Building better wildlife-habitat models

Wildlife-habitat models are an important tool in wildlife management toda?, and by far the majority of these predict aspects of species distribution ( abundance or presence) as a proxy measure of habitat quality. Unfortunately , few are tested on independent data, and of those that are, few show usefu l predictive st;ill. We demonstrate that six critical assumptions underlie distribution based wildlife-habitat models, all of which must be valid for the model to predict habitat quality. We outline these assumptions in a met e-model, and discuss methods for their validation. Even where all sis assum ptions show a high level of validity, there is still a strong likelihood th at the model will not predict habitat quality. However, the meta-model does suggest habitat quality can be predicted more accurately if distributional data are ignored, and variables more indicative of habitat quality are mod elled instead.