Projects in the Field is a series of independently produced articles profiling work supported by NFWF’s Electronic Monitoring & Reporting Grant Program, and is meant to raise awareness and support for these important initiatives. As always, your questions and comments are welcome.
In 2018 electronic monitoring (EM) was fully implemented in Alaska’s pot cod fishery, and fishermen, for the first time, could choose between carrying an EM system or an onboard observer to collect the data required to manage the fishery. This is the story of how industry got the EM program they wanted, and the National Marine Fisheries Service (NMFS) got the data they needed.
The North Pacific Fisheries Association (NPFA) is a fishermen’s organization that represents the interests of salmon, crab, and groundfish fishermen. Buck Laukitis and Malcolm Milne are former and current Presidents of NPFA. Both are active fishermen, and both thought Alaska’s pot fishery for Pacific cod could be monitored effectively with EM because of the way fish are brought onboard and the limited amount of bycatch. They also thought that EM might provide an alternative to the inherent liabilities, costs, and logistics of carrying onboard observers.
Making EM “work” in a new fishery is rarely easy. Fishermen are often reluctant to volunteer because of concerns about “cameras in our living rooms” and the confidentiality of data. Fishery managers are often not sure EM can produce the data they need. In the case of pot cod, all parties were concerned about whether discards could be identified, what the onboard handling procedures would need to be, and the care, maintenance, reliability of the EM systems.
NPFA partnered with Saltwater Inc., a company that provides both EM and observer services, to explore the feasibility of using EM in the pot cod fishery. 
With support from the National Fish and Wildlife Foundation (NFWF), NPFA-Saltwater deployed EM systems on six pot cod vessels from 2014-2016. Fishermen and Saltwater technicians collaboratively tested camera angles, handling protocols, and camera types and positioning to improve the consistency of image quality. Based on the data collected by onboard observers, Saltwater developed a list of key data requirements that would be needed from the fishery to meet NMFS’ and the North Pacific Fishery Management Council’s (NPFMC) catch accounting goals.
The results of the feasibility studies confirmed fishermen’s hypothesis that the pot cod fishery was well suited to EM. Saltwater EM systems collected high quality digital imagery and sensor data on 13,098 pot hauls, providing an accurate record of fishing effort and catch composition. Saltwater data reviewers were able to identify 99.6% of the 55, 212 catch items to a species or species group level.
Encouraged by the success of the feasibility studies NPFA-Saltwater wrote a plan to further test EM during 2017 in a pre-implementation phase with the goal of full implementation in 2018. The plan included a suggested sampling design, data collection fields, and described protocols for catch handling, data retrieval, and EM data review. The plan was reviewed and approved by NMFS and the North Pacific Fishery Management Council (NPFMC), and formed the basis of a proposal that received funding from NFWF. The plan was successfully tested in 2017, and is now part of full implementation for Alaska’s fixed gear fleet.
What Strategies were Tested?
The plan focused on two key objectives: providing high quality data and cost efficiency. To achieve those goals NPFA-Saltwater tested a distinct service delivery model that included:
Skippers responsible for mailing data
To avoid the cost of in-port coordinators or electronic transmission, skippers are responsible for retrieving and mailing the data from their boats. All of the EM data stored on the onboard hard drives (HDDs) is encrypted and password protected, and the “system log” records the serial number of the HDDs that are in the system, and the dates of install and removal. This enables the establishment of a chain of custody, and documents that the HDD installed is the one removed and the one received. The practice worked well – skippers were timely in sending in their data, the process was cost effective, and the model has now been adopted in other parts of the EM program for fixed gear in Alaska.
Integrating experienced NMFS-trained observers into the EM program
The EM data was reviewed by trained Saltwater data reviewers, all of whom are current or prior NMFS fishery observers who have been trained in sampling design, fish identification, and who have hands-on experience
collecting data onboard commercial fishing boats. At the request of Saltwater, NMFS approved the ability of qualified NMFS observers to rotate between reviewing EM data onshore and observing at sea. Integrating observers into the review process has promoted consistency between the EM and observer data streams, and observer insights have increased the quality of the review process. It has also provided increased and varied work opportunities for observers, which encourages their retention in the overall monitoring program. Having observers review data provides excellent review talent, and also allows for cost efficiencies as the workforce can fluctuate with seasonal needs. In a time of declining budgets and increased demand for data, this model promotes cost-effective use of observer time and talent.
