Attachment_1-_Herring_and_Mackeral_Fishery_EM_Project_Final_Report.pdf

Herring and Mackerel Fishery Electronic Monitoring Project

Contract EA-133F-16-SE-1143

Prepared for:

U.S. Department of Commerce

National Oceanic and Atmospheric Administration

National Marine Fisheries Service

Northeast Fisheries Science Center

Woods Hole, MA 02543

Prepared By:

Morgan C. Wealti, Kathryn Carovano1

Andrew W. Jones, Glenn Chamberlain, Justin Potter 2

1 Saltwater Inc., 733 N. Street Anchorage, AK 99501 www.saltwaterinc.com

2 Northeast Fisheries Science Center, 166 Water Street, Woods Hole, MA 02543

ATTACHMENT 1

Final Report

Table of Contents

Acronyms 7

I. Executive Summary 9

Summary Overview 9

Summary of Data Findings 10

Summary of Lessons Learned 11

II. Project Background & Overview 13

The Atlantic Herring and Mackerel Fisheries 13

Electronic Monitoring (EM) 15

Project Goals & Objectives 18

III. Project Implementation 20

Project Timeline 20

Outreach 20

EM systems Installations & Field Support 21

NEFOP Data Collection 23

VMS Data Collection 24

Data Management & Review 24

Data Transfer 24

Data Security & Storage 24

Data Review 24

IV. Data Analysis/Findings 26

Project Data Overview 26

Comparison of Review Time 33

Comparison of Reviewer Annotations 36

Comparison to NEFOP Data 40

Additional Data 43

V. Cost Drivers 46

Start-Up Costs 46

Ongoing Costs 49

VI. Lessons Learned/Recommendations 53

Implementation 53

System Components 53

System Use & Reliability 53

Compliance 55

Incentives to Participate 57

Data Management & Review 58

Data Retrieval 58

Data Processing Times 59

Modified Catch Handling 60

Distinguishing Between Discarding Events 61

Methods To Address Undetected Discard Events 61

Fish Disposition Categorization For Fish Not Brought Onboard 62

Applicability To Other Greater Atlantic Fisheries 62

VII. Conclusions 64

VIII. Acknowledgements 65

IX. References 66

X. Appendices 67

Appendix 1: Discarding Event Descriptions 67

Appendix 2: EM Study Outreach Sheet 69

Appendix 3: Vessel Reference Sheet 71

Appendix 4: Trip Feedback Form 77

Appendix 5: Quarterly Summary Form 80

Appendix 6: Vessel Representative Exit Interview 81

Appendix 7: Vessel Monitoring Plan 86

Appendix 8: NMFS/NEFSC EM System Specifications 107

Appendix 9: Image Quality Assessment 113

Appendix 10: Data Release Form 114

Table of Figures

Figure 1: Graphic User Interface (GUI) as seen on EM system monitor in wheehouse 16

Figure 2: Diagram of the electronic monitoring system used in this project 17

Figure 3: Project Timeline 20

Figure 4: Boom arm installed on one of the participating vessels. 22

Figure 5: View of fish pumping before and after installation of the boom arm. 22

Figure 6: EM data displayed using Saltwater’s Open Source Review Software. 25

Figure 7: A timeline of trips included in the analysis. 27

Figure 8: Haul-level agreement throughout the course of the project. 28

Figure 9: The spatial distribution of discard records. 30

Figure 10: Breakdown of event types by vessel. 31

Figure 11: Effect of trip effort on the number of discard events. 32

Figure 12: Summary of review time for trips sampled. 34

Figure 13: Effect of trip effort on review time. 34

Figure 14: Total review time for trips with varying numbers of events. 35

Figure 15: Percent of trip length that is active fishing. 36

Figure 16: Comparison of EM reviewers’ slippage annotations. 38

Figure 17: Comparison of EM reviewers’ slippage annotations by vessel. 38

Figure 18: Comparison of EM reviewers’ operational discard annotations. 39

Figure 19: Comparison of EM reviewers’ operational discard annotations by vessel. 39

Figure 20: The distributions of estimated weights for each discard event type. 44

Figure 21: A comparison of reviewer notes of protected species interactions. 45

Figure 22: A comparison of reviewer notes of individual animal interactions. 45

Figure 23: Overall EM program costs broken down into startup costs and ongoing costs by category. 46

Figure 24: Startup costs for EM Project broken down into categories. 49

Figure 25: EM implementation ongoing costs broken down into categories. 52

Figure 26: Reasons for missing video data. 54

Figure 27: Image quality assessment. 55

Figure 28: Project Participation. 56

Figure 29: Days between trip end and receipt of trip data 59

Table of Tables

Table 1: IFM alternatives for the herring fishery 19

Table 2: Observer coverage rates. 23

Table 3: The summary of camera installation information. 26

Table 4: Summary of events recorded by each review type. 29

Table 5: Partial and full release agreement between census reviewer and NEFOP observer. 40

Table 6: Agreement among three sources of data: census EM, audit EM, and NEFOP observer. 41

