Reference Document Home | Publications Home

I. Bottom Trawl Fishery
II. Mid-water Trawl Fishery

List of Acronyms

Northeast Fisheries Science Center Reference Document 07-15

Characterization of the Northeast and Mid-Atlantic Bottom and Mid-water Trawl Fisheries Based on Vessel Trip Report (VTR) Data

Christopher D. Orphanides and Gisele M. Magnusson
National Marine Fisheries Serv, Woods Hole Lab, 166 Water St, Woods Hole MA 02543-1026

Web version posted September 20, 2007

Citation: Orphanides CD, Magnusson GM. 2007. Characterization of the Northeast and Mid-Atlantic Bottom and Mid-water Trawl Fisheries Based on Vessel Trip Report (VTR) Data. U.S. Dep. Commer., Northeast Fish. Sci. Cent. Ref. Doc. 07-15; 127 p.

Information Quality Act Compliance: In accordance with section 515 of Public Law 106-554, the Northeast Fisheries Science Center completed both technical and policy reviews for this report. These predissemination reviews are on file at the NEFSC Editorial Office.

Acrobat Download complete PDF/print version

ABSTRACT: An overview of the United States Northeast and Mid-Atlantic trawl fisheries was created to support the development of a Take Reduction Plan (TRP) by the Atlantic Trawl Gear Take Reduction Team (ATGTRT). The analysis focused on aspects of the fisheries that correlated with marine mammal bycatch and characterized the trawl fisheries from Maine to North Carolina for the period 1996 to 2004 by temporal and spatial effort distributions, harvest patterns, vessel characteristics, gear configurations, and environmental relationships. The primary data used were vessel trip reports (VTRs); additional data sources included remotely sensed data for environmental variables, federal commercial dealer records, and vessel permit data.

The number of vessels in the bottom trawl fishery declined by 28% between 1996 and 2004. Over the same period total trips declined by 15%, days absent declined by 26%, and days fished declined by 40%. Total landings varied from a low of 92,159 metric tons (live) in 2002 to a high of 116,911 metric tons in 2004. Based on a 9-year average of landings, the top 5 species were Loligo squid, silver hake, Illex squid, skates, and monkfish. Inflation adjusted value declined for the fishery, reaching a low in 2003 but increasing in 2004. Based on a 9-year average of adjusted values, the top 5 species were Loligo squid, fluke, monkfish, cod, and winter flounder. Over all years, the majority of bottom trawl effort occurred north and east of the Hudson Canyon, although some significant effort occurred along the shelf break south of Hudson Canyon. Effort shifted tothe south during the winter. Southern regions tended to use nets with a smaller mesh than northern areas, while gear size, measured as foot rope length, was closely linked to vessel size.

The number of vessels in the mid-water trawl fishery declined by 18% between 1996 and 2004. Effort measured in trips declined by 1%, days absent declined by 5% and days fished declined by 38%. There were clear differences in data on the fishery after the implementation of the Atlantic Herring Fishery Management Plan in 2001. The 9-year average landings were composed of 80% herring and 18% mackerel, while herring accounted for 65% and mackerel 27% of the average inflation adjusted value. Total landings grew steadily and peaked in 2004; the inflation adjusted value doubled over the 9 years. The mackerel and herring fisheries had a small overlap in effort south of Rhode Island. Herring was harvested year round, with effort focused in the Gulf of Maine (GOM), the northern edge of Georges Bank (GB), and south of Rhode Island. Mackerel were fished from south of New England down through the mid-Atlantic from November through May with larger gear (foot rope length) and larger mesh size than herring.

Both fisheries landed the majority of their catch in Massachusetts; other important states included Rhode Island, Maine, New York, and New Jersey. Effort in both fisheries was concentrated in waters where the sea surface temperature (SST) was less than 10 °C, depth was less than 100 meters and bottom slope was less than 0.5°.

VTRs are compulsory for nearly all vessels participating in the U.S. Northeast and Mid-Atlantic bottom and mid-water trawl fisheries; thus, the records should nearly provide a census of fishing effort for the fisheries. Data accuracy, in particular with measures of effort, was examined; over 90% of the records fell within the norms for the variables. It appears that the VTR data provides a reasonably accurate overall picture of the trawl fishery.


Bottom and mid-water trawl fisheries are a major component of the United States Northeast and Mid-Atlantic fisheries. In September 2006, the National Oceanic and Atmospheric Administration’s (NOAA) National Marine Fisheries Service (NMFS) convened the Atlantic Trawl Gear Take Reduction Team (ATGTRT) to address incidental takes of pilot whales (Globicephala spp.), common dolphins (Delphinus delphis), and white-sided dolphins (Lagenorhynchus acutus) in U.S. Atlantic trawl fisheries. The goal of the ATGTRT is to develop a Take Reduction Plan to limit the serious injury and mortality of these marine mammals due to accidental entanglement in commercial fishing gear. This paper characterizes a subset of the U.S. Northeast and Mid-Atlantic trawl fisheries to aid in the development of a take reduction plan. Specifically this paper aims to assess characteristics of the fishery that were identified as correlated to marine mammal bycatch. However, it is also expected that this characterization will be comprehensive enough to serve as a reference for other research on these fisheries.

The trawl fisheries havebeen the subject of much analysis; however, most studies have focused on subsets of the fisheries. As such, the recent literature lacks an overall characterization or description of the U.S. Northeast and Mid-Atlantic trawl fisheries for use by the ATGTRT. Studies within the peer-reviewed literature tend to focus on small subsets of the fisheries and issues may not be relevant for management considerations under a take reduction team process. A significant amount of information on the fisheries is contained in various fishery management plans (FMPs)[1], in particular the Northeast Multispecies FMP documents. However, trawl gear is one of several gear types examined and vessels identified for inclusion in the analysis may be limited to those that hold permits for the species of interest. Rountree (1997) summarized the spatial and temporal patterns for otter trawl vessels for 1982 to 1992. The age of the analysis limits comparisons with this study[2]. A study by Stevenson et al. (2004) focused on gear impacts on fish habitat, and mapped effort in days absent for bottom trawl vessels. The maps allow for partial validation of the results of this study; however, characteristics important to bycatch reduction are not examined[3].

Historically, reductions in bycatch have been achieved through methods such as gear modifications and time and/or area closures, which require knowledge of the relationship between bycatch, the fleet, and fishing characteristics. Recent research suggests that trawl bycatch of pilot whales, common dolphins, and white-sided dolphins are primarily a function of the spatial and temporal aspects of fishing effort (Rossman [in prep]; Palka [in prep]). However, numerous other fishing characteristics such as gear fished, vessel characteristics, and species targeted have been shown to be correlated with bycatch and could play a role in mitigation measures. As such, this paper covers a broad range of characteristics including the spatial and temporal aspects of fishing effort, the gear fished, the species harvested, and the vessels involved. In addition, this paper summarizes fishing effort and revenues to provide information that could help assess the impacts of proposed mitigation measures.

The organization of the paper is as follows. First, to provide context for the variables examined and the changes identified, information is provided on trawl bycatch calculations and relevant changes in management regimes. Second, fishing effort and variables correlated with bycatch are summarized for 1996 to 2004 and presented separately for bottom and mid-water trawl fisheries. Third, the discussion identifies data concerns and areas for future study.


I. Variables important to assessing marine mammal bycatch in trawl fisheries

Studies on bycatch of marine mammals in the U.S. Northeast and Mid-Atlantic bottom and mid-water trawl fisheries were used to select the fishing characteristics summarized in this paper. The primary predictor variables identified varied by bycatch species and fishery. In the bottom trawl fishery the predictor variables were target species[4], vessel horsepower, and bottom slope for pilot whale bycatch, SST, and bottom depth for white-sided dolphin bycatch, and statistical area for common dolphin bycatch (Rossman [in prep]). In the mid-water trawl fishery bottom depth and latitude were the most important factors in estimating white-sided dolphin and pilot whale bycatch (Palka [in prep]). The primary predictor variables were used to estimate bycatch rates that were expanded to estimate total bycatch using stratified effort measured in days fished.

In addition to the predictor variables, there were other variables that had significant correlations to bycatch rates but were not used as predictor variables. The correlations do not necessarily indicate a cause and effect relationship, but do point to possible research areas that could lead to effective ways to limit bycatch. The bycatch models use data from the NMFS observer database. The Observer records include many variables that are not included in the vessel trip reports (VTRs) or permit databases nor can they be derived from VTR data. For the bottom trawl fishery[5], significant correlated variables in addition to the predictor variables mentioned above were:

For the mid-water trawl fishery, additional significant correlated variables included:

II. Relevant fisheries management actions, 1996–2004

Management actions have the potential to affect a number of variables examined in this report[7]. The diversity of harvest by the U.S. Northeast and Mid-Atlantic trawl fishery puts it in contact with a large body of management measures. Management authority for the species caught by the trawl fisheries resides with the New England Fishery Management Council (NEFMC), Mid-Atlantic Fishery Management Council (MAFMC), and the Atlantic States Marine Fisheries Council (ASMFC). There are 37 FMPs between the 3 organizations, covering about 60 species (Appendix A). As well, the national Highly Migratory Species Plan covers sharks, swordfish, tuna, and billfish. Despite this range of regulations, the principal species harvested suggests a smaller number of FMPs are of major importance to the U.S. Northeast and Mid-Atlantic trawl fishery; in particular, the FMPs for groundfish (multispecies), Atlantic herring, monkfish, mackerel/squid/butterfish, and summer flounder/scup/black sea bass. Aspects of these FMPs relevant to this report are described below.

i. Northeast Multispecies (Groundfish) FMP

By 1996, some management measures were already in place under the Northeast Multispecies FMP. These measures included a minimum mesh size (5.5”), a ban on the use of paired trawls, seasonal closures in the Gulf of Maine (GOM), Georges Bank (GB), and Southern New England (SNE), and year-round closures in Closed Area (CA) I, CA II, and the Nantucket Lightship CA. Amendment 5 (1996) placed a moratorium on new entrants to the fishery and developed a plan to reduce effort by decreasing historical days-at-sea (DAS) by 50%. Mandatory VTR and observer program requirements increased the information available on the fishery. Minimum mesh size for the GOM/GB Regulated Mesh Area (RMA) (Figure 1) was increased to 6.0”.

