A primary objective of the Fisheries Monitoring and Analysis (FMA) Division is to provide high resolution data from the North Pacific Groundfish Observer Program to meet end user needs. We collect and process a large volume of data each year and we use a combination of manual data review methods, computer programs, and input from end users to help us detect and correct data errors. All detected errors aid us in the continued development of computer programs to help identify problems before they reach a data user. This report summarizes our commitment to data quality with a focus on procedures to ensure our species identifications are correct, followed by an update on work to upgrade the program's administrative structure.
Information and Monitoring Technologies
Staff from the Information and Monitoring Technologies program, the Groundfish Assessment Program of the Resource Assessment & Conservation Engineering (RACE) Division, and Office of Fisheries Information Systems (OFIS) are collaborating to develop a system to improve the quality of the observer species identifications. The system will combine technology from Environmental Systems Research Institute (ESRI) GIS software and Oracle Corporation's relational database.
Using ESRI's ArcView application and ArcSDE server, the group has developed a process for creating spatial representations of the historical distribution of fish species based on the information collected by observers and from RACE surveys. These spatial representations are then stored as a specialized data type within the Oracle database where they become part of the computer programs that screen the data collected by observers.
When the data transmitted from the field by observers arrives in the FMA offices, the computer programs will identify data records that contain species that are outside the known distribution of each species encountered. Observers will then be requested to verify their identification of species outside their normal range. When implemented, this process will supplement the manual processes currently used by staff to verify species identification and will further improve the quality of the species composition data.
By Michael Brown
Observer Services
Correct species identification is a cornerstone of high quality observer data, and staff have recently been working on improvements to observer training materials and how we track each observer's species identification history. The major improvement made to observer species identification training materials this quarter was the creation of a waterproof fish identification key. The new key includes major finfish families, flatfish, salmonids, gadids, skates, crabs and a few select invertebrates. The new keys complement a waterproof "observer edition" of the Guide to Rockfishes of the Northeast Pacific Ocean by Orr et al., in use since 2000.
Observers work on deck and in factory processing facilities, where water, fish slime, and fish blood are abundant. However, with over 250 observers at sea during peak fishing seasons, we did not have sufficient resources to provide observers with dichotomous keys that could hold up well in this wet environment. Observers were issued plain paper guides, and instructed to keep them as dry as possible. This task was often difficult, because we also require observers to key out species with the fish "in hand." The new waterproof guides were enthusiastically greeted by the observer community, and we are confident that their use will increase our certainty of observer's species identifications in the future.
We require observers to complete a "Species Identification Form" for each species encountered so that we can verify these species identifications which are made in the field with no direct supervision. These paper forms were kept on file here at the AFSC, but were not readily available to observers, field staff, or data users with questions regarding a specific identification. New this year, we are using a computer application to record each observer's species identification history. The application allows us to create a "life list" of species described by each observer.
The database allows for improved data quality by enabling debriefing staff to ensure that an observer has adequately described each species reported during a deployment. It also enables us to include images such as digital photos of the actual fish or scanned images of the species ID form and debriefer comments regarding the identification. This new application will allow the FMA Division to confidently stand behind our data when identification questions arise, such as "Was that really a green sturgeon?
