Evidence of absence is not proof of absence: the case of the New Brighton katipō

Abstract

ABSTRACTThe katipō is an endemic New Zealand spider that was previously common in the sand dunes at New Brighton. At sites on Banks Peninsula, katipō were detected under dried seaweed on the strandline 70% of the time. However, we detected no katipō among strandlines at New Brighton after 382 sampling visits. Incorporating these results into binomial and iterative Bayesian sampling models, it appeared highly unlikely that katipō still existed at New Brighton given so many non-detection events. However, when re-visiting the site, katipō were observed in the dunes at two locations, although they were still not found on the strandline. This specific habitat may be avoided at New Brighton due to high exposure to the prevalent strong easterly winds that occur at this site. The results emphasise that sampling models that use non-detection to indicate the likelihood of species absence can be highly specific to the sampling method used.