Small cetacean bycatch as estimated from stranding schemes: The common dolphin case in the northeast Atlantic

Death in fishing gear of non-target species (called ‘bycatch’) is a major concern for marine wildlife, and mostly worrying for long-lived species like cetaceans, considering their demographic characteristics (slow population growth rates and low fecundity). In European waters, cetaceans are highly impacted by this phenomenon. Under the Common Fishery Policy, the EC 812/2004 regulation constitutes a legal frame for bycatch monitoring on 5–10% of fishing vessels >15 m. The aim of this work was to compare parameters and bycatch estimates of common dolphins (Delphinus delphis) provided by observer programmes in France and UK national reports and those inferred from stranding data, through two approaches. Bycatch was estimated from stranding data, first by correcting effectives from drift conditions (using a drift prediction model) and then by estimating the probability of being buoyant. Observer programmes on fishing vessels allowed us to identify the specificity of the interaction between common dolphins and fishing gear, and provided low estimates of annual bycaught animals (around 550 animals year−1). However, observer programmes are hindered by logistical and administrative constraints, and the sampling scheme seems to be poorly designed for the detection of marine mammal bycatches. The analyses of strandings by considering drift conditions highlighted areas with high levels of interactions between common dolphins and fisheries. Since 1997, the highest densities of bycaught dolphins at sea were located in the southern part of the continental shelf and slope of the Bay of Biscay. Bycatch numbers inferred from strandings suggested very high levels, ranging from 3650 dolphins year−1 [2250–7000] to 4700 [3850–5750] dolphins year−1, depending on methodological choices. The main advantage of stranding data is its large spatial scale, cutting across administrative boundaries. Diverging estimates between observer programmes and stranding interpretation can set very different management consequences: observer programmes suggest a sustainable situation for common dolphins, whereas estimates based on strandings highlight a very worrying and unsustainable process