Table 7: Discarding event descriptions 67

ACRONYMS

ASM At-Sea Monitor

AWS Amazon Web Service

EM Electronic Monitoring

ESA Endangered Species Act

ET Electronic Technology

FMP Fishery Management Plan

FOIA Freedom of Information Act

GARFO Greater Atlantic Regional Fisheries Office

GPS Global Positioning System

GUI Graphical User Interface

HDD Hard Disk Drive

IFM Industry Funded Monitoring

IP Internet Protocol

MADMF Massachusetts Division of Marine Fisheries

MAFMC Mid-Atlantic Fisheries Management Council

MEDMR Maine Department of Marine Resources

MMPA Marine Mammal Protection Act

NAS Network Attached Storage

NEFMC New England Fishery Management Council

NEFOP Northeast Fisheries Observer Program

NEFSC Northeast Fisheries Science Center

NMFS National Marine Fisheries Service

NFWF National Fish and Wildlife Foundation

PoE Power-over-Ethernet

RSA Research Set-Aside

SBRM Standardized Bycatch Reporting Methodology

UPS Uninterruptible Power Supply

USB Universal Serial Bus

VMP Vessel Monitoring Plan

VMS Vessel Monitoring System

VTR Vessel Trip Report

I. EXECUTIVE SUMMARY

SUMMARY OVERVIEW

Electronic Monitoring (EM) is increasingly being used as a tool for catch monitoring and reporting compliance in fisheries around the world. There are several EM initiatives and programs underway in the United States, but full program implementation in the Northeast remains limited. As part of the Greater Atlantic Region’s Electronic

Technology (ET) Implementation Plan, the New England Fishery Management Council (NEFMC) and the National

Marine Fisheries Service (NMFS) are considering implementation of EM in the Atlantic herring midwater trawl fishery to improve catch monitoring. In the Industry Funded Monitoring (IFM) omnibus amendment, the New

England Fishery Management Council recommended increased monitoring in the herring fishery to address the following goals: 1) accurate estimates of catch (retained and discarded), 2) accurate catch estimates for incidental species for which catch caps apply, and 3) affordable monitoring for the herring fishery. The IFM amendment evaluates how different coverage target alternatives meet the specific monitoring goals identified by the New

England Council while comparing the costs of the monitoring programs, particularly costs that would be borne by the fishing industry. The herring coverage target action alternatives include Northeast Fisheries Observer

Program-level (NEFOP-level) observer, at-sea monitoring (ASM), EM, and portside sampling (PS) coverage. Because midwater trawl vessels discard only a small percentage of catch at sea, EM and portside sampling have the potential to be a cost effective way to address monitoring goals for the midwater trawl vessels harvesting herring.

EM would be used to verify retention of catch on the midwater trawl fleet and portside sampling would be used to verify amount and species composition of landed catch.

Additionally, the Mid-Atlantic Fishery Management Council is considering EM as a monitoring option in the mackerel fishery pending the results of this study. While EM has been successfully deployed in other fisheries, its suitability for use in the Atlantic herring (and potentially mackerel) has not been explored. To this end, the NMFS

Northeast Fisheries Science Center (NEFSC) and the Greater Atlantic Regional Fisheries Office (GARFO) designed a project to simulate, test and refine an operational EM program.

In August 2016, the NEFSC contracted Saltwater Inc. (Saltwater) to conduct a project to determine if EM is an appropriate tool to improve monitoring and address bycatch issues in the Atlantic herring and Atlantic Mackerel midwater trawl fisheries. Specifically, the goals of this project were to inform:

● Development of EM program requirements;

● Development of a data program and EM service provider performance standards;

● The establishment of roles and responsibilities for the fishing industry, service providers, and NMFS;

● How EM data collected in this project could be integrated into other reporting requirements; and

● How information could be provided to enhance fishery-wide implementation requirements.

This information will assist NMFS and the Fishery Management Councils in the development of EM program requirements and EM performance standards. To achieve these objectives, NMFS identified the following

Contractor specific project deliverables:

● Installation and deployment of EM systems on up to twelve (12) Atlantic herring midwater trawl vessels;

● Develop local infrastructure for vessel and program support;

● Develop Vessel Monitoring Plans (VMPs) and establish standards and procedures for approving VMPs and equipment installations;

● Use EM to monitor fishing activity to determine if there are discards on herring and mackerel trips;

● Review sensor and video data; and

● Work with NMFS to review program performance for refinement.

NMFS and Saltwater staff conducted industry outreach and recruited volunteers willing to participate in the EM study during the 2016 and 2017 fishing years. Saltwater EM technicians installed systems on eleven commercial herring and mackerel midwater trawl fishing vessels in Maine, Massachusetts and New Jersey. Video and sensor data were collected for over 12 months on 192 trips and reviewed by Saltwater and NEFSC staff. Using the collected data, the project team evaluated the EM system’s ability to capture data to meet the forthcoming monitoring requirements in the herring fishery, evaluated the major drivers that could impact the costs associated with full implementation, and looked for applicability to other Northeast fisheries.

SUMMARY OF DATA FINDINGS

As a result of collaborative voluntary participation by the fleet and the diligent work of Saltwater and NMFS staff, an expansive and unique data set was collected as a part of this project.

● Data was collected on 192 trips across the 11 actively fishing midwater trawl vessels.

● These data were initially reviewed by both Saltwater and a secondary review was performed by NMFS reviewers; Saltwater staff performed a comprehensive ‘census’ review while NMFS staff performed a shorter ‘audit’ that focused exclusively on fishing events.

