Use of Catenary Geometry to Estimate Hook Depth during Near-Surface Pelagic Longline Fishing: Theory versus Practice

Abstract

Management and conservation of many highly migratory fish species are based on population assessments that rely heavily on catch and effort data from the pelagic longline fishing industry. In 2003, we monitored hook time at depth for shallow-set commercial longlines (i.e., four hooks between surface buoys) targeting swordfish Xiphias gladius in the Windward Passage between Haiti and Cuba. We deployed temperature–depth recorders (TDRs) on about every 13th hook and attached them to branchlines just above the hook. Most TDRs were placed on branchlines that were predicted by catenary geometry to be at the deepest hook position between floats. Additional TDRs were also placed at the shallowest predicted hook position. We monitored 10 pelagic longline sets with a length (mean ± SE) of 44.9 ± 2.0 km. Time at depth for each TDR was binned into 5-m depth intervals. The expected bimodal distributions of hook time at depth were not observed; modes were 40 m for both the shallowest and deepest predicted hook positions. The majority of the hook depth distributions for shallow and deep hook positions achieved only 43% and 31%, respectively, of the depths predicted by catenary equations (i.e., <92 and <127 m). Individual TDRs were poor estimators of hook time at depth for other TDRs in the same catenary hook position during the same set (significant mean depth differences = 76.2–100%) and were even worse predictors of the depths fished during other sets (significant mean depth differences = 100%). Hook depth predictions based on catenary geometry drastically overestimated actual fishing depth in this study. These results indicate that the use of catenary geometry for estimating hook depth and subsequent vertical fishing effort is inadequate and fails to capture both within- and among-set variability, potentially resulting in biased stock assessments.