Trends in small mammals derived from owl pellet data using occupancy modelling

Abstract

The EU Habitats Directive prescribes the monitoring of several small mammals on a national scale. A cost-effective way of monitoring these species is by using owl pellet data. Unfortunately, owl pellet data suffer from several methodological difficulties, all associated with the imperfect detection of the presence of prey species. Occupancy models can overcome the difficulties by taking detection into account, but they require temporal replicates. To supply these, we created replicates by splitting each pellet batch into two equal parts. A pellet batch consists of a number of pellets collected together in the field.Here we show how occupancy models can be applied to derive trend estimates from owl pellet data using such half batches as temporal replicates. We justify this approach by showing that the results from occupancy models treating half batches as temporal replicates in a test dataset were similar to the results of treating individual pellets as temporal replicates.The owl species and the number of prey individuals examined were included in the occupancy model applied to all data. We studied eleven small mammal species, two of which showed positive trends in occupancy. The confidence intervals of the trend estimates were satisfactorily small. Our methodological innovations reinforce the usefulness of pellet data for trend estimation in small mammals and increase the feasibility of large-scale monitoring of such species under the Habitats Directive.