REDNET - A Network to Redevelop a Sustainable Redfish (Sebastes fasciatus) Trawl Fishery in the Gulf of Maine

DRAFT FINAL REPORT

Date: 16 May 2016

Written by: Michael Pol Massachusetts Division of Marine Fisheries

On behalf of co-Network Coordinators

Pingguo He UMass Dartmouth School for Marine Science and Technology

and

Sally Sherman (beginning June 2013) Kohl Kanwit (until June 2013) Maine Department of Marine Resources

Funded by: Northeast Cooperative Research Partners Program Northeast Fisheries Science Center National Marine Fisheries Service

NOAA Contract Number EA133F10CN0323

1

REDNET Draft Final Report 2

I. Abstract

The goal of the REDNET project was to devise strategies and means to sustainably harvest the abundant Acadian redfish resource in the Gulf of Maine through a network approach, including fishing enterprises, gear manufacturers, researchers, social and economic experts and managers. The network defined the research pathway for conceiving, developing and implementing research, development and outreach strategies to sustainably access to the redfish resource under the current sector and ACL management regime. The vision was a comprehensive, integrated, well-planned project where the intent of every step is toward an environmentally and economically sustainable redfish fishery.

A network of forty or more participants from all facets of fishing was developed, along with strategies of consensus and openness, to facilitate and to improve the effectiveness of REDNET activities and research. Major information gathering, sharing and decision making occurred at six major meetings during the life of the REDNET contract, and in-between via email and telephone calls. The network was largely highly flexible, expanding and contracting as needs warranted. The completion of a substantial amount of work indicates that the network was an overall success.

Structured and monitored exploratory fishing using five commercial fishing vessels in every two to three months between the end of May 2011 and January 2012 revealed low amounts of bycatch, with redfish comprising 94.9% of all catch. The vessels used their own commercial groundfish trawls, but a relatively small mesh codend (114 mm or 4.5” mesh size, double twine diamond). Each vessel fished 4-5 days with a total of 85 tows, ranging from 15 to 20 tows per trip (per vessel). Fishing was carried out on traditional redfish grounds throughout the central portion of the Gulf of Maine at depths ranging from 77 fm to 138 fm.

The amount of redfish caught per trip ranged from about 35,000 to 67,000 lbs and the catch per tow ranged from 0 to 13,655 lb, averaging 2,766 lb. Total kept and landed redfish were 221,957 lbs, or 95.5% of all redfish caught. Pollock (Pollachius virens) was the main landed bycatch species (7,307 lb), with 13 other species landed in smaller amounts with only spiny dogfish exceeding an SBRM of 0.01.

Results indicated that harvesting redfish with a 4.5 in diamond mesh could be conducted without substantial catch of undersized redfish and other commercially important species. The size of redfish caught may be further reduced through codend mesh size adjustment or using size-sorting devices.

Measurement of codend mesh size selectivity of three sizes of mesh opening (4.5, 5.5 and 6.5 in double 5 mm twine diamond) was conducted using a trouser trawl on a commercial fishing vessel. Fishing off Provincetown, Massachusetts, 56 tows were completed in March and April 2013, catching over 42,000 kg of redfish and about 6,000 kg of other species. Adequate length frequencies of redfish and pollock (Pollachius virens) were collected to produce selectivity models; only redfish results are reported here. Neither species has been the subject of a trawl selectivity study in the Northeast US before. REDNET Draft Final Report 3

Robust models for the mean L50s and selection ranges, and confidence intervals, were developed for all three tested codends, incorporating both within and between haul variability. All measures of model validity were positive. These models are fully adequate to provide guidance to managers and fishermen on size retention of redfish and appropriate codend mesh size.

L50 and selection ranges were determined for 4.5 in (L50: 22.3 cm (8.8 in); SR: 4.5 cm), 5.5 in (L50: 29.2 cm (11.5 in); SR: 4.4 cm), and 6.5 in (L50: 33.6 cm (13.2 in); SR: 5.1 cm) codends. Simulation of fishing of the three tested codends on the observed population indicated that substantial escape of redfish through codend meshes occurs (48-94%), suggesting that investigation of escape of redfish is warranted to support a sustainable fishery. The observed population also indicates that inadequate numbers of larger redfish may be available to support a higher-priced market.

Choice of an appropriate mesh size, special access program, marketing strategy, and other factors are linked to mortality of small redfish. Small redfish that are excluded or escape during fishing and die as a result are unobserved and unquantified and this unaccounting can impact stock assessments. Small redfish brought to the surface and onto the deck likely die, but can be quantified. In order to weigh the costs and benefits of management choices, more information on location of redfish escapes was sought. Redfish that escape through codend meshes at bottom experience fewer sources of injury and mortality. To provide more information on the fate of small redfish, we deployed cameras on redfish codends, with the goal of determining when and where escapes occur. To investigate a means of encouraging escape at bottom, we tested a dual-grid system designed to exclude smaller redfish and allow larger redfish to pass to the codend.

Testing occurred on a commercial redfish trawler northeast of Provincetown, Massachusetts in July and August 2014. A trouser trawl section attached to a redfish net allowed simultaneous comparison of catches in two codends. For escapement, catches of redfish in a 2.5 in codend were compared to catches in a 4.5 in codend to estimate the number of escapes. Video cameras on the top and bottom of the codend were used to count escapes and the depths at which they occurred. For dual-grid testing, a 4.5 inch codend was attached to one leg, and the dual-grid system plus a 4.5 inch codend was attached to the other.

Results from escapement trials indicated that a small portion of total escapes were captured on video, perhaps due to sediment or other factors. Due to camera failures, the sample size was small. Most observed escapement occurred during haulback (on average, over 70%), a relatively brief portion of the haul. Results are preliminary, and video and other data require further analysis.

Dual grid testing showed no reduction or length effect due to the 40 mm grid spacing. The 50 mm grid spacing showed a reduction at all lengths, suggesting that a grid system could be tuned to reduce only small fish. Adjusting and deploying the grid system presented some challenges.

REDNET Draft Final Report 4

Further refinement of a dual grid system would be necessary for commercial application, but preliminary results indicate that a large proportion of redfish escapes likely occurs during haulback, suggesting that a grid system could reduce injury and delayed mortality of small redfish.

Results from the processing and marketing component revealed large price variation when landings of redfish were low, and higher variation when landings were higher. Processing capacity did not limit the price of redfish. It was suggested from the results of this component that an alternative to reduction of the catch of smaller redfish (7-8 in) is the development of human consumption instead of a bait market for these fish. This type of development would require processor and consumer testing beyond the duration of REDNET. Based on our results, most restaurant managers do not serve redfish but prefer locally landed (Massachusetts) fish. This finding implies that consistent redfish landings may encourage more restaurant owners to serve locally landed fresh redfish. Marketing programs, including eco-labeling, to raise the price of redfish would be challenged by redfish imports from Canada, which sell for less than half the price of redfish landed in New England. The consensus of marketing experts in the group was that continuity and stability (i.e. confidence) of supply would create demand. Markets are known (and kept as trade secrets) by marketers and could be developed. Sectors or other industry groups could coordinate activities to keep markets supplied consistently without over- or under-supply. However, a networked approach may not facilitate marketing due to competitive concerns and proprietary information.

An outreach plan was developed, but only partly implemented due in part to the completion of experiments late in the project and the ending of the contract.

Tradeoffs from possible management actions were described based on output from the multiple components of REDNET.

REDNET Draft Final Report 5

II. Goals, Objectives, and Value

An Acadian redfish (Sebastes fasciatus – hereafter referred to as “redfish”) Annual Catch Limit (ACL) of 6,848 mt available to sectors beginning May 1, 2010 remained largely untapped due to the lack of development of a sustainable and responsible harvest strategy and trawl gear technology. The implementation of sectors with catch retention rules and accountability measures made targeting redfish difficult. We proposed to establish and use the resources of a redfish network to develop a comprehensive research plan that draws on wide-ranging expertise and to conduct comprehensive research including bycatch assessment, gear testing, processing and marketing, outreach and implementation.

Goal: The goal of the project was to devise strategies and means to sustainably harvest the redfish resource in the Gulf of Maine through a network approach, including fishing enterprises, gear manufacturers, researchers, social and economic experts and managers. The network defined the research pathway for conceiving, developing and implementing research, development and outreach strategies to sustainably access to the redfish resource under the current sector and ACL management regime. The vision was a comprehensive, integrated, well- planned project where the intent of every step is toward an environmentally and economically sustainable redfish fishery.

Objectives: The specific objectives of the project were to: • assemble a network of experts dedicated to developing a redfish fishery through a multi-component approach; • convene strategic planning meetings to identify approaches and means for research, development and outreach related to the project; • establish baseline information on temporal and spatial distribution of redfish and major bycatch species through test fishing targeting redfish using commercial trawlers; • evaluate appropriate codend mesh sizes and shapes for retention of redfish while reducing non-target species and undersized redfish through a codend selectivity study; • develop bycatch reduction methods, devices or fishing strategies through behavioral observation, prototype testing and spatial analysis; • evaluate potential markets and the processing capacity for redfish; • provide outreach and recommendations to members of the industry, sector managers, fishery managers and regulators.

REDNET Draft Final Report 6

The proposed cooperative research project sought to achieve three fishery conservation and management goals:

1. Redirecting fishing effort in the multispecies fishery away from stocks that are overfished (e.g. pollock, yellowtail and winter flounder) to stocks that are considered rebuilt (e.g. redfish); 2. Achieving optimum yield, by increasing commercial landings of redfish through development of a directed fishery under the adaptive management ability of groundfish sectors, and 3. Increasing the economic viability of groundfish sectors by providing access to the ACL of a recovered species and thus generating much-needed revenue for the industry.

Groundfish management in the Northeast made a dramatic shift in May of 2010 from primary input-based controls to primary output based controls. Amendment 16 to the Northeast Multispecies Fishery Management Plan (FMP) established the rules for sector management, as well as catch limits and accountability measures mandated by the reauthorization of the Magnuson-Stevens Act. Framework 44 set the ACLs for the next three years and included an annual allocation of 6,848 mt of redfish in 2010. This allocation was second only to Georges Bank haddock (40,440 mt) and Gulf of Maine cod (7,240) and represented 9% of the total multispecies yield (21% if the extraordinarily high ACL for Georges Bank haddock is removed). The ACLs for several other stocks were much smaller and therefore severely limiting. If multispecies sector management is to be successful, the fleet must be able to catch and land stocks of high abundance like redfish while exercising their ability to avoid limiting species like pollock, winter flounder and yellowtail flounder.

Historically, redfish represented a significant fishery in the region, and the best available estimates assert that the resource can support a larger fishery. The directed redfish fishery began in the 1930’s and total landings rose from 100 mt to a peak of over 117,000 mt in 1951 and then steadily declined. By 1983, the total US landings of redfish were 5,328 mt.; 2008 landings were 1,189 mt (NEFMC 2009). The redfish fishery in the Gulf of Maine was traditionally prosecuted by vessels using otter trawls with relatively small mesh, in the range of 70-80 mm (2.5-3 in). The population of redfish declined, likely as a result of small mesh and overexploitation. In 1977, the minimum mesh size increased from 114 to 130 mm (4.5 and 5 in) and increased again in 1994 to 152 mm (6 in: NEFMC, 1985). Today the minimum mesh size mandated by the Multispecies FMP is 165 mm (6.5 in). These mesh restrictions, combined with low biomass levels between 1980 and 1995, have nearly eliminated the prosecution of a directed redfish fishery in the Northeastern United States, although one of our project partners has continued to target redfish with this large mesh. In recent years the combined restrictions in the multispecies FMP have resulted in the recovered status of the redfish resource. The most recent stock assessment of redfish was completed and reviewed at the 2008 Groundfish Assessment Review Meeting (GARM III). It indicated that redfish is not overfished and overfishing is not occurring (NEFSC, 2008).

REDNET Draft Final Report 7

Redfish (Sebastes spp.) are harvested in directed and non-directed fisheries throughout their global range. In Iceland, the fishery for Sebastes marinus (golden redfish) is prosecuted with bottom trawls and a minimum codend mesh size of 100 mm (4 in) in the directed fishery. A suite of management measures including sorting grids are required in other small mesh fisheries to minimize redfish bycatch and exclusion zones are used for the protection of juveniles, and temporary closures if juvenile catches are high. Sebastes mentella or oceanic redfish is also targeted in Iceland, with pelagic trawls. Oceanic redfish are also called the “deepwater” redfish, inhabiting waters 350-700 m and are considered exclusively pelagic. This behavior makes them targetable by midwater gears, unlike Acadian redfish (Sebastes fasciatus) which are primarily benthic (http://www.fisheries.is, site visited May 5, 2010). Norwegian fishery management has banned the directed fishery for golden redfish since 2003 but allows a bycatch limit of 15% in the mixed trawl fishery which uses 135 mm mesh (5.3 in: http://www.fisheries.no, accessed May 5, 2010). Other countries that operate in the eastern Atlantic and harvest Sebastes spp. include Russia, Germany and Denmark which operate under European Union regulations.

The Canadian redfish fishery in the northwestern Atlantic takes place in deep waters from Nova Scotia to Newfoundland. They primarily manage the resource with quotas in three stock units. The fishery is permitted to use 110 mm (4.3 in) diamond mesh versus the 130 mm (5.1 in) square mesh required in the groundfish fishery. To avoid large catches of small redfish there are permanent closed areas. It is not clear if the small redfish in the closed areas are segregated juveniles or just a smaller subpopulation of adults (Peter Comeau, Department of Fisheries and Oceans, pers. com.). Bycatch is managed with quotas and a percent of the total catch.

We believe redevelopment of a redfish fishery is critical to economic survival of fishermen and to the success of sectors. The problem is defined as follows: sectors were recently implemented through Amendment 16 to the multispecies FMP, and are assigned percent allocations based on the historic catches of their members. Allocations of redfish cannot be achieved under current regulations because the minimum mesh size is too large to effectively retain redfish. At the same time, allocations of other groundfish species are extremely small due to the very low ACLs recommended. These limiting allocations (“choke” species) will affect the fishing behavior of the sectors and potentially shut them down before more abundant allocations are realized. It is critically important to the success of sectors to find a way to allow them to access allocations of healthy stocks while avoiding those which are depressed.

REDNET was formed as a NMFS Cooperative Research Partners (CRP)-funded network of gear researchers, netmakers, fishermen, NMFS regional office and science center members, Council staff, fish processors and others working to reestablish the redfish trawl fishery in the Gulf of Maine.

REDNET sought to achieve three fishery conservation and management goals: • Redirecting fishing effort in the multispecies fishery away from stocks that are overfished to stocks that are considered rebuilt (e.g. redfish). REDNET Draft Final Report 8

• Achieving optimum yield, by increasing commercial landings of redfish through development of a directed fishery under the adaptive management ability of groundfish sectors. • Increasing the economic viability of groundfish sectors by providing access to the ACL of a recovered species and thus generating much‐needed revenue for the industry.

REDNET proceeded through several stages, or components: network building; exploratory fishing; codend selectivity; bycatch reduction; processing and marketing; outreach and implementation. Each of the components after network building was planned independently and collaboratively with a broad consensus of network participants; each was also led by different members of the network

III. Component 1: Network building

REDNET participants were solicited and attracted from a broad range of disciplines, with the intent of gathering as much expertise as possible to choose and to improve the conduct and outcomes of the project. These disciplines included processors and fish dealers, fishermen and fishing representatives, netmakers, fishery regulation promulgators and enforcers, gear and other scientists and including representatives from CRP. All parties interested in the fishery were invited to the REDNET meetings, beginning with the list of participants included in the proposal, and then by word of mouth; the membership of REDNET was open, and changed and adapted based on the needs and interests of the participants and of the project.

The network functioned via an email list maintained by network coordinators, phone calls, smaller informal meetings, and by periodic network meetings, typically to approve plans of work before research was conducted, when a significant decision was needed, to explain results, to resolve significant issues, or other as need arose. Other issues arising from the network meetings that required action included investigating confidentiality of the data, requirements for age structures, and marketing possibilities. Network coordinators typically initiated, planned and usually coordinated these meetings. For each meeting, agendas and meeting minutes/summaries were produced and distributed via email to the network.

Timing and location of meetings were selected to minimize inconvenience and improve participation. Meetings were held on the following dates and locations:

Meeting I: 24 November 2010, Boston Fish Pier, Boston, MA Meeting II: 24 October 2011, Gulf of Maine Research Institute, Portland, Maine Meeting III: 19 March 2012, School for Marine Science and Technology, New Bedford, MA Meeting IV: 27 February 2013, School for Marine Science and Technology, New Bedford, MA Meeting V: 22 August 2013, Annisquam River Marine Fisheries Field Station, Gloucester, MA Meeting VI: 19 December 2013, Urban Forestry Center, Portsmouth, NH

REDNET Draft Final Report 9

Meetings were attended by an average of 15 individuals, with as many as 21 and as few as 6 (Table 1).

Table 1: List of participants and affiliations in REDNET network meetings and date of attendance

Name Organization 11/24/2010 10/22/2011 3/19/2012 2/27/2013 8/23/2013 12/19/2013 Amanda Parks UNH-EcoGastronomy x Brett Alger NMFS NERO Permit Office x x x Carolyn Woodhead NMFS Coop. Research x x Chris Rillihan SMAST x x x x Daniel Georgianna UMASS x x x x x David Chosid MA DMF x x x x Don Frei NMFS x Douglas Christel NMFS NE Regional Office x x x x Steve Eayrs Gulf of Maine Research Institute x Earl Meredith NMFS Cooperative Research x x x x Elizabeth Etrie NESSN x Erik Chapman NH Sea Grant x x Gabby Bradt NH Sea Grant x J. Kohl Kanwit ME DMR x x Jamie Cournane NEFMC x Jennifer Bichrest Purse Line Bait, Inc. x Jerry McCarthy Cape Ann Seafood Exchange x x John Sackton Seafood.com x Jon Knight Superior Trawl x x Joshua Wiersma NEFS XI x Ken LaValley New Hampshire Sea Grant x Luke Szymanski AIS Inc. x x Associated Fisheries of Maine/ Sustainable Maggie Raymond Harvest Sector / Salmon Falls Trawlers, Inc. x x x Mark Grant NERO/NMFS x Mark Lussier Northcoast Seafood x Mark Szymanski MA DMF x x x Michael Pol MA DMF x x x x x x Michael Walsh Integrity Fishing Corporation x x x x x Nathatlie Berthiaume NMFS x Gloucester Seafood Display Auction/Cape Ann Nina Jarvis Seafood Exchange x x x Pingguo He SMAST x x x x x x Rick Usher AIS Inc. x Russell A. Sherman F/V Lady Jane or F/V SCM Inc. x Sally Sherman Maine Dept of Marine Resources x x Steve Cadrin SMAST x Steve Eayrs Gulf of Maine Research Institute x Associated Fisheries of Maine/ Sustainable Terry Alexander Harvest Sector x x Tim Miller NMFS/NEFSC x Tom Nies NEFMC x Tor Bendiksen Reidar's Manufacturing, Inc. x Tory Bramante Atlantic Coast Seafood, Inc. x x x

The meetings were the most visible expression of the functioning of the network, but multiple communication channels were used to build trust, exchange information, arrange smaller meetings, make decisions, and other actions. An email list of 46 individuals was developed. Email and phone communications were the most frequently used modalities.

The network coordinators worked to create and maintain the network as it formed, developing strategies in an ad hoc manner as best they could. They also took leadership and responsibility for completing the proposed work, and fulfilling reporting requirements. As the network functioned, it became clear that openness and honesty were necessary for the successful REDNET Draft Final Report 10

functioning of the network. Unspoken assumptions about each participant’s motives, skills and requirements must be discussed openly for the network to be effective. This level of openness is not quickly or easily achieved, and can be damaged by single bad missteps or actions.

Difficulties and conflicts occasionally arose, although they were usually resolved using consensus and broad agreement. However, the network is not structured to prevent individuals from acting on their own interests, which also occurred during the activities of this program.

MA DMF, as the funding awardee, was responsible for establishing payment to other network participants, receiving invoices, and other logistical and reporting requirements. This arrangement inevitably created an imbalance of power, or the perception of an imbalance.

The relatively small size of this network avoided problems seen in other networks with slow decision-making due to consensus building. Similarly to other networks, extensive participation of CRP personnel in the network required an adjustment by network coordinators, who had previously conducted research projects with limited input into priority-setting and research methods. A perception of an imbalance of power and authority within the network was an inevitable result of this level of participation.

Overall, the network method was demonstrated to be an effective mechanism for incorporating a broad range of wisdom, focusing research on relevant topics, shaping research to be most effective, and sharing information revealed by REDNET, although the latter was mostly in an informal manner. While not a perfect mechanism, the network appeared to foster buy-in and commitment to REDNET, and participants were generous with their time and knowledge.

REDNET Draft Final Report 11

IV. Component 2: Exploratory Fishing

A draft final report for Component 2 was submitted to CRP on 23 January 2013. A version revised to incorporate questions and suggestions in a review by the NEFSC dated August 1, 2014 is attached as an appendix, along with a response to reviewers.

Executive Summary

Acadian redfish (Sebastes fasciatus) is among a few groundfish stocks in the northeast US that have a relatively large Annual Catch Limit (ACL) that have not been fully utilized during the last several years. With reduced ACLs for many important groundfish species, full and sustainable utilization of the redfish resource is thus critical to the economic survival of fishermen and to the success of groundfish sectors. The allocations of redfish could not be fully utilized under current regulations because the minimum codend mesh size is too large to effectively retain redfish. The REDNET project seeks to achieve three fishery conservation and management goals:

I. Redirecting fishing effort in the multispecies fishery away from stocks that are overfished to stocks that are considered rebuilt (e.g. redfish). II. Achieving optimum yield, by increasing commercial landings of redfish through development of a directed fishery under the adaptive management ability of groundfish sectors. III. Increasing the economic viability of groundfish sectors by providing access to the ACL of a recovered species and thus generating much‐needed revenue for the industry.

REDNET includes six components:

Component 1: Network Meetings Component 2: Baseline Catch and Bycatch Evaluation Component 3: Codend Selectivity Component 4: Conservation Engineering and Bycatch Reduction Component 5: Processing/Marketing Component 6: Outreach/Implementation

This report describes work carried out in Component 2 - Baseline Catch and Bycatch Evaluation.

The work was carried out with a structured and monitored exploratory fishing method using five commercial fishing vessels in every two to three months between the end of May 2011 and January 2012. The vessels used their own commercial groundfish trawls, but a relatively small mesh codend (114 mm or 4.5” mesh size, double twine diamond) provided by the project. Each vessel fished 4-5 days with a total of 85 tows, ranging from 15 to 20 tows per trip (per vessel). Fishing was carried out on traditional redfish grounds throughout the central portion of the Gulf REDNET Draft Final Report 12

of Maine at depths ranging from 77 fm to 138 fm. Two technicians (or NOAA-approved observers) were on each of the trips to record operational conditions and to sample catch and bycatch.

The total catch of all species was just over 234,000 lb; with redfish comprising 232,380 lb (94.9% of all catch). The amount of redfish caught per trip ranged from about 35,000 to 67,000 lbs and the catch per tow ranged from 0 to 13,655 lb, averaging 2,766 lb. Total kept and landed redfish were 221,957 lbs, or 95.5% of all redfish caught. Pollock (Pollachius virens) was the main landed bycatch species (7,307 lb), with 13 other species landed in smaller amounts.

Thirty-four species were discarded, totaling almost 41,000 lb (14.9% of total catch), with spiny dogfish (Squalus acanthias) accounting for the majority of discards (>26,000 lbs, or 63.4%). Discards of undersized redfish (MLS: 9” or 23 cm) totaled 10,423 lb or 4.5% of the total redfish catch. Total discards of undersized pollock were 2,744.8 lbs. Total discards were less than 200 lbs per species for other species.

Discards analyzed using the Standard Bycatch Reporting Method revealed that almost all median SBRM ratios below 0.01. Only spiny dogfish had SBRM exceeded 0.01. Discard weights of redfish were higher in summer (July and September) and highest at shallower water depths (between 84 and 107 fm).

The exploratory fishing results indicate that it is possible to harvest redfish using 4.5” diamond mesh without substantial catch of undersized redfish and other commercially important groundfish species. Depth, time-of-year, or some other factors appear to have an impact on the catch of undersized redfish thought we were not able to isolate these factors.

The size of redfish caught may be further reduced through codend mesh size adjustment or using size-sorting devices. A size-sorting grid system may allow smaller redfish to escape at depth which will result in higher survival rates than through codend meshes at the surface. These aspects of research will be carried out in Component 3 and 4 of this project.

See the Appendices for the full revised report.

REDNET Draft Final Report 13

IV. Component 3: Codend Selectivity

A draft final report on this component was submitted to the CRP on 13 December 2013, with subsequent revisions in response to reviews on 28 July 2014 and 10 November 2014. The final version of the executive summary is below, and the full report is attached in the Appendices.

Executive Summary

Acadian redfish (Sebastes fasciatus) is among a few groundfish stocks in the northeast US that have a relatively large Annual Catch Limit (ACL) that have not been fully utilized during the last several years. With reduced ACLs for many important groundfish species, full and sustainable utilization of the redfish resource is thus critical to the economic survival of fishermen and to the success of groundfish sectors. The allocations of redfish could not be fully utilized under current regulations because the minimum codend mesh size is too large to effectively retain redfish. Beginning in 2010, the REDNET project seeks to achieve three fishery conservation and management goals:

Redirecting fishing effort in the multispecies fishery away from stocks that are overfished to stocks that are considered rebuilt (e.g. redfish).

Achieving optimum yield, by increasing commercial landings of redfish through development of a directed fishery under the adaptive management ability of groundfish sectors.

Increasing the economic viability of groundfish sectors by providing access to the ACL of a recovered species and thus generating much‐needed revenue for the industry.

REDNET includes six components: Component 1: Network Meetings Component 2: Baseline Catch and Bycatch Evaluation Component 3: Codend Selectivity Component 4: Conservation Engineering and Bycatch Reduction Component 5: Processing/Marketing Component 6: Outreach/Implementation

This report describes work carried out in Component 3 – Codend Selectivity. The final report for Component 2 has been submitted.

A trouser trawl was used to determine the size selectivity of three sizes of mesh opening (4.5, 5.5 and 6.5 in double 5 mm twine diamond) on a commercial fishing vessel. Fishing off Provincetown, Massachusetts, 56 tows were completed in March and April 2013, catching over 42,000 kg of redfish and about 6,000 kg of other species. Adequate length frequencies of redfish and pollock (Pollachius virens) were collected to produce selectivity models; only redfish results are reported here. Neither species has been the subject of a trawl selectivity study in the Northeast US before. REDNET Draft Final Report 14

Robust models for the mean L50s and selection ranges, and confidence intervals, were developed for all three tested codends, incorporating both within and between haul variability. All measures of model validity were positive. These models are fully adequate to provide guidance to managers and fishermen on size retention of redfish and appropriate codend mesh size.

L50 and selection ranges were determined for 4.5 in (L50: 22.3 cm (8.8 in); SR: 4.5 cm), 5.5 in (L50: 29.2 cm (11.5 in); SR: 4.4 cm), and 6.5 in (L50: 33.6 cm (13.2 in); SR: 5.1 cm) codends. Simulation of fishing of the three tested codends on the observed population indicated that substantial escape of redfish through codend meshes occurs (48-94%), suggesting that investigation of escape of redfish is warranted to support a sustainable fishery. The observed population also indicates that inadequate numbers of larger redfish may be available to support a higher-priced market.

REDNET Draft Final Report 15

V. Component 4: Escape Mortality/Bycatch Reduction

Executive Summary Escape location of small redfish was identified as a vital area of research in previous components. Choice of an appropriate mesh size, special access program, marketing strategy, and other factors are linked to mortality of small redfish. Small redfish that are excluded or escape during fishing and die as a result are unobserved and unquantified and this unaccounting can impact stock assessments. Small redfish brought to the surface and onto the deck likely die, but can be quantified. In order to weigh the costs and benefits of management choices, more information on location of redfish escapes was sought. Redfish that escape through codend meshes at bottom experience fewer sources of injury and mortality. To provide more information on the fate of small redfish, we deployed cameras on redfish codends, with the goal of determining when and where escapes occur. To investigate a means of encouraging escape at bottom, we tested a dual-grid system designed to increase escape of smaller redfish during towing and allow larger redfish to pass to the codend.

Testing occurred on a commercial redfish trawler northeast of Provincetown, Massachusetts in July and August 2014. A trouser trawl section attached to a redfish net allowed simultaneous comparison of catches in two codends. For escapement, catches of redfish in a 2.5 in codend were compared to catches in a 4.5 in codend to estimate the number of escapes. Video cameras on the top and bottom of the codend were used to count escapes and the depths at which they occurred. For dual-grid testing, a 4.5 inch codend was attached to one leg, and the dual-grid system plus a 4.5 inch codend was attached to the other.

Results from escapement trials indicated that a small portion of total escapes were captured on video, perhaps due to sediment or other factors. Due to camera failures, the sample size was small. Most observed escapement occurred during haulback (on average, over 70%), a relatively brief portion of the haul. Results are preliminary, and video and other data require further analysis.

Dual grid testing showed no reduction or length effect due to the 40 mm grid spacing. The 50 mm grid spacing showed a reduction at all lengths, suggesting that a grid system could be tuned to reduce only small fish. Adjusting and deploying the grid system presented some challenges.

Further refinement of a dual grid system would be necessary for commercial application.

Introduction Results from Components 2 (bycatch) and 3 (selectivity) of REDNET indicated that capture of small redfish is a major concern in the fishery, especially when a smaller mesh codend (e.g. 4.5 inches) is used. Component 5 (marketing) results showed that small redfish are also an issue in markets.

REDNET Draft Final Report 16

The selectivity experiment from Component 3 provided clear information on the probability of retention of different lengths of redfish for a wide range of mesh sizes. It further demonstrated that large numbers of redfish escaped through codend meshes; even for the smallest mesh (4.5 inches), 49.7% of redfish entering the codend would exit through the codend meshes.

These escapees experience some level of mortality, which can be difficult to quantify and is not accounted for in the stock assessment. It is known that the timing of this escape has a great impact on escapee survival. If escape occurs at the surface or in the water column while hauling, mortality rates due to predation, decompression, thermal shock, displacement from habitat, and other factors are likely to be far greater than if escape occurs at bottom.

It has been observed by fishermen and scientists that substantial escape of redfish occurs at the surface. Redfish experience barotrauma, which increases opportunity for predation and exposure induced mortality. These fish are wasted, providing neither additional growth and reproduction, nor revenue through landings. Knowing where and how much escape occurs at different locations will illustrate the magnitude of unaccounted mortality in the fishery. This information is vital to sustainable exploitation and management of this resource.

Depending on the location of escape, gear measures can be incorporated to promote escape at the bottom prior to mesh passage and prior to the start of haul. Use of sorting grids for other Sebastes documented an 85% grid escape rate. Use of a sorting grid, which likely has lower physical damage to fish than passage through codend meshes, can greatly reduce wastage and unaccounted mortality from escape through the codend, at the same time reduce the unmarketable small fish.

The following research plan was developed under direction of the REDNET network, as endorsed at the December 13, 2014 meeting.

Objective The research in this component includes two segments: 1) video observation to identify the time and the water depth that fish escapes from the codend, and 2) test of a grid device to encourage fish escape at depth and to reduce small fish landed.

• To monitor and identify redfish escape from the codend at various stages of capture – at depth, during hauling or at surface • To release small redfish at fishing depth to reduce escapee mortality, to reduce discards, and improve value of landed fish

Materials and Methods The F/V Guardian (80 ft LOA; 425 hp), a commercial groundfish trawler with recent experience targeting redfish, was chosen to conduct the research. The participating vessel provided a REDNET Draft Final Report 17

balloon trawl front end (ground gear, wings, and net mouth) to be attached to a “trouser trawl” section. The headline of the trawl was 33.4 m in length with 100 plastic floats 20.3 cm in diameter. The footrope was 42.5 m in length and attached to a rockhopper groundgear. The front end of the net had 152 mm diamond mesh openings constructed of 4.0 mm diameter braided twine. The fishing circle was 190 meshes across the bottom panel and 240 meshes across the top.

The trouser section of the trawl was also constructed of 152 mm diamond mesh, 3.6 mm diameter braided twine. It was designed with a 47.5 meshes deep common “mixing area” that was then separated uniformly into two lateral equal circumference legs (130 meshes across the bottom; 161 meshes across the top). One leg of the trouser trawl was lengthened by 25 meshes of double 4 mm 165 mm mesh to avoid contact or inhibition of escape by one codend on the other (Figure 1).

To monitor and to identify redfish escapement, a non-selective control codend, constructed of double 4 mm diamond shaped twine with a nominal mesh size of 2.5 in (65 mm), 125 meshes long and 125.5 meshes around was attached to one leg. Test codends of diamond double 5 mm 4.5 in (114 mm) or 6.5 in (165 mm) were attached to the other leg. Mesh openings in codends were measured prior to and after the experiment using an ICES OMEGA mesh gauge and associated protocols (Fonteyne, 2005). The number of meshes for each test codend was adjusted so that the same diameter and overall length were maintained for all codends. The side of the test and control codends was switched to avoid possible side-based effects. Video for escapement was collected with two low-cost action cameras (GoPro Hero2 and Hero3, San Mateo, CA), placed in deepwater housings and attached to custom-built aluminum frames. Frames were placed on top of and below the codend, near the attachment of the codend to the trouser trawl section.

Fifteen tows were conducted to estimate escapes. All used lights except the final three, which were not lit to provide an indication of the effect of lights on escape. The third of these three had neither lights nor cameras, only camera frames. Numbers of redfish within each codend were estimated by raising the number at length from length-frequency samples of each codend based on the weight of the subsample, and the total weight of redfish in each codend. The number of escapes from the larger mesh codend was estimated as the difference in the raised numbers between the smaller mesh codend and the larger mesh codend.

Videos were overlain with depth recorded from the temperature depth sensor attached to the codend and with time of day and date. Videos were viewed in a darkened room on a MacBook Pro 13 in laptop computer (2.6 Ghz processor, Intel HD Graphics 4000, 1024 MB card, OSX) using VLC media player (version 2.2.1) at 2x. Each video was viewed in its entirety. When necessary to accurately count or to identify species, the video was slowed. For each redfish escape, the time of day, tape time, depth, and escape count were recorded.

To test a release mechanism at depth, a dual grid system was designed, built, and attached to one leg of the trouser section. The system was constructed of three hinged steel frames that REDNET Draft Final Report 18

permitted attachment of high density polyethylene (HDPE) grids in the first two frames. The first two frames were 91 cm x 146 cm and 91 X 110 cm. The first panel had three rows each with 16 openings (40 mm grid) or 14 openings (50 mm grids) in each row, 45.7 mm long with 12 mm bars in between and with 25 mm horizontal bars between rows. The second panel had two rows of openings of the same numbers and sizes. The third panel was a steel frame covered with small square mesh PE netting to slow the flow of water and to prevent escape once fish reached the codend. The grid system was installed on one leg of the trouser section, with 4.5 mm codends on both ends of the trouser section (Figure 1).

Two chains connected the top of the first frame to the top of the third frame. At the hinge between the first and second grid, a 30 cm by 95 cm frame was added to maintain an opening under the grid system for continuing to the codend. The entire system was fitted into a two seam section of 4 in PE webbing. The grid system clogged after the first deployment due to clogging by dogfish. To avoid a repeat of this occurrence, the webbing in the grid system was replaced with 2.5 in green double diamond meshes to prevent gilling.

Figure 1: Diagram of trouser trawl section in redfish trawl net, with details of the dual-grid (Sort-X) system inset.

After the first deployment of this revised grid system, clogging once again occurred. In reaction, a slit was made along the bottom midline and more 2.5 in webbing was added to allow easier passage under the grids. During testing, chain lengths were also adjusted based on video observations.

Thirteen tows were used to test the 40 mm bar spacing in the dual grid system; eleven were valid. A catastrophic failure of the system terminated the first trip; tests of the 40 mm bar spacing occurred at the beginning and end of the second trip. Fourteen valid tows were used to test the 50 mm bar spacing in the middle of the second trip.

REDNET Draft Final Report 19

Effectiveness of the two sizes of grid was measured by length-based effects using a generalized linear mixed model and Wald’s t-test, with tow as a random effect, an offset based on the subsampling ratio, and length as a fixed effect. The link function was binomial.

For both escapement and dual-grid testing, tow locations were selected by the captain’s knowledge, echo sounder signals (including bottom topography), and a goal of a mix of redfish sizes. Tows were only made in daylight hours following the practice of the fishery. Tow durations varied based on the captain’s assessment of the volume of fish in the net and fishing ground conditions, and were consistent with commercial practice. Very large catches were avoided due to catch processing delays that might reduce the number of tows and affect quality of fish retained and survival of fish escaped.

The length of warp used was set by the captain based on the water depth and the bottom topography of the tow track. The range of tow speed was within normal operational conditions for the species and was mainly influenced by tidal conditions, as was typical in commercial operations.

Net geometry was measured using a trawl monitoring system (Notus Electronics, St. John’s, Newfoundland) with sensors on both doors, the trawl’s wing ends, just behind the headrope, and on the 64-mm codend. The sensors were set to provide bottom temperature, door spread, door heel (angle of the door to the right or left of the direction of travel), wing spread and to indicate when the control codend was full. In addition, these sensors can provide distance from the sensor to the hydrophone. Bottom temperature was also recorded with previously calibrated TidBit temperature recorders (Onset Computers, Inc., Pocasset, Massachusetts). Depth was collected using two temperature depth recorders, one on the codend and one on a trawl door.

Codends were hauled on deck one at a time, with the codend attached to the shorter “leg” of the trouser section hauled and emptied first. Catches from the experimental and control codends were deposited in separate areas on deck, and processed separately. The total catch of redfish per tow was determined to 0.1 kg with subsampling when there was a large amount of catch and, on some tows, legal and sublegal catch amounts were quantified. Lengths (measured as midline length, MLL) of a random subsample of more than 100 redfish (if possible) from each codend from each tow were measured to the nearest cm. For length- frequency (LF) analysis, counts at each length were multiplied by the subsample weight divided into the total weight. Other organisms were also identified, and weighed to the nearest 0.1 kg. Weights were directly measured or quantitatively determined; for example, by basket counts.

Results (as of reporting date) Tows were conducted generally east and northeast of Provincetown, Massachusetts, USA (Figure 1). Forty-three tows were completed in two trips carried out between 12-14 July, and 29 July and 3 August 2014. Overall, 15 tows were completed with cameras to view codend escapes, with 9 looking at 4.5 in mesh (three on the port side, all without lights) and 6 looking REDNET Draft Final Report 20

at 6.5 in mesh (2 on the port side); 14 tows (11 valid) with the 40 mm bar spacing in the modified Sort-X grid; 15 tows with the 50 mm bar spacing in the modified Sort-X grid. For safety and handling reasons, the modified dual-grid was always on the starboard side.

Tow duration ranged from 0.2 to 2.6 h with a mean tow duration of 1.1 h. Mean depth fished was 105 fm, ranging from 89-128 fm; mean towing speed was 3.3 knots (range: 2.9-3.7 knts); mean warp length (wire out) was 259 fm (range: 250-275 fm). Mean wave height experienced was 1.4 ft with a maximum of 5 ft.

Figure 2: Locations of Component 4 hauls

Escapement The total catch of all species during tows examining escape with cameras was just over 16,400 kg, with redfish comprising 10,252 kg or 62% of all catch. Pollock was the main bycatch species (2849 kg), with 12 other species with catches greater than 10 kg total (Table 2).

REDNET Draft Final Report 21

Table 2: Catch in escapement tows by codend

Codend mesh size (in) Species 2.5 4.5 6.5 Redfish Sebastes sp 7015.76 3014.238 222 Pollock Pollachius virens 1690.9 1006 152 Spiny dogfish Squalus acanthias 1495.99 1170.5 79.7 Haddock Melanogrammus aeglefinus 110 79.5 13.6 Atlantic cod Gadus morhua 69.1 18.5 9.7 Red hake Urophycis chuss 63.7 11.2 Monkfish Lophius americanus 21.1 8.8 16.1 Whiting Merluccius bilinearis 19 1.6 River herring Alosa 12.64 0.3 1.2 Barndoor skate Raja laevis 9.3 American lobster Homarus americanus 9.1 2.7 5 American plaice Hippoglossoides platessoides 8 0.9 1 White hake Urophycis tenuis 7.8 15.6 11.3 Atlantic herring Clupea harengus 7.5 2 0.5

Estimates of the number of redfish escaping from the larger mesh codends per tow averaged 44.6% (4.5 in; range: -0.53-75.0) and 77.6% (6.5 in; range: 53.0-95.3) (Table 3). Tows with cameras and lights had escape rates from the 4.5 in codend averaging 65.6% (range: 28.3-96.6); two tows with cameras and no lights had escape rates of 50.17% and 61.8%. The one tow where only camera frames were attached, curiously, had an escape rate of near 0, indicating the same numbers of fish were caught in both the 2.5 in and 4.5 in codend.

Of 28 total possible sets of recordings (14 filmed tows with two cameras), 24 were collected - on four hauls, the bottom camera failed for unknown reasons. Fifteen recordings were analyzed at the time of reporting, all from the 4.5 inch codend. One haul (8) was invalidated due to a fouling of the gear; on another (5), video collection was interrupted for an unknown reason. Suspended sediment, likely from the doors and mouth of the trawl net, occluded each of the videos at times. Of the four analyzed tows without camera or gear failure, the average detection rate (number observed on video/number estimated from catches) was 22.2%, with a range of 9.8-55%. Total detected escapes from top v. bottom codend for all analyzed video, were approximately equal (910 v. 913). The average percentage of all observed escapes that occurred during haulback was 71.5% (range: 46.8-100). No escapes were observed during setting. A cursory review of video from 6.5 in mesh did not show any dramatic differences from 4.5 in codend results.

Codends were observed to be in poor position to selectively sort fish when first deployed, with folds and little tension in the webbing. Tension increased as fish entered the codend. Codends were also observed to pulse rhythmically, and the contents of the codend were mixed continuously. Escapement appeared to only occur during the slack phase of the pulse. REDNET Draft Final Report 22

Escapement was particularly observed to occur when haulback was initiated; pulsing continued during hauling. No very large escapement events were observed, at any depth.

Table 3: Summary of escapement video analysis. Yellow cells mark complete camera failure; the orange bar indicates a gear malfunction. Hauling is the percent of observed escapes that occurred once the codend began ascending to the surface.

Catch Count Escapes Video Count Small Large Detection Haul Mesh Mesh No. % Top Bottom Sum Rate Hauling Mesh (in) Comment 1 2616 1877 739 28.25 162 162 21.9% 54.9% 4.5 Bottom camera failed 2 2717 1365 1352 49.76 310 440 750 55.5% 63.7% 4.5 3 5727 1712 4015 70.11 165 203 368 9.2% 92.1% 4.5 4 4919 1231 3688 74.97 267 261 528 14.3% 46.8% 4.5 5 535 363 172 32.15 0 2 2 1.2% 4.5 Partial camera failure 6 118 77 41 34.75 0 4 4 9.8% 100.0% 4.5 7 88 3 85 96.59 9 9 10.6% 6.5 Bottom camera failed 8 1706 80 1626 95.31 6.5 Codend twisted 9 1984 116 1868 94.15 6.5 Not analyzed 10 243 67 176 72.43 6.5 Not analyzed 11 220 59 161 73.18 6.5 Not analyzed; bottom camera failed 12 217 102 115 53.00 6.5 Not analyzed 13 791 302 489 61.82 0 0.0% 4.5 No lights; bottom camera failed 14 588 293 295 50.17 0 0 0.0% 4.5 No lights 15 563 566 -3 -0.53 4.5 Camera frames only; no video

Dual-grid Testing The total catch of all species during all tows testing the grid system was just over 52,000 kg, with redfish comprising 11,945 kg or 68% of all catch. Spiny dogfish was the main bycatch species (7242 kg), with 15 other species with catches greater than 10 kg total (Table 4).

A GLMM analysis using length as a fixed effect indicated that the 40 mm grid had no impact on amount or length of redfish (p>0.05; Figure 4). The 50 mm grid spacing significantly reduced all sizes of redfish caught (p<0.05, linear model; Figure 5).

Preliminary analysis of video monitoring indicated that redfish escaped through the grid, and appeared unharmed (Figure 6). Handling of the grid system resulted in bending and breaking of system elements during normal handling.

REDNET Draft Final Report 23

Figure 3: The dual-grid system being retrieved, with control codend also visible in orange (lower right). A camera frame is also in view.

Table 4: Catch (kg) by species during dual-grid testing. Sort40 and Sort50 indicate the two tested bar spacings.

Species Control (4.5 in) Sort40 Sort50 Redfish Sebastes sp 26,199.12 5,689.92 3,946.20 Spiny dogfish Squalus acanthias 5,064.54 1,296.20 881.71 Pollock Pollachius virens 4,265.85 1,529.55 511.70 Haddock Melanogrammus aeglefinus 606.34 126.60 198.74 American plaice Hippoglossoides platessoides 359.40 70.20 58.20 White hake Urophycis tenuis 238.10 46.60 34.00 Atlantic cod Gadus morhua 168.30 41.20 39.80 Monkfish Lophius americanus 159.80 31.10 64.40 Herring Clupeidae 88.50 1.00 0.80 Atlantic herring Clupea harengus 86.40 4.10 0.30 American lobster Homarus americanus 52.10 6.80 14.50 Whiting Merluccius bilinearis 40.60 4.50 4.30 Red hake Urophycis chuss 22.20 4.00 7.90 Sea raven Hemitripterus americanus 13.40 3.80 3.50 Thorny skate Raja radiata 11.30 1.10 0.80 Longfin squid Doryteuthis pealeii 10.70 0.60 0.80 Winter skate Raja ocellata 8.50 4.00 2.00 REDNET Draft Final Report 24

Figure 4: Proportion at length for Acadian redfish in dual-grid system with 40 mm bar spacings. Vertical red line is the MLS. The horizontal green line denotes equal proportions in both codends. Bubbles are proportional to numbers of fish measured. REDNET Draft Final Report 25

Figure 5: Proportion-at-length for Acadian redfish in dual grid system with 50 mm bar spacing. Vertical red line is the MLS. The horizontal green line denotes equal proportions in both codends. Bubbles are proportional to numbers of fish measured.

Figure 6: Redfish escaping from the dual-grid system. REDNET Draft Final Report 26

Discussion Our preliminary results from the escapement analysis suggest that a large fraction of the escapement occurs during hauling, a relatively brief event during the haul. Once fish are removed from the bottom, they are exposed to additional sources of mortality, including increased exposure to predation, barotrauma, and temperature changes. The video was able to capture only a small fraction of escapes, at least partly but not entirely due to obscuring by suspended sediment. Therefore, it is not possible to make conclusive statements about the timing of escape. Further investigation of these data should include analysis of the remaining tows, and a finer examination of escape depth and timing. It is tedious, but it may be necessary to quantify the amount of video time lost due to visibility.

The use of video was problematic as many opportunities were lost due to camera failure. Low cost action cameras provide a reasonably priced tool for capturing underwater video, but occasional failure may be a hidden cost of inexpensive cameras.

The use of depth recorders on the doors and on the net provides an opportunity to more accurately determine timing of escape, as well as time on bottom. These data should be further analyzed. Net geometry and bottom temperature were also collected, and should be analyzed in conjunction with depth.

The 50 mm grid system increased escapement of redfish at depth, possibly protecting fish from additional mortality if they had remained in the codend and were discarded at the surface or escaped from the codend during haulback. Although the grid itself may impart some damage to redfish, it is likely less to cause mortality than the combination of fatigue and mesh contact a redfish escaping from the codend experiences.

The 40 mm grid did not improve escapement for any size of redfish, while the 50 mm reduced all sizes. An ideal grid size, perhaps 45 mm, might provide a size-based selection where larger, landable fish are not reduced, and therefore no economic loss is experienced. A 45 mm grid would require additional testing.

To date, a dual grid system is the only functioning way to improve escape of smaller target fish at depth beyond the codend meshes. It is an unanswered question whether the challenges of rigging and deploying a dual grid system is worth the benefits arising from it. Dual grid systems are in commercial use in Scandinavian countries; however, fishing vessels of that region are larger and also do not use net reels, instead winching the entire trawl net including the sorting grids onto the deck in a “trawl alley.” A commercialized system for our region requires durability and greater simplicity, and should allow for verification of correct deployment. Analysis of remaining video of the grid system in action may provide additional insights into improvement of the design. Additionally, flume tank testing of the system, which could occur at full-scale, would provide an excellent means for adjusting or improving the system.

REDNET Draft Final Report 27

In summary, preliminary results from this experiment indicate that a large proportion of redfish escapes likely occurs during haulback, suggesting that a grid system could reduce injury and delayed mortality of small redfish.

Acknowledgements Primary thanks are due once again to the F/V Guardian captains and crew: Capts. Billy and Mike Walsh, Séan Farren, Richard Walsh, Carlos Acosta, Jake Almeida, and Alex Koeberle, for their wisdom, patience, friendship, and participation. We appreciate assistance of our staff biologists who collected data onboard from DMF: David Chosid, and SMAST: Chris Rillahan, and to Mark Szymanski of DMF for contract and other support. We acknowledge the extensive involvement of the REDNET network. Funding was provided by the NOAA Fisheries Northeast Cooperative Research Partners Program.

REDNET Draft Final Report 28

V. Component 5 (Processing and Marketing)

Executive Summary The market for redfish prefers large fish (4 oz fillet or 1lb fish); the current legal size (7 inches) is much smaller than the fillet market is demanding. Landings and retail price data showed some increase in pricing and landings in recent years. Demand is uncertain due to lack of knowledge of redfish, as revealed by a telephone survey (report attached). Pricing appears to be overwhelmingly dependent on the quantity of groundfish landed, not on Canadian redfish landings (as only 600MT of redfish comes into the US from Canada). Canadian prices for redfish are also very low.

It was agreed that REDNET could or should not embark on a marketing plan for redfish although some network members argued that “brand recognition” could create a market for fresh local fish. However, these types of marketing plans are expensive and difficult, and would compete with landings from Canada. It was also recognized that REDNET funds might not be acceptable for this type of activity.

The consensus of marketing experts in the group was that continuity and stability (i.e. confidence) of supply would create demand. Markets are known (and kept as trade secrets) by marketers and could be developed. Sectors or other industry groups could coordinate activities to keep markets supplied consistently without over- or under-supply.

Most redfish landings are sorted into large and small and the smaller fish are sold as lobster bait for $0.50-0.60. The larger fish got to auction and sold for $ 0.95 for an 8” fish. Price for bait is driven by loss of herring as bait so 70% goes to the bait market and 30% to the fillet market.

Processing appears not to be a limiting factor.

Future research is best directed toward finding better markets than bait for the smaller redfish landed.

REDNET Draft Final Report 29

Final Report on processing/marketing component of RedNet

By Dan Georgianna, School for Marine Science and Technology, UMass Dartmouth

Proposal & Work Plan The original proposal for REDNET contained a section (Component 5) for Marketing Study of Redfish, which was later amended with a Statement of Work. (See Appendix A)

Methods and Data The methods that were used in this market study included graphical analysis of annual landings and prices from 1950 through 2012, linear regression of quantity on price for daily landings from 2007 through 2010, and a telephone market survey of fish restaurants and retail fish markets. Data included annual quantities and ex-vessel revenues for redfish from the NOAA Office of Science and Technology, daily redfish landings and prices from the Whaling City Seafood Display Auction in New Bedford, daily redfish landings and prices from the Northeast Fishery Science Center, interviews with restaurant and fish market personnel, and interviews with redfish processors and visits to redfish processing plants.

Products Delivered and Conclusions I delivered several presentations to RedNet meetings and workshop, including the Redfish Symposium at UNH on November 5, 2010 (attached), the Redfish Workshop in Portland on October 24, 2011 (attached), and the RedNet Meeting at the Urban Forestry Center in Portsmouth NH on December 19, 2013.

The central point of the 2010 presentation was that the redfish fishery had been dormant at low landings for 30 years, but was beginning to revive from 2007 -2009 with rapidly increasing quantity of landings and rapidly declining exvessel prices. I had thought at that time that thin markets for redfish meant that the market could not absorb increased redfish landings at sustained prices. I also argued that market development for redfish required regulation that stabilized landings without large fluctuations, government subsidies for generic marketing of redfish, industry-wide investments in marketing, firm-specific marketing investments, and landing relatively large fish (min 9” or 10” that equaled or exceeded 1 pound). The latter was determined by a fillet large enough (3oz-4oz) for restaurant portions.

The 2011 presentation was based on the results of a market survey produced by the UMass Dartmouth Center for Market Research. (See attached)

The 2013 presentation updated the 2010 presentation showing continuing increases in N.E. redfish landings from 2007 through 2012 but with increasing ex-vessel prices. This anomaly of declining prices with increasing landings from 2007-2009 and increasing prices with larger REDNET Draft Final Report 30

increases in landings from 2009-2012 could not be easily explained using the market only for redfish. My colleague Min-Yang Lee, Economist at the Northeast Fishery Science Center, who had helped with the econometric modeling of redfish (see below), argued that prices for groundfish species depended mainly on landings of all similar groundfish species because they are close substitutes for consumers. This implies that the price of redfish would move in the same direction as the prices of all similar groundfish.

6,000 1.40 NE Redfish landings 1.20 5,000

1.00 4,000

0.80 3,000 0.60

2,000 0.40

1,000 0.20

0 0.00

Figure 1. Annual Landings (MT) and price per pound for NE landings from 1984 to 2012

As evidenced by simple correlation, average groundfish prices declined from 2007-2009 but increased from 2009-2012 (Georgianna et al, p. 17).

Sai Sravanthi Pasumarthia, who was a summer intern in 2011 from the School of Mechanical Sciences at the Indian Institute of Technology Bhubaneswar, and I analyzed the daily prices of redfish from 1997 to 2010. We tested the hypothesis that marketing capacity for redfish limited the price of redfish based on a model of a market for redfish rather than a market for groundfish. In other words, low daily redfish landings would yield a high price with gradually declining prices as landings quantities increased but would drop sharply when the market for redfish was saturated.

A graphical analysis showed that the data did not support this hypothesis. Rather, the data showed large price variation with low redfish landings with less price variation with large REDNET Draft Final Report 31

redfish landings. This result is consistent with the model of redfish prices depending on total landings of similar groundfish rather than price simply depending on redfish landings. However, this analysis left out other independent demand and supply variables such as prices and quantities of other groundfish and other fish landings.

I also contracted with the Center for Market Research at UMass Dartmouth to conduct a telephone survey of restaurants and fish markets in New England of their demand for redfish (See attached). Most fish restaurants did not serve redfish. Most restaurant managers or other respondents rated fillet size as very or somewhat important but they rated redfish fillet size as fair or good. There were some positive responses: most respondents preferred local seafood. Most thought eco-labeling valuable but few would definitely pay a higher price for eco-labeled fish.

Most studies of Eco-Labeling test their effects on fish conservation rather than on consumer demand, i.e., does eco-labeling increase consumer demand for seafood products. While I have not made a complete study of economic effects of eco-labeling, most studies show the same result as the restaurant survey described above. Price weighs more heavily in consumer choice than eco-labeling. Customers who prefer fishery products that are eco-label are typically higher income, concerned with conservation, who do not trust fishery regulation (Wessels, et al and Brécarda et al). Increasing demand for eco-labeled seafood requires consumer education on both quality of seafood and the importance of conservation practices.

Discussion The central question remains, “Would marketing programs increase redfish exvessel prices?” The restaurant survey indicated that most restaurant managers do not serve redfish but prefer locally landed (Massachusetts) fish. This implies that consistent redfish landings may encourage more restaurant owners to serve locally landed fresh redfish. Marketing programs to raise the price of redfish runs up against redfish imports from Canada, which sell for less than half the price of redfish landed in New England. We may not be asking the right question. For example a better question may be, “Can smaller redfish (7” to 8”) redfish be sold for human consumption rather that for lobster bait?” Answering this question would require processor and consumer testing.

Proposal and Statement of Work. Component 5: Processing/Marketing (text from final REDNET proposal) Redfish fresh and frozen fillets and other products were major products of Maine and Gloucester processors until the late 1970s or early 1980s. Sold as “ocean perch”, redfish competed as lower price whitefish in U.S. markets. When that market was lost due to low landings, other products, including fresh and frozen fillets from other Sebastes species from Canada and Northern Europe, filled those market niches. Once given up, markets are difficult to re‐establish because markets are open access resources subject to congestion. In other words, investing in new markets typically doesn’t yield sufficient returns because other sellers enter the market without paying market development costs (Hogan & Georgianna, 1989). Selling REDNET Draft Final Report 32

fresh products into new markets is especially costly due to increased transaction costs due to perish ability of product (Georgianna & Hogan, 1986).

