No abstract

Despite no obvious barrier to gene flow, historical environmental processes and ecological specializations can lead to genetic differentiation in highly mobile animals. Ecotypes emerged in several large mammal species as a result of niche specializations and/or social organization. In the North-West Atlantic, two distinct bottlenose dolphin (Tursiops truncatus) ecotypes (i.e. 'coastal' and 'pelagic') have been identified. Here, we investigated the genetic population structure of North-East Atlantic (NEA) bottlenose dolphins on a large scale through the analysis of 381 biopsy-sampled or stranded animals using 25 microsatellites and a 682-bp portion of the mitochondrial control region. We shed light on the likely origin of stranded animals using a carcass drift prediction model. We showed, for the first time, that coastal and pelagic bottlenose dolphins were highly differentiated in the NEA. Finer-scale population structure was found within the two groups. We suggest that distinct founding events followed by parallel adaptation may have occurred independently from a large Atlantic pelagic population in the two sides of the basin. Divergence could be maintained by philopatry possibly as a result of foraging specializations and social organization. As coastal environments are under increasing anthropogenic pressures, small and isolated populations might be at risk and require appropriate conservation policies to preserve their habitats. While genetics can be a powerful first step to delineate ecotypes in protected and difficult to access taxa, ecotype distinction should be further documented through diet studies and the examination of cranial skull features associated with feeding.

Ecological indicators for monitoring strategies are expected to combine three major characteristics: ecological significance, statistical credibility, and cost-effectiveness. Strategies based on stranding networks rank highly in cost-effectiveness, but their ecological significance and statistical credibility are disputed. Our present goal is to improve the value of stranding data as population indicator as part of monitoring strategies by constructing the spatial and temporal null hypothesis for strandings. The null hypothesis is defined as: small cetacean distribution and mortality are uniform in space and constant in time. We used a drift model to map stranding probabilities and predict stranding patterns of cetacean carcasses under H0 across the North Sea, the Channel and the Bay of Biscay, for the period 1990–2009. As the most common cetacean occurring in this area, we chose the harbour porpoise Phocoena phocoena for our modelling. The difference between these strandings expected under H0 and observed strandings is defined as the stranding anomaly. It constituted the stranding data series corrected for drift conditions. Seasonal decomposition of stranding anomaly suggested that drift conditions did not explain observed seasonal variations of porpoise strandings. Long-term stranding anomalies increased first in the southern North Sea, the Channel and Bay of Biscay coasts, and finally the eastern North Sea. The hypothesis of changes in porpoise distribution was consistent with local visual surveys, mostly SCANS surveys (1994 and 2005). This new indicator could be applied to cetacean populations across the world and more widely to marine megafauna.