● ‘Dual reviews’ were successfully completed on 126 trips (i.e., both ‘census’ and ‘audit’ reviews were completed).

● Of the 126 dual reviewed trips in this study, 32 trips (25%) had overlapping Northeast Fisheries Observer

Program (NEFOP) coverage.

● Video reviewers were tasked with identifying and documenting discard events to determine what information could be consistently gathered and which types of discard events could be accurately categorized using EM. Please refer to Appendix 1 for descriptions used by reviewers to categorize discard events.

● In total, review staff performed more than 1,000 hours of video review and catalogued 1,461 discard records (902 census reviewer records, 559 audit reviewer records).

● Of the the discard events as reported by the audit review, the most frequently assigned category was

“discarded after being brought onboard,” followed by “operational discards,” “other,” “unknown,”

“partial release,” and “full release.”

● Fishing activity made up approximately 23% of trips, suggesting that a reduced portion of the total video could be reviewed in detail to detect discard events.

Following the completion of the data collection period, the project team compiled the data and performed a series of summaries and analyses. Initial results of this work suggest that video-based EM has potential to be an effective monitoring tool in this fishery.

● Census and audit EM reviewers agreed that approximately 41 slippage events (26 partial release and 15 full release) had occurred in addition to another estimated 88 operational discard events.

● There was a high level of agreement among EM reviewers in categorizing full release events (94%).

● For smaller release events reviewers were generally able to identify that a release event had occurred, but often did not use the same classification to describe the events. For partial release events reviewers agreed in approximately 55% of the cases. In cases of disagreement, one reviewer typically classified a discard event as a partial release and the other reviewer classified the event as operational discards. The comments entered by reviewers suggested that in many of these events, reviewers were viewing similar releases of catch but categorizing them differently.

● Data comparisons between EM reviewers and NEFOP observers showed general agreement in identifying and categorizing slippage events. A close comparison of these events highlights the strengths and weaknesses of each data stream.

● Agreement between reviewers (our primary metric of performance in this study) was often impacted by factors such as the total number and placement of cameras on a vessel; factors that could be better controlled in an operational program where vessels would be expected to meet required standards and protocols regarding camera set-up (EM system set-up varied by participating vessel as participation was voluntary and vessels have different layouts).

In addition to comparisons of event categorization, data collected in this project assisted with the development of recommended operational considerations to maximize the effectiveness of video-based monitoring systems in this fleet. Specifically, results provide valuable information on the average times for EM video review and potential drivers of increased review time (mainly individual annotations of discard after being brought on board events).

Further, our results suggest that an audit approach to video review may be sufficient, and may substantially reduce total review time, program costs, and storage requirements.

SUMMARY OF LESSONS LEARNED

A primary goal of this project was to determine if EM technology was a suitable monitoring option for this fishery.

Throughout the project, feedback was collected from project participants and with that input, the project team identified what worked well and where improvements were needed. Overall, EM was effective in detecting and categorizing full release slippage events when EM cameras were appropriately situated and used as recommended. Furthermore, EM was effective in detecting and categorizing catch discarded after being brought onboard. While EM was effective in the detection of discard events, reviewers had some difficulty in differentiating between operational discards and partial release slippage events consistently. Incorporating a mechanism which allows vessel operators to provide information regarding discard events throughout the trip may further aid when distinguishing among these events. The following are recommendations to promote a successful EM program in the herring midwater trawl fishery.

Implementation

System components: The EM systems provided by Saltwater functioned reliably and captured high quality data that allowed reviewers to identify discard events. Unnecessary recording occurred when the vessels engaged in non-fishing activity at the dock that incorporated the vessel hydraulics and initiated camera recording. The incorporation of using geofencing technology to restrict the onset and completion of video recording eliminated these unintended recordings and should be required in an operational program.

System use and reliability: Power interruptions to the EM system caused incidences of data loss. The use of voltage conditioners and uninterruptible power supplies (UPS) decreased the risk of power loss to the EM system.

Camera connectivity issues that occurred were due to high vibration on the rail mounted cameras. Vibration resistant cameras are recommended for boom mounted cameras in this fishery.

Compliance: The project had lower participation in the last quarter for a variety of reasons that are addressed later in this report, but ultimately were a result of the voluntary structure of the study. In an operational program, vessels would be required to operate their systems or would be subject to consequence measures. Another common issue we encountered was a lack of proper training for the vessel personnel responsible for operating the EM system. The vessel representative trained by Saltwater during the install was often the vessel owner or fleet manager, not the captain. For this reason, captains and crews did not always fully understand their responsibilities.

Under full implementation, the captain should be present during the install for operator training.

To maximize the ability of EM reviewers to view all discards, we determined that cameras should be installed to capture all 4 possible discard locations as listed below;

● Fish pumping

● Dewatering box

● Full deck

● Stern

These four views can generally be captured by three properly placed cameras. On most vessels, getting the required views will require the installation of a boom arm mount (as described in section III below).

Data Management and Review:

Data review: It is important for all project partners to work together at the onset of any EM project to determine which data fields should be collected and how they should be reported. Doing this early ensures the EM systems are installed with the best possible configuration to collect the necessary information and that data is properly documented in the review process. All events of interest should be clearly defined to prevent variation in the classification of discard events among reviewers. Data reviewers should be trained to ensure categorization of events and species identification is standardized.

