How much time is required to survey land birds in forest-dominated atlas squares?

Abstract

Using bird atlas data from south-central Sweden and information on survey effort for the corresponding 5 x 5 km grid squares, we assessed how much time is required to get reliable presence/absence data for forest birds. We restricted the analyses to a suite of 28 species considered common enough and sufficiently generalised in their requirements to be present in all survey squares largely dominated by forest. For a set of 183 squares having more than 2/3 of their area covered by forest, the non-linear relationship between survey duration and the number of species detected showed that a total survey duration of less than ca. 16 h was often insufficient for getting a reliable species list. There was, however, much variation in efficiency for such shorter survey durations. We conclude that a total survey duration of ca. 40-45 h should constitute a fair compromise between maximising survey quality and minimising effort, if the fieldwork is done by a range of birdwatchers of varying competence. However, longer surveys may be required for uncommon or specialised species. These findings have implications for the planning of further atlas work in forested areas and for the use of bird atlas data in various research fields.