Trace-contrast models for capture-recapture without capture histories

Capture-recapture studies increasingly rely upon natural tags that allow animals to be identified by features such as coat markings, DNA profiles, acoustic profiles, or spatial locations. These innovations greatly increase the number of capture samples achievable and enable capture-recapture estimation for many inaccessible and elusive species. However, natural features are invariably imperfect as indicators of identity. Drawing on the recently developed Palm likelihood approach to parameter estimation in clustered point processes, we propose a new estimation framework based on comparing pairs of detections, which we term the trace-contrast framework. Importantly, no reconstruction of capture histories is needed. We show that we can achieve accurate, precise, and computationally fast inference. We illustrate the methods with a camera-trap study of a partially marked population of ship rats (Rattus rattus) in New Zealand.