Closed areas have been expanding in size and duration since 1996, although the general locations were consistent; Figure 2 (seasonal) and Figure 3 (year-round) represent groundfish closed areas as of 2007. Under Amendment 7 (1996), the GOM seasonal closures were expanded to cover all gear types, including trawl gear. The Cashes Ledge CA was implemented as a seasonal closure in 1998, but became year-round in 2002. The Western GOM year-round closure was implemented as a temporary measure in 1998, but remained beyond 2004. In 2004, Amendment 13 created 7 essential fish habitat (EFH) closure areas, which prohibited bottom trawl gear year-round.

Numerous amendments and framework adjustments continued to cap and reduce DAS for vessels. Amendment 13 (2004) resulted in significant changes to DAS, setting unrestricted DAS at 60% of the vessel’s baseline DAS.

Regulated changes in mesh size have been limited. In 1999, the minimum mesh size was increased to 6.5” square for the GOM, GB, and SNE RMAs. This has remained unchanged except for SNE where the minimum was increased to 7.0” square or 6.5” diamond in 2002. Amendment 13 (2004) clarified the mesh size requirements, indicating that vessels with large mesh permits were to have minimum mesh size of 8.5” diamond or square throughout the net in GOM, GB, and SNE RMA, and 7.5” in Mid-Atlantic RMA.

ii. Atlantic Herring FMP

When the Atlantic Herring FMP was developed, the resource was considered under-exploited (New England Fisheries Management Council 1999). The ASMFC Herring FMP was implemented in 2000, while the NEFMC Herring FMP was effective for the 2001 fishing year. Thus, prior to 2001, herring records are considered incomplete (New England Fisheries Management Council 2006), as not all herring vessels were engaged in other fisheries that had VTR requirements.

The FMP limited herring permitted vessels to less than 165 feet in length overall, less than 750 Gross Registered Tons (GRT) or with propulsion less than 3,000 shaft horsepower. The fishery was divided into 4 fishery management areas (FMAs; Figure 4). Each year a target total allowable catch (TAC) was determined for each management area. When catch reached 95% of the FMA’s TAC, targeted harvesting would be halted. Closures, in particular for FMA 1A, were common. In 2002, Framework Adjustment 1 created two quota periods in FMA 1A, winter/spring (January–May) and summer/fall (June–December), in an effort to increase the amount of quota available during the summer/fall peak demand period.

iii. Monkfish FMP

The monkfish fishery is jointly managed by the NEFMC and the MAFMC. The monkfish FMP came into effect November 1999 and thus data may be incomplete for earlier years. Limited access permits differentiated between those in the directed fishery and those for which monkfish was primarily a bycatch (holders of multispecies and scallop DAS). Permit holders were each allocated 40 Monkfish DAS. The TAC was to be achieved through a combination of limited access and trip limits. The fishery was divided into the Northern FMA and the Southern FMA; the dividing line runs from the intersection of 70° W. longitude and the south-facing shoreline of Cape Cod, MA southward along longitude 70°W to latitude 41°N, then eastward to the U.S.-Canada maritime boundary. Trip limits have been more restrictive in the Southern FMA. When using only a monkfish DAS (i.e., directed fishing), the minimum mesh size for trawls was set at 10” square or 12” diamond[8].

In May 2003, Framework Adjustment 2 increased trip limits for the Southern FMA for monkfish-only vessels, while there remained no trip limit for the Northern FMA. The 2004 rule restricted use of a vessel’s DAS in the Southern FMA to a maximum of 28 DAS and reduced trip limits for the area.

iv. Atlantic Mackerel, Squid, and Butterfish FMP

The MAFMC plan covers all Atlantic mackerel, butterfish, Loligo pealei (long-finned squid), and Illex illecebrosus (short-finned squid) under U.S. jurisdiction. In May 1996, Amendment 5 included a joint moratorium permit for Loligo squid and butterfish, a minimum mesh size of 1 7/8” diamond for Loligo vessels, and seasonal quotas for Loligo. Vessels fishing for Illex squid during June–September seaward of a line roughly matching the 50-fathom line were exempt from the minimum mesh size requirements; at that time the Illex squid fishery was considered underutilized. Closures for the directed fisheries in Loligo, Illex, butterfish, and mackerel are based on a percentage of the Domestic Annual Harvest level; the trigger is 80% for mackerel and 95% for all others. Over the years both Loligo and Illex squid fisheries have been closed as the triggers were reached.

In the late 1990s, mackerel was considered underutilized, and a declining annual allocation to joint venture processing sector was provided. Amendment 8 (1999) placed a size restriction on vessels that could fish in the Atlantic mackerel fishery to less than or equal to 165 feet in length and 750 GRT or less than or equal to 3,000 shaft horsepower.

v. Summer Flounder, Scup, and Black Sea Bass FMP

The MAFMC was responsible for the summer flounder (fluke) FMP; in 1996, scup and black sea bass were added to the FMP. Beginning in 1997, vessels issued permits for these species were required to submit VTRs.

All 3 species have been managed with commercial landings quotas that have resulted in closures of the fishery. Fluke used state level quotas, and state level closures have been frequent. Black sea bass initially operated under a coast-wide quarterly quota, but that was converted to a state level quota in 2003. Scup has operated under a coast-wide trimester quota.

Minimum mesh size requirements for trawl vessels differed by species. In 1997, trawls holding a scup moratorium permit were required to use nets with 4” or more diamond mesh. Black sea bass moratorium permit holders were required to use nets with a minimum of 4” diamond or 3.5” square mesh. For fluke, mesh had to be 5.5” diamond or 6.0” square mesh. In 1998, the scup minimum mesh size requirement was set at 4.5” diamond. In 2001, mesh size for black sea bass was increased to 4.5” diamond.


The primary source of data for this analysis was the fishing VTR database. Additional data sources included the Commercial Fisheries Dealer Database (CFDBS), often termed “dealer data,” and the NMFS Northeast Regional Office’s (NERO) permit database.

The VTR is logbook information provided by fishermen, after completing a fishing trip. The number of vessels required to submit reports have grown over the years, and most federally managed fisheries in the Northeast required VTRs by 2004[9]. Owners or operators of federally permitted vessels are required to provide reports for all trips, including trips targeting species without VTR requirements and trips with no catch. A separate report is required for each portion of a trip when changes occur in gear (mesh or gear size) or location (based on chart or statistical area). Thus, a single trip may have multiple reports, or sub-trips. Each report (Figure 5) includes: (i) vessel and operator identification; (ii) dates of sailing and landing and port landed; (iii) type of trip; (iv) gear characteristics; (v) location fished; (vi) descriptors of effort; (vii) details on species caught, including kept and discarded, and (viii) to whom sold.

VTR data from 1996 through 2004 was retrieved for otter trawl fish (OTF), otter trawl other (OTO), otter trawl mid-water (OTM), and paired mid-water trawl (PTM) gear types. These fisheries had documented marine mammal bycatch and were part of the ATGTRT.

Key variables, including gear size, mesh size, tow time, and number of hauls, were checked for errors, and unreasonable outliers were replaced with median values. Most outliers were determined using a table of reasonable expected values developed for data quality checks. The table of reasonable expected values was developed by examining both observer and VTR data to determine the real world maximum and minimum values for particular variables in each fishery. However, some variables, such as the number of hauls per trip, were not included in this table. In this case, the number of hauls per trip was examined through calculation of the number hauls per day, since the length of trips vary. After examining the distribution of both VTR and observer data, 10 was set as the maximum number of hauls per day. Tow hours and number of hauls were also compared to the amount of fish kept as a means of checking for invalid zero values. For example, if a trip had positive pounds kept but no hauls, then the number of hauls was replaced with a median value. Tow hours were dealt with in a similar manner; for instance, positive tow hours with no hauls were not considered valid.

To calculate median values, trips were stratified by hull number, gear type, FMP group[10], year, and state; each unique combination of these variables had a matching median. Outliers were substituted with the stratified median values; each replaced value was representative of similar trips by the same vessel. In cases where an outlier needed to be replaced and no median was available based on the stratification, the hull number was dropped from the stratification criteria and a median was created based on the remaining variables. About 2% of the data for these variables were replaced with median values, although roughly 6% of the gear size observations were replaced (Table 1).

