Using eDNA and SCUBA surveys for detection and monitoring of a threatened marine cryptic fish

Abstract

Abstract Quantification of species’ spatial distributions and population trends is crucial for successful conservation efforts. Obtaining sufficient population data, however, is often difficult in the marine environment, especially for rare fish and invertebrate species that are small, cryptic and very difficult to detect. This study sought to understand the effort required to search for undiscovered populations of small, cryptic, marine species in shallow vegetated coastal habitats and track population numbers, using the Critically Endangered red handfish (Thymichthys politus) as a representative species. A sampling strategy was designed using a combination of environmental DNA (eDNA) and structured underwater scuba surveys of life‐like 3D‐printed ‘handfish replicas’ to estimate detectability and ultimately determine whether current population monitoring is adequately covering the remaining habitat occupied by the species. Tested over scales of hundreds of metres to kilometres, the eDNA assays performed relatively poorly in situ, detecting red handfish presence in only ~13% of samples collected from the centre of a known, yet low‐density, red handfish site. In contrast, underwater searches for independently placed handfish replicas by scuba divers indicated that mean detection probabilities at finer scales (~100 m) ranged from 57 to 97%. Near certain (95% probability) detection of an adult handfish was achievable with only one to three surveys (300 m2 belt transects), depending on the habitat complexity. While other species will vary in detectability using eDNA and underwater searches, these findings give insight into the general effort required for the detection of small, rare species inhabiting vegetated coastal environments. Such knowledge not only helps to refine monitoring and conservation efforts for known threatened species, but may also assist in identifying other inconspicuous species whose population declines may otherwise go unnoticed.