history of the trawl surveys, only three research vessels -- NOAA's FSV Henry B. Bigelow , FRV Albatross IV and the FRV Delaware II -- have been used to conduct these surveys.
A small-mesh cod-end liner (1/2 inch mesh) is used to retain prerecruits. All species in each tow are weighed and counted, and all or a sub-sample is measured to determine the length composition of the catch. Hard parts (scales, ear stones, fin rays) are removed from some of the fish of each species taken in the trawl, and then cataloged. These hard parts are used to determine the age of each fish selected. The age distribution of the whole catch can then be estimated by expanding the sub-sample. Fish are also examined to determine sex and state of sexual maturity. Stomach contents are identified and any obvious disease-related conditions of the fish are recorded. All data are brought back to the laboratory, where they are entered into computer files. The accumulated trawl survey data set (1963 to present) represents over 20,000 stations, with millions of individual pieces of information concerning fishery resources of the region. The entire data series is available to fishery scientists wishing to examine trends in abundance, distribution, species associations, or numerous other scientific questions.
(2) Sea scallop dredge survey: Each summer, the Northeast Fisheries Science Center conducts a research survey targeting sea scallops. This survey is used to assess scallop abundance, distribution, size/age composition and other factors. The survey encompasses the continental shelf from Cape Hatteras through Georges Bank and the Gulf of Maine. The scallop survey uses the same stratified random survey design as that used in trawl surveys.
Scallop surveys began in 1975, and have been conducted every year since 1977. Since 1979, the standard gear has been an 8 ft-wide commercial scallop dredge equipped with a 2 in. ring bag and a 1-1/2 in. mesh liner. The dredge is towed for 15 minutes. The depths surveyed range from 15 to 60 fathoms. In addition to sea scallops, the surveys catch significant numbers of flounders (primarily yellowtail), hakes and goosefish. Results from the sea scallop survey have been used in assessing the other species.
(3) Hydraulic clam dredge for surf clam and ocean quahog: Since 1965, the Center has conducted hydraulic clam dredge surveys aimed at the surfclam and ocean quahog. The surveys are not conducted every year for several reasons The exploitation rates of these species are low: the time from first appearance in the survey gear until commercial size is attained is long: and fisheries are not generally dependent on the incoming year classes. The current schedule calls for the clam survey to be conducted every third year and the last survey was conducted in 2005.
A 5 ft-wide hydraulic clam dredge is used on the surveys. The dredge has a submersible electric pump that uses high pressure water jets to loosen substrate and animals in the path of the dredge. The submersible pump allows the dredge to fish in depths greater than those now commercially exploited (based on commercial standard deck-mounted dredge pumps). The 2 in. aqua mesh-lined dredge retains small-sized clams, along with large quantities of shell debris and live clams and associated invertebrates, all of which are included in the database. The areas primarily surveyed are from Cape Hatteras to Georges Bank. Additionally, some surveying has occurred in Massachusetts Bay (to evaluate fishery potential for the Arctic surfclam, Mactromeris polynyma) and off the Maine coast (to evaluate populations of ocean quahog).
(4) Summer Gulf of Maine trawl survey: Beginning in 1991, the Center has conducted a special bottom trawl survey directed to near shore areas in the Gulf of Maine. Traditionally, it has proved impractical to survey Maine, New Hampshire, and northern Massachusetts coastal waters with bottom trawls based on random station selections because of large areas of hard, rocky bottom and fixed fishing gear such as lobster pots and gill nets. Hence, large concentrations of inshore juvenile groundfish have not been adequately included in trawl catches. The survey methods used in the summer Gulf of Maine survey are a modification of the stratified random technique. In spring 1991, Center personnel met with fishermen and others knowledgeable about locations of towable areas in the Gulf of Maine. From these meetings, a master list of towable areas was generated and based on this list sites are chosen randomly as the stations to be sampled. This way, expensive and time-consuming gear damage is minimized, but the statistical criteria necessary for the survey are met. Prior to the 1992 survey, NMFS again consulted with Maine fishermen and expanded the master list of potential tows. The survey has generated important new information on the inshore distribution and abundance of groundfish, and has added to the biological data on the timing of spawning for a number of species.
Apart from the new Gulf of Maine survey, a monthly sampling program has been established in conjunction with the Maine Department of Marine Resources to improve the quality of groundfish population biology data.
(5) Winter trawl survey along the continental shelf: (NOTE: Discontinued after 2007) Initiated in 1992, a winter trawl survey along the Middle Atlantic, Southern New England and southern Georges Bank continental shelves is specifically directed to improve the quality of flatfish assessments. Standard groundfish surveys use rollers along the foot rope to minimize gear damage in rough bottoms. Although the roller-rigged gear catches flatfish, many of the animals pass under the foot rope, and thus abundance measures from the survey tend to be highly variable. The new survey employs a chain sweep (a "flat net") to minimize the escapement of flatfish under the foot rope. Accordingly, abundance indices from the new survey are likely to provide more accurate assessments than those derived from roller-rigged gear. Specifically, the survey has been used to improve assessments for summer flounders. Assessments of other species caught along with this flatfish have also been improved.