In order to estimate the feasibility of processing and marketing for some projected level of landings, we will analyze the pattern of landings, wholesale and retail sales, prices, and margins during the period when redfish was landed. We will then interview the processors that sold redfish in the 1980s and the processors who seem likely to start if redfish is landed in sufficient quantity, including Canadian processors of other Sebastes species. As part of this feasibility study, we will estimate minimum efficient scale of redfish landings for processing and wholesale sales.

Statement of work (Additional clarification to implement REDNET proposal) In order to attain the objective “evaluate potential markets and the processing capacity for redfish” (Rednet Proposal, p. 1) I will conduct a marketing study on the feasibility of increased sales of fresh redfish fillets, analyze effects of large scale landings on exvessel prices, investigate the possibility of eco-labeling, and contribute to all network meeting of Rednet, especially in the areas of processing and marketing redfish. Marketing study: I am currently leading a group of five marketing students in their project to estimate consumer preference for fillet size, local fresh product, eco-labeling, and fillet quality. We will meet every other week until the project is completed at the end of the spring semester. Thus far, we are considering interviewing consumers and/or wholesale buyers. My preference is to survey and interview restaurant owners (and chefs) and supermarket buyers. We are also working on a survey instrument. This project is under the direction of the Center for Marketing Research at UMass Dartmouth, who have completed several marketing studies for seafood products. The Center will charge Rednet $2,500 for this study.

In order to analyze effects of large-scale landings on exvessel prices, I will correlate price to quantity using daily auction prices and quantities from the Whaling City Display Auction. I have just received access to these data. My plan is to use statistical techniques to determine a break point (shift in coefficient) in the price-quantity relationship. I will also interview processors and dealers to estimate the extent of the market and problems that they have faced with quality and timing of landings, spatial extent of the market, and competition with low priced frozen fillets. In order to investigate the possibility of eco-labeling, I will summarize the literature on eco- labeling, including the Marine Stewardship Council and other agencies, into a description of the costs and benefits of each labeling service. As stated above, we will also investigate the influence of eco-labeling on demand from buyers. We will attend all meetings with special consideration to processing and marketing. Thus far, members of the group have begun an analysis of the size structure of redfish stocks and its changes through time. It is possible that the influence of size on price will inform the size selectivity of the trawl.

References

REDNET Draft Final Report 33

Brécarda, Dorothée, Boubaker Hlaimib, Sterenn Lucasa, Yves Perraudeaua, Frédéric Salladarréa. 2009. Determinants of demand for green products: An application to eco-label demand for fish in Europe. Ecological Economics. Volume 69, Pages 115–125

Georgianna D., R. Avila, D. Bethoney, A. Cass, S. Challingsworth, and W. Hogan. 2014. Groundfish Port Recovery and Revitalization Plan for the Port of New Bedford/Fairhaven. Mimeo.

Lee, Min-Yang and Eric Thunberg. An Inverse demand System for New England Groundfish: Welfare analysis of the transition To Catch Share management. Amer. J. Agr. Econ. 95(5): 1178– 1195.

Wessels, C. R. Johnston, and H. Donath. 1999. Assessing Consumer Preferences for Ecolabeled Seafood: The Influence of. Species, Certifier, and Household Attributes. Amer. J. Agr. Econ. 81(5): 1084-1089.

VI. Component 6: Outreach and Implementation

Summary Development and implementation of an outreach plan and scope of work was led by Erik Chapman of New Hampshire Sea Grant and was reviewed and approved by REDNET. However, timing of events, particularly the completion of research and unavailability of results, hampered implementation of the outreach plan as of October 2014.

One particular aspect of the outreach was a meeting facilitated by Chapman in response to concerns that participation in REDNET, particularly by industry members, had declined. NH SG Extension Specialists organized and facilitated a meeting in June, 2014 attended by 35 fishermen, scientists and managers. The meeting reviewed the progress of the REDNET project (A regional project designed to conduct research to inform development of a redfish fishery in the Northeast). Meeting was held in Portsmouth, NH. REDNET members discussed possible projects and prioritized their ideas. Meeting results were shared with REDNET participants. Outreach efforts were focused toward the end of the project which was not ideal for meeting outreach goals of the project. Instead, it is recommended that outreach specialists are engaged in the full process of collaborative research projects in order to realize opportunities for engagement with NOAA and industry personnel. This approach will maximize the chances for successful transfer of knowledge to inform adaptive management, science and fishing practices.

Scope of Work and Proposed Strategy

REDNET Outreach Plan – Scope of Work REDNET Draft Final Report 34

August 22, 2013

Overview

Outreach Goals 1. Industry uptake of gear that is effective in targeting redfish and minimizing bycatch (if/when gear is developed that is ready for transfer to industry) 2. Increased awareness of redfish processing, stock status, and market potential 3. Facilitate regulatory/policy change 4. Improved networking among fishermen, scientists, managers, eNGOs associated with redfish fishery

Audiences • REDNET Network • Fishermen and related industry representatives o Sector managers • Fisheries Managers • Fisheries Scientists • General public (eNGOs) • Dock-side industry and Value-Chain

Activities . Articles in industry publications . Website/portal o Short video segments/podcasts o Project documents o Blog . Dock-side workshops/roundtable meetings . Information/demonstration at meetings, festivals etc . REDNET final workshop . A REDNET Google Community (to facilitate file sharing, discussion, video chatting, information exchange)

Evaluation • # of new entrants into redfish fishery • # Policy/regulatory changes • Increase in overall utilization of redfish o Landings o Revenue

REDNET Draft Final Report 35

Proposed Strategy and Scope of Work

A series of 4-month plans that includes a timeline and specific strategies and actions will be developed in coordination with REDNET Coordinating Committee beginning in November 2013 and ending in October 2014. The overall, general plan for meeting outreach goals is provided below and all the work will be carried out by Dr. Erik Chapman and Dr. Gabriela Bradt:

Goal 1. Industry uptake of gear that is effective in targeting redfish and minimizing bycatch (if/when gear is developed that is ready for transfer to industry)

Audience: Fishermen

Activities: • Technology transfer program for gear developed as part of the project o Broadcast results of gear development research . Articles in Industry Publications (e.g. Commercial Fisheries News) . One-page fact-sheets distributed to dock-side meeting-points for fishermen . Dock-side workshops (planned when/where upon request by fishermen) . Video and article summaries on website . REDNET final workshop o Make gear available through some form of sub-subsidized trial period . Transfer will include some continued data collection and feedback on effectiveness and modifications made by fishermen

Goal 2. Increased awareness of redfish processing, stock status, and market potential

Audience: Fishermen, dock-side businesses (dealers, retailers), General Public, eNGOs

Activities: • Marketing of methods and opportunity for redfish in the marketplace o Broadcast methods of processing and availability to industry . Articles in Industry Publications (e.g. Commercial Fisheries News) . Develop web content that provides information on tools etc that are used in redfish processing o Broadcast stock status, availability and opportunity for consumers . Develop web content that provides information about redfish to consumers . Provide information at regional seafood festivals o Provide relevant information at REDNET final workshop

Goal 3. Facilitate regulatory/policy change

REDNET Draft Final Report 36

Audience – Fisheries managers

Activities: o Hold conference calls/webinars/roundtables as needed to communicate project results, and relevant policy concerns from industry to managers.

Goal 4. Improved networking among fishermen, scientists, managers, eNGOs associated with redfish fishery

Audience: See above.

Activities: . Develop web portal with blog and discussion capacity. . Hold webinars as needed to broadcast project accomplishments to industry parties . Present information at final REDNET workshop and facilitate discussion among interested parties about actions/next steps. . Establish a REDNET Google Community

A website was established: http://redfishrednet.weebly.com/about.html

REDNET Draft Final Report 37

VII. Conclusions and Recommendations

Note: These conclusions and recommendations are preliminary and may change following more data analysis. Also, they do not reflect the opinions or endorsements of the members of the REDNET network.

Discussion of individual components is included within the separate reports of those activities, and will not be repeated or reiterated in this section. This section briefly outlines tradeoffs in different potential implementation strategies or management options, based on results and input from the components of REDNET.

Options to Reestablish a Sustainable Redfish Fishery

Status Quo • Large mesh (6.0/6.5 inch codends) o Network members indicated only vessels with winch engines or high horsepower can haul back fast enough to avoid loss of catch with this mesh size, and that smaller boats cannot. o Component 3 results verified that use of this mesh results in incomplete selectivity even at the largest sizes of fish. o Ninety-five percent of all redfish that enter the net pass through the codend meshes and suffer an unknown level of injury and morality. o Video escapement data for 6.5 in codends were not yet processed to show where escape occurs o Markets (Component 5) would prefer fish of the larger sizes (>30 cm), but it appears that the population does not include many fish of this size based on catches in Components 2, 3, and 4.

• Minimum landing size (7 in) o Component 5 (Marketing) found that prices are low for smaller sizes of fish and recommended efforts to rebrand this fish for human consumption and presumably a higher price. o Component 3 results found that a 4.5 inch codend retains only 5 percent of the 7 in fish. To reach an L25 of seven inches would require a mesh smaller than 4.5 in. o Component 3 results indicated that 51% of fish redfish entering the net escaped through a 4.5 in codend. o Network members considered 4.5 in to be the smallest mesh possible to be approved in a special access program.

• Sector Exemptions o Sector requests for 6.0” mesh exemption and 4.5” exemptions, made in part by network members, were granted but not used because of the 100% industry-funded observer coverage stipulation, which was triggered primarily due to the potential to switch between 4.5” and 6.0” mesh in a trip. Some sector members felt codend REDNET Draft Final Report 38

switching was essential to profitability. Other sector members indicated that future exemptions might have relaxed requirements, which occurred in 2014.

Special Access Program(s) • 4.5 diamond mesh o Network members indicated that the yield-per-recruit would be reduced with smaller mesh. Some network members supported this size; others opposed it. One argument made in support was that mortality could be more fully determined by bringing catch on deck. Network members described the need to be able to use 4.5 in and larger mesh on the same trip, which triggered a need for 100% observer coverage, and industry’s costs to implement it. o Component 2 demonstrated that bycatch across the stock area and across 5 vessels with different experience in the fishery resulted in very low bycatch with this mesh size. Some network members indicated that the level of bycatch was biased low because testing vessels were highly experienced. o Component 3 results indicated that 51% of fish redfish entering the net escaped through a 4.5 in codend and experienced an unknown level of injury and mortality. o Component 4 results indicated that fish escape from the 4.5 in codend mostly during haulback, and are exposed to more threats than escape at bottom. o Component 4 found improvement in escapement at bottom using a dual-grid system. o Component 5 found that prices are low for smaller sizes of fish and recommended efforts to rebrand this fish for human consumption and presumably a higher price.

• Time/Area Restrictions o Component 5 results indicated that a steady, predictable supply was best for market development and expansions. o Network members largely opposed time/area restrictions, with some members supporting them based on enforcement or implementation issues. Also, the historic fishery was time

• Bycatch Reduction Devices o A 40 mm bar spacing in a dual grid system was not more size-selective than a 4.5 in codend, while a 50 mm spacing reduced all sizes. o Dual grid systems represent challenges, but could likely be made to work o Methods to improve selection at depth in trawl gears are limited. o Component 4 results were not presented to the network. The participating captain in dual grid testing was interested in using it commercially.

Opportunities and mechanisms to implement changes to the redfish fishery include individual action, sector action, sector exemptions, NMFS GARFO rules, and NEFMC action. No particular mechanism was identified by the network members. Network members of sectors REDNET Draft Final Report 39

independently used results from REDNET to support exemption requests. Future outreach should include sharing of results broadly, once final data analysis is complete.

REDNET Draft Final Report 40

Additional Material

Component 2: Exploratory Fishing Revised Final Report

Component 3: Codend Selectivity 1) Revised Final Report 2) Pol, M. V., Herrmann, B., Rillahan, C., He, P., 2015. Selectivity and retention of pollock Pollachius virens in a Gulf of Maine trawl fishery. Fish. Res. 1–7. doi:10.1016/j.fishres.2015.07.029 3) Pol, M. V., Herrmann, B., Rillahan, C., He, P., In revision. Impact of codend mesh sizes on selectivity and retention of Acadian redfish Sebastes fasciatus in the Gulf of Maine trawl fishery. Fish. Res.

Component 5 (Processing/Marketing)

1) Market survey produced by the UMass Dartmouth Center for Market Research 2) Processing and Marketing Capacity for Acadian Redfish in New England – D. Georgianna presentation to 2010 Redfish Symposium, 5 November 2010 3) Processing and Marketing Capacity for Acadian Redfish in New England – D. Georgianna presentation to Redfish Network Meeting II, 24 October 2011 4) S.S. Pasumarthi, University of Massachusetts Dartmouth Summer Internship Project Report, 9 May – 8 July 2011 REDNET - A Network to Redevelop a Sustainable Redfish (Sebastes fasciatus) Trawl Fishery in the Gulf of Maine

FINAL REPORT Component 2 – Baseline Catch and Bycatch Evaluation

Revision 1: 18 May 2016 Date: 23 January 2013

Written by: Kohl Kanwit1, Mike Pol2 and Pingguo He3

1. Maine Department of Marine Resources 2. Massachusetts Division of Marine Fisheries 3. UMass Dartmouth School for Marine Science and Technology

Funded by: Northeast Cooperative Research Partners Program Northeast Fisheries Science Center National Marine Fisheries Service

Executive Summary Acadian redfish (Sebastes fasciatus) is among a few groundfish stocks in the northeast US that have a relatively large Annual Catch Limit (ACL) that have not been fully utilized during the last several years. With reduced ACLs for many important groundfish species, full and sustainable utilization of the redfish resource is thus critical to the economic survival of fishermen and to the success of groundfish sectors. The allocations of redfish could not be fully utilized under current regulations because the minimum codend mesh size is too large to effectively retain redfish. The REDNET project seeks to achieve three fishery conservation and management goals:

1) Redirecting fishing effort in the multispecies fishery away from stocks that are overfished to stocks that are considered rebuilt (e.g. redfish).

Page 1 2) Achieving optimum yield, by increasing commercial landings of redfish through development of a directed fishery under the adaptive management ability of groundfish sectors. 3) Increasing the economic viability of groundfish sectors by providing access to the ACL of a recovered species and thus generating much‐needed revenue for the industry.

REDNET includes six components:

Component 1: Network Meetings Component 2: Baseline Catch and Bycatch Evaluation Component 3: Codend Selectivity Component 4: Conservation Engineering and Bycatch Reduction Component 5: Processing/Marketing Component 6: Outreach/Implementation

This report describes work carried out in Component 2 - Baseline Catch and Bycatch Evaluation.

The work was carried out with a structured and monitored exploratory fishing method using five commercial fishing vessels in every two to three months between the end of May 2011 and January 2012. The vessels used their own commercial groundfish trawls, but a relatively small mesh codend (114 mm or 4.5” mesh size, double twine diamond) provided by the project. Each vessel fished 4-5 days with a total of 85 tows, ranging from 15 to 20 tows per trip (per vessel). Fishing was carried out on traditional redfish grounds throughout the central portion of the Gulf of Maine at depths ranging from 77 fm to 138 fm. Two technicians (or NOAA-approved observers) were on each of the trips to record operational conditions and to sample catch and bycatch.

The total catch of all species was just over 234,000 lb; with redfish comprising 232,380 lb (94.9% of all catch). The amount of redfish caught per trip ranged from about 35,000 to 67,000 lbs and the catch per tow ranged from 0 to 13,655 lb, averaging 2,766 lb. Total kept and landed redfish were 221,957 lbs, or 95.5% of all redfish caught. Pollock (Pollachius virens) was the main landed bycatch species (7,307 lb), with 13 other species landed in smaller amounts.

Thirty-four species were discarded, totaling almost 41,000 lb (14.9% of total catch), with spiny dogfish (Squalus acanthias) accounting for the majority of discards (>26,000 lbs, or 63.4%). Discards of undersized redfish (MLS: 9” or 23 cm) totaled 10,423 lb or 4.5% of the total redfish catch. Total discards of undersized pollock were 2,744.8 lbs. Total discards were less than 200 lbs per species for other species.

Discards analyzed using the Standard Bycatch Reporting Method revealed that almost all median SBRM ratios below 0.01. Only spiny dogfish had SBRM exceeded 0.01. Discard

Page 2 weights of redfish were higher in summer (July and September) and highest at shallower water depths (between 84 and 107 fm).

The exploratory fishing results indicate that it is possible to harvest redfish using 4.5” diamond mesh without substantial catch of undersized redfish and other commercially important groundfish species. Depth, time-of-year, or some other factors appear to have an impact on the catch of undersized redfish thought we were not able to isolate these factors.

The size of redfish caught may be further reduced through codend mesh size adjustment or using size-sorting devices. A size-sorting grid system may allow smaller redfish to escape at depth which will result in higher survival rates than through codend meshes at the surface. These aspects of research will be carried out in Component 3 and 4 of this project.

Introduction

Acadian redfish (Sebastes fasciatus – “redfish”) is the only groundfish species in the northeast US with a relatively large Annual Catch Limit (ACL) but has not been fully utilized by the current multispecies fishery participants during the last several years. The requirement for codends with large mesh sizes (6.5”) and lost market may be the primary reason for the underutilization. The implementation of sector system with catch retention rules and accountability measures makes targeting redfish with codend mesh sizes smaller than the current regulated mesh size possible. The Maine Department of Marine Resources, the Massachusetts Division of Marine Fisheries and the University of Massachusetts School for Marine Science and Technology joined with other members of the scientific community and the industry have developed a research plan that draws on wide‐ranging expertise in order to conduct comprehensive research on the development of a sustainable directed redfish trawl fishery in the Gulf of Maine. The project, called “REDNET” includes components of catch and bycatch assessment through exploratory fishing, evaluation of codend mesh size, testing of size and species selective design and devices, processing and marketing, outreach, and implementation. The project is funded by NOAA Fisheries Northeast Cooperative Research Partners Program’s 2010 funding cycle.

The redfish cooperative research project or “REDNET” seeks to achieve three fishery conservation and management goals:

• Redirecting fishing effort in the multispecies fishery away from stocks that are overfished to stocks that are considered rebuilt (e.g. redfish).

Page 3 • Achieving optimum yield, by increasing commercial landings of redfish through development of a directed fishery under the adaptive management ability of groundfish sectors. • Increasing the economic viability of groundfish sectors by providing access to the ACL of a recovered species and thus generating much‐needed revenue for the industry.

Groundfish management in the Northeast made a dramatic shift in May of 2010 from primary input controls to primary output controls. Amendment 16 to the Northeast Multispecies Fishery Management Plan (FMP) established the rules for sector management, as well as catch limits and accountability measures mandated by the 2006 reauthorization of the Magnuson‐Stevens Act. If multispecies sector management is to be successful, the fleet must be able to catch and land stocks of high abundance such as redfish while exercising their ability to avoid quota-limiting species (e.g. most recently, Gulf of Maine cod).

Historically, redfish represented a significant fishery and income in the region. The current best available estimates assert that the resource can support a larger fishery. The directed redfish fishery in the northeast began in the 1930s and total landings rose from 100 mt to a peak of over 117,000 mt in 1951 and then steadily declined. By 1983, the total US landings of redfish were 5,328 mt, and in 2008 landings were only 1,189 mt. The redfish fishery in the Gulf of Maine was traditionally prosecuted by vessels using otter trawls with relatively small mesh codends in the range of 70 ‐ 80 mm (2.5 ‐ 3”). Since 1950s and through 60s and 70s, the population of redfish declined, likely as a result of overexploitation. In 1977, the minimum codend mesh size increased from 114 to 130 mm (4.5 to 5”) and increased again in 1994 to 152 mm (6”). Today the minimum codend mesh size mandated by the Multispecies FMP is 165 mm (6.5”) , though groundfish sector vessels may use 152 mm (6”) codends provided an observer has been assigned to the fishing trip. These mesh restrictions, combined with low biomass levels between 1980 and 1995 eliminated a directed redfish fishery in the Northeastern United States. In recent years, the combined restrictions in the multispecies FMP have resulted in the recovered status of the redfish resource. A stock assessment of redfish was completed and reviewed at the 2008 Groundfish Assessment Review Meeting (GARM III) and updated through 2010 at a 2012 review. IThe redfish stock is not overfished and overfishing is not occurring.

Redfish (Sebastes spp.) are harvested in directed and non‐directed fisheries throughout their global range. In Iceland, the fishery for Sebastes marinus (golden redfish) is prosecuted with bottom trawls and a minimum codend mesh size of 100 mm (4”) in the directed fishery. A suite of management measures including sorting grids is required in other small mesh fisheries to minimize redfish bycatch and exclusion zones are used for the protection of juveniles, and temporary closures if juvenile catches are high. Sebastes mantella or oceanic redfish is targeted in Iceland with pelagic trawls. Oceanic redfish are

Page 4 also called the “deepwater” redfish, inhabiting waters 350‐700 m and are considered exclusively pelagic. Unlike Acadian redfish (Sebastes fasciatus) that are primarily benthic, their pelagic distribution makes them susceptible to midwater gears. Norwegian fishery management has banned the directed fishery for golden redfish since 2003 but allows a bycatch limit of 15% in the mixed trawl fishery which uses 135 mm mesh (5.3”). Other countries that operate in the eastern Atlantic and harvest Sebastes spp. include Russia, Germany and Denmark; the latters operate under European Union regulations.

The Canadian redfish fishery in the northwestern Atlantic takes place in deep waters from Nova Scotia to Newfoundland. They primarily manage the resource with quotas in three stock units. The fishery is permitted to use 110 mm (4.3 in) diamond mesh codend versus the 130 mm (5.1 in) square mesh codend required in the groundfish fishery. To avoid large catches of small redfish there are permanent closed areas. It is not clear if the small redfish in the closed areas are segregated juveniles or just a smaller subpopulation of adults. Bycatch is managed with quotas and a percent of the total catch.

Development of a redfish fishery in the Northeastern United States is critical to the economic survival of fishermen and to the success of groundfish sectors. Sectors were recently implemented through Amendment 16 to the multispecies FMP. The sectors are assigned percent allocations based on the historic catches of their members. However, the allocations of redfish could not be fully utilized under current regulations because the minimum codend mesh size is too large to effectively retain redfish. At the same time, allocations of other groundfish species are extremely small due to the very low ACLs recommended. These limiting allocations (“choke” species) will affect the fishing behavior of the sectors and potentially shut them down before more abundant allocations are realized. It is critically important to the success of sectors to find a way to allow them to access allocations of healthy stocks such as redfish while avoiding those that are depressed.

Page 5 Project Design

REDNET is a multifaceted, comprehensive project to determine how to best access the redfish ACL for the groundfish sectors. The goal is to execute a complete project, from conception to regulatory implementation, to marketing, that supports environmentally and economically sustainable harvesting of redfish. REDNET includes individuals with various expertise to accomplish this vision. There are six components of the project:

Component 1: Network Meetings Component 2: Baseline Catch and Bycatch Evaluation Component 3: Codend Selectivity Component 4: Conservation Engineering and Bycatch Reduction Component 5: Processing/Marketing Component 6: Outreach/Implementation

This report summarizes the results from Component 2: Baseline Catch and Bycatch Evaluation. While the entire REDNET project is still ongoing, Component 2 is complete. This is the final report on the data gathered through this yearlong effort.

Redfish are rarely targeted in the modern groundfish fishery due to codend mesh size restrictions (6.5” minimum), and therefore there are no recent data to describe commercial availability of the target species and associated non-target species. Before Component 2 was initiated, existing databases on occurrence and catch rates of redfish were examined from Vessel Trip Reports (VTR), Observer data (OBDBS) and NMFS survey database (SVDBS) (see Attachment A).

Methods

An Experimental Fishing Permit was received in March 2011 (Attachment B). REDNET industry partners shared trips with a total of five vessels participating, each making a five-day trip over the course of a year. Trips were made every two to three months; attempting to capture seasonal variations in availability or occurrence of the target species and/or incidental bycatch. The participating vessels used their own commercial groundfish trawls with two 114 mm (4.5” nominal) mesh size, double twine diamond mesh codends provided by the project. One codend was used during each tow while the other was kept as spare. Trawls varied between vessels. See Attachment C for a table of gear characteristics.

The entire Gulf of Maine was defined as the study area and year‐round sampling was targeted. Fishermen were asked to fish in a commercial manner to maximize their redfish catches and minimize discards. In some cases, vessels emphasized broader spatial coverage over catch rates. All regulated groundfish were counted against the sector annual catch limits (ACLs). Participating sector members’ redfish allocations were

Page 6 used to offset the guaranteed daily vessel rate through sales of redfish. The project also reimbursed the participating vessels for fuel in order to encourage the vessel to fish at different locations throughout the Gulf of Maine. Vessels retained revenue from sales of fish other than redfish.

Two observers were on board the vessels during each and every trip and documented all trip, tow, environmental, catch, and bycatch data following National Marine Fisheries Service (NMFS) Fisheries Observer Program protocols. The observers determined the total catch of legal and sublegal redfish per tow and then identified, and weighed all other species. A special protocol was agreed upon between the NMFS habitat group and the project participants in the event that deep-sea corals were encountered.

Lengths (measured as total length, TL) of a random subsample of approximately 100 redfish from each tow were measured; other important species were measured opportunistically. For redfish length-frequency (LF) analysis, counts at each length were multiplied by the subsample weight divided into the total weight. On two trips, lengths were subsampled before redfish were separated into kept and discard, and on three trips, lengths were subsamples after sorting into kept and discard. Where kept and discard lengths were measured separately, counts were scaled to the appropriate weight and the kept and discard counts were added together. The resulting estimated total number per tow by length was then used, resulting in larger tows being weighted more heavily.

All catch data (along with trip and gear data) were entered and uploaded into the Maine Department of Marine Resources (DMR) biological database (MARVIN). Catch and bycatch data were plotted in a Geographic Information System (GIS) so that seasonal target species concentration and bycatch “hot spots” could be identified. Data were downloaded and converted into a Microsoft Access database. Collected data were analyzed using Microsoft Excel and R statistical software (R Development Core Team, 2009), primarily using the lattice package (Sarkar, 2009). Catch and bycatch were characterized using the Standard Bycatch Reporting Method developed as part of the national bycatch initiative by NMFS (Wigley et al. 2007). The ratio of landed or discarded pounds of each species (bspecies_x) to kept pounds of all species (kall_species) for each tow sampled by observers at sea.

b species _ x kall−species

The ratio can be expanded to total bycatch of a fleet according to commercial landings of all species (Kall_species) by fleet, and its variance can be estimated according to Wigley et al. (2007). The sample variance for the experimental fishery was zero.

Page 7

b B = species _ x K species _ x k all _ species all−species

Results

Codends were measured prior to fishing using an OMEGA mesh gauge following International Council for Exploration of the Seas (ICES) protocols (Fonteyne, 2005). Mean mesh sizes of the two codends were measured to be 110 mm (4.3 inches; n= 60, var = 5.1) and 111 mm (4.4 inches, n= 60, var = 4.7). The five vessels that participated (Table 1) are members of two sectors in the northeast multispecies fishery, each of which have considerable amount of allocations of redfish. Two of the participating vessels are among those currently targeting redfish with 6.0 or 6.5 inch mesh size codends. The first trip departed at the end of May 2011, followed by trips in July, September, December of 2011, and the last trip was completed in January 2012.

Table 1: Characteristics of participating vessels, dates, and effort. Vessel Length (ft) Departure Date/Time Return Date/Time No. of tows Olympia 71 05/31/11, 20:40 06/05/11, 15:12 17 Jocka 61 07/05/11, 19:15 07/10/11, 13:48 17 Black Beauty 63 09/06/11, 15:53 09/11/11, 13:48 16 America 76 12/05/11, 09:58 12/10/11, 07:45 15 Guardian 72 02/15/12, 13:00 02/21/12, 04:00 20 Total number of tows 85

Vessels fished 4-5 days with a total of 85 tows, ranging from 15 to 20 tows per trip. Minimum depths fished for each vessel ranged from 77 fm to 101 fm and maximum depths ranged from 102 fm to 138 fm. The average depth fished over a five-day trip ranged from 89 fm to 112 fm with the deepest trip fished occurring in the winter (February) and the shallowest trip occurring in the summer (July) (Figure 1). Median towing speeds by vessels ranged from 2.9 to 3.5 knots. Median warp lengths by vessels ranged from 225 to 300 fm. Median wave heights experienced by vessels during towing were between 2 - 3 ft (Figure 1).

Tow locations were distributed throughout the central portion of the Gulf of Maine (Figure 2), and were based on captains’ previous knowledge, echo sounder signals, and a goal of wide coverage. Only daylight tows were made with the earliest start time at 4:38 am and the latest start time at 5:33 pm, both on the June trip. Tow duration ranged from 15 minutes to a little over an hour and a half. The average tow duration was 37 minutes, but varied by trip from an average tow time of 26 minutes to 46 minutes. Tow durations varied based on captains’ assessment of the volume of fish in the net and fishing ground conditions.

Page 8

Wire Out (fm)

350

300

250

200 Wave Ht (ft) 10 8 6 4 2 0 Tow Speed (knt)

3.5

3.0

Tow Duration (h) 1.5

1.0

0.5

Depth (fm) 140 130 120 110 100 90 80

OLYMPIA JOCKA BLACK BEAUTY AMERICA GUARDIAN May-2011 Jul-2011 Sep-2011 Dec-2011 Feb-2012

Figure 1: Operational and environmental variables by vessel, in order of occurrence by start date. Red dots are individual medians; dashed black horizontal lines are panel medians. Data points are jittered horizontally for clarity.

Page 9

Figure 2: Tow locations in the Gulf of Maine by month (June: purple; July: green; September: blue; December: red; February: yellow). The black rectangle is the Western Gulf of Maine Closed Area.

All bycatch and incidental catches were identified and enumerated. One piece of suspected coral was also caught, preserved and transported to the Smithsonian Institute for identification as per agreed coral bycatch protocol. The total catch of all species was just over 234,000 lb; with redfish comprising 232,380 lb or 94.9% of all catch (Table 2). The amount of redfish caught per trip ranged from about 35,000 lbs to about 67,000 lbs and the catch per tow ranged from 0 lb (4 tows with no redfish) to 13,655 lb, averaging 2,766 lb (Figure 3). The catch per unit effort (CPUE; pounds of redfish per minute towed) averaged 84 lb/min with a maximum of 701 lb/min. Total kept and landed redfish were 221,957 lbs, or 95.5% of all redfish caught. Pollock (Pollachius virens) was the main landed bycatch species (7,307 lb), with 13 other species landed in smaller amounts (Figure 3).

Page 10

Table 2: Total catches (lb) by species and by disposition from all five trips.

Species Kept Discard redfish acadian ocean perch Sebastes fasciatus 221,957.1 10,423.1 pollock Pollachius virens 7,307.6 2,744.8 hake white Urophycis tenuis 2,027.6 43.2 cod atlantic Gadus morhua 1,039.4 140.6 haddock Melanogrammus aeglefinus 972.1 87.3 cusk Brosme brosme 251.0 monkfish Lophius americanus 182.2 3.8 flounder atlantic witch (gray sole) Glyptocephalus cynoglossus 117.2 7.4 hake silver (whiting) Merluccius bilinearis 53.6 71.0 lobster american Homarus americanus 44.8 192.4 plaice american (dab) Hippoglossoides platessoides 33.6 18.1 rosefish black bellied Helicolenus dactylopterus 8.3 skate winter Leucoraja ocellata 5.5 28.6 squid short-finned Illex illecebrosus 0.8 47.9 hake atlantic red Urophycis chuss 0.4 81.0 dogfish spiny Squalus acanthias 26,376.9 sea raven Hemitripterus americanus 197.5 skate barndoor Dipturus laevis 123.7 rock rock 57.5 skate thorny Amblyraja radiata 55.0 wolffish atlantic Anarhichas lupus 41.7 skate smooth Malacoraja senta 35.8 shad american Alosa sapidissima 24.7 crab northern stone Lithodes maja 8.0 anemone frilled Metridium dianthus 5.0 cucumber sea Cucumaria frondosa 2.2 anemones Actiniaria 2.0 flounder fourspot Paralichthys oblongus 1.7 cunner Tautogolabrus adspersus 1.5 lumpfish Cyclopterus lumpus 1.5 skate little Leucoraja erinacea 1.5 starfish Asteroidae 1.2 sponge finger Haliclona oculata 1.0 scad round Decapterus punctatus 0.8 herring atlantic Clupea harengus 0.8 alewife Alosa pseudoharengus 0.4 Totals 234,001.0 40,829.4

Page 11 hake silver (whiting) pollock redfish acadian ocean perch 54.0 800 10000 53.8 600 53.6 400 5000 53.4 200 53.2 0 0 haddock cod atlantic hake white 300 150 100 200 100

50 100 50

0 0 0 skate winter cusk monkfish

5.8 30 50 40 5.6 20 30 5.4 20 10 5.2 10 0 plaice american (dab) flounder atlantic witch (gray sole) lobster american 15 12 4 10 10 3 8 6 2 5 4 1 2 hake atlantic red squid short-finned rosefish black bellied 0.8 1.2 3.0 0.6 1.0 2.5 0.4 0.8 2.0 0.6 0.2 1.5 0.4 0.0 1.0 OLYMPIA JOCKA BLACK BEAUTYAMERICA GUARDIAN OLYMPIA JOCKA BLACK BEAUTYAMERICA GUARDIAN OLYMPIA JOCKA BLACK BEAUTYAMERICA GUARDIAN Jun-2011 Jul-2011 Sep-2011 Dec-2011 Feb-2012 Jun-2011 Jul-2011 Sep-2011 Dec-2011 Feb-2012 Jun-2011 Jul-2011 Sep-2011 Dec-2011 Feb-2012

Figure 3: Catch (lb) by tow and vessel for all kept species. Red dots are medians for each trip; red dashed lines are species medians. Points have been jittered horizontally for clarity. Scales for weights (vertical axis) vary between panels.

Thirty-four species were discarded, totaling almost 41,000 lb (14.9% of total catch), with spiny dogfish (Squalus acanthias) accounting for over 26,000 lb (Table 2) and not landed. Discards of undersized redfish (MLS: 9” or 23 cm) totaled 10,423 lb or about 4.5% of the total redfish catch. Total discards of pollock were 2,744.8 lb and were due to small size. Total discards were less than 200 lb per species for the other species caught.

Discards were further analyzed using the Standard Bycatch Reporting Method (SBRM). SBRM is the ratio of the weight of individual species and dispositions to the total kept catch. Ratios were calculated for each tow, and summary statistics (median and quartile values, minimums and maximums) were calculated across all tows (Table 3, Table 4). All

Page 12 median SBRM ratios besides kept redfish were below 0.05. Only one discarded species, spiny dogfish, exceeded 0.01. Maximum values of 1.0 for kept species indicate tows composed entirely of that species. High maximum values for discard species (for example, >2.8 for spiny dogfish) illustrate tows where discards were very high or kept amounts were very low.

Table 3: Median SBRM values in descending order, for species with 6 or more observations (n). Full boxplot statistics are also reported. Species Disposition Min 1st Qu Median 3rd Qu Max n redfish acadian ocean perch Kept 0.1313 0.8642 0.9443 0.9784 1.0000 80 dogfish spiny Discard 0.0004 0.0064 0.0414 0.3043 2.8288 68 redfish acadian ocean perch Discard 0.0000 0.0046 0.0229 0.0721 1.4905 60 pollock Kept 0.0010 0.0069 0.0221 0.0680 1.0000 69 hake white Kept 0.0001 0.0046 0.0101 0.0367 0.2500 61 skate barndoor Discard 0.0009 0.0067 0.0099 0.0194 0.0476 8 pollock Discard 0.0001 0.0022 0.0063 0.0227 0.7095 39 haddock Kept 0.0008 0.0027 0.0061 0.0171 0.1456 44 wolffish atlantic Discard 0.0009 0.0009 0.0059 0.0139 0.0404 6 cod atlantic Kept 0.0006 0.0028 0.0053 0.0116 0.1466 34 monkfish Kept 0.0003 0.0028 0.0042 0.0245 0.0347 9 cusk Kept 0.0004 0.0014 0.0041 0.0067 0.1708 29 lobster american Discard 0.0003 0.0015 0.0035 0.0064 0.0302 32 cod atlantic Discard 0.0001 0.0010 0.0020 0.0039 0.0183 31 skate thorny Discard 0.0003 0.0007 0.0019 0.0030 0.0120 15 lobster american Kept 0.0001 0.0009 0.0016 0.0038 0.0058 12 hake atlantic red Discard 0.0001 0.0007 0.0015 0.0027 0.0635 23 sea raven Discard 0.0002 0.0006 0.0014 0.0021 0.1671 15 skate smooth Discard 0.0002 0.0005 0.0011 0.0016 0.0170 17 flounder atlantic witch (gray sole) Kept 0.0001 0.0005 0.0010 0.0027 0.0606 39 haddock Discard 0.0000 0.0003 0.0009 0.0032 0.0157 25 hake white Discard 0.0001 0.0003 0.0008 0.0016 0.0058 13 shad american Discard 0.0001 0.0002 0.0008 0.0023 0.0046 11 hake silver (whiting) Discard 0.0000 0.0001 0.0007 0.0011 0.0202 22 squid short-finned Discard 0.0003 0.0005 0.0006 0.0011 0.0315 15 plaice american (dab) Kept 0.0001 0.0003 0.0005 0.0017 0.0056 18 flounder atlantic witch (gray sole) Discard 0.0000 0.0001 0.0004 0.0007 0.0253 10 plaice american (dab) Discard 0.0000 0.0002 0.0004 0.0007 0.3182 18 crab northern stone Discard 0.0001 0.0002 0.0003 0.0005 0.0007 10

Page 13 Table 4: Median SBRM values, sorted by median, for species with 5 or less observations (n). Full boxplot statistics are also reported. Species Disposition Min 1st Qu Median 3rd Qu Max n hake silver (whiting) Kept 0.0169 0.0169 0.0169 0.0169 0.0169 1 skate winter Kept 0.0079 0.0079 0.0079 0.0079 0.0079 1 unidentified catch Discard 0.0001 0.0021 0.0077 0.0140 0.0168 4 anemone frilled Discard 0.0002 0.0002 0.0058 0.0309 0.0505 4 rock Discard 0.0005 0.0005 0.0051 0.0097 0.0097 2 cucumber sea Discard 0.0018 0.0018 0.0018 0.0018 0.0018 1 skate winter Discard 0.0008 0.0011 0.0014 0.0014 0.0014 3 lumpfish Discard 0.0013 0.0013 0.0013 0.0013 0.0013 1 starfish Discard 0.0002 0.0005 0.0008 0.0022 0.0035 3 flounder fourspot Discard 0.0000 0.0003 0.0006 0.0022 0.0039 3 cunner Discard 0.0001 0.0001 0.0005 0.0009 0.0009 2 hake atlantic red Kept 0.0004 0.0004 0.0004 0.0004 0.0004 1 rosefish black bellied Kept 0.0001 0.0003 0.0004 0.0005 0.0076 5 squid short-finned Kept 0.0002 0.0002 0.0002 0.0002 0.0002 1 monkfish Discard 0.0001 0.0001 0.0002 0.0008 0.0013 4 anemones Discard 0.0002 0.0002 0.0002 0.0002 0.0002 1 skate little Discard 0.0001 0.0001 0.0001 0.0001 0.0001 1 sponge finger Discard 0.0001 0.0001 0.0001 0.0001 0.0001 1 scad round Discard 0.0001 0.0001 0.0001 0.0001 0.0001 1 herring atlantic Discard 0.0001 0.0001 0.0001 0.0002 0.0004 3

SBRM values were more fully explored on a tow-by-tow and vessel-by-vessel basis. While “vessel” is used in the plots, it encompasses a number of potential factors: season (which itself includes several factors), vessel, and net. When interpreting results, the multidimensionality of “vessel” should be considered.

A subset of species, of which at least 10 lbs were caught, was also examined (Figure 4, Figure 5, Figure 6). Several tows showed high SBRM ratios for spiny dogfish both within and across trips. Redfish discards were notably high for one tow in September 2011. On the same trip, one tow showed a high SBRM ratio for discards of pollock. For the other 32 species discarded, SBRM ratios were consistently close to zero.

Page 14 species. each for included also are observations of Numbers clarity. for horizontally jittered been deleted Figure Species lbs caught/total lbs kept 4 . Red dots are medians for each trip each for medians are dots Red . : SBRM ratios by tow and vessel for for vessel and tow by ratios SBRM : 0.0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 2.5

OLYMPIA 0 n = 1 n = 1 n = 0 n = 8 n = 16 n = Jun-2011 unidentified catch anemone frilled

JOCKA 0 n = 0 n = 5 n = 4 n = 8 n = 14 n = dogfish spiny Jul-2011 sea raven pollock rock n = 0 n = 2 n = 0 n = 8 n = 10 n = BLACK BEAUTY 4 n = Sep-2011 n = 3 n = n = 0 n = 4 n = n = 0 n = 14 n = AMERICA 0 n = Dec-2011 ; red d red ; 11 primary 11 n = 0 n = 1 n = 1 n = GUARDIAN 2 n = 1 n = 1 n = Feb-2012 ash 0.0 0.5 1.0 1.5 2.0 2.5 ed linesed are species medians. Points have n = 0 n = 14 n = redfish acadian ocean perch OLYMPIA 7 n = 1 n = 0 n =

discard species discard Jun-2011 plaice american (dab) lobster american skate barndoor wolffish atlantic wolffish n = 5 n = 14 n = n = 2 n = JOCKA 5 n = 3 n = Jul-2011 n = 5 n = n = 6 n = 4 n = BLACK 3 n = BEAUTY 2 n =

; zero ; Sep-2011 - n = 2 n = 12 n = n = 12 n = 0 n = 1 n = catch AMERICA Dec-2011

tows are are tows n = 5 n = 1 n = GUARDIAN 5 n = 0 n = 0 n = Feb-2012 Page Page 0.0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 2.5 15

jittered horizontally for clarity. Numbers of observations are also included for each species. each for included also are observations of Numbers clarity. for horizontally jittered d trip; red each medians for are dotsRed Figure Species lbs caught/total lbs kept 5 : 0.0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 2.5 12 secondary 12 for vessel and tow by ratios SBRM

OLYMPIA 0 n = 3 n = 0 n = 2 n = 2 n = 4 n = Jun-2011

JOCKA 0 n = 0 n = 4 n = 9 n = 7 n = 5 n = skate smooth skate thorny cod atlantic

Jul-2011 white hake haddock starfish n = 1 n = 2 n = 2 n = 2 n = 7 n = BLACK BEAUTY 5 n = Sep-2011 n = 2 n = n = 8 n = 7 n = n = 4 n = 5 n = AMERICA 0 n = Dec-2011 ash n = 8 n = 2 n = 1 n = GUARDIAN 2 n = 0 n = 1 n = Feb-2012 ed lines are species medians. Points have been flounder atlantic witch (gray sole) n = 0 n = 0 n = 3 n = OLYMPIA 0 n = 1 n = 3 n = Jun-2011

discard species; z species; discard hake silver (whiting) squid short-finned squid hake atlantic red flounder fourspot n = 2 n = 4 n = 0 n = cucumber sea 0 n = JOCKA 0 n = 0 n = Jul-2011 n = 5 n = n = 7 n = 3 n = 3 n = BLACK 0 n = BEAUTY 0 n = Sep-2011 ero tows are deleted. n = 6 n = 0 n = 5 n = AMERICA 0 n = 0 n = 7 n = Dec-2011 n = 0 n = 3 n = 1 n = GUARDIAN 3 n = 0 n = 7 n = Feb-2012 Page Page 0.0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 2.5

16

spe each for included also are observations of Numbers clarity. for horizontally jittered a dotsRed Figure Species lbs caught/total lbs kept 6 : SBRM ratio : SBRM 0.0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 2.5 re medians for each trip; red dash trip; red each medians for re

OLYMPIA 0 n = 0 n = 0 n = 0 n = 0 n = 0 n = Jun-2011 discard species;zero tows tertiary are deleted. 12 for vessel and tow by s

JOCKA 0 n = 0 n = 0 n = 1 n = 0 n = 1 n = scad round

Jul-2011 anemones skate little skate monkfish lumpfish alewife n = 1 n = 0 n = 1 n = 0 n = 0 n = BLACK BEAUTY 0 n = Sep-2011 n = 0 n = n = 0 n = 0 n = n = 0 n = 0 n = AMERICA 0 n = Dec-2011 n = 2 n = 1 n = 0 n = GUARDIAN 0 n = 1 n = 1 n = Feb-2012 ed lines are species medians. Points have been n = 0 n = 0 n = 1 n = OLYMPIA 0 n = 0 n = 2 n = Jun-2011 crab northern stone shad american herring atlantic n = 2 n = 0 n = 4 n = n = 2 n = JOCKA 1 n = 0 n = sponge finger

Jul-2011 skate winter cunner n = 2 n = n = 0 n = 0 n = 0 n = BLACK 0 n = BEAUTY 0 n = Sep-2011 n = 0 n = 1 n = 3 n = AMERICA 0 n = 3 n = 3 n = Dec-2011 n = 0 n = 2 n = 3 n = GUARDIAN 0 n = 0 n = 1 n = Feb-2012 cies. Page Page 0.0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 1.5 2.0 2.5

17

redfish acadian ocean perch

2.5

2.0

1.5

1.0

0.5

0.0 pollock

2.5

2.0

Standard Bycatch Reporting Method Reporting Bycatch Standard lbs kept species lbs caught/total 1.5

1.0

0.5

0.0 dogfish spiny

2.5

2.0

1.5

1.0

0.5

0.0

80 90 100 110 120 130 140 Depth (fm) loess smoother, span=0.3, degree = 2 Figure 7: SBRM ratio over depth for all vessels combined, with trends depicted by loess smoother. The three species plotted had the highest overall amounts of discards.

Depth is reported by network members to be an important factor in redfish size. The effect of depth on SBRM ratio for the three major discard species was assessed visually using loess smoother (Figure 7), and pointed at depth as playing a possible role in catch of small redfish. Redfish discard patterns were also examined for temporal and vessel patterns. Discard weights of redfish were higher in the July and September trips (Figure 8), and highest at depths between 84 and 107 fm (Figure 9).

Page 18

70000

60000

50000

40000

POUNDS 30000

20000

10000

0

FEB DEC MAY JULY JUNE SEPT

DISCARD KEPT

Figure 8: Kept and discarded redfish (lb) by month fished.

30000

25000

20000

15000 POUNDS 10000

5000

0 80 84 87 92 94 96 100 102 104 107 109 111 113 115 117 123 125 128 136 DEPTH

DISCARD KEPT

Figure 9: Kept and discarded redfish (lb) by depth fished (fm).

Page 19 The length frequencies of the redfish catch were plotted for each trip. These frequencies are based on a sample of 100 fish per tow and then summed over the five- day trip to represent the total catch distribution (Figure 10).

GUARDIAN Feb-2012 15 10 5 0 AMERICA Dec-2011 15 10 5 0 BLACK BEAUTY Sep-2011 15 10 Percent of Total of Percent 5 0 JOCKA Jul-2011 15 10 5 0 OLYMPIA Jun-2011 15 10 5 0 10 20 30 40 Total length (cm)

Figure 10: Redfish length frequencies raised to the total catch and summed over each trip. Red dashed line is minimum landing size.

The effect of depth on the size of redfish was further examined through histograms of length distributions in relation to the depth fished and the minimum landing size (Figure 11). Results support the prior observation that shallower tows included greater amounts of smaller redfish, particularly those less than 90 fm.

Page 20

(80,90] 15 10 5 0 (90,100] 15 10 5 0 (100,110] 15 10 5 0 Percent of Total of Percent (110,120] 15 10 5 0 (120,130] 15 10 5 0 (130,140] 15 10 5 0 10 20 30 40 Total length (cm)

Figure 11: Histograms of redfish length frequencies by depth interval (fm). The red dashed line is the minimum landing size.

All tows less than 90 fm (n=9) occurred in the July 2011 trip. Length distributions in this depth range were more closely inspected (Figure 12). Seven of nine tows caught redfish less than the minimum landing size. The estimated total count of sublegal redfish was approximately 3900 individuals in these tows.

Page 21 1000 800 600 400 200 0 1000 800 600 400 200 0 1000 800 600 400 200 0 1000 800 600 400 200 0 1000 800 Estimated Count Estimated 600 400 200 0 1000 800 600 400 200 0 1000 800 600 400 200 0 1000 800 600 400 200 0 1000 800 600 400 200 0

10 20 30 40 Total length (cm)

Figure 12: Histograms (estimated counts) of length distributions of redfish for tows less than 90 fm;each panel represents one tow. The red dashed line is the minimum landing size.

The ratio of redfish caught to multispecies caught (both kept and discarded) was plotted spatially, showing both depth and physical location (Figure 8).

Page 22

Figure 13: The amount and proportion of redfish and other multispecies catch by tow location. The size of circle indicates the amount of catch. The black rectangle is Western Gulf of Maine Closed Area.

Discussion

All five exploratory fishing trips resulted in substantial catches of redfish using a 4.5” mesh codend with low levels of incidental catch or bycatch of regulated species. Vessels conducted successful operations over a range of seasons and depths, and trawl designs. The spatial distribution of the exploratory fishing recreated to a large extent the geographic spread of the historic fishery.

Variation was apparent in catch per tow for redfish, including tows made up exclusively of legal-sized redfish and multiple tows above 10,000 lb. Other tows resulted in no redfish. This variation reflects the patchy nature of the species, attempts by network captains to explore a wider geographical range, and natural variation. Further sources of variation are likely due to gear, skill, seasonal differences and randomness individually or in combination. Catch per unit of effort was not closely examined as tow length or time was not driven exclusively by commercial considerations – vessels in some cases

Page 23 husbanded hold space, or avoided excessive deck loading to facilitate catch sampling by observers.

Commercially important species such as pollock, white hake, and haddock of legal size were caught and landed without substantial discards, despite the smaller mesh size. Pollock is the most significant incidental catch, as was anticipated based on historic knowledge of mixing of the two species and other redfish fisheries (e.g. Canada). Two tows resulted in relatively large numbers of sublegal pollock (~ 500 and ~1500 lb respectively).

The primary discard, spiny dogfish, was a regulatory discard, and not due to catch of undersized fish in the small mesh codends. This species has large commercial quotas, suggesting that the use of small mesh for redfish would not likely significantly impact this species. Also, commercial operations typically move away from high dogfish concentrations as large dogfish catches demand intense labor, dogfish are low value especially on multi-day trips, and dogfish reduce quality through direct abrasion. As spiny dogfish populations increase, consideration may be given to a grid system that could eliminate dogfish, similar to the device tested in the silver hake fishery (Chosid et al. 2011).

It is clear from the length frequency distributions that catches of undersized redfish are possible with 4.5” diamond mesh. Depth, time-of-year, or some other factors appear to have had an impact on catches of undersized redfish; the patterns we observed in catches of undersized redfish could not be isolated to area, time, vessel, net or depth, in part because many of these factors were confounded by design. Further, the lower proportion of undersized redfish in deeper tows may be related to the speed or time for haulback. It is possible that deeper tows allow more time for small redfish to escape slacker meshes in the codend.

Refinement of the size of redfish caught may be achieved through codend mesh size adjustment. A trouser trawl codend selectivity study, the third component of REDNET, will provide guidance to fishermen, netmakers, processors, and regulators on retention likelihoods and sizes of redfish for different codend mesh sizes. The appropriate size of redfish to be landed should involve consideration of biological characteristics of redfish (length-at-age, maturity-at-length, fecundity), market preferences, and commercial preferences, with an overall goal of sustainability of the fishery.

A fourth component, bycatch reduction may be useful for development of exclusionary grids for dogfish, or testing of a size-sorting grid to reduce undersized redfish. A grid system may allow smaller redfish to escape on-bottom with higher survival rates than through codend meshes at the surface. The importance of this fourth component will be clarified by the results from the size selectivity study.

Page 24 The data collected through this effort show promise and suggest a targeted redfish fishery could be successful using a small mesh codend, without large catches of undersized regulated species.

Acknowledgements

Primary thanks are due to the participating vessel owners, captains, and crews for their wisdom, patience and participation, and for the use of their redfish quota. Our thanks also go to biologists who collected data onboard from AIS, Inc. and SMAST, and to the staff of the Division of Marine Fisheries for contract and other support. We acknowledge the extensive involvement of the REDNET network, which at present includes over 30 members. Steve Cadrin of SMAST contributed to the SBRM calculations and analysis. Funding was provided by the National Marine Fisheries Service Cooperative Research Partnership Program.

References

Chosid, D.M., M. Pol, M. Szymanski, F. Mirarchi, and A. Mirarchi. 2012. Development and observations of a spiny dogfish Squalus acanthias reduction device in a raised footrope silver hake Merluccius bilinearis trawl. Fisheries Research 114: 66-75. Fonteyne, R. 2005. Protocol for the Use of an Objective Mesh Gauge for Scientific Purposes. ICES Cooperative Research Report, (279). R Development Core Team, 2009. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, http://www.R- project.org. Sarkar, D., 2009. Lattice: Lattice Graphics. R package version 0.17-26, http://CRAN.Rproject.org/package=lattice. Wigley S.E., P.J. Rago, K.A. Sosebee and D.L. Palka. 2007. The Analytic Component to the Standardized Bycatch Reporting Methodology Omnibus Amendment: Sampling Design, and Estimation of Precision and Accuracy (2nd Edition). NEFSC Ref. Doc. 07- 09.

Additional Material

Attachment A: Review of Existing Information on Redfish Occurrence and Catch Rates Attachment B: Experimental Fishing Permit Attachment C: Additional Data

Page 25 Attachment A: Review of Existing Information on Redfish Occurrence and Catch Rates

Page 26 Attachment B: Experimental Fishing Permit

Page 27 Attachment C: Additional Data

Table 5: Net characteristics for participating vessels.

Vessel

Olympia Jocka Black Beauty America Guardian 360 Balloon 2-seam Separator 2-seam 2-seam Net Name trawl Balloon trawl trawl Balloon trawl Balloon trawl Door weight 750 kg 640 kg 750 kg 630 kg Headline length 60' 60' 103' 90' 96' Number of floats 80 100 45 71 20 Float diameter 8" 6" 12" 8" 11" Footrope length 80' 80' 119' 130' 120' Footrope type Rockhopper Rockhopper Rockhopper Rockhopper Rockhopper Rockhopper size 12" 16" 18" 18" 12" Ground cable length 30 fm 15 fm 40 fm 43 fm 75 fm Ground cable type Cookie Cookie Cookie Cookie Cookie Bridle length 15 fm 15 fm 30 fm 10 fm 15 fm Bridle (lower/upper) Cookie/wire Cookie/wire Cookie/wire Cookie/wire Cookie/wire Codend type Double Double Double Double Double Codend mesh size Nominal 4.5” 4.5” 4.5” 4.5” 4.5” Actual (Codend 1) 4.34” 4.34” 4.34” 4.34” 4.34” Actual (Codend 2) 4.38” 4.38” 4.38” 4.38” 4.38”

Page 28

redfish acadian ocean perch dogfish spiny 2000 4000

1500 3000

1000 2000

500 1000

0 0 rock sea raven pollock 50 150

40 1000 100 30 500 20 50

10 0 0 lobster american squid short-finned haddock 25 30 30

20 25 25 20 20 15 15 15 10 10 10 5 5 5 0 0 hake white skate barndoor hake silver (whiting) 25 15 20 20

10 15 15 10 10 5 5 5 0 0 OLYMPIA JOCKA BLACK BEAUTYAMERICA GUARDIAN OLYMPIA JOCKA BLACK BEAUTYAMERICA GUARDIAN OLYMPIA JOCKA BLACK BEAUTYAMERICA GUARDIAN Jun-2011 Jul-2011 Sep-2011 Dec-2011 Feb-2012 Jun-2011 Jul-2011 Sep-2011 Dec-2011 Feb-2012 Jun-2011 Jul-2011 Sep-2011 Dec-2011 Feb-2012

Figure 14: Catch (lb) by tow and vessel for each discard species. Red dots are medians for each trip; red dotted lines are species medians. Points have been jittered horizontally for clarity. Scales for weights vary by panel.