Data retrieval: In fisheries with complex logistics where the vessels are not all located in the same port, in person data retrieval can be costly and logistically complicated. Mailing EM data to the review center can simplify this process and result in cost savings. However, mailing the data diminishes the opportunity for face to face contact, which allows vessel operators to ask questions, build working relationships with technicians and facilitates advantageous system performance checks. This issue may be mitigated by more frequent communication with the vessel operators early in the project (after the first few trips) to ensure EM responsibilities are understood and data collection is optimized. In an operational EM program with required compliance, vessel operators would be required to perform a “system check” prior to each trip, and ensure that any issues with the system are reported immediately to the EM service provider. In a fleet that makes frequent, short trips and is somewhat migratory, sufficient spare hard drives should be made available to the vessels to ensure data collection is not hindered due to HDD resource limitations.

II. PROJECT BACKGROUND & OVERVIEW

Saltwater Inc. is an observer and EM service provider headquartered in Anchorage, Alaska with field offices in

Massachusetts, Washington, and Kodiak, Alaska. Saltwater has collected high quality data on fisheries and oceans for government agencies, research organizations, and fishermen for nearly 30 years. Since 2010, Saltwater has provided electronic monitoring services on vessels in the Atlantic pelagic longline fishery, the North Pacific longline and pot cod fisheries, the Pacific/West Coast groundfish fishery, the Pacific Islands shallow and deep set longline fishery, and the Gulf of Mexico reef and shrimp fisheries. The EM systems in these fisheries consistently produce high quality digital imagery integrated with precise Global Positioning System (GPS) and sensor data. These projects have allowed Saltwater to engage with fishery managers, the fishing industry, and other stakeholders to identify key challenges and define the best means of implementing EM in each fishery. This includes hardware and software decisions, data retrieval and management planning, data review protocols and procedures, skipper engagement and training, and infrastructure development.

THE ATLANTIC HERRING AND MACKEREL FISHERIES

Atlantic herring (Clupea harengus) are a schooling fish distributed throughout the North Atlantic and Arctic Oceans, including the eastern seaboard of North America, where they migrate between Canada and North Carolina to feed and spawn . These herring are a slow-growing species, generally reaching maturity at age 3 (~10 in) and attaining lengths of 15 inches . Herring are an important forage species for tuna, sharks, haddock, flounder, squid, and marine mammals . They are commercially valuable as bait fish, for fish oil, fish meal, and for human consumption.

The fishery is managed through a stock-wide annual catch limit allocated to four separate management areas overseen by NMFS and the NEFMC (for Federal waters), and the Atlantic States Marine Fisheries Commission and individual states (for coastal waters). Atlantic herring are caught with a variety of gear types, including trawl, purse seines, and gillnets, with midwater trawls (paired and single) accounting for the majority of the catch, 35,074 metric tons of herring from all areas, landed in 2016. This amount includes Research Set-Aside (RSA) quota.

Atlantic mackerel (Scomber scombrus) have a similar distribution and life history as Atlantic herring, and the two are often caught in conjunction. Mackerel grow to lengths of up to 16 inches; like herring, they feed on zooplankton and small crustaceans, and are an important forage species for other animals. There is a recreational fishery for mackerel, and their stock-wide annual catch limit is divided between the commercial and recreational fishery, managed entirely by NMFS and the Mid-Atlantic Fisheries Management Council (MAFMC) . Mackerel are commercially fished with a wide variety of gears, such as handline, longline, purse seine, pot/trap, gillnets, and trawls, with midwater trawls accounting for the majority of the catch.

Under the Northeast Fisheries Observer Program (NEFOP), when selected by NMFS, vessels operating in the herring and mackerel fisheries must carry at-sea observers who document catch and discards, economic information, gear characteristics, fishing location, and biologically sample the catch. . The herring fishery is not currently characterized as overfished, or as experiencing overfishing , but stakeholders have expressed concerns with bycatch and interactions with marine mammals. Given the the findings of the recent Atlantic Mackerel

1 http://s3.amazonaws.com/nefmc.org/herring_FMP.PDF 2 https://www.nefsc.noaa.gov/publications/tm/tm126/tm126.pdf 3 https://www.mass.gov/service-details/learn-about-atlantic-herring 4 https://www.greateratlantic.fisheries.noaa.gov/sustainable/species/atlherring/index.html 5 https://www.fisheries.noaa.gov/species/atlantic-mackerel 6 https://www.greateratlantic.fisheries.noaa.gov/regs/2015/June/15sbrmomnibusamendea.pdf, pg. 125, 4.5.1.1 Federal Observer Program 7 https://www.fisheries.noaa.gov/species/atlantic-herring assessment (available at https://www.nefsc.noaa.gov/saw/reports.html), its official status will soon change to overfished with overfishing occurring. Plans are underway to implement a rebuilding program commencing

January 2019 (pers com Jason Didden, MAFMC staff). The Standardized Bycatch Reporting Methodology (SBRM) sets the rate at which each fishery is covered; the actual rate depends on the amount of variability in the observer data used to complete the analysis. As a result, the coverage rate changes from year to year. The sea day schedule is released to the public in the spring each year.

