Document Type


Date of Degree Completion

Spring 2018

Degree Name

Master of Science (MS)



Committee Chair

R. Steven Wagner

Second Committee Member

Megan K. Walsh

Third Committee Member

Wayne Quirk


Detection of stream-associated amphibians in visual encounter surveys is challenging due to their cryptic nature; however, occupancy models were developed to deal with these detectability problems and provide estimates of occupancy that can also be related to site characteristics. Highway crossing risks and habitat isolation were mitigated for in recent construction of wildlife underpasses, where creeks cross Interstate 90 east of Snoqualmie Pass in Washington State. The effects of these restored underpasses on stream-associated amphibians were evaluated across 8 creeks, some with and some without restored underpasses, by comparing modeled occupancy of 3 amphibian species in stream habitat upstream, under, and downstream of Interstate 90. The amphibians modeled in this study are Coastal Giant Salamander (Dicamptodon tenebrosus), Cascades Frog (Rana cascadae), and Coastal Tailed Frog (Ascaphus truei). Multiple visual encounter surveys were conducted in the 8 creeks over two years during July-September. Over all surveys, D. tenebrosus, R. cascadae, and A. truei had detection probabilities of 0.66, 0.51, and 0.39/survey, respectively. Average occupancy probabilities were similar among these 3 species: 0.54/surveyfor D. tenebrosus, 0.55/survey for R. cascadae, and 0.52/surveyfor A. truei. Creek section occupancy model estimates support the use of underpass-culverts by all three amphibians in this study. Although highway underpass renovation occurred fairly recently, R. cascadae and D. tenebrosus are already being found within newly completed underpasses with rock substrate that matches the surrounding habitat. Recommended features that should be incorporated into future crossing structures to enhance connectivity between amphibian populations in the I- 90 Snoqualmie Pass East Project include (1) incorporating rock substrate that mimics surrounding stream habitat as much as possible, (2) planting native vegetation that will eventually provide canopy cover, and (3) manipulating the creek’s overall slope or gradient as little as possible, as this will retain vital pools and small waterfalls.