8177 - Laserfiche WebLink

Home Browse Search

classes were collected in the lower 80 Ian of the Green River (River Kilometer [RK] 0-80), and data for the 1989-94 year classes were collected in the lower 193 km of the Green River (RK 0-193). Sampling methodology was similar to that described in the ISMP (U.S. Fish and Wildlife Service 1987, 1994). Two backwaters (i.e., "primary" - first encountered; "secondary" - second encountered) were sampled every 8 km within each sample reach using beach seines, 4.6 m long by 1.2 m deep with 3.2 mm ace mesh. Two seine hauls were taken at each backwater and the fish in each haul were counted. All endangered fish were measured to total length (TL) and released. Some samples were preserved and the fish identified and enumerated at the Larval Fish Laboratory (Colorado State University, Fort Collins, Colorado) to evaluate the accuracy offield counts. Depths and dimensions of each backwater and seine haul were recorded. Water depth and temperature were recorded at all seine haul locations. Catch,.per-effort (CPUE), as number of fish per 100 m2, was computed for each seine haul. Efforts to enumerate backwaters with aerial photography and ground census during the fall and spring sampling periods were abandoned because of variability in flow that flooded or desiccated some backwaters and created others during a given sample period. I nstead, average backwater depth was recorded for every backwater sampled in Reach 3 during 1987- 92. and pooled geometric mean CPUE for fall and spring were compared for backwaters with average depths <30 cm, 30-60 em, 60-90 em, 90-120 em, and> 120 em. Catch- Per-U nit-Effort Geometric mean CPUE was used to compare densities of age-O Colorado pikeminnow. Geometric mean involves a log transformation ofCPUE's, primarily because catch data for Colorado pikeminnow in backwaters are not normally distributed. The geometric mean calculation method lIsed was as follows: CPUE for age-O Colorado pikeminnow w~ computed for each seine haul (number of age-Oil 00 m2). A '1' was added to the CPUE of each seine haul to account for zero catches. This adjusted CPUE was transformed to a natural logarithm. The transformed values were averaged by reach and year, and the antilog of the average minus '1 'yielded the geometric mean.. Standard deviation was computed from log-transformed values and the antilog taken to provide bounds around the geometric mean. The advantage of the geometric mean is that it considers non-normally distributed data and reduces variation by making the variance independent of the mean, allowing application of standard parametric statistics to comparative analyses. However, log transformations tend to weight higher CPUE's less. The computation is expressed as follows (Sokal and Rohlf 1987): GMcPE = antilog [(Vn) ~ In (CPE + 1)] - 1 where: GMCPE CPE = geometric mean catch per effort, catch per effort of age-O pikeminnow (age-Oil 00 m2), and total number of seine hauls (with and without target fish). n = 5