Are Block Nets Necessary?: Movement of Stream-Dwelling Salmonids in Response to Three Common Survey Methods

—Fish movement during sampling may negatively bias sample data and population estimates. We evaluated the short-term movements of stream-dwelling salmonids by recapture of marked individuals during day and night snorkeling and backpack electrofishing. Bull trout Salvelinus confluentus and rainbow trout Oncorhynchus mykiss were captured in sufficient numbers to evaluate instream movements and the influence of stream habitat on movement. Salmonids moved upstream more often than downstream, which suggests that fish fled in response to sampling procedures. Fish movement rates were related to sampling method and stream habitat characteristics. The proportion of bull trout (mean total length, 148 mm) moving upstream from a 50-m reach was 28% for day snorkeling, 25% for night snorkeling, and 18% for single-pass electrofishing. The likelihood of rainbow trout movement upstream during sampling declined as the percentage of rubble substrate increased. Rainbow trout (mean total length, 139 mm) traveled short distances; 67% of mobile rainbow trout traveled 25 m, whereas 59% of mobile bull trout traveled more than 50 m. Our results suggest that fish movement out of sample sites during sampling may negatively bias abundance estimates and positively bias estimates of detection probability. We suggest that block nets are necessary to ensure population closure when estimating population abundance for stream-dwelling salmonids. We further caution that, when fish flee in response to sampling, observations of habitat use may only apply to the proportion of fish that remained stationary and that were observed. Various sampling approaches are commonly used to assess presence, abundance, and status of stream-dwelling fishes. These surveys are often accomplished with the use of block nets to prevent the movement of individuals into or out of a sample unit (Li and Li 1996), and block nets are believed to be essential for ensuring fish population closure. For example, fish movement can negatively bias sampling data and population estimates (Kendall 1999). The magnitude of bias, however, will depend on the distances fish move and the number of fish moving relative to the length of the reach being sampled. This bias is additional to that * Corresponding author: [email protected]. uga.edu Received March 30, 2004; accepted August 12, 2004 Published online May 16, 2005 associated with the probability of capture or observation of fish remaining in the sample site. Fish movement out of sampling sites can influence species presence and absence data as well as abundance estimates because species detection depends in part on fish abundance (Bayley and Peterson 2001). If fish movement is affected by the size and species of fish or by physical habitat characteristics, then fish habitat studies and multispecies studies may be confounded. Thus, failure to use block nets may lead to biased data and hence poor management decisions. Block nets, however, are not always feasible to install and maintain, particularly in streams that are larger, that have high water velocities, or that carry debris. Financial (e.g., personnel) or logistical (e.g., stream access) constraints also may preclude use of block nets during surveys. Even when block nets are used, fish movement 733 EFFECTS OF SURVEY METHODS ON SALMONIDS in response to sampling may confound attempts to assess fish habitat utilization. Underwater observation, for example, is a common technique for assessing relations between stream-dwelling fish and habitat characteristics (Dolloff et al. 1996). In typical studies, one or more underwater observers traverse the sample unit and mark the locations of and habitats used by fish. Previous investigations have assessed fish–habitat relations by recording observations for individuals they judged to be undisturbed, but not for those that exhibited a flight response (Grossman and Freeman 1987; Petty and Grossman 1996). Inferences from these fish habitat studies may only be applicable for fish that actually did not move in response to the observation tech-