Two additional variables were created from existing VTR variables: “days fished” and “main species.” A “days fished” effort variable was created using tow hours, tow minutes, and number of hauls recorded. The average tow time per haul for a trip was converted to decimal days and multiplied by the number of hauls for that trip. A variable describing the “main species” landed on a trip was based on the species with the largest percentage of the trip’s live weight[11]. For example, if a trip caught more pounds of fluke than any other individual species, the characteristics of that trip were fully attributed to fluke. The implicit assumption was that vessel captains were targeting the species of which they caught the most, having modified gear, season, and location accordingly. The main species categorization was used to examine gear characteristics and effort measures. For some species (e.g., mackerel) more than 80% of the live weight reported in the VTR was harvested by trips where that species (e.g., mackerel) was identified as the main species (Table 2). However, for some species (e.g., cod, plaice, and witch flounder) less than 50% of the total reported live weight landed was harvested on trips where that species (i.e., cod, plaice or witch flounder) was identified as the main species. The impact of this difference on the summary statistics is unclear.

Environmental variables, including SST, bottom depth, and bottom slope, were created based on location information recorded in the VTR dataset. SST data for each trip was obtained from 5-day SST composites derived from several satellite imagery sources, or 5-day climatology images obtained from NASA’s Jet Propulsion Laboratory[12]. Satellite SST imagery sources included AVHRR Pathfinder Version 5, Modis Aqua, Modis Terra, and GOES satellites[13]. These sources were combined when available to create one 5-day median composite image for each day. A Visual Basic for Applications routine in ArcGIS 9.1 extracted SST values at point locations, and using a 3x3 cell window, for both the 5-day median composites and the climatology. The VTR data location variable is reported as the center point of a trip that occurs in one area and in which a consistent gear type and size was used. When choosing which SST to use in the analysis, the 5-day medians were preferred over the climatology, and point locations were preferred over the 3x3 cell medians. Lastly, the resulting SST value was compared to the climatology data to screen for incorrect temperature values.

Depth data for each sub-trip was obtained using bathymetry data acquired from the National Geophysical Data Center[14]. Like SST, bottom depth (Figure 6A) was sampled with ArcGIS at each fishing location recorded by the vessel log. This depth data was preferred over that reported in the VTR for its consistency both in location and in its source[15]. Bottom slope in degrees was calculated using the ArcGIS slope function after projecting the bathymetry data to a Lambert Conformal map projection customized for the area off the Northeast U.S. coast (Figure 6B). Bottom depth and bottom slope are closely correlated.

To estimate revenue, a price series was created from the CFDBS database. For each species, an average weighted price per live pound landed was calculated by year and month. The weighted price took into account the allocation of the species between different grades (e.g., round, dressed) and between ports of landing. For species where the majority of the landings were “unclassified” for a species group (e.g., skates) in the CFDBS, the VTR records were grouped similarly. The price is an average by month over all grade types and ports from Maine to North Carolina, so it does not reflect spatial price differences but does reflect seasonal differences. The value is referred to as the “calculated value” to indicate that it was calculated using the VTR data, rather than taken directly from the CFDBS which is considered a more accurate determination of trip value. Unless indicated, calculated values are in nominal dollars (i.e., not adjusted for inflation), meaning that averages across years are not appropriate. However, where indicated adjusted calculated values were adjusted for inflation using the annual GDP implicit price deflator[16] with a base year of 2000 (i.e., 2000=100).

The NERO permit database was the source of information on vessel characteristics such as GRT, vessel horsepower (VHP), and vessel length. While the majority of vessels within the VTR dataset have associated records in the permit dataset, occasionally a match will not be found. Rather than exclude these vessels from the results, these records are identified separately.

To organize the data spatially, 5 fishing regions were defined by grouping statistical areas (Figure 7A). The regions are based on the fishing areas used by NERO and the NEFMC. The NEFMC regions are not exactly aligned with the statistical area boundaries, so adjustments were made. These adjustments were, in part, guided by the ecosystem regions (Figure 7B) identified based on fish communities (Gabriel 1992). The goal was to define regions that were meaningful from both management and ecosystems perspectives. The fishing regions include: GOM, GB, North Southern New England (NSNE), South Southern New England (SSNE), and Mid-Atlantic (MA). Records that lack location information are included in the “unknown” category rather than excluded. The fishing regions are very different in size, and are subject to different management regulations, so comparisons based on total values should be done with caution.


The results are reported separately for the bottom trawl and mid-water trawl fisheries; however, each section follows a similar format. Each fishery is examined in terms of characteristics of effort, harvest, vessels, gear, and environmental variables.

I. Bottom Trawl Fishery (OTF and OTO)

Only vessels reporting the use of OTF and OTO gear on the VTRs are considered in this analysis. While most of these vessels indicate otter trawl gear on all their VTR records, vessels that occasionally report use of otter trawl gear are also included. The number of vessels using bottom (otter) trawl gear decreased between 1996 and 2004 from 944 to 732 (Table 3), a 23% decline. Despite an increase in the number of fisheries covered by VTR requirements during this period, the number of reported trips decreased from 38,748 in 1996 to 33,089 in 2004, a decline of 15%.

i. Effort characteristics

Bycatch rates are multiplied by stratified effort to calculate total bycatch; thus, it is important to understand the distribution of effort relative to factors correlated with bycatch rates. For the Atlantic trawl bycatch estimates, bycatch rates are expanded by days fished. “Days fished” is an encompassing measure of effort that accounts for tow time and hauls and, to some degree, days absent. The focus of this section is the general spatial and temporal distribution of effort, in terms of days fished and its components. Effort is also considered later when fishing and vessel characteristics are examined in other sections, to provide additional information that could be helpful in the assessment of impacts of proposed TRP mitigation measures.

a. Days fished

Over the entire 9-year period, 1996 to 2004, effort in the bottom trawl fishery was focused from the Hudson Canyon north, with areas of high density along the axis of the Hudson Canyon, along the shelf edge, on GB, and in the GOM, particularly off the Massachusetts coast in the transition zone between the GOM, and GB (Figure 8). Some effort extended south of the Hudson Canyon along the shelf break, and along the mid-coast of Virginia and North Carolina. This overall distribution illustrates some of the key fishing grounds for the bottom trawl fishery; however, some of the discontinuities in the distribution of effort closely match closures for management purposes, discussed and illustrated above. The overall distribution of effort also masks changes over the 9 years in terms of the relative importance of areas.

Between 1996 and 2004, the total days fished by the bottom trawl fishery decreased from 37,381 to 22,324 (Table 4), a 40% decline. Days fished declined in all fishing regions, although the declines were not evenly distributed. The greatest impact was in the GOM, where days fished declined in all years except 2001; the total days fished decreased from 15,806 in 1996 to 8,262 in 2004, representing a 48% decline. This was followed closely by NSNE, where there was a decline in every year except 1998; the total days fished fell from 12,896 in 1996 to 6,926 in 2004, a decrease of 46%. The MA had moderate increases in days fished in 1998 and 2002, but still had an overall decrease from 2,328 to 1,522, for a total decline of 35% over the 9 years. The least affected region was GB, where there were slight increases in days fished in 6 of the 9 years; however, the total decreased between 1996 and 2004 from 5,871 to 5,565, a 5% decline.

As a result of the differences in the decline of days fished by fishing region, the relative importance of the regions in terms of effort has shifted (Table 4). In 1996, The GOM accounted for 42% of total bottom trawl effort (=15,806/37,381), NSNE for 34% (=12,896/37,381), and GB accounted for 16% (=5,871/37,381). By 2004, days fished in the GOM were 37% of the total (=8,262/22,324), NSNE accounted for 31% (=6,926/22,324), and GB accounted for 25% (=5,565/22,324).

The average amount of time spent fishing per trip, as measured by days fished per trip, declined by about 30% between 1996 and 2004 (Figure 9). This decline extended to all fishing regions (Table 5). The relative decline was greater in NSNE where mean days fished per trip decreased from 0.65 in 1996 to 0.36 in 2004, a 45% decline. The smallest decline was in GB, where mean days fished per trip decreased from 3.55 to 3.05, a 14% decline. On average, trips to GB fished the most, while vessels on trips in the NSNE fished the least amount of time per trip.

The amount of effort used in the pursuit of different species has declined overall (Table 6), with a few notable exceptions. Main species trips for haddock increased from 30 days fished in 1996 to 1,594 in 2004, a 52-fold increase, while days fished on Atlantic croaker main species trips doubled from 42 in 1996 to 86 in 2004. There were also small increases in the total days fished for yellowtail flounder and Loligo squid trips. The most significant declines in days fished were for American plaice trips, which declined by 90% from 4,517 days to 424 days between 1996 and 2004, and cod, which declined from 4,863 to 1,484 days, a 70% decline. Silver hake, winter flounder, and monkfish trips also exhibited declines in days fished between 1996 and 2004. These changes resulted in some shifts in the relative fishing effort between species (Table 7). In 1996 cod and monkfish main species trips both accounted for approximately 13% of effort in days fished, while American plaice was second at 12% and fluke third at 10% of days fished (Table 7). In 2004, monkfish accounted for 14% of the days fished, fluke accounted for 13%, and Loligo squid for 11%; cod had fallen to 7% and plaice to 2%.

Seasonality for the bottom trawl fishery is evident in the average days fished per trip by month (Figure 10). The highest means occurred in the first quarter, fell to fell until July then increased through December; the average for February was almost twice that for July. This seasonality could be in part due to seasonal differences in the species targeted.

