Estimating abundance based on time‐to‐detection data

Many studies in ecology and management aim at quantifying absolute abundance based on counts a set of surveyed sites. As time for data collection is typically limited, methods reliable estimation occupancy or from low-cost are desirable. Time-to-detection (TTD) models have shown promise the occupancy. However, they remain heavily underutilized, restricted to inference about occupancy, rather than abundance. We developed binomial N-mixture model species-level TTD protocols that allows with multiple- single-visit data. An extension multi-visit version estimating availability per visit, given temporary emigration random. provide JAGS code new function (nmixTTD) R package unmarked fitting variety such models. Simulations showed accurate parameter if individual detection probability high (≥~0.7) number visited sites hundreds (≥~300). Additional visits improved accuracy estimates considerably. A comparison Royle-Nichols- classic N-mixture-model revealed performance our between these two, but requires less expensive error-prone count required N-mixture-models. In case study, we found similar results when analysing Royle-Nichols-, model. Analysing yielded lower higher detectability estimates. Presumably differences due emigration, as method individuals available one sampling occasion, whereas refer superpopulation, is, present over study period. Our TTD-N-mixture shows because it enables abundance, corrected imperfect detection, single- multiple-visit data, will be large quantities near future thanks technical advances like autonomous recording units. The effects unmodelled heterogeneity rate require further study.