Comparing avian diversity of a northern hardwood forest and nearby agricultural grasslands to illustrate the shortfalls of point counts.

Abstract

Point counts are the most used surveys used to measure avian diversity and abundance. In recent years, however, the validity of using point counts has come under questioning. Two types of error are imvolved with point counts: bias and variance. Tias occurs when there is any kind of over- or underestimation during the survey: when the counts are not conducted from random places, a bird is identified incorrectly, data is recorded improperly, or is simply not seeing every bird in an area. Variance occurs in many population studies: birds often change position or there could be background noise affecting how many birds occur in the same area in repeated counts (Thompson 2002). Other problems include differences in weather, observer ability and performance, and the appearance and behavior of birds that makes them hard to observe (Rosenstock 2002). Despite these problems, point counts remain the most popular surveys because they are easy and cost effective. The importance of monitoring bird populations - for management, understanding reactions to environmental problems, and studying population trends - demands more precise techniques should be developed to measure the true density of bird populations. For a number of years the Biology of Birds class at the University of Michigan Biological Station in Pellston, MI, has recorded richness and abundance in two locations near the station, one a forest, the other agricultural grasslands. Using this data, I will show how one variable, feeding flocks, affects the measures of diversity, abundance, richness, and evenness of the two habitats to elucidate the inefficiency of point counts.