In California, Ixodes pacificus Cooley & Kohls nymphs have been implicated as the primary bridging vectors to humans of the spirochetal bacterium causing
Lyme disease (Borrelia burgdorferi). Because the nymphs typically do not ascend emergent vegetation, risk of human exposure is minimal in grasslands, chaparral, and woodland-grass. Instead, woodlands with a ground cover dominated by leaf litter (hereinafter referred to as woodland-leaf) have emerged as a primary risk habitat for exposure to B. burgdorferi-infected nymphs. As a means of differentiating woodland-leaf habitats from others with minimal risk (e.g., chaparral, grassland, and woodland-grass), we constructed a maximum likelihood model of these habitat types within a 7,711-ha area in southeastern Mendocino County based on the normalized difference vegetation index derived from Landsat 5 Thematic Mapper imagery (based on a 30 by 30-m pixel size) over four seasons. The overall accuracy of the model to discriminate woodland-leaf, woodland-grass, open grassland, and chaparral was 83.85% (Kappa coefficient of 0.78). Validation of the accuracy of the model to classify woodland-leaf yielded high values both for producer accuracy (93.33% of validated woodland-leaf pixels correctly classified by the model) and user accuracy (96.55% of model-classified validation pixels correctly categorized as woodland-leaf). Woodland-leaf habitats were found to be highly aggregated within the examined area. In conclusion, our model successfully used remotely sensed data as a predictor of habitats where humans are at risk for
Lyme disease in the far-western United States.