Patterns of relative trout abundance were used to select study section boundaries for subsequent monitoring. A total of 13 mainstem and 6 tributary study sections, ranging in length from 30 m to 298 m, were selected. All study sections were located upstream of the steelhead trout distribution. Study sections consisting of multiple channel-units (i.e., pool, riffle, cascade, and step) were classified into high and low relative abundance categories by graphically interpreting the moving average (window size = 6 channel-units) of coastal cutthroat trout relative abundance patterns obtained from the previous electrofishing survey. Mainstem stream sections having relative abundances of < 6 trout and tributary sections with < 3 trout were classified as low abundance.
In order to monitor coastal cutthroat trout movement into or out of individual study sections, 24 stationary PIT-tag antennas were installed at the upstream and downstream boundaries of each stream section. Seven stationary PIT-tag antennas were already located at mainstem-tributary junctions throughout the South Fork Hinkle Creek watershed (Gresswell et al. 2003). The stream-width antennas were oriented in a "swim-through" position (Castro-Santos et al. 1996; Barbin-Zydlewski et al. 2001) perpendicular to stream flow in relatively high-velocity habitats (e.g., riffles or pool tails). Antenna locations within individual channel-units were chosen to minimize multiple PIT-tag detections due to trout remaining in the antenna field for extended periods of time.
Antenna tuning, maximum read range, and detection efficiency were measured weekly. Antenna tuning was checked by calculating wattage (from voltage and amperage measurements) and comparing it to a standard for each stationary antenna. Maximum read range was defined as the furthest distance that a submerged PIT-tag, oriented perpendicular to the antenna, could be detected from the vertical axis of a stationary antenna. Detection efficiency was measured by calculating the proportion of submerged test tags that were detected at each antenna. Neutrally buoyant test tags were released upstream of antennas and allowed to drift through the antenna field at ambient water velocities. Time of each test was recorded so that results could be tested for correlations with stream temperature, stage height, and discharge. Both stage height and discharge were recorded to test for potential effects of water depth and velocity on antenna efficiency.
In addition to fixed monitoring stations, portable PIT-tag antennas (Roussel et al. 2000) were used to relocate trout bimonthly at the channel-unit scale. Portable antennas resembled backpack electrofishing anodes and were operated by a single observer wading upstream. In addition, watershed-scale portable antenna surveys occurred in December 2003, April 2004, and June 2004 (Gresswell et al. 2003). During each survey the PIT-tag code, date, time, channel-unit type, and location were recorded for each trout detected. Portable PIT-tag antennas had an omni directional read range of about 1 m. Sample efficiency was evaluated by using mobile antennas to detect PIT-tagged trout in block-netted stream sections with known numbers of PIT-tagged trout.
Twice during the study additional trout were marked in each stream section to mitigate for fish mortality and possible tag loss. On June 7 and 8, 2004 trout were captured by angling and PIT-tagged. In August 2004 an electrofishing census was conducted to recapture previously marked trout and PIT-tag additional trout.
Habitat Monitoring A watershed-scale habitat census was conducted concurrently with the August 2003 electrofishing survey, and channel-units were classified as pool, riffle, cascade, or step habitats (Bisson et al. 1981). To track habitat locations over subsequent sampling, individual channel-units located within study sections were numbered and marked with survey flagging. Furthermore, each channel-unit was linked to an existing watershed-scale geographic reference system comprised of numbered tree tags spaced about 20 m apart. This system was used to locate channel-units and fish positions by estimating the upstream or downstream distance to the nearest visible tree tag.
To evaluate possible physical factors that influence trout abundances, for each channel-unit we measured length, active channel width, substrate composition, canopy cover, shrub cover, stream gradient, pool spacing, boulder abundance, and large wood abundance. We chose these habitat variables based on their association with trout abundance and the observed habitat preference of stream trout as reported previously in the literature. Individual channel-unit measurements and counts were averaged for stream sections.
In order to understand how changes in habitat may have affected coastal cutthroat trout distributions we remeasured specific habitat features in December and April 2003, and August 2004. Each resurvey included identification of discrete channel-units (i.e., pool, riffle, cascade, and step), number of large wood pieces, number of boulders, and the maximum depth of each channel-unit. Another physical variable that we considered important to habitat quality was stream discharge. Stage height was recorded at a gauging station located near the mouth of South Fork Hinkle Creek (about 1 km below the mainstem study section), and discharge was calculated every 0.5 h (USGS station 14319830). During lapses in data recording, stage height was estimated by simple linear regression between gauges at South Fork Hinkle Creek and Little River near Peel, Oregon (USGS station 14318000; about 15 km southeast of the South Fork Hinkle Creek gauge).