Thompson et al. • LANDSCAPE EFFECTS ON CERULEAN WARBLER ABUNDANCE 725 
Percent bottomland forest in 250-m buffer Percent upland forest in 250-m buffer 
20 40 60 80 100 
Percent forest in 10-km buffer 
FIG. 2. Effects of forest composition at a local and landscape scale on the number of singing Cerulean Warblers along 
rivers in Missouri and Arkansas, 1999-2006. 
to be essentially zero below 50% forest cover in 
the 10-km buffer (Fig. 2). Other studies of forest 
songbirds have found support for landscape versus 
patch or habitat effects bill the strength of these 
effects varies considerably (McGarigal and 
McComb 1995, Howell et al. 2000, Hagan and 
Meehan 2002, Lichstein et al. 2002. Betts el al. 
2006). Landscape may have had a larger effect 
on abundance than habitat factors because our 
sampling was constrained to riparian areas where 
forest is generally suitable for Cerulean Warblers. 
The importance of landscape is also consistent 
with the theory that habitat selection is hierarchi¬ 
cal anti landscape provides an important proxi¬ 
mate cue (Hilden 1965). Potential ultimate factors 
affecting selection of more forested landscapes in 
this region include lower nest predation and brood 
parasitism (Robinson et al. 1995, Thompson et al. 
2002 ). 
Our finding that Cerulean Warbler abundance 
was greater in extensively forested landscapes is 
consistent with results elsewhere in their range. 
TABLE 3. Parameter estimates for the most supported generalized Poisson model used to model habitat effects on the 
number of singing male Cerulean Warblers along 5-km river segments in Missouri, 1999-2006. 
Effect 
Estimate 
SE 
95% CL 
Intercept 
-7.999 
1.495 
-11.187. -4.812 
Developed in 250 m 
-0.642 
0.494 
-1.622. 0.339 
Upland forest in 250 m 
0.017 
0.011 
-0.006, 0.040 
Bottomland forest in 250 m 
0.060 
0.019 
-0.021. 0.099 
Forest in 10 km 
0.088 
0.016 
0.056, 0.119 
