US Fish & Wildlife Service


Date of this Version



Published in THE PRAIRIE NATURALIST 36(4): December 2004 Published by the North Dakota Natural Science Society.


We investigated potential biases of using roadside point counts to sample breeding bird populations in the northern mixed··grass prairie. In 1995, a breeding bird inventory and monitoring program was initiated at Arrowwood National Wildlife Refuge, east-central North Dakota. Surveys were conducted annually through 1998 from an extensive point count array permanently located across the refuge (n = 162). Fifty-nine percent of the point counts were established directly on tertiary dirt roads and trails, while the remaining were interspersed at randomly selected distances of 100, 200, 300, and 400 m away from roads. Roadside point counts commonly consisted of a mosaic of habitat types, partly due to the inclusion of woodlands (riparian, shelterbelt or farmstead groves) and wetlands. The number of bird species observed at point counts tended to be greater on-road and the effect did not differ over time. Relative abundance estimates for most commonly observed species were not related to distance from road, either spatially or temporally. Only the relative abundance of savannah sparrow (Passerculus sandwichensis), grasshopper sparrow (Ammodramus savannarum), and western meadowlark (Sturnella neglecta) was associated with distance from road, but the effect was small (R2 < 0.06). In addition, there was little evidence that the annual variation in counts was related to distance from the road, which suggested that trend estimates derived during a population-monitoring program from roadside counts would not differ from off-road counts at Arrowwood National Wildlife Refuge.