Paired and single midwater trawling for Atlantic herring and mackerel are characterized as High Volume Fisheries, defined by large catch of many small fish, which are typically brought onboard using a high powered vacuum pump. Catches flow at a rate of 3,600 - 8,900pounds per minute (final loaded NEFOP data, 2010-2017, midwater trawl vessels only; flow rate estimate based on observed pump times and weights by permit) directly through a series of chutes into the vessel’s hold, which contains refrigerated sea water. Currently monitored bycatch includes haddock, and river herring (alewife and blueback herring), and shads.

There are 12 vessels that actively fish for herring and mackerel with midwater trawl gear, and the vessels fish out of 5 main ports: Rockland, ME; Portland, ME; Gloucester, MA; New Bedford, MA; and Cape May, NJ. The vessels in these fisheries are usually large (80-150 feet or more in length), and in roughly half the fleet, two vessels fish together, which is referred to as “pair trawling”. Trips typically last 3-5 days, though much of that time may be spent searching for fish via electronic sounders. The vessels rely heavily on their sonar systems (sounders, fish finders) to locate the targeted schools of fish.

Paired midwater trawling is only possible in relatively good weather because the paired vessels must maintain a uniform distance while towing and hauling back . Setting begins with one vessel putting the net over the stern. The second vessel then approaches, pulling alongside its sister ship to retrieve a line attached to the net bridle from the vessel setting the net. The receiving vessel will then attach the bridle to their steel wire and, at the specified signal, both vessels begin to pay out a certain amount of wire in unison. Communications throughout the operation are maintained over VHF radio. Once the wire is paid out, towing begins, with the vessels on parallel courses and about one half to one third the warp length apart. Depth can be modified by increasing or decreasing wire length and towing speed.

One of the challenges to fishing with two vessels is that they need to be of similar size and power, and the captains must work in close cooperation for successful fishing to occur. Sensors are used to ensure the gear is fishing correctly and to monitor the catch in the net. Haul back begins at a given signal, with the warps being pulled in until the legs are brought up to the vessels. The vessels come alongside each other once more, when one of the vessels releases its cable and throws the line attached to the net bridle back to the hauling vessel, and the net is brought alongside at the surface to have the catch pumped onboard, into refrigerated saltwater tanks.

A primary objective of the study was to determine if EM technology could monitor catch retention in the midwater trawl fishery. The NEFOP provides at-sea coverage in the high volume fisheries and classifies discards into three categories: 1) Catch discarded after being brought on board 2) Slippage, referred to as the partial or full release of catch 3) Operational Discards. This classification structure was utilized by video reviewers throughout the EM project. Once a discard event was identified, reviewers selected one of five options to document the discard event type. These discard event types are defined in Appendix 1.

8 Thomson, D. 1978 “Pair trawling and pair seining; the technology of two-boat fishing.” Fishing New Books. 0-85238-087-9

ELECTRONIC MONITORING (EM)

The use of EM systems on fishing vessels is an increasingly popular addition to the monitoring portfolio of fishery management tools. The systems are comprised of video cameras, sensors, and data collection software. Data generated by the EM system includes vessel location, speed, gear deployment logs, and video of fishing activity, but does not collect audio. The data can augment or replace human observers to meet certain monitoring goals.

EM has advanced significantly in the past 10 years, making high-quality, consistent data capture a cost-effective option to at-sea observers in some, if not all, fisheries. The following is a more detailed description of the EM system components:

Digital Cameras: Saltwater’s internet protocol (IP) digital cameras have an ingress protection rating of IP67, which means they are manufactured with housings designed to protect against water immersion up to 1 meter and up to

30 minutes. Each camera is capable of capturing high-resolution imagery (1920 x 1080) at a rate of up to 30 frames per second. The cameras were selected for their low lux rating, ensuring clear imagery in low light, and use wide dynamic range technology to better handle fluctuating lighting conditions. Saltwater has deployed these cameras on vessels from Alaska to Hawaii, fishing in all kinds of challenging conditions. For this project, video resolution was set to 720p (1280x720) at a 15 fps, which is the recommended setting for smooth video playback.

GPS Receiver: Saltwater’s GPS receiver is integrated within the control box which is coupled with an external antenna. It provides heading, velocity, latitude, longitude, and time/date. It begins producing data the moment it receives power and records continuously as long as there is power to the EM system. For this project, GPS data was logged every 10 seconds, though the capture rate can easily be adjusted to as often as every second (1 Hz) or as seldom as every hour. This information is used during data review to track vessel activity and to identify fishing effort.

Hydraulic & Drum Rotation Sensors: Hydraulic pressure transducers and rotation sensors are central to Saltwater’s

EM systems. Hydraulic pressure transducers monitor a vessel’s hydraulic pressure and rotation sensors monitor mechanics, like line drums or warp winches. Sensor status is logged continuously from the start to the end of each fishing trip. This information helps reviewers identify fishing activity during the review process. The sensors can also be used to trigger video recording.

Control Center: Saltwater’s small footprint EM control center (12” x 7.5” x 4.5”) is fanless, noiseless, and ruggedized to withstand a wide temperature range, as well as shock and vibration. It includes a built in GPS receiver and two hard drive bays. All data collected by the EM system is recorded to high capacity (1 TB) hard disk drives (HDDs). The system is configured to write all information on encrypted HDDs to ensure data security and confidentiality. The system software monitors storage capacity and when one drive fills up, the system automatically stores additional data on the second drive.