Disaggregating average days fished per trip by main species illustrates the high degree of variability within the broadly defined bottom trawl fishery (Figure 11). Differences exist between species, and within a species between months. Among the Northeast multispecies, witch flounder, silver hake (Figure 11A), and other groundfish (Figure 11D) exhibited the same seasonal pattern as the overall fishery, while winter flounder (Figure 11A) had an opposite pattern. Haddock (Figure 11A) and monkfish (Figure 11C) had the highest days fished per trip, but with opposite seasonal patterns; thus, haddock was highest in the third and fourth quarters while monkfish was highest in the first 2 quarters.

The spatial distribution of days fished by main species (Figure 12) closely matches the distribution of the species[17]. An examination of the maps of effort indicates that the species harvested break into 5 groups spatially:

b. Tow time and number of hauls

The examination of the haul and tow time variables was limited as they were used to calculate the days fished variable, which was examined in detail above. The total number of hauls decreased from 301,241 to 191,032 between 1996 and 2004 (Table 8), a 37% decline. The number of hauls decreased for all main species, except for Atlantic croaker, witch flounder, yellowtail flounder, haddock, and scup. The allocation of hauls between species (Table 9) is similar to that for days fished by species. However, fluke trips generally have a higher percentage of hauls than days fished, while monkfish is the opposite (Table 9 and Table 7). This suggests that fluke tow times are shorter than average, and monkfish tow times are longer than average.

The mean number of hauls per trip declined by about 30% over the 9 years (Figure 13), while mean average tow time per haul both declined by about 10% (Figure 14). Both variables exhibited seasonality, although there were some differences. The mean number of hauls per trip peaked in February and was at its lowest level in July (Figure 15), with about a 40% difference between the two. Mean average tow time per haul was highest in March and lowest in July–September (Figure 16), with about a 30% difference.

c. Days absent

“Days absent” is calculated from the VTR as the number of days between departure and landing, rounded up to the nearest whole day. Days absent is a combination of steam time and fishing time. In the bottom trawl fishery, days absent declined from 101,790 in 1996 to 74,852 in 2004 (Table 10), a 26% decline. This is less than the decline in days fished discussed above. Days absent could increase relative to days fished if vessels must travel further to fish.

The fishing regions had different patterns in terms of reported days absent (Table 10). The changes in days absent were not consistent between years or in direction; however, the overall trend was downward. The GOM exhibited the largest overall decline in days absent, from 35,635 in 1996 to 20,861 in 2004 (a 42% decline). In the NSNE region, days absent declined from 44,805 to 33,074 (26% decrease). The GB and MA regions both experienced small increases in days absent over the 9 years. The GB region’s days absent increased from 12,839 to 13,085, a 2% increase, while the MA went from 7,294 to 7,600, a 4% increase.

Seasonal variation in length of trips, measured by days absent, is evident (Table 11), however, high coefficient of variation (CV) values mean the differences are not statistically significant. Trips in the first quarter (January–March) were slightly longer, while trips in the second and third quarters were shorter. There appears to a less variation in the mean value, as measured by the CV, in January and February, when trip length is near its maximum mean value.

ii. Harvest characteristics

Landings and value for the fishery are important from the perspective of potential impacts from proposed TRP actions. The spatial and temporal patterns of harvest help identify groups of fishermen and regions that may be affected by TRP mitigation actions, while value provides an indication of the level of economic impacts.

a. Species

Over the 9-year period 1996 to 2004, the main species algorithm classified the majority of the trips as fluke trips, followed by Loligo squid, cod, winter flounder, and silver hake trips (Table 12). The species making up the majority of the trip’s catch by live weight was considered the “main species.” These rankings differed from those based on total value and total kept weight, though the top species are similar. In terms of total kept weight (live) the top 5 species based on a 9-year average were Loligo squid, silver hake, Illex squid, skates, and monkfish (Table 13). The difference in ranking suggests that fluke, cod, and winter flounder trips may be frequent with a small harvest, and indeed there were small trips limits for each species over much of the time examined.

The importance of species by value is expected to differ from measures based on weight as many high volume species (e.g., herring, mackerel, skates, and Illex squid) have relatively lower prices per pound, either nominal (Table 14) or adjusted for inflation (Table 15). Nominal calculated value illustrates within-year differences in value (Table 16); however, to compare values between years the adjusted value must be used (Table 17). The top 5 species based on an average of adjusted values are Loligo squid, fluke, monkfish, cod, and winter flounder, respectively. Most species have lost real value over the years, although haddock and Illex squid posted gains.

b. Fishing regions

The NSNE region, which encompasses the northern Mid-Atlantic Bight, has consistently harvested the most fish by weight (Table 18). The second most productive region for the bottom trawl fishery was more variable, shifting between GB and the GOM. In terms of calculated value (Table 18), the ranking was consistent from 1996–2000, with NSNE first, followed by the GOM, GB, and the MA. Since 2001, the calculated values for the GOM, GB, and NSNE have been similar with each taking a turn at the top valued region.

As the fishing regions are based on ecosystem regions, one would expect differences in the species harvested by fishing region. All the top species were harvested in all fishing regions (Table 19) except for SSNE; however, the importance of individual species to the regional total harvests differed. For the GOM the top 5 species harvested were other groundfish, cod, plaice, monkfish, and winter flounder (Table 19). For GB the top 5 were silver hake, yellowtail flounder, cod, monkfish, and haddock (Table 19). For NSNE the top 5 were Loligo squid, Illex squid, skates, silver hake, and mackerel, and for the MA the top 5 were croaker, Illex squid, fluke, mackerel, and Loligo squid (Table 19). The overlap in the top 5 between adjacent regions may be partly explained by statistical areas that include more than one eco-region, but there are is also seasonal movement by species. This becomes more apparent when landings by species are examined by region and month (Appendix B). Examination at the month-region-species level illustrates shifts between species that allow harvest levels to be maintained throughout the year.

c. Port of landing

The top states in terms of landings are those closest to the GOM and the NSNE fishing regions (Table 20), where effort is concentrated. Massachusetts was usually the top state for bottom trawl landings, with the exception of 1998 and 2004. In terms of inflation adjusted calculated value, Massachusetts was consistently the top state by a significant margin (Table 20). Over the 9-year period, the average adjusted value indicates Massachusetts was the top state, with almost twice the value of second place Rhode Island, followed by Maine as third, and New York and New Jersey nearly tied as fourth.

iii. Vessel characteristics

Most bottom trawl vessels only report using OTF gear, and dependence on this single gear increased between 1996 and 2004 (Appendix C). In 1996, of the 944 vessels reporting use OTF gear, 623 (66%) used only that gear type. By 2004, 588 of the 732 vessels (80%) reporting use of OTF used only that gear type. Only OTF and OTO VTR records are used in the calculation of the information in this report.

a. Fleet

Most average measures of effort and landings per vessel have remained steady across the 9 years, with small differences rendered insignificant by large CVs. While the number of vessels has steadily declined, the average number of trips increased from 42 per vessel per year to 45 (Table 21); however, the CV remained around 100% throughout the period. While the CVs for average days absent per year and days fished were below 100%, the small declines are insignificant (Table 21). All weight-based measures – including average landings per year (live weight), landings per day absent, and landings per day fished – increased (Table 21), although much of the gain occurred in the last year. Both nominal and inflation adjusted values increased over the years, despite drops in 2002–03 (Table 21). The real value per vessel (revenue, base year 2000) increased from $184,728 in 1996 to $222,160 in 2004 (Table 21), a 20% increase.

b. Ton class

A possible reason for the large CVs for the averages per vessel, discussed above, is a high degree of heterogeneity within the bottom trawl fishery. To address this heterogeneity, vessels may be categorized by a number of categories including size characteristics, depending on the type of analysis. The only “standard” set of categories is that for GRT; the ton class (TC) categories for vessels are: TC1 (1–4 GRT); TC2 (5–50 GRT); TC3 (51–150); TC4 (151–500 GRT); and TC5 (>500 GRT). The majority of bottom trawl vessels were of TC2 and TC3 (Table 22), with small numbers in TC4 and very few in TC1. The number of vessels within each class declined over the 9 years, although not equally.

Within TC categories, vessel size measures exhibited a high degree of homogeneity (Table 23). The average VHP had the highest overall level of variation within a TC; however the means were relatively stable over the years. Average vessel length was very stable within a TC, and had CVs below 20% for all groups except TC1, which had a slightly more erratic pattern. As with length, the average GRT within each class was very stable, both in terms of the mean and CV; the CV was larger for TC2 than TC3, which had more vessels.

Despite the number of vessels in TC2 being slightly smaller than TC3, the TC2 group has consistently taken more total trips than TC3 vessels (Table 24). The trips by TC2 vessels are substantially shorter than TC3 trips as the total number of days absent by TC3 vessels is nearly double that of the TC2 group. Additionally, the total days fished by the TC3 group are almost 3 times greater than that for the TC2 group. The small TC4 group has a higher number of total days absent and total days fished relative to the number of trips than any of the other size groups. The percentage of time spent fishing is greater for the larger TC groups. For example, in 2004 TC2 vessels spent 23% of days absent fishing (=4,950 days fished/21,054 days absent) while TC4 vessels spent 39% of the time fishing (4,546 days fished/11,748 days absent).

