Variation in Juvenile Steelhead Density in Relation to Instream Habitat and Watershed Characteristics

Physical habitat characteristics can affect the abundance and distribution of organisms, and are frequently used to predict the standing crop of stream fish for purposes of understanding their ecology and better direct management. However, the spatial scale of the investigation and the resolution of the data can affect the outcome of such analyses. In this study we coupled watershed-level characteristics with instream habitat variables to model the density of two age-classes of juvenile steelhead Oncorhynchus mykiss in a watershed in the Clearwater River basin, Idaho. Density varied considerably across time and space. Variance partitioning showed that 41–50% of the variances in density were due to unexplained differences between sampling occasions (residual variance), and the rest resulted from variation at the site and site-and-year levels, depending on the age-class. Instream habitat variables better explained the variation in density than did models that included watershed-level characteristics. The density of subyearling steelhead was best explained by stream discharge, with a negative relationship. The density of yearling steelhead was best explained by a negative relationship with average weekly temperature; however, this relationship was statistically indistinguishable from zero. Finally, total density (subyearlings and yearlings combined) was best explained by discharge and average daily temperature. We believe that our approach is useful for identifying the physical factors associated with the density of stream salmonids, but we stress that findings from correlative studies should be interpreted in concert with detailed knowledge about life history variation in the study system. Identifying the principal determinants of population abundance is one of the main concerns in ecology (Hixon et al. 2002). Much of the discussion has centered on how populations are regulated at their upper limits by factors that slow down the intrinsic rate of increase (Begon et al. 1996). While it is clear that density-dependent processes are necessary to regulate a population (Murdoch 1994; Turchin 1995), the species’ life history characteristics and the conditions in the environment it inhabits are more important in determining its abundance (Sinclair et al. 2005). For example, cohort strength in Atlantic Salmon Salmo salar has been shown to depend negatively on discharge (Jensen and Johnsen 1999). At any given time, abundance is determined by the combined effect of all of the factors and processes that act on the population, dependent or independent of its density (Begon et al. 1996; Sinclair et al. 2005). Highly fecund organisms such as anadromous salmonids are particularly suitable for testing the role of environmental *Corresponding author: [email protected] Received June 20, 2014; accepted February 16, 2015 577 Transactions of the American Fisheries Society 144:577–590, 2015 American Fisheries Society 2015 ISSN: 0002-8487 print / 1548-8659 online DOI: 10.1080/00028487.2015.1022220 D ow nl oa de d by [ U ni ve rs ity o f Id ah o] a t 1 3: 45 2 4 A pr il 20 15 constraints on population size. The high fecundity of these organisms, realized by productive feeding opportunities in the ocean, has the potential to saturate the stream environment with offspring (Bjornn and Reiser 1991). Further, juveniles have specific habitat requirements depending on their age (Heggenes 1988; Rosenfeld and Boss 2001), and their life history causes some fraction of the population to out-migrate prior to the recruitment of a new cohort. Although density alone cannot be used as a metric of habitat quality (Van Horne 1983; Gaillard et al. 2010), if it varies consistently with certain habitat factors, this suggests that these factors are important (Pess et al. 2002; Fausch 2010). The basic premise when studying habitat relationships is that certain habitat characteristics produce a response in an organismal metric, such as individual growth or population density, because they confer some fitness advantage (MacArthur and Pianka 1966; Rosenzweig 1981; Gaillard et al. 2010). Much effort has thus been devoted to understanding the mechanisms through which habitat conditions affect the growth, survival, and reproduction of stream salmonids (Bjornn and Reiser 1991; Elliott 1994; Heggenes et al. 1999). The influence of physical habitat characteristics has primarily been studied at scales ranging from microhabitats (0.1 m) to stream segments (100 m), and the most important variables include depth, flow velocity, substrate size, and cover (Everest and Chapman 1972; Heggenes 1988; Beecher et al. 