The Van Dyke's salamander (Plethodon vandyke,) is a rare species endemic to Washington State. It has been found in cool moist microhabitats along streams, splash zones of waterfalls, and headwater seeps. We explored the association of the Van Dyke's salamander with hydrologic condition, geomorphology, disturbance characteristics, and vegetation structure in first- and second-order streams, and headwater seeps in the Cascade Range of Washington. We conducted salamander surveys and measured habitat characteristics at 50 streams and 40 seeps May-October 2000-2002. We described Van Dyke's salamander occurrence in stream and seep sites at three spatial scales: between sites, within sites, and between microhabitat sites. Using presence and absence as the response, we fit logistic regression models predicting Van Dyke's salamander occurrence. To identify the model that best fit the data, we ranked a priori models using Bayesian Information Criterion (BIC). Results were consistent for both stream and seep sites, at all three spatial scales. Best approximating models indicated that Van Dyke's salamander occurrence at sites was related to geological and hydrological habitat characteristics that provided hydnc and thermal stability. The probability of Van Dyke's salamander occurrence along streams was associated with habitat characteristics that protected salamanders from exposure, provided a source cover, and stream habitat types providing splash zone areas. Between streams, Van Dyke's salamander occurrence was positively associated with the proportion of valley walls with canopy cover <5%, the proportion of the stream channel dominated by bedrock, boulder, or soil substrates, and additional stream channels entering the main channel. Within streams, the probability of Van Dyke's salamander occurrence increased with the presence of non-forested areas, the presence of bedrock dominated stream habitat types, and the presence of vertical or V-shaped valley wall morphology. Between microhabitat sites, the probability of Van Dyke's salamander occurrence increased with an absence of trees, the presence of seeps, and the presence of small cobble sized substrates. The probability of Van Dyke's salamander occurrence in seeps was associated with habitat characteristics that protected salamanders within the larger landscape, provided a moisture gradient from dry to saturated, and the presence of cover objects. Between seeps, Van Dyke's salamander occurrence was positively associated with seep faces having a dry and sheeting hydrology, and with seep faces >5 m high. Within seeps, the probability of Van Dyke's salamander occurrence was negatively associated with seeps that had proportionately more point measures of total overhead cover that were >25%. Between microhabitat sites, the probability of Van Dyke's salamander occurrence was positively associated with an increase in the percent cover of small cobble, small gravel, and bedrock substrates. We conducted mark-recapture surveys of the Van Dyke's salamander at two high-gradient stream sites located within the Cascade Range of Washington State, June-November 2002. Sites known to support populations of the Van Dyke's salamander were chosen, and were ecologically different. One site, lacking significant overstory and located within the blast zone created by the 1980 eruption of Mount St. Helens, was surveyed 10 times. The other site, located in an old-growth coniferous stand, was surveyed 11 times. Abundance of salamanders at the blast zone site was estimated to be 458 (95% Cl: 306-739). Abundance of salamanders at the old-growth site was estimated to be 100 individuals (95% Cl: 61-209). Capture probabilities were extremely low (5 = <0.10) for all trapping occasions at both sites, with an average capture probability for the two sites of 0.038 (range = 0.02-0.09). Analysis of movement patterns suggested that most individual salamanders had home ranges <2 m, at least when moving on or near the surface. Individuals were recaptured under the same cover object as initial capture 36% of the time, and 89% of the recaptured individuals moved <2 m. Our results indicated that populations of the Van Dyke's salamander were rare on the landscape, even within the species documented range. Van Dyke's salamander occurrence was associated with geological and hydrological habitat characteristics that created microhabitats favorable for a species that is especially sensitive to heat and drying due to physiological constraints. Animals were difficult to detect due to fossorial habits and low capture probabilities, and it is likely that the Van Dyke's salamander was not detected even at sites where it existed. Life history characteristics, such as lunglessness and fossorial habits, low capture probabilities, and low abundances make it difficult to manage for and protect the Van Dyke's salamander. However, habitat associations may be used to identify and protect habitats suitable for Van Dyke's salamander occurrence. / Graduation date: 2004

No description available.