The control center has eight built-in Power-over-Ethernet (PoE) ports for IP camera connectivity, which eliminates the need for additional power delivery to an external switch. Universal serial bus (USB) ports allow for simple system software upgrades via flash drive and the opportunity for the integration of other electronic devices. The control box and all of Saltwater’s EM system components have low power requirements and can run off either alternating current (AC) or direct current (DC) power.

Wheelhouse Monitor: A wheelhouse monitor with waterproof keyboard allows vessel crew to see current date, time and location data, live camera feeds that show camera views at all times, even when the cameras are not recording, camera status (e.g., recording or not), and view what is being recorded in real-time. Our graphical user interface (GUI) is designed to provide vessel operators with a clear, simple way to confirm that all system components are operating correctly, the presence of a functioning HDD, and the percentage of disk space used.

The GUI also allows viewers to determine which HDD has data--and how much--and which drive to send in at the end of a trip.

Figure 1: Graphic User Interface (GUI) as seen on the Saltwater EM system wheelhouse monitor .

Figure 2: Diagram of the Saltwater EM system used in this project.

EM has been suggested to monitor gear interactions with protected species, compliance with discard and retention regulations, to account for catch and bycatch, and to validate vessel landing and logbook information. A key constraint to effective EM implementation in many fisheries is the cost of data review and storage. Operational implementation of EM requires not only collecting hours of video and sensor data, but also the ability to efficiently extract from that data, the meaningful information needed to manage a particular fishery. Saltwater has developed open-source review software that integrates video and sensor data for efficient data review and analysis. Open-source software avoids the limitations and expense associated with proprietary code, encourages collaboration and innovation, and will speed the development of cost effective review solutions. A key constraint to successful EM program implementation is the cost of data review. The use of open-source data review software will be critical to the long-term success and sustainability of EM programs. Open source software can be used, changed, and shared (in modified or unmodified form) by anyone. The open-source movement promotes collaborative development of computer source code by multiple independent developers. It is among the most transformative trends shaping technology in the 21st century and companies such as Google, Oracle, IBM, LinkedIn, Square, Twitter, Netflix and others already rely on it.

PROJECT GOALS & OBJECTIVES

The overarching objectives were to evaluate the utility of EM for monitoring catch retention and to detect discard events in the Atlantic herring and mackerel midwater trawl fisheries according to the proposed use of EM in the

IFM (as component of a monitoring option that would complement a portside sampling program). To assess the potential of EM to serve as a means for ensuring catch retention this project sought to 1) Compare annotations associated with discard event events among EM reviewers to quantify how consistently events can be detected by

EM, 2) Compare the review times among review types to determine the review time needed, 3) Compare EM discard event detections to those noted by a NEFOP observer to assess the relative efficacy of the technology relative to existing monitoring options, and 4) Refine industry and NMFS EM cost estimates by identifying the variables that impacted the total project cost. The design of the program involved installing EM systems on volunteer vessels actively participating in these fisheries, collecting the video data after each trip, and reviewing

100% of the collected video data. The project team sought to determine if the collected data met specific monitoring needs, what factors were critical to successful program implementation, whether there were ways to improve any aspect of the program, and to identify the primary cost drivers and any cost efficiencies.

The NEFMC developed the Industry Funded Monitoring (IFM) Omnibus Amendment to increase monitoring and/or other types of data collection in some of its Fishery Management Plans (FMP). The goal of the IFM Amendment is to improve the amount and type of catch data, to more precisely monitor annual catch limits, and/or provide other information for management. This increased monitoring will be in addition to coverage required through SBRM, the Endangered Species Act (ESA) or Marine Mammal Protection Act (MMPA). The IFM Amendment specifically addresses increased monitoring in the Atlantic Herring FMP, and contains alternatives that maintain or increase observer coverage in the herring fishery. Although the IFM Amendment was originally intended to be a joint venture between the NEFMC and the MAFMC, the MAFMC has decided to delay any decision on an industry-funded monitoring action until the results of this study are complete.

At its April 2017 meeting, the NEFMC selected IFM Herring Alternative 2.7, which will implement a combined coverage target of 50% for all herring Category A and B vessels, and will provide herring midwater trawl vessels the option to choose between monitoring with At-Sea Monitors (ASM) or with the combination of EM and Portside monitoring. Final approval of EM as a monitoring option will be made at its April, 2018 meeting, and will be based largely upon results and recommendations of this study.

If approved by the New England Council, midwater trawl vessels could choose EM in association with portside monitoring. Since 2008, the Massachusetts Division of Marine Fisheries (MADMF) and the Maine Department of

Marine Resources (MEDMR) have collected data (species, morphometrics) on the herring and mackerel midwater trawl fishery through independent portside data sampling programs. MADMF has opportunistically targeted 50% sampling of midwater trawl trips landed in MA, while MEDMR generally samples 5-10% of trips landed in ME.

Though these programs are not currently used in NMFS quota monitoring, they have provided valuable and expedited information on catch composition to fishermen and fishery managers. In conjunction with EM, these states will continue to collect data through the first year of IFM implementation in 2018, after which a federal program (modeled after the state programs) will commence in 2019.

Table 1: IFM alternatives for the herring fishery.

