Using Landscape Analysis to Test Hypotheses about Drivers of Tick Abundance and Infection Prevalence with Borrelia burgdorferi

Patterns of vector-borne disease risk are changing globally in space and time and elevated disease risk of vector-borne infection can be driven by anthropogenic modification of the environment. Incidence of Lyme disease, caused by the bacterium Borrelia burgdorferi sensu stricto, has risen in a number of locations in North America and this increase may be driven by spatially or numerically expanding populations of the primary tick vector, Ixodes scapularis. We used a model selection approach to identify habitat fragmentation and land-use/land cover variables to test the hypothesis that the amount and configuration of forest cover at spatial scales relevant to deer, the primary hosts of adult ticks, would be the predominant determinants of tick abundance. We expected that land cover heterogeneity and amount of forest edge, a habitat thought to facilitate deer foraging and survival, would be the strongest driver of tick density and that larger spatial scales (5–10 km) would be more important than smaller scales (1 km). We generated metrics of deciduous and mixed forest fragmentation using Fragstats 4.4 implemented in ArcMap 10.3 and found, after adjusting for multicollinearity, that total forest edge within a 5 km buffer had a significant negative effect on tick density and that the proportion of forested land cover within a 10 km buffer was positively associated with density of I. scapularis nymphs. None of the 1 km fragmentation metrics were found to significantly improve the fit of the model. Elevation, previously associated with increased density of I. scapularis nymphs in Virginia, while significantly predictive in univariate analysis, was not an important driver of nymph density relative to fragmentation metrics. Our results suggest that amount of forest cover (i.e., lack of fragmentation) is the most important driver of I. scapularis density in our study system.