Factors Affecting Robust Estimates of the Catch-and- Release Mortality Using Pop-Off Tag Technology
C. P. Goodyear
Abstract.—Most billfish caught by recreational and U.S. longline fishermen are returned to the sea and, because of their overfished status, the United States has urged that all live billfish taken in Atlantic longline fisheries be released. Knowledge of the proportion of these fish that die due to the catch-and-release process, is important both for stock assessment, and to know the potential benefit of releasing fish taken as bycatch in commercial fisheries. Existing information indicates that the magnitude of this mortality is low, but comes from a limited number of studies using small numbers of ultrasonic tags. Recent technology that uses tags that release from the fish after a preprogrammed time, and then transmit data to satellites, offers the potential for developing better estimates of release mortality. This paper uses simulation techniques to examine factors leading to robust estimates of release mortality. Most sources of error in tagging experiments will lead to upward bias in the estimates. These include tag failure, tagging induced mortality, natural mortality, and tag shedding. Given the importance of the estimate to future billfish management, initial studies should focus on proving the technology. Tag failures produce ambiguous results and should be minimized, to the extent possible, or eliminated from the analysis where appropriate. Under perfect conditions (no tag failure, no tag induced mortality, and no tag shedding), individual experiments should apply a minimum of about 100 tags. The length of time from tagging until the tag releases from the fish should only be long enough for the catch-and-release mortality to be fully expressed. Because each fishing mode is likely to have a different release mortality rate, each experiment only estimates the release mortality rate for the species, gear, and fishing method employed in the fishery studied. The number of tags required to estimate the total number of deaths of released fish, of all species, could be in the tens of thousands. However, a well-researched experimental design might reduce the required number of tags significantly.