Stream-associated amphibians are sensitive bioindicators in headwater streams across the Pacific Northwest moist coniferous forests of North America. Much of this landscape is intensively managed for timber. Forest Practices (FP) rules determine harvest prescriptions on most private lands in Washington State and cover over 3.7 million hectares. Under these rules, non-fish-bearing headwater streams receive buffers on at least 50% of the stream length, including FP Sensitive Sites that receive 15 to 17-meter radius no-cut patch buffers. We evaluated how torrent (Rhyacotriton spp.) and giant (Dicamptodon spp.) salamander relative abundance is influenced by headwater stream network features that correspond to FP Sensitive Sites. In particular, we examined how salamander relative abundance in the two most common FP Sensitive Sites, tributary junctions (TJs) and perennial initiation points (PIPs), compared to densities in non-Sensitive Site stream reaches, hereafter branches. We also evaluated salamander relative abundance and two hydrologic characteristics, dry channel and seeps. We analyzed data collected in 2006 and 2007 from 17 amphibian-occupied, non-fish-bearing basins in Western Washington with managed forest ages 30 to 80. We found no relationship between torrent salamander relative abundance and PIPs and TJs or between giant salamander relative abundance and TJs, when compared to branches. Consistent with expectations, giant salamander relative abundance was less in PIPs than in TJs and branches, and less in first-order than second- and third-order streams. Conversely, torrent salamander relative abundance lacked a clear relationship to stream order. Giant and torrent salamander relative abundance showed a negative relationship with proportion of dry channel, but torrent salamanders were observed in short reaches of surface water located in predominantly dry channels. Importantly, reaches with seeps had 123% (CI: +103% to +146%) and 81% (CI: +49% to +121%) greater relative abundance of torrent and giant salamanders, respectively, than reaches without seeps. Current FP rules protect select side-slope seeps as another category of Sensitive Sites but may too narrowly define the criteria of seeps to protect the full range of those being utilized by stream-breeding salamanders. Studies focused on seeps and adjacent stream channel characteristics may better inform features important to stream-breeding salamanders.

Stream reaches evaluated in this study were located in 17 amphibian-occupied non-fish-bearing stream basins in western Washington that were part of the Type N Experimental Buffer Treatment Study (McIntyre et al. 2018; see McIntyre et al. 2009 for detailed account of site selection process that screened for the following criteria). We collected salamander count data from streams that were sampled each year using a modified light-touch technique (Quinn et al. 2007; Lowe and Bolger 2002). We conducted an active search as we moved upstream, turning all moveable surface substrates ≥64 mm [2.5 inches] within the ordinary high-water mark, including in dry reaches, and within the saturated area of channel-connected seeps.  Amphibian observations were checked for errors and counted by reach type, study site and detection segement.