Commercial fishing has been identified as one of the main threats affecting the survival of most seabird species. Although seabird mortality in Argentine longline and demersal trawl fisheries has already been characterized and quantified, the interactions with pelagic trawl fisheries targeting anchovy (Engraulis anchoita Hubbs & Marini, 1935) remains unknown. The goal of this study was to characterize seabird assemblages attending pelagic trawl vessels and to analyse their interactions (i.e. contact of the birds with the vessel and/or fishing gear and by‐catch). Data were obtained by on‐board observers during three consecutive fishery runs, 2011–2013. From a total of 333 observations, seabird abundance averaged 157.3 ± 229.7 birds per haul (totalling 23 species). Procellariiform followed by Charadriiform birds were the more frequent and abundant groups. The black‐browed albatross (Thalassarche melanophris (Temminck, 1828)), shearwaters (Ardenna spp. and Puffinus spp.), white‐chinned petrel (Procellaria aequinoctialis Linnaeus, 1758), and the kelp gull (Larus dominicanus Lichtenstein, 1823) were the most frequent and abundant attending species. The seabird abundance increased when the swell and the number of neighbouring vessels decreased. Seabird interactions with the vessel and/or fishing gear occurred in approximately 70% of the observations, with most of these representing interactions with the net (92%). The estimated contact rate was 16.7 birds h−1 per haul. A total of 121 birds were by‐caught and the average mortality rate was 0.55 birds h−1 per haul. Shearwaters and Magellanic penguins (Spheniscus magellanicus (Forster, 1781)) were the main by‐caught species (101 and 12 individuals, respectively). Lower levels of mortality were recorded in black‐browed albatrosses and white‐chinned petrels. The interactions increased in the presence of fishing discards and during haulback operations. This study is relevant to the implementation of the Argentine National Plan of Action – Seabirds, as well as for the continuing certification process in the anchovy fishery.

The food and feeding of Flemish Cap cod are described for 5 years based on 3921 stomachs collected in the fishing grounds off the Flemish Cap, Newfoundland in summer 1989-1993. Feeding intensity was high but the prey spectrum was narrow in all years with hyperiids and redfish (Sebastes sp.) predominating. Squid and polychaetes had a high inter-annual variability. Juvenile cod diet was dominated by crustaceans, mainly hyperiids, and polychaetes, while in adult cod diet the most important prey were fish, mainly redfish. The maximum size of redfish eaten increased with cod size, but prey-predator size relationships showed weak correlation. Cannibalism increased in 1991 (mainly upon 1-year-olds), coinciding with the appearance of a large year class in 1990. In the years 1992 and 1993, a change in the diet was observed involving an increase of hyperiids in the adult cod diet and a decline of redfish.

Oceanographic variables are important to determine oceanographic processes such as the distribution of water bodies, identification of marine fronts, areas of high primary productivity and/or high availability of prey (Pereira et al. 2018). Anomalies and gradients are used for identifying marine fronts at various temporal and spatial scales. Anomalies are used to identify mesoscale fronts as eddies, while the gradients allow for the identification of the confluence zones where the larger scale fronts are generated. Anomalies of Sea Surface Temperature (SSTa), Cholorophyll-a concentration (CHLORa) and Sea Surface Hight (SSHa) for each month were calculated as the difference between the average value for a given month and year, and the average for that month over a 8-year (for Sea Surface Temperature [SST] and Cholorophyll-a concentration [CHLOR]) and 4-year (for Sea Surface Hight [SSH]) period in each grid cell. Gradients were generated by calculating the standard deviation of each cell in relation to the adjacent cell values (Sidhu et al. 2012;Li et al. 2015). For identifying fronts, eddies with values of SSHa < -10 cm were considered as a cyclone (cold core eddy), SSHa > +10 cm as anticyclone (warm core eddy), and SSHa between < 10 cm and > -10 cm as confluence zones (Davis et al. 2002). Habitat selection modellingThe platform BIOMOD allows the use different modelling techniques, their evaluation, and the construction of different averaged outputs of the single-model predictions (Thuiller et al. 2009;Ducci et al. 2015). The outstanding features of this modelling technique are: the generation of random pseudo-absences (when data are presence-only), the use of different statistical methods (highly recommended because the prediction discrepancies between modelling techniques can be very large) (Araujo et al. 2005;Thuiller et al. 2009), and the use of data-splitting procedures, whereby a set of the original data are used of training the models and the remainder for model evaluation.