A0620
Title: On the occupancy models with time-to-detection data
Authors: Wen-Han Hwang - National Tsing Hua University (Taiwan) [presenting]
Abstract: The detection time occupancy model introduced as an alternative method can estimate occupancy parameters even with a single visit, which is beneficial to improving the estimation efficiency of the common presence/absence occupancy model. However, the efficiency obtained from this method is only based on some examples and simulation studies in the literature. Compared with the presence/absence model, we derive the asymptotic relative efficiency and conduct simulation studies to strictly evaluate the effectiveness of the exponential detection time model. We also propose a mixed exponential detection time model, which can be linked to a negative binomial model if count data is available. The results obtained by analyzing the data of 63 bird species collected in the Karoo region of South Africa show that the proposed detection time model is generally more suitable for the data than the presence/absence model and the exponential detection time model.