Gear Type Midwater Trawl Purse Seine Small Mesh

Bottom Trawl

Herring Alternative 1: No Coverage Target for

IFM Program (No Action)

SBRM

Herring Alternative 2: Coverage Targets for IFM

Program

Includes Sub-Options: 1) Waiver Allowed, 2) Wing Vessel

Exemption, 3) 2-Year Sunset, 4) 2-Year Re-evaluation, and

5) 25 mt or 50 mt Exemption Threshold

Herring Alternative 2.1: 100% NEFOP-Level

Coverage on Category A and B Vessels

100% NEFOP-Level Observer

Herring Alternative 2.2: ASM Coverage on

Category A and B Vessels

25%, 50%, 75% or 100% ASM

Herring Alternative 2.3: Combination Coverage on Category A and B Vessels and Midwater

Trawl Fleet

50% or 100%

EM/Portside

25%, 50%, 75% or 100% ASM

Herring Alternative 2.4: EM and Portside

Coverage on Midwater Trawl Fleet

50% or 100%

EM/Portside

SBRM (No Action)

Herring Alternative 2.5: 100% NEFOP-Level

Coverage on Midwater Trawl Fleet in

Groundfish Closed Areas*

100%

NEFOP-Level

Coverage

SBRM (No Action)

Herring Alternative 2.6: Combination Coverage on Midwater Trawl Fleet in Groundfish Closed

Areas

Coverage would match selected alternative 2.1-2.4

SBRM (No Action)

Herring Alternative 2.7: ASM Coverage on

Category A and B Vessels, then Vessels may choose either ASM or EM/Portside Coverage

50% ASM or

EM/Portside

50% ASM 50% ASM

* Sub-Options do not apply to Herring Alternative 2.5.

III. PROJECT IMPLEMENTATION

PROJECT TIMELINE

The contract was awarded to Saltwater and signed on August 12, 2016. The timeline included a 2 month project setup period which included clarifying project objectives, hardware and software specifications, vessel requirements, defining roles and responsibilities, and development of a project plan. In addition, Saltwater setup local capacity and infrastructure to provide and support field services. The EM operational and data collection period consisted of 13 months (October 2016 - October 2017). Activities included; vessel visits, installation and servicing of EM units, field and project support and data collection. Analysis and report writing occurred over 3 months.

Figure 3: Project Timeline

OUTREACH

Beginning in 2016, NMFS conducted initial outreach, alerting the herring and mackerel midwater trawl fleet vessel owners that a voluntary EM project had been funded and would be taking place in 2016-17. Vessel owners were informed about the objectives of the project, the timeline, and their roles and responsibilities if they chose to volunteer. NMFS created an EM study outreach sheet (Appendix 2) to explain the project goals. Participation in the project was not remunerated and did not remove vessel owners from the requirement to carry observers on selected trips. Saltwater prepared a Vessel Reference Sheet (Appendix 3), outlining vessel power requirements and basic EM system operations that was provided to potential volunteer participants, along with a detailed description of the install process.

In order to prepare for the installations, an in-person vessel assessment was conducted on all but one vessel in

October 2016. In order to make the best use of the limited time in port, the remaining vessel assessment took place in November, on the first day of the EM system installation on that vessel. During the vessel assessment meeting, members of NMFS and Saltwater met with the vessel owner and/or captain to create an EM installation plan. Saltwater worked with the owner or captain to determine where each of the components should be installed for optimal function. Project goals and objectives were also discussed at this meeting.

Throughout the project, vessel feedback was provided in a variety of formats. Saltwater staff contacted vessel owners or operators if there was something that needed more immediate feedback, such as keeping the EM system powered for the entirety of the trip. Video reviewers also filled out trip feedback forms (Appendix 4) to effectively address issues on a particular trip. Quarterly evaluations summarizing all declared herring/mackerel trips and several metrics evaluating adherence to EM project requirements were sent to the vessel owners

(Appendix 5). Throughout the project, representatives from NMFS and Saltwater met with vessel owners and operators to review the quarterly evaluations and trip feedback forms as well as go over video footage from that quarter and address any questions or concerns the owners or operators had about the project. At the end of the project, NMFS and Saltwater conducted exit interviews (Appendix 6) with representatives from each vessel to discuss the project and gather feedback. A total of 37 outreach meetings took place during the course of the project, including 10 vessel assessments/pre-install meetings, 17 quarterly feedback meetings and 10 exit interviews.

EM SYSTEMS INSTALLATIONS & FIELD SUPPORT

Saltwater developed an EM system that has been successfully deployed in the past six years on well over 150 vessels in multiple fisheries, including over 100 vessels across the Atlantic Coast fishing in the Pelagic Longline fleet for Tuna and Swordfish. This EM system used for this project has proved capable of collecting high-quality digital video data and accurate, supporting sensor data. Saltwater’s EM system consists of a control box with two bays for removable hard drives, one to three Internet protocol (IP) cameras per vessel, hydraulic pressure sensors, drum rotation sensors, a GPS receiver, and a monitor and keyboard for the vessel’s wheelhouse. Saltwater’s data acquisition software utilizes open-source code to log and process GPS, sensor, and video data into usable EM datasets. All of the EM systems were leased by NMFS for the duration of project. Industry participants were not required to pay for any of the components or service costs, but were given the option to purchase the systems for a nominal fee at the end of the project.