Differences in the proportion of time spent fishing may, in part, explain some of the differences in the distribution of harvest and revenues between the ton classes (Table 25). Vessels in TC3 landed 4–6 times the weight of fish as did the similar number of vessels in TC2, while the smaller group of vessels in TC4 vessels landed 3–5 times as much fish as the TC2 vessels. The pattern for nominal and inflation adjusted value (i.e., revenue) is similar (Table 25).

Differences in the species composition of landings by vessels in the TC categories (Table 26) may partly explain differences in both effort and landings. The TC4 vessels had higher average annual landings of squid (Illex and Loligo), silver hake, and mackerel, relative to other species, than did other TC groups. For TC4 vessels the 4 species accounted for 57% (=7,504 + 6,682 + 4,606 + 3,759/39,778) of average kept live weight landings. For TC3 vessels, Loligo squid and silver hake are also important, but other top species include skates, other groundfish, monkfish, fluke, and yellowtail flounder, while the top species for TC2 vessels include skates, cod, plaice, fluke, monkfish, Loligo squid, and winter flounder. The smallest vessels, TC1, are almost fully dependent on groundfish, including cod, witch flounder, plaice, yellowtail flounder, monkfish, and other groundfish.

iv. Gear characteristics

Two gear characteristics are available in the VTR records: codend mesh size, and gear size as measured by footrope length. Mesh size is determined by target species and fishing method, which in turn are limited by regulations. Minimum mesh size restrictions have been in place for most species fished by the bottom trawl fishery for the entire 9-year period. While numerous exemptions exist for specific times, areas, and target species, one would anticipate that the self-reported data (i.e., VTRs) would closely follow the regulated sizes for the associated main species. In contrast, gear size in terms of footrope length for trawl nets for the most part is not regulated but is limited by fishing conditions and vessel characteristics.

a. Mesh size (codend)

The mean codend mesh (inches) increased over the years (Figure 17A); however, the total increase has been a little over 1/2 inch since the smallest average in 1998. There is a greater difference between months in average mesh size (Figure 17B), perhaps illustrating shifts in targeted species. There are two peaks in mesh size; generally mesh size was large in the winter and early spring, and decreased in size from June through October (Figure 17B).

There is almost no seasonality present in codend mesh size for a given species (Figure 18), though differences are evident between species. Many of the northeast demersal species have mesh sizes of about 6 inches including haddock, yellowtail flounder, American plaice, winter flounder, witch flounder, cod (Figure 18A), monkfish, and skate (Figure 18B). Scup and fluke, which have distributions from Rhode Island and south, have average mesh sizes of roughly 4.5 and 5.5, respectively (Figure 18B). Spiny dogfish, whose distribution spans the Northeast and MA had a mesh size between 4–5 inches. Squid, herring, mackerel, and butterfish, which are more pelagic and are often caught together, had mesh sizes between 2–3 inches. Silver hake, which often feed on herring and young mackerel (Bigelow and Schroeder 1953), also had an average mesh size of about 3 inches.

When examined by fishing region, the GOM and GB used larger size mesh than NSNE and MA (Table 27). The SSNE region, which is primarily offshore, generally had the smallest mesh size. The general pattern is a decrease in size from north to south, which generally corresponds to the spatial pattern seen by species. All fishing regions except SSNE exhibited an increase in mesh size over the 9 years.

While mean mesh size appears to get smaller as the vessel size increases (Table 28), the differences are small and the CV is large enough to render the sizes indistinguishable. The relatively smaller CV in the TC1 group may suggest a more uniform targeting of species than in the TC4 group, which has a much larger CV.

b. Gear size (footrope length [18])

The average footrope length for the bottom trawl fleet was about 84 feet from 1996– 1999; in 2000 there was a sharp increase to almost 88 feet followed by a steady decline to 85 feet in 2004 (Figure 19A). Seasonality was evident (Figure 19B), with larger footrope lengths in the first quarter, which drop sharply from March to the low in May, and followed by a steady increase in size until December. There are some differences in mean gear size between species (Figure 20). Compared to other species, gear size was smaller for trips that caught winter flounder, cod, yellowtail flounder (Figure 20A), fluke (Figure 20B), skate, dogfish, and Atlantic herring (Figure 20C). Trips that caught haddock (Figure 20A), Illex squid (Figure 20B), and monkfish (Figure 20C) tended to have larger gear. For most species, seasonal variation was limited (Figure 20). Seasonality was evident for witch flounder, American plaice (Figure 20A), scup, butterfish, both squid species (Figure 20B), and monkfish (Figure 20C).

Gear size varied by fishing region, with the largest gear used in GB, followed by GOM, NSNE, and MA (Table 29). Within each region the CV remained around 30% over the years. The GOM and GB both exhibited a steady increase in the gear size over the 9 years except for 2004, when the average size used decreased in GB region. The NSNE and MA regions had declines in gear size over the years, with the difference larger for the MA. The SSNE region exhibited no clear pattern for gear size, although this may be influenced by the limited number of trips in the fishing region.

The average size of gear (in feet of footrope) was larger for larger vessels (Table 30), and appears to have increased over the 9-year period for all categories.

v. Environmental and Habitat Characteristics

Fish and marine mammals have particular habitat requirements which may depend in part on environmental variables such as SST, water/bottom depth and bottom slope.

a. Sea Surface Temperature

The annual mean SST for bottom trawl trips varied considerably over the 9-year period from a low of about 13°C in 1996, 1997, and 1998, to a peak of 14°C in 2000 (Figure 21A). This could be the result of inter-annual variability, but may also reflect changes in the location of effort over time, which in turn is related to the species harvested. As one would anticipate, seasonal differences in water temperature were evident (Figure 21B); the monthly mean SST for trips was lowest in the first quarter, increased steadily until August, and then declined to the end of the year. This shift in SST occurred despite shifts in effort between the seasons (Figure 22). There was a greater concentration of effort in the south during the winter (Figure 22A), while during the spring (Figure 22B) and summer (Figure 22C) effort shifted north, and then began to shift back to the south in the fall (Figure 22D).

The mean monthly SST did not differ greatly between species (Figure 23), with patterns that closely reflect the overall seasonal pattern. During the winter and spring, trips that caught predominantly silver hake had a warmer temperature profile than the rest of the Northeast multispecies group (Figure 23A). The mean temperature signature for Illex squid (Figure 23B) and Atlantic croaker (Figure 23C) was warmer than almost all other species throughout the year; however, the series for croaker was discontinuous due to the limited number of main species trips. The Mid-Atlantic FMP species (Figure 23B) had warmer temperatures than most of the Northeast multispecies group (Figure 23A). Dogfish, herring, and monkfish (Figure 23C) all had similar temperature distributions, though dogfish had a slightly warmer signature than the other two species during the winter and spring. The temperature distribution for the “other finfish” category (Figure 23D) was warmer in the spring due to several species captured at the southern edge of the MA.

The effort within different SST groups has shifted over the years (Table 31), which in part explains the annual changes noted above. Between 1996 and 2004, the days fished in waters with a SST between 0–5°C declined from 9,010 to 3,984, a 56% decline. As effort overall was declining, other SST groups also saw declines in days fished; however, they were not as great as that for the 0–5°C group. The result was a shift in the distribution of effort to warmer waters (Table 31), which was more apparent in 2001 and 2002.

The number of hauls within each SST groups also declined and shifted (Table 32), with the 0–5°C group shifting from 24% of the total in 1996 to 14% in 2004. The shift from the lowest SST group to higher temperature groups is more pronounced when measured by hauls than by days fished, suggesting that tow time may have increased in the 0–5°C group, shortened in other SST groups, or both.

b. Bottom Depth

The distribution of effort in days fished relative to bottom depth indicates that the majority of effort is between 0 and 100 meters, but there was a shift over the 9-year period (Table 33). In 1996 42% of the days fished were in the 0–50 meter depth category, but by 2004 this had dropped to 33%. The majority of the effort shifted to the 50–100 meter category, which increased from 21% to 29% of effort; there was also a small increase in effort in the 200–2000 meter category.

The pattern of change was very similar for effort measured by number of hauls (Table 34), although hauls were more heavily concentrated in the 0–50 meter depth category than were days fished. In 1996, 51% of the hauls were in waters 0–50 meter deep. By 2004, the share had declined to 43%; the decline was matched by an increase in the 50–100 meter depth category. When considered by bottom depth categories, shallow waters (100 meters or less) had a higher percentage of hauls than days fished, suggesting that tow times were longer for trips in deeper waters.

c. Bottom Slope

The sea floor has relatively little bottom slope over the continental shelf and significantly more slope along the shelf break. The slope decreases again on the continental slope, though it generally is not as flat as on the continental shelf. Effort, in days fished, for the bottom trawl fishery overall and by species is concentrated on the continental shelf where bottom slope is slight. Almost 90% of days fished have been concentrated in the 0–0.5° slope category for the 9-year period (Table 35), and the pattern is similar for effort measured by hauls (Table 36).

II. Mid-water Trawl Fishery (OTM and PTM)

The mid-water trawl fishery consists of both single and pair mid-water trawl vessels. The fishery is smaller and its harvest less diverse than the bottom trawl fishery. All trip reports where the gear was listed as OTM or PTM were included in the analysis. This included vessels that also fished in other fisheries, including the bottom trawl; however, only the records relating to mid-water trawl activity are included.

