A simulation-based evaluation of sea sampling design for American lobster in the Gulf of Maine

Fishery-dependent monitoring programs provide valuable information for the assessment and management of fishery stocks. The quality of the data and subsequent stock assessment are largely influenced by sampling design and sample size. The Lobster Sea Sampling Program in the Gulf of Maine yields critical information for the assessment of American lobster (Homarus americanus). To facilitate the use of an on-boat tablet-based data entry system, which will replace the current voice recording system, the sampling program requires development of a subsampling procedure to account for a decrease in the data recording speed after the new data entry system is implemented. We evaluated the performance of four potential subsampling strategies, including possible combinations of simple random vs. systematic subsampling with fixed subsample size vs. subsample size in proportion to reference data. Proportional subsampling showed lower relative estimation error and relative bias compared to subsampling with fixed subsample size. Within the two proportional subsampling strategies, systematic subsampling showed similar precision and accuracy compared to random subsampling, but systematic subsampling is more feasible to be conducted onboard. We concluded that the systematic subsampling of 50% of reference data could reduce current sampling effort but still capture the temporal trend of reference data. The simulation approach developed in this study can also be applied to optimize other fisheries-dependent monitoring programs.