The influence of fish farm activity on the social structure of the common bottlenose dolphin in Sardinia (Italy)

Abstract

In a wide variety of habitats, including some heavily urbanised areas, the adaptability of populations of common bottlenose dolphin (Tursiops truncatus) may depend on the social structure dynamics. Nonetheless, the way in which these adaptations take place is still poorly understood. In the present study we applied photo-identification techniques to investigate the social structure of the common bottlenose dolphin population inhabiting the Gulf of Alghero (Sardinia, Italy), analysing data recorded from 2008 to 2019. The social structure analysis showed a division of the entire population into five different communities and the presence of non-random associations, while there was no evidence of segregation between sexes. Furthermore, results highlighted an important change in social structure through time, likely due to a reduction in fish farm activity since 2015. The division of the population into different communities, the presence of segregation based on the foraging strategy (inside or outside the fish farm area) and the social network measures were evaluated by analysing independently the two datasets: the intense and low farm activity periods: 2008–2014 and 2015–2020, respectively. Segregation among individuals belonging to the same foraging strategy class was found only in the earlier period, and the composition of the four communities was consistent with this result. Our study improves the knowledge about bottlenose dolphin adaptation, as a lower complexity in social structure was linked to a reduction in anthropogenic food availability.

Accuracy and power analysis of social networks built from count data

“…We have shown that the required level of network correlation is context dependent. This contrasts with some studies that have used Whitehead (2008)'s method to calculate network correlation, and suggest a minimum acceptable level of network correlation at 0.40 (Chabanne et al, 2017;Frau et al, 2021;Hawkins et al, 2020). Our results suggest a threshold value of 0.40 is not generally optimal when conducting nodal regression on social networks built from count data, and that the true optimal will depend on the level of social differentiation and sampling effort.…”

Accuracy and Power Analysis of Social Networks Built From Count Data

“…We have shown that the required level of network correlation is context dependent. This contrasts with some studies that have used Whitehead (2008)'s method to calculate network correlation, and suggest a minimum acceptable level of network correlation at 0.40 (Chabanne et al, 2017;Hawkins et al, 2020;Frau et al, 2021). Our results suggest a threshold value of 0.40 is not generally optimal when conducting nodal regression on social networks built from count data, and that the true optimal will depend on the level of social differentiation and sampling effort.…”

First Report of Epimeletic and Acoustic Behavior in Mediterranean Common Bottlenose Dolphins (Tursiops truncatus) Carrying Dead Calves

“…However, anthropogenic disturbances and threats (vessel traffic, reduced prey availability caused by overfishing, plastic and chemical pollution, and habitat degradation, including noise) may impact distribution and individuals’ interactions, thus altering the social structure and influencing how the population responds to changes to its environment [ 57 , 58 , 59 , 60 , 61 ]. Future investigations should address the structure of the relationships between individuals as it is an essential aspect affecting animals’ responses to both human-related pressures and management actions [ 57 , 62 ].…”