(6) Marine mammal sighting surveys: Shipboard sighting surveys of marine mammals involve directing the vessel along a predetermined transect and counting the number of each marine mammal species sighted. The interpretation of the data is complicated by the range and bearing of the sighting from the vessel track line, and the fact that the probability of sighting a mammal falls off with its distance from the vessel. These effects are well known when conducting sighting surveys, and are being evaluated by Center scientists. An intensive survey to estimate the abundance of harbor porpoise in the Gulf of Maine began in 1991. Other sighting surveys have been conducted to evaluate the abundance and distribution of the marine mammal community in coastal shelf waters and those associated with the western boundary of the Gulf Stream. Marine mammal sightings are routinely conducted as a piggy-back activity on bottom trawl and larval fish survey cruises.
(7) Surveys of fish eggs and larvae: Surveys of the distribution and abundance of "baby fish" (the free-floating eggs and new hatchlings called larvae) are conducted several times per year to evaluate the timing and distribution of spawning. Another important function of these surveys is to estimate the quantity of baby fish spawned, and thereby be able to work back to the size of the female population that must have been present to produce the numbers of small fish counted. This back-calculation of the spawning stock is an important tool used in assessing fish populations worldwide, and in some cases may be the only reliable information about the size of the spawning stock. For Northeast fishes, egg and larval studies have been used to evaluate populations such as Atlantic mackerel , yellowtail flounder , sand lance, and bluefish.
Egg and larval surveys are conducted using a pair of very fine-mesh nets towed in a frame resembling a bongo drum (hence the name "bongo nets"). The nets are pulled with varying amounts of scope on the towing line so the bongo net samples all layers of the water column from the surface to just off the bottom. Contents of the net are preserved at sea for later intensive analysis in the laboratory. These fine-mesh plankton nets also sample the community of free-floating plants and animals that support the base of marine food chains. These animals are assessed to examine the production and distribution of zooplankton (animals) and phytoplankton (plants).
(8) Special experiments: A variety of special experiments are conducted to augment the standard monitoring surveys. For example, when vessels or surveying gear have to be changed, it is necessary to estimate conversion factors accounting for differences in fishing power. Several years ago the standard trawl net doors had to be changed, since fishing gear suppliers could no longer manufacture doors to 1963 specifications. A polyvalent door was chosen for the new standard, and a series of research vessel experiments was conducted to estimate the effect that the door change alone had on catch rates. Similar experiments have been conducted to relate catches between the FSV Henry B. Bigelow and the FRV Albatross IV
Other special experiments using vessel surveys have been directed at evaluating feeding interactions among species; relating oceanographic processes to the survival of eggs and larvae; and evaluating effects of area closures and openings on fish, benthic fauna, and sediment geochemistry.
The fundamental objective of the surveys is to provide realistic estimates of relative abundance of populations over time. By only fishing where fishing is exceptionally good (the "hot spots"), we would potentially miss large fractions of the stock that may occur in lower density areas. This would be akin to estimating the population of Massachusetts or New Jersey by counting only people that live in the eastern counties of these states, and extrapolating the rest based on numbers of people-per-square-mile.
By selecting random stations within certain depth zones, any fish that exists in that zone has an equal probability of being caught in the survey. Thus, the method produces unbaised estimates of the relative stock size. One of the downsides to this procedure is that for fish stocks highly clumped in their distributions, the error bars around the estimates tend to be rather wide. Alternative survey methods that produce more precise error bars include a grid of fixed-station locations that are fished each year. However, this design may provide biased abundance measures if changes in the abundance at these locations do not reflect the actual changes in the stock as a whole. Some countries use the stratified random technique, and others use fixed stations (for example the English groundfish survey of the North Sea; ICES 1992). There appears to be no sampling design that covers all cases. Rather it depends on the particular situation of the species distribution and the number of sampling sites that are included in the survey. Because of the large number of species and stocks analyzed from the bottom trawl surveys and the different bottom types and habitats in the Northeast, the stratified random sampling plan appears to be the most efficient design.
For more information:
Azarovitz, TR. 1981. A brief historical review of the Woods Hole Laboratory trawl survey time series. In: Doubleday, WG and Rivard, D, eds. Bottom trawl surveys. Canadian Special Publication of Fisheries and Aquatic Sciences 58, p. 62-67.
Clark, S. 1981. Use of trawl survey data in assessments. In: Doubleday, WG and Rivard, D, eds. Bottom trawl surveys. Canadian Special Publication of Fisheries and Aquatic Sciences 58, p. 82-92.
International Council for the Exploration of the Sea. 1992. Report of the workshop on the analysis of trawl survey data. ICES C.M. 1992:D:6.