Estimates of population size are a fundamental basis for any conservation plan. As the number of threatened species is increasing in all natural ecosystems and resources available for their conservation are limited, there is a need for practical, rapid and inexpensive methods to provide reliable metrics of animal population size and trends in abundance. Parrots are among the world’s most threatened groups of birds, and they are difficult to survey accurately. Today, distance sampling is the most reliable method for parrot abundance estimation; however, it is time-consuming and resource-demanding, thus often unfeasible where required resources and skills are lacking. We explored the efficacy of a simple method to assess parrots abundance, using Grey Parrot Psittacus erithacus as model species. Specifically, we compared local density estimates for Grey Parrot in the island of Príncipe (Gulf of Guinea) calculated by distance sampling, and the number of parrots recorded during long watches counts from vantage points overlooking forest in the same areas. We counted parrots seen flying in, out and over the surveyed areas during four different three-hour time slots, to account for variation in detectability at different times. We used linear mixed effects models to explore the relationship between the two methods.
Results/Conclusions
We found a significant correlation between local densities and the number of outbound flights recorded (i.e. parrots flying outward the surveyed area) for observations made in the last three hours of the day (i.e. 3PM – 6PM). This time slot corresponds to the activity peak of grey parrots in Príncipe at dusk. However, no relationship was found between local densities and the number of outbound, inbound or total flights in all the other time slots. These results suggest that flight frequencies themselves, as counted from vantage points, are not a reliable representation of local densities as calculated from distance sampling surveys. The correlation we found with outbound flights between 3PM and 6PM is likely to be affected by the high number of individuals travelling through the area to the roosting sites. We conclude that the use of long watches as an index of abundance is not recommendable if the relationship with actual local abundance has not been tested. However, long watches counts might provide a broad indication of abundance where no other option is available.