Counting birds in urban areas: A review of methods for the estimation of abundance

Abstract

Counts of birds can inform studies with different goals, such as estimating population size, monitoring populations over time and in response to environmental change, and estimating vital rates to model population dynamics. Because estimates need to be reasonably accurate and precise, considerable thought has gone into developing counting techniques that enable robust estimation of abundance, taking into account probability of detection, which can vary between species, land cover types and over time. In recent years these have been applied to over 60% of studies estimating bird abundance conducted in non-urban landscapes. However, robust estimation techniques are not being similarly applied to studies in urban areas. We reviewed 162 articles in which birds had been counted and abundance and/or occupancy reported in urban areas, spanning the years 1991 to 2015, and found that only 11% attempted to account for variable detectability; few of these had modelled detectability satisfactorily. There was no indication of increasing methodological rigour over time. Counting birds in urban areas poses significant challenges; robust techniques are constrained by limitations imposed by built structures, social factors and a mosaic of many small private parcels of land. We present a framework for estimating bird abundance and discuss the strengths and weaknesses of the different approaches, relating each to the urban context. Citizen science initiatives are considered as a good fit in urban areas and are increasing in number: sampling designed for all landscapes might be inappropriate in urban areas, but counting protocols should allow the modelling of detection probability.