|
Résumé :
|
Studying social behavior has traditionally been labor-intensive, limiting the scale of datasets and therefore the questions that can be asked. We explore the potential for large individual capture datasets, which arise from population monitoring projects such as bird banding stations, to fill this gap. Using a 23-year dataset from San Jose, CA, comprised of six passerine bird species from four families, we constructed social networks to investigate whether age-, sex-, and taxon-related patterns of social association could be detected. We found support for age-related patterns in social behavior and patterns of overall species differences that agreed with past work. We then explored interactions between species and patterns in social behavior over time, two questions that have historically been difficult to address. We detected positive assortment by taxon, and we observed a significant multi-decade decrease in mean degree. Using an agent-based simulation to allow us to independently vary behavior and aspects of the capture dataset, we found that sample size and capture probability affected the metrics that were intended to only measure social behavior, but that these confounding effects could be effectively managed in many cases. We conclude that while analysis of individual capture datasets should not replace more intensive studies of social behavior, the utilization of these data could considerably expand the scope of our knowledge of social behavior, particularly at large temporal scales and for species that are difficult to observe.
|