Determining Presence of Rare Amphibian Species: Testing and Combining Novel Survey Methods
Data deficiencies exist for ∼20–25% of amphibian and reptile species worldwide, typically excluding them from conservation consideration and funding opportunities. Current species status is often unknown at historic locations or previously unsurveyed areas. We present an iterative study to assess and optimize novel survey methods for a declining amphibian species (Boreal Toad, Anaxyrus boreas boreas) using a combination of methods. We found that if toads are present in a drainage, searching riparian areas multiple times during the active season is an efficient way of detecting the occurrence of previously unknown breeding populations. Once a breeding population has been located, traditional visual encounter surveys yield high probabilities of species detection for monitoring efforts (p̂vis ≈ 0.80). Supplementing streamside surveys when toads are not detected with other survey methods (e.g., environmental DNA [eDNA] samples) at suitable breeding locations can help confirm the species is absent, provided ≥10 eDNA samples are collected. Moreover, employing both visual surveys and eDNA samples can simultaneously yield distributional information on amphibian species and target pathogens, if pathogen presence is evaluated for all captured amphibian species and environmental samples are tested for both amphibian and pathogen DNA. Our iterative process of designing, testing, optimizing, and combining sampling methods to determine current species distribution should serve as a model for other rare amphibian and reptile species and provide managers better information with which to plan mitigation and conservation efforts.Abstract

Map of sampled stream drainages in Colorado. Inset gives an example of a subset of hoop nets including the adjacent net location (star location on each stream) and additional hoop net locations (inset) set at 500-m intervals up and downstream of the adjacent net (nmax = 17 nets on a given stream).

Photo of modified Nalgene bottles used to collect eDNA samples.

Boreal Toad occupancy probability estimates (ψ̂) for stream reaches (units) sampled in 2011–2012 in high- and low-population drainages. Estimates (solid line) and 95% confidence intervals (dotted lines) are obtained from the best supported model, ψ(Dist × Direction) p(Pop + StrGrad) (w = 0.34).

Boreal Toad detection probability estimates, p̂ for stream reaches (units) sampled in 2011–2012 in high- and low-population drainages as a function of stream slope. Estimates are obtained from the best-supported model, ψ(Dist × Direction) p(Pop + StrGrad) (w = 0.34); dotted and dashed lines represent 95% confidence intervals.

Boreal Toad eDNA detection probability estimates across dates for samples collected in 2014–2016. Estimates are obtained from the best supported model, ψunk (Yr) peDNA (Prev_det + Date + Yr) pvis (Sitetype). Detection probabilities are given for eDNA samples taken when toads were and were not observed via visual surveys in the same 2-wk period (dotted and solid lines, respectively).
Contributor Notes