Capture�mark�recapture, removal sampling, and occupancy models

Abstract Understanding the distribution and abundance of organisms is frequently stated as the objective of ecological investigations (Elton 1927; Krebs 1972). Similarly, distribution and abundance are primary criteria used to classify the status of species (e.g. threatened, endangered) for conservation purposes (Gardenfors et al. 2001). Thus, both scientific and conservation perspectives lead us to select abundance and occupancy as reasonable state variables for investigation. We define abundance as the number of organisms present in some area of interest and occupancy as the proportion of sample units or, more generally, of area, that is occupied by a species of interest. Estimation of these quantities is typically based on count statistics of some sort. For example, abundance estimation may be based on the number of animals caught in traps, detected by visual encounter surveys, or perhaps by auditory means. Estimation of occupancy is typically based on detection/non-detection data (frequently referred to as presence/absence data) representing counts of sample units at which the focal species is (or is not) detected. Use of count statistics to draw inferences about such quantities as abundance and occupancy requires attention to two important sources of variation in counts (Pollock et al. 2002; Williams et al. 2002; Lancia et al. 2005).