Page 29 hake atlantic red wolffish atlantic cod atlantic 15 15 15

10 10 10

5 5 5

0 0 skate thorny skate winter unidentified catch 10 15 14 8 12 10 6 10 4 8 5 2 6 0 0 plaice american (dab) shad american skate smooth 3.0

2.5 6 6 2.0 4 1.5 4 1.0 2 2 0.5 0 0 anemones cucumber sea anemone frilled 2.4 2.6 2.5

2.2 2.4 2.0

2.0 2.2 1.5

1.8 2.0 1.0

1.6 1.8 0.5 OLYMPIA JOCKA BLACK BEAUTYAMERICA GUARDIAN OLYMPIA JOCKA BLACK BEAUTYAMERICA GUARDIAN OLYMPIA JOCKA BLACK BEAUTYAMERICA GUARDIAN Jun-2011 Jul-2011 Sep-2011 Dec-2011 Feb-2012 Jun-2011 Jul-2011 Sep-2011 Dec-2011 Feb-2012 Jun-2011 Jul-2011 Sep-2011 Dec-2011 Feb-2012

Figure 15: Catch (lb) by tow and vessel for each discard species. Red dots are medians for each trip; red dotted lines are species medians. Points have been jittered horizontally for clarity. Scales for weights vary by panel.

Page 30 monkfish skate little crab northern stone

1.4 1.8 1.5 1.2 1.6 1.0 1.0 1.4 0.8 1.2 0.6 0.5

sponge finger flounder atlantic witch (gray sole) lumpfish

1.4 1.4 1.8 1.2 1.2 1.0 1.6 1.0 0.8 1.4 0.8 0.6 1.2 0.6 0.4 flounder fourspot scad round cunner 0.8 1.2 1.0 0.9 0.7 1.0 0.8 0.6 0.8 0.7 0.5 0.6 0.6

0.4 0.4 0.5 alewife herring atlantic starfish 0.8 0.40 0.50

0.6 0.35 0.45

0.4 0.30 0.40

0.2 0.25 0.35

0.0 0.20 0.30 OLYMPIA JOCKA BLACK BEAUTYAMERICA GUARDIAN OLYMPIA JOCKA BLACK BEAUTYAMERICA GUARDIAN OLYMPIA JOCKA BLACK BEAUTYAMERICA GUARDIAN Jun-2011 Jul-2011 Sep-2011 Dec-2011 Feb-2012 Jun-2011 Jul-2011 Sep-2011 Dec-2011 Feb-2012 Jun-2011 Jul-2011 Sep-2011 Dec-2011 Feb-2012

Figure 16: Catch (lb) by tow and vessel for each discard species. Red dots are medians for each trip; red dotted lines are species medians. Points have been jittered horizontally for clarity. Scales for weights vary by panel.

Page 31 hake white pollock redfish acadian ocean perch 1.0 0.8 0.6 0.4 0.2 0.0 haddock cod atlantic hake silver (whiting) 1.0 0.8 0.6 0.4 0.2 0.0 skate winter monkfish cusk 1.0 0.8 0.6 0.4 0.2 0.0 rosefish black bellied lobster american flounder atlantic witch (gray sole) 1.0 0.8 0.6 0.4 0.2 0.0 squid short-finned hake atlantic red plaice american (dab) 1.0 0.8 0.6 0.4 0.2 0.0 OLYMPIAJOCKA BLACK BEAUTYAMERICAGUARDIANOLYMPIAJOCKA BLACK BEAUTYAMERICAGUARDIANOLYMPIAJOCKA BLACK BEAUTYAMERICAGUARDIAN Jun-2011 Jul-2011 Sep-2011 Dec-2011 Feb-2012 Jun-2011 Jul-2011 Sep-2011 Dec-2011 Feb-2012 Jun-2011 Jul-2011 Sep-2011 Dec-2011 Feb-2012

Figure 17: SBRM ratios by tow and vessel for all kept species; zero tows are deleted. Red dots are medians for each trip; red dotted lines are species medians. Points have been jittered horizontally for clarity.

Page 32

Figure 18: Length frequencies of all species with 5 or greater length measurements, pooled for the entire component. Red dotted lines, where present, are the minimum landing sizes. Note changes in scales.

Page 33 Response to NEFSC Review Comments (August 1, 2014) on REDNET Component 2 - Exploratory Fishing Final Report Thank you for the review of this report. Our actions in response to each comment are detailed below. Comment 1: On page 18 they state that the discarded weights of redfish were higher in July and August trips and at depths between 94-107 fm. The data and figures show July and September (No August trips) and the discard rate was highest in approximately 84-107 (not 94-107). Response: We have replaced “August” with “September” and “94” with “84” in all places: the Executive Summary and on page 18.

Comment 2: In many of the plots, the Y-axis values are covered by the figures.

The figures became distorted during conversion from Word to Adobe.

Comment 3: If I look at the length frequency histograms from the first draft report (Figure 4) and compare to this one report (Figure 10), there are large differences, especially for February and December. Which one is correct?

Response: The length frequency histograms are correct in the reviewed report. The length frequency histograms from the preliminary report were incorrect.

Comment 4: On Figure 10 it lists trips in June, July, September, December and February yet Figure 8 and the text the dates list a trip in May and not in June. I would suggest that they make them so they are consistent. I see from the data that the trip started on 5/31and ended on 6/5.

The text was revised in one place to state that testing began at the end of May. Three figures have been revised to be consistent in referring to the trip that started on 5/31 as “June”.

Page 34 REDNET - A Network to Redevelop a Sustainable Redfish (Sebastes fasciatus) Trawl Fishery in the Gulf of Maine

FINAL REPORT Component 3 – Codend Selectivity

Revision 2: 10 November 2014 Revision 1: 28 July 2014 Original: 13 December 2013

Written by: Michael Pol1 and Pingguo He2

1. Massachusetts Division of Marine Fisheries 2. UMass Dartmouth School for Marine Science and Technology

Funded by: Northeast Cooperative Research Partners Program Northeast Fisheries Science Center National Marine Fisheries Service

Revision Notes: Revision 1 was produced in response to a first round of comments provided by the NOAA Fisheries Northeast Cooperative Research Partners Program (CRPP) on 9 July 2014. Revision 1 was reviewed and accepted by the NEFMC Research Steering Committee (RSC) on 8 August 2014. Revision 2 responds to minor issues in Revision 1 reported by the CRPP, and to remove title page restrictions on use of the document as presentation of it to the RSC made it a public document.

Executive Summary

Acadian redfish (Sebastes fasciatus) is among a few groundfish stocks in the northeast US that have a relatively large Annual Catch Limit (ACL) that have not been fully utilized during the last several years. With reduced ACLs for many important groundfish species, full and sustainable utilization of the redfish resource is thus critical to the economic survival of fishermen and to the success of groundfish sectors. The allocations of redfish could not be fully utilized under current regulations because the minimum codend mesh size is too large to effectively retain redfish. Beginning in 2010, the REDNET project seeks to achieve three fishery conservation and management goals:

1) Redirecting fishing effort in the multispecies fishery away from stocks that are overfished to stocks that are considered rebuilt (e.g. redfish). 2) Achieving optimum yield, by increasing commercial landings of redfish through development of a directed fishery under the adaptive management ability of groundfish sectors. 3) Increasing the economic viability of groundfish sectors by providing access to the ACL of a recovered species and thus generating much‐needed revenue for the industry.

REDNET includes six components:

Component 1: Network Meetings Component 2: Baseline Catch and Bycatch Evaluation Component 3: Codend Selectivity Component 4: Conservation Engineering and Bycatch Reduction Component 5: Processing/Marketing Component 6: Outreach/Implementation

This report describes work carried out in Component 3 – Codend Selectivity. The final report for Component 2 has been submitted.

A trouser trawl was used to determine the size selectivity of three sizes of mesh opening (4.5, 5.5 and 6.5 in double 5 mm twine diamond) on a commercial fishing vessel. Fishing off Provincetown, Massachusetts, 56 tows were completed in March and April 2013,

REDNET Component 3 Draft Final Report – Codend Selectivity Page 2

catching over 42,000 kg of redfish and about 6,000 kg of other species. Adequate length frequencies of redfish and pollock (Pollachius virens) were collected to produce selectivity models; only redfish results are reported here. Neither species has been the subject of a trawl selectivity study in the Northeast US before.

Robust models for the mean L50s and selection ranges, and confidence intervals, were developed for all three tested codends, incorporating both within and between haul variability. All measures of model validity were positive. These models are fully adequate to provide guidance to managers and fishermen on size retention of redfish and appropriate codend mesh size.

L50 and selection ranges were determined for 4.5 in (L50: 22.3 cm (8.8 in); SR: 4.5 cm),

5.5 in (L50: 29.2 cm (11.5 in); SR: 4.4 cm), and 6.5 in (L50: 33.6 cm (13.2 in); SR: 5.1 cm) codends. Simulation of fishing of the three tested codends on the observed population indicated that substantial escape of redfish through codend meshes occurs (48-94%), suggesting that investigation of escape of redfish is warranted to support a sustainable fishery. The observed population also indicates that inadequate numbers of larger redfish may be available to support a higher-priced market.

Introduction

Acadian redfish (Sebastes fasciatus – “redfish”) is one of very few groundfish species in the northeast US with a relatively large Annual Catch Limit (ACL) but has not been fully utilized by the current multispecies fishery participants during the last several years. The requirement for codends with large mesh sizes (6.5”) and a lack of market may be the primary reason for the underutilization. The implementation of a sector management system with catch retention rules and accountability measures makes it possible to target redfish with codend mesh sizes smaller than the current regulated mesh size. The Maine Department of Marine Resources, the Massachusetts Division of Marine Fisheries and the University of Massachusetts School for Marine Science and Technology joined with other members of the scientific community and the industry have developed a research plan that draws on wide‐ranging expertise in order to conduct comprehensive research on the development of a sustainable redfish trawl fishery in the Gulf of Maine. The project, called “REDNET”, includes components of catch and bycatch assessment REDNET Component 3 Draft Final Report – Codend Selectivity Page 3

through exploratory fishing, evaluation of codend mesh size, testing of size and species selective design and devices, processing and marketing, outreach, and implementation. The project is funded by NOAA Fisheries Northeast Cooperative Research Partners Program’s 2010 funding cycle.

The redfish cooperative research project or “REDNET” seeks to achieve three fishery conservation and management goals:

• Redirecting fishing effort in the multispecies fishery away from stocks that are overfished to stocks that are considered rebuilt (e.g. redfish).

• Achieving optimum yield, by increasing commercial landings of redfish through development of a directed fishery under the adaptive management ability of groundfish sectors.

• Increasing the economic viability of groundfish sectors by providing access to the ACL of a recovered species and thus generating much‐needed revenue for the industry.

Groundfish management in the Northeast made a dramatic shift in May of 2010 from primary input controls to primary output controls. Amendment 16 to the Northeast Multispecies Fishery Management Plan (FMP) established the rules for sector management, as well as catch limits and accountability measures mandated by the 2006 reauthorization of the Magnuson‐Stevens Act. If multispecies sector management is to be successful, the fleet must be able to catch and land stocks of fish with high abundance such as redfish while exercising their ability to avoid quota-limiting species (e.g. most recently, Gulf of Maine cod).

Historically, redfish represented a significant fishery and income in the region. The directed redfish fishery in the northeast began in the 1930s and total landings rose from 100 metric tons (mt) to a peak of over 117,000 mt in 1951 and then steadily declined. By 1983, the total US landings of redfish were 5,328 mt, and in 2008 landings were only 1,189 mt. The current best available estimates assert that the resource can support a larger fishery.

REDNET Component 3 Draft Final Report – Codend Selectivity Page 4

The redfish fishery in the Gulf of Maine was traditionally prosecuted by vessels using otter trawls with relatively small mesh codends in the range of 70 ‐ 80 mm (2.5 ‐ 3 in). Since 1950s and through 60s and 70s, the population of redfish declined, likely as a result of overexploitation. In 1977, the minimum codend mesh size increased from 114 to 130 mm (4.5 to 5 in) and increased again in 1994 to 152 mm (6 in). Today the minimum codend mesh size mandated by the Multispecies FMP is 165 mm (6.5 in), though groundfish sector vessels may use 152 mm (6”) codends provided an observer has been assigned to the fishing trip. These mesh restrictions, combined with low biomass levels between 1980 and 1995 eliminated a directed redfish fishery in the Northeastern United States.

In recent years, the combined restrictions in the multispecies FMP have resulted in the recovered status of the redfish resource. A stock assessment of redfish was completed and reviewed at the 2008 Groundfish Assessment Review Meeting (GARM III) and updated through 2010 at a 2012 review. The assessments and reviews concluded that the redfish stock is not overfished and overfishing is not occurring.

Development of a redfish fishery in the Northeastern United States is critical to the economic survival of fishermen and to the success of groundfish sectors. Sectors were implemented through Amendment 16 to the multispecies FMP. The sectors are assigned percent allocations based on the historic catches of their members. However, the allocations of redfish could not be fully utilized under current regulations because the minimum codend mesh size is too large to effectively retain redfish. At the same time, allocations of other groundfish species are extremely small due to the very low ACLs recommended. These limiting allocations (“choke” species) will affect the fishing behavior of the sectors and potentially shut them down before more abundant allocations are realized. It is critically important to the success of sectors to find a way to allow them to access allocations of healthy stocks such as redfish while avoiding those that are depressed.

Size selection is the process that describes why the size distribution of fish caught by a fishing gear is different than the size distribution of a fish population. The concept behind size selectivity in trawls is that the retention rate of fish across lengths can be determined through a comparison of the numbers of fish caught per length in the test

REDNET Component 3 Draft Final Report – Codend Selectivity Page 5

codend and the numbers per length in a non-selective (smaller mesh) control codend. Since the relationship is largely mechanical (what size fish can fit through a certain-sized hole?), the resulting relationship is considered robust, stable, and adequate for

management purposes. Typically, the L50 (the length at which 50% of fish in the trawl are retained by the codend) is used to characterize the codend performance, and can be used to set codend size limits. Other metrics (L25, for example) can also be input into the model or used as management benchmarks.

The methods, statistical model and methodology to describe this relationship have evolved, but have become fairly standardized in part due to increased computing speed and power. Trawl hauls are made simultaneously (parallel, twin, or trouser trawls) or alternatively with the test codend and the control codend with the aim of sampling the same population (one of identical or similar sizes of fish). A logistic curve that implies

increasing retention, reaching a maximum of 100% (L100) is fit to individual hauls (if possible) or pooled data if not. The between and within haul variation can then be combined to produce an average curve, with error estimates or confidence limits. During this process the random effects of other factors on the size selection can be assessed. Additionally, over the years a body of research has been developed that has identified certain codend characteristics and other factors that can alter the size selection of the same mesh opening.

Size selectivity specifically for redfish species in codends has been examined across the North Atlantic since 1961 with varying levels of rigor. A recent review of this topic (Herrmann et al. 2012) found 21 trials of codend meshes, mostly diamond, mostly for redfish relatives Sebastes marinus and S. mentella, with only three testing S. fasciatus, our species of interest, and was combined with S. mentella. Prior to this study, we lacked vital information necessary to redevelop a sustainable trawl fishery because managers and fishermen would not be able to reliably predict the effect of different codend mesh sizes on commercial catches and associated discards of small redfish and other groundfish species.

REDNET Component 3 Draft Final Report – Codend Selectivity Page 6

Project Design

REDNET is a multifaceted, comprehensive project to determine how to sustainably access the redfish ACL for the groundfish sectors in the northeast US. The goal is to execute a complete project, from conception to regulatory implementation, to marketing, that supports environmentally and economically sustainable harvesting of redfish. REDNET includes individuals with various expertise to accomplish this vision. There are six components of the project:

Component 1: Network Meetings Component 2: Baseline Catch and Bycatch Evaluation Component 3: Codend Selectivity Component 4: Conservation Engineering and Bycatch Reduction Component 5: Processing/Marketing Component 6: Outreach/Implementation

This report describes experimental methods and results from Component 3: Codend selectivity. While the entire REDNET project is still ongoing, Components 2 and 3 (this report) are complete. This report should be considered the final report on the data gathered for this component.

Methods

Fishing vessel and gear

After a competitive bid process among REDNET active industry participants, the F/V Guardian (80 ft LOA; 425 hp) was chosen to conduct the research. An Experimental Fishing Permit was received in December 2012. Candidate mesh sizes of 4.5, 5.5 and 6.5 in diamond mesh codends were selected by REDNET members This range incorporates the current minimum mesh size and the likely smallest acceptable mesh size for a special access program, with one mesh at the midpoint. These sizes are also readily available commercially. Mesh sizes were measured prior to and after the

REDNET Component 3 Draft Final Report – Codend Selectivity Page 7

experiment using an ICES OMEGA mesh gauge and associated protocols (Fonteyne, 2005). Square mesh codends were dismissed based on impracticality due to sticking of fish, as advised by the industry members of the REDNET and researchers of ICES-FAO Working Group on Fishing Technology and Fish Behavior (WGFTFB). Sticking results when fish pass partway through meshes, and are gilled in the codend. A large number of fish sticking in the codend requires a great deal of time removing them from the netting, some of which became damaged and unsalable. Two six-day trips were made during March-April 2013 with the goal of comparing the sizes of fish caught in 3 different commercial-sized codends to the catch in a 2.5 in control codend using a trouser trawl. Fishing trials were carried out in southern Gulf of Maine (Figure 1).

The participating vessel provided a balloon trawl front end (ground gear, wings, net mouth) to be attached to a “trouser trawl”; both were designed by Tor Bendiksen of Reidar’s Manufacturing (now Reidars Trawl Gear and Marine Supply, New Bedford, MA) (Figure 2). The headline of the trawl was 109 ft 10 in (33.4 m) in length with 100 plastic floats 8 in (20.3 cm) in diameter. The footrope was 138 ft 6 in (42.5 m) in length and attached with a rockhopper sweep. The front end of the net was uniformly 6.0 in (152 mm) mesh openings constructed of 4.0 mm diameter braided twine. The fishing circle was 190 meshes across the bottom panel and 240 meshes across the top. The trouser part of the trawl was also constructed of 6.0 in (152 mm) mesh size, 3.6 mm diameter braided twine. It was designed with a 47.5 meshes deep common “mixing area” that was then separated uniformly into two lateral equal circumference legs (130 meshes across the bottom; 161 meshes across the top) (Figure 2). A test codend (4.5, 5.5, or 6.5 in double 5 mm twine) was attached to one leg of the trousers and a 2.5 in diamond mesh control codend (double 4 mm) on the other side (Table 1). The side of the test and control codends was switched regularly to avoid possible side-based effects. The number of meshes for each test codend was adjusted so that the same diameter and overall length were maintained for all codends. One leg of the trouser trawl was lengthened by 25 meshes of double 4 mm 6.5 in mesh to avoid contact or inhibition between codends. A GoPro Hero2 high-definition camera (San Mateo, CA) with a deepwater underwater housing and lights was mounted to view fish reaction in the mixing area during some tows.

Onboard sampling

REDNET Component 3 Draft Final Report – Codend Selectivity Page 8

All regulated groundfish were counted against the sector annual catch limits (ACLs). The vessel was reimbursed for any quota purchases. Net revenue from sale of redfish was divided between the vessel and the project. The project also reimbursed the participating vessels for fuel in order to encourage the vessel to fish at different locations throughout the Gulf of Maine. Other expenses incurred by the scientific crew (insurance, food) were also reimbursed to the vessel.

Two or three scientists were on board the vessels and worked alongside the vessel crew during both trips and documented all trip, tow, environmental, catch, and bycatch data following National Marine Fisheries Service (NMFS) Fisheries Observer Program protocols. The total catch of redfish per tow was determined (on some tows, legal and sublegal catch amounts were quantified). Other organisms were also identified, and weighed; weights (kg) were directly collected or quantitatively determined; for example, by basket counts. A special protocol was agreed upon between the NMFS habitat group and the project participants in the event that deep-sea corals were encountered.

Codends were hauled on deck one at a time, with the codend attached to the shorter “leg” hauled first. Catches from the experimental and control codends were deposited in separate areas on deck, and processed separately. The same test codend was used for approx. three days before switching to a new mesh size.

Lengths (measured as midline length, MLL) of a random subsample of more than 100 redfish from each codend from each tow were commonly collected. Collection of pollock lengths was also prioritized. Other important species were measured opportunistically. For length-frequency (LF) analysis, counts at each length were multiplied by the subsample weight divided into the total weight.

Net geometry was measured using a trawl monitoring system (Notus Electronics, St. John’s, Newfoundland) with sensors on both doors, the trawl’s wing ends, just behind the headrope, and on the 2.5-in codend. The sensors were set to provide bottom temperature, door spread, door heel (angle of the door to the right or left of the direction of travel), wing spread and to indicate when the control codend was full. In addition, these sensors can provide distance from the sensor to the hydrophone.

REDNET Component 3 Draft Final Report – Codend Selectivity Page 9

Bottom temperature was also recorded with previously calibrated TidBit temperature recorders (Onset Computers, Inc., Pocasset, Massachusetts).

Analysis

All catch data (along with trip and gear data) were entered and uploaded into a customized relational database in Microsoft Access 2007. Collected data were analyzed using Microsoft Excel and R statistical software (R Development Core Team, 2009), primarily using the lattice package (Sarkar, 2009) and SELNET, a selectivity analysis program. SELNET was developed to acquire and analyze size selectivity and catch data for towed fishing gears, both at the haul level and for a group of hauls (Frandsen et al., 2011; Herrmann et al., 2012; Herrmann et al., 2013). The methods implemented in SELNET comply with accepted recommendations for the analysis of size selectivity data (Wileman et al. 1996; Fryer 1991).

To model the size selection first we used a logistic curve described by the parameters L50

and the selection range SR (= L75 – L25) (Wileman et al., 1996). For each haul, the

number of fish counted in the experimental codend is described as ntl for the count of

fish at each length l, and in the control codend ncl. The proportion of the total catch actually measured for lengths is described by the sampling rates qt (experimental) and qc (control). The size selection in each haul can then be obtained by minimizing the following function with respect to the parameters L50, SR and SP:

× ( , 50, , ) × + × ( , 50, , ) + × 1 ( , 50, , ) 푙 푞푡 휑 푙 퐿 푆푅 푆푃 푙 − � �푛푡 푙푛 � × 1 ( , 50, , ) � 푛푐 푙 × 푞푡 휑 푙 퐿 푆푅 푆푃 푞푐 � − 휑 푙 퐿 푆푅 푆푃 � × ( , 50, , ) + × 1 ( , 50, , ) 푞푐 � − 휑 푙 퐿 푆푅 푆푃 � 푙푛 � �� 푞푡 휑 푙 퐿 푆푅 푆푃 푞푐 � − 휑 푙 퐿 푆푅 푆푃 � With:

× ( , 50, ) ( , 50, , ) = 푙표푔푖푡 ( ) 1 푠푝× 1푟 푙 퐿 , 푆푅50, 휑 푙 퐿 푆푅 푆푃 − 푠푝 � − 푟푙표푔푖푡 푙 퐿 푆푅 � REDNET Component 3 Draft Final Report – Codend Selectivity Page 10

SP is defined as the split parameter and expresses the assumed length independent relative entry of fish to the test or control side of the gear during the fishing process. SP needs to be estimated to assess the values of the selection parameters L50 and SR.

Fit statistics (i.e., the p-value and model deviance versus degrees of freedom (DOF)) were inspected for individual hauls (Wileman et al. 1996). Where the p-value < 0.05 or the deviance >> DOF, the residuals were examined for patterns or structural problems. Where no pattern was seen, the poor fit was considered overdispersion in the data and the data were included.

The second step considered between-haul variation (Fryer 1991) using the results from all the individual hauls simultaneously for the L50, SR and SP, together with their covariance matrix and information on the values of the mesh size, m. In addition, we considered the effect of w (total control codend catch weight at end of haul in kg) and S2, which side of the twin trawl the test codend was attached to. (The trouser trawl was constructed to be longer in one leg to avoid interference underwater between the codends. Since one codend stayed in the water longer, and could potentially lose more fish, we wanted to test whether this longer hauling time might impact size selectivity.)

A model considering the potential effect of the parameters m, w and S2 was constructed with the following form and applied in SELNET for each species separately.

50 = + × + × + × 2 = + × + × + × 2 퐿 푓0 푓1 푚 푓2 푤 푓3 푆 = + × + × + × 2 푆푅 푔0 푔1 푚 푔2 푤 푔3 푆

푆푃 ℎ0 ℎ1 푚 ℎ2 푤 ℎ3 푆 The parameters f0…f3, g0…g3 and h0…h3 are estimated while fitting the model to the data with values for L50, SR and SP based on the selectivity results from the individual hauls. Models were selected based on the AIC value (Akaike, 1974), while considering every possible simpler sub-model following the procedure described in Wienbeck et al. (2011) and Herrmann et al. (2013). A total of 4096 model runs were completed.

REDNET Component 3 Draft Final Report – Codend Selectivity Page 11

Individual haul results were plotted for the L50 and SR with 95% CI versus the mean model estimated values and the predicted 95% CI for the total variation (between-haul variation + uncertainty around the mean). The lower and upper 95% CI for the estimated between-haul variation in the selection parameters (lim L50, lim SR) were calculated by:

50 = 50 ± 1.96 × ( 50 + )

푚푒푎푛 ( 푚푒푎푛 ) 11 푙푖푚 퐿 = 퐿 ± 1.96 × � 푉푎푟퐿 + 퐷 (x) 푙푖푚 푆푅 =푆푅푚푒푎푛 ± 1.96 ×� 푉푎푟( 푆푅푚푒푎푛 +퐷22 )

푙푖푚 푆푃 푆푃푚푒푎푛 � 푉푎푟푆푃푚푒푎푛 퐷33 where L50mean , SRmean and SPmean are the predictions based on the selected submodel

based on (1), and D11 , D22 and D33 are the diagonal elements in the estimated between haul-variation matrix for the selected model (Fryer, 1991).

These plots were inspected to see if the model predictions appeared to reflect the main trends for the effects of catch size on the results for each codend to see if it was necessary to consider the estimates of the between-haul variation and uncertainty on the means in the selection process in addition to the uncertainty of the haul results.

After successful model validation based on the above procedure the models were applied to predict size selection for codend mesh sizes between 80 and 170 mm.

Results and Discussion

The codend mesh sizes measured prior to and following fishing using an OMEGA mesh gauge indicated a slight change during the experiment (Table 1). All test and control codends except the largest mesh codend (6.5”) showed some increase. The average of the two measurements was used for analysis.

Fifty-six tows were completed in two trips carried out between 27 March and 1 April, and 3 April and 8 April 2013. Overall, 18 tows were completed pairing the 4.5 in mesh,

REDNET Component 3 Draft Final Report – Codend Selectivity Page 12

with 10 on the starboard side; 16 tows with the 5.5 inch with 9 on the starboard side; 22 tows with the 6.5 in codend with 9 on the starboard side.

Tows were conducted generally east and northeast of Provincetown, Massachusetts over an area of approximately 4700 nm2 (Figure 1). Tow locations were based on captains’ previous knowledge, echo sounder signals (including bottom topography), and a goal of wide coverage and fish size mix. Tows were only made in daylight hours. Median tow duration was 0.6 h. Tow durations varied based on the captain’s assessment of the volume of fish in the net and fishing ground conditions, and were consistent with commercial practice. Duration decreased over time as the captain narrowed the search area and successfully found fish. Very large catches were avoided if possible due to catch processing delays that might reduce the number of tows and affect quality of fish retained and survival of fish escaped.

Median depth fished was 100 fm (Figure 3), deeper than is typical for other groundfish. Median towing speed was 3.0 knots. Median warp length (wire out) was 225 fm. The length of warp used was determined by the captain based on the water depth and the bottom topography of the tow track. Median wave height experienced was 4 ft with a maximum of 10 ft; these heights are unlikely to substantially affect net performance both based on the captain’s decisions and as they are within typical commercial operational conditions. The range of tow speed was also within normal operational conditions for the species and was mainly influenced by tidal conditions, as was typical in commercial operations.

Median bottom temperatures (°C) recorded by the TidBit recorder did not show differences between mesh sizes (4.5: 6.9°C (inter quartile range, IQR: 6.7-7.0°C); 5.5: 6.9 °C (IQR: 6.7-7.0°C); 6.5: 6.7 °C (IQR: 6.5-6.7°C). Median bottom temperatures as recorded by the Notus system were very similar to that from the Tidbit recorder (4.5: 7.0 °C (IQR: 6.7-7.3°C); 5.5: 6.7 °C (IQR: 6.6-6.8°C); 6.5: 6.8 °C (IQR: 6.4-7.0°C)). The small differences in median temperature are not likely to affect selectivity of redfish between codends.

Adequate trawl monitoring sensor readings were obtained to estimate distance from a hydrophone towed off the port stabilizer to both trawl doors, one wing end, the

REDNET Component 3 Draft Final Report – Codend Selectivity Page 13

headline and the codend (Figure 4). Additional data on door heel (angle of the door to the right or left of the direction of travel) (Figure 5), distance of the headline to the bottom, and some information on door spread (Figure 6) were collected. The distance measurements in Figure 6 are separated by wire out as the sensors are further away when there is more wire out. No anomalies are apparent; some filtering of clearly erroneous readings was used in the headline height plot. Additionally, since the split parameter indicated good net performance (see below) and both codends were on the same net, any differences in geometry would have minimal impact.

The lack of substantial variation in geometry, geographical area, weather conditions, depth, and temperature suggests that uncontrollable sources of variability were limited during the testing, and that the comparability of the codends was high. Additionally, since each tow consisted of a selective and a non-selective codend, the results are consistent and accurate within each tow.

The total catch of all species was just over 47,900 kg, with redfish comprising 42,482.9 kg or 89.7% of all catch (Table 2.) Pollock (Pollachius virens) was the main bycatch species (3390.5 kg), with 21 other species with catches greater than 10 kg total. One porbeagle shark was captured and then released with no reported weight.

Approximately 19,600 kg of redfish were sold, combined over both trips, at $0.50 per pound. Discards were not a primary focus during the study. They were quantified on some tows during the trip, and were primarily driven by a market that discourages the smaller legal-sized fish.

Over 18,000 redfish were measured. Lengths ranged from 13 to 40 cm with distributions differing between codends (Figure 7). Sufficient numbers of pollock were also caught to construct length frequencies and to conduct selectivity analyses. Over 1800 pollock were measured; sizes ranged from 18 to 86 cm (Figure 8). Size selectivity analysis for pollock will be reported separately.

Fifty-four hauls were included in redfish size selectivity analyses; two hauls could not be included due to zero catches of redfish in one of the codends (experimental or control).

REDNET Component 3 Draft Final Report – Codend Selectivity Page 14

Testing of the full model revealed a simplified version where the L50 and the SR depend only on mesh size (p>>0.001; AIC = 1467.35).

50 = 0.209 × (CI: 0.206, 0.216) = 0.0315 × esh size (CI: 0.029, 0.034) 퐿 푚푒푠ℎ 푠푖푧푒 = 0.556 (CI: 0.509, 0.602) 푆푅 푚

푆푃 The model estimated a split parameter near the ideal value of 0.5 with the 95% confidence interval (CI) overlapping 0.5. This parameter indicated that fish were equally likely to enter either codend and demonstrates that the trouser trawl was functioning properly. Wileman et al. (1996) recommend the use of a trouser trawl with a full vertical split, but our design avoids a vertical panel based on the gear designer’s recommendation and experiences in a shrimp trawl codend research by one of the authors (He and Balzano 2011). Vertical panels are very difficult or impossible to rig without causing deformity or distracting motion in the net and panel. Our design is similar to a pair of trousers, with a “mixing area” in the forward part of the net and no complex panel. Video recordings collected from this area also seemed to indicate no unusual fish behavior. The result of SP value indicates that this design can be considered an effective means of testing codend selectivity.

We originally considered use of a covered codend method. In consultation with industry partners and others in the network, it was felt that a trouser trawl avoided some deck- handling problems associated with covers, reduced risk of masking of codend meshes, and risk of over-filling the cover, since we did not know what size catches might result from using a 2.5 inch mesh. Few difficulties were encountered using the trouser trawl, as the crew was adaptable to the hauling of two codends, and operationally it was not difficult to retrieve, and empty two codends in a controlled and safe manner.

The lack of significant impact of catch size or which side the codends were on also provided evidence for good functioning of the gear and random distribution of fish

between the test and control codends. The absence of any intercept terms in the L50 and SR provided the logical result that as the mesh size is reduced to zero, the fish length would similarly be reduced to zero.

REDNET Component 3 Draft Final Report – Codend Selectivity Page 15

Full logistic curves that illustrate the catch curves for the codends as measured are shown in Figure 9, along with 95% confidence intervals. Also included in the figure are

the current and more recent minimum landing sizes (MLs). L50 and selection ranges

were determined for 4.5 in (L50: 22.3 cm (8.8 in); SR: 4.5 cm), 5.5 in (L50: 29.2 cm (11.5

in); SR: 4.4 cm), and 6.5 in (L50: 33.6 cm (13.2 in); SR: 5.1 cm) codends. Further validation of the selectivity model was demonstrated by plotting L50 and SR values for each individual haul, along with 95% confidence intervals as error bars, along with overall mean values predicted by the final model, against the size of the codend catch (Figure

10). Only six hauls were found to have values of L50 outside the 95% CIs; all of these hauls overlapped their error bars with the confidence band indicating that the model is an excellent fit to the data.

A similar comparison was made for the selection range (Figure 11); eight individual hauls had SR values outside the overall CI for the mean SR. All of these also had error bars overlapping the CI band. These combined results indicated that the fit of the individual hauls to the overall results were excellent.

Model results were used to produce estimates of mean L50 (Figure 12) and SR (Figure 13) across a broad range of mesh sizes to support the choice of appropriate mesh sizes. Model results can be used to estimate these values for both larger and smaller meshes, but expansion outside the tested range is less reliable and likely unnecessary.

For further consideration of the impact of different choices of mesh sizes, we used the length distribution found in the 2.5 in codend as a representation of the overall population size structure available to trawl gear. (This distribution was compared to NMFS NEFSC Groundish Survey data and found to be similar). A simulation using our model results was then developed using SELNET that estimated the distribution and number of fish predicted to be caught from a similarly structured theoretical population of 1000 redfish (Figure 14) using codends ranging from 4.5 in to 6.5 in, in 0.5 in steps. Intermediate mesh sizes can also be extrapolated using this procedure.

Additionally, the estimated escape of fish through the codend meshes can be inferred from the difference between the predicted number of fish caught and the theoretical population of 1000 fish. It is predicted that only 5.1% redfish in numbers would be

REDNET Component 3 Draft Final Report – Codend Selectivity Page 16

retained by a 6.5-in codend, 19.5% for a 5.5-in codend, and 51.3% by a 4.5-in codend. It is not known whether these fish would escape during fishing at the depth, during hauling in midwater, or at surface. Mortality rates of escapees that exit the net during different stages of fishing (towing, hauling, at surface) are likely different. It is generally agreed that fish that escape during towing may suffer less mortality than during hauling or at surface (Madsen et al. 2008). Escape at the surface increases mortality due to predation by other fish and by seabirds (Grimaldo et al. 2009; Madsen et al. 2008). Redfish may be unable to leave the surface for some time, if at all, due to eversion of swim bladders.

The simulations illustrate challenges associated with sustainable harvest of this species. First, the size distribution of redfish S. fasciatus is truncated compared to its near relatives S. mentella and S. marinus. This limitation may prevent exploitation of certain markets.

The current minimum landing size of 7 in (18 cm) represents the tail of the distribution of redfish, even when using the 2.5 in mesh control codend. Any mesh size larger than 2.5 will result in legal-sized fish passing through the codend meshes, and becoming subject to escape mortality.

The simulation results also suggest that some mesh sizes, while yielding larger sized fish, may result in catch rates are too low to be commercially viable. Additionally, the simulation emphasizes that substantial numbers of fish are passing through codend meshes and not being caught. The process of pursuit, exhaustion, and mesh passage has a definite, but difficult to quantify, level of mortality on fish. The range of possible escape mortality is quite wide; however, even at mortality rates of 10%, a significant number of redfish would die and be wasted as a result of the capture process.

Selectivity studies relate mesh opening to fish length, which serves as a proxy for fish girth, a more important morphological characteristic for determining codend escape. Fish girth can vary due to condition and breeding status, and thus the retention probabilities can vary over time (Wileman et al. 1996; Özbilgin et al. 2006). Additionally, twine type and thickness and other gear parameters may also influence retention probabilities (Wileman et al. 1996). Variation in selection between seasons, gears,

REDNET Component 3 Draft Final Report – Codend Selectivity Page 17

weather and other factors is therefore expected (Pope et al. 1975; DeAlteris and Grogan 1997). While this potential for variability presents a challenge for selecting an appropriate mesh size for this fishery, Herrmann et al. (2012) identified two cross- sections of Sebastes encompassing the hard parts of the skull as important to codend selectivity; their work suggests that size selectivity of Sebastes species may be subject to less variation attributable to changes in girth. Nevertheless, consideration of possible variation in retention curves should be given.

Conclusions and Recommendations

Testing and analysis of the selectivity of the codends provided results that by all and multiple measures appear robust. These results should be used to identify appropriate codend mesh sizes for sustainable harvesting of the species and for incorporating into stock assessment models.

The relatively small amount of bycatch in this experiment, even using a very small mesh control codend, indicates the power of choosing appropriate time, depth, and areas to sustainably target redfish with minimal bycatch.

The relatively small size range of this species of redfish should be a consideration in determining sustainable strategies for harvesting redfish. Unlike species with a larger maximum size, implementation of larger mesh sizes will not yield substantially greater growth.

The biology and population dynamics of this species should be used to help identify an appropriate size of redfish to be targeted, including whether this target should be at the

L25, L50, or other retention level.

The results of this study should be communicated to processors and marketers, in addition to the fishing fleet and managers, to emphasize the catch size and volume of fish at length available to the market.

REDNET Component 3 Draft Final Report – Codend Selectivity Page 18

Emphasis should be placed on determination of when escapement occurs. Research on other species has indicated that substantial escape could occur at the surface with all trawl gears. Unobserved escape mortality can be minimized by investigating when or if escape occurs in the codend during towing, and further if the use of a sorting grid, similar to the Nordmøre grate used in shrimp trawl fisheries, can be used to efficiently and effectively exclude small fish during the capture process with minimal mortality. This work is the focus of Component 4 of REDNET.

Acknowledgements

Primary thanks are due to the F/V Guardian captain and crew: Capt. Bradford Horrell, Bobby Blethen, José Garcia, Séan Farren, and Richard Walsh, and its owner, Mike Walsh, for their wisdom, patience, friendship, and participation. Thanks are also go to Sally Sherman, Co-PI of the REDNET project, for her support and advice. We appreciate assistance of our staff biologists who collected data onboard from DMF: David Chosid, and SMAST: Chris Rillahan and to Mark Szymanski of DMF for contract and other support. We acknowledge the extensive involvement of the REDNET network, which at present includes over 30 members. Bent Herrmann of SINTEF Fisheries and Aquaculture is kindly thanked for the use of SELNET and advices for the analysis using the software package. Funding was provided by the NOAA Fisheries Northeast Cooperative Research Partners Program.

References

Akaike, H., 1974. A new look at the statistical model identification. IEEE Trans. Auto. Control 19, 716-722. DeAlteris, J., Grogan, C., 1997. An Analysis of Harvesting Gear Size Selectivity for Eight Demersal Groundfish Species in the Northwest Atlantic Ocean. Fisheries Technical Report #1, University of Rhode Island Fisheries Center, Kingston, RI. Fonteyne, R., 2005. Protocol for the Use of an Objective Mesh Gauge for Scientific Purposes. ICES Coop. Res. Rep. No. 279. Frandsen, R.P., Herrmann, B., Madsen, N., 2010. A simulation-based attempt to quantify the morphological component of size selection of Nephrops norvegicus in trawl codends. Fish. Res. 101,156–167.

REDNET Component 3 Draft Final Report – Codend Selectivity Page 19

Fryer, R., 1991. A model of the between-haul variation in selectivity. ICES J. Mar. Sci. 48, 281–290. Grimaldo, E., Larsen, R.B., Sistiaga, M., Madsen, N., Breen, M., 2009. Selectivity and escape percentages during three phases of the towing process for codends fitted with different selection systems. Fish. Res. 95, 198–205. He, P., Balzano, V,. 2011. Rope grid: A new grid design to further reduce finfish bycatch in the Gulf of Maine pink shrimp fishery. Fish. Res. 111, 100-107. doi:10.1016/j.fishres.2011.07.001

Herrmann, B., Sistiaga, M., Nielsen, K.N., Larsen, R.B., 2012. Understanding the size selectivity of redfish (Sebastes spp.) in North Atlantic trawl codends. J. Northw. Atl. Fish. Sci. 44, 1–13. Herrmann, B., Wienbeck, H., Moderhak, W., Stepputtis, D., Krag, L., 2013. The influence of twine thickness, twine number and netting orientation on codend selectivity. Fish. Res. 145, 22–36. Madsen, N., Skeide, R., Breen, M., Krag, L.A., Huse, I., Soldal, A.V., 2008. Selectivity in a trawl codend during haul-back operation—An overlooked phenomenon. Fish. Res. 91,168–174. Özbilgin, H., Ferro, R.S.T., Robertson, J.H.B., Holtrop, G., Kynoch, R.J., 2006. Seasonal variation in trawl codend selection of northern North Sea haddock. ICES J. Mar. Sci. 63, 737–748. doi:10.1016/j.icesjms.2005.01.025 Pope, J.A., Margetts, A.R., Hamely, J.M., Akyüz, E.F., 1975. Manual of methods for fish stock assessment, Part III - Selectivity of fishing gear. Rome. R Development Core Team, 2009. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, http://www.R- project.org. Sarkar, D., 2009. Lattice: Lattice Graphics. R package version 0.17-26, http://CRAN.Rproject.org/package=lattice. Wienbeck, H., Herrmann, B., Moderhak, W., Stepputtis, D., 2011. Effect of netting direction and number of meshes around on size selection in the codend for Baltic cod (Gadus morhua). Fish. Res. 109, 80–88. Wileman, D., Ferro, R.S.T., Fonteyne, R., Millar, R.B. (Eds.), 1996. Manual of methods of measuring the selectivity of towed fishing gears. ICES Coop. Res. Rep. No. 215.

REDNET Component 3 Draft Final Report – Codend Selectivity Page 20

TABLES

REDNET Component 3 Draft Final Report – Codend Selectivity Page 21 Table 1: Codend mesh measurements Mesh opening (mm) Mesh Size Diameter Length Circumference Pre-experiment Post-experiment Average (in) (mm) (meshes) (meshes) Mean SD Mean SD Mean SD 2.5 Double 4 125 125.5 64 1.6 67.2 1.6 65.6 2.3 4.5 Double 5 70 70.5 106.3 2.5 110.6 2.8 108.4 3.4 5.5 Double 5 60 60.5 141.1 3 142.7 3.1 141.9 3.1 6.5 Double 5 50 50.5 163.4 3.1 163.1 4.1 163.2 3.6

Table 2: Catch weights (kg) by species and by mesh size (in) where total catch exceeded 5 kg, sorted by total weight. Note that the 2.5 in mesh catches are separate for each tested codend mesh size. Codend Mesh Size (in) Species 2.5 4.5 2.5 5.5 2.5 6.5 Total (kg) Redfish,Acadian Sebastes fasciatus 13,974.4 10,829.4 9,469.5 824.6 6,976.0 710.7 42,784.6 Pollock Pollachius virens 524.1 357.1 861.6 476.3 868.9 302.6 3,390.5 Cod, Atlantic Gadus morhua 62.2 94.7 67.9 96.8 233.2 85.8 640.6 Monkfish ( Goosefish) Lophius americanus 11.7 32.0 4.3 10.2 86.0 81.0 225.2 Lobster, American Homarus americanus 17.8 24.7 11.0 12.2 43.3 87.0 196.0 Dogfish, Spiny Squalus acanthias 32.5 38.8 12.9 0.7 91.8 176.7 Haddock Melanogrammus aeglefinus 9.1 15.0 22.9 22.2 34.7 9.5 113.4 Skate, Nk Rajidae 21.4 13.6 18.2 21.5 17.3 18.1 110.0 Seal, Gray Halichoerus grypus 100.0 100.0 Hake, Silver (Whiting) Merluccius bilinearis 11.5 4.6 24.7 0.4 44.6 1.7 87.4 Hake, White Urophycis tenuis 9.2 20.8 12.7 6.5 22.8 13.4 85.4 Hake, Red (Ling) Urophycis chuss 13.1 8.8 40.8 0.5 13.8 1.0 77.9 Flounder, American Plaice Hippoglossoides platessoides 15.0 11.8 7.6 7.0 14.2 9.1 64.6 Herring, Atlantic Clupea harengus 8.5 0.3 4.1 44.5 0.3 57.7 Herring, River, Nk* Alosa 9.6 0.8 12.7 33.6 0.7 57.4 Mackerel, Atlantic Scomber scombrus 4.2 2.4 0.4 35.7 0.5 43.2 Flounder, Witch Glyptocephalus cynoglossus 6.0 3.6 4.1 1.7 17.5 4.7 37.7 Cusk Brosme brosme 11.3 13.7 5.9 4.3 2.4 37.6 Halibut, Atlantic Hippoglossus hippoglossus 9.3 0.9 1.0 6.1 17.5 34.8 Squid, Atl Long-fin Doryteuthis pealeii 7.5 4.0 3.2 0.6 8.9 1.1 25.3 Sea Raven Hemitripterus americanus 4.3 1.9 2.5 3.9 2.9 4.6 20.1 Shad, American Alosa sapidissima 1.5 0.3 9.4 11.2 Ocean Pout Macrozoarces americanus 5.4 1.9 1.9 1.0 10.2 *Blueback or Alewife

REDNET Component 3 Draft Final Report – Codend Selectivity Page 22 FIGURES

REDNET Component 3 Draft Final Report – Codend Selectivity Page 23 US Canada Gulf of Maine

Study Area

Figure 1: Tow start locations by mesh size tested: blue circles = 4.5 in; red x = 5.5 in; black crosses = 6.5 in.

REDNET Component 3 Draft Final Report – Codend Selectivity Page 24 Figure 2a: Net diagram of the trouser trawl (top panel).

REDNET Component 3 Draft Final Report – Codend Selectivity Page 25 Figure 2b: Net diagram of the trouser trawl (lower panel).

REDNET Component 3 Draft Final Report – Codend Selectivity Page 26 Wire Out (fm)

260

240

220

200 Tow Speed (knt) Wave Ht (ft) 10

3.2 8

3.0 6

2.8 4

2 2.6 Depth (fm) Duration (h)

120 1.5

100 1.0

80

0.5 60

10 20 30 40 50 Tow Number

Figure 3: Operational and environmental variables in chronological order. Green dots are tows testing the 6.5 in codend; pink is the 5.5 in codend; blue dots are the 4.5 in codend. Dashed red horizontal lines are panel medians.

REDNET Component 3 Draft Final Report – Codend Selectivity Page 27 Figure 4: Distance (m) to net sensors by amount of wire out (fm)

REDNET Component 3 Draft Final Report – Codend Selectivity Page 28 Figure 5: Door heels (negative = outward lean ; positive = inward lean ) in degrees, separated by trip and wire out (fm)

REDNET Component 3 Draft Final Report – Codend Selectivity Page 29 Figure 6: Headline height (m) and door spread (m) shown for each amount of wire used during selectivity trials. Insufficient readings were obtained when 275 fm of wire was used. Red horizontal line is the overall median. Sample sizes proportional to box width.

REDNET Component 3 Draft Final Report – Codend Selectivity Page 30 Figure 7: Redfish length frequencies (cm) by mesh size (in). Red dashed lines are the older and newer minimum landing sizes

REDNET Component 3 Draft Final Report – Codend Selectivity Page 31 Figure 8: Pollock length frequencies (cm) by mesh size (in). Red dashed line is the minimum landing size.

REDNET Component 3 Draft Final Report – Codend Selectivity Page 32 Figure 9: Selection curves for 4.5 (blue), 5.5 (red), and 6.5 (black) (nominal inches) codends, with 95% confidence bands in stippled lines.

REDNET Component 3 Draft Final Report – Codend Selectivity Page 33 Figure 10: L50 estimates by mesh size (in) from individual hauls (blue circles) with 95% confidence intervals (error bars) compared to modeled mean L50 (solid horizontal line) and 95% confidence intervals (stippled lines), depicted by catch weight (kg).

REDNET Component 3 Draft Final Report – Codend Selectivity Page 34 Figure 11: Selection range (cm) estimates by mesh size (in) from individual hauls (blue circles) with 95% confidence intervals (error bars) compared to modeled mean L50 (solid horizontal line) and 95% confidence intervals (stippled lines), depicted by catch weight (kg).

REDNET Component 3 Draft Final Report – Codend Selectivity Page 35 Figure 12: Predicted mean L50 v mesh size for redfish (solid line) with 95% confidence intervals (stippled lines).

REDNET Component 3 Draft Final Report – Codend Selectivity Page 36 Figure 13: Predicted mean selection range (cm) v. mesh size for redfish (solid line) with 95% confidence intervals (stippled lines).

REDNET Component 3 Draft Final Report – Codend Selectivity Page 37 Figure 14: Simulated catch distributions based on selectivity analysis using five different mesh sizes (4.5; 5.0; 5.5; 6.0; 6.5 in) from the observed population distribution, scaled to 1000 fish. Numbers indicate the estimated number of fish retained by that codend mesh size.

REDNET Component 3 Draft Final Report – Codend Selectivity Page 38

G Model

FISH-4196; No. of Pages 7 ARTICLE IN PRESS

Fisheries Research xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Fisheries Research

journal homepage: www.elsevier.com/locate/fishres

Selectivity and retention of pollock Pollachius virens in a Gulf of Maine

trawl fishery

a,b,∗ c,d a a

Michael V. Pol , Bent Herrmann , Chris Rillahan , Pingguo He

a

School for Marine Science and Technology, University of Massachusetts Dartmouth, 706 South Rodney French Boulevard, New Bedford, MA 02744, USA

b

Massachusetts Division of Marine Fisheries, 1213 Purchase St — 3rd Floor, New Bedford, MA 02740, USA

c

SINTEF Fisheries and Aquaculture, Fishing Gear Technology, Willemoesvej 2, Hirtshals 9850, Denmark

d

University of Tromsø, Breivika, N-9037 Tromsø, Norway

a r t i c l e i n f o a b s t r a c t

Article history: Measuring the size selectivity of pollock, or saithe, Pollachius virens, has been challenging due to the

Received 13 March 2015

patchy, elusive nature of the species. By using data collected opportunistically during a redfish Sebastes

Received in revised form 16 July 2015

fasciatus selectivity study, selectivity of three sizes of mesh opening (114, 140 and 165 mm double 5 mm

Accepted 26 July 2015

twine diamond) for pollock was determined using a commercial fishing vessel fishing off Provincetown,

Available online xxx

Massachusetts, USA. Fifty-six tows were completed in March and April 2013; 21 included sufficient

catches of pollock to estimate size selectivity. Robust, simple models for the mean L50s (50% selection

Keywords:

length) and SR (selection ranges), and confidence intervals, were developed for all three tested codends,

Pollock

Saithe incorporating both within and between haul variability. Selection parameters and selection ranges were

determined for codends with nominal mesh sizes of 114 mm (L50: 34.8 cm; SR: 2.4 cm), 139 mm (L50:

Bottom trawl

Mesh selectivity 45.6 cm; SR: 3.1 cm), and 165 mm (L50: 52.4 cm; SR: 3.6 cm). All measures of model validity were positive,

indicating robust and reliable findings that can be used to provide guidance to fishery managers, stock

assessment scientists, and fishermen on size-dependent retention of pollock by codend mesh size.

© 2015 Elsevier B.V. All rights reserved.

1. Introduction our search of the literature revealed only a handful of attempts to

quantify size selectivity for pollock, and most of these studies were

Pollock, or saithe, Pollachius virens, is an important commer- “gray” literature and/or suffered from inadequate numbers of fish.

cial and sport fish along continental margins throughout the North In one particular example, a gillnet mesh selectivity analysis for

Atlantic (Robins and Ray 1986; Collette and Klein-MacPhee 2002; pollock was completed opportunistically during a haddock gillnet

Mayo et al., 1989). It is a fast and efficient swimmer, capable of mesh study (Marciano et al., 2006).

sustained swimming speeds of 0.89 m per second (He and Wardle, No pollock trawl selectivity results were found for the range of

1988; Collette and Klein-MacPhee 2002). They are patchily dis- sizes most relevant to current fisheries in the Gulf of Maine and

tributed, and notably active and elusive, often using the entire Georges Bank. DeAlteris and Chosid (2008) tested trawl codend

water column (Collette and Klein-MacPhee 2002). The elusiveness mesh sizes of 165 mm square and diamond, and even larger dia-

of this species is belied by the occurrence of directed fisheries — mond and square meshes of 178 and 203 mm in the region.

typically, pollock is a welcome bycatch during multispecies fish- Smolowitz (1983) used both a covered codend and alternate hauls

eries, and not a target (Mayo et al., 1989). But, targeted fisheries for to test a 138 mm diamond mesh codend, among other sizes, on

this species do occur in the North Sea (Holley and Marchal 2004). A nearby Georges Bank. Dahm (1998) (reported in Halliday et al.,

targeted fishery also briefly existed in the Gulf of Maine, USA during 1999) tested 121 mm diamond using a covered codend in European

the 1980’s (Mayo et al., 1989). waters.

The elusiveness, activity and patchy distribution of the species Pollock is one of three groundfish species in the Northeast

has apparently hindered the conduct of mesh selectivity studies — United States with Annual Catch Limits (ACL) in excess of 10,000

metric tons (Greater Atlantic Regional Fisheries Office, 2014).

In 2014, only 38.1 % of the commercial allowable catch limit

∗ was caught (Greater Atlantic Regional Fisheries Office, 2014). The

Corresponding author at: Massachusetts Division of Marine Fisheries, 1213 Pur-

biomass of the population exceeds target levels and the exploita-

chase St — 3rd Floor, New Bedford, MA 02740, USA. Fax: +1 15089900449.

E-mail address: [email protected] (M.V. Pol). tion rate is at a sustainable level (Mayo and Terceiro 2005).

http://dx.doi.org/10.1016/j.fishres.2015.07.029

0165-7836/© 2015 Elsevier B.V. All rights reserved.

Please cite this article in press as: Pol, M.V., et al., Selectivity and retention of pollock Pollachius virens in a Gulf of Maine trawl fishery.

Fish. Res. (2015), http://dx.doi.org/10.1016/j.fishres.2015.07.029

G Model

FISH-4196; No. of Pages 7 ARTICLE IN PRESS

2 M.V. Pol et al. / Fisheries Research xxx (2015) xxx–xxx

Therefore, this stock offers an opportunity for continued and of the trouser trawl and the control codend was attached to the

increased exploitation. other side. The side of the test and control codends was switched

Our goal was to develop length retention curves for pollock to regularly to avoid possible side-based effects, including the effect of

advise and to inform fishing industry members, fishery managers, the longer leg of the trouser trawl, described above. The same test

and assessment biologists of appropriate mesh sizes to harvest and codend was used for approximately three days before switching to

manage this species while maintaining and sustaining the health of another codend with a new mesh size.

this stock. Working in collaboration with the fishing industry and Tow locations were based on the captain’s knowledge of loca-

other interested parties, three candidate mesh sizes were tested: tions of Acadian redfish, echo sounder signals (including bottom

114, 140, and 165 mm. This size range was chosen for relevance topography), and a mix of redfish sizes. Tows were only made in

in the Acadian redfish fishery in which pollock is taken as a wel- daylight hours following the practice of the redfish fishery. Control

come bycatch. The tested sizes include the mandated minimum of tow durations was not necessary for valid comparisons, and var-

and two sizes at equal increments below that size that represent ied based on the captain’s assessment of the volume of fish in the

candidate mesh sizes for a regulatory exemption. The mesh selec- net and fishing ground conditions, as consistent with commercial

tivity data and analysis of this paper is a result of opportunistic practice. Very large catches were avoided if possible to reduce catch

capture of pollock during a redfish mesh selectivity study (see Pol handling time and maximize the number of tows.

et al., 2015; this volume). For this paper, we summarize only those The length of warp used was set by the captain based on the

tows where pollock was captured in sufficient numbers for size water depth and the bottom topography of the tow track. The

selectivity analysis. range of tow speed was restricted to within normal operational

conditions for Acadian redfish.

