Dinâmica Temporal e Espacial em um Megadormitório da Andorinha-azul (Progne subis), uma Ave Migratória Invernando na Amazônia

Abstract

Wintering is the least known part of the annual cycle of many migratory birds and can be an important phase for those species experiencing population declines. The Purple Martin (Progne subis) breeds in North America, where an alarming population reduction has already been observed, and then migrates to South America for the non-breeding period. Here, individuals gather in large numbers every night at roosts, which offer a unique opportunity to study them during their wintering. We studied a P. subis roost near the city of Manaus (Amazonas - Brazil) during three winters (2019, 2020, 2022) to determine number of individuals, age and sex composition of the population, and why and for how long they use the roost. During two seasons, we did morning counts of birds leaving the roost, resulting in estimates of 244-320 thousand martins. We also captured and banded 198 individuals to assess their physical condition, and in some of them we placed Motus radio transmitters to quantify the length of their stay at the roost. We analyzed the relationship between length of stay and body variables, such as weight, fat, molt, muscle, sex and age (young and adult). Of the 59 swallows tracked, the longest stay was one month, but 96% stayed for just two weeks or less. Among the variables analyzed, fat accumulation, pectoral muscle size, molting and age were most related to length of stay. When they complete molt of their flight feathers (remiges and rectrices) and accumulate adequate fat and muscle, they leave the roost. We believe that the Manaus roost serves as a pre-migratory staging area where the birds arrive almost ready for migration, and when ready they leave for their breeding areas. The short permanence of individuals in the dormitory, relative to the active period of the dormitory (2-3 months), suggests that there is an almost constant flux, with replacement of individuals throughout the period, and consequently that the daily counts drastically underestimate the total number of individual martins that use the roost in a season.