The number of vessels participating in the mid-water trawl fishery declined from 40 in 1996 to 33 in 2004 (Table 37), an 18% decline. The decline in the number of trips was small, from 1,247 to 1,231 (Table 37), a 1% decline.

i. Effort characteristics

Bycatch rates are multiplied by effort to calculate total bycatch; thus, it is important to understand the distribution of effort relative factors correlated with bycatch rates. For the Atlantic trawl bycatch estimates, bycatch rates are expanded by days fished. “Days fished” is an encompassing measure of effort that accounts for tow time and hauls and, to some degree, days absent. The focus of this section is the general spatial and temporal distribution of effort, in terms of days fished and its components. Effort is also considered later when fishing and vessel characteristics are examined in other sections, to provide additional information that could be helpful in the assessment of impacts of proposed TRP mitigation measures.

a. Days fished

Over the entire 9-year period of 1996–2004, effort in terms of days fished for the mid-water trawl fishery was focused on the shelf south of New England, on the northeastern edge of GB, and in the western GOM (Figure 24). The total number of days fished by the mid-water trawl fishery declined from 484 in 1996 to 302 in 2004 (Table 38), a 38% decline. The decline has not been equal across fishing regions; in fact GB had an increase in days fished from 27 to 40 (Table 38), a 48% increase. However, GB remains a distant third in terms of days fished for the mid-water fishery. In 2001, the GOM became the most important region in terms of days fished, replacing the NSNE. The MA region, which accounted for about 20% of effort in 1996 (=94/484), declined to the point where by 2004 only 7% (=21/302) of days fished were in the region.

The mean days fished per trip for the fishery decreased after the late 1990s (Figure 25), although the true difference was only a little over 0.1 day fished or a little over 2 hours. This decline extended to all fishing regions (Table 39). Overall declines from 1996 to 2004 mask fairly steady days fished per trip for the top 3 regions, GOM, GB, and NSNE, since 2001. In 2004, trips to GB spent the most time fishing at 0.51 days per trip, while those in the GOM spent the least time fishing at 0.19 days per trip.

The amount of effort in days fished declined for all main species trips, except for mackerel, which increased from 35 to 71 from 1996 to 2004 (Table 40), a 106% increase. The mid-water fishery was a more diverse fishery in the 1990s when several different main species were evident. However, by 2004 herring and mackerel trips accounted for 93% (=70%+23%) of all days fished in the fishery (Table 41).

The mean number of days fished per trip varied with month (Figure 26). Fishing effort per trip peaked in March–April and September–October, and reaching lows during January and June. When considered by main species and month (Figure 27), the mean days fished per trip are fairly constant through out the year for Atlantic herring. Atlantic mackerel, however, shows a discontinuous series with a sharp increase in May.

The spatial distribution of effort, in days fished, for mid-water herring and mackerel are distinct, except for a small overlap south of Rhode Island (Figure 28). Herring was fished in the GOM, northern edge of GB and south of Rhode Island (Figure 28A), while mackerel was fished from south of New England down through the mid-Atlantic (Figure 28B).

b. Tow time and number of hauls

The examination of the haul and tow time variables was limited as they were used to calculate the days-fished variable, which was examined in detail above. The total number of hauls in the mid-water fishery decreased from 4,162 in 1996 to 2,632 in 2004 (Table 42), a 37% decline, which closely mirrors the decline in days fished. The distribution of hauls between main species trips (Table 43) shows increased concentration in the herring and mackerel fisheries. The distribution of hauls between fisheries is almost identical to that of days fished, suggesting tow times may not differ between herring or mackerel trips.

c. Days absent

Days absent is calculated from the VTR as the number of days between departure and landing, rounded up to the nearest whole day. In the mid-water trawl fishery, the total number of days absent declined from 3,534 in 1996 to 3,350 in 2004 (Table 44), a 5% decline. The decline in days absent was substantially less than that for days fished, suggesting less time was spent fishing in 2004 than in 1996. All fishing regions experienced declines in days absent except the GB region, which showed an increase (Table 44); however, in 2004 the GB region remained a region with limited use by mid-water trawls, with only 328 days absent of the total 3,350 or about 10% of the total.

There appear to be seasonal differences in the length of mid-water trawl trips, with shorter trips in the summer and early fall (Table 45); however, the high CV makes it unlikely that these differences are statistically significant. There appears to less variation in the mean value in August–September and November–December, when trip length is near the lower end of its range.

ii. Harvest characteristics

a. Species

Between 1996 and 2004, nearly 83% of mid-water trawl trips were classified as Atlantic herring main species trips (Table 46). The species making up the majority of the trip’s catch by live weight was considered the “main species.” Atlantic mackerel is the main species on almost 10% of the trips. Other species that are occasionally determined to be the main species on a trip include Atlantic croaker, bluefish, Illex squid, Loligo squid, and scup, although this is largely an artifact of data from 1996–2000. Landings have been quite volatile but have grown over the 9 years (Table 47), and herring has accounted from the majority of landings for the entire period. However, the dominance of herring began to erode in 2002, and by 2004 herring accounted for only 71,446 of the 124,493 metric tons (live weight) landed by the fleet, or 57% of the total. Mackerel landings accounted for an increasing share of the growing landings for the mid-water trawl fishery.

The price for mackerel was higher than that for herring throughout the 9-year period both in nominal (Table 48) and inflation-adjusted prices (Table 49), although the difference narrowed from 2001 onward. The nominal value of the mid-water trawl fishery showed a steady increase from almost $11 million in 1996 to almost $25 million in 2004 (Table 50), a 133% increase. In inflation-adjusted terms (2000=100), the value of the fishery doubled from a little over $11 million to a little over $22 million (Table 51). During this time the value of other species dropped to under $200,000, while the value of mackerel increased to nearly equal that of herring.

b. Fishing regions

In terms of fishing regions, the share of landings from GB and MA has grown (Table 52); however, the GOM and NSNE continue to be the major source of landings. Despite having similar landings in the last few years of the time series, the NSNE region provided a slightly higher share of calculated value than the GOM (Table 52). In nominal and real terms, the combined value of landings from GB and the MA have accounted for approximately one-third of value throughout the time series, with the GOM and NSNE accounting for approximately one-third each.

The species composition of landings by fishing region (Table 53) illustrates a north–south pattern. In the GOM and GB herring dominate. In the NSNE region, landings of herring and mackerel were similar. In the MA, mackerel is the dominant species harvested. Most of the Illex and Loligo squid was landed in NSNE, while the croaker was predominately landed in the MA. As well as spatial differences, there are seasonal differences in landings when examined by fishing region and month (Appendix D).

c. Port of landing

The port location where mid-water trawl vessels landed their harvest was closely linked to the location of the harvest, with the top states being those closest to the GOM and NSNE fishing regions, which includes Maine, Massachusetts, and Rhode Island (Table 54). On average over the 9-year period, Massachusetts was the top state in terms of landings and value for the mid-water trawl fishery (Table 54); however, the top state within a give year was variable.

iii. Vessel characteristics

The majority of mid-water trawl vessels participate in other trawl fisheries, in particular the bottom trawl fishery (Appendix C); however, by 2004 the majority of pair-trawl vessels used only that gear type. The number and share of vessels dependent only on mid-water trawl has increased, even while the total number of vessels in the mid-water trawl fishery declined from 40 to 33 from 1996 to 2004 (Table 55), a 18% decline.

a. Fleet

Most average measures of effort and landings per vessel show a clear difference between the periods from 1996–2000 and from 2001–2004 (Table 55). In 2001 the average number of trips per year taken by a mid-water trawl vessel increased from 29 to 36, while days absent increased from 65 to 94 (Table 55). Landings per vessel, per day absent, and per days fished all increased sharply, as did nominal and inflation-adjusted value per vessel. This suggests a sharp change in the fishery, and in 2001 the Atlantic Herring FMP came into effect, requiring VTRs from all vessels permitted for herring. The only measure that did not appear to change was the total days fished per year per vessel, which remained around 9–10 days from all periods after 1996 (Table 55).

b. Ton class

Large CVs for these values may make the measures more apparent than significant. High CVs can be a sign of a heterogeneous fleet. Categorizing vessels may provide for more uniform strata with smaller CV values. One standard set of categories is for GRT (see “Ton Class” under the “Bottom Trawl Fishery” section, above). The distribution of vessels between the size categories has changed between 1996 and 2004 (Table 56). There were no vessels in TC1 or TC5. The number of vessels in TC2 remained relatively constant, while those in TC3 declined and those in TC4 increased (Table 56). By 2004, 20 of the 33 vessels were in TC4, which accounted for over 60% of the mid-water fleet.

Vessel size measures have relatively low CV values within a TC (Table 57). While there is some variation between years, average VHP has increased for all 3 TCs represented. Average vessel length and average GRT have remained steady over the years within TC2 and TC3 (Table 57); however, both measures have increased for TC4. The CV for average length and GRT have been small and steady within TC2 and TC3, but growing in TC4.

Looking at total effort measures by ton class shows strong differences between the different groups of vessels (Table 58). Since 2001, vessels in TC2 have operated as day boats where the number of trips is equal to the number of days absent. There has also been a sharp contraction for TC3 vessels in the number of trips and days absent, which in part reflects a decrease in the number of vessels. Vessels in TC4 showed a decline in trips and days absent in the 1997–2000 period, but both measures increased since 2001. In terms of actual fishing effort, TC4 vessels dominate the fishery. As mentioned above, in 2004 vessels in TC4 made up 61% of the mid-water vessels; however, in 2004 they accounted for 1,089 of the 1,231 trips (88%), 3,078 of 3,332 days absent (92%), and 278 of 302 days fished (92%).