2. Materials and methods

2.2. Gear monitoring

The fishing gear, codends, fishing operation and sampling are the

A GoPro Hero2 high-definition camera (San Mateo, CA) with a

same as described in Pol et al. (2015; this volume) in which detailed

deepwater underwater housing and lights (Sartek Industries, Port

descriptions are provided. Some key elements are summarized for

completeness. Jefferson, NY) was mounted to view fish behavior in the mixing

area during some tows. Lights were pointed aft to minimize effects

on fish behavior. Net geometry was measured using a trawl mon-

2.1. Fishing vessel and gear

itoring system (Notus Electronics, St. John’s, Newfoundland) with

sensors on both doors, the trawl’s wing ends, just behind the head-

Sea trials were carried out on board F/V Guardian (25 m LOA;

rope, and on the control codend. The sensors were set to provide

425HP), a commercial groundfish trawler based in Boston, MA, USA.

bottom temperature, door spread, door heel (angle of the door to

The trouser trawl method was used for the experiment; a small

the right or left of the direction of travel), wing spread and to indi-

mesh control codend was attached to one leg of the trouser, and an

cate when the control codend was full. In addition, these sensors

experimental codend to the other leg. The trouser trawl method is

provided distance from the sensor to a towed hydrophone. Bottom

considered one of the accepted methods for conducting selectivity

temperature was also recorded with calibrated TidBit temperature

studies for towed gears by the International Council for the Explo-

recorders (Onset Computers Inc., Pocasset, Massachusetts).

ration of Sea (Wileman et al., 1996). The vessel provided a balloon

trawl front end (ground gear, wings, and net mouth) to be attached

2.3. Catch sampling

to a “trouser trawl” section. The headline of the trawl was 33.4 m in

length with 100 floats 20.3 cm in diameter. The footrope was 42.5 m

Codends were hauled on deck one at a time, with the codend

in length and attached to 30 cm diameter rockhopper groundgear.

attached to the shorter “leg” hauled and emptied first. Catches from

The front end of the net had 152 mm diamond mesh constructed

the experimental and control codends were deposited in separate

of 4.0 mm diameter braided PE twine. The fishing circle was 190

areas on deck, and processed separately.

meshes across the bottom panel and 240 meshes across the top.

Catch by species was weighed to 0.1 kg with subsampling when

The trouser section of the trawl was constructed of 152 mm dia-

there was a large amount of catch. Lengths (measured as midline

mond mesh, 3.6 mm diameter braided PE twine. It was designed

length, MLL) of a random subsample of more than 100 pollock

with a 47.5 meshes deep common “mixing area” that was then

(if possible) from each codend from each tow were measured to

separated uniformly into two lateral equal circumference legs (130

the nearest cm. For length-frequency (LF) analysis, counts at each

meshes across the bottom; 161 meshes across the top). One leg of

length were multiplied by the subsample weight divided into the

the trouser trawl was longer by 25 meshes of double 4 mm 165 mm

total weight.

mesh to avoid contact or inhibition of escape by one codend by the

other. Other organisms were also identified, and weighed to the

nearest 0.1 kg. Weights were directly measured or quantitatively

Mesh openings in the codends were measured prior to and after

determined; for example, by basket counts.

the experiment using an ICES OMEGA mesh gauge and associated

protocols (Fonteyne, 2005). The number of meshes for each test

codend and the control codend were adjusted so that the same 2.4. Analysis

diameter and overall length were maintained for all codends. The

control codend was constructed of double 4 mm diamond shaped All catch data (along with trip and gear data) were entered and

twine with a nominal mesh size of 64 mm, and was 125 meshes long uploaded into a customized relational database in Microsoft Access

and 125.5 meshes around. The test codends (114, 140, and 165 mm 2007 and were analyzed using Microsoft Excel and R statistical

nominal) were all constructed of diamond double 5 mm, and were software (R Development Core Team, 2009), primarily using the lat-

70, 60, and 50 meshes long and 70.5, 60.5, and 50.5 meshes around, tice package (Sarkar, 2009) and SELNET. SELNET was developed to

respectively. The actual mean mesh sizes taken before and after the acquire and analyze size selectivity and catch data for towed fishing

experiments were 65.6 mm, 108.4 mm, 141.9 mm and 163.2 mm gears, both at the haul level and for a group of hauls (Frandsen et al.,

for the codends of nominal mesh sizes of 64 mm, 114 mm, 140 mm, 2010; Herrmann et al., 2012; Herrmann et al., 2013). The methods

and 165 mm, respectively. A test codend was attached to one leg implemented in SELNET comply with accepted recommendations

Please cite this article in press as: Pol, M.V., et al., Selectivity and retention of pollock Pollachius virens in a Gulf of Maine trawl fishery.

Fish. Res. (2015), http://dx.doi.org/10.1016/j.fishres.2015.07.029

G Model

FISH-4196; No. of Pages 7 ARTICLE IN PRESS

M.V. Pol et al. / Fisheries Research xxx (2015) xxx–xxx 3

for the analysis of size selectivity data of towed fishing gears (Fryer, in the estimated between-haulvariation matrix for the selected

1991; Wileman et al., 1996). model (Fryer, 1991).

First, a logistic curve described by the parameters L50 (50%

These plots were inspected to see if the model predictions

selection length) and the SR (selection range) was used to model

appeared to reflect the main trends for the effects of catch size on

the size selection (Wileman et al., 1996). For each haul, the num-

the results for each codend and to see if it was necessary to con-

ber of fish counted at each length l in the experimental codend was

described as ntl, and in the control codend as ncl. The proportion sider the estimates of the between-haul variation and uncertainty

of the total catch measured for lengths is described by the subsam- on the means in the selection process in addition to the uncertainty

pling rates qt (experimental) and qc (control). The size selection

of the haul results. After successful model validation based on the

in each haul can then be obtained by minimizing the following

above procedure, the models were applied to predict size selection

function with respect to the parameters L50, SR and SP:

   for codend mesh sizes between 80 and 170 mm.

qt × ϕ(l, L50, SR, SP)

− ntl × ln

qt × ϕ(l, L50, SR, SP) + qc × (1 − ϕ(l, L50, SR, SP))

l 3. Results

 

qc × (1 − ϕ(l, L50, SR, SP))

+

nc × ln

l Tows were conducted generally east and northeast of Province-

qt × ϕ(l, L50, SR, SP) + qc × (1 − ϕ(l, L50, SR, SP))

2

town, Massachusetts, USA over an area of approximately 3800 nm .

With: Fifty-six tows were completed in two trips between 27 March and

1 April, and between 3 April and 8 April, 2013. Twenty-one tows

×

SP rlogit(l,L50,SR)

=

ϕ(l, L50, SR, SP) included sufficient numbers of pollock for analysis. Overall, 6 tows

− × −

1 SP (1 rlogit(l, L50, SR))

were included pairing the 114 mm mesh, with 2 on the starboard

SP is defined as the split parameter and expresses the assumed side; 8 tows with the 140 mm with 6 on the starboard side; 7 tows

length independent relative entry of fish to the test or control side with the 165 mm codend with 4 on the starboard side. Only four

of the trouser trawl during the fishing process. SP needs to be esti- tows were subsampled for pollock; subsampling fractions ranged

mated to assess the values of the selection parameters L50 and from 0.29 to 0.52. Catches of pollock in individual hauls ranged from

SR. 0.6 to 437 kg in a codend.

Fit statistics (i.e., the p-value and model deviance versus degrees Durations for tows with pollock catches ranged from 0.3 to 1.1 h

of freedom (DOF)) were inspected for individual hauls (Wileman with a median tow duration of 0.5 h. Median depth fished was

et al., 1996). Where the p-value < 0.05 or the deviance » DOF, the 99 fm, ranging from 50 to 117 fm; median towing speed was 3.0

residuals were examined for patterns or structural problems. knots (range: 2.9–3.3 kts); median warp length was 225 fm (range:

Where no pattern was seen, the poor fit was considered overdis- 225–250 fm). Median wave height experienced was 1.2 m with a

persion in the data and the data were included. maximum of 2.4 m; these heights are unlikely to substantially affect

Second, between-haul variation was considered (Fryer, 1991) net performance as they are within typical commercial operational

using the results from all the individual hauls simultaneously for conditions.

the L50, SR and SP, together with their covariance matrix and infor- A median headrope height of 4.1 m (Interquartile range (IQR):

mation on the values of the mesh size, m. In addition, the effect 3.8–5.1 m, n = 17 tows), door spread of 79 m ((IQR): 71–87 m,

of w (total test codend catch weight, in kg) and S (the side of the n = 15), and door heel medians of 3.4 degrees (IQR: 0.8–5.5 degrees,

trouser trawl the test codend was attached to). Since one codend n = 15) (port) and 5.1 degrees (IQR: 0.9–6.3 degrees, n = 13) (star-

necessarily stayed in the water longer during haulback, and could board) inward were recorded. Median distance to doors was 423 m

potentially lose more fish, we wanted to test whether this longer (IQR: 420–468 m, n = 15); to the headrope 556 m (IQR: 554–567 m,

hauling time might impact size selectivity. n = 16); to the codend 603 m (IQR: 598–608 m, n = 9) — these dis-

A model considering the potential effect of the parameters m, tances increased with depth and longer warp length. Based on net

w and S was constructed with the following form and applied in geometry, no anomalous tows were identified. Median temper-

SELNET. ature was 6.9 degrees (IQR: 6.4–7.1 degrees) and did not differ

between mesh sizes.

L50 = f0 + f1 × m + f2 × w + f3 × S

The total catch of all species was just over 47,900 kg, with redfish

SR = g0 + g1 × m + g2 × w + g3 × S comprising 42,482.9 kg or 89.7% of all catch. Pollock was the main

SP = h0 + h1 × m + h2 × w + h3 × S bycatch species (3390.5 kg), with 21 other species with catches

greater than 10 kg total (See Table 2 in Pol et al., 2015, this vol-

The parameters f0. . .f3, g0. . .g3 and h0. . .h3 were estimated while

ume). Over 1800 pollock were measured. Lengths ranged from 18

fitting the model to the data with values for L50, SR and SP based

to 86 cm with distributions differing between codends (Fig. 1).

on the selectivity results from the individual hauls. Models were

Twenty-one hauls were included in the pollock size selectivity

selected based on the lowest AIC value (Akaike, 1974), while consid-

analyses. Testing of the full model and all simpler models – a total

ering every possible simpler sub-model following the procedure

of 4096 models in all – resulted in selection of a sub-model where

described in Wienbeck et al. (2011) and Herrmann et al. (2013).

the L50 and the SR depend only on mesh size, m:

Individual haul results were plotted for the L50 and SR with

95% CI versus the mean model estimated values and the predicted L50 = 0.3212 × m(CI : 0.3011, 0.3413)

95% CI for the total variation (between-haul variation + uncertainty

around the mean). The lower and upper 95% CI for the estimated =

SR 0.0221 × m(CI : 0.0120, 0.0322)

between-haul variation in the selection parameters (lim L50, lim

SR) were calculated by:

 The model estimate of SP = 0.5359 (CI: 0.4389, 0.6328) was near

the ideal value of 0.5 with the 95% confidence interval (CI) over-

limL50 = L50mean ± 1.96 × (VarL50mean + D11)

 lapping 0.5. Video recordings collected from this area indicated no

unusual fish behavior. The side the codend was on (S) and the total

= ± × +

limSR SRmean 1.96  (VarSRmean D22)

weight of the catch (w) were not present in the final model.

limSP = SPmean ± 1.96 × (VarSPmean + D33)

Full logistic curves with 95% confidence intervals illustrating

where L50mean, SRmean and SPmean are the predictions based on the the catch curves for the codends as measured were constructed

selected submodel and D11, D22 and D33 are the diagonal elements from model results (Fig. 2). L50 (in MLL) and selection ranges

Please cite this article in press as: Pol, M.V., et al., Selectivity and retention of pollock Pollachius virens in a Gulf of Maine trawl fishery.

Fish. Res. (2015), http://dx.doi.org/10.1016/j.fishres.2015.07.029

G Model

FISH-4196; No. of Pages 7 ARTICLE IN PRESS

4 M.V. Pol et al. / Fisheries Research xxx (2015) xxx–xxx

64 mm 12 10 8 6 4 2 0 114 mm 12 10 8 6 4 2 otal T 0 140 mm 12

ercent of 10 P 8 6 4 2 0 165 mm 12 10 8 6 4 2 0

20 40 60 80

Midline length (cm)

Fig. 1. Pollock length frequencies (cm) by mesh size (mm). Vertical dashed lines mark the minimum landing size, converted to FL using Pol et al. (2011).

Only six hauls were found to have values of L50 outside the 95% CIs;

1.0 all of these hauls overlapped their error bars with the confidence

band indicating that the model is an appropriate fit to the data.

A similar comparison was made for the selection range (not pic-

tured); six individual hauls had SR values outside the overall CI for

0.8

the mean SR. All of these also had error bars overlapping the CI band.

These combined results indicated that the fit of the individual hauls

to the overall results were excellent.

0.6

Model results were used to produce estimates of mean L50 and tion

r

SR (Fig. 4) across a broad range of mesh sizes to support the choice of

appropriate mesh sizes. Model results can be used to estimate these

Propo

0.4 values for both larger and smaller meshes, but expansion outside

the tested range is less reliable and likely unnecessary.

0.2 4. Discussion

Testing and analysis of the selectivity of the codends provided

robust results with good measures of validation and a simple

0.0

MLS selection model. These results can be used to identify appropri-

ate codend mesh sizes for sustainable harvesting of the species and

20 30 40 50 60

Midline length (cm) for incorporating into stock assessment models and other man-

agement measures. While it is not frequently targeted, pollock is a

Fig. 2. Selection curves (solid lines) for 114 (left), 140 (center), and 165 (right) mm primary bycatch species when targeting Acadian redfish and other

codends, with 95% confidence bands in dashed lines. MLS converted from TL using

groundfish species such as haddock; our results can be incorpo-

Pol et al. (2011).

rated with selectivity results for redfish to anticipate the effects of

different choices of codends in the current and any future redfish

were determined for codends with nominal mesh sizes of 114 mm directed fishery and other groundfish fisheries.

(L50: 34.8 cm; SR: 2.4 cm), 139 mm (L50: 45.6 cm; SR: 3.1 cm), and Testing of codends was sequential; that is, for operational rea-

165 mm (L50: 52.4 cm; SR: 3.6 cm). Further validation of the selec- sons, codends were used in blocks, and not randomized or rotated

tivity model was demonstrated by plotting L50 (Fig. 3) and SR per haul. Geometry, geographical area, weather conditions, depth,

values (not shown)) for each individual haul, with 95% confidence and temperature did not vary substantially, limiting uncontrolled

intervals as error bars, along with overall mean values predicted sources of variability, and supporting comparability of the codends.

by the final model, against the catch weight of the codend (Fig. 3). Aft-pointing lights in the mixing area did not appear to affect behav-

Please cite this article in press as: Pol, M.V., et al., Selectivity and retention of pollock Pollachius virens in a Gulf of Maine trawl fishery.

Fish. Res. (2015), http://dx.doi.org/10.1016/j.fishres.2015.07.029

G Model

FISH-4196; No. of Pages 7 ARTICLE IN PRESS

M.V. Pol et al. / Fisheries Research xxx (2015) xxx–xxx 5

Fig. 3. L50 estimates by mesh size (mm) from individual hauls (circles) with 95% confidence intervals (error bars) compared to modeled mean L50 (solid horizontal line) and

95% confidence intervals (dashed lines), depicted by catch weight (kg).

ior and likely had little or no impact on codend selection. The mechanical factors such as the opening angles of the meshes or

employment of a trouser trawl, which simultaneously tows a selec- the stiffness of twines were identical during the selection process

tive test codend and a non-selective codend, helps obtain results for both species, and therefore permit comparisons between the

that are consistent and accurate within each tow, as the same school results for the two species. Under these conditions, the SR depends

of fish is exposed to both codends. And, while the trouser trawl is only on morphological variation (Herrmann et al., 2009). Thus, the

not used commercially in the fishery studied, the results are likely lower SR coefficient (by more than a factor of 10) implies that the

applicable to commercial gear. escape ability of pollock changes quickly with size. While behavior

The total catch weight in the codend or which side the codends undoubtedly plays a role in escape ability, girth is more likely the

were attached to were not significant terms in the model. This dominating factor. Therefore, the narrow SR illustrates greater uni-

lack of effect helps demonstrate good functioning of the trouser formity in the length-girth relationship for pollock than for redfish.

trawl and randomization of the distribution of fish between the Our results were compatible with earlier studies. Using an alter-

test and control codends. Also, this result suggests that selectivity nate tow method, DeAlteris and Chosid (2008) found an identical

may not vary greatly across tows. While Wileman et al. (1996) rec- L50 for 165 mm diamond mesh (52.4 cm), but with a SR of 12.33 cm,

ommended the use of a trouser trawl with a full vertical split, our compared to our SR of 3.6 cm using similar analytical methods. This

design avoided a vertical panel (similar to He and Balzano, 2011) to difference may be explained by the smaller sample size in the ear-

reduce possible deformation or distracting motion in the net and lier study (11 pairs vs. 21 in our study). DeAlteris and Chosid (2008)

panel. A trouser trawl can avoid deck-handling problems associated also tested square mesh of the same dimension, which resulted,

with covers, reduces risk of masking of codend meshes, and risk as expected, in a much larger L50 (67.6 cm, SR: 7.6 cm). The L50

of over-filling the cover. Few difficulties were encountered using reported by Smolowitz (1983) for 138 mm diamond mesh using a

the trouser trawl, as the crew was adaptable to the hauling of two covered codend on Georges Bank was 45 cm, with a selection fac-

codends. Operationally it was not difficult to retrieve and empty tor of 3.26. Our results for the L50 of 140 mm mesh were similar

two codends in a controlled and safe manner. to that study: 45.6 cm and a SR of 3.1 cm. Dahm (1998) reported a

The absence of any intercept terms in the L50 and SR pro- L50 of 51.8 for 121 diamond; this size is intermediate to our tested

vided the logical result that as the mesh size is reduced to zero, codends, and does not match well with our results, which predict

the fish length would similarly be reduced to zero. The selec- a L50 of 38.9 cm for this mesh size.

tion ranges were narrow compared to those found for redfish The similarity of the L50 across studies further implies that the

during the same study (Pol et al., 2015, this volume). The SR coeffi- fish length to girth relationship for pollock is steady. Fish girth can

cient for pollock we determined (0.0221 × mesh size) is notably vary due to condition and breeding status, changing the retention

lower than that found for redfish during the same experiment probabilities (Wileman et al., 1996; Özbilgin et al., 2006). Addition-

(0.310 × mesh size) (Pol et al., 2015; this volume). It is reasonably ally, twine type and thickness and other gear parameters may also

safe to presume, since these results come from the same tows, that influence retention probabilities (Wileman et al., 1996). Variation

Please cite this article in press as: Pol, M.V., et al., Selectivity and retention of pollock Pollachius virens in a Gulf of Maine trawl fishery.

Fish. Res. (2015), http://dx.doi.org/10.1016/j.fishres.2015.07.029

G Model

FISH-4196; No. of Pages 7 ARTICLE IN PRESS

6 M.V. Pol et al. / Fisheries Research xxx (2015) xxx–xxx

two species, as a 22.9 cm redfish is above the MLS of 17.8 TL; how-

ever, the predicted L50 for pollock is below the MLS of 48.3 cm TL

for pollock. This incongruity highlights one challenge of managing

a mixed species fishery. Our results provide valuable information

50 for managing that fishery.

MLS (cm)

50 Acknowledgements

L 40

Primary thanks are due to the F/V Guardian captain and crew:

Mean

Capt. Bradford Horrell, Bobby Blethen, José Garcia, Séan Farren,

and Richard Walsh, and its owner, Mike Walsh, for their wisdom,

30

patience, friendship, and participation. Thanks are also go to Sally

4.5 5.0 5.5 6.0 6.5 Sherman and Kohl Kanwit of the Maine Department of Marine

Resources for their support and advice. We appreciate assistance

from Division of Marine Fisheries staff biologists David Chosid for

110 120 130 140 150 160

data collection and Mark Szymanski for contract and other sup-

Mesh size (mm and in) port. We acknowledge and appreciate the extensive involvement of

the REDNET network, which at present includes over 30 members.

Funding was provided by the NOAA Fisheries Northeast Cooper-

ative Research Partners Program and a NOAA Interstate Fisheries

Management Support Grant.

5

References

4 Akaike, H., 1974. A new look at the statistical model identification. IEEE Trans. Auto

ange (cm)

r Control 19, 716–722.

Collette, B., Klein-MacPhee, G., 2002. Bigelow and Schroeder’s Fishes of the Gulf of

Maine, third ed. Smithsonian Institution Press, Washington.

3

Dahm, E., 1998. Measurements of codend selectivity for North Sea saithe by the

covered codend and divided trawl methods. Arch. Fish. Marine Res. 46, 43–59.

DeAlteris, J.T., Chosid, D.M., 2008. Effects of codend mesh size on the multispecies

2 yield and spawning stock biomass in the western Georges Bank trawl fisheries.

In: Completion Report. University of Rhode Island.

Mean selection DeAlteris, J., Grogan, C., 1997. An Analysis of Harvesting Gear Size Selectivity for

Eight Demersal Groundfish Species in the Northwest Atlantic Ocean. Fisheries

1 4.5 5.0 5.5 6.0 6.5

Technical Report #1. University of Rhode Island Fisheries Center, Kingston, RI.

Fonteyne, R., 2005. Protocol for the use of an objective mesh gauge for scientific

110 120 130 140 150 160 purposes. ICES Coop. Res. Rep. No. 279.

Frandsen, R.P., Herrmann, B., Madsen, N., 2010. A simulation-based attempt to

Mesh size (mm and in) quantify the morphological component of size selection of Nephrops

norvegicus in trawl codends. Fish. Res. 101, 156–167.

Fryer, R., 1991. A model of the between-haul variation in selectivity. ICES J. Mar.

Fig. 4. Predicted mean L50 (top) and SR (bottom) v mesh size for pollock (solid line) Sci. 48, 281–290.

with 95% confidence intervals (dashed lines). MLS converted using Pol et al. (2011). Greater Atlantic Regional Fisheries Office, 2014. Northeast multispecies fishery

final year-end results for fishing year 2013. http://www.greateratlantic.

fisheries.noaa.gov/aps/monitoring/nemultispecies.html Accessed 12

in selection between seasons, gears, weather and other factors is November 2014.

also expected (Pope et al., 1975; DeAlteris and Grogan 1997). How- Halliday, R.G., Cooper, C.G., Fanning, P., Hickey, W.M., Gagnon, P., 1999. Size

selection of Atlantic cod, haddock and pollock (saithe) by otter trawls with

ever, pollock seems less susceptible to change in the mesh size to

square and diamond mesh codends of 130–155 mm mesh size. Fish. Res. 41,

length relationship. This stability may also be due to the location 255–271.

on the fish of the maximum girth – for example, rigid structures at He, P., Balzano, V., 2011. Rope grid: a new grid design to further reduce finfish

bycatch in the Gulf of Maine pink shrimp fishery. Fish. Res. 111, 100–107,

this point would lead to less variability in the length-girth relation-

http://dx.doi.org/10.1016/j.fishres.2011.07.001

ship. Further studies with pollock could determine the important

He, P., Wardle, C.S., 1988. Endurance at intermediate swimming speeds of Atlantic

cross-sections for pollock using the FISHSELECT method (Sistiaga mackerel, Scomber scombrus L., herring, Clupea harengus L., and saithe,

Pollachius virens L. J. Fish Biol. 33, 255–266.

et al., 2011).

Herrmann, B., Krag, L.A., Frandsen, R.P., Madsen, N., Lundgren, B., Staehr, K.-J.,

The pollock and redfish selectivity models (Pol et al., 2015; this

2009. Prediction of selectivity from morphological conditions: Methodology

volume) for L50 and SR both rely only on mesh size. This simple rela- and a case study on cod (Gadus morhua). Fish. Res. 97, 59–71.

Herrmann, B., Sistiaga, M., Nielsen, K.N., Larsen, R.B., 2012. Understanding the size

tionship can assist in choosing an appropriate mesh size in the Gulf

selectivity of redfish (Sebastes spp.) in North Atlantic trawl codends. J. Northw.

of Maine redfish fishery, where pollock is a primary bycatch species

Atl. Fish. Sci. 44, 1–13.

(although below 5% of total catch (Kanwit et al., 2013)). Using the Herrmann, B., Wienbeck, H., Moderhak, W., Stepputtis, D., Krag, L., 2013. The

influence of twine thickness, twine number and netting orientation on codend

relationship for redfish of L50 = 0.206 × mesh size (Pol et al., 2015;

selectivity. Fish. Res. 145, 22–36.

this volume) and for pollock of L50 = 0.3212 × mesh size, we find

Holley, J., Marchal, P., 2004. Fishing strategy development under changing

×

the L50 for pollock = 1.56 L50 for redfish. This relationship assists conditions: examples from the French offshore fleet fishing in the North

in judging an appropriate mesh size when targeting redfish. (It also Atlantic. ICES J. Mar. Sci. 61, 1410–1431, http://dx.doi.org/10.1016/j.icesjms.

2004.08.010

indirectly quantifies differences between these two species in body

Kanwit, K., Pol, M., He, P., 2013. REDNET — A network to redevelop a sustainable

shape: a pollock of 50% retention probability is 1.56 times longer

redfish (Sebastes fasciatus) trawl fishery in the Gulf of Maine. In: Component 2

than a redfish of the same retention probability.) For example if a — baseline catch and bycatch evaluation. Report to the Northeast Cooperative

Research Partners Program, National Marine Fisheries Service.

22.9 cm (9 inch) MLL L50 is desired for redfish, our results allow

Marciano, D., Rosen, S., Pol, M.V., Szymanski, M., 2006. Testing the selectivity of

quick calculation of an L50 of 35.7 cm MLL for pollock. This result

gillnets to target haddock in the Gulf of Maine. Final Report to the NMFS

highlights an inconsistency in the minimum landing sizes of these Cooperative Research Partners Initiative.

Please cite this article in press as: Pol, M.V., et al., Selectivity and retention of pollock Pollachius virens in a Gulf of Maine trawl fishery.

Fish. Res. (2015), http://dx.doi.org/10.1016/j.fishres.2015.07.029

G Model

FISH-4196; No. of Pages 7 ARTICLE IN PRESS

M.V. Pol et al. / Fisheries Research xxx (2015) xxx–xxx 7

Mayo, R.K., McGlade, J.M., Clark, S.H., 1989. Patterns of exploitation and biological R Development Core Team, 2009. R: A Language and Environment for Statistical

status of pollock (Pollachius virens L.) in the Scotian Shelf, Georges Bank, and Computing. R Foundation for Statistical Computing, Vienna, Austria, http://

Gulf of Maine Area. J. Northwest Atl. Fish. Sci. 9, 13–36. www.R-project.org

Mayo R.K., Terceiro M., Editors, 2005. Assessment of 19 Northeast groundfish Robins, C.R., Ray, G.C., 1986. A field guide to Atlantic coast fishes. Houghton Mifflin,

stocks through 2004. 2005 Groundfish Assessment Review Meeting (2005 Boston.

GARM). Northeast Fisheries Science Center, Woods Hole, Massachusetts, Sarkar, D., 2009. Lattice: Lattice Graphics. R package version 0. 17-26, http://CRAN

15–19 August 2005. U.S. Dep. Commer., Northeast Fish. Sci. Cent. Ref. Doc. Rproject.org/package=lattice.

05-13; 499 p. Sistiaga, M., Herrmann, B., Nielsen, K.N., Larsen, R.B., 2011. Understanding limits to

Özbilgin, H., Ferro, R.S.T., Robertson, J.H.B., Holtrop, G., Kynoch, R.J., 2006. Seasonal cod and haddock separation using size selectivity in a multispecies trawl

variation in trawl codend selection of northern North Sea haddock. ICES J. Mar. fishery: an application of FISHSELECT. Can. J. Fish. Aquat. Sci. 68, 927–940,

Sci. 63, 737–748, http://dx.doi.org/10.1016/j.icesjms.2005.01.025 http://dx.doi.org/10.1139/F2011-017

M.V. Pol, B. Herrmann, C. Rillahan, P. He, (submitted). Impact of codend mesh sizes Smolowitz, R.N. 1983. Mesh size and the New England groundfishery —

on selectivity and retention of Acadian redfish Sebastes fasciatus in the Gulf of applications and implications. NOAA Tech. Rpt. NMFS SSRF-776. pp. 121.

Maine trawl fishery Fish. Res. 2015;. Wienbeck, H., Herrmann, B., Moderhak, W., Stepputtis, D., 2011. Effect of netting

Pol, M.V., Szymanski, M., Chosid, D.M., Salerno, D.J., 2011. Fork length — total direction and number of meshes around on size selection in the codend for

length conversions for haddock and pollock, North Am. J. Fish. Manag 31, Baltic cod (Gadus morhua). Fish. Res. 109, 80–88.

427–430, http://dx.doi.org/10.1080/02755947.2011.590115 Wileman, D., Ferro, R.S.T., Fonteyne, R., Millar, R.B., 1996. Manual of methods of

Pope, J.A., Margetts, A.R., Hamely, J.M., Akyüz, E.F., 1975. Manual of methods for measuring the selectivity of towed fishing gears. In: ICES Coop. Res. Rep., pp.

fish stock assessment, Part III - Selectivity of fishing gear. Rome. No 215.

Please cite this article in press as: Pol, M.V., et al., Selectivity and retention of pollock Pollachius virens in a Gulf of Maine trawl fishery.

Fish. Res. (2015), http://dx.doi.org/10.1016/j.fishres.2015.07.029 Elsevier Editorial System(tm) for Fisheries Research Manuscript Draft

Manuscript Number:

Title: Impact of codend mesh sizes on selectivity and retention of Acadian redfish Sebastes fasciatus in the Gulf of Maine trawl fishery

Article Type: SI: Balanced fishing

Keywords: Acadian redfish; Sebastes fasciatus; Bottom trawl; Size selectivity

Corresponding Author: Mr. Michael Pol,

Corresponding Author's Institution: Massachusetts Division of Marine Fisheries

First Author: Michael Pol

Order of Authors: Michael Pol; Bent Herrmann; Chris Rillahan; Pingguo He

Manuscript Region of Origin: USA

*Cover Letter

Commonwealth of Massachusetts Division of Marine Fisheries 1213 Purchase St – 3rd Floor New Bedford, Massachusetts 02740 (508) 990-2860 Paul J. Diodati Director

15 January 2015

Editorial Board Fisheries Research

Dear Sir/Madam,

On behalf of my co-authors, it is a pleasure to submit the manuscript, “Impact of codend mesh sizes on selectivity and retention of Acadian redfish Sebastes fasciatus in the Gulf of Maine trawl fishery” for consideration in the special Balanced Harvest volume. The manuscript is original research that has not been published elsewhere, and is not under consideration for publication elsewhere.

The information contained was collected in close collaboration with a broad network of scientists, commercial fishermen, processors, netmakers, regulators, and others in response to a clear need, and has been robustly analyzed with the latest techniques. I hope you find the manuscript to your satisfaction.

My co-authors have been closely involved with the preparation of the manuscript, and have consented to publication.

Please let me know if you have any questions regarding this submission.

Sincerely,

Michael Pol Sr. Marine Fisheries Biologist/Program Leader Conservation Engineering Division of Marine Fisheries 1213 Purchase St - 3rd Floor New Bedford, MA 02740 USA +1.508.990.2860 x116 [email protected]

Highlights (for review)

Highlights

 Simple, robust codend selectivity models were developed for three mesh openings for Acadian redfish (Sebastes fasciatus).

 Size of codend catch and codend side were not significant factors in the final model. All measures of model validity were positive.

 The minimum mandatory mesh size retained less than 5% of the fish that entered the net.

 Location of escape (bottom, midwater, surface) likely affects mortality rates and should be investigated.

*Manuscript including abstract Click here to download Manuscript including abstract: PoletalRedfishFishRes_submitted.docClick here to view linked References

1 Impact of codend mesh sizes on selectivity and retention of Acadian redfish Sebastes

2 fasciatus in the Gulf of Maine trawl fishery

3

4 Michael V. Pol*a,b, Bent Herrmannc,d, Chris Rillahanb, Pingguo Heb

5 *corresponding author: [email protected]; +1 508 990 2860 x116

6 aMassachusetts Division of Marine Fisheries, 1213 Purchase St – 3rd Floor, New Bedford,

7 Massachusetts, USA 02740

8 bSchool for Marine Science and Technology, University of Massachusetts Dartmouth, 706 South

9 Rodney French Boulevard, New Bedford, MA 02744 USA

10 cSINTEF Fisheries and Aquaculture, Fishing Gear Technology, Willemoesvej 2, 9850 Hirtshals,

11 Denmark

12 dUniversity of Tromsø, Breivika, N-9037 Tromsø, Norway

13

14 Keywords:

15 Acadian redfish

16 Sebastes fasciatus

17 Bottom trawl

1

2

18 Size selectivity

19

20

21

22

23

24

25

26

27

28 Abstract

29 A trouser trawl was used to determine the size selectivity of three sizes of mesh opening (114,

30 140 and 165 mm double 5 mm twine diamond) on a commercial fishing vessel fishing off

31 Provincetown, Massachusetts, USA. Fifty-six tows were completed in March and April 2013,

32 catching over 42,000 kg of Acadian redfish (Sebastes fasciatus) and about 6,000 kg of other

33 species. Robust models for the mean L50s and selection ranges, and confidence intervals, were

34 developed for all three tested codends, incorporating both within and between haul variability.

35 L50 and selection ranges were determined for the nominal 114 mm (L50: 22.3 cm; SR: 3.3 cm), 3

36 139 mm (L50: 29.2 cm; SR: 4.4 cm), and 165 mm (L50: 33.6 cm; SR: 5.0 cm) codends. All

37 measures of model validity were positive. These models are fully adequate to provide guidance

38 to managers and fishermen on size retention of redfish and appropriate codend mesh size.

39 Additionally, simulation of fishing of the three tested codends on the observed population

40 indicated that substantial escape of redfish through codend meshes occurs (51-96%),

41 suggesting that investigation of escape of redfish is warranted to support a sustainable fishery.

42 The observed population also indicates that inadequate numbers of larger redfish may be

43 available to support a higher-priced market.

44

45 1. Introduction

46 Acadian redfish (Sebastes fasciatus – “redfish”) is one of three groundfish species in the

47 Northeast United States with Annual Catch Limits (ACL) in excess of 10,000 metric tons (GARFO,

48 2014a). However, the ACL has not been fully utilized over the last few years because the

49 mandatory minimum codend mesh size is too large to effectively retain redfish.

50 Historically, redfish represented a sizeable fishery and income in the region. The directed

51 commercial fishery for redfish in the Northeast US began in the 1930s with the advent of

52 freezing technology (Mayo et al., 2006). Total landings from the Gulf of Maine and Georges

53 Bank rose from 100 metric tons (mt) to a peak of 56,000 mt in 1942 and then steadily declined

54 (Mayo et al., 2006). By comparison, annual landings in the 1930-40 period of the iconic Atlantic

55 cod (Gadus morhua) averaged approximately 9,000 mt (Mayo et al., 2006). By 1989, the total 4

56 US landings of redfish fell below 1,000 mt, and remained at that level throughout the 2000s

57 (Mayo et al., 2006). Research vessel survey indices begun in 1963 showed a 90% decline in

58 redfish per tow between 1968 and 1985 during which time the catch rate reached their lowest

59 point (Collette and Klein-MacPhee, 2002; Mayo et al., 2006). Since 1995, survey catches have

60 steadily increased and remain high to this day (Mayo et al., 2006; Northeast Fisheries Science

61 Center (NEFSC), 2012). Based on the most recent assessment, the stock has reached and

62 exceeded its target biomass for maximum sustainable yield of 238,000 mt and the exploitation

63 rate remains below its target of 0.04 (NEFSC, 2012).

64 Mesh restrictions, combined with low biomass levels between 1980 and 1995, eliminated the

65 directed redfish fishery in the Northeastern United States. The decline in abundance of redfish

66 likely resulted from overexploitation, and recovery was likely encouraged by a mismatch

67 between mandatory minimum mesh sizes for the multispecies fishery and the smaller size of

68 redfish compared to other target species in the groundfish complex. In 1977, the minimum

69 codend mesh size for redfish (and all groundfish species) was increased from 114 to 130 mm

70 (Anthony, 1990) and increased again in 1994 to 152 mm (Fogarty and Murawski, 1998;

71 Murawski et al., 2000). Currently, the mandated minimum codend mesh size is 165 mm. This

72 minimum mesh size is applied to all managed groundfish species, including Atlantic cod,

73 haddock (Melanogrammus aeglefinus), and pollock (Pollachius virens), as well as several species

74 of flatfish. These species have a range of minimum landing sizes (MLS) from 30.5 to 48 cm;

75 redfish is the smallest at 22.9 cm (recently lowered to 17.8 cm). Individual redfish are typically

76 in the 20-30.5 cm range (Collette and Klein-MacPhee, 2002) and redfish at the minimum size 5

77 are too small to be retained in 165 mm codend. Consequently, the use of one mesh size for the

78 multispecies fishery results in a mismatch between the mesh size and the minimum landing size

79 for redfish. As a result, only 38.5% of the annual catch limit was landed in 2013 (GARFO, 2014b).

80 Indeed, fishing industry collaborators within the REDNET network, a research network

81 established to redevelop the redfish trawl fishery, have reported that the majority of redfish in

82 the codend cannot be retained in 165 mm mesh unless the vessel maintains a constant vessel

83 headway during retrieval of the net.

84 Acadian redfish is one of three Sebastes species, all very similar in morphology, exploited in

85 fisheries across the North Atlantic (Herrmann et al., 2012). Experiments on codend mesh

86 selectivity for these species started in the 1960s with varying levels of rigor. A recent thorough

87 review of this topic (Herrmann et al. 2012) found 21 investigations of codend mesh selectivity,

88 mostly for diamond-shaped meshes, and mostly for redfish congeners Sebastes marinus and S.

89 mentella. Only three of these studies involved S. fasciatus, which was combined with S.

90 mentella during those studies as one species group (Herrmann et al. 2012).

91 Our goal was to develop length retention curves for Acadian redfish to advise and to inform

92 fishermen, processors, fishery managers, and assessment biologists on selection of an

93 appropriate mesh size to increase codend retention an, therefore, landings while maintaining

94 and sustaining the health of the stock. Working in collaboration with the REDNET network,

95 three candidate mesh sizes were chosen: 114, 140, and 165 mm. This range incorporates the

96 current minimum mesh size and the likely smallest acceptable mesh size for a special access

97 program, with one mesh at the midpoint. Diamond meshes were preferred over square mesh 6

98 due to “sticking” by redfish in square mesh codends (ICES, 2012). Sticking results when fish pass

99 partway through meshes, and are gilled in the codend. A large number of fish sticking in the

100 codend requires a great deal of time to remove them; many of these fish become damaged and

101 unsalable.

102

103 2. Materials and Methods

104 2.1 Fishing vessel and gear

105 The F/V Guardian (80 ft LOA; 425 hp), a commercial groundfish trawler with recent experience

106 targeting redfish, was chosen to conduct the research. The participating vessel provided a

107 balloon trawl front end (ground gear, wings, and net mouth) to be attached to a “trouser trawl”

108 section. The headline of the trawl was 33.4 m in length with 100 plastic floats 20.3 cm in

109 diameter. The footrope was 42.5 m in length and attached to a rockhopper groundgear. The

110 front end of the net had 152 mm diamond mesh openings constructed of 4.0 mm diameter

111 braided twine. The fishing circle was 190 meshes across the bottom panel and 240 meshes

112 across the top.

113 The trouser section of the trawl was also constructed of 152 mm diamond mesh, 3.6 mm

114 diameter braided twine. It was designed with a 47.5 meshes deep common “mixing area” that

115 was then separated uniformly into two lateral equal circumference legs (130 meshes across the

116 bottom; 161 meshes across the top). One leg of the trouser trawl was lengthened by 25 meshes 7

117 of double 4 mm 165 mm mesh to avoid contact or inhibition of escape by one codend on the

118 other.

119

120 Mesh openings in codends were measured prior to and after the experiment using an ICES

121 OMEGA mesh gauge and associated protocols (Fonteyne, 2005). The number of meshes for

122 each test codend was adjusted so that the same diameter and overall length were maintained

123 for all codends. The non-selective control codend was constructed of double 4 mm diamond

124 shaped twine with a nominal mesh size of 64 mm, 125 meshes long and 125.5 meshes around

125 (Table 1). The test codends (114, 140, and 165 mm nominal) were all constructed of diamond

126 double 5 mm, and were 70, 60, and 50 meshes long and 70.5, 60.5, and 50.5 meshes around,

127 respectively. A test codend was attached to one leg of the trouser trawl and the control codend

128 was attached to the other side. The side of the test and control codends was switched regularly

129 to avoid possible side-based effects. The same test codend was used for approximately three

130 days before switching to a new mesh size.

131 Tow locations were based on the captain’s knowledge, echo sounder signals (including bottom

132 topography), and a goal of a mix of redfish sizes. Tows were only made in daylight hours

133 following the practice of the fishery. Tow durations varied based on the captain’s assessment of

134 the volume of fish in the net and fishing ground conditions, and were consistent with

135 commercial practice. Duration decreased over time as the captain narrowed the search area

136 and successfully found fish. Very large catches were avoided due to catch processing delays 8

137 that might reduce the number of tows and affect quality of fish retained and survival of fish

138 escaped.

139

140 The length of warp used was set by the captain based on the water depth and the bottom

141 topography of the tow track. The range of tow speed was also within normal operational

142 conditions for the species and was mainly influenced by tidal conditions, as was typical in

143 commercial operations.

144

145 2.2 Gear monitoring

146 A GoPro Hero2 high-definition camera (San Mateo, CA) with a deepwater underwater housing

147 and lights was mounted to view fish reaction in the mixing area during some tows. Net

148 geometry was measured using a trawl monitoring system (Notus Electronics, St. John’s,

149 Newfoundland) with sensors on both doors, the trawl’s wing ends, just behind the headrope,

150 and on the 64-mm codend. The sensors were set to provide bottom temperature, door spread,

151 door heel (angle of the door to the right or left of the direction of travel), wing spread and to

152 indicate when the control codend was full. In addition, these sensors can provide distance from

153 the sensor to the hydrophone. Bottom temperature was also recorded with previously

154 calibrated TidBit temperature recorders (Onset Computers, Inc., Pocasset, Massachusetts).

155

156 2.3 Catch Sampling 9

157 Codends were hauled on deck one at a time, with the codend attached to the shorter “leg”

158 hauled and emptied first. Catches from the experimental and control codends were deposited

159 in separate areas on deck, and processed separately.

160 The total catch of redfish per tow was determined to 0.1 kg with subsampling when there was

161 a large amount of catch and, on some tows, legal and sublegal catch amounts were quantified.

162 Lengths (measured as midline length, MLL) of a random subsample of more than 100 redfish (if

163 possible) from each codend from each tow were measured to the nearest cm. For length-

164 frequency (LF) analysis, counts at each length were multiplied by the subsample weight divided

165 into the total weight.

166 Other organisms were also identified, and weighed to the nearest 0.1 kg. Weights were directly

167 measured or quantitatively determined; for example, by basket counts.

168

169 2.4 Analysis

170 All catch data (along with trip and gear data) were entered and uploaded into a customized

171 relational database in Microsoft Access 2007. Collected data were analyzed using Microsoft

172 Excel and R statistical software (R Development Core Team, 2009), primarily using the lattice

173 package (Sarkar, 2009) and SELNET, a selectivity analysis program. SELNET was developed to

174 acquire and analyze size selectivity and catch data for towed fishing gears, both at the haul level

175 and for a group of hauls (Frandsen et al., 2011; Herrmann et al., 2012; Herrmann et al., 2013). 10

176 The methods implemented in SELNET comply with accepted recommendations for the analysis

177 of size selectivity data (Fryer, 1991; Wileman et al., 1996).

178

179 To model the size selection first we used a logistic curve described by the parameters L50 and

180 the selection range SR (= L75 – L25) (Wileman et al., 1996). For each haul, the number of fish

181 counted in the experimental codend is described as ntl for the count of fish at each length l, and

182 in the control codend ncl. The proportion of the total catch measured for lengths is described by

183 the sampling rates qt (experimental) and qc (control). The size selection in each haul can then

184 be obtained by minimizing the following function with respect to the parameters L50, SR and

185 SP:

186

187

188

189

190 With:

191

192

193 11

194 SP is defined as the split parameter and expresses the assumed length independent relative

195 entry of fish to the test or control side of the gear during the fishing process. SP needs to be

196 estimated to assess the values of the selection parameters L50 and SR.

197

198 Fit statistics (i.e., the p-value and model deviance versus degrees of freedom (DOF)) were

199 inspected for individual hauls (Wileman et al. 1996). Where the p-value < 0.05 or the deviance

200 >> DOF, the residuals were examined for patterns or structural problems. Where no pattern

201 was seen, the poor fit was considered overdispersion in the data and the data were included.

202

203 The second step considered between-haul variation (Fryer 1991) using the results from all the

204 individual hauls simultaneously for the L50, SR and SP, together with their covariance matrix

205 and information on the values of the mesh size, m. In addition, we considered the effect of w

206 (total control codend catch weight (kg)) and S2, which side of the twin trawl the test codend

207 was attached to. Since one codend necessarily stayed in the water longer during haulback, and

208 could potentially lose more fish, it is prudent to test whether this longer hauling time might

209 impact size selectivity.

210

211 A model considering the potential effect of the parameters m, w and S2 was constructed with

212 the following form and applied in SELNET. 12

213

214

215

216

217

218 The parameters f0…f3, g0…g3 and h0…h3 are estimated while fitting the model to the data with

219 values for L50, SR and SP based on the selectivity results from the individual hauls. Models were

220 selected based on the AIC value (Akaike, 1974), while considering every possible simpler sub-

221 model following the procedure described in Wienbeck et al. (2011) and Herrmann et al. (2013).

222 Individual haul results were plotted for the L50 and SR with 95% CI versus the mean model

223 estimated values and the predicted 95% CI for the total variation (between-haul variation +

224 uncertainty around the mean). The lower and upper 95% CI for the estimated between-haul

225 variation in the selection parameters (lim L50, lim SR) were calculated by:

226

227

228

229

230 13

231 where L50mean , SRmean and SPmean are the predictions based on the selected submodel and D11 ,

232 D22 and D33 are the diagonal elements in the estimated between haul-variation matrix for the

233 selected model (Fryer, 1991).

234

235 These plots were inspected to see if the model predictions appeared to reflect the main trends

236 for the effects of catch size on the results for each codend and to inspect if the model was able

237 to describe the results for the individual hauls, considering the estimated between-haul

238 variation and uncertainty on the means in the selection process, in addition to the uncertainty

239 of the haul results. After successful model validation based on the above procedure, the models

240 were applied to predict size selection for codend mesh sizes between 80 and 170 mm.

241 For further consideration of the impact of different choices of mesh sizes, we used the length

242 distribution found in the 64 mm control codend as a representation of the overall population

243 size structure available to trawl gear. A simulation using our model results was then developed

244 using SELNET that estimated the distribution and number of fish predicted to be caught from a

245 similarly structured theoretical population of 1000 redfish using codends ranging from 114 mm

246 to 165 mm, in 12.7 mm steps.

247

248 3. Results

249 Tows were conducted generally east and northeast of Provincetown, Massachusetts, USA over

250 an area of approximately 4700 nm2 (Figure 1). Fifty-six tows were completed in two trips 14

251 carried out between 27 March and 1 April, and 3 April and 8 April 2013. Overall, 18 tows were

252 completed pairing the 114 mm mesh, with 10 on the starboard side; 16 tows with the 140 mm

253 with 9 on the starboard side; 22 tows with the 165 mm codend with 9 on the starboard side.

254

255 Tow duration ranged from 0.1 to 1.8 h with the median tow duration of 0.6 h (Figure 2). Median

256 depth fished was 100 fm, ranging from 50-131 fm; median towing speed was 3.0 knots (range:

257 2.6-3.3 knts); median warp length (wire out) was 225 fm (range: 200-275 fm). Median wave

258 height experienced was 1.2 m with a maximum of 3 m (Figure 2); these heights are unlikely to

259 substantially affect net performance both based on the captain’s decisions and as they are

260 within typical commercial operational conditions.

261

262 Trawl monitoring sensor readings indicated a median headrope height of 4.6 m (Interquartile

263 range (IQR): 3.8-9.0 m), a door spread of 80 m, and door heel medians of 6.2 degrees (IQR: 0.5-

264 6.25 degrees) (port) and 4.6 degrees (IQR: 1.5-7.25 degrees) (starboard) inward. Median

265 distance to doors was 423 m (IQR: 420-468 m); to the headrope 555 m (IQR: 553-561 m); to the

266 codend 603 m (IQR: 598-605 m) – these distances increased with depth and more wire out.

267 Based on net geometry, no anomalous tows were identified. Median temperature was 6.7

268 degrees (IQR: 6.5-7.1 degrees) and did not differ between mesh sizes.

269

270 The total catch of all species was just over 47,900 kg, with redfish comprising 42,482.9 kg or

271 89.7% of all catch. Pollock was the main bycatch species (3390.5 kg), with 21 other species with 15

272 catches greater than 10 kg total (Table 2). Over 18,000 redfish were measured. Lengths ranged

273 from 13 to 40 cm with distributions differing between codends (Error! Reference source not

274 found.).

275

276 Fifty-three hauls were included in redfish size selectivity analyses; two hauls could not be

277 included due to zero catches of redfish in one of the codends (experimental or control). Testing

278 of the full model, and all simpler sub-models (4096 models in all), resulted in selection of a sub-

279 model where the L50 and the SR depend only on mesh size:

280

281 L50 = 0.206 x mesh size (CI: 0.194, 0.217)

282 SR = 0.310 x mesh size (CI: 0.027, 0.035)

283 SP = 0.533 (CI: 0.458, 0.609)

284

285 The model estimated a split parameter near the ideal value of 0.5 with the 95% confidence

286 interval (CI) overlapping 0.5, indicating that redfish were equally likely to enter either codend

287 and that the trouser trawl was functioning properly. Video recordings collected from this area

288 also seemed to indicate no unusual fish behavior. The side the codend was on and the size of

289 the catch were not significant in the model and were removed.

290 16

291 The effect of w (total control codend catch weight (kg)) and S2, which side of the twin trawl the

292 test codend was attached to were not significant and were removed from the model.

293 Full logistic curves with 95% confidence intervals illustrating the catch curves for the codends as

294 measured were constructed from model results (Figure 4). L50 and selection ranges were

295 determined for the nominal 114 mm (L50: 22.3 cm; SR: 3.3 cm), 139 mm (L50: 29.2 cm; SR: 4.4

296 cm), and 165 mm (L50: 33.6 cm; SR: 5.0 cm) codends. Further validation of the selectivity model

297 was demonstrated by plotting L50 (Figure 5) and SR values (not shown)) for each individual haul,

298 with 95% confidence intervals as error bars, along with overall mean values predicted by the

299 final model, against the size of the codend catch (Error! Reference source not found.). Only six

300 hauls were found to have values of L50 outside the 95% CIs; all of these hauls overlapped their

301 error bars with the confidence band indicating that the model is an excellent fit to the data.

302

303 A similar comparison was made for the selection range. Eight individual hauls had SR values

304 outside the overall CI for the mean SR. All of these also had error bars overlapping the CI band.

305 These combined results indicated that the fit of the individual hauls to the overall results were

306 excellent.

307

308 Model results were used to produce estimates of mean L50 and SR (Error! Reference source not

309 found.) across a broad range of mesh sizes to support the choice of appropriate mesh sizes.

310 Model results can be used to estimate these values for both larger and smaller meshes, but

311 expansion outside the tested range is less reliable and likely unnecessary. 17

312

313 Additionally, the estimated escape of fish through the codend meshes can be inferred from the

314 difference between the predicted number of fish caught and the theoretical population of 1000

315 fish (Error! Reference source not found.). It is predicted that only 3.6% (36/1000) of redfish in

316 numbers would be retained by a 165 mm codend, 14.9% (149/1000) for a 140 mm codend, and

317 49.2% (492/1000) by a 114 mm codend.

318

319 4. Discussion

320 Testing and analysis of the selectivity of the codends provided results that by all and multiple

321 measures are robust. These results support observations by industry members and others that

322 the current mandatory minimum mesh size of 165 mm is a mismatch to the MLS. These results

323 should be used to identify appropriate codend mesh sizes for sustainable harvesting of the

324 species and for incorporating into stock assessment models.

325

326 Additionally, this study provides information on potential bycatch in smaller mesh fisheries.

327 Despite the use of a non-selective codend much smaller than the mandatory minimum (64

328 mm), bycatch levels were very low. Since the gear used is substantially similar to groundfish

329 trawls that target Atlantic cod, haddock, and pollock, our results illustrate the power of

330 choosing appropriate time, depth, and areas to sustainably target redfish with minimal bycatch.

331 18

332 The lack of substantial variation in geometry, geographical area, weather conditions, depth, and

333 temperature further suggests that uncontrollable sources of variability were limited during the

334 testing, and that the comparability of the codends was high. Additionally, since each tow

335 consisted of a selective and a non-selective codend, the results are consistent and accurate

336 within each tow. The lack of significant impact of catch size or which side the codends were on

337 also provided evidence for good functioning of the trouser trawl and randomization of the

338 distribution of fish between the test and control codends. The absence of any intercept terms

339 in the L50 and SR provided the logical result that as the mesh size is reduced to zero, the fish

340 length would similarly be reduced to zero. Wileman et al. (1996) recommend the use of a

341 trouser trawl with a full vertical split, but our design avoids a vertical panel based on the gear

342 designer’s recommendation and our prior experience (He and Balzano 2011). Vertical panels

343 are very difficult or impossible to rig without causing deformity or distracting motion in the net

344 and panel.

345 Selectivity studies relate mesh opening to fish length, which serves as a proxy for fish girth, a

346 more important morphological characteristic for determining codend escape. Fish girth can vary

347 due to condition and breeding status, and thus the retention probabilities can vary over time

348 (Wileman et al. 1996; Özbilgin et al. 2006). Additionally, twine type and thickness and other

349 gear parameters may also influence retention probabilities (Wileman et al. 1996). Variation in

350 selection between seasons, gears, weather and other factors is therefore expected (Pope et al.

351 1975; DeAlteris and Grogan 1997). While this potential for variability presents a challenge for

352 selecting an appropriate mesh size for this fishery, Herrmann et al. (2012) identified two cross- 19

353 sections of Sebastes encompassing the hard parts of the skull as important to codend

354 selectivity; their work suggests that size selectivity of Sebastes species may be subject to less

355 variation attributable to changes in girth. Nevertheless, consideration of possible variation in

356 retention curves should be given.

357

358 We originally considered use of a covered codend method. In consultation with industry

359 partners and others in the network, it was felt that a trouser trawl avoided some deck-handling

360 problems associated with covers, reduced risk of masking of codend meshes, and risk of over-

361 filling the cover, since we did not know what size catches might result from using a 64 mm

362 mesh. Few difficulties were encountered using the trouser trawl, as the crew was adaptable to

363 the hauling of two codends, and operationally it was not difficult to retrieve, and empty two

364 codends in a controlled and safe manner.

365 The size distribution of redfish S. fasciatus available to the fishery, as determined by the catch

366 in the 64 mm codend, is truncated compared to its near relatives S. mentella and S. marinus

367 (Robins and Ray, 1986). This limitation may prevent exploitation of broader markets as the

368 processing costs are higher and price lower for smaller fish. The relatively small size range of

369 this species of redfish should be a consideration in determining sustainable strategies for

370 harvesting and marketing redfish. Unlike species with a larger maximum size, implementation

371 of larger mesh sizes will not yield substantially greater growth with age.

372 20

373 Also, the current minimum landing size of 7 in (18 cm) represents the tail of the available size

374 distribution of redfish. Any mesh size larger than 64 mm will result in potentially large numbers

375 of legal-sized fish passing through the codend meshes, and becoming subject to escape

376 mortality. It is not currently known whether these fish would escape during fishing at the

377 depth, during hauling in midwater, or at surface, although some escape was observed at

378 surface during our fieldwork. Mortality rates of redfish escapees that exit the net during

379 different stages of fishing (towing, hauling, at surface) are likely different. It is generally agreed

380 that fish that escape during towing may suffer less mortality than during hauling or at surface

381 (Madsen et al. 2008). Redfish are notably more vulnerable than many other species to capture

382 and escape (Benoît et al., 2013) Escape at the surface increases mortality due to predation by

383 other fish and by seabirds and injury due to barotrauma and solar radiation (Grimaldo et al.

384 2009; Madsen et al. 2008). Escape is also prolonged by the impacts of barotraumas: everted

385 stomachs and an inability to control buoyancy.

386

387 The small maximum size of the population also suggests that larger mesh sizes, while yielding

388 larger sized fish, may result in catch rates too low to be commercially viable due to escape

389 through the codend meshes. The biology and population dynamics of this species should be

390 used to help identify an appropriate size of redfish to be targeted, including whether this target

391 should be at the L25, L50, or other retention level. The results of this study are also useful to

392 processors and marketers, in addition to the fishing fleet and managers, in terms of catch size

393 and volume of fish at length available to the market.

