Publication Title: Biases in estimating fish populations due to large wood and other physical habitat in Oregon Cascade streams [Abstract]
Year: 2000 Status: Published Publication Type: Abstract
H. J. Andrews Publication Number: 2842
Citation: Bayley, P. B.; Burgess, J.; Gregory, S. V. 2000. Biases in estimating fish populations due to large wood and other physical habitat in Oregon Cascade streams [Abstract]. In: International conference on wood in world rivers; 2000 October 23-27; Corvallis, OR. Corvallis, OR: Oregon State University: 113.
Online PDF: http://andrewsforest.oregonstate.edu/pubs/pdf/pub2842.pdf
Abstract: The efficiency of an intensive capture process (catchability) was determined from 3 backpack units used simultaneously, including one applied by a snorkeler underwater, in pools of Lookout and Tidbits Creeks, Oregon Cascades. Efficiency was expressed as numbers caught from a single pass from all units, divided by the original number present, which was determined from a known number of marked fish released in their locations of capture in a blocked off pool the previous day. Each stream received four treatments of distinct quantities of large wood, and were sampled during two summers, resulting in 32 calibrations. Catchability was modeled for salmonids, dominated by coastal cutthroat (Onchorynchus clarki) and cottids. For salmonids fish size and amount of large wood were highly significant effects. No species effects (cutthroat versus rainbow) could be detected among salmonids. The fitted model of salmonid catchability was a set of unimodal curves as a function of fish length, with peak catchabilities of 0.74 for zero wood, and 0.57 for the densest wood treatment (6.2 mean wetted wood area as percent of water surface area) that corresponded to a fish length (fork) of 135 mm. The salmonid model predicted corresponding values of 0.58 and 0.39 respectively for fish 50 mm long. As with salmonids less than 135 mm, cottid catchability increased with fish size (27 - 81 mm). However, in contrast to salmonids, catchability of cottids (dominated by C. bairdi) was unaffected by the amounts of large wood tested, but was affected negatively by the proportion of substrate that was cobble and boulder combined. For cottids 50 mm long the model predicted values of0.49 and 0.19 corresponding to cobble and boulder coverages of 50% and 100% respectively. These models demonstrate quantitatively what many experienced field biologists suspect, and conform to known refuge habitat preferences of each taxon. The results also demonstrate the danger of presuming abundance or relative abundance offish from catch data only, which can lead to serious under estimates of, for example, the effect of large wood on salmonid abundance.
Personnel and Keyword Links