Timely Data Review and Feedback Memos
The project team believed that timely data review, and a tight feedback loop between the boats and the data would contribute to data quality. After the review of each trip, Saltwater emailed a feedback memo to the vessel and NMFS that described the data from that particular trip. Many issues that can compromise data quality can only be recognized once a reviewer looks at the EM video. The feedback memos included assessments of data quality, photos where appropriate, and notes on problems affecting data quality that needed to be addressed (e.g., slime on lens, crew blocking view of camera, etc.). Generally, skippers appreciated knowing what data was coming from their boats, welcomed the photos, and responded well to comments about areas that needed improvement (better cleaning of the sorting table, discarding in view of the camera). This tight feedback loop between the boats and the data meant that issues that can affect data quality – whether an EM system issue or a catch handling problem – could be quickly resolved before more problematic data is collected.
Use of Open Source Software
Saltwater used the open source EM review software developed by Saltwater/Chordata/Sea State with support from NFWF. Using the EM open source software saves on the licensing costs and limitations common to proprietary software, and allows the incorporation of other evolving open source developments. The EM review software had been used in multiple NMFS’ EM programs throughout the U.S., but this was the first time it was used at a production scale. It worked well and proved itself adaptable to ongoing changes and refinements to the data requirements and review protocols.
Data Transmission to NMFS
A critical piece of the infrastructure in an EM program is getting reviewed EM data to fishery managers in a format they can use. Early in the project Saltwater worked closely with NMFS on issues related to the uptake of EM data into NMFS’ databases and ultimately the catch accounting system. Saltwater submitted to NMFS a draft “Data Design” with suggested data fields and formats to allow integration into NMFS’ observer program databases. NMFS completed a “final” database schema, and Saltwater developed an in-house database that mirrored the NMFS database.
NMFS and Saltwater collaboratively developed protocols that allow NMFS to pick up data from the Saltwater database at NMFS’ convenience. The development of the database and “data pipeline” has increased the capacity for transmission of EM data in Alaska and could be an important part of the infrastructure for full-scale EM implementation.
Saltwater managers/reviewers also used the in-house database to conduct quality checks on the data during the season. Similar to onboard observer “debriefings,” Saltwater developed and ran a series of queries on the reviewed data to check for mistakes. Did fish names match with species codes? Were anomalous fish or rare event species correctly identified? Was the sampling protocol followed correctly? Did descriptions of data gaps match with the sensor logs? These quality assurance checks allowed Saltwater to spot and, as appropriate, correct anomalies in reviewed data prior to submission to NMFS.
What about Costs?
Because industry pays for the costs of observers in the North Pacific, there is strong interest in determining whether EM can provide a less expensive alternative to onboard observers. Boats under 60 feet length overall in the North Pacific are part of the partial coverage (under 100% coverage) observer program and pay a fee on landings to support the program. In the partial coverage program, costs have been about $1,000 per observed day over the last several years and are projected to be approximately $1,400 per day in 2018.
In order to compare observer costs with EM costs, the NPFMC’s EM work group established a structure for estimating the cost of an ongoing EM program. The structure distinguishes between one-time, amortized, and recurring costs. One-time costs that occur at the beginning of a project include start-up planning, fleet outreach, the development of data review protocols, and database design. Amortized costs are largely the cost of installed EM equipment that is amortized over a five-year period, recognizing that the actual equipment life may be longer. Recurring costs include ongoing project coordination, equipment services, technical support, and data review. The approach results in an Adjusted Annual Cost that includes the recurring costs and amortized costs as a way to project the costs for an ongoing EM program.
The Adjusted Annual Cost for the EM program in the pot cod fishery in 2017 is estimated to be approximately $455 per day of monitoring including data review, and is expected to be similar in 2018. This is less than half of the cost of an onboard observer day through the partial coverage program.
It should be noted that this projection is sensitive to the number of fishing days selected for monitoring. Monitoring of the partial coverage sector is based on a trip selection system with a current goal of 30% of fishing trips monitored. If fishery managers decided to increase the number of days selected for monitoring from the current rule of 30% of selected trips to 50% of selected trips, the cost per day would decrease. Likewise, if fishing effort dropped and resulted in fewer fishing days, the projected ongoing cost per day would increase.
Lessons Learned
One: The coordinated efforts and support of industry, NFWF, the Council, and NMFS have driven the successful adoption of EM in Alaska’s pot cod fleet.