1993; Chun et al. 2011). However, watershed-scale physical factors such as geology and topography can control the distribution of these sitespecific stream habitat characteristics (Frissell et al. 1986; Richards et al. 1996; Wiley et al. 1997; Johnson et al. 2000; Wiens 2002; Allan 2004). For example, watershed geology can influence salmonid rearing habitat potential by constraining the morphological characteristics of stream reaches, resulting in spatial variation in habitat productivity beyond that which can be predicted by channel-level metrics alone (Burnett 2001; Hicks and Hall 2003; CoulombePontbriand and Lapointe 2004; Montgomery 2004; Deschênes and Rodriguez 2007). In this study, we quantified the variation in the density of juvenile steelhead Oncorhynchus mykiss across a watershed and modeled the variation in space and time using both instreamand watershed-level variables that describe physical habitat characteristics. The data on juvenile steelhead density and age structure came from a spatially extensive and temporally intensive monitoring approach in a watershed tributary to the Clearwater River, Idaho. While acknowledging that habitat factors alone cannot explain all the variation in density, our goal was to identify the levels at which the variation occurs and which habitat factors can explain this variation during the summer and fall seasons. We expected the densities of the two age-classes in the study to be associated with different factors pertinent to both their habitat preferences and ontogenetic stage. METHODS Study Area and Population The four main streams in the Lapwai Creek watershed (694 km) drain the north slopes of Craig Mountain (elevation, 1,530 m), carve steep canyons through the landscape, and empty into the Clearwater River (elevation, 237 m) (Figure 1). The predominant geology in the watershed is Columbia River basalt, with a band of Idaho Batholith in the upper, high-elevation portion. The plateau above the escarpment is overlain with loess, and the predominant land use is dryland grain agriculture, which covers 34% of the watershed. Coniferous forests cover 29%, primarily at higher elevations above the prairie, and grasslands dominate the steep canyon sides and valley floors. Mean annual precipitation is 490 mm, with higher amounts falling at higher elevations. Five percent of the watershed area is classified as urban, i.e., developed for housing and infrastructure (Homer et al. 2007). Flood control levees and infrastructure have channelized the main-stem reaches (Richardson and Rasmussen 2007) and prevent connectivity with the floodplain (Williams 2011). The selection of study sites (approximately 100 m in length) was done in a stratified random fashion (Frissell et al. 1986). At the scale of each of the four subwatersheds we identified three regions (1 km in length) that spanned a gradient of physiographic (topography, geology, and land cover) and land use FIGURE 1. Map showing the four major streams of the Lapwai Creek watershed and the locations of the study sites. The first letter of each site code refers to the stream segment (upper, middle, or lower), the second to the stream (Sweetwater, Webb, Mission, or Lapwai), and the third to the site position in that stream (upper, middle, or lower). For example, site USL D upper Sweetwater lower. All 16 sites were sampled in 2010 and 2011, whereas 7 of these (open circles) were not sampled in 2012. 578 MYRVOLD AND KENNEDY D ow nl oa de d by [ U ni ve rs ity o f Id ah o] a t 1 3: 45 2 4 A pr il 20 15 conditions. Within each region we identified several 100-m sections that were representative of the stream channel in the region, from which we randomly selected one as our study site (described in Myrvold and Kennedy, in press). Field data were collected on average five times per year at each study site, from early June to the end of October. We sampled 16 sites in 2010 and 2011 and 9 sites in 2012 (a subset of the 16 sites). Due to unforeseen events such as equipment failure, we were unable to sample on 7 of these occasions, which resulted in a total of 198 sampling events. These occasions occurred at random. Other fish species at the study sites include (in order of abundance) Longnose Dace Rhinichthys cataractae, sculpins Cottus spp., Bridgelip Sucker Catostomus columbianus, Redside Shiner Richardsonius balteatus, Northern Pikeminnow Ptychocheilus oregonensis, and Chiselmouth Acrocheilus alutaceus. In recent years juvenile Coho Salmon Oncorhynchus kisutch have been stocked as part of a reintroduction program in lower reaches of the system, but they are generally not sympatric with steelhead in space or time. No hatchery supplementation exists for steelhead in the watershed.