A comparison of approaches for estimating Hochstetter's frog (Leiopelma hochstetteri) abundance : a thesis presented in partial fulfilment of the requirements for the degree of Master of Science in Conservation Biology at Massey University, Palmerston North, New Zealand
Loading...
Date
2022
DOI
Open Access Location
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Massey University
Rights
The Author
Abstract
The Hochstetter’s frog (Leiopelma hochstetteri) population of Aotea / Great Barrier Island is an evolutionarily distinct unit of a nationally ‘At Risk – Declining’ species. The Te Paparahi region of Aotea is this species’ stronghold on Aotea. However, it is unknown whether the population is temporally stable. I used a range of hierarchical models on data collected from this population between 2012 and 2021 to assess both temporal and spatial patterns in abundance in this population. Population data were collected from 15 100-m stream transects using a double-observer protocol in 2012 and 2015 and a triple-observer protocol in 2021. Frogs were not marked, but the 2021 survey tentatively identified individuals from their body size and position within a transect. The relative occurrence of waterfalls, selected plant species, and rat tracking within each transect was recorded to examine spatial correlation with frog abundance. I compared the efficacy of capture-mark-recapture (2021 only), N-mixture, occupancy models, and Poisson regression of single-count data for interring patterns in abundance. Occupancy was modelled at the level of 10 m segments, and the occupancy probability in each transect used to derive abundance. All models were Bayesian formulations constructed in OpenBUGS. Single-year capture-mark-recapture and N-mixture models, and multi-year N-mixture models, yielded sensible abundance estimates. The estimates of abundance generated by the single-year CMR model ranged from 0 – 118 frogs across the 15 transects, whereas estimates generated by the single-year N-mixture model ranged from 0 – 98 frogs. The estimates generated by the CMR and N-mixture models were strongly correlated, with N-mixture estimates 83% as large as CMR estimates. Abundance estimates derived from single- or multi-year occupancy models were neither sensible nor precise. The multi-year N-mixture model and Poisson regression of single counts suggested a ca. 5% decrease in Hochstetter’s frog abundance between 2012 and 2021. However, the possibilities of a stable or increasing population were not discounted by the 95% credible intervals for the effect of time. The relative abundance of kanono (Coprosma grandifolia) was the only habitat variable positively correlated with variation in Hochstetter’s frog abundance among transects. Although highly uncertain, the estimated decrease in abundance is of concern, and continued monitoring of this population is recommended. Furthermore, this pattern reflects the gradual declines displayed by other long-lived New Zealand herpetofauna in mammal-invaded environments. Although no correlation between Hochstetter’s frog abundance and rat tracking rates was identified, this result may be due to the small sample size of 45 tunnels which were only run on a single occasion. Therefore, it is recommended that the rat tracking surveys be repeated and expanded to include environments other than stream-gully margins.