It was Alaskan fishermen who originally identified the potential of EM to 
effectively monitor the pot cod fishery. A strong partnership between industry and Saltwater, the EM service provider, resulted in the implementation design that was tested and proved successful in meeting the twin goals of high-quality data and cost efficiency. The NPFMC provided leadership and a forum for all parties to collaborate and cooperate. NMFS collaborated on the development of key data fields, review protocols, database structure, and data transmission. NFWF provided the funding that made the projects possible.
Two: Direct skipper engagement with both EM equipment and data can save money and contribute to data quality.
Saltwater technicians provided skippers with extensive training on the operation and maintenance of their EM system during the install. This training, the relationship with a technician, and the tech support “hotline” allowed for a strong system of remote support. Skippers or crew were able to resolve many EM system problems with help from a Saltwater technician by phone. Over the course of the project, the vast majority of system problems were resolved remotely, with very few requiring an in-port service call. In Alaska, where unpredictable weather and distance make transportation expensive, a strong system of skipper training and remote support are key to a cost-efficient program.
Engaging skippers with the onboard EM system was significant, but just as important was engaging skippers and crew with the data from their boat. To stimulate this engagement, NPFA-Saltwater created a very tight feedback loop between individual vessels and the overall EM program. Ongoing conversations throughout the early implementation phase covered everything from the EM equipment to handling procedures to the quality of data coming from individual boats. Engaging skippers and crew with both the EM systems and the data via timely feedback memos were key aspects of the program design, and contribute to both data quality and cost effectiveness.
Skippers reported some benefits from EM that they did not expect including:
- greater flexibility to schedule trips, including the ability to make quick turnarounds after offloads in port without having to wait for an observer; and
- improved ability to monitor deck operations from the wheelhouse via the multiple camera views provided by the EM system.
Three: Timely data review and feedback memos improve data quality and encourage skipper “ownership” of the data.
It became apparent over the course of the project that vessel operators do read feedback memos! Skippers often asked questions about the memos, sometimes complained, and nearly always appreciated seeing the photos from their trips. Most importantly, skippers made changes to improve data quality. As one Saltwater reviewer commented: “I have seen an improvement and increased involvement in making sure we’re able to get the data we need from their trips. Last year there was a vessel I would not have recommended for EM, but they have improved greatly”.
Four: NMFS-certified observers with onboard experience brought valuable insights and refinements to the data review process.
The North Pacific fishing industry has benefited greatly from the talent pool created by the North Pacific Observer Program. Observers with onboard 
experience in Alaska’s fisheries and knowledge from the observer program have gone on to fill key positions within NMFS, industry, academia, and the Council. For this project, Saltwater cross-trained NMFS-certified observers to accomplish the different functions necessary for EM program implementation including EM system installation, maintenance, in-port and remote service calls, as well as data processing and review. By integrating observers into the EM program, the overall North Pacific monitoring program retains the insights and knowledge of these biologists who have been trained in sampling protocols, species identification, and data integrity.
Saltwater reviewers worked closely with NMFS to refine and improve the review protocols for both pot and longline EM data over the course of the project. Because of their understanding of how the data is used by NMFS, and their knowledge of what observers are collecting onboard, the Saltwater review team was able to provide helpful insights into the review protocols so that NMFS managers could better understand “what’s behind the numbers” that come from EM data review. As an example, halibut viabilities determined via EM are not the same as viabilities determined by an onboard observer with a fish in hand. Thus, the metrics for assessing halibut viability via EM needed to be changed so the differences can be clearly understood by managers. Rather than a subjective assessment of image quality (high, medium, low) the Saltwater reviewers developed a quantitative assessment that allows fishery managers to know when, for example, water spots on a camera lens may, or may not, have affected species identification and by what degree.
Five: The open source review software was readily adaptable to changes in protocols or data fields, and its intuitive design made data review and the training of reviewers a cost-efficient process.
Saltwater used the open source EM review software developed with support from NFWF to review data for the project. The software uses templates, which are tailored to each fishery and easy to adapt. Saltwater developed training materials and quality control protocols to ensure data provided by different reviewers was standardized. The software’s intuitive design made training reviewers a relatively quick process. The end result allowed the review work force to efficiently fluctuate with seasonal demands and provide consistent, high quality data.
Nancy Munro is the founder of Saltwater Inc., and is focused on creating effective ways of collecting reliable data on fisheries and oceans to support sustainable management. She welcomes your thoughts and questions regarding this project. Please comment below or contact her directly at nancy.munro@saltwaterinc.com.