A similar pattern emerges when total landings and value are examined by ton class (Table 59); the TC4 vessels contributing the majority of landings and calculated value. Average landings by ton class illustrate those vessels in the largest class, TC4, predominately target herring while those in TC3 land a mix of herring and mackerel (Table 60).

iv. Gear characteristics

There are two gear characteristics available in the VTR records, codend mesh size, and gear size measured by footrope length.

a. Mesh size (codend)

The mean codend mesh decreased by approximately 0.5 inches from a peak in 1998 (Figure 29A), although size was steady around 1.5 inches since 2001. Monthly mean mesh size was highest during the January–April period, was steady around 1.5 inches from May–October, then increased through November and December (Figure 29B). The reason for this seasonal pattern is evident when mesh size was examined by species (Figure 30). Atlantic mackerel was fished with larger mesh, and only occurred in the first 5 months of the year. Herring is fished throughout the year and uses a smaller mesh size.

Average mesh size by fishing region shows differences that are the result of species targeted (Table 61). In the GOM and GB, where herring is the principal species, mesh size is the smallest and has been steady over the years. In NSNE, where both herring and mackerel are landed the average mesh size is slightly larger and more variable, illustrating the different mix of landings over the years. In the MA, where mackerel is the principal species landed, average mesh size is considerably larger than in the other fishing regions.

Mean codend mesh size by TC illustrates the focus of the different groups of vessels between herring and mackerel (Table 62). Vessels in TC4 have the smallest average mesh size, and it was shown earlier that landings by these vessels are dominated by herring. The average mesh size for vessels in TC3 is larger than that for TC4 vessels, reflecting the greater role of mackerel in the landings for these vessels.

b. Gear size (footrope length)

Gear size, measured as footrope length in feet, shows a sharp increase in 2000 (Figure 31A), after which it has been relatively stable around 200 feet. Footrope length showed a seasonal pattern (Figure 31B), with the shortest lengths in August, slowing increasing to January, sharply decreasing in February, and then slowly increasing to June. Atlantic mackerel trips were distinguished from Atlantic herring trips by their larger footrope length and their limited fishing season from January–May (Figure 32).

There have been changes in average gear size used in the fishing regions (Table 63), and much variation between years. In 1996, the shortest footrope was used in the MA and the longest in the NSNE region; by 2004, the shortest footrope was used in the GOM, and the longest in the MA.

The average footrope length is longer for larger vessels (Table 64), and increased over time for TC2 and TC4 vessels. Gear size appears to have decreased for TC3 vessels, although this group also has the highest CV values.

v. Environmental and Habitat Characteristics

Fish and marine mammals have particular habitat requirements which may depend in part on environmental variables such as SST, water/bottom depth and bottom slope.

a. Sea Surface Temperature

As with the bottom trawl fishery, there was considerable variability in the mean SST by year for the mid-water trawl fishery (Figure 33A). This similarity suggests that the variation by year was likely due to inter-annual variability of SST rather than changes in fishing practices. However, the mid-water trawl annual SST presents a cooler profile than the bottom trawl fishery, with a low of ~11°C in 1999 and a high of 12°C in 2000 (Figure 33A). The monthly average SST for the mid-water trawl harvest followed a pattern of declining temperatures from January–March, followed by a period of warming until August, then a decline for the remainder of the year (Figure 33B). Effort within the fishery shifted north and south with the seasons, with most effort taking occurring in the SSNE and in the MA in the first quarter (Figure 34A), and in the GOM in other quarters (Figure 34B-D).

When examined by main species, the average SST followed a pattern similar to that for the entire fishery, with Atlantic mackerel showing a slightly warmer winter–spring temperature signature than herring (Figure 35). This is a result of the more southerly harvest areas for mackerel, identified above.

Between 1996 and 2004, effort in days fished decreased in waters less than 20°C (Table 65), and was sustained in the colder waters below 10°C. This resulted in a slight shift in the distribution of effort to colder waters (Table 65). The relationship between days fished and hauls is highly correlated for SST groups, resulting in a similar pattern when effort is examined by number of hauls (Table 66).

b. Bottom depth and slope

Bottom depth and bottom slope are closely related, as illustrated earlier (Figure 6). In the mid-water trawl fishery, Atlantic herring is caught primarily south of Rhode Island, along the northern edges of GB, and in the GOM. These areas are relatively shallow, with a flat to moderate bottom slope. Atlantic mackerel is primarily caught on the mid-Atlantic shelf, extending north to east of Long Island, especially in the area of the Hudson Canyon. The majority of these areas are less than 100 meters deep and have little bottom slope, except for the localized topography of the Hudson Canyon, where the walls are too close together to be well represented in the bottom slope data source used.

An examination of effort by bottom depth categories shows that the majority of effort in days fished is in waters less than 100 meters (Table 67), with about 60% of days fished. This has changed little over the 9-year period. When effort is measured by number of hauls, we see that effort is slightly more concentrated in the less than 100 meter categories (Table 68); this suggests shorter hauls in the more shallow waters.

Days fished by bottom slope categories shows that a vast majority of effort is concentrated in areas where slope is less than 0.5° (Table 69); this concentration has been growing. By 2004, 92% of days fished were in the category of slope less than 0.5° (Table 69), up from 69% in 1996. A very similar patter is evident when effort is measured based on the number of hauls (Table 70).


Between 1996 and 2004, effort in the bottom trawl fishery decreased between 15% and 40%, depending on the unit of measured. Despite these declines, the total harvest kept (live weight) was relatively stable until 2004, when there was a substantial increase in 2004. The nominal value of the fishery increased by about 9% over this time; however, the inflation-adjusted value decreased by 5% and the inflation-adjusted value per vessel increased by about 20%. Relative shifts in effort meant that GB became a more important fishing region, and increased shares of effort were directed to yellowtail flounder, fluke, haddock, and Loligo squid trips compared to other species. These shifts were accompanied by a larger share of effort in deeper and warmer waters.

The mid-water trawl fishery data shows a sharp division in 2001, the year the Atlantic Herring FMP came into effect. The fishery focuses on herring and mackerel, which are spatially distinct for most of the year. Both landings and value for the fishery have been increasing, although there is a high degree of inter-annual variability. In 2004, the value of the fishery was almost evenly split between herring and mackerel, although landings for herring were significantly higher. In contrast to the bottom trawl fishery, the GB region has become less important to the mid-water trawl fishery as the SSNE and MA regions have increased in importance both in landings and value. Large vessels (TC4) increasingly dominate the fishery, measured in effort, landings, and value.

In general, the results of this report will not surprise those who are in close contact with the bottom and mid-water trawl fishery. However, the members of the ATGTRT represent a broad cross-section of stakeholders including fishermen, environmental organizations, state agencies, and federal government representatives. Thus a disparate level of knowledge of the fishery may exist; even fishermen who are aware of changes within their fishery or region may be unaware of the broader picture. This report attempted to create a base level of information for all participants in the process. This report is presented at a high level of aggregation, and thus is limited in its ability to determine the potential impact from proposed actions. However, it does illustrate changes that are occurring within the fisheries that may help to identify priorities for actions. For example, shifts in effort to different regions, main species or depths could result in increases in bycatch which may need to be addressed proactively.

The data also highlight the importance of understanding changes in FMPs. Effort patterns within the bottom trawl fishery (Figure 8) illustrate the impact of year-round closures (Figure 3); while the closed areas themselves are clear of effort, effort is highly concentrated along their edges. The expansion of areas closed to bottom trawl gear with the creation of essential fish habitat (EFH) areas could have potential impacts on the location and intensity of bottom trawl effort, and potential bycatch. Similarly, the mid-water trawl fishery showed a high concentration of effort (Figure 24) in herring management area 1A (Figure 4); in the future mid-water trawls will be excluded from this area. The relocation of this effort could have implications for bycatch. This report cannot comment on how effort will shift with management actions, but highlights the need to support existing and future modeling efforts to capture these shifts.

The data presented in this paper was drawn primarily from VTRs. Because these trip reports are mandatory, this data source provides a theoretically complete survey of all bottom and mid-water trawl effort. However, data accuracy may not always be reliable. This may be due to errors on trip reports, data entry errors, or purposeful misrepresentation of trip data by the fishers themselves. We have taken steps to ensure data accuracy, though the true accuracy is unknown. Checks of gear size, mesh size, tow time, and number of hauls against reasonable expected values has shown that in most cases 98% of the data fell within expected limits (Table 1). Comparisons of recent VTR data against observer data have also shown a good match (Rago et al. 2005). Within the Northeast Fisheries Science Center and Regional Office, steps have been taken to improve the accuracy of both incoming and existing VTR data. Despite these efforts and generally good comparisons with other data sources, some data fields such as latitude and longitude were missing values, especially in earlier years, and given the size of the database some errors are nearly undetectable and unknowable. Given the available data, we believe that though each trip record may not be exactly correct, the VTR data are representative as a whole. Upcoming efforts to compare VTR data to dealer data (Magnusson in prep) and a more detailed comparison with observer data (Orphanides in prep) should shed light on this issue and provide a better framework for assessing the VTR data.

