http://hdl.handle.net/11375/26954

Luymes, Nicholas (2021) Toolkit to manage key habitat for amphibians in Ontario forests, MacSphere Open Access Dissertations and Theses, http://hdl.handle.net/11375/26954

Forest-dwelling amphibians contribute to diverse ecosystem services in Ontario but are threatened by habitat degradation and fragmentation. My thesis investigated key amphibian habitats in Ontario forests with the goal of providing resource managers with tools and techniques to protect and restore amphibian populations. I primarily focussed on amphibians that breed in temporary forested wetlands known as vernal pools, as these wetlands are overlooked in provincial legislature and particularly sensitive to changes in land-use and climate. First, I investigated the distribution and community structure of vernal pools in forests of eastern Georgian Bay, Lake Huron. I developed an accurate remote sensing technique to map vernal pool habitat using readily available spatial data and found that undocumented vernal pools accounted for over half of the wetlands in the region. I documented the importance of the length of pool inundation (hydroperiod) and canopy openness in determining the composition of vernal pool amphibian assemblages. In particular, pools with short hydroperiods and closed canopies tended to support only early breeders and canopy generalists. Next, I used two case studies to demonstrate the usefulness of existing amphibian occurrence datasets, specifically for the improvement of habitat mapping and monitoring. For the first case study, I created habitat suitability models using known locations of the endangered Jefferson salamander (Ambystoma jeffersonianum). Models yielded good discriminatory ability between presence and pseudo-absence data and confirmed the importance of deciduous/mixed forests as key habitats. Habitat suitability maps revealed potential undocumented habitat in the Greenbelt region of Ontario. For the second case study, I developed optimizations of time and effort for a salamander monitoring program. I verified the need for at least ten years’ worth of monitoring data for reliable trend detection and demonstrated that the precision and accuracy of occupancy estimates are dependent on the allocation of effort across monitoring sites and repetitions.

https://macsphere.mcmaster.ca/bitstream/11375/26954/2/Luymes_Nicholas_W_202109_PhDBiology.pdf