Development and implementation of a fully randomized sampling design for a fishery monitoring program

Fishery observer programs in the United States are mandated to collect fishery-dependent data. These monitoring programs must have clear objectives and be statistically sound despite the constraints imposed by commercial fishing operations. In addition, successful programs continuously self evaluate and improve. Decisions regarding sampling design must integrate the concerns and needs of industry and fisheries managers while ensuring the scientific integrity of the data. Here, we use examples from a large observer program to illustrate how the elements of a statistical design can be effectively implemented. With the rich economic resources of the fishing fleet in Alaska and a long history of observer coverage, the North Pacific Observer Program currently has high overall coverage rates and employs a statistically rigorous sampling design. The North Pacific Fishery Management Council receives 2 annual reports on the objectives and performance of fishery monitoring in the region. As fisheries management has become more sophisticated, data needs have increased. Balancing the amount and types of data collected with observer workloads is a critical component of designing sampling methods.