394 21

395

396 Emphasis should be placed on determination of when escapement occurs. Research on other

397 species has indicated that substantial escape could occur at the surface with all trawl gears.

398 Unobserved escape mortality can be minimized by investigating when or if escape occurs in the

399 codend during towing, and further if the use of a sorting grid, similar to the Nordmøre grate

400 used in shrimp trawl fisheries, can be used to efficiently and effectively exclude small fish while

401 the net is still on the seabed, resulting in minimal mortality.

402

403 Acknowledgements

404

405 Primary thanks are due to the F/V Guardian captain and crew: Capt. Bradford Horrell, Bobby

406 Blethen, José Garcia, Séan Farren, and Richard Walsh, and its owner, Mike Walsh, for their

407 wisdom, patience, friendship, and participation. Thanks are also go to Sally Sherman, Co-PI of

408 the REDNET project, for her support and advice. We appreciate assistance of Division of Marine

409 Fisheries staff biologists David Chosid for data collection and Mark Szymanski for contract and

410 other support. We acknowledge and appreciate the extensive involvement of the REDNET

411 network, which at present includes over 30 members. Funding was provided by the NOAA

412 Fisheries Northeast Cooperative Research Partners Program and a NOAA Interstate Fisheries

413 Management Support Grant.

414

415 References

416 22

417 Akaike, H., 1974. A new look at the statistical model identification. IEEE Trans. Auto. Control 19,

418 716-722.

419 Anthony, V.C., 1990. The New England groundfish fishery after 10 years under the Magnuson

420 Fishery Conservation and Management Act. N. Am. J. Fish. Mgmt. 10:175-184.

421 Benoît, H., Plante, S., Kroiz, M., Hurlburt, T., 2013. A comparative analysis of marine fish species

422 susceptibilities to discard mortality: effects of environmental factors, individual traits, and

423 phylogeny. ICES J. Mar. Sci. 70, 99–113.

424 Collette, B., Klein-MacPhee, G. 2002. Bigelow and Schroeder’s fishes of the Gulf of Maine, third

425 ed. Smithsonian Institution Press, Washington.

426 DeAlteris, J., Grogan, C., 1997. An Analysis of Harvesting Gear Size Selectivity for Eight Demersal

427 Groundfish Species in the Northwest Atlantic Ocean. Fisheries Technical Report #1,

428 University of Rhode Island Fisheries Center, Kingston, RI.

429 Fonteyne, R., 2005. Protocol for the Use of an Objective Mesh Gauge for Scientific Purposes.

430 ICES Coop. Res. Rep. No. 279.

431 Fogarty, M.J., Murawski, S. A., 1998. Large-Scale Disturbance and the Structure of Marine

432 Systems: Fishery Impacts on Georges Bank. Ecol. Appl. 8, S6-S22.

433 Frandsen, R.P., Herrmann, B., Madsen, N., 2010. A simulation-based attempt to quantify the

434 morphological component of size selection of Nephrops norvegicus in trawl codends. Fish.

435 Res. 101,156–167. 23

436 Fryer, R., 1991. A model of the between-haul variation in selectivity. ICES J. Mar. Sci. 48, 281–

437 290.

438 Greater Atlantic Regional Fisheries Office, 2014a. Updated annual catch limits for sectors,

439 common pool and herring mid-water trawl vessels for fishing year 2014.

440 http://www.greateratlantic.fisheries.noaa.gov/nr/2014/October/14mulcatchspecscarryoverphl.pdf.

441 Accessed 12 November 2014.

442 Greater Atlantic Regional Fisheries Office, 2014b . Northeast multispecies fishery final year-end

443 results for fishing year 2013.

444 http://www.greateratlantic.fisheries.noaa.gov/aps/monitoring/nemultispecies.html

445 Accessed 12 November 2014.

446 Grimaldo, E., Larsen, R.B., Sistiaga, M., Madsen, N., Breen, M., 2009. Selectivity and escape

447 percentages during three phases of the towing process for codends fitted with different

448 selection systems. Fish. Res. 95, 198–205.

449 He, P., Balzano, V,. 2011. Rope grid: A new grid design to further reduce finfish bycatch in the

450 Gulf of Maine pink shrimp fishery. Fish. Res. 111, 100-107.

451 doi:10.1016/j.fishres.2011.07.001

452 Herrmann, B., Sistiaga, M., Nielsen, K.N., Larsen, R.B., 2012. Understanding the size selectivity

453 of redfish (Sebastes spp.) in North Atlantic trawl codends. J. Northw. Atl. Fish. Sci. 44, 1–13. 24

454 Herrmann, B., Wienbeck, H., Moderhak, W., Stepputtis, D., Krag, L., 2013. The influence of

455 twine thickness, twine number and netting orientation on codend selectivity. Fish. Res. 145,

456 22–36.

457 ICES, 2012. Report of the ICES-FAO Working Group on Fishing Technology and Fish Behaviour

458 (WGFTFB). ICES CM. SSGESST:07, 214 p.

459 Madsen, N., Skeide, R., Breen, M., Krag, L.A., Huse, I., Soldal, A.V., 2008. Selectivity in a trawl

460 codend during haul-back operation—An overlooked phenomenon. Fish. Res. 91,168–174.

461 Mayo, R.K., Col, L., Traver, M., 2006. Acadian Redfish, in: Mayo, R., Serchuk, F., Holmes, E.

462 (Eds.), Status of Fishery Resources off the Northeastern United States. NEFSC Resource

463 Evaluation and Assessment Division. pp. 1–16.

464 Murawski, S., Brown, R., Lai, H., 2000. Large-scale closed areas as a fishery-management tool in

465 temperate marine systems: the Georges Bank experience. Bull. Mar. Sci. 66, 775–798.

466 Northeast Fisheries Science Center (NEFSC). 2012. Assessment or Data Updates of 13

467 Northeast Groundfish Stocks through 2010. US Dept Commer, Northeast Fish Sci Cent Ref

468 Doc. 12-06; 789 p.

469 Özbilgin, H., Ferro, R.S.T., Robertson, J.H.B., Holtrop, G., Kynoch, R.J., 2006. Seasonal variation

470 in trawl codend selection of northern North Sea haddock. ICES J. Mar. Sci. 63, 737–748.

471 doi:10.1016/j.icesjms.2005.01.025

472 Pope, J.A., Margetts, A.R., Hamely, J.M., Akyüz, E.F., 1975. Manual of methods for fish stock

473 assessment, Part III - Selectivity of fishing gear. Rome. 25

474 R Development Core Team, 2009. R: A Language and Environment for Statistical Computing. R

475 Foundation for Statistical Computing, Vienna, Austria, http://www.R-project.org.

476 Robins, C.R., Ray, G.C., 1986. A field guide to Atlantic coast fishes. Houghton Mifflin, Boston.

477 Sarkar, D., 2009. Lattice: Lattice Graphics. R package version 0.17-26,

478 http://CRAN.Rproject.org/package=lattice.

479 Wienbeck, H., Herrmann, B., Moderhak, W., Stepputtis, D., 2011. Effect of netting direction and

480 number of meshes around on size selection in the codend for Baltic cod (Gadus morhua).

481 Fish. Res. 109, 80–88.

482 Wileman, D., Ferro, R.S.T., Fonteyne, R., Millar, R.B. (Eds.), 1996. Manual of methods of

483 measuring the selectivity of towed fishing gears. ICES Coop. Res. Rep. No. 215.

484

485 Table Labels

486 Table 1: Codend mesh measurements

487 Table 2: Catch weights (kg) by species and by mesh size (mm) where total catch exceeded 5 kg,

488 sorted by total weight. Note that the 64 mm mesh catches are separate for each tested codend

489 mesh size.

490

491 Figure Captions 26

492 Figure 1: Tow start locations by mesh size tested: blue circles = 114 mm; red x = 140 mm; black

493 crosses = 165 mm.

494 Figure 2: Operational and environmental variables in chronological order. Green arrowheads (<)

495 are tows testing the 165 mm codend; pink crosses is the 140 mm codend; blue circles are the

496 114 mm codend. Dashed red horizontal lines are panel medians.

497 Figure 3: Redfish length frequencies (cm) by mesh size (mm). Red dashed lines are the older

498 (larger) and newer (smaller) minimum landing sizes.

499 Figure 4: Selection curves (solid lines) for 114 (blue, left), 140 (red, center), and 165 (black,

500 right) mm codends, with 95% confidence bands in stippled lines.

501 Figure 5: L50 estimates by mesh size (mm) from individual hauls (blue circles) with 95%

502 confidence intervals (error bars) compared to modeled mean L50 (solid horizontal line) and

503 95% confidence intervals (stippled lines), depicted by catch weight (kg).

504 Figure 6: Predicted mean L50 (top) and SR (bottom) v mesh size for redfish (solid line) with 95%

505 confidence intervals (stippled lines).

506 Figure 7: Simulated catch distributions based on selectivity analysis using five different mesh

507 sizes (from bottom: 114; 127; 140; 152; 165 mm) from the observed population distribution,

508 scaled to 1000 fish. Numbers indicate the estimated number of fish retained by that codend

509 mesh size. Table 1 Codend Measurements

Mesh opening (mm) Mesh Size Diameter Length Circumference Pre-experiment Post-experiment Average (mm) (mm) (meshes) (meshes) Mean SD Mean SD Mean 64 Double 4 125 125.5 64 1.6 67.2 1.6 65.6 114 Double 5 70 70.5 106.3 2.5 110.6 2.8 108.4 140 Double 5 60 60.5 141.1 3 142.7 3.1 141.9 165 Double 5 50 50.5 163.4 3.1 163.1 4.1 163.2 Table 2 Catch

Codend Mesh Size (mm) Species 64 114 64 140 64 165 Total (kg) Redfish,Acadian Sebastes fasciatus 13,974.4 10,829.4 9,469.5 824.6 6,976.0 710.7 42,784.6 Pollock Pollachius virens 524.1 357.1 861.6 476.3 868.9 302.6 3,390.5 Cod, Atlantic Gadus morhua 62.2 94.7 67.9 96.8 233.2 85.8 640.6 Monkfish ( Goosefish) Lophius americanus 11.7 32.0 4.3 10.2 86.0 81.0 225.2 Lobster, American Homarus americanus 17.8 24.7 11.0 12.2 43.3 87.0 196.0 Dogfish, Spiny Squalus acanthias 32.5 38.8 12.9 0.7 91.8 176.7 Haddock Melanogrammus aeglefinus 9.1 15.0 22.9 22.2 34.7 9.5 113.4 Skate, Nk Rajidae 21.4 13.6 18.2 21.5 17.3 18.1 110.0 Seal, Gray Halichoerus grypus 100.0 100.0 Hake, Silver (Whiting) Merluccius bilinearis 11.5 4.6 24.7 0.4 44.6 1.7 87.4 Hake, White Urophycis tenuis 9.2 20.8 12.7 6.5 22.8 13.4 85.4 Hake, Red (Ling) Urophycis chuss 13.1 8.8 40.8 0.5 13.8 1.0 77.9 Flounder, American Plaice Hippoglossoides platessoides 15.0 11.8 7.6 7.0 14.2 9.1 64.6 Herring, Atlantic Clupea harengus 8.5 0.3 4.1 44.5 0.3 57.7 Herring, River, Nk* Alosa 9.6 0.8 12.7 33.6 0.7 57.4 Mackerel, Atlantic Scomber scombrus 4.2 2.4 0.4 35.7 0.5 43.2 Flounder, Witch Glyptocephalus cynoglossus 6.0 3.6 4.1 1.7 17.5 4.7 37.7 Cusk Brosme brosme 11.3 13.7 5.9 4.3 2.4 37.6 Halibut, Atlantic Hippoglossus hippoglossus 9.3 0.9 1.0 6.1 17.5 34.8 Squid, Atl Long-fin Doryteuthis pealeii 7.5 4.0 3.2 0.6 8.9 1.1 25.3 Sea Raven Hemitripterus americanus 4.3 1.9 2.5 3.9 2.9 4.6 20.1 Shad, American Alosa sapidissima 1.5 0.3 9.4 11.2 Ocean Pout Macrozoarces americanus 5.4 1.9 1.9 1.0 10.2 *Blueback or Alewife Figure 1

US Canada Gulf of Maine

Study Area

Figure 1: Tow start locations by mesh size tested: blue circles = 114 mm; red x = 140 mm; black crosses = 165 mm.

REDNET Component 3 Draft Final Report – Codend Selectivity Page 26 Figure 2 Tow Variables

Wire Out (fm)

<<<<< +

260

<< < < < oo o oo + + < <<

240

<< << oooo oooo ooo o++++++++++ ++ +< << < 220

200 o < Tow Speed (knt) Wave Ht (ft) o < 10 +

3.2 o o + + 8 ++ < o o + + + + 3.0 << <<<<<<<<<< o oo oo oo ooo+ + + ++ ++ ++++<<<< < < 6 < + < < < < o o < < << < oooo + + <<

2.8 + 4 o +++ ++ + +

<<<<<<< o + <<< o o o 2 o oooo ooo << 2.6 o o Depth (fm) Duration (h)

+ < + 120 oo + <<<< < 1.5 < + < < < < < o o + + < < o + < 100 << o < < o+ + < o o o < o < + + < < < o + < <<< < < o ++ 1.0 o oo o ++ < o o o o + + 80 o < o o o + < + < o + o + + 0.5 < o + + + < o + < o o < < o o + << 60 ++ + < o + < o o

10 20 30 40 50 Tow Number Figure 3 Histogram

64 mm

10

5

0 165 mm

10

5

0 140 mm

Percent of Total Percent 10

5

0 114 mm

10

5

0

15 20 25 30 35 40 Midline length (cm) Figure 4 3 Curves

1.0

0.8

0.6 Proportion 0.4

0.2

0.0 7 9

20 25 30 35 Midline length (cm and in) Figure 5 ObsModeled ● Mesh Size 165 mm 60

●

50 ● ●

● ● ● 40 ●

● ● ● ● ● ● 30 ● ● ● ● ●

20 ● 10 0

0 100 200 300 Mesh Size 140 mm

60 ●

●

50 ●

●

40 ● ● ● ● ● ●

30 ● (cm) ● ● ● ● 50 L 20 10 0

0 100 200 300 Mesh Size 114 mm 60 50 40

● ●

30 ● ●● ● ● ● ● ● ● ● ● ● ● ● 20 ● ● 10 0

0 1000 2000 3000 Catch weight (kg) Figure 6 Predicted

35

30 (cm and in) 50 L 25

9 Mean

20

7 4.5 5.0 5.5 6.0 6.5

110 120 130 140 150 160 Mesh size (mm and in)

6

5

4 Mean selection range (cm) Mean selection range

3

4.5 5.0 5.5 6.0 6.5

110 120 130 140 150 160 Mesh size (mm and in) Figure7 Simulated

200

1000

150

492

100 Number

288

50

149

76 36 0 7 9

15 20 25 30 35 40 Midline length (cm and in) Telephone Survey of the New England Retail Fresh Fish Sector

A report produced for REDNET

UMass Dartmouth Center for Market Research

May 26, 2011

Introduction

Seafood today has become a direct competitor to meat products which includes poultry as well as red meat. In 2010 fresh fish and seafood generated the largest share of meat products in the retail market with 53%, resulting in $8.3 billion in sales. Popular seafood items found on restaurant menus include salmon, tuna, shrimp and crab. Since 2005, U.S. retail sales for fish and seafood has grown 17% and forecasted to continue to grow at over 4% each year.

According to a study conducted by a global research firm, America’s diet is changing.

People are becoming more health conscious and trying to decrease intake of fatty foods and focus on more healthy ones such as fish. People of all demographics consider healthy eating extremely important and 37% of people say their diet had become healthier in recent years. The fishing industry as a whole has a chance to connect with consumers by encouraging this health conscious trend and converting them to lifetime fish consumers. Expectations are that fishing industry will grow with a compound annual growth rate of 4.7% during 2011-2015.

Ocean Perch, also known as redfish, are slow to mature and were overfished in the

1930’s. They reside off the Northeast coast of the United States up to Canada as well as the

Pacific coast up into Alaska. The perch is a small fillet fish in comparison to the more popular types of fish but are 1/2 the price of flounder. One consumer survey shows 44% of consumers would try out a new species of fish/shellfish.

Today, the Ocean Perch is the 3rd most abundant fish in the Northeast. Local fishermen want to find a market for this fish because of the vast amount in this area. This creates opportunity for significant sales from bringing this new fish into the market. The Ocean Perch will sell for approximately 50-60 cents a pound. This means that in order to make a profit, a large

1 quantity would have to be caught and sold. This fish lives near underwater hills which makes it difficult to fish and requires boats with a lot of power to drag their nets along these hills.

A problem with this fish is that it is a relatively small fish measuring from 8-20 inches, and also consumers are used to a large fillet when they order so this could be an issue. This small size makes filleting the fish labor intensive and difficult. Another problem is that many fisheries do not have the proper equipment to cut this type of fish. Also, there is a catch limit of 15 million pounds for redfish. This is not a problem yet because last year they only caught 3.7 million pounds. There are also regulations and fishing permits needed to fish.

Forty-six percent of the permits to fish Ocean Perch go to boats 78 ft. or longer. These same boats hold 4% of total fishing permits in the area. The government gave the permits to boats that had already caught some of this fish and most are larger vessels. If government restrictions can be mitigated then there can be a huge profitable incentive to sell this fish.

There are many health benefits for people that enjoy the taste of fish. Although fish is one of the most popular mealtime foods for many Americans in the Northeast, in the past the market has not done well.

Currently, we are seeing an overabundance of redfish. This fish is one of the major upcoming sellers for fisheries in this region. One of the strong selling points for this fish when comparing it to an equivalent serving of Salmon, is that it has only one-tenth the amount of saturated fat and approximately one-third the amount of calories from fat. The protein content of

3.5 ounces of fish varies from Salmon’s 19.9 grams to 17.8 grams in the redfish. Another beneficial aspect of the redfish is the amount of omega-3 fatty acids to help lower blood

2 pressure, cholesterol, and prevent heart attacks. A 3.5 ounce serving of redfish contains .22 grams of omega-3, which can be up to 100 times more than chicken or beef.

With a vast amount of redfish potentially flooding the fish market in the area, we examine the local businesses to better understand the potential size of the market. With a new fish on the market, we expect an increase in demand of fish in this region. Our study investigates the attitudes and beliefs of local businesses towards redfish.

Statistics provided by Mintel (Global Consumer Product and Market Research). Available at Mintel.com

3

Executive Summary

• All respondents carry fish other than shellfish.

• Fifty-five percent of businesses that carry fish mostly come from New England (outside Massachusetts), 19% comes from outside New England, 14% comes from Massachusetts, 10% comes from outside the U.S. and 1% did not know where the fish they carry comes from.

• Fifty-five percent of businesses carry 0%-10% frozen fish, 13% carry 11%-24% frozen fish, 14% carry 25%-49% frozen fish, 10% carry 50%-74% frozen fish and 8% carry 75%-100% frozen fish.

• Nine percent of businesses carry 0-10% fresh fish, 7% carry 11-24% fresh fish, 0% carry 25-49% fresh fish, 15% carry 50-74% fresh fish and 69% carry 75-100% fresh fish.

• Ninety–four percent of businesses carry 0%-10% live fish, 3% carry 11%-24% live fish, 3% carry 25%-49% live fish and no one carries more than 49% live fish.

• Twenty-four percent of businesses carry under five types of fish, 23% carry 5-10 types of fish, 24% carry 11-15 types of fish, 13% carry 16-20 types of fish and 16% carry 21 or more types of fish.

• Thirty-seven percent of businesses said it is very important that the fish they carry is from Massachusetts when selecting which fish to carry, 35% said it is somewhat important, 19% said it is somewhat unimportant and 10% said it is very unimportant.

• Forty-one percent of businesses said that the size of the fillet is very important in selecting which fish to carry, 37% said it is somewhat important, 14% said it is somewhat unimportant and 8% said it is very unimportant.

• Eighty-six percent of businesses say that the smell of the fish is very important when selecting which fish to carry, 12% said it is somewhat important, 1% said it is somewhat unimportant and 1% said it is very unimportant.

• Seventy-eight percent of businesses said that the consumer demand is very important when selecting which fish to carry, 20% said it is somewhat important, 2% said it is somewhat unimportant and 0% said it is very unimportant.

• Eighty-six percent of businesses said that the appearance of the fish is very important when selecting which fish to carry, 12% said it is somewhat important, 1% said it is somewhat unimportant and 0% said it is very unimportant.

4

• Fifty-five percent of businesses said that the price of the fish is very important when selecting which fish to carry, 36% said it is somewhat important, 7% said it is somewhat unimportant and 2% said it is very unimportant.

• Sixty percent of businesses said that the place of origin is very important when selecting which fish to carry, 29% said it is somewhat important, 10% said it is somewhat unimportant and 1% said it is very unimportant.

• Thirty-seven percent of businesses are very familiar with redfish, 25% are somewhat familiar, 22% are somewhat unfamiliar and 16% are very unfamiliar.

• Thirty-five percent of businesses carry Redfish and 66% do not.

• Fourteen percent of businesses who do not currently carry Redfish would definitely consider carrying it, 39% said probably, 26% said probably not, 20% said definitely not and 1% did not respond.

• Twenty-eight percent of businesses sell 1-25 pounds of Redfish per month, 20% sell 26- 50 pounds, 4% sell 51-75 pounds, 46% sell 76 pounds or more and 2% did not respond.

• Twenty-two percent of businesses rate the taste of Redfish as excellent, 42% rate it good, 22% rate it fair, 4% rate it poor and 10% did not respond.

• Forty-two percent of businesses rate the freshness of Redfish as excellent, 36% rate it good, 16% rate it fair, 2% rate it poor and 4% did not respond.

• Forty percent of businesses rate the availability of Redfish as excellent, 24% rate it good, 22% rate it fair, 8% rate it poor and 6% did not respond.

• Twelve percent of businesses rate the green/eco labels as excellent, 24% rate it good, 12% rate it fair, 24% rate it poor and 28% did not respond.

• Twenty-eight percent of businesses rate the price of Redfish as excellent, 40% rate it good, 24% rate it fair, 2% rate it poor and 6% did not respond.

• Sixteen percent of businesses rate the size of the fillet for Redfish as excellent, 32% rate it good, 38% rate it fair, 6% rate it poor and 8% did not respond.

• Forty percent of businesses rate the smell of Redfish as excellent, 40% rate it good, 12% rate it fair, 2% rate it poor and 6% did not respond.

5

• Thirty-five percent of businesses are very aware of eco-labeling, 24% are somewhat aware, 15% are somewhat unaware and 24% are very unaware.

• Forty-seven percent of businesses said they definitely see value in carrying a fish with an eco-label, 35% said probably, 10% said probably not and 8% said definitely not.

• Fourteen percent of businesses said they are definitely willing to pay more for a fish that has an eco-label, 42% said probably, 28% said probably not and 16% said definitely not.

• Two percent of businesses have been at their current location for less than one year, 8% have been there 1-3 years, 13% have been there 4-6 years, 6% have been there for 7-9 years and 72% said they have been there 10 or more years.

• Seventy-two percent of businesses said they are an independently owned business, 20% said they are a large chain and 8% of businesses said they were something other than those mentioned.

• Forty-eight percent of businesses said they would be interested in obtaining more information on Redfish.

• Thirty-nine percent of restaurants said they are a Family Restaurant, 13% said they are a Bar and Grill, 26% said they are a Seafood Restaurant, 13% said they are Fine Dining and 9% of restaurants said they are classified as something other than those mentioned.

• Eleven percent of restaurants said they have a seating capacity of 26-50, 11% have a seating capacity of 51-70, 20% have a seating capacity of 76-100 and 58% have a seating capacity of 101 people or more.

• Thirty-three percent of restaurants said they have 1-2 chefs, 20% said they have 3-4, 30% said they have 5-6, 0% said they have 7-8 and 17% said they have 9 or more chefs.

• Four percent of restaurants said they are open seasonally and 96% said they are open year round.

• Seventy percent of restaurant said they are the Manager, 20% said they are the Owner, 7% said they are the Chef and 4% said they are the Server/Waitress.

6

Research Objective

RESEARCH OBJECTIVE: To investigate the demand for redfish among retailers along coastal Maine, Massachusetts, New Hampshire and Rhode Island.

The research goals aim to provide the following critical insights necessary to redfish:

 Preference towards size of fish  Freshness of product  Awareness of Redfish  Willingness to pay more for local (fresher fish)  Impressions of Eco-labeling  Encountered negative experiences with redfish

7

Methodology

RESEARCH DESIGN Survey

RESEARCH METHOD Telephone

SAMPLING DESIGN Probability

SAMPLING METHOD Systematic random sample through phonebook

SAMPLE Seafood restaurants and fish markets along coastal POPULATION Maine, Massachusetts, New Hampshire and Rhode Island. (N=1,509)

DATA COLLECTION March 2011 PERIOD

SAMPLE SIZE (ERROR AT THE 144 (+ 6.5%) 90% CONFIDENCE LEVEL)

8

Redfish Study

Conducted by the University of Massachusetts Center for Marketing Research

1. Does your business carry any fish other than shellfish? Yes ____ No___ (End Survey)

2. Where does the majority of the fish you carry come from? Massachusetts___ New England (outside MA)___ Outside NE in the US___ Outside US__ Don’t Know____

3. What percentage of your fish is: Frozen___ Fresh___ Live___ Other (please specify) ______

4. How many types of fish other than shellfish do you carry? Under 5 5-10 _____ 11-15 _____ 16-20 _____ 21 or More _____

5. What is your best selling fish? ______

6. How important are the following to you in selecting which fish to carry? Very Somewhat Somewhat Very Important Important Unimportant Unimportant Massachusetts Fish Size of Fillet Smell Consumer Demand Appearance Price Place of Origin

7. How familiar are you with Redfish? Very Familiar___ Somewhat Familiar____ Somewhat Unfamiliar____ Very Unfamiliar___ (Read following paragraph to everyone EXCEPT THOSE THAT ARE VERY FAMILIAR)

9

Redfish is a small fish yielding small fillets. This fish is mild tasting yet a bit sweet, with a moderately firm texture. The meat is lean, moist and flaky. Large Redfish develop a coarse texture. Deep-skinned Redfish with the fat line removed has the most delicate flavor. The flesh is white, though not as light as cod, and it turns opaque white when cooked.

8. Do you carry Redfish? Yes___ (skip to question 10) No___ If no, why not? ______

9. After hearing about Redfish would you consider carry it? Definitely ____ Probably____ Probably Not ___ Definitely Not____

If probably not or definitely not, why not? ______(Skip to question 12)

10. How many pounds of Redfish do you sell per month? 1-25___ 26-50___ 51-75___ 76 or More___

11. How would you rate Redfish on the following? Excellent Good Fair Poor Taste Freshness Availability Green/Eco Labels Price Size of Fillet Smell

12. How aware are you of eco-labeling? Very Aware___ Somewhat Aware__ Somewhat Unaware ___ Very Unaware _____

(Read following paragraph to everyone EXCEPT THOSE THAT ARE VERY FAMILIAR)

Eco-labels are affixed to products that pass eco-friendly criteria laid down by government, association or standards certification bodies. The criteria utilize extensive research based on the product's life cycle impact on the environment.

10

13. Do you see value in carrying fish that has an eco-label? Definitely____ Probably____ Probably Not____ Definitely Not____

14. Are you willing to pay more for fish that has an eco label? Definitely _____ Probably ____ Probably Not ____ Definitely Not ___

15. How many years have you been in business at this location?

Less than 1 year ____ 1-3____ 4-6____ 7-9 ____ 10 or More____

16. What best describes your business?

Large Chain___

Independently Owned____

Other (please specify) ______

17. Would you be interested in obtaining more information on redfish? Yes___ No___ (Restaurants continue below, fish/supermarkets end here)

18. What type of restaurant is this?

Family Restaurant___

Bar and Grill___

Seafood Restaurant___

Fine Dinning___

Ethnic Restaurant___

Other (please specify)______

19. What is your seating capacity? 1-25____ 26-50_____ 51-75_____ 76-100_____ 101 or More______

20. How many chefs do you have? 1-2____ 3-4_____5-6___ 7-8_____ 9 or More______

11

21. Are you open seasonally or year round? Seasonally ____ Year Round______

22. What is your position at the restaurant? Owner____ Manager____ Chef____ Other (please specify)______

To be filled out by student team:

Name of Business: ______

Full Address: ______

Phone Number: ______

12

Significance Test

Test Statistics

Where does Do you see value majority of the in carrying fish Do you carry fish you carry that has an eco-

Redfish? come from? label?

Chi-Square 13.966a 124.414b 63.855c

df 1 4 3 Asymp. Sig. .000 .000 .000

a. 0 cells (.0%) have expected frequencies less than 5. The minimum expected cell frequency is 72.5.

b. 0 cells (.0%) have expected frequencies less than 5. The minimum expected cell frequency is 29.0.

c. 0 cells (.0%) have expected frequencies less than 5. The minimum expected cell frequency is 36.3.

A Chi-Square test was performed on this data. The data in this study test significant at .000, which indicates the findings are statistically valid.

13

Question 1: Does your business carry any fish other than shellfish?

Does your business carry any fish other than shellfish? Cumulative Frequency Percent Valid Percent Percent

Valid Yes 145 100.0 100.0 100.0

Does your business carry any other fish other than shellfish? 100%

100% 90% 80% 70% 60% 50% 40% 30% 0% 20% 10% 0% Yes No

All respondents carry fish other than shellfish.

14

Question 2: Where does the majority of the fish you carry come from?

Where does majority of the fish you carry come from? Cumulative Frequency Percent Valid Percent Percent Local 20 13.8 13.8 13.8 Valid New England 80 55.2 55.2 69.0 Outside New England 28 19.3 19.3 88.3 Outside US 15 10.3 10.3 98.6 Don't Know 2 1.4 1.4 100.0 Total 145 100.0 100.0

Where does the majority of the fish you carry come from?

100% 90% 80% 70% 55% 60% 50% 40% 30% 19% 14% 10% 20% 1% 10% 0% Local New Outside New Outside Don't England England the US Know

Fourteen percent of businesses that carry fish receive it from local sources, 55% receive fish from New England (outside Massachusetts), 19% receive fish from outside New England, 10% receive fish from outside of the United States and 1% do not know where the fish they carry comes from.

15

Question 3a: What percentage of your fish is frozen?

What percentage of your fish is frozen? Cumulative Frequency Percent Valid Percent Percent Valid 0%-10% 79 54.5 54.9 54.9 11%-24% 19 13.1 13.2 68.1

25%-49% 20 13.8 13.9 81.9 50%-74% 15 9.7 9.7 91.7 75%-100% 12 8.3 8.3 100.0 Total 144 99.3 100.0 Total 145 100.0

What percentage of your fish is frozen?

100% 90% 80% 70% 55% 60% 50% 40% 30% 13% 14% 10% 20% 8% 10% 0% 0-10% 11-24% 25-49% 50-74% 75-100%

Fifty-five percent of businesses carry 0%-10% frozen fish, 13% carry 11%-24% frozen fish, 14% carry 25%-49% frozen fish, 10% carry 50%-74% frozen fish and 8% carry 75%-100% frozen fish.

16

Question 3b: What percentage of your fish is fresh?

What percentage of your fish is fresh? Cumulative Frequency Percent Valid Percent Percent 0%-10% 13 9.0 9.0 9.0 Valid 25%-49% 10 6.9 6.9 16.0 50%-74% 22 15.2 15.3 31.3 75%-100% 99 68.3 68.8 100.0 Total 144 99.3 100.0 N/A 1 .7

Total 145 100.0

What percentage of your fish is fresh?

100% 90% 69% 80% 70% 60% 50% 40%

30% 15% 20% 9% 0% 7% 10% 0% 0-10% 11-24% 25-49% 50-74% 75-100%

Nine percent of businesses carry 0%-10% fresh fish, 0% carry 11%-24% fresh fish, 7% carry 25%-49% fresh fish, 15% carry 50%-74% fresh fish and 69% carry 75%-100% fresh fish.

17

Question 3c: What percentage of your fish is live?

What percentage of your fish is live?

Cumulative Frequency Percent Valid Percent Percent

0%-10% 136 93.8 94.4 94.4 Valid 11%-24% 4 2.8 2.8 97.2 25%-49% 4 2.8 2.8 100.0

Total 144 99.3 100.0 N/A 1 .7

Total 145 100.0

What percentage of your fish is live?

94% 100% 90% 80% 70% 60% 50% 40% 30% 20% 3% 0% 0% 10% 3% 0% 0%-10% 11%-24% 25%-49% 50%-74% 75%-100%

Ninety–four percent of businesses carry 0%-10% live fish, 3% carry 11%-24% live fish, 3% carry 25%-49% live fish and no one carries more than 49% live fish.

18

Question 4: How many types of fish other than shellfish do you carry?

How many types of fish other than shellfish do you carry?

Cumulative Frequency Percent Valid Percent Percent Under 5 35 24.1 24.1 24.1 Valid 5-10 33 22.8 22.8 46.9 11-15 35 24.1 24.1 71.0 16-20 19 13.1 13.1 84.1 21 or More 23 15.9 15.9 100.0

Total 145 100.0 100.0

How many types of fish other than shellfish do you carry?

100% 90% 80% 70% 60% 50% 40% 24% 24% 30% 23% 13% 16% 20% 10% 0% Under 5 5 to 10 11 to 15 16 to 20 21 or More

Twenty-four percent of businesses carry under five types of fish, 23% carry 5-10 types of fish, 24% carry 11-15 types of fish, 13% carry 16-20 types of fish and 16% carry 21 or more types of fish.

19

Question 5: What is your best selling fish?

Response Frequency

Haddock (68)

Salmon (34)

Cod (20)

Tuna (4)

Flounder (2)

Tilapia (2)

Alaskan Pollock (1)

Fluke (1)

Hake (1)

Halibut (1)

Hiramatsu (1)

North Atlantic Ground (1)

Pollock (1)

Red Perch (1)

Scrod (1)

Sea bass (1)

Skate Dog Fish (1)

Squid (1)

Striped Bass (1)

Swordfish (1)

White Fish Blend (1) ______Total Responses 145

20

Question 6a: How important are the following to you in selecting which fish to carry? (From Massachusetts)

How important are the following to you in selecting which fish to carry? (From Massachusetts) Cumulative Frequency Percent Valid Percent Percent Very Important 53 36.6 36.6 36.6 Valid Somewhat Important 50 34.5 34.5 71.0

Somewhat Unimportant 27 18.6 18.6 89.7 Very Unimportant 15 10.3 10.3 100.0 Total 145 100.0 100.0

How important are the following to you in selecting which fish to carry? (From Massachusetts)

100% 90% 80% 70% 60% 37% 50% 35% 40% 19% 30% 10% 20% 10% 0% Very Somewhat Somewhat Very Important Important Unimportant Unimportant

Thirty-seven percent of businesses said it is very important that the fish they carry is from Massachusetts when selecting which fish to carry, 35% said it is somewhat important, 19% said it is somewhat unimportant and 10% said it is very unimportant.

21

Question 6b: How important are the following to you in selecting which fish to carry? (Size of Fillet)

How important are the following to you in selecting which fish to carry? (Size of Fillet)

Cumulative Frequency Percent Valid Percent Percent

Very Important 59 40.7 40.7 40.7 Valid Somewhat Important 54 37.2 37.2 77.9 Somewhat Unimportant 20 13.8 13.8 91.7 Very Unimportant 12 8.3 8.3 100.0 Total 145 100.0 100.0

How important are the following to you in selecting which fish to carry? (Size of Fillet)

100% 90% 80% 70% 60% 41% 50% 37% 40% 30% 14% 8% 20% 10% 0% Very Somewhat Somewhat Very Important Important Unimportant Unimportant

Forty-one percent of businesses said that the size of the fillet is very important in selecting which fish to carry, 37% said it is somewhat important, 14% said it is somewhat unimportant and 8% said it is very unimportant.

22

Question 6c: How important are the following to you in selecting which fish to carry? (Smell)

How important are the following to you in selecting which fish to carry? (Smell) Cumulative Frequency Percent Valid Percent Percent Very Important 124 85.5 85.5 85.5 Valid Somewhat Important 18 12.4 12.4 97.9 Somewhat Unimportant 2 1.4 1.4 99.3 Very Unimportant 1 .7 .7 100.0 Total 145 100.0 100.0

How important are the following to you in selecting which fish to carry? (Smell)

100% 86% 90% 80% 70% 60% 50% 40% 30% 12% 20% 1% 1% 10% 0% Very Somewhat Somewhat Very Important Important Unimportant Unimportant

Eighty-six percent of businesses say that the smell of the fish is very important when selecting which fish to carry, 12% said it is somewhat important, 1% said it is somewhat unimportant and 1% said it is very unimportant.

23

Question 6d: How important are the following to you in selecting which fish to carry? (Consumer Demand)

How important are the following to you in selecting which fish to carry? (Consumer Demand) Cumulative Frequency Percent Valid Percent Percent Valid Very Important 113 77.9 77.9 77.9

Somewhat Important 29 20.0 20.0 97.9 Somewhat Unimportant 3 2.1 2.1 100.0 Total 145 100.0 100.0

How important are the following to you in selecting which fish to carry? (Consumer Demand)

100% 78% 90% 80% 70% 60% 50% 40% 20% 30% 20% 2% 0% 10% 0% Very Somewhat Somewhat Very Important Important Unimportant Unimportant

Seventy-eight percent of businesses said that the consumer demand is very important when selecting which fish to carry, 20% said it is somewhat important, 2% said it is somewhat unimportant and 0% said it is very unimportant.

24

Question 6e: How important are the following to you in selecting which fish to carry? (Appearance)

How important are the following to you in selecting which fish to carry? (Appearance) Cumulative Frequency Percent Valid Percent Percent Very Important 125 86.2 86.2 86.2 Valid Somewhat Important 18 12.4 12.4 98.6

Somewhat Unimportant 2 1.4 1.4 100.0 Total 145 100.0 100.0

How important are the following to you in selecting which fish to carry? (Appearance)

100% 86% 90% 80% 70% 60% 50% 40% 30% 12% 20% 1% 0% 10% 0% Very Somewhat Somewhat Very Important Important Unimportant Unimportant

Eighty-six percent of businesses said that the appearance of the fish is very important when selecting which fish to carry, 12% said it is somewhat important, 1% said it is somewhat unimportant and 0% said it is very unimportant.

25

Question 6f: How important are the following to you in selecting which fish to carry? (Price)

How important are the following to you in selecting which fish to carry? (Price) Cumulative Frequency Percent Valid Percent Percent Very Important 80 55.2 55.2 55.2 Valid Somewhat Important 52 35.9 35.9 91.0 Somewhat Unimportant 10 6.9 6.9 97.9 Very Unimportant 3 2.1 2.1 100.0

Total 145 100.0 100.0

How important are the following to you in selecting which fish to carry? (Price)

100% 90% 80% 70% 55% 60% 50% 36% 40% 30% 7% 20% 2% 10% 0% Very Somewhat Somewhat Very Important Important Unimportant Unimportant

Fifty-five percent of businesses said that the price of the fish is very important when selecting which fish to carry, 36% said it is somewhat important, 7% said it is somewhat unimportant and 2% said it is very unimportant.

26

Question 6g: How important are the following to you in selecting which fish to carry? (Place of Origin)

How important are the following to you in selecting which fish to carry? (Place of Origin) Cumulative Frequency Percent Valid Percent Percent Valid Very Important 87 60.0 60.0 60.0 Somewhat Important 42 29.0 29.0 89.0 Somewhat Unimportant 14 9.7 9.7 98.6 Very Unimportant 2 1.4 1.4 100.0 Total 145 100.0 100.0

How important are the following to you in selecting which fish to carry? (Place of Origin)

100% 90% 80% 60% 70% 60% 50% 29% 40% 30% 10% 20% 1% 10% 0% Very Somewhat Somewhat Very Important Important Unimportant Unimportant

Sixty percent of businesses said that the place of origin is very important when selecting which fish to carry, 29% said it is somewhat important, 10% said it is somewhat unimportant and 1% said it is very unimportant.

27

Question 7: How familiar are you with Redfish? (Ocean Perch)

How familiar are you with Redfish? (Ocean Perch) Cumulative Frequency Percent Valid Percent Percent Very Familiar 54 37.2 37.2 37.2 Valid Somewhat Familiar 36 24.8 24.8 62.1 Somewhat Unfamiliar 32 22.1 22.1 84.1 Very Unfamiliar 23 15.9 15.9 100.0 Total 145 100.0 100.0

How familiar are you with Redfish (Ocean Pearch)?

100% 90% 80% 70% 60% 37% 50% 40% 25% 22% 30% 16% 20% 10% 0% Very Somewhat Somewhat Very Familiar Familiar Unfamiliar Unfamiliar

Thirty-seven percent of businesses are very familiar with redfish, 25% are somewhat familiar, 22% are somewhat unfamiliar and 16% are very unfamiliar.

28

Question 8a: Do you carry Redfish?

Do you carry Redfish? Cumulative Frequency Percent Valid Percent Percent Valid Yes 50 34.5 34.5 34.5 No 95 65.5 65.5 100.0 Total 145 100.0 100.0

Do you carry Redfish?

100% 90% 66% 80% 70% 60% 35% 50% 40% 30% 20% 10% 0% Yes No

Thirty-five percent of businesses carry Redfish and 66% do not.

29

Question 8b: Why don’t you carry Redfish?

Response Frequency No Demand (36) No Market (22) Lack of availability (3) Never heard of it (3) Don’t know (2) Not popular (2) Not their decision (2) Bad shelf life, dry skin (1) Can’t make decision (1) Commodity fillet (1) Corporate- if they can get to all restaurants (1) Customers don’t want it (use as lobster bait) (1) Deal with local and more familiar fish (1) Demand is low in area because of Italian neighborhood (1) Don’t carry (1) Don’t do those types of dishes (1) Don’t want it (1) Hasn’t been promoted enough (1) Little menu for fish (1) Low demand in area because of Irish neighborhood (1) No orders for it (1) No requests, no knowledge (1) Not a fish they specialize in (1) Not available in the area (1) Not enough volume or demand (1) Not in their business (1) Not local, customers don’t like it (1) Only deal with Cod (1)

30

Question 8b: Why don’t you carry Redfish? (Continued)

Owner only wants Haddock (1) Satisfied with Haddock (1) Traditional restaurant doesn’t vary from familiar species (1) ______Total Responses 94

31

Question 9a: After hearing about Redfish would you consider carrying it?

After hearing about Redfish would you consider carrying it? Cumulative Frequency Percent Valid Percent Percent Definitely 13 9.0 13.7 13.7 Valid Probably 37 25.5 38.9 52.6 Probably Not 25 17.2 26.3 78.9 Definitely Not 19 13.1 20.0 98.9 No Response 1 .7 1.1 100.0 Total 95 65.5 100.0 N/A 50 34.5

Total 145 100.0

After hearing about Redfish would you consider carrying it?

100% 90% 80% 70% 60% 39% 50% 40% 26% 20% 30% 14% 20% 1% 10% 0% Definitely Probably Probably Not Definitely Not No Response

Fourteen percent of businesses who do not currently carry Redfish would definitely consider carrying it, 39% said probably, 26% said probably not, 20% said definitely not and 1% did not respond.

32

Question 9b: After hearing about Redfish, why wouldn’t you consider carrying it?

Response Frequency No Demand (14) No Market (6) Maybe for a special (2) Not popular (2) Doesn’t make decision (2) Already have a good set (1) Business does not carry that fish (1) Can’t make decision (1) Carry it randomly (1) Comfortable with what they offer (1) Demand is low due to Italian neighborhood (1) Don’t do those types of dishes (1) Don’t want it (1) Fresh and local (1) High demand for Haddock (1) Low demand (1) Menu change too costly (1) Needs more info (1) Not a fish they specialize in (1) Only want Cod (1) Owner only wants Haddock (1) Slow sales (1) No Response (1) ______Total Responses 45

33

Question 10: How many pounds of Redfish do you sell per month?

How many pounds of Redfish do you sell per month? Cumulative Frequency Percent Valid Percent Percent 1-25 14 9.7 28.0 28.0 Valid 26-50 10 6.9 20.0 48.0 51-75 2 1.4 4.0 52.0 76 or More 23 15.9 46.0 98.0 No Response 1 .7 2.0 100.0 Total 50 34.5 100.0 N/A 95 65.5

Total 145 100.0

How many pounds of Redfish do you sell per month?

100% 90% 80% 70% 60% 46% 50% 40% 28% 30% 20% 20% 4% 2% 10% 0% 1-25 lbs 26-50 lbs 51-75 lbs 76 lbs or No More Response

Twenty-eight percent of businesses sell 1-25 pounds of Redfish per month, 20% sell 26-50 pounds, 4% sell 51-75 pounds, 46% sell 76 pounds or more and 2% did not respond.

34

Question 11a: How would you rate Redfish on taste?

How would you rate Redfish on taste? Cumulative Frequency Percent Valid Percent Percent Excellent 11 7.6 22.0 22.0 Valid Good 21 14.5 42.0 64.0 Fair 11 7.6 22.0 86.0 Poor 2 1.4 4.0 90.0 No Response 5 3.4 10.0 100.0 Total 50 34.5 100.0 N/A 95 65.5

Total 145 100.0

How would you rate Redfish on taste?

100% 90% 80% 70% 60% 42% 50% 40% 22% 22% 30% 10% 20% 4% 10% 0% Excellent Good Fair Poor No Response

Twenty-two percent of businesses rate the taste of Redfish as excellent, 42% rate it good, 22% rate it fair, 4% rate it poor and 10% did not respond.

35

Question 11b: How would you rate Redfish on freshness?

How would you rate Redfish on freshness? Cumulative Frequency Percent Valid Percent Percent Valid Excellent 21 14.5 42.0 42.0 Good 18 12.4 36.0 78.0 Fair 8 5.5 16.0 94.0 Poor 1 .7 2.0 96.0 No Response 2 1.4 4.0 100.0 Total 50 34.5 100.0 N/A 95 65.5 Total 145 100.0

How would you rate redfish on freshness?

100% 90% 80% 70% 60% 42% 50% 36% 40%

30% 16% 20% 2% 4% 10% 0% Excellent Good Fair Poor No Response

Forty-two percent of businesses rate the freshness of Redfish as excellent, 36% rate it good, 16% rate it fair, 2% rate it poor and 4% did not respond.

36

Question 11c: How would you rate Redfish on availability?

How would you rate Redfish on availability? Cumulative Frequency Percent Valid Percent Percent Excellent 20 13.8 40.0 40.0 Valid Good 12 8.3 24.0 64.0 Fair 11 7.6 22.0 86.0 Poor 4 2.8 8.0 94.0 No Response 3 2.1 6.0 100.0 Total 50 34.5 100.0 N/A 95 65.5

Total 145 100.0

How would you rate Redfish on the availability?

100% 90% 80% 70% 60% 50% 40% 40% 24% 30% 22% 8% 20% 6% 10% 0% Excellent Good Fair Poor No Response

Forty percent of businesses rate the availability of Redfish as excellent, 24% rate it good, 22% rate it fair, 8% rate it poor and 6% did not respond.

37

Question 11d: How would you rate Redfish on green/eco-labels?

How would you rate Redfish on green/eco labels? Cumulative Frequency Percent Valid Percent Percent Excellent 6 4.1 12.0 12.0 Valid Good 12 8.3 24.0 36.0 Fair 6 4.1 12.0 48.0 Poor 12 8.3 24.0 72.0 No Response 14 9.7 28.0 100.0 Total 50 34.5 100.0 N/A 95 65.5

Total 145 100.0

How would you rate Redfish on the green/eco labels?

100% 90% 80% 70% 60% 50% 40% 28% 24% 24% 30% 12% 20% 12% 10% 0% Excellent Good Fair Poor No Response

Twelve percent of businesses rate the green/eco labels as excellent, 24% rate it good, 12% rate it fair, 24% rate it poor and 28% did not respond.

38

Question 11e: How would you rate Redfish on price?

How would you rate Redfish on price? Cumulative Frequency Percent Valid Percent Percent Valid Excellent 14 9.7 28.0 28.0 Good 20 13.8 40.0 68.0 Fair 12 8.3 24.0 92.0 Poor 1 .7 2.0 94.0 No Response 3 2.1 6.0 100.0 Total 50 34.5 100.0 N/A 95 65.5 Total 145 100.0

How would you rate Redfish on the price?

100% 90% 80% 70% 60% 40% 50% 40% 28% 24% 30% 20% 2% 6% 10% 0% Excellent Good Fair Poor No Response

Twenty-eight percent of businesses rate the price of Redfish as excellent, 40% rate it good, 24% rate it fair, 2% rate it poor and 6% did not respond.

39

Question 11f: How would you rate Redfish on size of the fillet?

How would you rate Redfish on size of fillet? Cumulative Frequency Percent Valid Percent Percent Excellent 8 5.5 16.0 16.0 Valid Good 16 11.0 32.0 48.0 Fair 19 13.1 38.0 86.0 Poor 3 2.1 6.0 92.0 No Response 4 2.8 8.0 100.0 Total 50 34.5 100.0 N/A 95 65.5

Total 145 100.0

How would you rate Redfish on the size of fillet?

100% 90% 80% 70% 60% 38% 50% 40% 32% 30% 16% 20% 8% 6% 10% 0% Excellent Good Fair Poor No Response

Sixteen percent of businesses rate the size of the fillet for Redfish as excellent, 32% rate it good, 38% rate it fair, 6% rate it poor and 8% did not respond.

40

Question 11g: How would you rate Redfish on smell?

How would you rate Redfish on smell? Cumulative Frequency Percent Valid Percent Percent Valid Excellent 20 13.8 40.0 40.0 Good 20 13.8 40.0 80.0 Fair 6 4.1 12.0 92.0 Poor 1 .7 2.0 94.0 No Response 3 2.1 6.0 100.0 Total 50 34.5 100.0 N/A 95 65.5 Total 145 100.0

How would you rate Redfish on the smell?

100% 90% 80% 70% 60% 40% 40% 50% 40% 30% 12% 6% 20% 2% 10% 0% Excellent Good Fair Poor No Response

Forty percent of businesses rate the smell of Redfish as excellent, 40% rate it good, 12% rate it fair, 2% rate it poor and 6% did not respond.

41

Question 12: How aware are you of eco-labeling?

How aware are you of eco-labeling?

Cumulative Frequency Percent Valid Percent Percent Very Aware 50 34.5 34.5 34.5 Valid Somewhat Aware 39 26.9 26.9 61.4 Somewhat Unaware 21 14.5 14.5 75.9 Very Unaware 35 24.1 24.1 100.0 Total 145 100.0 100.0

How aware are you of eco-labeling?

100% 90% 80% 70% 60% 50% 35% 27% 40% 24% 30% 15% 20% 10% 0% Very Somewhat Somewhat Very Aware Aware Unaware Unaware

Thirty-five percent of businesses are very aware of eco-labeling, 24% are somewhat aware, 15% are somewhat unaware and 24% are very unaware.

42

Question 13: Do you see value in carrying fish that has an eco-label?

Do you see value in carrying fish that has an eco-label? Cumulative Frequency Percent Valid Percent Percent Valid Definitely 68 46.9 46.9 46.9 Probably 51 35.2 35.2 82.1 Probably Not 15 10.3 10.3 92.4 Definitely Not 11 7.6 7.6 100.0

Total 145 100.0 100.0

Do you see value in carrying a fish that has an eco-label?

100% 90% 80% 70% 60% 47% 50% 35% 40% 30% 10% 8% 20% 10% 0% Definitely Probably Probably Not Definitely Not

Forty-seven percent of businesses said they definitely see value in carrying a fish with an eco- label, 35% said probably, 10% said probably not and 8% said definitely not.

43

Question 14: Are you willing to pay more for fish that has an eco-label?

Are you willing to pay more for fish that has an eco-label? Cumulative Frequency Percent Valid Percent Percent Definitely 20 13.8 13.8 13.8 Valid Probably 61 42.1 42.1 55.9 Probably Not 41 28.3 28.3 84.1 Definitely Not 23 15.9 15.9 100.0 Total 145 100.0 100.0

Are you willing to pay more fore a fish that has an eco-label?

100% 90% 80% 70%

60% 42% 50% 28% 40% 30% 14% 16% 20% 10% 0% Definitely Probably Probably Not Definitely Not

Fourteen percent of businesses said they are definitely willing to pay more for a fish that has an eco-label, 42% said probably, 28% said probably not and 16% said definitely not.

44

Question 15: How many years have you been in business at this location?

How many years have you been in business at this location?

Cumulative Frequency Percent Valid Percent Percent

Less than 1 year 3 2.1 2.1 2.1 Valid 1-3 years 11 7.6 7.6 9.7 4-6 years 19 13.1 13.1 22.8 7-9 years 8 5.5 5.5 28.3 10 or more years 104 71.7 71.7 100.0 Total 145 100.0 100.0

How many years have you been in business at this location?

100% 90% 72% 80% 70% 60% 50% 40% 30% 8% 13% 20% 2% 6% 10% 0% Less than 1 1- 3 4-6 7-9 10 or more year years years years years

Two percent of businesses have been at their current location for less than one year, 8% have been there 1-3 years, 13% have been there 4-6 years, 6% have been there for 7-9 years and 72% said they have been there 10 or more years.

45

Question 16: What best describes your business?

What best describes your business? Cumulative Frequency Percent Valid Percent Percent Large Chain 29 20.0 20.0 20.0 Valid Independently Owned 104 71.7 71.7 91.7 Other 12 8.3 8.3 100.0 Total 145 100.0 100.0

What best describes your business?

100% 90% 72% 80% 70% 60% 50% 40% 20% 30% 8% 20% 10% 0% Large Chain Independently Owned Other

Twenty percent of businesses said they are a large chain, 72% said they are independently owned and 8% of businesses said they were something other than those mentioned.

46

16a: What best describes your business? (Other)

Response Frequency

Family Owned (4)

Corporation (1)

Importer (1)

Privately owned seafood (1)

Publicly Traded (1)

Small Chain (1)

Small Corporation (1)

Small Seafood Broker (1)

Wholesale Distributor (1) ______Total Responses 12

47

Question 17: Would you be interested in obtaining more information on Redfish?

Would you be interested in obtaining more information on Redfish?

Cumulative Frequency Percent Valid Percent Percent

Valid Yes 69 47.6 47.6 47.6

No 76 52.4 52.4 100.0

Total 145 100.0 100.0

Would you be interested in obtaining more information on Redfish?

100% 90% 80% 52% 70% 48% 60% 50% 40% 30% 20% 10% 0% Yes No

Forty-eight percent of businesses said they would be interested in obtaining more information on Redfish.

48

Question 18a: What type of restaurant is this?

What type of restaurant is this? Cumulative Frequency Percent Valid Percent Percent Valid Family Restaurant 18 12.4 39.1 39.1 Bar and Grill 6 4.1 13.0 52.2 Seafood Restaurant 12 8.3 26.1 78.3 Fine Dinning 6 4.1 13.0 91.3 Other 4 2.8 8.7 100.0 Total 46 31.7 100.0 N/A 99 68.3 Total 145 100.0

What type of restaurant is this?

100% 90% 80% 70% 60% 39% 50% 40% 26% 30% 13% 13% 9% 20% 10% 0% Family Bar Seafood Fine Other Restaurant and Grill Restaurant Dining

Thirty-nine percent of restaurants said they are a Family Restaurant, 13% said they are a Bar and Grill, 26% said they are a Seafood Restaurant, 13% said they are Fine Dining and 9% of restaurants said they are classified as something other than those mentioned.

49

Question 18b: What type of restaurant is this? (Other)

Response Frequency Counter service (1) High end steakhouse (1) Quick service (1) Steakhouse (1) ______Total Responses 4

50

Question 19: What is your seating capacity?

What is your seating capacity? Cumulative Frequency Percent Valid Percent Percent Valid 26-50 5 3.4 10.9 10.9 51-75 5 3.4 10.9 21.7 76-100 9 6.2 19.6 41.3 101 or More 27 18.6 58.7 100.0 Total 46 31.7 100.0 N/A 99 68.3 Total 145 100.0

What is your seating capacity?

100% 90% 80% 58% 70% 60% 50% 40% 20% 30% 11% 11% 20% 10% 0% 26-50 51-70 76-100 101 or more

Eleven percent of restaurants said they have a seating capacity of 26-50, 11% have a seating capacity of 51-70, 20% have a seating capacity of 76-100 and 58% have a seating capacity of 101 people or more.

51

Question 20: How many chefs do you have?

