Fishing effort as a covariate in predicting the at-sea distribution of a vessel-attending seabird

• Main Authors: 3rd World Seabird Conference 2021, Darby, Jamie
• Format: Article in Journal/Newspaper
• Language: unknown
• Published: Underline Science Inc. 2021
• Subjects: Environmental Sustainability; Fulmar; Bolton; Darby; Fulmarus glacialis; Northern Fulmar
• Online Access: https://dx.doi.org/10.48448/fg8r-n539; https://underline.io/lecture/34645-fishing-effort-as-a-covariate-in-predicting-the-at-sea-distribution-of-a-vessel-attending-seabird

Abstract: The Northern fulmar (Fulmarus glacialis) is a far-ranging pelagic-foraging seabird that can scavenge offal and discards from fishing vessels, and whose at-sea distribution is poorly understood. We modelled the foraging distribution of fulmar using GPS tracks (n=64, map attached) from nine colonies across Southeast Ireland and Scotland using biotic, abiotic and anthropogenic covariates. Behaviours were annotated to tracks using non-interpolated Hidden Markov Models to detect commuting, resting and foraging states. Areas foraged in were seen as preferred feeding habitat and tested against areas that were passed through with no foraging behaviour using a GLMM with an autoregressive correlation structure. Covariates included sea surface temperature, chlorophyll-a, depth, slope, fishing effort (mobile gears) and fishing vessel presence. Fisheries data were aggregated across multiple years across the breeding season (May-August, 2012-16) to provide generalised intensity per unit area rather than assuming direct vessel interactions in time and space. Fishing effort was an important predictive variable for habitat selection; others included bathymetric slope, sea surface temperature and chlorophyll-a concentration. Vessel presence and depth were dropped from the final model. The power of fishing effort to predict fulmar foraging distribution is likely due fulmars and fisheries targeting the same areas, as well as fulmar attendance at vessels to consume discards and offal. Finally we used this model to predict the at-sea distribution of fulmars breeding across the UK and Ireland. Our study demonstrates how including habitat variables based on human activity can provide greater predictive power than biotic and abiotic covariates alone for species feeding at higher trophic levels, and is one of a few to do so at a large spatial scale. Authors: Jamie Darby¹, Sophie de Grissac², Ellie Owen³, Mark Bolton³, Emily Shepard², Luca Borger², John Quinn⁴, Mark Jessopp⁴ ¹University College Cork - MaREI and School of BEES, ²Swansea University, ³Royal Society for the Protection of Birds, ⁴University College Cork