Avances en foto-identificación automatizada de fauna silvestre

La foto-identificación de especies silvestres es un recurso base para la obtención de información necesaria para diversas tareas de investigaciones biológicas. Hoy en día el crowdsourcing y la ciencia ciudadana están comenzando a desempeñar un rol importante en la recopilación de datos científicos. Esta fuente de datos permite aumentar considerablemente el número de registros en la base de muestreo de diferentes proyectos científicos, especialmente los relacionados con los modelos de capturarecaptura fotográfica de vida silvestre. No obstante, mientras que se aumenta la cantidad de datos recopilados de fuentes no científicas, se presenta un nuevo desafío, el procesamiento masivo de manera ágil y eficiente, que permita limpiar y seleccionar los datos relevantes para las siguientes etapas.Este trabajo aborda la automatización de la primer etapa del proceso de foto-identificación de cetáceos, el cual se trata de la detección de la presencia o ausencia de la región de interés en la imagen (ROI). Para ello, se especializó una red neuronal convolucional de propósito general (Mask R-CNN) con imágenes de delfines de la especie Cephalorhynchus commersonii recolectadas en diferentes sitios de la costa patagónica durante un período de siete años.

Development of a site fidelity index based on population capture-recapture data

BackgroundSite fidelity is considered as an animal’s tendency to return to a previously occupied place; this is a component of animal behaviour that allows us to understand movement patterns and aspects related to the animal’s life history. Although there are many site fidelity metrics, the lack of standardisation presents a considerable challenge in terms of comparability among studies.MethodsThis investigation focused on the theoretical development of a standardised composite site fidelity index and its statistical distribution in order to obtain reliable population-level site fidelity comparisons. The arithmetic and harmonic means were used as mathematical structures in order to create different indexes by combining the most commonly used indicators for site fidelity such as Occurrence, Permanence and Periodicity. The index performance was then evaluated in simulated populations and one real population of Commerson’s dolphins (Cephalorhynchus commersonii(Lacépède 1804)). In the first case, the indexes were evaluated based on how they were affected by different probability values such as the occurrence of the individual within the study area (φ) and capture probability (p). As a precision measure for the comparison of the indexes, the Wald confidence interval (CI) and the mean square error were applied. Given that there was no previous data concerning the distribution parameters of this population, bootstrap CIs were applied for the study case.ResultsEight alternative indexes were developed. The indexes with an arithmetic mean structure, in general, had a consistently inferior performance than those with a harmonic mean structure. The index IH4, in particular, achieved the best results in all of the scenarios and in the study case. Additionally, this index presented a normal distribution. As such, it was proposed as a standardised measure for site fidelity (Standardised Site Fidelity Index—SSFI).DiscussionThe SSFI is the first standardised metric that quantifies site fidelity at a populational level. It is an estimator that varies between zero and one and works in situations where detection is not perfect and effort can be constant or not. Moreover, it has an associated CI that allows users to make comparisons.