How many chefs do you have? Cumulative Frequency Percent Valid Percent Percent Valid 1-2 15 10.3 32.6 32.6 3-4 9 6.2 19.6 52.2 5-6 14 9.7 30.4 82.6 9 or more 8 5.5 17.4 100.0 Total 46 31.7 100.0 N/A 99 68.3 Total 145 100.0

How many chefs do you have?

100% 90% 80% 70% 60% 50% 33% 30% 40% 20% 17% 30% 20% 10% 0% 0% 1-2 3-4 5-6 7-8 9 or more

Thirty-three percent of restaurants said they have 1-2 chefs, 20% said they have 3-4, 30% said they have 5-6, 0% said they have 7-8 and 17% said they have 9 or more chefs.

52

Question 21: Are you open seasonally or year round?

Are you open seasonally or year round? Cumulative Frequency Percent Valid Percent Percent Valid Seasonally 2 1.4 4.3 4.3 Year Round 44 30.3 95.7 100.0 Total 46 31.7 100.0 N/A 99 68.3 Total 145 100.0

Are you open seasonally or year round?

96%

100% 90% 80% 70% 60% 50% 40% 30% 4% 20% 10% 0% Seasonally Year Round

Four percent of restaurants said they are open seasonally and 96% said they are open year round.

53

Question 22a: What is your position at the restaurant?

What is your position at the restaurant? Cumulative Frequency Percent Valid Percent Percent Valid Owner 9 6.2 19.6 19.6 Manager 32 22.1 69.6 89.1 Chef 3 2.1 6.5 95.7 Other 2 1.4 4.3 100.0 Total 46 31.7 100.0 N/A 99 68.3 Total 145 100.0

What is your position at the restaurant?

100% 90% 70% 80% 70% 60% 50% 40% 20% 30% 20% 7% 4% 10% 0% Manager Owner Chef Other

Seventy percent of restaurant respondents said they are the Manager, 20% said they are the Owner, 7% said they are the Chef and 4% said they are the Server/ Waitress.

54

Additional Analysis (Selected Cases)

This is a report of all cases in which businesses that do carry Redfish responded to this survey. The question was: “Do you carry Redfish?” SPSS then eliminated the respondents who do not carry Redfish, leaving only respondents that do carry Redfish.

Where does majority of the fish you carry come from? Cumulative Frequency Percent Valid Percent Percent Valid Local 6 12.0 12.0 12.0 New England 18 36.0 36.0 48.0 Outside New England 21 42.0 42.0 90.0 Outside US 4 8.0 8.0 98.0 Don't Know 1 2.0 2.0 100.0 Total 50 100.0 100.0

How aware are you of eco-labeling? Cumulative Frequency Percent Valid Percent Percent Valid Very Aware 20 40.0 40.0 40.0 Somewhat Aware 8 16.0 16.0 56.0 Somewhat Unaware 7 14.0 14.0 70.0 Very Unaware 15 30.0 30.0 100.0 Total 50 100.0 100.0

Would you be interested in obtaining more information on redfish?

Cumulative Frequency Percent Valid Percent Percent Valid Yes 25 50.0 50.0 50.0

No 25 50.0 50.0 100.0

Total 50 100.0 100.0

Are you willing to pay more for fish that has an eco-label? Cumulative Frequency Percent Valid Percent Percent Valid Definitely 9 18.0 18.0 18.0 Probably 16 32.0 32.0 50.0 Probably Not 14 28.0 28.0 78.0 Definitely Not 11 22.0 22.0 100.0 Total 50 100.0 100.0 Of those who carry redfish about half get the fish they carry from inside New England and half outside New England.

55

Additional Analysis (Selected Cases)

This is a report of all cases in which businesses that do not carry Redfish responded to this survey. The question was: “Do you carry Redfish?” SPSS then eliminated the respondents who do carry Redfish, leaving only respondents that do not carry Redfish.

Where does majority of the fish you carry come from? Cumulative Frequency Percent Valid Percent Percent Valid Local 14 14.7 14.7 14.7 New England 62 65.3 65.3 80.0 Outside New England 7 7.4 7.4 87.4 Outside US 11 11.6 11.6 98.9 Don't Know 1 1.1 1.1 100.0 Total 95 100.0 100.0

How aware are you of eco-labeling? Cumulative Frequency Percent Valid Percent Percent Valid Very Aware 30 31.6 31.6 31.6 Somewhat Aware 31 32.6 32.6 64.2 Somewhat Unaware 14 14.7 14.7 78.9 Very Unaware 20 21.1 21.1 100.0 Total 95 100.0 100.0

Would you be interested in obtaining more information on redfish?

Cumulative Frequency Percent Valid Percent Percent Valid Yes 44 46.3 46.3 46.3 No 51 53.7 53.7 100.0

Total 95 100.0 100.0

56

Are you willing to pay more for fish that has an eco-label? Cumulative Frequency Percent Valid Percent Percent

Valid Definitely 11 11.6 11.6 11.6 Probably 45 47.4 47.4 58.9 Probably Not 27 28.4 28.4 87.4 Definitely Not 12 12.6 12.6 100.0 Total 95 100.0 100.0

The majority of the respondents who do not carry redfish, carry local or New England fish (80%).

About half of those who do not carry redfish are interested in obtaining more information on redfish and would probably pay more for Eco-labeled fish.

57

Recommendations and Conclusions

The purpose of this study is to investigate the demand for redfish among retailers along coastal Maine, Massachusetts, New Hampshire and Rhode Island.

1. Preference towards size of fish Forty-one percent of businesses said that the size of the fillet is very important in selecting which fish to carry, 37% said it is somewhat important, 14% said it is somewhat unimportant and 8% said it is very unimportant. We recommend that if possible fish be caught at full maturity. Most respondents prefer a fish of a larger fillet. 2. Awareness of redfish Thirty-seven percent of businesses are very familiar with redfish, 25% are somewhat familiar, 22% are somewhat unfamiliar and 16% are very unfamiliar. There is a majority of companies that know about redfish but it appears that not a lot of them carry it due to lack of a market. We recommend that you get this fish into the consumers’ mind in order to raise awareness as well as consumer demand. We also concluded that the lack of demand is directly correlated with the low availability of this fish. 3. Willingness to pay more for local (fresher fish) Thirty-seven percent of businesses said it is very important that the fish they carry is from Massachusetts when selecting which fish to carry, 35% said it is somewhat important, 19% said it is somewhat unimportant and 10% said it is very unimportant. It is evident that small businesses and fish markets are willing to sacrifice price to a certain extent for local fish. However, larger businesses such as large chains and restaurants are more price-sensitive but still see the need to have New England caught fish.

4. Impressions of Eco-labeling Fourteen percent of businesses said they are definitely willing to pay more for a fish that has an eco-label, 42% said probably, 28% said probably not and 16% said definitely not. The majority of companies see a value in selling fish that has an eco-label. More than half of them are willing to pay more so we recommend that getting the eco-label would increase sales as well as boost the image of the fish. 5. Taste of redfish

Twenty-two percent of businesses rate the taste of redfish as excellent, 42% rate it good, 22% rate it fair, 4% rate it poor. After evaluating the qualities of redfish, we determined that the knowledgeable business owners that were talked to had a positive experience towards this fish. After talking to these people we realized that the bad stigma the fish receives is through the customers. We see the need to improve the image of redfish in the customers’ eyes as opposed to the retailers.

58

Respondent Contact Information

Would Like More Information Name Address Phone Number Yes Amaral's Fish Market New Bedford, MA 02746 508 - 996 - 1222 365 Hope Street Yes Leo's Ristorante Bristol , RI 02809 401 - 253 - 9300 1946 Post Rd. Yes Zeria Warwick, RI 02886 401 - 732 - 4343 484 Cambridge Street Yes Court Hose Fish Market Cambridge, MA 617 - 876 - 6716 622 Cambridge Street Yes New Deal Fish Market Inc. Cambridge, MA 617 - 876 - 8227 407 Harvard Street Yes Wulf's Fish Market Brookline, MA 02446 617 - 277 - 2560 2 Fish Pier Yes Bramante Seafood Boston, MA 617 - 261 - 0383 689 Salem Street Yes Fisherman's Fleet Malden , MA 781 - 322 - 5200 506 Main Street Yes Turners Seafood Grill and Market Melrose , MA 781 - 662 - 0700

Yes Anonymous Anonymous Anonymous 8 Seafood Way Yes Boston Sword and tuna Boston , MA 617 - 946 - 9850 140 Northern Ave Yes Anthony's Pier 4 Boston , MA 617 - 482 - 6262 800 Boylston Street Yes Top of the Hub Boston , MA 617 - 536 - 1775 283 Summer Street Yes Papagyo Boston , MA 617 - 423 - 10 00 78 Hathaway Road Yes Mozzarella's Pizza and Seafood New Bedford, MA 02746 508 - 997 - 5555 24 Washburn Street Yes Lou-Joe's Fresh Seafood New Bedford , MA 02740 508 - 993 - 0030 1105 State Road Yes Harbor Light Seafood Market Westport , MA 02790 508 - 677 - 9400 459 Washington Street Yes Snug Harbor Fish Company Duxbury, MA 02332 781 - 934 - 8167 1365 Fall River Ave Yes Tony's Seafood Seekonk , MA 02771 508 - 336 - 6800 74 Railroad Ave Yes The Town Dock Johnston, RI 02919 401 - 223 - 0540 45 Water Street Suite 1 Yes Sea-Trek Enterprise Inc. East Greenwich , RI 02818 401 - 884 - 3814 8 New England Way Yes Metro Lobster Seafood Inc. Warwick , RI 02886 401 - 737 - 5250

59

Respondent Contact Information

Would Like More Information Name Address Phone Number 365 Atwells Ave Yes Pane Vino Providence Providence , RI 02903 401 - 223 - 2230 1 Union Station Yes Capital Grille Providence Providence RI, 02903 401 - 521 - 5600 1450 Hartford Ave. Yes South Stream Seafood Johnston RI, 02919 401 - 272 - 6378 Yes Cape Quality Seafood Restaurant South Dartmouth , MA 0 2748 508 - 996 - 6725 77 Daniel Street Yes Press Room Restaurant Portsmouth, NH 03801 603 - 431 - 5186 Portsmouth Brewery and 56 Market Street Yes Restaurant Portsmouth, NH 03801 603 - 431 - 1175 409 The Hill Yes Blue Mermaid World Grille Portsmouth, NH 03801 603 - 4 27 - 2583 490 Lafayette Road Yes Old Salt at Lamies Inn Hampton, NH 03842 603 - 926 - 8322 379 Ocean Blvd. Yes Ron's Landing Hampton, NH 03842 603 - 929 - 2122 881 Lafayette Road Yes Wilbur's Family Dining Hampton, NH 03842 603 - 929 - 2962 595 Ocean blvd. Yes Lupo's Seafood and Grille Hampton, NH, 03842 603 - 926 - 3910 2263 Ocean Blvd, Yes Carriage House Restaurant Rye,NH 03870 603 - 964 - 8251 58 New Rochester Road Yes Strafford Farm Restaurant Dover, NH, 03820 603 - 742 - 7012 1 Washington street, Yes Fury's Public House Dover, NH 03820 603 - 617 - 3633 River Works Restaurant and 164 Main Street, Yes Tavern Newmarket, NH 03857 603 - 659 - 6119 9 Custom House Wharf Yes Harbor Fish Market Portland Me 207 - 772 - 6557 Yes Holyoke Wharf, Portland , ME 207 - 774 - 3411 55 Long Pond Drive Yes Stop and Shop, South Yarmouth, MA 508 - 394 - 1234 2 Portland Fish Pier Yes M F Foley Inc. Portland , ME 207 - 761 - 0818 Yes Stop and Shop South Dennis, MA 1 - 508 - 394 - 4535 510 Alfred Road Yes Shaw's Supermarket Biddleford, ME 1 - 207 - 283 - 8616 230 Fresser Blvd. Yes Mitchell's Fish Market Stamford, CT 06901 1 - 203 - 323 - 3474 99 Linwood Ave Yes Stop and Shop Colchester , CT 1 - 860 - 537 - 2361

60

Respondent Contact Information

Would Like More Information Name Address Phone Number 1400 Boston Post Rd. Yes Atlantic Seafood Old Saybrook , CT 1860 - 388 - 4527 2239 State Street Yes Number 1 Fish Market Hamden , CT 1 - 203 - 624 - 6171 226 East Main St. Yes ShopRite Supermarket Clinton , CT 860 - 669 - 0107 32 Halls Hill Road Yes John's Fish Market Colchester , CT 86 0 - 537 - 6635 399 Post Rd. Yes Whole Foods Westport , CT 203 - 227 - 6858 55 Welles Street Yes Whole Foods Galstonbury , CT 860 - 652 - 9800 1686 Boston Post Rd. Yes Whole Foods Milford, CT 06460 203 - 874 - 0883 118 South Main Street Yes Jess's Market Rockland, ME 207 - 596 - 6068 235 Camden Street Yes Shaw's Supermarket Rockland, ME 207 - 594 - 8615 455 Riverside Drive Yes Red Barn Augusta, ME 207 - 623 - 9485 2002 N Belfast Ave Yes Augusta Seafood Augusta, ME 207 - 6 22 - 6951 150 Western Ave Yes Shaw's Supermarket Augusta, ME 207 - 623 - 4520 11 Traders Way Yes Shaw's Supermarket Salem, MA 978 - 741 - 8660 252 Huron Ave Yes Fish Monger Cambridge, MA 617 - 661 - 4834 3 Parker Street Yes Northern Ocean Marine Glouster, MA 978 - 283 - 0222 54 Bridge Road Yes David's Fish Market Sallisbury , MA 978 - 462 - 2504 2 Cabot Street, Yes Rowand Fisheries Beverly, MA 978 - 927 - 1871 13 Sagamore Road, Yes Seaport Fish Rye, NH 603 - 436 - 7286 591 North Ave. Yes Whitecap International Seafood Wakefield, MA 781 - 245 - 4901 146 Front Street, Yes Fresh Fish Daily Exeter, NH 603 - 772 - 5011 20 Plaistow Road, Yes Donahues Fish Market Plaistow, NH 603 - 382 - 6181 28 Chestnut Street, Yes Hometown Seafoods Andover, MA 978 - 475 - 7426

61

Respondent Contact Information

Would Like More Information Name Address Phone Number 29 Whitten Road, Yes Hannaford's Supermarket Augusta, ME 207 - 622 - 3137 1250 A Eastern Ave Yes Molly's Seafood Chelsea, ME 207 - 588 - 0099 15 Alfred Street No Wharf Seafood Warwick , RI 401 - 732 - 4303 160 Charles Street No Savenor's Market Boston , MA 617 - 732 - 6328 1401 Washington Street No Morse Fish Co Boston , MA 617 - 262 - 9375 199- Clarendon Street No Skipjack's Seafood Emporium Boston , MA 617 - 536 - 3500 370 East Eagle Street No Channel Fish Boston, MA 617 - 569 - 3200 339 Northern Ave No East Bay Seafood Boston , MA 617 - 737 - 0890 377 Main Street No Captain Boston Fish Company Medford , MA 781 - 395 - 8138 400 American Legion Hwy. No Pine River Fish Market Revere, MA 781 - 284 - 9863 350 Copeland Street No Maloney Seafood Corp. Quincy , MA 617 - 472 - 1004 730 Centre Street No J.P. Seafood Cafe Jamaica Plain , MA 617 - 993 - 1327 783 Adams Street No Adams Fish Market Dorchester Center , MA 617 - 825 - 2090 21 Fish Pier No Ideal Seafood Boston, MA 617 - 482 - 9160 351 Hanover Street No Allora Restaurant Boston, MA 508 - 485 - 4300 63 Salem Street No Neptune Oyster Boston , MA 617 - 742 - 3474 41 Union Street No Union Oyster House Boston , MA 617 - 227 - 2750 1 Appleton Street No 28 Degrees Boston , MA 617 - 728 - 0728 360 Hathaway Road No Captain Frank's Seafood New Bedford, MA 508 - 992 - 3000 57 Hassey Street No Big G Seafood Inc. New Bedford, MA 02740 508 - 991 - 3200 50 Homers Wharf No International C Food Inc. New Bedford, MA 02704 508 - 996 - 4377 350 South Front Street No Carlos Seafood New Be dford, MA 02740 508 - 997 - 8971

62

Respondent Contact Information

Would Like More Information Name Address Phone Number 2234 State Road No Rock Bottom Seafood Plymoth MA, 02360 508 - 888 - 6678 136 Silver Lake Ave. No Quality Seafood Providence RI, 02909 401 - 437 - 6688 400 Massasoit Ave. Ste 103 No Seasource Imports Ltd. East Providence RI, 02914 401 - 435 - 4500 267 Ocean Street No Brant Rock Fish Market Marshfield MA, 02020 781 - 834 - 6231 127 Waterman Ave. No Atlantic Fish Market East Providence RI, 02914 401 - 490 - 2266 1702 Mineral Spring Ave No Captain's Catch North Providence , RI 02904 401 - 353 - 6350 473 East Washington Street No Fresh Catch North Attleboro, 02760 50 8 - 695 - 4182 40 Macondray Street No Cumberland Seafood Cumberland RI, 02864 401 - 728 - 6088 456 East Washington Street No Norm's Seafood Restaurant North Attleboro , 02760 508 - 645 - 3722 61 State Road No Not Your Average Joe North Dartmouth MA, 02747 508 - 992 - 5637 146 Huttleston Ave No Genes Famous Seafoods Fairhaven MA, 02719 508 - 996 - 5127 No Anonymous Anonymous Anonymous 189 State Street No Four Portsmouth, NH 03801 603 - 319 - 1547 28 Woodbury Ave No Shaw's Supermarket Portsmouth, NH 03801 603 - 436 - 0323 80 State Street No Rosa Restaurant Portsmouth, NH 03801 603 - 436 - 9715 55 Ceres Street No Oar House Portsmouth, NH 03801 603 - 436 - 4025 Herbert's Seafood and Steak 1500 Lafayette Road #8 No Restaurant Portsmouth, NH 03801 603 - 431 - 5882 127 Ocean Blvd. No Sea Ketch Restaurant Hampton, NH 03842 603 - 926 - 0324 94 New Rochester Road No Crossroads Restaurant Dover, NH 03820 603 - 74 0 - 0233 431 Dover Point Road No Newick's Lobster House Dover, NH 03820 603 - 742 - 3205 134 Front Street No Supreme Pizza and Seafood Exeter, NH 03833 603 - 778 - 8304 315 Us Route 1 No Bob's Clam Hut Kittery, ME 03904 207 - 439 - 4233

63

Respondent Contact Information

Would Like More Information Name Address Phone Number 33 Wharf Street No Street and Company Portland, ME 04101 207 - 77 5 - 0887 92 Commercial Street No Gilbert's Chowder House Portland, ME 04101 207 - 871 - 5636 88 Middle Street No Hugo's Restaurant Portland, ME 04101 207 - 774 - 8538 540 Forest Avenue No Great Lost Bear Portland, ME 04101 207 - 772 - 0300 617 Congress Street No Norm's Bar and Grill Portland, ME 207 - 828 - 9944 22 Monument Square No David's Restaurant Portland, ME 207 - 773 - 4340 5 Portland Pier No J's Oyster Portland, ME 207 - 772 - 4828 9 Mills Road, No Nunan's Lobster Hut Kennebunkport, ME 12 Portland Fish Pier No North Atlantic Inc. Portland , ME 207 - 774 - 6025 65 Independence Drive No Stop and Shop, Hyannis , MA 508 - 778 - 1774 30 Kildee Road No Georges Place Fish Market Chatham, MA 1 - 508 - 432 - 5493 995 Popuonnock Road No Big Y Food Stores Groton , CT 1 - 860 - 449 - 1229 248 Ilanders Road No Stop and Shop Niantic , CT 1 - 860 - 739 - 0024 535 Munroe Turnpike No Big Y Supermarket Munroe , CT 1 - 203 - 452 - 9002 100 Franklin Street No Shaw's Supermarket Westerly , RI 02891 401 - 348 - 8254 18 Water Street No Stop and Shop Vineyard Haven , MA 508 - 696 - 6047 9 West Road No Shaw's Supermarket Orleans , MA 508 - 240 - 1021 351 interstate 95frontage Road No Shoprite Supermarket New London , CT 860 - 447 - 1424 830 Boston Post Road No Big Y Supermarket Guilford , CT 203 - 453 - 3267 155 Cherry Street No Shoprite Supermarket Milford, CT 203 - 876 - 7868 355 Hawley Lane No Big Y Supermarkets Stratford, CT 203 - 386 - 0576 214 Spencer Street No Shoprite Supermarket Central Manchester , CT 860 - 645 - 3240

64

Respondent Contact Information

Would Like More Information Name Address Phone Number 75 Maverick street Suite3 No Hannaford's Supermarket Rockland, ME 207 - 594 - 2173 26 Hobart Street No Cherry Street Fish Market Danvers, MA 978 - 777 - 3449 801 Jubilee Drive No Product International Fishery Peabody, MA 978 - 750 - 5135 227 Highland Ave No Market Basket Salem, MA 978 - 741 - 0986 450 Paradise Road No Super Stop and Shop Swampscott, MA 781 - 596 - 1445 139 Endicott Street No Market Basket Danvers, MA 978 - 774 - 3683 30 Western Avenue Suite 301 No Yeoman Seafoods Inc Glouster, MA 978 - 283 - 7422 46 Mass Ave No Frankie's Catch of the Day Lexington, MA 617 - 484 - 6460 83 Hammond Street No Savage Carl W Seafoods Rowley, MA 978 - 948 - 7951 637 Lowell Street No Hannaford's Supermarket Peabody, MA 978 - 536 - 3200 2090 Atlantic Highway No Delano Seafood Markets Waldoboro, ME 207 - 832 - 7902 211 Union Street No Graffam Bros Rockport, ME 207 - 236 - 8391

65

66

Processing and Markeng Capacity for Acadian Redfish in New England

Dan Georgianna Department of Fishery Oceanography School for Marine Science and Technology University of Massachuses Dartmouth

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Definion of Economic Capacity

Economic Capacity is the maximum producon rate that allows firms to break even without sharp decreases in input prices

Assumes that exvessel buyers are compeve but have market power as a group

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Central Quesons

1. Is there sufficient market demand for redfish 2. Does processing and markeng capacity constrain supply 3. Can management regulaons promote or harm redfish harvesters or processors

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY OUTLINE

1. Historical paern of redfish supply & prices 2. Historical Paern of redfish processing 3. Current supply & demand factors 4. Market development for Redfish

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY 1.40 120,000

1.20 100,000 NE Exvessel 1.00 Landings prices 80,000

0.80

60,000 0.60

40,000 0.40

20,000 0.20

0 0.00 1950 1952 1954 1956 1958 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008

NE REDFISH LANDINGS AND EXVESSEL PRICES (ADJUSTED FOR INFLATION)

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY 6,000 1.40

1.20 5,000

1.00 4,000

Exvessel prices 0.80

3,000

0.60

2,000 NE Landings 0.40

1,000 0.20

0 0.00

NE REDFISH LANDINGS AND EXVESSEL PRICES (ADJUSTED FOR INFLATION)

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Historic paern of processing & markeng • 1950s to 1970s Filleted, frozen into blocks, cut, breaded, cooked & frozen into fillets, scks & proons • Sold as low price product in supermarkets • Marketed as Ocean Perch into U.S. Midwest markets

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Current paern of processing & markeng • High end: sell large fresh fillets into mid-west markets (retail fillet price from $3.99 to $4.99 per pound) • Low end: sell headed & gued frozen redfish to Chinese or other low cost producon centers, who export frozen fillets back to US (fillet price < $2.50 per pound)

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Current Market for Redfish

• Thin market for landings in high end market • Lile incenve to develop markets in compeve markets • Current investments in processing

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY

Thin market for landings

• Increases in landings has large negave effect on exvessel price (capacity) • Inelasc demand: increase in quanty caused decrease in revenue • Capacity constraints

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY

Market as Congesble Resource

• Congesble resources have open access but every persons use decrease usefulness (producvity) to other users • E.G. highways, fisheries (open access or user right system) • Market is open access resource

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Investments in markeng

• Differences between sales & markeng • Investors can’t secure gain full gain from markeng – Free riders • Lile incenve for firms to invest in markeng products in compeve markets

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Investment in markeng

Develop brand names Low input prices finance investment in markeng

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Current investments in processing

• Some purchases of Baader filleng machines

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Market development for Redfish

Sensible regulaon Government subsidies Industry-wide investments Firm-specific markeng investments Land large fish (> 1 pound)

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Sensible regulaon

• Stable or connuous increase in landings – No surprises or sudden shis • Posive examples: scallop area rotaon & markets • Negave example: 2004 SAP program for YT Flounders (June – August) Average Exvessel price during SAP was $0.43 per pound down from $0.92 per pound for same period in 2003

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY 2000 $1.20 Exvessel 1800 prices $1.00 1600

1400 $0.80 1200

1000 $0.60

800 NE Landings $0.40 600

400 $0.20 200

0 $0.00 Apr May Jun Jul Aug Sep

2004 NE YELLOWTAIL FLOUNDER LANDINGS AND EXVESSEL PRICES

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Government support for infrastructure

• Transportaon: highways, airports etc. • Subsidies for markeng

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Industry-wide investments

• “Got Milk?” • New Bedford Seafood Council from 1950s through 1970s – Adverzing, health reports, TV Chefs, recipes • Maine lobsters • Sectors? Markeng cooperaves (Ocean Spray)

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Firm-specific investments in processing • Redfish lines • Baader filleng & skinning machines

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Land large fish (> 1 pound)

• Only fillets > 2-3 oz sell in fresh market • Smaller fillets sell to China processors • Frozen into Blocks? for scks & porons

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Thanks to Cate O’Keefe & Steve Cadrin from SMAST; Chris Curran from Ocean Choice & Mark Lussier from Northcoast Seafood

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Processing and Markeng Capacity for Acadian Redfish in New England

Dan Georgianna Department of Fishery Oceanography School for Marine Science and Technology University of Massachuses Dartmouth

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Definion of Economic Capacity

Economic Capacity of harvesng sector is the maximum producon rate that allows firms normal profits without sharp decreases in input prices

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Central Quesons

1. Is there sufficient market demand for redfish to pay exvessel prices sufficient to pay all costs & normal profits? 2. Does processing and markeng capacity constrain supply prices? 3. Do management regulaons promote or harm redfish harvesters or processors?

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY Processing Capacity does not Constrain Supply Prices

Evidence • past performance • Current performance • Technological improvements • Interviews Does markeng capacity constrain supply prices?

• Telephone Survey by Center for Market Research at UMass Dartmouth

• Linear regression analysis on price determinaon Survey by Center for Market Research

• Survey done by 5 senior markeng majors directed by markeng professor & Director of Center for Markeng Research and me. • Several meengs to describe fishery & fish markets • Develop quesonnaire • Telephone survey of fish retail markets & restaurants • Interpret results Quesonnaire

17 quesons for all respondents with addional 4 quesons for restaurants (10-15 minutes)

• Preference towards size of fish • Freshness of product • Awareness of Redfish • Willingness to pay more for local (fresher fish) • Impressions of Eco-labeling • Encountered negave experiences with redfish

Survey Sample

• Seafood restaurants and retail fish markets along coastal Maine, Massachuses, New Hampshire and Rhode Island. (N=1,509) • Completed interviews with 99 retail fish markets (including super markets) • Completed interviews with 47 restaurants What is your position at the restaurant?

100% 90% 70% 80% 70% 60% 50% 40% 20% 30% 20% 7% 4% 10% 0% Manager Owner Chef Other What type of restaurant is this?

100% 90% 80% 70% 60% 39% 50% 40% 26% 30% 13% 13% 9% 20% 10% 0% Family Bar Seafood Fine Other Restaurant and Grill Restaurant Dining How many years have you been in business at this location?

100% 90% 72% 80% 70% 60% 50% 40% 30% 8% 13% 20% 2% 6% 10% 0% Les s than 1 1- 3 4-6 7-9 10 or more year years years years years How important are the following to you in selecting which fish to carry? (Smell)

100% 86% 90% 80% 70% 60% 50% 40% 30% 12% 20% 1% 1% 10% 0% Very Somewhat Somewhat Very Important Important Unimportant Unimportant How important are the following to you in selecting which fish to carry? (Consumer Demand)

100% 78% 90% 80% 70% 60% 50% 40% 20% 30% 20% 2% 0% 10% 0% Very Somewhat Somewhat Very Important Important Unimportant Unimportant How important are the following to you in selecting which fish to carry? (Price)

100% 90% 80% 70% 55% 60% 50% 36% 40% 30% 7% 20% 2% 10% 0% Very Somewhat Somewhat Very Important Important Unimportant Unimportant How important are the following to you in selecting which fish to carry? (Size of Fillet)

100% 90% 80% 70% 60% 41% 50% 37% 40% 30% 14% 8% 20% 10% 0% Very Somewhat Somewhat Very Important Important Unimportant Unimportant How aware are you of eco-labeling?

100% 90% 80% 70% 60% 50% 35% 27% 40% 24% 30% 15% 20% 10% 0% Very Somewhat Somewhat Very Aware Aware Unaware Unaware Do you see value in carrying a fish that has an eco-label?

100% 90% 80% 70% 60% 47% 50% 35% 40% 30% 10% 8% 20% 10% 0% Definitely Probably Probably Not Definitely Not Are you willing to pay more fore a fish that has an eco-label?

100% 90% 80% 70%

60% 42% 50% 28% 40% 30% 14% 16% 20% 10% 0% Definitely Probably Probably Not Definitely Not How familiar are you with Redfish (Ocean Pearch)?

100% 90% 80% 70% 60% 37% 50% 40% 25% 22% 30% 16% 20% 10% 0% Very Somewhat Somewhat Very Familiar Familiar Unfamiliar Unfamiliar How would you rate Redfish on taste?

100% 90% 80% 70% 60% 42% 50% 40% 22% 22% 30% 10% 20% 4% 10% 0% Excellent Go o d Fair Poor No Response How would you rate redfish on freshness?

100% 90% 80% 70% 60% 42% 50% 36% 40%

30% 16% 20% 2% 4% 10% 0% Excellent Go o d Fair Poor No Response How would you rate Redfish on the availability?

100% 90% 80% 70% 60% 50% 40% 40% 24% 30% 22% 8% 20% 6% 10% 0% Excellent Go o d Fair Poor No Response How would you rate Redfish on the price?

100% 90% 80% 70% 60% 40% 50% 40% 28% 24% 30% 20% 2% 6% 10% 0% Excellent Go o d Fair Poor No Response How would you rate Redfish on the size of fillet?

100% 90% 80% 70% 60% 38% 50% 40% 32% 30% 16% 20% 8% 6% 10% 0% Excellent Go o d Fair Poor No Response Do you carry Redfish?

100% 90% 66% 80% 70% 60% 35% 50% 40% 30% 20% 10% 0% Yes No

4% of restaurants carry redfish, 49% of fish markets carry redfish After hearing about Redfish would you consider carrying it?

100% 90% 80% 70% 60% 39% 50% 40% 26% 20% 30% 14% 20% 1% 10% 0% Definitely Probably Probably Not Definitely Not No Response

4% & 22% of restaurants & fish markets would consider carrying redfish

Would you be interested in obtaining more information on Redfish?

100% 90% 80% 52% 70% 48% 60% 50% 40% 30% 20% 10% 0% Yes No

45 % & 50% of restaurants & fish markets want more informaon

Does markeng capacity constrain supply prices?

Linear regression analysis on price determinaon

Simple relaonship between Quanty & Price Q = a +bP

y = ‐6E‐06x + 0.756 2010 R² = 0.186 1.8

1.6

1.4

1.2

1 2010 0.8 Linear(2010)

0.6

0.4

0.2

0 0 20000 40000 60000 80000 100000 120000 140000 Quanty as funcon of price

Demand equaon

QRF = f(priceRF, prices of substutes, taste, income, etc)

If quanty supplied is independent of price, then

PRF =g(QRF, Qsubstutes, other factors)

Preliminary Strategies

Improve product: size selecon Markeng: educaon, PR, instuonal scale buyers Timing: target redfish to match low quanes of substutes Thanks to Professor Nora Barnes and her students, Sravanthi Pasumarthi, Min-Yang Lee from NEFSC & Mark Lussier from Northcoast Seafood

UMASS SCHOOL OF MARINE SCIENCE & TECHNOLOGY

University of Massachusetts Dartmouth Summer Internship Project Report

Sai Sravanthi Pasumarthi Final Year Undergraduate Student School of Mechanical Sciences Indian Institute of Technology Bhubaneswar

9th May 11 – 8th July 11

Introduction

Commercial fishing is known to be America’s first industry, dating back to over 400 years. The tradition of commercial fishing in this country has a long and proud heritage. This industry continues to provide the economic, social and environmental backbone of the coastal communities.

As per definition each species of fish that is fished for by humans is called a fishery. There are hundreds of fisheries in the United States, all managed by the state or federal authorities. These fish under jurisdiction of the United States are considered a common property resource. Seafood is one of the most traded primary commodities in world. Despite the amount of seafood located in US waters, the US currently imports 80% of the seafood consumed in the US. This makes seafood the second most imported commodity in the US after oil (NOAA). The dockside seafood sales each year amount to over $28 billion and this only shows part of the nationwide importance of the commercial fishing industry. The revenue generated here on the landings supports innumerable jobs, from boat builders to bait suppliers, which strengthen the nation’s economy each year.

Talking about the North East Coast of America, particularly the New England region, commercial fishing industry is the third highest in terms of economic impact. The Massachusetts commercial fishing industry alone supports more than 80,000 jobs.

Fig.1 Species of fish found in the New England region

• The Acadian Redfish or the Ocean Perch is the 3rd most abundant fish in the North Eastern Coast of the USA.

• New England landings of the Acadian Redfish fell from more than 100,000 mt in the 1950s to less than 300 mt 1990s as shown in the graph below.

6,000 1.40

5,000 1.20

1.00 4,000 Ex-Vessel Prices 0.80

3,000 0.60 2,000 Net Ex vessel Landings 0.40

1,000 0.20

0 0.00

1985 2003 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2004 2005 2006 2007 2008 2009 1984 Fig.2 NET EXVESSEL REDFISH LANDINGS AND EXVESSEL PRICES

 As the landings fell considerably and the ex vessel prices remained high, the market for Redfish slowly diminished.  The cause for concern now is that the population of the Redfish in the waters has been increasing and promises to grow at a very high rate but fishermen aren’t very enthusiastic to catch this fish because of the lack of market.

 Since 2006, landings of Ocean Perch have increased by 40% per year.

 Though Ocean Perch has the advantage of being caught locally and close to the consumer, one would expect the consumer would favour domestically caught products as there is a growing concern among consumers about the freshness, nutritional value and environmental sustainability. But, this is not the case.

 The current pattern of processing and marketing of this fish include selling large fillets into the Midwest markets and selling headed and gutted frozen redfish to Chinese or other low cost manufacturers , who in turn export the frozen products back to the USA.

 The above processes have a lot of negative implications towards the quality of product and whether the low end manufacturers stick to the environmentally sustainable processes.

 The consumer would prefer to buy food that is produced close to home and is not shipped half way around the world for processing. This also helps in reducing energy used on food production and eliminates sustainability issues.

Our main objective is to determine the Economic Capacity of the market for this fish so as to encourage processors and manufacturers to set up sustainable units for the processing of this fish in the US.

Hypothesis:

 We initially assumed that there would be a break point in the data of the net landing verses the net ex vessel prices and upon multiple segmented regressions we could find this break point which would give us the economic capacity.

60 50 Breakpoint or 40 Economic Capacity 30 $ 20 pred2$ 10 0 -10 1 2 3 4 5 6 7 8 9 10 11 12 -20

Results:

 Having obtained the daily data of the years 1997 to 2010 from a local auction and performing regression analysis gave some interesting results.

 The figure below shows us a plot of the annual average price verses the annual average quantity for the years 1997 to 2010. One would expect that the graph should be downward sloping in accordance with any simple economic model represented by P= a Q + b Where P= Price Q= Quantity a & b= coefficients

Avg. Price vs Quantity 0.9 0.8 y = 5E-07x + 0.6464 0.7 R² = 0.131 0.6 0.5 Avg. Price vs Quantity 0.4 Linear (Avg. Price vs Quantity) 0.3 0.2 0.1 0 0 50000 100000 150000 200000 250000 300000

Fig.3 Annual Average Prices vs Quantity

 Also on plotting the data for every year separately we find that there are clumps of data points to the left side of the graph indicating that for the same ex vessel landings the prices of the fish have been varying sharply.

 This tells us that the Price of the fish in the market is not just a function of the quantity but there are other factors involved.

 Though it is impossible to pinpoint the exact factors affecting the price of the ocean perch in the market, it is safe to assume that factors such as the price of substitutes, landings of the fish on the previous day and their prices would affect the current price of the fish.

The following graphs show the plots of the yearly data (neglecting very low landings)

y = -2E-05x + 0.699 1997 R² = 0.025 1.8 1.6 1.4 1.2 1 1997 0.8 Linear (1997) 0.6 Linear (1997) 0.4 0.2 0 0 5000 10000 15000 20000 25000

Fig.4

1998 y = -3E-05x + 0.725 R² = 0.084 1.6 1.4 1.2 1

0.8 1998 0.6 Linear (1998) 0.4 0.2 0 0 2000 4000 6000 8000 10000 12000 14000 16000 18000

Fig.5 y = -2E-05x + 0.648 1999 R² = 0.050 1.4 1.2 1 0.8 1999 0.6 0.4 Linear (1999) 0.2 0 0 2000 4000 6000 8000 10000 12000 14000 16000 18000

Fig.6

y = -3E-05x + 0.577 2001 R² = 0.087 1.4 1.2 1 0.8 2001 0.6 0.4 Linear (2001) 0.2 0 0 2000 4000 6000 8000 10000 12000 14000

Fig.7

2003 y = -2E-05x + 0.620 R² = 0.052 1.6 1.4 1.2 1 0.8 2003 0.6 Linear (2003) 0.4 0.2 0 0 2000 4000 6000 8000 10000 12000 14000 16000

Fig.8 y = -3E-05x + 0.668 2004 R² = 0.154 1.6 1.4 1.2 1 0.8 2004 0.6 Linear (2004) 0.4 0.2 0 0 2000 4000 6000 8000 10000 12000 14000 16000 18000

Fig.9

y = -1E-05x + 0.683 2005 R² = 0.071 1.4 1.2 1 0.8 2005 0.6 0.4 Linear (2005) 0.2 0 0 5000 10000 15000 20000 25000 30000

Fig.10

y = -1E-05x + 0.835 2006 R² = 0.062 2

1.5

1 2006 Linear (2006) 0.5

0 0 5000 10000 15000 20000 25000 30000

Fig.11 y = -1E-05x + 0.732 2007 R² = 0.161 1.4 1.2 1 0.8 2007 0.6 Linear (2007) 0.4 0.2 0 0 5000 10000 15000 20000 25000 30000 35000 40000 45000

Fig.12

2008 y = -9E-06x + 0.738 R² = 0.222 2.5

2

1.5 2008 1 Linear (2008) 0.5

0 0 10000 20000 30000 40000 50000 60000

Fig.13

y = -9E-06x + 0.722 2009 R² = 0.140 5

4

3 2009 2 Linear (2009) 1

0 0 10000 20000 30000 40000 50000 60000 70000 80000

Fig.14 y = -6E-06x + 0.756 2010 R² = 0.186 1.8 1.6 1.4 1.2 1 0.8 2010 0.6 Linear (2010) 0.4 0.2 0 0 20000 40000 60000 80000 100000 120000 140000

Fig.15 All the above figures have clumps of data to the left and because of the very low R2 value performing a segmented regression would not yield any results.

We need a better mathematical model

Assumptions:

 The Price of fish depends on the Price of substitutes (flounders)  The quantity of fish landed in the previous days also affects the price of fish. Taking into account the quantities of the last 3 day should be a reasonable assumption.  Considering dummy variables for the years.

Working on the assumption of:

P = a1 Qt + a2 Qt-1 + a3 Q t-2 + a4 Pf + a5 D07 + a6 D08 + a7D09 + a8D10

Where P= Price of Redfish a1, a2 … a8 = Coefficients

D07 ….. D10 = Dummy Variables

Pf = Price of Flounders Q= Quantity Data for the second assumption:

