Characteristics of nocturnal roosts of house crows in Singapore

Abstract

Roosts of house crows (Corvus splendens) in urban Singapore are perceived as a nuisance. Public complaints primarily focus on the fouling of gardens, pedestrian paths, buildings, and vehicles beneath or near roosts. We investigated their roost-site-selection patterns to help develop management recommendations for controlling urban crows. Roost sites (n = 30) were better illuminated by 89%, closer to food centers by 672.5 m, 33.1% more enclosed by tall buildings, and located in areas of greater human activity (measured using variables such as pedestrian traffic) by 47.0% more than randomly selected non-roost sites (n = 30). Roost trees, which primarily consisted of angsana (Pterocarpus indicus), were taller by 66.7%, spaced farther apart by 4.1 m, had larger crown volume by 275.9%, and had denser crowns by 15.4% more than non-roost trees. Crow roosts were surrounded by 76.3% more built-up urban environments with 90.3% less vegetated areas and 89.8% fewer water bodies than non-roost trees. However, discriminant analysis (DA) identified only 4 variables that clearly distinguished roost from non-roost sites: average distance to surrounding buildings, clear bole height (the visible portion of tree trunk), distance between adjacent roost trees, and percentage of the area around the roost tree covered with vegetation. Discriminant analysis correctly classified 100% of roost trees and 80% of non-roost trees. We recommend measures designed to avoid the creation of new roost sites (e.g., planting short [