We introduce a Bayesian multivariate framework to investigate the sitefidelity patterns and estimate the population size of bottlenose dolphins (Tursiops truncatus) at the Tiber River estuary (central Mediterranean, Tyrrhenian Sea, Rome, Italy) between 2017 and 2020. In order to compare the results obtained through a distancebased clustering (Pace et al., 2021), a model-based clustering is performed using the same site-fidelity metrics: in particular, a multivariate finite mixture model is assumed for the vector of metrics (McLachlan et al., 2019). The proposed approach consists of two steps. We start with a Bayesian model-based classification of individuals in three different clusters labeled resident, part-time and resident using 347 unique individuals identified. Each individual is allocated to the group with the greatest estimated posterior probability. Finally, for each group, we estimate the corresponding population size via a capture-recapture analysis based on the Jolly-Seber model (Schwarz & Arnason, 1996): this kind of model allows to take into account the apparent survival probability of the animal in the population along with the capture probability. The results are compared to those obtained by the distance-based classification provided by Pace et al. (2021).

Model-based clustering for monitoring cetaceans population dynamics / Panunzi, G.; Caruso, G.; Mingione, M.; Alaimo di Loro, P.; Moro, S.; Bompiani, E.; Lanfredi, C.; Pace, D. S.; Tardella, L.; Jona Lasinio, G.. - (2021). ((Intervento presentato al convegno GRASPA 2021 tenutosi a Rome; Italy.

Model-based clustering for monitoring cetaceans population dynamics

Panunzi G.;Caruso G.;Mingione M.;Alaimo di Loro P.;Bompiani E.;Pace D. S.;Tardella L.;Jona Lasinio G.
2021

Abstract

We introduce a Bayesian multivariate framework to investigate the sitefidelity patterns and estimate the population size of bottlenose dolphins (Tursiops truncatus) at the Tiber River estuary (central Mediterranean, Tyrrhenian Sea, Rome, Italy) between 2017 and 2020. In order to compare the results obtained through a distancebased clustering (Pace et al., 2021), a model-based clustering is performed using the same site-fidelity metrics: in particular, a multivariate finite mixture model is assumed for the vector of metrics (McLachlan et al., 2019). The proposed approach consists of two steps. We start with a Bayesian model-based classification of individuals in three different clusters labeled resident, part-time and resident using 347 unique individuals identified. Each individual is allocated to the group with the greatest estimated posterior probability. Finally, for each group, we estimate the corresponding population size via a capture-recapture analysis based on the Jolly-Seber model (Schwarz & Arnason, 1996): this kind of model allows to take into account the apparent survival probability of the animal in the population along with the capture probability. The results are compared to those obtained by the distance-based classification provided by Pace et al. (2021).
979-12-200-8496-3
File allegati a questo prodotto
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/11573/1644910
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact