The annual International Council for the Exploration of the Sea (ICES) conference was held in Reykjavik, Iceland, on 23-27 September 2013. ICES was created in 1902 to “coordinate and promote marine research on oceanography, the marine environment, the marine ecosystem, and living marine resources.” The international conference provides an annual venue to present new marine research as well as discuss future scientific opportunities with colleagues in other agencies and academia. Three scientists from the FMA Division presented multiple papers at the fisheries sampling session entitled “What’s the Catch? Designing and implementing statistically sound fishery sampling schemes in the real world.” The objective of this session was to review the development of statistically sound sampling schemes which have incorporated practical challenges into their design and implementation, with a focus on the practicality of landing and discard estimations.
In 2013, the North Pacific Observer Program was restructured to expand the scope and improve the quality of information collected from vessels fishing in the federal groundfish and halibut fisheries of Alaska. This action provided the National Marine Fisheries Service (NMFS) the flexibility to decide when and where to deploy observers using a statistically sound sample design. The ICES conference provided an excellent opportunity for FMA to present results from several observed fisheries in Alaska with random catch sampling and several components of the 2013 sampling scheme implemented by the restructured Observer Program, including a review of current and past electronic monitoring systems utilized in lieu of human observers.
Using Random Sampling Methods to Sample Catch at Sea
Figure 1. A North Pacific groundfish observer measuring flathead sole (Hippoglossoides elassodon) randomly selected from the catch on a catcher/processor trawler.
Jennifer Cahalan, a NOAA affiliate from the Pacific States Marine Fisheries Commission, presented a paper discussing the viability of observers randomly sampling catch at sea on different types of vessels using several different fishing gear types. Sampling catches on commercial fishing vessels that are actively engaged in fishing activities is a difficult task, and observers do not always have the time, storage space, or tools available to sample according to a rigorous sample design. In 2010 observers deployed into the federal Alaska groundfish fisheries began recording the sample design used for each sampled haul as part of their standard data collections. Observers also recorded instances in which they were unable to collect their intended sample as planned. Data from 2010 through 2012 clearly showed that observers are able to implement random sampling methods. The sampling designs used tended to vary by vessel type, gear type, and vessel operations. Observers most often employed a systematic random or simple random sample design to collect samples used to determine the species composition of the catches. However, in difficult sampling situations such as trawl catcher vessels, observers had to resort to opportunistically collecting samples since randomized methods could not be consistently implemented. Ms. Cahalan presented a second paper in this session on catch estimations in federal trawl fisheries, co-authored by Jennifer Mondragon and Jason Gasper from the NMFS Alaska Regional Office.
Developing and Implementing the Annual Deployment Plan
Dr. Craig Faunce presented two papers in the fisheries sampling session summarizing the design and implementation of the restructured observer program. Prior to 2013, observer coverage in Alaskan fisheries was funded by commercial fishers at coverage rates set in federal regulation. Operators “self-selected” which fishing events would carry observers, providing a mechanism for biased observer data. A change in observer deployment and funding was initiated in 2008, which involved 53 persons and took 5 years to complete. Observers are now deployed onto vessels and into processing facilities according to an Annual Deployment Plan (ADP) that is developed annually by the Federal Government with public review. Although various concessions to the design were made to improve the salability of the new program, the 2013 ADP still incorporates sampling (deployment) frames with randomization to control bias. The second paper described the online implementation of the ADP using computer applications developed by coauthors from FMA; Paul Packer, Martin Park, Glenn Campbell, and Doug Turnbull. This ADP assigns fishing operations to deployment strata which randomly selects fishing trips and vessels for observer coverage. The Observer Declare and Deploy System (ODDS) facilitates the logging of trips by vessel operators in trip-selection, determines the protocol and appropriate probability of selection for that trip, and returns the outcome to the user. In contrast, vessel-selections are performed every two months based on past fishing activity. Selected vessels may petition the Federal Government for an assessment and a conditional release from coverage may be granted. The Vessel Assessment Logging System (VALS) facilitates the petition, stores information on the site visit, and constitutes a database on conditional releases. These computer applications provide the necessary data to evaluate observer deployment (and sampling) efficiency of the observer program in near real-time.
Electronic Monitoring Systems in the North Pacific
Figure 2. A North Pacific fisheries observer collecting random species composition samples from the catch.
Farron Wallace presented a fifth paper on electronic monitoring systems as a potential alternative to human observers in the North Pacific. Electronic monitoring (EM) is considered an important potential tool for fisheries monitoring onboard small catcher vessels due to the logistical and economic difficulties of accommodating an observer. In an effort to better inform future implementation of EM, a literature review was conducted on all past EM studies. Of the 59 EM study summaries located, 6 were published in peer review journals with the remaining either presented to the funding organization or published as a technical report. Sixty-three percent of the studies focused on catch composition, which requires species composition identification and quantification. In nearly all cases where species composition data was collected, video data quality was found to be inconsistent or missing for at least a portion of the study, thus degrading the ability to reliably distinguish species. Addressing these issues will be an important consideration prior to implementing an EM program in the North Pacific.
Conference summary and theme session reports can be found at: 2013 ICES annual science conference. Full conference proceedings and extended abstracts will be available soon.
