Figure 1

Conceptual figure of the experiment and influence of different errors on the structure of the capture histories (CH) in photograph-based population abundance estimation. Here, the true CH contains an individual (A) who was captured twice using a camera-trap and another individual (B) who was captured once. Capture-recapture methods use the number of individuals and the proportion of captures (1) and non-captures (0) to estimate the population abundance; thus anything influencing either of these factors will influence the population estimate. A shift error moves a capture event from one individual to another, but does not change the total number of individuals or the number of 1’s and 0’s (middle left). A combination error combines the captures from two individuals into one, reducing the number of individuals and the total number of zeros (lower left). A splitting error splits the captures from one individual into two and creates a “ghost” individual, increasing the number of individuals and the total number of 0’s (top right). A capture exclusion where identification is possible, is a form of identification error that changes a 1 to 0 in that individual’s CH (bottom right). Population abundance estimates will be underestimated by combinations and overestimated by splits. In a conventional capture-recapture framework, shifts will largely not affect population estimates. Exclusions may over- or underestimate the abundance, depending on whether they result in the loss of individuals from the CH or if they are non-random relative to individual identity (bottom right). The image was created in the software OmniGraffle 7 (https://www.omnigroup.com/omnigraffle).