A Regression-Based Spatial Capture-Recapture Model for Estimating Species Density (303221)

Keywords: SECR, Density estimation

Data obtained from capture-recapture studies are essentially spatial in nature. The spatial proximity of the activity center, of an animal, and the trap location determines how likely the concerned individual will be captured. In order to incorporate the spatial information in the inference about the relative abundance of a species in the study region, Borchers and Efford (2008) proposed the spatially explicit capture-recapture (SECR) model. In its original form, SECR allowed the state-space of the activity centers of the individuals to arise from a non-homogeneous Poisson process (NHPP). However, in practise, complete spatial randomness (CSR) is generally assumed for the distribution of the activity centers. We perform simulation studies to demonstrate that an NHPP is better suited to capture the underlying intensity that generates the activity centers in SECR specification. We then illustrate our methodology on camera-trap survey data obtained during a survey of the swift foxes in north-west Texas. Then the relative abundance of swift foxes in north-west Texas was estimated.