Optimizing sampling interval length in species abundance comparison using the Royle-Nichols occupancy model

The Royle-Nichols occupancy model is commonly used for comparison of wildlife abundance between sites or years with biomonitoring data, such as camera-trap or bioacoustics records. The method requires that records of species are subdivided into sampling intervals with detection or non-detection. Here, we evaluated how the choice of interval length - which is commonly chosen arbitrarily - affects the outcomes of such comparisons. Using simulations and field data, we find that interval length can affect model results substantially. In some cases, different interval lengths can yield opposite differences in abundance between locations. Our results indicate that the interval length should be carefully selected based on properties of the data rather than arbitrarily chosen. Careful selection increases the accuracy of species’ abundance comparisons, and makes population comparisons more effective. We provide guidelines for optimizing the choice of interval length.