Year Month Date Pt Qt Qt-1 Qt-2 Qt-3 Pf

2007 1 0 0 0 0 0 0 1.955166 1 0 0 0 2007 1 1 0.8 970 0 0 0 2.308799 1 0 0 0 2007 1 2 0.9112628 3516 970 0 0 2.747655 1 0 0 0 2007 1 3 0.54863003 6241 3516 970 0 2.73527 1 0 0 0 2007 1 4 0.59283626 12312 6241 3516 970 2.561105 1 0 0 0 2007 1 5 0.5114663 1439 12312 6241 3516 2.279829 1 0 0 0 2007 1 6 0 0 1439 12312 6241 2.196883 1 0 0 0 2007 1 7 0.89805689 3551 0 1439 12312 2.48094 1 0 0 0 2007 1 8 0.8744079 7389 3551 0 1439 2.357102 1 0 0 0 2007 1 9 0.97991071 448 7389 3551 0 2.416031 1 0 0 0 2007 1 10 0.77487897 3718 448 7389 3551 2.193522 1 0 0 0 2007 1 11 0.85707915 4485 3718 448 7389 2.18755 1 0 0 0 2007 1 12 0.71655953 3267 4485 3718 448 2.0934 1 0 0 0 2007 1 13 0 0 3267 4485 3718 1.926373 1 0 0 0 2007 1 14 1.0078555 2546 0 3267 4485 2.270965 1 0 0 0 2007 1 15 0.72446374 3916 2546 0 3267 2.477099 1 0 0 0 2007 1 16 0.99691174 10038 3916 2546 0 2.37689 1 0 0 0 2007 1 17 0 0 10038 3916 2546 2.362434 1 0 0 0 2007 1 18 0.90206054 7862 0 10038 3916 2.173397 1 0 0 0 2007 1 19 0.81905679 1039 7862 0 10038 2.165594 1 0 0 0 2007 1 20 0.88930937 2114 1039 7862 0 3.198432 1 0 0 0 2007 1 21 0.86394236 8467 2114 1039 7862 2.937115 1 0 0 0 2007 1 22 0.86346301 2622 8467 2114 1039 2.457848 1 0 0 0 2007 1 23 0.73429793 3041 2622 8467 2114 2.08373 1 0 0 0 2007 1 24 0.75702723 13590 3041 2622 8467 2.545226 1 0 0 0 2007 1 25 0.76144286 9897 13590 3041 2622 2.316327 1 0 0 0 2007 1 26 0.58445146 2534 9897 13590 3041 2.164539 1 0 0 0 2007 1 27 0.57604723 3557 2534 9897 13590 2.132141 1 0 0 0 2007 1 28 0.962 500 3557 2534 9897 2.153229 1 0 0 0 2007 1 29 0.71986817 2124 500 3557 2534 2.200932 1 0 0 0 2007 1 30 0.59299098 2882 2124 500 3557 2.302959 1 0 0 0 2007 1 31 0.57070017 19024 2882 2124 500 1.968738 1 0 0 0 2007 2 0 0 0 19024 2882 2124 1.817597 1 0 0 0 2007 2 1 0.48355739 9670 0 19024 2882 1.931447 1 0 0 0 2007 2 2 0.50244499 818 9670 0 19024 1.960611 1 0 0 0 2007 2 3 0.87043189 602 818 9670 0 2.188963 1 0 0 0 2007 2 4 0.94752187 2058 602 818 9670 1.900958 1 0 0 0 2007 2 5 0.73329652 5433 2058 602 818 2.076105 1 0 0 0 2007 2 6 0.90654086 7843 5433 2058 602 2.51548 1 0 0 0 2007 2 7 0.87087995 6614 7843 5433 2058 2.227357 1 0 0 0 2007 2 8 0.99559256 2042 6614 7843 5433 2.042685 1 0 0 0 2007 2 9 0 0 2042 6614 7843 2.09157 1 0 0 0 2007 2 10 0 0 0 2042 6614 2.148177 1 0 0 0 2007 2 11 1.2032176 1181 0 0 2042 2.290679 1 0 0 0 2007 2 12 0.87293442 11498 1181 0 0 2.101177 1 0 0 0 2007 2 13 0.66952032 14718 11498 1181 0 2.325844 1 0 0 0 2007 2 14 0.6423757 12611 14718 11498 1181 2.018564 1 0 0 0 2007 2 15 0.58551001 15735 12611 14718 11498 2.68226 1 0 0 0 2007 2 16 0 0 15735 12611 14718 1.771734 1 0 0 0 2007 2 17 0.6973001 1926 0 15735 12611 2.136919 1 0 0 0 2007 2 18 0.71389271 7997 1926 0 15735 2.125849 1 0 0 0 2007 2 19 0.90760711 3431 7997 1926 0 2.373836 1 0 0 0 2007 2 20 0.83499171 1206 3431 7997 1926 2.39396 1 0 0 0 2007 2 21 1.0070806 5508 1206 3431 7997 2.19411 1 0 0 0 2007 2 22 0.99119963 4318 5508 1206 3431 2.412082 1 0 0 0 2007 2 23 0.95063376 1499 4318 5508 1206 1.709306 1 0 0 0 2007 2 24 0.90425532 282 1499 4318 5508 1.872453 1 0 0 0 2007 2 25 0.67875798 3446 282 1499 4318 2.464802 1 0 0 0 2007 2 26 0.54379076 29287 3446 282 1499 1.99415 1 0 0 0 2007 2 27 0.70139427 2582 29287 3446 282 2.373216 1 0 0 0 2007 2 28 0.7577333 7694 2582 29287 3446 2.249182 1 0 0 0 2007 3 0 0 0 7694 2582 29287 1.975917 1 0 0 0 2007 3 1 0.6929258 13316 0 7694 2582 1.900722 1 0 0 0 2007 3 2 0.60469723 7068 13316 0 7694 1.807308 1 0 0 0 2007 3 3 0.88484848 165 7068 13316 0 1.61306 1 0 0 0 2007 3 4 0.94859155 1420 165 7068 13316 1.691894 1 0 0 0 2007 3 5 0.49564681 12979 1420 165 7068 1.455974 1 0 0 0 2007 3 6 0.82494279 7866 12979 1420 165 2.094712 1 0 0 0 2007 3 7 0.6972973 1110 7866 12979 1420 1.870031 1 0 0 0 2007 3 8 0.73119797 7087 1110 7866 12979 1.80794 1 0 0 0 2007 3 9 0 0 7087 1110 7866 2.095065 1 0 0 0 2007 3 10 0.65644576 3033 0 7087 1110 2.063221 1 0 0 0 2007 3 11 0.84969697 2475 3033 0 7087 2.255633 1 0 0 0 2007 3 12 0.83092784 485 2475 3033 0 2.194554 1 0 0 0 2007 3 13 0.72468676 11812 485 2475 3033 1.995548 1 0 0 0 2007 3 14 0.57763065 18427 11812 485 2475 1.993515 1 0 0 0 2007 3 15 0.49515121 13921 18427 11812 485 1.894428 1 0 0 0 2007 3 16 0.41515152 3300 13921 18427 11812 1.944519 1 0 0 0 2007 3 17 0 0 3300 13921 18427 2.030683 1 0 0 0 2007 3 18 0.73649485 4850 0 3300 13921 2.297105 1 0 0 0 2007 3 19 0.64377183 859 4850 0 3300 1.964771 1 0 0 0 2007 3 20 0.73040936 1710 859 4850 0 2.222992 1 0 0 0 2007 3 21 0.93604651 172 1710 859 4850 2.184136 1 0 0 0 2007 3 22 0.68383828 9473 172 1710 859 2.202726 1 0 0 0 2007 3 23 0.77517564 427 9473 172 1710 2.181921 1 0 0 0 2007 3 24 1.1084337 166 427 9473 172 2.127642 1 0 0 0 2007 3 25 1.0956023 523 166 427 9473 2.193152 1 0 0 0 2007 3 26 0.72033372 11267 523 166 427 2.221154 1 0 0 0 2007 3 27 0.78471567 11659 11267 523 166 2.404332 1 0 0 0 2007 3 28 0.55194969 11925 11659 11267 523 1.976662 1 0 0 0 2007 3 29 0.69320123 5854 11925 11659 11267 1.923602 1 0 0 0 2007 3 30 0.60275019 1309 5854 11925 11659 2.023504 1 0 0 0 2007 3 31 0 0 1309 5854 11925 2.016613 1 0 0 0 2007 4 0 0 0 0 1309 5854 2.068975 1 0 0 0 2007 4 1 0.6538835 6180 0 0 1309 2.09013 1 0 0 0 2007 4 2 0.57234213 8757 6180 0 0 2.075238 1 0 0 0 2007 4 3 0.74524232 3941 8757 6180 0 2.148951 1 0 0 0 2007 4 4 0.5668905 7744 3941 8757 6180 2.108318 1 0 0 0 2007 4 5 0.7170418 1555 7744 3941 8757 2.265385 1 0 0 0 2007 4 6 0.82185273 842 1555 7744 3941 2.113631 1 0 0 0 2007 4 7 0.71543408 622 842 1555 7744 2.142347 1 0 0 0 2007 4 8 0 0 622 842 1555 1.75061 1 0 0 0 2007 4 9 0.81342183 1356 0 622 842 2.247019 1 0 0 0 2007 4 10 0.84870317 10410 1356 0 622 2.474336 1 0 0 0 2007 4 11 1.0567412 5904 10410 1356 0 2.643849 1 0 0 0 2007 4 12 1.0131168 3202 5904 10410 1356 2.570544 1 0 0 0 2007 4 13 0.80430569 14446 3202 5904 10410 2.723484 1 0 0 0 2007 4 14 0 0 14446 3202 5904 2.236264 1 0 0 0 2007 4 15 0.9226895 3971 0 14446 3202 3.228051 1 0 0 0 2007 4 16 0.79356846 964 3971 0 14446 2.770971 1 0 0 0 2007 4 17 1.0358672 1868 964 3971 0 3.186055 1 0 0 0 2007 4 18 0 0 1868 964 3971 2.5348 1 0 0 0 2007 4 19 0 0 0 1868 964 2.93741 1 0 0 0 2007 4 20 0 0 0 0 1868 2.531942 1 0 0 0 2007 4 21 0 0 0 0 0 2.659661 1 0 0 0 2007 4 22 0 0 0 0 0 2.937336 1 0 0 0 2007 4 23 0.91578947 1140 0 0 0 2.676143 1 0 0 0 2007 4 24 0.93376866 1072 1140 0 0 2.440287 1 0 0 0 2007 4 25 1.0537634 651 1072 1140 0 2.256396 1 0 0 0 2007 4 26 0.88189655 1160 651 1072 1140 2.200668 1 0 0 0 2007 4 27 0.5086753 3631 1160 651 1072 2.103592 1 0 0 0 2007 4 28 0.49972238 1801 3631 1160 651 1.663612 1 0 0 0 2007 4 29 0.44819099 8126 1801 3631 1160 1.749891 1 0 0 0 2007 4 30 0.47232446 13857 8126 1801 3631 1.525462 1 0 0 0 2007 5 0 0 0 13857 8126 1801 1.447912 1 0 0 0 2007 5 1 0.60033141 12673 0 13857 8126 1.481596 1 0 0 0 2007 5 2 0.58922692 7593 12673 0 13857 1.561512 1 0 0 0 2007 5 3 0.21619586 5310 7593 12673 0 1.057761 1 0 0 0 2007 5 4 0.60888118 7499 5310 7593 12673 1.433495 1 0 0 0 2007 5 5 0.09223301 206 7499 5310 7593 1.595223 1 0 0 0 2007 5 6 0.59732143 2240 206 7499 5310 1.960705 1 0 0 0 2007 5 7 0.47807018 2280 2240 206 7499 1.819073 1 0 0 0 2007 5 8 0.52516148 6657 2280 2240 206 1.95808 1 0 0 0 2007 5 9 0.35503064 16807 6657 2280 2240 2.089706 1 0 0 0 2007 5 10 0.31486426 13482 16807 6657 2280 1.960064 1 0 0 0 2007 5 11 0 0 13482 16807 6657 1.800781 1 0 0 0 2007 5 12 0 0 0 13482 16807 2.053275 1 0 0 0 2007 5 13 0 0 0 0 13482 2.471508 1 0 0 0 2007 5 14 0.70098039 204 0 0 0 2.21465 1 0 0 0 2007 5 15 0.83634139 3269 204 0 0 2.230316 1 0 0 0 2007 5 16 0.76569038 478 3269 204 0 2.341219 1 0 0 0 2007 5 17 0.50163093 41694 478 3269 204 1.974364 1 0 0 0 2007 5 18 0.88803089 1036 41694 478 3269 2.492615 1 0 0 0 2007 5 19 0 0 1036 41694 478 2.308104 1 0 0 0 2007 5 20 0.90547798 931 0 1036 41694 2.353414 1 0 0 0 2007 5 21 0.85090909 550 931 0 1036 2.352579 1 0 0 0 2007 5 22 0.75010643 2349 550 931 0 2.164351 1 0 0 0 2007 5 23 0.60961092 6888 2349 550 931 2.180982 1 0 0 0 2007 5 24 0.64809536 3859 6888 2349 550 1.837173 1 0 0 0 2007 5 25 0 0 3859 6888 2349 1.657427 1 0 0 0 2007 5 26 0 0 0 3859 6888 1.650876 1 0 0 0 2007 5 27 0.65909091 572 0 0 3859 2.093165 1 0 0 0 2007 5 28 0.53720276 15644 572 0 0 1.906019 1 0 0 0 2007 5 29 0.39533574 30187 15644 572 0 1.524274 1 0 0 0 2007 5 30 0.6 1600 30187 15644 572 1.441749 1 0 0 0 2007 5 31 0.63584475 1752 1600 30187 15644 1.295328 1 0 0 0 2007 6 0 0 0 1752 1600 30187 1.155294 1 0 0 0 2007 6 1 0 0 0 1752 1600 1.239977 1 0 0 0 2007 6 2 0 0 0 0 1752 1.267092 1 0 0 0 2007 6 3 0.61783901 5516 0 0 0 1.211234 1 0 0 0 2007 6 4 0.47870412 4414 5516 0 0 1.120102 1 0 0 0 2007 6 5 0.39224138 464 4414 5516 0 1.072743 1 0 0 0 2007 6 6 0.45625692 2709 464 4414 5516 1.093139 1 0 0 0 2007 6 7 0.49928952 11260 2709 464 4414 1.105938 1 0 0 0 2007 6 8 0.45133992 709 11260 2709 464 1.115732 1 0 0 0 2007 6 9 0.47544643 448 709 11260 2709 1.020663 1 0 0 0 2007 6 10 0.51974146 7117 448 709 11260 1.193725 1 0 0 0 2007 6 11 0.48125378 9922 7117 448 709 1.229676 1 0 0 0 2007 6 12 0.45630015 9222 9922 7117 448 1.462242 1 0 0 0 2007 6 13 0.48707358 26148 9222 9922 7117 1.429778 1 0 0 0 2007 6 14 0.35104401 15565 26148 9222 9922 1.409296 1 0 0 0 2007 6 15 0 0 15565 26148 9222 1.547968 1 0 0 0 2007 6 16 0 0 0 15565 26148 1.722239 1 0 0 0 2007 6 17 0.33380848 19643 0 0 15565 1.63951 1 0 0 0 2007 6 18 0.4162103 4565 19643 0 0 1.755753 1 0 0 0 2007 6 19 0.45897064 8957 4565 19643 0 2.057488 1 0 0 0 2007 6 20 0.48343611 6973 8957 4565 19643 2.111522 1 0 0 0 2007 6 21 0.22142756 28384 6973 8957 4565 2.062455 1 0 0 0 2007 6 22 0.32188841 233 28384 6973 8957 1.946739 1 0 0 0 2007 6 23 0.52873563 174 233 28384 6973 1.920866 1 0 0 0 2007 6 24 0.42420575 3305 174 233 28384 1.873152 1 0 0 0 2007 6 25 0.44628957 1927 3305 174 233 1.962673 1 0 0 0 2007 6 26 0.3472066 5334 1927 3305 174 1.90396 1 0 0 0 2007 6 27 0.34992517 30738 5334 1927 3305 1.863457 1 0 0 0 2007 6 28 0.375 3224 30738 5334 1927 1.855245 1 0 0 0 2007 6 29 0.36325023 5415 3224 30738 5334 1.969895 1 0 0 0 2007 6 30 0 0 5415 3224 30738 1.707707 1 0 0 0 2007 7 0 0 0 0 5415 3224 1.764078 1 0 0 0 2007 7 1 0.35227779 6739 0 0 5415 1.732403 1 0 0 0 2007 7 2 0.36257583 2143 6739 0 0 1.867939 1 0 0 0 2007 7 3 0.53240741 648 2143 6739 0 1.930516 1 0 0 0 2007 7 4 0.49545455 440 648 2143 6739 1.695597 1 0 0 0 2007 7 5 0.56280992 1210 440 648 2143 1.925873 1 0 0 0 2007 7 6 0.63528139 924 1210 440 648 1.719868 1 0 0 0 2007 7 7 0.76769912 452 924 1210 440 1.912522 1 0 0 0 2007 7 8 0.79002876 1043 452 924 1210 1.613675 1 0 0 0 2007 7 9 0.60966142 4637 1043 452 924 1.716942 1 0 0 0 2007 7 10 0.60957309 3865 4637 1043 452 1.855325 1 0 0 0 2007 7 11 0.68938434 3947 3865 4637 1043 2.08271 1 0 0 0 2007 7 12 0.55069124 3472 3947 3865 4637 2.047206 1 0 0 0 2007 7 13 0.52207637 1676 3472 3947 3865 2.070843 1 0 0 0 2007 7 14 0.74563107 515 1676 3472 3947 1.857123 1 0 0 0 2007 7 15 0.79627518 2631 515 1676 3472 1.796188 1 0 0 0 2007 7 16 0.46200856 9818 2631 515 1676 2.02748 1 0 0 0 2007 7 17 0.47573341 7397 9818 2631 515 2.011776 1 0 0 0 2007 7 18 0.47101449 3312 7397 9818 2631 1.696546 1 0 0 0 2007 7 19 0.40099683 15449 3312 7397 9818 1.845375 1 0 0 0 2007 7 20 0.63276836 177 15449 3312 7397 1.869893 1 0 0 0 2007 7 21 0.6835443 237 177 15449 3312 2.04099 1 0 0 0 2007 7 22 0.6102071 1352 237 177 15449 2.122644 1 0 0 0 2007 7 23 0.42616959 8208 1352 237 177 1.903156 1 0 0 0 2007 7 24 0.47633136 676 8208 1352 237 2.149239 1 0 0 0 2007 7 25 0.74820831 2093 676 8208 1352 2.258171 1 0 0 0 2007 7 26 0.48623446 2252 2093 676 8208 2.236827 1 0 0 0 2007 7 27 0.65359477 153 2252 2093 676 1.906725 1 0 0 0 2007 7 28 0 0 153 2252 2093 2.013376 1 0 0 0 2007 7 29 0.70173536 3688 0 153 2252 1.941826 1 0 0 0 2007 7 30 0.54612811 8368 3688 0 153 2.025288 1 0 0 0 2007 7 31 0.41939577 25222 8368 3688 0 2.001138 1 0 0 0 2007 8 0 0 0 25222 8368 3688 1.911645 1 0 0 0 2007 8 1 0.66423358 274 0 25222 8368 1.902087 1 0 0 0 2007 8 2 0.43365456 5396 274 0 25222 1.993381 1 0 0 0 2007 8 3 0 0 5396 274 0 2.094411 1 0 0 0 2007 8 4 0 0 0 5396 274 2.348145 1 0 0 0 2007 8 5 0.31265073 4561 0 0 5396 2.053757 1 0 0 0 2007 8 6 0.90748299 735 4561 0 0 2.190196 1 0 0 0 2007 8 7 0.53761643 4187 735 4561 0 2.354725 1 0 0 0 2007 8 8 0.75635448 4367 4187 735 4561 2.384203 1 0 0 0 2007 8 9 0.7274276 587 4367 4187 735 2.494142 1 0 0 0 2007 8 10 0.40564081 2411 587 4367 4187 2.465593 1 0 0 0 2007 8 11 0.6350365 137 2411 587 4367 2.599933 1 0 0 0 2007 8 12 0.40571891 3707 137 2411 587 2.861102 1 0 0 0 2007 8 13 0.5755627 3110 3707 137 2411 2.753244 1 0 0 0 2007 8 14 0.84731544 596 3110 3707 137 2.930099 1 0 0 0 2007 8 15 0.54912517 6687 596 3110 3707 2.747029 1 0 0 0 2007 8 16 0.4030722 22785 6687 596 3110 2.700335 1 0 0 0 2007 8 17 0.48353096 759 22785 6687 596 2.639266 1 0 0 0 2007 8 18 0 0 759 22785 6687 2.791319 1 0 0 0 2007 8 19 0.80879121 1820 0 759 22785 2.820064 1 0 0 0 2007 8 20 0.76576302 3283 1820 0 759 2.657459 1 0 0 0 2007 8 21 0.6746089 4155 3283 1820 0 2.652952 1 0 0 0 2007 8 22 0.81980198 1010 4155 3283 1820 2.714331 1 0 0 0 2007 8 23 0.72737368 1801 1010 4155 3283 2.857506 1 0 0 0 2007 8 24 0.68695652 230 1801 1010 4155 2.203174 1 0 0 0 2007 8 25 0 0 230 1801 1010 1.82345 1 0 0 0 2007 8 26 0.77144842 2879 0 230 1801 1.917954 1 0 0 0 2007 8 27 0.54299363 1884 2879 0 230 1.830766 1 0 0 0 2007 8 28 0.49961939 7882 1884 2879 0 1.846766 1 0 0 0 2007 8 29 0.43601032 6587 7882 1884 2879 1.663893 1 0 0 0 2007 8 30 0.48127854 7665 6587 7882 1884 1.620875 1 0 0 0 2007 8 31 0 0 7665 6587 7882 1.609351 1 0 0 0 2007 9 0 0 0 0 7665 6587 1.533654 1 0 0 0 2007 9 1 0.8021978 182 0 0 7665 1.622161 1 0 0 0 2007 9 2 0.80882353 476 182 0 0 1.980852 1 0 0 0 2007 9 3 0.83928571 112 476 182 0 2.071598 1 0 0 0 2007 9 4 0.60744939 6175 112 476 182 1.738745 1 0 0 0 2007 9 5 0.50170931 9653 6175 112 476 1.602301 1 0 0 0 2007 9 6 0.27775704 13393 9653 6175 112 1.699215 1 0 0 0 2007 9 7 0.415398 2299 13393 9653 6175 1.756407 1 0 0 0 2007 9 8 0 0 2299 13393 9653 1.580084 1 0 0 0 2007 9 9 0.74210374 3451 0 2299 13393 2.080518 1 0 0 0 2007 9 10 0.6107164 3378 3451 0 2299 1.517171 1 0 0 0 2007 9 11 0.54259465 7184 3378 3451 0 1.828473 1 0 0 0 2007 9 12 0.51974386 5622 7184 3378 3451 1.854924 1 0 0 0 2007 9 13 0.61564771 2339 5622 7184 3378 2.080119 1 0 0 0 2007 9 14 0.856102 549 2339 5622 7184 1.763558 1 0 0 0 2007 9 15 0 0 549 2339 5622 2.064486 1 0 0 0 2007 9 16 0.67334168 5593 0 549 2339 1.896932 1 0 0 0 2007 9 17 0.5814405 5151 5593 0 549 2.034994 1 0 0 0 2007 9 18 0.62966805 4820 5151 5593 0 1.789395 1 0 0 0 2007 9 19 0.53712667 8471 4820 5151 5593 1.70139 1 0 0 0 2007 9 20 0.57776094 5279 8471 4820 5151 1.632773 1 0 0 0 2007 9 21 0.47037422 1924 5279 8471 4820 1.614435 1 0 0 0 2007 9 22 0 0 1924 5279 8471 1.37702 1 0 0 0 2007 9 23 0.6329246 3833 0 1924 5279 1.697888 1 0 0 0 2007 9 24 0.57758621 10324 3833 0 1924 1.631625 1 0 0 0 2007 9 25 0.45491275 11633 10324 3833 0 1.573076 1 0 0 0 2007 9 26 0.31858475 25946 11633 10324 3833 1.467414 1 0 0 0 2007 9 27 0.35524051 12203 25946 11633 10324 1.635661 1 0 0 0 2007 9 28 0.40949555 2022 12203 25946 11633 1.461779 1 0 0 0 2007 9 29 0.40296595 5462 2022 12203 25946 2.235294 1 0 0 0 2007 9 30 0.42174899 7936 5462 2022 12203 1.964298 1 0 0 0 2007 10 0 0 0 7936 5462 2022 1.238596 1 0 0 0 2007 10 1 0.57257683 2115 0 7936 5462 1.722584 1 0 0 0 2007 10 2 0.49960651 7624 2115 0 7936 2.026055 1 0 0 0 2007 10 3 0.55796767 4330 7624 2115 0 2.205237 1 0 0 0 2007 10 4 0.44403013 13141 4330 7624 2115 2.126372 1 0 0 0 2007 10 5 0.52191781 730 13141 4330 7624 1.932885 1 0 0 0 2007 10 6 0 0 730 13141 4330 2.401402 1 0 0 0 2007 10 7 0.49276305 6149 0 730 13141 1.929529 1 0 0 0 2007 10 8 0.50103681 3858 6149 0 730 2.322675 1 0 0 0 2007 10 9 0.42794416 17764 3858 6149 0 2.364402 1 0 0 0 2007 10 10 0.42075212 15237 17764 3858 6149 2.360075 1 0 0 0 2007 10 11 0.46593757 4389 15237 17764 3858 2.60971 1 0 0 0 2007 10 12 0 0 4389 15237 17764 2.920247 1 0 0 0 2007 10 13 0 0 0 4389 15237 2.900645 1 0 0 0 2007 10 14 0.60526316 6308 0 0 4389 2.618965 1 0 0 0 2007 10 15 0.5356446 6663 6308 0 0 2.815996 1 0 0 0 2007 10 16 0.68973472 1734 6663 6308 0 2.930163 1 0 0 0 2007 10 17 0.52964824 8955 1734 6663 6308 3.360778 1 0 0 0 2007 10 18 0.57237386 6045 8955 1734 6663 3.024652 1 0 0 0 2007 10 19 0.43897216 2802 6045 8955 1734 2.575895 1 0 0 0 2007 10 20 0 0 2802 6045 8955 3.018097 1 0 0 0 2007 10 21 0.6250556 6745 0 2802 6045 2.478754 1 0 0 0 2007 10 22 0.6045273 7510 6745 0 2802 2.26192 1 0 0 0 2007 10 23 0.63721036 4402 7510 6745 0 1.843272 1 0 0 0 2007 10 24 0.52608569 13724 4402 7510 6745 1.946993 1 0 0 0 2007 10 25 0.5660462 8918 13724 4402 7510 2.09435 1 0 0 0 2007 10 26 0.63018868 2385 8918 13724 4402 2.23928 1 0 0 0 2007 10 27 0 0 2385 8918 13724 2.464398 1 0 0 0 2007 10 28 0.55970898 7972 0 2385 8918 2.614707 1 0 0 0 2007 10 29 0.75214133 1868 7972 0 2385 2.452179 1 0 0 0 2007 10 30 0.65450644 2796 1868 7972 0 2.334668 1 0 0 0 2007 10 31 0.69111462 3804 2796 1868 7972 2.046664 1 0 0 0 2007 11 0 0 0 3804 2796 1868 1.393401 1 0 0 0 2007 11 1 0.63763225 15407 0 3804 2796 2.091776 1 0 0 0 2007 11 2 0 0 15407 0 3804 2.08376 1 0 0 0 2007 11 3 0 0 0 15407 0 2.706564 1 0 0 0 2007 11 4 0.85037731 3843 0 0 15407 2.626326 1 0 0 0 2007 11 5 0.8195317 3502 3843 0 0 2.426978 1 0 0 0 2007 11 6 0 0 3502 3843 0 2.190286 1 0 0 0 2007 11 7 1.2867133 429 0 3502 3843 2.56733 1 0 0 0 2007 11 8 0.67641287 10192 429 0 3502 2.369728 1 0 0 0 2007 11 9 0.87439222 1234 10192 429 0 2.021536 1 0 0 0 2007 11 10 0.98 300 1234 10192 429 2.371 1 0 0 0 2007 11 11 0.99038882 4578 300 1234 10192 2.444381 1 0 0 0 2007 11 12 1.0093573 4061 4578 300 1234 1.990177 1 0 0 0 2007 11 13 0.7030525 4095 4061 4578 300 1.857911 1 0 0 0 2007 11 14 0.72318678 7142 4095 4061 4578 1.81483 1 0 0 0 2007 11 15 0.61187653 11468 7142 4095 4061 2.004401 1 0 0 0 2007 11 16 0.60043431 921 11468 7142 4095 1.801584 1 0 0 0 2007 11 17 0 0 921 11468 7142 1.356802 1 0 0 0 2007 11 18 1.0004032 2480 0 921 11468 1.897816 1 0 0 0 2007 11 19 0.910492 1687 2480 0 921 1.478424 1 0 0 0 2007 11 20 0.80656238 2621 1687 2480 0 1.993454 1 0 0 0 2007 11 21 0.65499316 3655 2621 1687 2480 2.129853 1 0 0 0 2007 11 22 0 0 3655 2621 1687 2.568548 1 0 0 0 2007 11 23 0.80099502 201 0 3655 2621 2.269048 1 0 0 0 2007 11 24 0 0 201 0 3655 2.142696 1 0 0 0 2007 11 25 0.7633434 1986 0 201 0 2.314328 1 0 0 0 2007 11 26 0.51952782 22788 1986 0 201 2.233255 1 0 0 0 2007 11 27 0.57465468 4561 22788 1986 0 2.235227 1 0 0 0 2007 11 28 0.66968326 1768 4561 22788 1986 2.143552 1 0 0 0 2007 11 29 0.63243653 5278 1768 4561 22788 2.378254 1 0 0 0 2007 11 30 0.8041958 143 5278 1768 4561 2.165829 1 0 0 0 2007 12 0 0 0 143 5278 1768 2 1 0 0 0 2007 12 1 0 0 0 143 5278 2.440882 1 0 0 0 2007 12 2 0.90651907 813 0 0 143 2.787796 1 0 0 0 2007 12 3 0.67572229 2319 813 0 0 2.89298 1 0 0 0 2007 12 4 0.74689071 7558 2319 813 0 3.088069 1 0 0 0 2007 12 5 0 0 7558 2319 813 2.819509 1 0 0 0 2007 12 6 0.98876404 178 0 7558 2319 2.817319 1 0 0 0 2007 12 7 1.079602 804 178 0 7558 2.763419 1 0 0 0 2007 12 8 0 0 804 178 0 2.465654 1 0 0 0 2007 12 9 1.1092706 3743 0 804 178 2.537371 1 0 0 0 2007 12 10 0.97990431 2090 3743 0 804 2.538923 1 0 0 0 2007 12 11 0.8519049 3491 2090 3743 0 2.289992 1 0 0 0 2007 12 12 0.68076877 8273 3491 2090 3743 2.209963 1 0 0 0 2007 12 13 0.65956454 6430 8273 3491 2090 1.864303 1 0 0 0 2007 12 14 1.242921 671 6430 8273 3491 2.136051 1 0 0 0 2007 12 15 0.61898774 3181 671 6430 8273 2.418971 1 0 0 0 2007 12 16 0.49055542 15988 3181 671 6430 2.984535 1 0 0 0 2007 12 17 0.42108576 6355 15988 3181 671 2.62002 1 0 0 0 2007 12 18 0.56534802 1997 6355 15988 3181 2.200334 1 0 0 0 2007 12 19 0.58726359 4177 1997 6355 15988 2.528227 1 0 0 0 2007 12 20 0.58794788 614 4177 1997 6355 2.023687 1 0 0 0 2007 12 21 0.45404046 9541 614 4177 1997 2.018594 1 0 0 0 2007 12 22 0.69196025 1107 9541 614 4177 2.459022 1 0 0 0 2007 12 23 0.75322442 1163 1107 9541 614 3.210132 1 0 0 0 2007 12 24 0.724309 1411 1163 1107 9541 2.489057 1 0 0 0 2007 12 25 0 0 1411 1163 1107 2.2 1 0 0 0 2007 12 26 0.75254611 3633 0 1411 1163 2.705808 1 0 0 0 2007 12 27 0.84139265 1034 3633 0 1411 2.27196 1 0 0 0 2007 12 28 0 0 1034 3633 0 2.656669 1 0 0 0 2007 12 29 0 0 0 1034 3633 2.192124 1 0 0 0 2007 12 30 0.67827869 1464 0 0 1034 2.39859 1 0 0 0 2007 12 31 0.69796337 7807 1464 0 0 2.065795 1 0 0 0 2008 1 0 0 0 7807 1464 0 1 0 1 0 0 2008 1 1 0 0 0 7807 1464 2.904309 0 1 0 0 2008 1 2 0.52573499 12143 0 0 7807 2.50103 0 1 0 0 2008 1 3 0.60680448 4644 12143 0 0 2.107387 0 1 0 0 2008 1 4 0 0 4644 12143 0 1.954806 0 1 0 0 2008 1 5 0 0 0 4644 12143 1.683708 0 1 0 0 2008 1 6 0.97758703 2097 0 0 4644 1.852717 0 1 0 0 2008 1 7 0.93044496 4270 2097 0 0 1.662325 0 1 0 0 2008 1 8 0.93518519 1944 4270 2097 0 1.540527 0 1 0 0 2008 1 9 0.70813727 17426 1944 4270 2097 1.717945 0 1 0 0 2008 1 10 0.71891892 1665 17426 1944 4270 1.290976 0 1 0 0 2008 1 11 0.61314825 3818 1665 17426 1944 1.244135 0 1 0 0 2008 1 12 0 0 3818 1665 17426 1.465383 0 1 0 0 2008 1 13 0.65825178 8992 0 3818 1665 1.642206 0 1 0 0 2008 1 14 0.6680436 8715 8992 0 3818 1.610712 0 1 0 0 2008 1 15 0.30701296 49708 8715 8992 0 1.778323 0 1 0 0 2008 1 16 0.88 250 49708 8715 8992 1.730976 0 1 0 0 2008 1 17 0.73400366 2188 250 49708 8715 1.54536 0 1 0 0 2008 1 18 0.46079882 1352 2188 250 49708 1.506914 0 1 0 0 2008 1 19 0.7140255 1647 1352 2188 250 1.460641 0 1 0 0 2008 1 20 0.55002296 15243 1647 1352 2188 1.650038 0 1 0 0 2008 1 21 0.49741461 27462 15243 1647 1352 1.791332 0 1 0 0 2008 1 22 0.63649025 1436 27462 15243 1647 1.882084 0 1 0 0 2008 1 23 0.75098935 3285 1436 27462 15243 2.080246 0 1 0 0 2008 1 24 0.85430586 5388 3285 1436 27462 2.241212 0 1 0 0 2008 1 25 0.50942484 14483 5388 3285 1436 2.245745 0 1 0 0 2008 1 26 0.51953852 2687 14483 5388 3285 2.30841 0 1 0 0 2008 1 27 0.82415601 6102 2687 14483 5388 2.016708 0 1 0 0 2008 1 28 0.5102717 34707 6102 2687 14483 2.386421 0 1 0 0 2008 1 29 0.31750715 22031 34707 6102 2687 2.40415 0 1 0 0 2008 1 30 0.71922828 9278 22031 34707 6102 2.217865 0 1 0 0 2008 1 31 1.0559441 286 9278 22031 34707 2.488975 0 1 0 0 2008 2 0 0 0 286 9278 22031 3.6 0 1 0 0 2008 2 1 0 0 0 286 9278 2.709641 0 1 0 0 2008 2 2 0.80302115 1655 0 0 286 2.900497 0 1 0 0 2008 2 3 0.76717917 6374 1655 0 0 2.869839 0 1 0 0 2008 2 4 0.70861175 18707 6374 1655 0 2.835919 0 1 0 0 2008 2 5 0.67905022 8802 18707 6374 1655 2.778721 0 1 0 0 2008 2 6 0.63923241 11725 8802 18707 6374 2.583312 0 1 0 0 2008 2 7 0.4804493 9437 11725 8802 18707 2.522916 0 1 0 0 2008 2 8 0.4017094 351 9437 11725 8802 1.89523 0 1 0 0 2008 2 9 0.62096156 9131 351 9437 11725 1.89763 0 1 0 0 2008 2 10 0.64484946 9931 9131 351 9437 2.164617 0 1 0 0 2008 2 11 0.47575453 9940 9931 9131 351 2.148486 0 1 0 0 2008 2 12 0 0 9940 9931 9131 1.965305 0 1 0 0 2008 2 13 0.61045965 15947 0 9940 9931 2.081616 0 1 0 0 2008 2 14 0.61590214 3270 15947 0 9940 2.197291 0 1 0 0 2008 2 15 0 0 3270 15947 0 2.107096 0 1 0 0 2008 2 16 0 0 0 3270 15947 2.064127 0 1 0 0 2008 2 17 0 0 0 0 3270 1.993352 0 1 0 0 2008 2 18 0.6573211 27250 0 0 0 2.225267 0 1 0 0 2008 2 19 0.70217471 2713 27250 0 0 2.023454 0 1 0 0 2008 2 20 0.61666263 12399 2713 27250 0 2.235246 0 1 0 0 2008 2 21 0.57880435 4048 12399 2713 27250 2.290948 0 1 0 0 2008 2 22 0.52945253 1443 4048 12399 2713 2.232898 0 1 0 0 2008 2 23 0.47340922 10718 1443 4048 12399 2.260405 0 1 0 0 2008 2 24 0.59689753 4577 10718 1443 4048 2.316638 0 1 0 0 2008 2 25 0.59072234 13387 4577 10718 1443 2.222454 0 1 0 0 2008 2 26 0.54002675 20936 13387 4577 10718 2.213847 0 1 0 0 2008 2 27 0.44270074 32243 20936 13387 4577 2.267236 0 1 0 0 2008 2 28 0.61382296 6699 32243 20936 13387 2.19717 0 1 0 0 2008 2 29 0.60008254 2423 6699 32243 20936 2.092009 0 1 0 0 2008 3 1 0.6097561 205 2423 6699 32243 2.234075 0 1 0 0 2008 3 2 0.68000852 9394 205 2423 6699 2.187804 0 1 0 0 2008 3 3 0.65909501 5547 9394 205 2423 2.218163 0 1 0 0 2008 3 4 0.65188953 6880 5547 9394 205 2.269311 0 1 0 0 2008 3 5 0.71544085 9028 6880 5547 9394 2.65007 0 1 0 0 2008 3 6 0.60301508 398 9028 6880 5547 2.437055 0 1 0 0 2008 3 7 0.92586143 8039 398 9028 6880 2.36268 0 1 0 0 2008 3 8 1.2241899 2654 8039 398 9028 2.325522 0 1 0 0 2008 3 9 0.90349732 6891 2654 8039 398 2.336854 0 1 0 0 2008 3 10 0.83100414 15456 6891 2654 8039 2.302375 0 1 0 0 2008 3 11 0.85341615 2415 15456 6891 2654 2.271088 0 1 0 0 2008 3 12 0.77234247 22916 2415 15456 6891 2.319271 0 1 0 0 2008 3 13 0.79749057 11397 22916 2415 15456 2.434265 0 1 0 0 2008 3 14 0.68567312 4858 11397 22916 2415 2.30102 0 1 0 0 2008 3 15 0.91094494 5884 4858 11397 22916 2.186691 0 1 0 0 2008 3 16 0.71881787 12892 5884 4858 11397 2.271971 0 1 0 0 2008 3 17 0.68812969 13201 12892 5884 4858 2.398304 0 1 0 0 2008 3 18 0.59417476 1545 13201 12892 5884 2.292923 0 1 0 0 2008 3 19 0.48934856 40464 1545 13201 12892 2.325249 0 1 0 0 2008 3 20 0.48543907 4464 40464 1545 13201 2.602685 0 1 0 0 2008 3 21 0 0 4464 40464 1545 2.42507 0 1 0 0 2008 3 22 0.97869177 2018 0 4464 40464 2.489169 0 1 0 0 2008 3 23 0 0 2018 0 4464 2.383242 0 1 0 0 2008 3 24 0.8988482 6772 0 2018 0 2.47241 0 1 0 0 2008 3 25 0.79850426 11633 6772 0 2018 2.448463 0 1 0 0 2008 3 26 0.86895256 2129 11633 6772 0 2.642688 0 1 0 0 2008 3 27 0.56347516 23009 2129 11633 6772 2.132263 0 1 0 0 2008 3 28 0.71861304 7758 23009 2129 11633 2.115985 0 1 0 0 2008 3 29 0.41156202 2612 7758 23009 2129 1.866279 0 1 0 0 2008 3 30 0.49550479 31033 2612 7758 23009 2.095431 0 1 0 0 2008 3 31 0.46672497 22870 31033 2612 7758 2.109168 0 1 0 0 2008 4 1 0.44134136 38886 22870 31033 2612 2.128794 0 1 0 0 2008 4 2 0.51683778 9740 38886 22870 31033 2.116309 0 1 0 0 2008 4 3 0.41710611 7775 9740 38886 22870 2.101127 0 1 0 0 2008 4 4 0 0 7775 9740 38886 2.564975 0 1 0 0 2008 4 5 0 0 0 7775 9740 2.79323 0 1 0 0 2008 4 6 0.96199324 1184 0 0 7775 2.937948 0 1 0 0 2008 4 7 0.92312009 1782 1184 0 0 3.278841 0 1 0 0 2008 4 8 0.79219434 12555 1782 1184 0 3.170679 0 1 0 0 2008 4 9 0.77400362 2208 12555 1782 1184 3.511714 0 1 0 0 2008 4 10 0.78415132 16178 2208 12555 1782 3.127236 0 1 0 0 2008 4 11 0.74198932 2996 16178 2208 12555 3.067631 0 1 0 0 2008 4 12 0.79992485 5323 2996 16178 2208 2.996573 0 1 0 0 2008 4 13 0.55310517 32784 5323 2996 16178 2.507 0 1 0 0 2008 4 14 0.58034026 6348 32784 5323 2996 2.400797 0 1 0 0 2008 4 15 0.4239731 4163 6348 32784 5323 2.255863 0 1 0 0 2008 4 16 0.66203624 11590 4163 6348 32784 2.279683 0 1 0 0 2008 4 17 0.72184457 5443 11590 4163 6348 2.31587 0 1 0 0 2008 4 18 0.62837838 592 5443 11590 4163 2.390023 0 1 0 0 2008 4 19 0 0 592 5443 11590 2.315187 0 1 0 0 2008 4 20 0.82773109 2618 0 592 5443 2.11457 0 1 0 0 2008 4 21 0.8001628 2457 2618 0 592 2.271436 0 1 0 0 2008 4 22 0.81750925 811 2457 2618 0 2.448472 0 1 0 0 2008 4 23 0.58473205 9890 811 2457 2618 2.42498 0 1 0 0 2008 4 24 0.57710724 13157 9890 811 2457 2.124136 0 1 0 0 2008 4 25 0 0 13157 9890 811 2.069367 0 1 0 0 2008 4 26 0.70234114 299 0 13157 9890 2.010408 0 1 0 0 2008 4 27 0.55898123 29094 299 0 13157 2.01913 0 1 0 0 2008 4 28 0.66005553 2521 29094 299 0 1.888264 0 1 0 0 2008 4 29 0.82458951 2497 2521 29094 299 1.985879 0 1 0 0 2008 4 30 0.49067396 4021 2497 2521 29094 2.178536 0 1 0 0 2008 5 0 0 0 4021 2497 2521 4.752137 0 1 0 0 2008 5 1 0.76986301 3650 0 4021 2497 2.101405 0 1 0 0 2008 5 2 0.77672956 318 3650 0 4021 1.929706 0 1 0 0 2008 5 3 0.75767918 293 318 3650 0 2.080135 0 1 0 0 2008 5 4 0.48439866 40285 293 318 3650 2.40687 0 1 0 0 2008 5 5 0.71149228 2915 40285 293 318 1.809678 0 1 0 0 2008 5 6 0 0 2915 40285 293 1.996975 0 1 0 0 2008 5 7 0.7367387 4581 0 2915 40285 1.891021 0 1 0 0 2008 5 8 0.69852613 2239 4581 0 2915 2.114113 0 1 0 0 2008 5 9 0.94162437 394 2239 4581 0 1.857571 0 1 0 0 2008 5 10 0.34017595 341 394 2239 4581 1.669137 0 1 0 0 2008 5 11 0.53074083 34124 341 394 2239 2.033552 0 1 0 0 2008 5 12 0.52776149 10230 34124 341 394 1.843354 0 1 0 0 2008 5 13 0 0 10230 34124 341 2.825713 0 1 0 0 2008 5 14 0 0 0 10230 34124 2.689029 0 1 0 0 2008 5 15 0 0 0 0 10230 2.333536 0 1 0 0 2008 5 16 0 0 0 0 0 2.311687 0 1 0 0 2008 5 17 1 330 0 0 0 2.34195 0 1 0 0 2008 5 18 0 0 330 0 0 2.407727 0 1 0 0 2008 5 19 1.1653543 381 0 330 0 2.095097 0 1 0 0 2008 5 20 0.7 7470 381 0 330 2.216324 0 1 0 0 2008 5 21 0.80230643 607 7470 381 0 2.070725 0 1 0 0 2008 5 22 0.50562116 28286 607 7470 381 1.825774 0 1 0 0 2008 5 23 0 0 28286 607 7470 1.635249 0 1 0 0 2008 5 24 0 0 0 28286 607 1.657837 0 1 0 0 2008 5 25 0.39965473 6951 0 0 28286 1.510941 0 1 0 0 2008 5 26 0 0 6951 0 0 1.281419 0 1 0 0 2008 5 27 0.57435757 3658 0 6951 0 1.391584 0 1 0 0 2008 5 28 0 0 3658 0 6951 1.35915 0 1 0 0 2008 5 29 0.54201927 37578 0 3658 0 1.420061 0 1 0 0 2008 5 30 0 0 37578 0 3658 1.387167 0 1 0 0 2008 5 31 0 0 0 37578 0 1.901631 0 1 0 0 2008 6 1 0.42642226 35296 0 0 37578 2.016181 0 1 0 0 2008 6 2 0.52029312 10644 35296 0 0 1.67153 0 1 0 0 2008 6 3 0.78651026 1705 10644 35296 0 1.453747 0 1 0 0 2008 6 4 0.66815416 2465 1705 10644 35296 1.399025 0 1 0 0 2008 6 5 0.93290548 1714 2465 1705 10644 1.046263 0 1 0 0 2008 6 6 1.0590842 677 1714 2465 1705 0.743787 0 1 0 0 2008 6 7 0 0 677 1714 2465 1.069058 0 1 0 0 2008 6 8 0.80866228 3648 0 677 1714 1.111305 0 1 0 0 2008 6 9 0.49190283 988 3648 0 677 0.89283 0 1 0 0 2008 6 10 0.91176471 612 988 3648 0 0.901098 0 1 0 0 2008 6 11 0.69720497 1932 612 988 3648 0.924779 0 1 0 0 2008 6 12 0.50267123 18718 1932 612 988 1.088231 0 1 0 0 2008 6 13 0.50069252 1444 18718 1932 612 1.360779 0 1 0 0 2008 6 14 0.5 1502 1444 18718 1932 1.356659 0 1 0 0 2008 6 15 0.46830986 5112 1502 1444 18718 1.294996 0 1 0 0 2008 6 16 0.71729958 237 5112 1502 1444 1.282856 0 1 0 0 2008 6 17 0.66829413 6759 237 5112 1502 1.700571 0 1 0 0 2008 6 18 0.76744186 2752 6759 237 5112 1.899026 0 1 0 0 2008 6 19 0.58042677 11669 2752 6759 237 2.110816 0 1 0 0 2008 6 20 0.57735734 3277 11669 2752 6759 1.957465 0 1 0 0 2008 6 21 0 0 3277 11669 2752 1.46587 0 1 0 0 2008 6 22 0.59081042 8292 0 3277 11669 1.610898 0 1 0 0 2008 6 23 0.58255564 5572 8292 0 3277 1.514857 0 1 0 0 2008 6 24 0.67799642 2236 5572 8292 0 1.726162 0 1 0 0 2008 6 25 0.78075065 2691 2236 5572 8292 1.810862 0 1 0 0 2008 6 26 0.81125828 906 2691 2236 5572 1.705988 0 1 0 0 2008 6 27 0.89534884 258 906 2691 2236 1.69013 0 1 0 0 2008 6 28 0 0 258 906 2691 1.576626 0 1 0 0 2008 6 29 0.49518011 7884 0 258 906 1.895505 0 1 0 0 2008 6 30 0.46900935 13375 7884 0 258 1.737377 0 1 0 0 2008 7 0 0 0 13375 7884 0 4.889655 0 1 0 0 2008 7 1 0.5082544 3695 0 13375 7884 1.885165 0 1 0 0 2008 7 2 0.72606383 1880 3695 0 13375 1.759509 0 1 0 0 2008 7 3 0.5740458 655 1880 3695 0 1.690957 0 1 0 0 2008 7 4 0 0 655 1880 3695 1.483384 0 1 0 0 2008 7 5 0.53623188 414 0 655 1880 1.719481 0 1 0 0 2008 7 6 0.55735537 3025 414 0 655 1.521894 0 1 0 0 2008 7 7 0.39234783 2875 3025 414 0 1.904352 0 1 0 0 2008 7 8 0.45600824 37848 2875 3025 414 1.994756 0 1 0 0 2008 7 9 0.53558926 857 37848 2875 3025 2.10055 0 1 0 0 2008 7 10 0.07028568 15437 857 37848 2875 2.125039 0 1 0 0 2008 7 11 0 0 15437 857 37848 1.676976 0 1 0 0 2008 7 12 0 0 0 15437 857 2.033909 0 1 0 0 2008 7 13 0.45162387 5881 0 0 15437 2.091314 0 1 0 0 2008 7 14 0.38451756 18873 5881 0 0 2.260262 0 1 0 0 2008 7 15 0.53889561 4962 18873 5881 0 2.26617 0 1 0 0 2008 7 16 0.67870183 2465 4962 18873 5881 2.341203 0 1 0 0 2008 7 17 0.57149555 2133 2465 4962 18873 1.811528 0 1 0 0 2008 7 18 0 0 2133 2465 4962 1.823094 0 1 0 0 2008 7 19 0 0 0 2133 2465 2.015055 0 1 0 0 2008 7 20 0.75278121 2427 0 0 2133 2.032915 0 1 0 0 2008 7 21 0.65620094 4459 2427 0 0 1.686504 0 1 0 0 2008 7 22 0.53995608 15029 4459 2427 0 1.529002 0 1 0 0 2008 7 23 0.44546348 3053 15029 4459 2427 1.3697 0 1 0 0 2008 7 24 0.5623173 3763 3053 15029 4459 1.076442 0 1 0 0 2008 7 25 0.51136364 528 3763 3053 15029 1.324494 0 1 0 0 2008 7 26 0 0 528 3763 3053 1.828909 0 1 0 0 2008 7 27 0.67959002 2244 0 528 3763 1.576853 0 1 0 0 2008 7 28 0.57530231 2729 2244 0 528 1.866688 0 1 0 0 2008 7 29 0.49914749 9384 2729 2244 0 1.63447 0 1 0 0 2008 7 30 0.44149398 8059 9384 2729 2244 1.956726 0 1 0 0 2008 7 31 0.4278244 23713 8059 9384 2729 1.6766 0 1 0 0 2008 8 0 0 0 23713 8059 9384 4.361963 0 1 0 0 2008 8 1 0.7345679 648 0 23713 8059 1.451911 0 1 0 0 2008 8 2 1.0310559 322 648 0 23713 1.950289 0 1 0 0 2008 8 3 0.53476064 4157 322 648 0 1.799564 0 1 0 0 2008 8 4 0.76847729 1123 4157 322 648 1.724968 0 1 0 0 2008 8 5 0.50269318 9654 1123 4157 322 1.735767 0 1 0 0 2008 8 6 0.50619835 5324 9654 1123 4157 1.787289 0 1 0 0 2008 8 7 0.56049496 6546 5324 9654 1123 1.592729 0 1 0 0 2008 8 8 0.69424084 955 6546 5324 9654 1.597117 0 1 0 0 2008 8 9 0.67759563 183 955 6546 5324 1.90556 0 1 0 0 2008 8 10 0.56807626 4091 183 955 6546 1.329773 0 1 0 0 2008 8 11 0.67181565 3949 4091 183 955 1.653195 0 1 0 0 2008 8 12 0.552393 4743 3949 4091 183 1.738832 0 1 0 0 2008 8 13 0.41464815 27867 4743 3949 4091 1.647488 0 1 0 0 2008 8 14 0.48029729 7131 27867 4743 3949 1.699844 0 1 0 0 2008 8 15 0.64905363 1268 7131 27867 4743 1.619913 0 1 0 0 2008 8 16 0.61438202 2225 1268 7131 27867 2.087776 0 1 0 0 2008 8 17 0.51084723 4379 2225 1268 7131 2.021692 0 1 0 0 2008 8 18 0.61046318 3001 4379 2225 1268 2.331557 0 1 0 0 2008 8 19 0.46013323 9908 3001 4379 2225 1.902741 0 1 0 0 2008 8 20 0.4661795 5337 9908 3001 4379 1.998571 0 1 0 0 2008 8 21 0.56318538 3830 5337 9908 3001 1.808325 0 1 0 0 2008 8 22 0.73584906 1378 3830 5337 9908 1.789858 0 1 0 0 2008 8 23 1.1236443 461 1378 3830 5337 2.26449 0 1 0 0 2008 8 24 0.70283019 3392 461 1378 3830 10.81963 0 1 0 0 2008 8 25 0.58913481 2485 3392 461 1378 1.816793 0 1 0 0 2008 8 26 0.47401378 6388 2485 3392 461 1.820724 0 1 0 0 2008 8 27 0.45111525 20937 6388 2485 3392 1.755406 0 1 0 0 2008 8 28 0.58548753 4410 20937 6388 2485 2.013944 0 1 0 0 2008 8 29 0.53856749 2178 4410 20937 6388 1.9592 0 1 0 0 2008 8 30 0.75 220 2178 4410 20937 2.210701 0 1 0 0 2008 8 31 0.49647059 1275 220 2178 4410 1.823044 0 1 0 0 2008 9 0 0 0 1275 220 2178 4.882353 0 1 0 0 2008 9 1 0.69626697 10608 0 1275 220 2.457444 0 1 0 0 2008 9 2 0.41197296 9321 10608 0 1275 2.096875 0 1 0 0 2008 9 3 0.64336775 3777 9321 10608 0 2.087538 0 1 0 0 2008 9 4 0.78224199 6521 3777 9321 10608 2.289077 0 1 0 0 2008 9 5 0.89690722 1358 6521 3777 9321 2.237724 0 1 0 0 2008 9 6 0.87067545 3642 1358 6521 3777 2.66747 0 1 0 0 2008 9 7 0.48041349 1838 3642 1358 6521 2.783251 0 1 0 0 2008 9 8 0.88368056 576 1838 3642 1358 2.828531 0 1 0 0 2008 9 9 0.70501636 1834 576 1838 3642 2.521487 0 1 0 0 2008 9 10 0.47049901 32304 1834 576 1838 2.401764 0 1 0 0 2008 9 11 0.48706027 11399 32304 1834 576 2.47207 0 1 0 0 2008 9 12 0.81586402 706 11399 32304 1834 2.221746 0 1 0 0 2008 9 13 0 0 706 11399 32304 1.885736 0 1 0 0 2008 9 14 0.47178077 9816 0 706 11399 2.169419 0 1 0 0 2008 9 15 0.54967494 21842 9816 0 706 1.698295 0 1 0 0 2008 9 16 0.54257552 10957 21842 9816 0 1.678233 0 1 0 0 2008 9 17 0.60627423 2837 10957 21842 9816 1.427771 0 1 0 0 2008 9 18 0.47976955 22565 2837 10957 21842 1.090173 0 1 0 0 2008 9 19 0.63565891 1032 22565 2837 10957 1.210934 0 1 0 0 2008 9 20 0.53438557 1774 1032 22565 2837 1.454101 0 1 0 0 2008 9 21 0.54633648 3917 1774 1032 22565 2.338686 0 1 0 0 2008 9 22 0.5494664 3467 3917 1774 1032 1.152376 0 1 0 0 2008 9 23 0.51543366 9881 3467 3917 1774 1.149017 0 1 0 0 2008 9 24 0.46730028 15046 9881 3467 3917 1.676713 0 1 0 0 2008 9 25 0.37945451 30138 15046 9881 3467 1.748719 0 1 0 0 2008 9 26 0.38352002 33568 30138 15046 9881 1.752124 0 1 0 0 2008 9 27 0 0 33568 30138 15046 1.548611 0 1 0 0 2008 9 28 0.66097438 1991 0 33568 30138 1.741946 0 1 0 0 2008 9 29 0.76055017 3199 1991 0 33568 1.646229 0 1 0 0 2008 9 30 0.9476554 1834 3199 1991 0 1.794177 0 1 0 0 2008 10 0 0 0 1834 3199 1991 4.885714 0 1 0 0 2008 10 1 0.65072714 4057 0 1834 3199 1.694605 0 1 0 0 2008 10 2 0.84765069 2107 4057 0 1834 1.856111 0 1 0 0 2008 10 3 1.147301 1093 2107 4057 0 1.877949 0 1 0 0 2008 10 4 0.62078652 356 1093 2107 4057 1.247665 0 1 0 0 2008 10 5 0.65805727 7124 356 1093 2107 1.706438 0 1 0 0 2008 10 6 0.62809706 19575 7124 356 1093 1.230161 0 1 0 0 2008 10 7 1.0372208 806 19575 7124 356 1.275502 0 1 0 0 2008 10 8 0.8719457 4420 806 19575 7124 1.341428 0 1 0 0 2008 10 9 0.42160835 40725 4420 806 19575 1.204143 0 1 0 0 2008 10 10 0.64696268 7342 40725 4420 806 1.49729 0 1 0 0 2008 10 11 0.71040724 221 7342 40725 4420 1.954898 0 1 0 0 2008 10 12 1.9162562 203 221 7342 40725 1.816021 0 1 0 0 2008 10 13 0.42117832 34337 203 221 7342 1.447572 0 1 0 0 2008 10 14 0.77096115 1467 34337 203 221 1.853004 0 1 0 0 2008 10 15 0.78873239 781 1467 34337 203 1.74406 0 1 0 0 2008 10 16 0.48965754 20353 781 1467 34337 1.447912 0 1 0 0 2008 10 17 0.51625434 4891 20353 781 1467 1.714431 0 1 0 0 2008 10 18 0 0 4891 20353 781 2.091286 0 1 0 0 2008 10 19 0.5848713 16472 0 4891 20353 2.496216 0 1 0 0 2008 10 20 0.57273854 4035 16472 0 4891 1.128655 0 1 0 0 2008 10 21 0.61133239 7783 4035 16472 0 2.066289 0 1 0 0 2008 10 22 0.44207621 26086 7783 4035 16472 1.873593 0 1 0 0 2008 10 23 0.33255892 26687 26086 7783 4035 2.175946 0 1 0 0 2008 10 24 0 0 26687 26086 7783 1.924135 0 1 0 0 2008 10 25 0.6405959 1074 0 26687 26086 2.101145 0 1 0 0 2008 10 26 0.59855678 2633 1074 0 26687 1.677734 0 1 0 0 2008 10 27 0.38309898 29371 2633 1074 0 1.577089 0 1 0 0 2008 10 28 0.58305921 12160 29371 2633 1074 2.635406 0 1 0 0 2008 10 29 0.80635477 3997 12160 29371 2633 1.921309 0 1 0 0 2008 10 30 0.83018868 371 3997 12160 29371 1.60246 0 1 0 0 2008 10 31 0.53828306 862 371 3997 12160 1.590379 0 1 0 0 2008 11 0 0 0 862 371 3997 4.182927 0 1 0 0 2008 11 1 0 0 0 862 371 1.150689 0 1 0 0 2008 11 2 0.98853868 349 0 0 862 1.301387 0 1 0 0 2008 11 3 1.0634425 1923 349 0 0 1.187578 0 1 0 0 2008 11 4 0.65305672 4531 1923 349 0 1.386566 0 1 0 0 2008 11 5 1.0550953 2414 4531 1923 349 1.132758 0 1 0 0 2008 11 6 0.42752966 54029 2414 4531 1923 1.008201 0 1 0 0 2008 11 7 0.96661367 629 54029 2414 4531 1.332798 0 1 0 0 2008 11 8 0.85040984 488 629 54029 2414 1.147547 0 1 0 0 2008 11 9 0.80221424 3342 488 629 54029 1.845674 0 1 0 0 2008 11 10 0.72581528 12787 3342 488 629 1.140365 0 1 0 0 2008 11 11 0.56831263 13204 12787 3342 488 1.096131 0 1 0 0 2008 11 12 0.45718629 11842 13204 12787 3342 1.067515 0 1 0 0 2008 11 13 0.41517477 18768 11842 13204 12787 1.348147 0 1 0 0 2008 11 14 0.81405896 3087 18768 11842 13204 1.367967 0 1 0 0 2008 11 15 0 0 3087 18768 11842 1.56124 0 1 0 0 2008 11 16 0.80449658 1023 0 3087 18768 1.617886 0 1 0 0 2008 11 17 0.49588566 9236 1023 0 3087 1.732523 0 1 0 0 2008 11 18 0.70308123 4641 9236 1023 0 1.815413 0 1 0 0 2008 11 19 0.51111111 12015 4641 9236 1023 2.380401 0 1 0 0 2008 11 20 1.0199637 1653 12015 4641 9236 2.607551 0 1 0 0 2008 11 21 0.68832117 1370 1653 12015 4641 2.42177 0 1 0 0 2008 11 22 0 0 1370 1653 12015 2.89021 0 1 0 0 2008 11 23 0.77581864 1985 0 1370 1653 2.737656 0 1 0 0 2008 11 24 0.58421927 12040 1985 0 1370 2.998898 0 1 0 0 2008 11 25 0.79228792 1945 12040 1985 0 3.679722 0 1 0 0 2008 11 26 0.46221703 9409 1945 12040 1985 3.274931 0 1 0 0 2008 11 27 0 0 9409 1945 12040 2.20339 0 1 0 0 2008 11 28 0.96660959 1168 0 9409 1945 2.332255 0 1 0 0 2008 11 29 0 0 1168 0 9409 2.256736 0 1 0 0 2008 11 30 0.44171467 31423 0 1168 0 2.412934 0 1 0 0 2008 12 0 0 0 31423 0 1168 5.164557 0 1 0 0 2008 12 1 0.59224442 2553 0 31423 0 2.258952 0 1 0 0 2008 12 2 0.92135836 1119 2553 0 31423 2.159987 0 1 0 0 2008 12 3 1.1597387 6736 1119 2553 0 2.164919 0 1 0 0 2008 12 4 1.0419162 1670 6736 1119 2553 2.069746 0 1 0 0 2008 12 5 0.6858345 713 1670 6736 1119 1.514422 0 1 0 0 2008 12 6 0.69014085 355 713 1670 6736 1.185406 0 1 0 0 2008 12 7 0.45745942 53161 355 713 1670 1.915875 0 1 0 0 2008 12 8 0.74183351 2847 53161 355 713 1.416262 0 1 0 0 2008 12 9 0.77187807 3051 2847 53161 355 1.30201 0 1 0 0 2008 12 10 0.51492308 6500 3051 2847 53161 1.521209 0 1 0 0 2008 12 11 0.56686627 501 6500 3051 2847 1.52449 0 1 0 0 2008 12 12 1.1282288 1084 501 6500 3051 1.185816 0 1 0 0 2008 12 13 0 0 1084 501 6500 1.431553 0 1 0 0 2008 12 14 1.0334596 1584 0 1084 501 1.685842 0 1 0 0 2008 12 15 0.3952277 34365 1584 0 1084 1.724791 0 1 0 0 2008 12 16 0.82 450 34365 1584 0 2.138072 0 1 0 0 2008 12 17 1.3000464 2153 450 34365 1584 2.760896 0 1 0 0 2008 12 18 0.77947598 2748 2153 450 34365 3.107409 0 1 0 0 2008 12 19 0.61569389 7213 2748 2153 450 2.761955 0 1 0 0 2008 12 20 0.6605042 595 7213 2748 2153 2.820644 0 1 0 0 2008 12 21 0.63262572 9704 595 7213 2748 2.408631 0 1 0 0 2008 12 22 0.54501272 17695 9704 595 7213 2.382109 0 1 0 0 2008 12 23 0.3871037 13973 17695 9704 595 2.372947 0 1 0 0 2008 12 24 0.79372197 223 13973 17695 9704 2.470932 0 1 0 0 2008 12 26 0.38377393 4388 223 13973 17695 2.826341 0 1 0 0 2008 12 27 0.80973451 226 4388 223 13973 2.509842 0 1 0 0 2008 12 28 0.54623155 1287 226 4388 223 2.830909 