Researchers are challenged to understand the distribution of biodiversity on the landscape or determine if community-level assemblages respond to landscape variability. Canonical ordinations and multispecies occupancy modeling are often utilized and differ in their analytical scale and technique. Community-level data for carnivores (n=18) from camera trapping data (n=60) were tested for their collective response to covariates (distance to villages, roads, plantations, streams, and also topographic variables for elevation and slope) in Bukit Barsan Selatan National Park. Redundancy analysis (RDA) results confirmed that the carnivore community responded to the anthropogenic covariates, with carnivore detections being higher in areas at a high distance from roads and plantations medium distance villages. Then, multispecies occupancy models (MSOM), a comparatively finer temporal scale analysis that incorporated imperfect detections, were compared to RDA. We fit four models (full, anthropogenic, natural, and null) that were mainly inconclusive. In some MSOM models, the carnivore community did respond to environmental covariates, although the coefficients did not show a consistent response between seasons or years. These results indicate that RDA was able to detect broad-scale covariate effects that were not able to be modeled by MSOM and that these scaling and imperfect detection issues should be considered when attempting to understand landscape community diversity.