Humans turn natural landscapes into complex mosaics, which include agriculture, urban, developing, and natural remnants as land cover.  Population responses to these landscapes are just as varied.  Consequently, interactions between organisms are likely to be altered including parasite-host interactions.  Our goal was to quantify the spatial relationships between song sparrows (Melospiza melodia) and several hematozoan parasites in a complex landscape.  Fifty-four sparrows were captured in habitats varying from urban centers to relatively rural sites.  A blood smear was made from each bird and examined for 15 minutes (30 minutes total) by two observers.  Capture sites were georeferenced and served as centers of 90, 210 and 990 m diameter circular buffers from which we extracted the proportion of urban, scrub, and forest cover available from the 2001 USGS National Land Cover Database.  We created a number of biologically plausible logistic models with presence/absence as the response variable and arcsine-square-root transformed landscape variables as the explanatory variables.

Results/Conclusions

Using Akaiki Information Criterion (AIC) to rank models, the most parsimonious model (best fit with the fewest variables) included percent forest cover in 90 m buffers and percent urban in 210 m buffers.  Other top-ranking models only include urban and forest at smaller scales.  Models that included larger-scale land cover information or the amount of scrub performed worse than a null model.  These results conflict with most avian biodiversity studies that demonstrate the importance of larger-scale land use for determining species richness.