Rock Type and Channel Gradient Structure Salmonid Populations in the Oregon Coast Range

—The study objective was to investigate the response of salmonid populations to disturbance in Oregon Coast Range streams in two rock types, basalt and sandstone. Salmonid abundance was estimated in a total of 30 km of channel in 10 Oregon Coast Range streams with similar basin areas (14–20 km2). These basins had a range of disturbance caused by timber harvest, fire, and salvage logging. Mean channel gradient in sandstone was 0.012 m/m, and pools were the dominant habitat type. Mean channel gradient in basalt (0.025 m/m) was twice that in sandstone, and riffles were the dominant habitat type. Mean percentages by length of pools, glides, and riffles were 47, 33, and 20%, respectively, in sandstone, compared with 24, 27, and 50% in basalt. Channel gradient and channel morphology appeared to account for the observed differences in salmonid abundance, which reflected the known preference of juvenile coho salmon Oncorhynchus kisutch for pools. Coho salmon predominated in sandstone streams, whereas steelhead O. mykiss and cutthroat trout O. clarki predominated in basalt streams. In sandstone, juvenile coho salmon were four times more abundant than age-0 trout (steelhead and cutthroat trout combined). In basalt, age-0 trout were five times more abundant than juvenile coho salmon. Steelhead and cutthroat trout aged 1 or older were more abundant in basalt streams than in sandstone. However, mean densities of all salmonids combined were not different between rock types. We failed to find a clear fish response to disturbance, but our study shows the importance of geology in the design of studies investigating the response of salmonids to timber harvest and suggests that streams in basalt and sandstone have different potential capacities for salmonid communities. The effects on salmonid abundance of the frequent disturbance caused by forest practices have been a long-standing issue in the Pacific Northwest of the United States (Hicks et al. 1991b). Modifications to natural geomorphic processes attributable to timber harvest take place against a background of natural floods and wildfires, which have historically occurred less frequently than timber harvest but at unpredictable intervals (Reeves et al. 1995). The variable responses of salmonids have plagued investigations into the effects of timber harvest. In some circumstances, salmonids have been negatively affected (e.g., Hall et al. 1987; Schwartz 1991), whereas in others, there have been some positive responses (e.g., Murphy and Hall 1981; Hawkins et al. 1983; Hartman and Scrivener 1990). Still other studies have failed to find a salmonid response to logging (Bradford and Irvine 2000). Results from a long-term Carnation Creek study typify the variation: though logging appeared to increase the number of coho salmon * Corresponding author: [email protected] 1 Present address: Centre for Biodiversity and Ecology Research, Department of Biological Sciences, The University of Waikato, Private Bag 3105, Hamilton, New Zealand. Received October 30, 2001; accepted October 8, 2002 Oncorhynchus kisutch smolts leaving the basin, increased sedimentation of spawning gravel led to a substantial reduction in numbers of chum salmon O. keta fry moving to sea (Tschaplinski 2000). These variable responses may be attributable to differences in salmonid species, variation in forest practices, and differences in basin characteristics (Hicks et al. 1991b). For instance, rock type has a strong influence on channel morphology (Hack 1957; Keller and Tally 1979), as does large woody debris (Montgomery et al. 1995). Changes in channel morphology in response to timber harvest can affect salmonid distribution (Bisson and Sedell 1984; Tripp and Poulin 1986, 1992; Hogan and Church 1988), and latitude and climate can also modify the salmonid response (Hicks et al. 1991b). Study design has contributed to the apparently variable responses of salmonids. Short study reaches, often used because of limited resources, are generally assumed to be representative, but may not encompass the range of densities within the stream. We attempted to overcome this problem by choosing as the spatial scale stream segments that were 3 km long. These segments largely conformed to the criteria of Frissell et al. (1986), as they represented the variability in physical habitat while maintaining uniform channel slope and bedrock type. Our stream segments also contained 469 SALMONID POPULATION STRUCTURE enough channel units to provide a statistically valid application of double sampling (Hankin and Reeves 1988; Dolloff et al. 1993). Streams can widen in response to timber harvest (e.g., Lyons and Beschta 1983; Tripp and Poulin 1986). As stream width is dependent on hydrological characteristics such as basin area and discharge (Brush 1961; Osterkamp and Hedman 1977), the study segments were matched for basin size rather than for stream width. Basin size was controlled by positioning the downstream ends of the stream segments above a tributary junction, which is an accepted longitudinal boundary for stream segments (Frissell et al. 1986). The study objective was to investigate the response of salmonid populations to disturbance caused by timber harvest, fire, and salvage logging in Oregon Coast Range streams in two rock types, basalt and sandstone. The importance of this study is its attempt to isolate rock type as a factor causing some of the previously observed variable responses of salmonids to forest practices. Our hypothesis is that streams in basalt and sandstone have different potential capacities for salmonid communities, in the sense of Warren et al. (1979) and Frissell et al. (1986). By choosing streams with basins in uniform rock types and with a range of disturbance histories, but with similar basin areas, we also hoped to control for the effect of basin size on the response of salmonids to disturbance.