Premium
Coupling an oceanographic model to a Fishery Observing System through mixed models: the importance of fronts for anchovy in the Adriatic Sea
Author(s) -
Carpi Piera,
Martinelli Michela,
Belardinelli Andrea,
Russo Aniello,
Arneri Enrico,
Coluccelli Alessandro,
Santojanni Alberto
Publication year - 2015
Publication title -
fisheries oceanography
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.016
H-Index - 80
eISSN - 1365-2419
pISSN - 1054-6006
DOI - 10.1111/fog.12126
Subject(s) - anchovy , engraulis , akaike information criterion , pelagic zone , oceanography , stock assessment , mixed model , random effects model , environmental science , fishery , sea surface height , abundance (ecology) , sea surface temperature , geology , statistics , mathematics , biology , fish <actinopterygii> , medicine , fishing , meta analysis
Anchovy, Engraulis encrasicolus , forms the basis of Italian small pelagic fisheries in the Adriatic Sea. The strong dependence of this stock on environmental factors and the consequent high variability makes the dynamics of this species particularly complicated to model. Weekly geo‐referenced catch data of anchovy obtained by means of a Fishery Observing System ( FOS ) from 2005 to 2011 were referred to a 0.2 × 0.2 degree grid (about 20 km 2 ) and associated with the environmental parameters calculated by a Regional Ocean Modelling System, Adria ROMS . Generalized Additive Mixed Models ( GAMM ) with and without random effects were used to identify a relationship between abundance in the catch and oceanographic conditions. The outcomes of models with no random effects, with random vessel effects and with the random vessel and random week‐of‐the‐year effects were examined. The GAMM incorporating a random vessel and week‐of‐the‐year effect were selected as the best model on the basis of the Akaike information criteria ( AIC ). This model indicated that catches (abundance) of anchovy in the Adriatic Sea correlate well with low temperatures, salinity fronts and sea surface height, and allowed the identification of areas where high concentrations of this species are most likely to occur. The results of this study demonstrate that GAMM are a useful tool to combine geo‐referenced catch data with oceanographic variables and that the use of a mixed‐model approach with spatial and temporal random effects is an effective way to depict the dynamics of marine species.