During the initial vessel visits, Saltwater and NMFS discussed the installation with the vessel representative. After discussing fishing operations, NMFS and Saltwater communicated the preferred location for the cameras to capture the ideal views. Typically, these vessels pull the net along the starboard side of the vessel to pump the fish and the catch is pumped through the dewatering box where discarding may occur. Once pumping is complete, the net is pulled back around to the stern where the codend is opened up and the net is brought back onboard on the net reel. With these fishing operations, there are three locations where discards could occur; at the pumping location, at the dewatering box, and at the stern. It was determined that cameras should be installed in a way that captures all 3 of these possible discard locations, usually requiring 3 cameras. A full deck view camera (often used to capture a view of the dewatering box) also proved helpful to allow the reviewer to be able to see when the vessel was setting out the net and hauling it back. There was one vessel in the project that pumped at the stern and therefore did not require a starboard camera.

The typical installation took 3 - 3.5 days for a two technician team (or roughly 60 man hours). Much of this time was spent running wires since many wire runs were behind wall and ceiling panels and were difficult to access.

After reviewing the initial datasets from each vessel, it was determined that a camera mounted on a boom could provide a direct line of sight of the pump station, starboard rail, and water while eliminating physical obstructions unique to each vessel participant. It was determined that most pump views in this project, could be improved by the installation of a boom arm extending over the edge of the vessel, allowing the camera an unobstructed view of the pump in the water. Booms were installed on five vessels, however, all but one vessel would have benefited from the installation of a boom. In some situations, captains declined to install a boom because they felt that it would have been an obstruction during the offload process, and others felt the view without the boom should be sufficient for the project. Dewatering box/deck camera views were seen by some participants as unnecessary or intrusive to personal space for a project looking at documenting discard events, so those participants declined to have those views captured on their vessels. It was determined during the project that a deck view, dewatering box view, stern view and pump view were all required to accurately classify discard events. However, since this was a voluntary project, booms and cameras were installed to capture the best views possible while adhering to the vessel owner’s guidance and level of comfort.

Figure 4: Boom arm installed on one of the participating vessels. Boom arms were added to 5 vessels in this project to determine if a better view of pumping activity and the contents of the net could be obtained. Photograph used with vessel owner’s permission.

Figure 5: View of fish pumping before and after installation of the boom arm. Photograph used with vessel owner’s permission.

When the installation was complete, the vessel owner or operator was trained by the EM technician in system operation and maintenance. A system operating manual was provided at the time of install and a VMP (Appendix

7) was created and provided to the vessel operator once complete.

Saltwater established a call-in number and service tracking system that was used to meet the support needs of the

11 vessels participating in the project. During the system installs, Saltwater technicians provided vessel owners and/or operators the call-in number, available twenty-four hours a day, seven days a week, and appropriate email contact information. Trained Saltwater EM staff answered all calls and were able to carry out remote troubleshooting and help identify and/or resolve system issues. The majority of calls were answered by the same technicians who carried out the installs, had developed working relationships with the captains, and were familiar with the set up on the boats.

NEFOP DATA COLLECTION

The Northeast Fisheries Observer Program (NEFOP) collects fishery dependent data from the midwater trawl herring fleet. Observers collect a suite of data on herring trips, including: trip-level information (costs, vessel description, safety), gear, catch (kept and discarded), and data to assist NEFOP in the determination of slippage events, and incidental take information on protected species. For a complete description of all data collected by observers, refer to the observer program operations manual, data entry manual, and on-deck reference guide . 9

NEFOP data are used by GARFO to monitor catch caps for river herring/shad and haddock. The NEFSC uses NEFOP data in support of SBRM to estimate the amount of bycatch occurring in the fishery. The SBRM analysis is described in detail on the observer program website . 10

The SBRM sets the rate at which the fishery is covered; the actual rate is determined based on the amount of variability in the observer data used to complete the analysis. As a result, the coverage rate changes from year to year. The sea day schedule is released to the public in the spring each year. The table below demonstrates the observer coverage over the time-frame of the project.

Table 2. Midwater Trawl Quarterly Observer Coverage Rates, Nov. 2016-Oct. 2017. Includes observed (at least one observed haul per trip) single and paired midwater trawl trips divided by VTR trips reporting kept catch. Includes all fisheries, not just herring and mackerel fisheries. The herring EM project data collection period overlapped with two SBRM years, 2016-2017 and 2017-2018 (SBRM schedule runs from April-March). Observer coverage was relatively high Feb-Apr during the study as this aligned with increased SBRM coverage needs during this same period. SBRM coverage of the fleet can vary from year to year, driven by funding and data needs to monitor bycatch (Source:

GARFO’s DMIS and NEFSC’s OBDBS databases as of 2018-01-24).

Period Observer Coverage

NOV-JAN 14.60%

FEB-APR 31.30%

MAY-JUL 13.50%

AUG-OCT 11.00%

9 https://www.nefsc.noaa.gov/fsb/training/ 10 https://www.nefsc.noaa.gov/fsb/SBRM/

VMS DATA COLLECTION

All participants in the Atlantic herring fleet are required to carry a vessel monitoring system (VMS) for use by NMFS to track vessel activity as a condition of their federal fishing permit. VMS data were used during the project to aid in data retrieval. NMFS staff were able to verify vessel activity and communicate expected landing dates to EM technicians. This relay of information allowed the EM provider to plan ahead for data retrievals and system maintenance.