0 1 0 0 2008 12 29 0.99164345 359 1287 226 4388 2.632444 0 1 0 0 2008 12 30 0.95283549 1781 359 1287 226 2.658201 0 1 0 0 2008 12 31 0.97993311 1196 1781 359 1287 2.553609 0 1 0 0 2009 1 0 0 0 1196 1781 359 2.833333 0 0 1 0 2009 1 2 1.0270957 1181 0 1196 1781 2.042464 0 0 1 0 2009 1 3 0.70322581 155 1181 0 1196 1.949521 0 0 1 0 2009 1 4 0 0 155 1181 0 1.917182 0 0 1 0 2009 1 5 0.777999 4018 0 155 1181 1.945585 0 0 1 0 2009 1 6 0.46312655 30225 4018 0 155 1.64918 0 0 1 0 2009 1 7 0.68318876 4842 30225 4018 0 1.7198 0 0 1 0 2009 1 8 0.59325961 9495 4842 30225 4018 2.100696 0 0 1 0 2009 1 9 0 0 9495 4842 30225 1.589135 0 0 1 0 2009 1 10 0 0 0 9495 4842 1.457505 0 0 1 0 2009 1 11 0.75818374 947 0 0 9495 2.115539 0 0 1 0 2009 1 12 0.64998186 2757 947 0 0 1.794628 0 0 1 0 2009 1 13 0.80537132 9197 2757 947 0 1.958435 0 0 1 0 2009 1 14 0.89765505 4563 9197 2757 947 1.745874 0 0 1 0 2009 1 15 0.79778213 1533 4563 9197 2757 1.467945 0 0 1 0 2009 1 16 0.95014144 2828 1533 4563 9197 1.402383 0 0 1 0 2009 1 17 0.7020202 396 2828 1533 4563 1.31639 0 0 1 0 2009 1 18 0.79185919 4545 396 2828 1533 1.450239 0 0 1 0 2009 1 19 0.6886188 19664 4545 396 2828 1.368486 0 0 1 0 2009 1 20 0.43937834 50510 19664 4545 396 1.26493 0 0 1 0 2009 1 21 0.48142791 19357 50510 19664 4545 1.24839 0 0 1 0 2009 1 22 0.53902417 11211 19357 50510 19664 1.270134 0 0 1 0 2009 1 23 0.52459016 7808 11211 19357 50510 1.309838 0 0 1 0 2009 1 24 0 0 7808 11211 19357 1.224955 0 0 1 0 2009 1 25 0.57279579 8359 0 7808 11211 1.357619 0 0 1 0 2009 1 26 0.56968533 17701 8359 0 7808 1.414388 0 0 1 0 2009 1 27 0.57179555 11289 17701 8359 0 1.361471 0 0 1 0 2009 1 28 0.4462754 13290 11289 17701 8359 1.263351 0 0 1 0 2009 1 29 0.34987128 21365 13290 11289 17701 1.285926 0 0 1 0 2009 1 30 0.41964629 65081 21365 13290 11289 1.378318 0 0 1 0 2009 1 31 0.7501356 5531 65081 21365 13290 1.320483 0 0 1 0 2009 2 1 0.4648286 9539 5531 65081 21365 1.531378 0 0 1 0 2009 2 2 0.48674912 7924 9539 5531 65081 1.457213 0 0 1 0 2009 2 3 0.54602865 19477 7924 9539 5531 1.488232 0 0 1 0 2009 2 4 0.43808339 3214 19477 7924 9539 1.473995 0 0 1 0 2009 2 5 0.55375112 10037 3214 19477 7924 1.74203 0 0 1 0 2009 2 6 0.50495751 1412 10037 3214 19477 1.646557 0 0 1 0 2009 2 7 0 0 1412 10037 3214 1.823097 0 0 1 0 2009 2 8 0.48445931 9266 0 1412 10037 2.073786 0 0 1 0 2009 2 9 0.61337546 9495 9266 0 1412 2.244848 0 0 1 0 2009 2 10 0 0 9495 9266 0 1.904748 0 0 1 0 2009 2 11 0.47606583 61126 0 9495 9266 2.123883 0 0 1 0 2009 2 12 0.49464707 11022 61126 0 9495 2.015063 0 0 1 0 2009 2 13 0.65057471 870 11022 61126 0 1.734914 0 0 1 0 2009 2 14 0 0 870 11022 61126 1.581369 0 0 1 0 2009 2 15 0.64311341 10034 0 870 11022 1.748556 0 0 1 0 2009 2 16 0.72029061 1927 10034 0 870 1.614084 0 0 1 0 2009 2 17 0.74599051 8854 1927 10034 0 1.524216 0 0 1 0 2009 2 18 0.85254692 3357 8854 1927 10034 1.495352 0 0 1 0 2009 2 19 0.59657103 20006 3357 8854 1927 1.35 0 0 1 0 2009 2 20 0.5014245 1053 20006 3357 8854 1.501805 0 0 1 0 2009 2 21 0 0 1053 20006 3357 2.148356 0 0 1 0 2009 2 22 0.70344009 5058 0 1053 20006 2.613952 0 0 1 0 2009 2 23 0.64221436 16113 5058 0 1053 2.306675 0 0 1 0 2009 2 24 0.41487142 38342 16113 5058 0 2.015231 0 0 1 0 2009 2 25 0.73840345 18540 38342 16113 5058 2.049118 0 0 1 0 2009 2 26 0.81176187 4149 18540 38342 16113 1.828338 0 0 1 0 2009 2 27 0.63000766 3919 4149 18540 38342 1.647072 0 0 1 0 2009 2 28 0 0 3919 4149 18540 1.316646 0 0 1 0 2009 3 0 0 0 0 3919 4149 2.456 0 0 1 0 2009 3 1 0.78539547 11202 0 0 3919 1.480058 0 0 1 0 2009 3 2 0.62437023 19253 11202 0 0 1.512964 0 0 1 0 2009 3 3 0.57340626 5851 19253 11202 0 1.470818 0 0 1 0 2009 3 4 0.6160425 7717 5851 19253 11202 1.366951 0 0 1 0 2009 3 5 0.62394705 3324 7717 5851 19253 1.340146 0 0 1 0 2009 3 6 0.58132118 8780 3324 7717 5851 1.370915 0 0 1 0 2009 3 7 0.58159393 1054 8780 3324 7717 1.363596 0 0 1 0 2009 3 8 0.48687574 32078 1054 8780 3324 1.490342 0 0 1 0 2009 3 9 0.75832622 1171 32078 1054 8780 1.474976 0 0 1 0 2009 3 10 0.79456607 5116 1171 32078 1054 1.489544 0 0 1 0 2009 3 11 0.61263172 30178 5116 1171 32078 1.599898 0 0 1 0 2009 3 12 0.44076728 9488 30178 5116 1171 1.725167 0 0 1 0 2009 3 13 0.71932115 1532 9488 30178 5116 1.696881 0 0 1 0 2009 3 14 0.60820368 707 1532 9488 30178 1.486632 0 0 1 0 2009 3 15 0.54964254 24898 707 1532 9488 1.57426 0 0 1 0 2009 3 16 0.44109969 12258 24898 707 1532 1.751613 0 0 1 0 2009 3 17 0.44573112 29879 12258 24898 707 1.620752 0 0 1 0 2009 3 18 0.55157794 27124 29879 12258 24898 1.586539 0 0 1 0 2009 3 19 0.44218792 11079 27124 29879 12258 1.492593 0 0 1 0 2009 3 20 0.39863223 53079 11079 27124 29879 1.181227 0 0 1 0 2009 3 21 0 0 53079 11079 27124 1.149268 0 0 1 0 2009 3 22 0.60208127 4036 0 53079 11079 1.111227 0 0 1 0 2009 3 23 0.55318221 11470 4036 0 53079 1.02428 0 0 1 0 2009 3 24 0.47925817 18441 11470 4036 0 1.150219 0 0 1 0 2009 3 25 0.49878464 12342 18441 11470 4036 1.735267 0 0 1 0 2009 3 26 0.83802628 3729 12342 18441 11470 1.634192 0 0 1 0 2009 3 27 0.63239719 1994 3729 12342 18441 1.412134 0 0 1 0 2009 3 28 0.90971168 1318 1994 3729 12342 1.369499 0 0 1 0 2009 3 29 0.80990234 4403 1318 1994 3729 1.642972 0 0 1 0 2009 3 30 0.54543111 15518 4403 1318 1994 1.600344 0 0 1 0 2009 3 31 0.46450177 55707 15518 4403 1318 1.78174 0 0 1 0 2009 4 1 0.55217423 13637 55707 15518 4403 2.130685 0 0 1 0 2009 4 2 0.62804414 5256 13637 55707 15518 1.959098 0 0 1 0 2009 4 3 0.70810481 1641 5256 13637 55707 1.972383 0 0 1 0 2009 4 4 0 0 1641 5256 13637 1.589725 0 0 1 0 2009 4 5 0.4891145 18419 0 1641 5256 2.112026 0 0 1 0 2009 4 6 0.85459387 3755 18419 0 1641 2.306198 0 0 1 0 2009 4 7 0.73545372 13268 3755 18419 0 2.574916 0 0 1 0 2009 4 8 0.71933962 5512 13268 3755 18419 2.641889 0 0 1 0 2009 4 9 0.95232198 3230 5512 13268 3755 2.651562 0 0 1 0 2009 4 10 0.90909091 462 3230 5512 13268 2.269578 0 0 1 0 2009 4 11 0.60368011 1413 462 3230 5512 2.362803 0 0 1 0 2009 4 12 0 0 1413 462 3230 3.1 0 0 1 0 2009 4 13 0.82885086 5726 0 1413 462 2.660755 0 0 1 0 2009 4 14 0.87816412 3595 5726 0 1413 2.231105 0 0 1 0 2009 4 15 0.59645731 22243 3595 5726 0 2.505588 0 0 1 0 2009 4 16 0.83392808 3643 22243 3595 5726 2.242573 0 0 1 0 2009 4 17 1.2502806 891 3643 22243 3595 1.935805 0 0 1 0 2009 4 18 0 0 891 3643 22243 1.494834 0 0 1 0 2009 4 19 0.78683386 2233 0 891 3643 1.693904 0 0 1 0 2009 4 20 0.99025974 2156 2233 0 891 1.599789 0 0 1 0 2009 4 21 0.80155161 8894 2156 2233 0 1.283365 0 0 1 0 2009 4 22 0.76358712 16247 8894 2156 2233 1.280265 0 0 1 0 2009 4 23 0.70237184 1307 16247 8894 2156 1.162554 0 0 1 0 2009 4 24 0 0 1307 16247 8894 1.206172 0 0 1 0 2009 4 25 0 0 0 1307 16247 1.246319 0 0 1 0 2009 4 26 0.87715517 464 0 0 1307 1.149621 0 0 1 0 2009 4 27 0.74825175 6292 464 0 0 1.182663 0 0 1 0 2009 4 28 1.0644457 1831 6292 464 0 1.327964 0 0 1 0 2009 4 29 0.59002555 18006 1831 6292 464 1.159307 0 0 1 0 2009 4 30 0.46716927 19631 18006 1831 6292 1.124995 0 0 1 0 2009 5 0 0 0 19631 18006 1831 1.973516 0 0 1 0 2009 5 1 0.64052863 1135 0 19631 18006 1.166901 0 0 1 0 2009 5 2 0 0 1135 0 19631 1.345861 0 0 1 0 2009 5 3 0.57290052 10967 0 1135 0 1.129691 0 0 1 0 2009 5 4 0.95861643 1619 10967 0 1135 1.344645 0 0 1 0 2009 5 5 0.91947769 919 1619 10967 0 1.657106 0 0 1 0 2009 5 6 0.5658363 19389 919 1619 10967 1.322279 0 0 1 0 2009 5 7 0 0 19389 919 1619 1.569729 0 0 1 0 2009 5 8 0.5 5800 0 19389 919 1.766201 0 0 1 0 2009 5 9 0 0 5800 0 19389 1.502934 0 0 1 0 2009 5 10 0.8060321 3614 0 5800 0 1.867917 0 0 1 0 2009 5 11 0.44130631 43267 3614 0 5800 1.393221 0 0 1 0 2009 5 12 0.70103093 679 43267 3614 0 1.557498 0 0 1 0 2009 5 13 0 0 679 43267 3614 1.797931 0 0 1 0 2009 5 14 0.57345392 8182 0 679 43267 2.233103 0 0 1 0 2009 5 15 0 0 8182 0 679 1.757839 0 0 1 0 2009 5 16 0 0 0 8182 0 1.742173 0 0 1 0 2009 5 17 0.59281529 6486 0 0 8182 2.72697 0 0 1 0 2009 5 18 0.6306163 5030 6486 0 0 2.080027 0 0 1 0 2009 5 19 0.76484648 5439 5030 6486 0 2.075235 0 0 1 0 2009 5 20 0.76 1875 5439 5030 6486 1.745549 0 0 1 0 2009 5 21 0.6019729 8414 1875 5439 5030 1.521141 0 0 1 0 2009 5 22 1.0296296 135 8414 1875 5439 1.531352 0 0 1 0 2009 5 23 0 0 135 8414 1875 1.157855 0 0 1 0 2009 5 24 0 0 0 135 8414 1.188311 0 0 1 0 2009 5 25 0.47467167 6929 0 0 135 1.184314 0 0 1 0 2009 5 26 0.46118705 54067 6929 0 0 1.251704 0 0 1 0 2009 5 27 0.40549503 4222 54067 6929 0 1.245096 0 0 1 0 2009 5 28 0.44890269 73771 4222 54067 6929 1.265314 0 0 1 0 2009 5 29 0.44468736 2287 73771 4222 54067 1.220115 0 0 1 0 2009 5 30 0.44995915 9792 2287 73771 4222 1.120337 0 0 1 0 2009 5 31 0 0 9792 2287 73771 1.033454 0 0 1 0 2009 6 0 0 0 0 9792 2287 2.509942 0 0 1 0 2009 6 1 0 0 0 0 9792 1.389484 0 0 1 0 2009 6 2 0.97058824 306 0 0 0 1.404109 0 0 1 0 2009 6 3 0.57393687 19706 306 0 0 1.161783 0 0 1 0 2009 6 4 0.75379939 329 19706 306 0 0.831376 0 0 1 0 2009 6 5 0.40875141 41639 329 19706 306 0.552162 0 0 1 0 2009 6 6 0.5 380 41639 329 19706 0.761577 0 0 1 0 2009 6 7 0.43203592 31627 380 41639 329 0.893402 0 0 1 0 2009 6 8 0.42680592 5745 31627 380 41639 0.754725 0 0 1 0 2009 6 9 0.32731936 22226 5745 31627 380 0.753377 0 0 1 0 2009 6 10 0.34349384 20373 22226 5745 31627 0.892577 0 0 1 0 2009 6 11 0.40233294 15174 20373 22226 5745 1.083803 0 0 1 0 2009 6 12 0 0 15174 20373 22226 1.016215 0 0 1 0 2009 6 13 0.55434783 276 0 15174 20373 1.038847 0 0 1 0 2009 6 14 0.3935206 15557 276 0 15174 1.107165 0 0 1 0 2009 6 15 0.49629891 2837 15557 276 0 1.044568 0 0 1 0 2009 6 16 0.41678014 28641 2837 15557 276 1.567858 0 0 1 0 2009 6 17 0.39331721 50847 28641 2837 15557 1.850676 0 0 1 0 2009 6 18 0.59838895 6083 50847 28641 2837 1.657414 0 0 1 0 2009 6 19 0.47139142 3076 6083 50847 28641 1.626004 0 0 1 0 2009 6 20 0.50463576 755 3076 6083 50847 1.886899 0 0 1 0 2009 6 21 0.41884302 7122 755 3076 6083 1.888889 0 0 1 0 2009 6 22 0.29999759 41557 7122 755 3076 1.854631 0 0 1 0 2009 6 23 0.37964882 9169 41557 7122 755 2.308639 0 0 1 0 2009 6 24 0.35275533 16459 9169 41557 7122 2.345243 0 0 1 0 2009 6 25 0.46066253 966 16459 9169 41557 2.223151 0 0 1 0 2009 6 26 0 0 966 16459 9169 2.211298 0 0 1 0 2009 6 27 0.82242991 321 0 966 16459 2.370696 0 0 1 0 2009 6 28 0.63328197 649 321 0 966 2.210894 0 0 1 0 2009 6 29 0.35754476 7820 649 321 0 2.522859 0 0 1 0 2009 6 30 0.72970297 1010 7820 649 321 2.249052 0 0 1 0 2009 7 0 0 0 1010 7820 649 2.511111 0 0 1 0 2009 7 1 0.48720251 2227 0 1010 7820 2.470783 0 0 1 0 2009 7 2 0.38526619 31913 2227 0 1010 2.027213 0 0 1 0 2009 7 3 0 0 31913 2227 0 1.722913 0 0 1 0 2009 7 4 0.64066852 718 0 31913 2227 1.356681 0 0 1 0 2009 7 5 0.69801085 553 718 0 31913 1.266174 0 0 1 0 2009 7 6 0.54157064 5641 553 718 0 2.100377 0 0 1 0 2009 7 7 0.5443038 7900 5641 553 718 1.672494 0 0 1 0 2009 7 8 0.56108347 3618 7900 5641 553 1.380118 0 0 1 0 2009 7 9 0.32709017 5179 3618 7900 5641 1.312135 0 0 1 0 2009 7 10 0.35745565 13473 5179 3618 7900 1.088998 0 0 1 0 2009 7 11 0.40841584 808 13473 5179 3618 1.313146 0 0 1 0 2009 7 12 0.54082889 2437 808 13473 5179 0.919212 0 0 1 0 2009 7 13 0.34631005 10095 2437 808 13473 1.446815 0 0 1 0 2009 7 14 0.72394044 873 10095 2437 808 1.552833 0 0 1 0 2009 7 15 0.35521406 47439 873 10095 2437 1.645664 0 0 1 0 2009 7 16 0.36872837 18496 47439 873 10095 1.542585 0 0 1 0 2009 7 17 3.9985994 714 18496 47439 873 1.454802 0 0 1 0 2009 7 18 0 0 714 18496 47439 1.770549 0 0 1 0 2009 7 19 0.40529695 1246 0 714 18496 1.56162 0 0 1 0 2009 7 20 0.32491381 13342 1246 0 714 1.703285 0 0 1 0 2009 7 21 0.50910834 1043 13342 1246 0 1.68855 0 0 1 0 2009 7 22 0.44475921 1765 1043 13342 1246 1.704732 0 0 1 0 2009 7 23 0.46512105 4874 1765 1043 13342 1.299165 0 0 1 0 2009 7 24 0.45472837 1491 4874 1765 1043 1.884675 0 0 1 0 2009 7 25 0 0 1491 4874 1765 1.569077 0 0 1 0 2009 7 26 0.60519167 3313 0 1491 4874 1.710299 0 0 1 0 2009 7 27 0.35138674 42005 3313 0 1491 1.64308 0 0 1 0 2009 7 28 0.71788413 794 42005 3313 0 1.638137 0 0 1 0 2009 7 29 0.85905225 823 794 42005 3313 2.102109 0 0 1 0 2009 7 30 0.55169035 4082 823 794 42005 2.047672 0 0 1 0 2009 7 31 0.96394984 638 4082 823 794 2.027213 0 0 1 0 2009 8 0 0 0 638 4082 823 2.413043 0 0 1 0 2009 8 1 0 0 0 638 4082 1.707454 0 0 1 0 2009 8 2 0.42990917 10679 0 0 638 1.62592 0 0 1 0 2009 8 3 0.65340909 528 10679 0 0 2.137047 0 0 1 0 2009 8 4 0.55291411 2608 528 10679 0 1.836463 0 0 1 0 2009 8 5 0.72612514 3644 2608 528 10679 2.145053 0 0 1 0 2009 8 6 0.35550945 23133 3644 2608 528 1.533443 0 0 1 0 2009 8 7 0.89519651 458 23133 3644 2608 1.362439 0 0 1 0 2009 8 8 0.63562753 494 458 23133 3644 1.697329 0 0 1 0 2009 8 9 0.33777677 4426 494 458 23133 1.445664 0 0 1 0 2009 8 10 0.60441094 3582 4426 494 458 0.751248 0 0 1 0 2009 8 11 0.48854115 8116 3582 4426 494 0.989114 0 0 1 0 2009 8 12 0.4500846 16548 8116 3582 4426 1.01292 0 0 1 0 2009 8 13 0.47695778 3103 16548 8116 3582 1.015178 0 0 1 0 2009 8 14 0.47150259 386 3103 16548 8116 1.17527 0 0 1 0 2009 8 15 0 0 386 3103 16548 1.322131 0 0 1 0 2009 8 16 0.65538194 1152 0 386 3103 0.844622 0 0 1 0 2009 8 17 0.63940917 6567 1152 0 386 1.062032 0 0 1 0 2009 8 18 0.72054089 1553 6567 1152 0 1.016693 0 0 1 0 2009 8 19 0.57121567 4697 1553 6567 1152 0.936848 0 0 1 0 2009 8 20 0.35999504 32231 4697 1553 6567 0.910684 0 0 1 0 2009 8 21 0.40402318 2933 32231 4697 1553 1.094096 0 0 1 0 2009 8 22 0 0 2933 32231 4697 1.752885 0 0 1 0 2009 8 23 0.66554282 2966 0 2933 32231 1.614685 0 0 1 0 2009 8 24 0.54287198 13482 2966 0 2933 1.61321 0 0 1 0 2009 8 25 0.71018277 383 13482 2966 0 2.053375 0 0 1 0 2009 8 26 0.55374593 307 383 13482 2966 2.235251 0 0 1 0 2009 8 27 0.8 510 307 383 13482 2.389249 0 0 1 0 2009 8 28 0.84864865 370 510 307 383 2.327482 0 0 1 0 2009 8 29 0.6424581 1253 370 510 307 2.091907 0 0 1 0 2009 8 30 0.50124723 14432 1253 370 510 2.049923 0 0 1 0 2009 8 31 0.41547946 12003 14432 1253 370 2.269283 0 0 1 0 2009 9 0 0 0 12003 14432 1253 1.545455 0 0 1 0 2009 9 1 0.71428571 875 0 12003 14432 2.338963 0 0 1 0 2009 9 2 0.70331588 1719 875 0 12003 2.376891 0 0 1 0 2009 9 3 0.55868956 5708 1719 875 0 1.847242 0 0 1 0 2009 9 4 0.6875 576 5708 1719 875 1.845285 0 0 1 0 2009 9 5 0 0 576 5708 1719 1.352808 0 0 1 0 2009 9 6 0.60953347 5916 0 576 5708 1.645522 0 0 1 0 2009 9 7 0.41878173 788 5916 0 576 1.653469 0 0 1 0 2009 9 8 0.50442105 11875 788 5916 0 1.452203 0 0 1 0 2009 9 9 0.39229396 6722 11875 788 5916 1.057956 0 0 1 0 2009 9 10 0.33497144 56366 6722 11875 788 0.84911 0 0 1 0 2009 9 11 1.4638633 761 56366 6722 11875 0.750257 0 0 1 0 2009 9 12 0 0 761 56366 6722 1.154523 0 0 1 0 2009 9 13 0.35546983 3629 0 761 56366 1.007334 0 0 1 0 2009 9 14 0.71467991 3624 3629 0 761 1.244092 0 0 1 0 2009 9 15 0.67756037 4804 3624 3629 0 1.396181 0 0 1 0 2009 9 16 0.73175074 8425 4804 3624 3629 1.467332 0 0 1 0 2009 9 17 0.73328855 2977 8425 4804 3624 1.458569 0 0 1 0 2009 9 18 0.67082294 1604 2977 8425 4804 1.397686 0 0 1 0 2009 9 19 0 0 1604 2977 8425 0.87659 0 0 1 0 2009 9 20 0.56947162 3577 0 1604 2977 1.218227 0 0 1 0 2009 9 21 0.40335174 25360 3577 0 1604 1.218343 0 0 1 0 2009 9 22 0.29937128 24017 25360 3577 0 1.113311 0 0 1 0 2009 9 23 0.33670235 66884 24017 25360 3577 1.280655 0 0 1 0 2009 9 24 0.5128746 13748 66884 24017 25360 1.26332 0 0 1 0 2009 9 25 0.5125448 1953 13748 66884 24017 1.243721 0 0 1 0 2009 9 26 0 0 1953 13748 66884 1.395506 0 0 1 0 2009 9 27 0.56710775 529 0 1953 13748 1.647426 0 0 1 0 2009 9 28 0.52705171 14232 529 0 1953 1.222553 0 0 1 0 2009 9 29 0.65244171 4546 14232 529 0 1.338676 0 0 1 0 2009 9 30 0.77240941 3146 4546 14232 529 1.674476 0 0 1 0 2009 10 0 0 0 3146 4546 14232 1.766667 0 0 1 0 2009 10 1 0.72284644 267 0 3146 4546 2.015369 0 0 1 0 2009 10 2 0.76264997 3834 267 0 3146 1.800809 0 0 1 0 2009 10 3 0 0 3834 267 0 1.687805 0 0 1 0 2009 10 4 0.46120401 14950 0 3834 267 1.707775 0 0 1 0 2009 10 5 0.35060443 26885 14950 0 3834 1.910799 0 0 1 0 2009 10 6 0.46731668 10051 26885 14950 0 1.998874 0 0 1 0 2009 10 7 0.3385786 63740 10051 26885 14950 1.60498 0 0 1 0 2009 10 8 0.33072395 16479 63740 10051 26885 1.642841 0 0 1 0 2009 10 9 0.97333333 300 16479 63740 10051 2.157059 0 0 1 0 2009 10 10 0 0 300 16479 63740 1.644737 0 0 1 0 2009 10 11 0.48284385 6878 0 300 16479 2.061501 0 0 1 0 2009 10 12 0.62126246 4515 6878 0 300 2.149747 0 0 1 0 2009 10 13 0.40381612 14727 4515 6878 0 2.269781 0 0 1 0 2009 10 14 0.37669201 10786 14727 4515 6878 2.389301 0 0 1 0 2009 10 15 0.38079352 8393 10786 14727 4515 2.199415 0 0 1 0 2009 10 16 0.34878198 13054 8393 10786 14727 1.911475 0 0 1 0 2009 10 17 0 0 13054 8393 10786 2.333333 0 0 1 0 2009 10 18 0.55021624 8324 0 13054 8393 2.741708 0 0 1 0 2009 10 19 0.45484308 8858 8324 0 13054 2.300497 0 0 1 0 2009 10 20 1.0516529 484 8858 8324 0 2.690076 0 0 1 0 2009 10 21 0.53796625 4504 484 8858 8324 2.407053 0 0 1 0 2009 10 22 0.35092176 25983 4504 484 8858 2.45148 0 0 1 0 2009 10 23 0.44346766 11489 25983 4504 484 2.017818 0 0 1 0 2009 10 24 0.4 10730 11489 25983 4504 1.837977 0 0 1 0 2009 10 25 0.49930686 12263 10730 11489 25983 2.420694 0 0 1 0 2009 10 26 0.52609506 6438 12263 10730 11489 1.782691 0 0 1 0 2009 10 27 0.58778524 15031 6438 12263 10730 1.616831 0 0 1 0 2009 10 28 0.44080384 11445 15031 6438 12263 1.482124 0 0 1 0 2009 10 29 0.35420032 17332 11445 15031 6438 1.381197 0 0 1 0 2009 10 30 0.46341463 738 17332 11445 15031 1.175177 0 0 1 0 2009 10 31 0 0 738 17332 11445 1.136816 0 0 1 0 2009 11 0 0 0 0 738 17332 1.5 0 0 1 0 2009 11 1 0.57582688 5684 0 0 738 1.550312 0 0 1 0 2009 11 2 0.6486692 2630 5684 0 0 1.216257 0 0 1 0 2009 11 3 0.3246878 10410 2630 5684 0 1.309734 0 0 1 0 2009 11 4 0.34123699 40825 10410 2630 5684 1.219311 0 0 1 0 2009 11 5 0.41864187 26242 40825 10410 2630 1.389096 0 0 1 0 2009 11 6 0.51319261 758 26242 40825 10410 1.555733 0 0 1 0 2009 11 7 0.56481481 4536 758 26242 40825 2.022844 0 0 1 0 2009 11 8 0.40494135 17647 4536 758 26242 2.16248 0 0 1 0 2009 11 9 0.83020833 960 17647 4536 758 1.736138 0 0 1 0 2009 11 10 0.7123839 3230 960 17647 4536 1.753245 0 0 1 0 2009 11 11 0.56945946 11100 3230 960 17647 2.035961 0 0 1 0 2009 11 12 0.64315486 3119 11100 3230 960 2.212461 0 0 1 0 2009 11 13 0.47204861 11520 3119 11100 3230 2.250419 0 0 1 0 2009 11 14 0 0 11520 3119 11100 2.045833 0 0 1 0 2009 11 15 0.60529359 15566 0 11520 3119 2.275994 0 0 1 0 2009 11 16 0.86530612 2695 15566 0 11520 2.329157 0 0 1 0 2009 11 17 0.65873016 3402 2695 15566 0 2.702028 0 0 1 0 2009 11 18 0.36003453 35916 3402 2695 15566 2.395487 0 0 1 0 2009 11 19 0.5262041 10838 35916 3402 2695 2.583393 0 0 1 0 2009 11 20 0.70886076 316 10838 35916 3402 2.231734 0 0 1 0 2009 11 21 316 10838 35916 1.689437 0 0 1 0 2009 11 22 0.72865375 3818 316 10838 1.967643 0 0 1 0 2009 11 23 0.650471 6051 3818 316 1.624666 0 0 1 0 2009 11 24 0.51029369 10249 6051 3818 2.040842 0 0 1 0 2009 11 25 0.44578313 1162 10249 6051 3818 2.088652 0 0 1 0 2009 11 26 0 0 1162 10249 6051 1.138824 0 0 1 0 2009 11 27 0.57133432 6063 0 1162 10249 1.978227 0 0 1 0 2009 11 28 0 0 6063 0 1162 2.291667 0 0 1 0 2009 11 29 0.76236934 2870 0 6063 0 1.844196 0 0 1 0 2009 11 30 0.40569561 2142 2870 0 6063 2.184727 0 0 1 0 2009 12 0 0 0 2142 2870 0 1.617152 0 0 1 0 2009 12 1 0.4132226 15292 0 2142 2870 2.283134 0 0 1 0 2009 12 2 0.31084947 33209 15292 0 2142 2.191217 0 0 1 0 2009 12 3 0.58361181 14364 33209 15292 0 2.461418 0 0 1 0 2009 12 4 0.90039063 512 14364 33209 15292 2.104202 0 0 1 0 2009 12 5 0 0 512 14364 33209 1.775665 0 0 1 0 2009 12 6 0.84221526 957 0 512 14364 2.013843 0 0 1 0 2009 12 7 1.2243402 1364 957 0 512 1.984653 0 0 1 0 2009 12 8 1.1695474 7579 1364 957 0 2.260461 0 0 1 0 2009 12 9 0.94579646 9944 7579 1364 957 2.34071 0 0 1 0 2009 12 10 0.9716895 4380 9944 7579 1364 2.37095 0 0 1 0 2009 12 11 1.11 400 4380 9944 7579 2.454182 0 0 1 0 2009 12 12 0 0 400 4380 9944 2.120215 0 0 1 0 2009 12 13 1.1727273 660 0 400 4380 2.387078 0 0 1 0 2009 12 14 1.1536773 911 660 0 400 2.091851 0 0 1 0 2009 12 15 1.1646341 984 911 660 0 2.090876 0 0 1 0 2009 12 16 0.95186581 8227 984 911 660 2.421228 0 0 1 0 2009 12 17 0.90659213 4293 8227 984 911 2.987608 0 0 1 0 2009 12 18 0 0 4293 8227 984 2.763661 0 0 1 0 2009 12 19 0.90583804 531 0 4293 8227 2.367242 0 0 1 0 2009 12 20 1.232327 2716 531 0 4293 3.86763 0 0 1 0 2009 12 21 0.99485798 4084 2716 531 0 3.339201 0 0 1 0 2009 12 22 0.33218423 24947 4084 2716 531 2.969187 0 0 1 0 2009 12 23 0 0 24947 4084 2716 2.889906 0 0 1 0 2009 12 24 0.87312572 867 0 24947 4084 3.390178 0 0 1 0 2009 12 25 0 0 867 0 24947 2.381818 0 0 1 0 2009 12 26 0 0 0 867 0 2.626617 0 0 1 0 2009 12 27 1.3159743 1709 0 0 867 4.084115 0 0 1 0 2009 12 28 1.0213333 375 1709 0 0 3.060235 0 0 1 0 2009 12 29 1.1527308 2069 375 1709 0 2.923846 0 0 1 0 2009 12 30 1.2420814 442 2069 375 1709 3.221184 0 0 1 0 2009 12 31 2.4372439 2685 442 2069 375 2.384444 0 0 1 0 2010 1 1 0.86805556 288 2685 442 2069 2.085417 0 0 0 1 2010 1 2 0.93351648 1820 288 2685 442 2.220862 0 0 0 1 2010 1 3 1.0409683 1611 1820 288 2685 2.714309 0 0 0 1 2010 1 4 0.70461197 9909 1611 1820 288 2.51453 0 0 0 1 2010 1 5 1.1811518 955 9909 1611 1820 2.047789 0 0 0 1 2010 1 6 0.71772956 9126 955 9909 1611 1.94315 0 0 0 1 2010 1 7 0.77010018 3893 9126 955 9909 2.067289 0 0 0 1 2010 1 8 0.72769142 1737 3893 9126 955 1.726665 0 0 0 1 2010 1 9 0.80288462 208 1737 3893 9126 1.554674 0 0 0 1 2010 1 10 0.96695821 4116 208 1737 3893 1.855509 0 0 0 1 2010 1 11 0.7616336 16332 4116 208 1737 1.45779 0 0 0 1 2010 1 12 0.65626839 6796 16332 4116 208 1.224783 0 0 0 1 2010 1 13 0.69080235 5621 6796 16332 4116 1.182877 0 0 0 1 2010 1 14 0.48321436 28298 5621 6796 16332 1.048563 0 0 0 1 2010 1 15 0.57138404 3208 28298 5621 6796 1.155518 0 0 0 1 2010 1 16 0.58983051 885 3208 28298 5621 1.11416 0 0 0 1 2010 1 17 0.56650937 15682 885 3208 28298 1.188427 0 0 0 1 2010 1 18 0.59215514 18203 15682 885 3208 1.448997 0 0 0 1 2010 1 19 0.51404775 18010 18203 15682 885 1.451939 0 0 0 1 2010 1 20 0.53493405 14785 18010 18203 15682 1.722428 0 0 0 1 2010 1 21 0.5458601 5604 14785 18010 18203 1.471717 0 0 0 1 2010 1 22 0.70013254 3018 5604 14785 18010 1.56484 0 0 0 1 2010 1 23 0 0 3018 5604 14785 1.487578 0 0 0 1 2010 1 24 0.56081966 12981 0 3018 5604 1.738824 0 0 0 1 2010 1 25 0.55240848 5190 12981 0 3018 1.836988 0 0 0 1 2010 1 26 0.46871113 36978 5190 12981 0 1.667966 0 0 0 1 2010 1 27 0.4011896 23201 36978 5190 12981 1.682852 0 0 0 1 2010 1 28 0.44296408 14561 23201 36978 5190 1.832715 0 0 0 1 2010 1 29 0.54294748 5693 14561 23201 36978 1.926568 0 0 0 1 2010 1 30 0.55944056 286 5693 14561 23201 2.129372 0 0 0 1 2010 1 31 0.60964491 4168 286 5693 14561 1.924082 0 0 0 1 2010 2 1 0.47344167 8246 4168 286 5693 2.025357 0 0 0 1 2010 2 2 0.86385256 1682 8246 4168 286 1.900429 0 0 0 1 2010 2 3 0.47115593 18444 1682 8246 4168 2.088912 0 0 0 1 2010 2 4 0.50543478 10304 18444 1682 8246 1.816796 0 0 0 1 2010 2 5 0.77245509 668 10304 18444 1682 1.603679 0 0 0 1 2010 2 6 0.49906542 535 668 10304 18444 1.597403 0 0 0 1 2010 2 7 0.48112387 11496 535 668 10304 1.678242 0 0 0 1 2010 2 8 0.49832867 14659 11496 535 668 1.738337 0 0 0 1 2010 2 9 0.52507767 11265 14659 11496 535 1.690582 0 0 0 1 2010 2 10 0.47282563 11831 11265 14659 11496 1.817989 0 0 0 1 2010 2 11 0.45107389 24118 11831 11265 14659 1.681975 0 0 0 1 2010 2 12 0.71627907 215 24118 11831 11265 1.660167 0 0 0 1 2010 2 13 0 0 215 24118 11831 1.659363 0 0 0 1 2010 2 14 0 0 0 215 24118 1.684714 0 0 0 1 2010 2 15 1.2056921 773 0 0 215 2.021264 0 0 0 1 2010 2 16 1.0814234 3316 773 0 0 2.312347 0 0 0 1 2010 2 17 0.69764545 16861 3316 773 0 2.254911 0 0 0 1 2010 2 18 0.77431232 4617 16861 3316 773 2.189548 0 0 0 1 2010 2 19 0 0 4617 16861 3316 2.151767 0 0 0 1 2010 2 20 0 0 0 4617 16861 1.758597 0 0 0 1 2010 2 21 0.65436324 4091 0 0 4617 1.896123 0 0 0 1 2010 2 22 0.68477874 8429 4091 0 0 1.993786 0 0 0 1 2010 2 23 0.68249036 9332 8429 4091 0 1.964947 0 0 0 1 2010 2 24 0.55229178 19439 9332 8429 4091 1.946459 0 0 0 1 2010 2 25 0.44349494 23989 19439 9332 8429 1.402368 0 0 0 1 2010 2 26 0.29649123 570 23989 19439 9332 1.458376 0 0 0 1 2010 2 27 0 0 570 23989 19439 1.632032 0 0 0 1 2010 2 28 0.65033784 3552 0 570 23989 1.550133 0 0 0 1 2010 3 1 0.62339597 8728 3552 0 570 1.562306 0 0 0 1 2010 3 2 0.5758715 5852 8728 3552 0 1.594329 0 0 0 1 2010 3 3 0.89608177 1174 5852 8728 3552 1.813818 0 0 0 1 2010 3 4 1.2117528 987 1174 5852 8728 1.994391 0 0 0 1 2010 3 5 1.4521739 230 987 1174 5852 2.118846 0 0 0 1 2010 3 6 0 0 230 987 1174 2.031433 0 0 0 1 2010 3 7 0.89180328 610 0 230 987 2.192702 0 0 0 1 2010 3 8 0.78214736 13477 610 0 230 2.091705 0 0 0 1 2010 3 9 0.70939694 30478 13477 610 0 1.930551 0 0 0 1 2010 3 10 0.77498863 10995 30478 13477 610 1.932262 0 0 0 1 2010 3 11 0.69828135 7855 10995 30478 13477 1.676983 0 0 0 1 2010 3 12 0.6217512 7926 7855 10995 30478 1.433769 0 0 0 1 2010 3 13 0.62080537 298 7926 7855 10995 1.274348 0 0 0 1 2010 3 14 0.73377704 1202 298 7926 7855 1.521358 0 0 0 1 2010 3 15 0.52861121 38429 1202 298 7926 1.479957 0 0 0 1 2010 3 16 0.65486265 7681 38429 1202 298 1.591456 0 0 0 1 2010 3 17 0.64043382 7192 7681 38429 1202 1.555985 0 0 0 1 2010 3 18 1.2619808 313 7192 7681 38429 1.595012 0 0 0 1 2010 3 19 0.56781418 7233 313 7192 7681 1.64775 0 0 0 1 2010 3 20 0 0 7233 313 7192 1.308344 0 0 0 1 2010 3 21 0.61724717 13092 0 7233 313 1.458608 0 0 0 1 2010 3 22 0.78664752 6965 13092 0 7233 1.538302 0 0 0 1 2010 3 23 0.67724538 9642 6965 13092 0 1.577375 0 0 0 1 2010 3 24 0.51148711 36998 9642 6965 13092 1.502052 0 0 0 1 2010 3 25 0.4320233 41551 36998 9642 6965 1.309123 0 0 0 1 2010 3 26 0.33095578 701 41551 36998 9642 1.156113 0 0 0 1 2010 3 27 0 0 701 41551 36998 1.399756 0 0 0 1 2010 3 28 0.73316391 3148 0 701 41551 1.545872 0 0 0 1 2010 3 29 0.45082548 18353 3148 0 701 1.630849 0 0 0 1 2010 3 30 0.7328125 4480 18353 3148 0 1.539373 0 0 0 1 2010 3 31 0.56890501 17096 4480 18353 3148 1.642106 0 0 0 1 2010 4 1 0.94193548 930 17096 4480 18353 1.688129 0 0 0 1 2010 4 2 0.623069 1942 930 17096 4480 1.50973 0 0 0 1 2010 4 3 0.51681076 4164 1942 930 17096 1.446622 0 0 0 1 2010 4 4 0 0 4164 1942 930 1.928666 0 0 0 1 2010 4 5 0.88500624 5609 0 4164 1942 1.612722 0 0 0 1 2010 4 6 0.60219407 7657 5609 0 4164 1.649039 0 0 0 1 2010 4 7 0.49388492 27146 7657 5609 0 1.960325 0 0 0 1 2010 4 8 0.85467791 2608 27146 7657 5609 1.806374 0 0 0 1 2010 4 9 0.81506849 730 2608 27146 7657 1.539306 0 0 0 1 2010 4 10 0 0 730 2608 27146 1.579512 0 0 0 1 2010 4 11 0.68139224 2241 0 730 2608 1.67682 0 0 0 1 2010 4 12 1.0586895 3510 2241 0 730 1.501998 0 0 0 1 2010 4 13 0.64583631 14002 3510 2241 0 1.802177 0 0 0 1 2010 4 14 1.0616642 1346 14002 3510 2241 1.473785 0 0 0 1 2010 4 15 0 0 1346 14002 3510 1.699448 0 0 0 1 2010 4 16 0.73882763 5169 0 1346 14002 1.297251 0 0 0 1 2010 4 17 0.56108597 221 5169 0 1346 1.338983 0 0 0 1 2010 4 18 0.55461558 4019 221 5169 0 1.196435 0 0 0 1 2010 4 19 0.59360474 12478 4019 221 5169 1.238991 0 0 0 1 2010 4 20 0.9321663 457 12478 4019 221 1.304029 0 0 0 1 2010 4 21 0.50325028 70917 457 12478 4019 1.202053 0 0 0 1 2010 4 22 0.92915531 2569 70917 457 12478 1.199911 0 0 0 1 2010 4 23 0.84295261 3292 2569 70917 457 1.263848 0 0 0 1 2010 4 24 0.70180723 332 3292 2569 70917 1.265827 0 0 0 1 2010 4 25 0.78664384 2920 332 3292 2569 1.104819 0 0 0 1 2010 4 26 0.84176764 5612 2920 332 3292 1.222778 0 0 0 1 2010 4 27 0.89416232 2107 5612 2920 332 1.142558 0 0 0 1 2010 4 28 0.6894264 14470 2107 5612 2920 1.323045 0 0 0 1 2010 4 29 0.46539255 34429 14470 2107 5612 1.461259 0 0 0 1 2010 4 30 0.50143403 2092 34429 14470 2107 1.926123 0 0 0 1 2010 5 1 0 0 2092 34429 14470 1.601562 0 0 0 1 2010 5 2 0.70936439 2061 0 2092 34429 1.533883 0 0 0 1 2010 5 3 0.53014666 1841 2061 0 2092 1.748729 0 0 0 1 2010 5 4 0.99115044 339 1841 2061 0 2.00557 0 0 0 1 2010 5 5 0.38000294 34055 339 1841 2061 2.466594 0 0 0 1 2010 5 6 0.98305085 767 34055 339 1841 2.419542 0 0 0 1 2010 5 7 0 0 767 34055 339 2.120925 0 0 0 1 2010 5 8 0 0 0 767 34055 1.289465 0 0 0 1 2010 5 9 0 0 0 0 767 2.533812 0 0 0 1 2010 5 10 0 0 0 0 0 2.700275 0 0 0 1 2010 5 11 1.0034247 2044 0 0 0 2.669407 0 0 0 1 2010 5 12 0.48474714 69069 2044 0 0 2.254659 0 0 0 1 2010 5 13 0.46496376 5794 69069 2044 0 2.073566 0 0 0 1 2010 5 14 0.8204804 791 5794 69069 2044 2.247067 0 0 0 1 2010 5 15 0 0 791 5794 69069 1.976671 0 0 0 1 2010 5 16 0.78330914 6890 0 791 5794 2.754879 0 0 0 1 2010 5 17 0.86398581 1691 6890 0 791 2.211715 0 0 0 1 2010 5 18 0.9633111 3298 1691 6890 0 1.972249 0 0 0 1 2010 5 19 0 0 3298 1691 6890 2.208122 0 0 0 1 2010 5 20 0.86333333 600 0 3298 1691 1.889904 0 0 0 1 2010 5 21 1.2 2570 600 0 3298 1.734328 0 0 0 1 2010 5 22 0 0 2570 600 0 1.659541 0 0 0 1 2010 5 23 1.2181208 1490 0 2570 600 2.987324 0 0 0 1 2010 5 24 0.89348564 6478 1490 0 2570 1.483202 0 0 0 1 2010 5 25 0.78076063 2682 6478 1490 0 1.709666 0 0 0 1 2010 5 26 0.49957461 84628 2682 6478 1490 1.851916 0 0 0 1 2010 5 27 0.52199959 4841 84628 2682 6478 1.507255 0 0 0 1 2010 5 28 0 0 4841 84628 2682 1.314867 0 0 0 1 2010 5 29 0 0 0 4841 84628 0.957122 0 0 0 1 2010 5 30 0.89035917 529 0 0 4841 1.059361 0 0 0 1 2010 5 31 0.84070796 452 529 0 0 2.295489 0 0 0 1 2010 6 1 0.57173489 30780 452 529 0 1.707573 0 0 0 1 2010 6 2 0.60494845 2425 30780 452 529 2.044102 0 0 0 1 2010 6 3 0 0 2425 30780 452 1.988752 0 0 0 1 2010 6 4 0 0 0 2425 30780 1.630376 0 0 0 1 2010 6 5 0 0 0 0 2425 1.785003 0 0 0 1 2010 6 6 0.47611918 68130 0 0 0 2.057992 0 0 0 1 2010 6 7 0.51331084 2667 68130 0 0 1.880024 0 0 0 1 2010 6 8 1.1981609 2175 2667 68130 0 1.401553 0 0 0 1 2010 6 9 0.49198265 35735 2175 2667 68130 1.532282 0 0 0 1 2010 6 10 0.55942623 976 35735 2175 2667 1.357696 0 0 0 1 2010 6 11 0.59247573 4120 976 35735 2175 1.3391 0 0 0 1 2010 6 12 0 0 4120 976 35735 1.429951 0 0 0 1 2010 6 13 0.55080483 7952 0 4120 976 1.681166 0 0 0 1 2010 6 14 0.50371327 34067 7952 0 4120 1.943496 0 0 0 1 2010 6 15 0.44865906 13498 34067 7952 0 1.853281 0 0 0 1 2010 6 16 0.4688742 111467 13498 34067 7952 1.81695 0 0 0 1 2010 6 17 0.53750967 1293 111467 13498 34067 1.829418 0 0 0 1 2010 6 18 1.0204082 147 1293 111467 13498 1.38436 0 0 0 1 2010 6 19 0 0 147 1293 111467 1.201381 0 0 0 1 2010 6 20 0.97464342 631 0 147 1293 1.717221 0 0 0 1 2010 6 21 0.52164362 7739 631 0 147 1.353825 0 0 0 1 2010 6 22 0.46201782 51511 7739 631 0 1.5953 0 0 0 1 2010 6 23 0.52656641 41161 51511 7739 631 1.34329 0 0 0 1 2010 6 24 1.0675944 503 41161 51511 7739 1.178799 0 0 0 1 2010 6 25 0.51323829 491 503 41161 51511 1.127787 0 0 0 1 2010 6 26 0 0 491 503 41161 1.684263 0 0 0 1 2010 6 27 0.58166757 5553 0 491 503 1.443223 0 0 0 1 2010 6 28 0.53541707 10334 5553 0 491 1.339159 0 0 0 1 2010 6 29 0.44741882 73106 10334 5553 0 1.671346 0 0 0 1 2010 6 30 0.47189835 42969 73106 10334 5553 1.462489 0 0 0 1 2010 7 1 0.60267857 448 42969 73106 10334 2.107713 0 0 0 1 2010 7 2 0 0 448 42969 73106 1.885628 0 0 0 1 2010 7 3 0 0 0 448 42969 2.136952 0 0 0 1 2010 7 4 1.4785276 326 0 0 448 2.144978 0 0 0 1 2010 7 5 0.71067416 1068 326 0 0 2.303175 0 0 0 1 2010 7 6 0.52908416 24240 1068 326 0 2.222152 0 0 0 1 2010 7 7 0.5460885 7887 24240 1068 326 2.256838 0 0 0 1 2010 7 8 0.46658897 23166 7887 24240 1068 2.474485 0 0 0 1 2010 7 9 0 0 23166 7887 24240 2.276537 0 0 0 1 2010 7 10 0.00494493 4449 0 23166 7887 1.643021 0 0 0 1 2010 7 11 0.55294118 340 4449 0 23166 2.113863 0 0 0 1 2010 7 12 0.40011955 117105 340 4449 0 2.093115 0 0 0 1 2010 7 13 0.27013221 3328 117105 340 4449 2.199393 0 0 0 1 2010 7 14 0.36172629 7044 3328 117105 340 2.067204 0 0 0 1 2010 7 15 0.50652565 4444 7044 3328 117105 2.003424 0 0 0 1 2010 7 16 0.58290946 983 4444 7044 3328 2.037892 0 0 0 1 2010 7 17 0 0 983 4444 7044 2.010725 0 0 0 1 2010 7 18 0.53072219 6868 0 983 4444 2.135974 0 0 0 1 2010 7 19 0.40417597 69445 6868 0 983 1.409792 0 0 0 1 2010 7 20 0.51592469 5683 69445 6868 0 1.527191 0 0 0 1 2010 7 21 0.62412669 2147 5683 69445 6868 1.745812 0 0 0 1 2010 7 22 0.47540442 30290 2147 5683 69445 1.753833 0 0 0 1 2010 7 23 0.49808061 2084 30290 2147 5683 1.144449 0 0 0 1 2010 7 24 0.66197183 284 2084 30290 2147 1.854684 0 0 0 1 2010 7 25 0.72336449 535 284 2084 30290 2.088704 0 0 0 1 2010 7 26 0.53671365 4753 535 284 2084 1.485626 0 0 0 1 2010 7 27 0.51993413 11538 4753 535 284 1.423606 0 0 0 1 2010 7 28 0.40668299 73111 11538 4753 535 1.398416 0 0 0 1 2010 7 29 0.57592632 4669 73111 11538 4753 1.443721 0 0 0 1 2010 7 30 0.70314164 3756 4669 73111 11538 1.455902 0 0 0 1 2010 7 31 0 0 3756 4669 73111 1.692222 0 0 0 1 2010 8 1 0.41694915 590 0 3756 4669 2.588991 0 0 0 1 2010 8 2 0.48037563 4153 590 0 3756 1.558599 0 0 0 1 2010 8 3 0.76790123 405 4153 590 0 2.068287 0 0 0 1 2010 8 4 0.53740715 11308 405 4153 590 1.900984 0 0 0 1 2010 8 5 0.50673907 10313 11308 405 4153 2.048382 0 0 0 1 2010 8 6 0 0 10313 11308 405 2.101384 0 0 0 1 2010 8 7 0 0 0 10313 11308 2.140533 0 0 0 1 2010 8 8 0.55369979 4730 0 0 10313 2.025287 0 0 0 1 2010 8 9 0.50930762 22401 4730 0 0 2.271426 0 0 0 1 2010 8 10 0.39324552 107040 22401 4730 0 2.226626 0 0 0 1 2010 8 11 0.40998363 3666 107040 22401 4730 2.039968 0 0 0 1 2010 8 12 0.45797885 12577 3666 107040 22401 1.877574 0 0 0 1 2010 8 13 0.40949227 4530 12577 3666 107040 1.776402 0 0 0 1 2010 8 14 0 0 4530 12577 3666 1.467492 0 0 0 1 2010 8 15 0.57929227 1526 0 4530 12577 1.847571 0 0 0 1 2010 8 16 0.47210632 19413 1526 0 4530 1.750973 0 0 0 1 2010 8 17 0.61501869 2943 19413 1526 0 1.630055 0 0 0 1 2010 8 18 0.53166006 11355 2943 19413 1526 1.981355 0 0 0 1 2010 8 19 0.79011639 5241 11355 2943 19413 2.063577 0 0 0 1 2010 8 20 0 0 5241 11355 2943 2.160763 0 0 0 1 2010 8 21 0 0 0 5241 11355 1.806422 0 0 0 1 2010 8 22 0.58513189 2919 0 0 5241 1.667982 0 0 0 1 2010 8 23 0.41979103 60199 2919 0 0 1.709314 0 0 0 1 2010 8 24 0.45254339 25065 60199 2919 0 1.874703 0 0 0 1 2010 8 25 0.36617114 27580 25065 60199 2919 2.29971 0 0 0 1 2010 8 26 0.50833333 1200 27580 25065 60199 1.730535 0 0 0 1 2010 8 27 0 0 1200 27580 25065 2.595163 0 0 0 1 2010 8 28 0.53456221 217 0 1200 27580 1.813109 0 0 0 1 2010 8 29 0.77509035 3597 217 0 1200 2.361194 0 0 0 1 2010 8 30 0.65293162 6771 3597 217 0 2.255746 0 0 0 1 2010 8 31 0.51062399 14872 6771 3597 217 2.05232 0 0 0 1 2010 9 1 0.46303105 53856 14872 6771 3597 2.354407 0 0 0 1 2010 9 2 0.4525447 29080 53856 14872 6771 2.21172 0 0 0 1 2010 9 3 0 0 29080 53856 14872 1.453795 0 0 0 1 2010 9 4 0 0 0 29080 53856 1.925267 0 0 0 1 2010 9 5 0 0 0 0 29080 1.883186 0 0 0 1 2010 9 6 0.50275267 9627 0 0 0 2.682665 0 0 0 1 2010 9 7 0.60275296 12423 9627 0 0 2.614273 0 0 0 1 2010 9 8 0.64009199 10436 12423 9627 0 2.499795 0 0 0 1 2010 9 9 0.77349943 1766 10436 12423 9627 3.554189 0 0 0 1 2010 9 10 0 0 1766 10436 12423 3.319495 0 0 0 1 2010 9 11 0 0 0 1766 10436 1.968712 0 0 0 1 2010 9 12 0.64939535 10750 0 0 1766 2.810609 0 0 0 1 2010 9 13 0.45958172 51066 10750 0 0 2.244248 0 0 0 1 2010 9 14 0.51334604 8392 51066 10750 0 2.299134 0 0 0 1 2010 9 15 0.75704737 3441 8392 51066 10750 2.155101 0 0 0 1 2010 9 16 0.44721877 30598 3441 8392 51066 2.147152 0 0 0 1 2010 9 17 1.0062794 637 30598 3441 8392 2.100437 0 0 0 1 2010 9 18 0 0 637 30598 3441 1.80383 0 0 0 1 2010 9 19 0.45949493 38094 0 637 30598 1.823785 0 0 0 1 2010 9 20 0.45411627 22241 38094 0 637 2.006054 0 0 0 1 2010 9 21 0.61544172 4041 22241 38094 0 2.071839 0 0 0 1 2010 9 22 0.84099909 3283 4041 22241 38094 2.081613 0 0 0 1 2010 9 23 0.5137541 13705 3283 4041 22241 2.394724 0 0 0 1 2010 9 24 0.96013865 577 13705 3283 4041 1.960789 0 0 0 1 2010 9 25 0 0 577 13705 3283 1.815243 0 0 0 1 2010 9 26 0.47101637 30914 0 577 13705 2.707356 0 0 0 1 2010 9 27 0.71094628 4449 30914 0 577 2.238517 0 0 0 1 2010 9 28 0.88559483 2631 4449 30914 0 2.262676 0 0 0 1 2010 9 29 0.48485199 51855 2631 4449 30914 2.504768 0 0 0 1 2010 9 30 0.6610237 5021 51855 2631 4449 2.307892 0 0 0 1 2010 10 0 0 0 5021 51855 2631 1.545455 0 0 0 1 2010 10 1 0.67118227 812 0 5021 51855 2.157537 0 0 0 1 2010 10 2 0 0 812 0 5021 3.77242 0 0 0 1 2010 10 3 0.68586985 11387 0 812 0 2.374808 0 0 0 1 2010 10 4 0.95103448 2900 11387 0 812 2.510711 0 0 0 1 2010 10 5 0.63210843 14314 2900 11387 0 2.817149 0 0 0 1 2010 10 6 0.73473033 8399 14314 2900 11387 2.267488 0 0 0 1 2010 10 7 0.70071719 5159 8399 14314 2900 2.212316 0 0 0 1 2010 10 8 0.91851852 540 5159 8399 14314 2.443706 0 0 0 1 2010 10 9 0 0 540 5159 8399 1.983146 0 0 0 1 2010 10 10 0 0 0 540 5159 3.646665 0 0 0 1 2010 10 11 0.61810802 8869 0 0 540 3.410188 0 0 0 1 2010 10 12 0.5819987 27598 8869 0 0 2.499101 0 0 0 1 2010 10 13 0.69621692 6159 27598 8869 0 2.488742 0 0 0 1 2010 10 14 0.59177713 22255 6159 27598 8869 2.364818 0 0 0 1 2010 10 15 0.58306334 8951 22255 6159 27598 1.907881 0 0 0 1 2010 10 16 0 0 8951 22255 6159 2 0 0 0 1 2010 10 17 0.78186866 8284 0 8951 22255 2.324121 0 0 0 1 2010 10 18 0.57193974 15930 8284 0 8951 1.577322 0 0 0 1 2010 10 19 1.1892697 1342 15930 8284 0 2.886065 0 0 0 1 2010 10 20 1.1534268 1284 1342 15930 8284 2.364084 0 0 0 1 2010 10 21 0.82010406 9033 1284 1342 15930 2.173596 0 0 0 1 2010 10 22 0.80192835 12031 9033 1284 1342 2.151187 0 0 0 1 2010 10 23 0 0 12031 9033 1284 2.4 0 0 0 1 2010 10 24 0.50697098 78253 0 12031 9033 3.025806 0 0 0 1 2010 10 25 0.78353037 2453 78253 0 12031 2.212064 0 0 0 1 2010 10 26 0.49608043 40183 2453 78253 0 2.255528 0 0 0 1 2010 10 27 0.53053435 8646 40183 2453 78253 2.001252 0 0 0 1 2010 10 28 0.61999034 4142 8646 40183 2453 1.734869 0 0 0 1 2010 10 29 0.61074197 1806 4142 8646 40183 1.667303 0 0 0 1 2010 10 30 0 0 1806 4142 8646 1.442105 0 0 0 1 2010 10 31 0.76615494 2801 0 1806 4142 2.203565 0 0 0 1 2010 11 0 0 0 2801 0 1806 1.5 0 0 0 1 2010 11 1 0.51836568 22052 0 2801 0 2.068416 0 0 0 1 2010 11 2 0.75770457 941 22052 0 2801 2.307574 0 0 0 1 2010 11 3 0.59899638 13551 941 22052 0 2.087111 0 0 0 1 2010 11 4 0.53452516 27241 13551 941 22052 2.172353 0 0 0 1 2010 11 5 0.5110329 7387 27241 13551 941 1.876083 0 0 0 1 2010 11 6 0 0 7387 27241 13551 2.8509 0 0 0 1 2010 11 7 1.0110497 905 0 7387 27241 2.533017 0 0 0 1 2010 11 8 1.1317411 1298 905 0 7387 2.459712 0 0 0 1 2010 11 9 1.6135593 295 1298 905 0 2.504884 0 0 0 1 2010 11 10 0.84397596 11479 295 1298 905 2.957759 0 0 0 1 2010 11 11 1.3049422 1902 11479 295 1298 2.936036 0 0 0 1 2010 11 12 0 0 1902 11479 295 1.538793 0 0 0 1 2010 11 13 0 0 0 1902 11479 1.878947 0 0 0 1 2010 11 14 1.268319 928 0 0 1902 2.837915 0 0 0 1 2010 11 15 1.4872727 1375 928 0 0 3.013648 0 0 0 1 2010 11 16 0.97508101 12962 1375 928 0 3.071183 0 0 0 1 2010 11 17 0.68368377 11890 12962 1375 928 2.657021 0 0 0 1 2010 11 18 0.7309417 892 11890 12962 1375 2.554325 0 0 0 1 2010 11 19 1.2396313 217 892 11890 12962 2.179269 0 0 0 1 2010 11 20 0 0 217 892 11890 1.812834 0 0 0 1 2010 11 21 0.62701956 9408 0 217 892 1.641951 0 0 0 1 2010 11 22 0.86571315 2517 9408 0 217 1.465849 0 0 0 1 2010 11 23 0.5981204 25963 2517 9408 0 1.630727 0 0 0 1 2010 11 24 0.78320739 5955 25963 2517 9408 1.57094 0 0 0 1 2010 11 25 0 0 5955 25963 2517 0.286667 0 0 0 1 2010 11 26 0.69627647 4109 0 5955 25963 1.871503 0 0 0 1 2010 11 27 0 0 4109 0 5955 1.668965 0 0 0 1 2010 11 28 0.94042212 3743 0 4109 0 1.799093 0 0 0 1 2010 11 29 1.232381 525 3743 0 4109 1.726693 0 0 0 1 2010 11 30 0.89526242 4306 525 3743 0 1.585769 0 0 0 1 2010 12 0 0 0 4306 525 3743 1.5 0 0 0 1 2010 12 1 1.0390434 4098 0 4306 525 2.032573 0 0 0 1 2010 12 2 0.98458187 5967 4098 0 4306 1.591363 0 0 0 1 2010 12 3 0.97574627 536 5967 4098 0 2.071846 0 0 0 1 2010 12 4 0 0 536 5967 4098 2.549304 0 0 0 1 2010 12 5 0.60008137 17206 0 536 5967 1.776294 0 0 0 1 2010 12 6 0.66622963 10678 17206 0 536 2.073499 0 0 0 1 2010 12 7 0.73735016 9427 10678 17206 0 1.811509 0 0 0 1 2010 12 8 0.92503748 667 9427 10678 17206 1.81881 0 0 0 1 2010 12 9 0.87523208 2693 667 9427 10678 1.935358 0 0 0 1 2010 12 10 0.74637681 276 2693 667 9427 1.972216 0 0 0 1 2010 12 11 0.79399142 466 276 2693 667 1.775013 0 0 0 1 2010 12 12 0.63540011 3774 466 276 2693 2.053607 0 0 0 1 2010 12 13 0.89479212 11425 3774 466 276 2.216361 0 0 0 1 2010 12 14 0.81121076 6690 11425 3774 466 2.250629 0 0 0 1 2010 12 15 0.75025679 21418 6690 11425 3774 2.69373 0 0 0 1 2010 12 16 0.96144279 5628 21418 6690 11425 2.222453 0 0 0 1 2010 12 17 0 0 5628 21418 6690 2.466818 0 0 0 1 2010 12 18 0 0 0 5628 21418 1.555986 0 0 0 1 2010 12 19 0.61129125 22513 0 0 5628 2.598241 0 0 0 1 2010 12 20 0.59110502 20416 22513 0 0 2.184294 0 0 0 1 2010 12 21 0.66507381 11719 20416 22513 0 2.344841 0 0 0 1 2010 12 22 0.61749117 2264 11719 20416 22513 2.240615 0 0 0 1 2010 12 23 0.60141218 5665 2264 11719 20416 2.073276 0 0 0 1 2010 12 24 0 0 5665 2264 11719 2.062801 0 0 0 1 2010 12 25 0 0 0 5665 2264 1.909091 0 0 0 1 2010 12 26 0 0 0 0 5665 2.825366 0 0 0 1 2010 12 27 1.1779935 309 0 0 0 2.241541 0 0 0 1 2010 12 28 0.72588978 6968 309 0 0 2.918612 0 0 0 1 2010 12 29 0.79186603 418 6968 309 0 2.375409 0 0 0 1 2010 12 30 1.2886836 433 418 6968 309 2.479616 0 0 0 1 2010 12 31 0.71930435 5750 433 418 6968 2.091722 0 0 0 1