Assuming that the VTR data are generally representative, it is still possible to misrepresent the data. Much of the fisheries information was summarized according to what was termed the “main species” of a trip. This meant that if a trip landed more weight of a particular species than any other, then that trip was assigned to that main species and the trip variables were analyzed as being part of the group assigned to that particular species. Using this method each trip is only used once when summarizing variables by species. It is unclear whether this grouping method best represents trip variables for species that did not often make up the majority of the catch. An alternate method would be to summarize trips by all species caught on a particular trip; if a trip caught mostly haddock, but also contained cod and monkfish, then the trip would be included when summarizing variables for haddock, cod, and monkfish. Alternatively, the trips could be categorized based on the species with the highest value within a trip, based on the economic principle of revenue maximization. The question of interest however, is which version of “main species” most closely matches captains’ determination of “target species”; this requires additional comparisons with observer data.

It is also important to point out that the spatial and temporal distributions of fish catch shown in this paper may not necessarily reflect true fish distributions. Numerous regulations limit the times and areas where fishers can fish, particularly in the Northeast, and these regulations have changed over time. Regulations also affect the gear characteristics used in certain locations, times of the year, and target species. For example, the Northeast Multispecies FMP, perhaps the most important FMP for the bottom trawl fishery, has resulted in several changes in permitted mesh sizes, with spatial, temporal, and species exemptions. Limited access areas, days at sea, and trip limits have all been added to the FMP and changed over time. Several seasonal and year-round closed areas have been created, removed, and reconfigured resulting in a patchwork of year-round, seasonal, and rolling closed areas. These restrictions make it very difficult to determine the true species distribution from catch data. The ongoing changes in the regulations also make it difficult to compare catch data across years. In addition, these regulations have played a role in the decline of effort in trawl fisheries, which also makes it is difficult to compare across years. Despite all these changes, fishers still have managed to maintain or increase harvest and value.

The aim of this paper was to provide summary statistics on the bottom and mid-water trawl, so it lacks in-depth comparative statistical analysis. Given the size of the fisheries and the wealth of data available, we felt that this paper would best serve as a basis for subsequent analysis in other papers. Similarly, the biological and ecological reasons for the temporal, spatial, and environmental distributions of the fish were not discussed in this paper. Again, we felt that we had to limit the scope of this paper to provide baseline information for the take reduction team process and have left these topics for other possible subsequent papers.


First, we would like to thank the fishers for their honest reporting of their fishing effort, without which this study would not have been possible. We would also like to thank all who collected, edited, and archived these mammoth datasets. We would also like to thank D. Palka for her guidance and reviews during the preparation of this paper, K. Bisack for guidance in stratifying the data, and R. Merrick, F. Serchuk, and E. Thunberg for their helpful comments. All errors and omissions remain those of the authors.


Bigelow HB, Schroeder WC.1953.Fishes of the Gulf of Maine.US Fish Wildl Serv, Fish Bull. 53; 577 p.

Gabriel WL. 1992. Persistence of demersal fish assemblages between Cape Hatteras and Nova Scotia, Northwest Atlantic. J Northw Atlant Fish Sci. 14:29-46.

Magnusson GM.[in prep]. A comparison of VTR and commercial fisheries landing data for the Atlantic trawl fishery. NOAA NMFS Northeast Fisheries Science Center, 166 Water St, Woods Hole MA 02543.

New England Fisheries Management Council. 1999. Final Atlantic Herring Fishery Management Plan Incorporating the Environmental Impact Statement and Regulatory Impact Review (Including the Regulatory Flexibility Analysis) - Vol I. Saugus (MA): NEFMC; p 376.

New England Fisheries Management Council. 2006. Final Amendment 1 to the Fishery Management Plan (FMP) for Atlantic Herring Including a Final Supplemental Environmental Impact Statement (FEIS) and Initial Regulatory Flexibility Analysis (IRFA) - Vol I. Newburyport (MA): NEFMC; p. 713.

Orphanides, CD. [in prep.] A comparison of observer and VTR records for the Atlantic trawl fishery. NOAA NMFS Northeast Fisheries Science Center, 166 Water St, Woods Hole MA 02543.

Palka DL.[in prep].Mid-water trawl cetacean bycatch estimate.NOAA/NMFS/Northeast Fisheries Science Center, 166 Water St., Woods Hole MA 02543.

Rago PJ, Wigley SE, Fogarty MJ. 2005. NEFSC bycatch estimation methodology: allocation, precision, and accuracy. Northeast Fish Sci Cent Ref Doc. 05-09; p. 44.

Rossman M. [in prep]. Bottom trawl fishery cetacean bycatch estimates.NOAA NMFS Northeast Fisheries Science Center, 166 Water St, Woods Hole MA 02543.

Rountree BP. 1997. Individual Vessel Behavior in the Northeast Otter Trawl Fleet during 1982-92.NOAA Tech Memo NMFS NE 113; p. 50.

Stevenson D, Chiarella L, Stephan D, Reid R, Wilhelm K, McCarthy J, Pentony M. 2004. Characterization of the fishing practices and marine benthic ecosystems of the Northeast U.S. Shelf, and an evaluation of the potential effects of fishing on essential fish habitat.NOAA Tech Memo NMFS NE 181; p. 79.


[1] Sources for these documents include the New England Fishery Management Council ( and the Mid-Atlantic Fishery Management Council (

[2] Rountree (1997) summarized vessel level data for the otter trawl fishery for 1982-1992, prior to the availability of VTR data. Thus, her results are based on a combination of the commercial fisheries data and port agent reports. In 1992, a total of 1,012 vessels used otter trawl gear.

[3] Stevenson et al. (2004) used VTRs to map total days at sea (days absent) from 1995-2001 by 10-minute square. For the bottom otter trawl (fish) gear type, activity was concentrated in the western part of the Gulf of Maine, along the northwest edge of Georges Bank, east of Long Island, and along the shelf edge south of Rhode Island and Massachusetts.

[4] Target species is defined by the vessel captain prior to a fishing trip and is recorded in NMFS observer records. This differs from “main species” which is based on some measure of actual landings (e.g., greatest landed weight).

[5] The bottom trawl analysis focused on those variables that were in both the observer and VTR or permit records or could be determined from data within those records (Rossman [in prep])

[6] Variables with an asterisk (*) are available in the VTR or permit databases or can be derived from information within those databases (e.g. SST is derived using date, latitude and longitude from VTR records).

[7] For example, regulated minimum mesh sizes may create a truncated distribution when examined by species, and increases (decreases) in minimum mesh size may shift the mean value up (down).

[8] If a vessel was using both a monkfish DAS and a multispecies DAS the mesh size must conform to the multispecies minimum mesh size.

[9] The lobster fishery is an exception.

[10] FMP group was determined by summing landings for all species within a FMP (Appendix A). If landings for a FMP group of species was greater than or equal to 50% of landings, the trips was assigned to that FMP.

[11] From an economic perspective, it would be more reasonable to assume that the main species was that with the largest contribution to trip revenues. Historically, however, the bycatch analysis does not incorporate economic values, focusing instead on physical features such as weight. For consistency the same method was used in this study. This differs from target species, which is recorded in the observer data based on a vessel captain’s intentions at the beginning of the trip, but is not available for VTR records.

[12] Additional information on the climatology data source can be found at:

[13] Additional information on the satellite data sources can be found at the following links: AVHRR: MODIS Terra and Aqua, see products 162 and 184 ( and; GOES, see product 190 (

[14] Bathymetry data was acquired from ETOPO Global 2’ Elevations CD, available from the National Geophysical Data Center.

[15] There are also concerns about possible errors in the VTR depth data with regard to units.

[16] Source:; accessed September 1, 2005. This is the same deflator used in the annual Fisheries of the United States publication by NMFS.

[17] Information on the distribution of the individual species may be found in the Essential Fish Habitat Source Documents, which are part of the NOAA Technical Memorandum NMFS NE series. These documents are available through the NEFSC’s “EFH Source Documents: Life History and Habitat Characteristics” web page (; accessed July 20, 2007).

[18] Footrope length is also referred to as “sweep length” and is an indication of the width of the trawl opening.

ASMFC = Atlantic States Marine Fisheries Council
ATGTRT = Atlantic Trawl Gear Take Reduction Team
CA = closed area
CFDBS = Commercial Fisheries Dealer Database
CV = coefficient of variation
DAS = days at sea
EFH = essential fish habitat
FMA = fishery management area
FMP = fishery management plan
GB = Georges Bank
GOM = Gulf of Maine
GRT = gross registered tons
MA = Mid-Atlantic
MAFMC = Mid-Atlantic Fishery Management Council
NEFMC = New England Fishery Management Council
NERO = Northeast Regional Office
NMFS = National Marine Fisheries Service
NOAA = National Oceanic and Atmospheric Administration
NSNE = North Southern New England
OTF = otter trawl fish
OTM = otter trawl mid-water
OTO = otter trawl other
PTM = paired mid-water trawl
RMA = regulated mesh area
SNE = Southern New England
SSNE = South Southern New England
TAC = total allowable catch
TRC = take reduction plan
VHP = vessel horsepower
VTR = vessel trip report
NMFS Search
Link Disclaimer
Privacy Policy
(File Modified Jul. 01 2016)

This page has had 2 visits today, 12 visits this week, 57 visits this month, 160 visits this year