WSP12650 - Laserfiche WebLink

Home Browse Search

001560 SAMPLING VARIABILITY FOR RARE RIVERINE FISHES 1645 including the reach near the LCR, probably because of alterations caused by Glen Canyon Dam (124km upstream of the LCR confluence) and non-native species inter- actions (Minckley, 1991). Trammel nets were 22.9m long, 1.8 m deep, and consisted of two outer walls of 30.5 cm multifilament netting and one inner wall of 2.5 cm multi- filament netting. Nets were primarily tied to the shore and stretched across the river channel, but were ocCasionally suspended in the mid-water column of the river (Valdez & Ryel, 1995), usually in deepwater eddies or other areas previously known to be frequented by humpback chub, flannelmouth suckers and bluehead suckers (Hoffnagle et aI., 1999). To minimize fish mortality, nets were checked about every 2h (median = 2.21 h, range: l'5-3'35h), but remained in the water throughout several sets (total soak time mean: 4,9 h, range: 1.6 to 12.1 h). Each 2 h check was treated as one sampling effort. Electrofishing was conducted along the shoreline using a 5 m Achilles inflatable boat equipped with pulsed-DC current using a Coffelt Mark XX Complex Pulse System (CPS) used to minimize injuries to native fishes (Valdez & Ryel, 1995; Hoffnagle et aI., 1999). Output typically ranged from 200 to 250 V and 8 to 10 A (Valdez & Ryel, 1995). Sampling stations consisted of only one general habitat type (e.g. sand bar) and therefore sampling times were variable because of different habitat lengths. Variation in CPUE tends to increase with sample time duration (Miranda et al., 1996). To minimize this effect, only electrofishing samples that were between 0.15 and 0.75h in duration and trammel net samples between 1.5 and 3,5 h were used in the analysis. Catch per unit effort (CPUE) was calculated as the number of fish ~200mm Lr (for each species) collected per hour of trammel netting or electrofishing. A minimum size of 200 mm was used because this is the minimum size of adult humpback chub defined in the U.S. Fish and Wildlife Service recovery goals (USFWS, 2002). This same size was used for flannelmouth and bluehead suckers for consistency. To ensure that the analysis included sampling from both all gear types and diel periods, the samples were grouped by seasons (February, March and April: spring; May, June and July: summer; August, September and October: autumn). November, December and January were not used in the analyses because of limited sampling conducted during these months. Samples were also categorized by diel periods. Night samples were samples started from 1 h prior to sunset to I h prior to sunrise and day samples were samples started at 1 h prior to sunrise to 1 h prior to sunset. In addition, results were limited to the years 1992 and 1993, in which all gear types were used in all diel periods and seasons. STATISTICAL ANALYSIS Trammel net and electrofishing relative variability in CPUE were compared by using the coefficient of variation (CV, CV = 100S.D. X-I; Zar, 1996) calculated for each gear type, season, diel period and year for the entire 8 km area. An analysis of covariance (ANCOV A) was then used to determine the effects of gear type, diel period and season (with year as a covariate) on mean CV of CPUE for each species. Prior to analysis, homogeneity of variances was determined by Levene's test (Zar, 1996). The sample size needed to detect a 10 and 25% difference in mean CPUE for each gear was estimated using methods described by Allen et al. (1999) and Tate et al. (2003). To estimate the required sample sizes, the mean and S.D. estimates of CPUE for both years combined were used. Spearman rank correlations were used to compare 'paired' CPUE indices for trammel nets and electro fishing. This was done by using only trammel nets and electro- fishing stations that were conducted during the same date, year, month and hour at the same 0.16 rkm. In this analysis, only samples that did not have 0 CPUE for both gears were used. Although these samples were not intentionally 'paired', the samples were collected at the same time and location and offered insight into gear difference by minimizing spatial and temporal differences in sampling. To determine if the proportion of fishes collected ~200 mm Lr differed between gear types, diel periods and among seasons for each species, a likelihood-ratio X2 test (SAS, 1996) was used. In this analysis, only the gear type, season and diel period combinations where at least 10 fish of each species were collected were used. When the low numbers of fishes collected precluded the analysis by diel period and season (which occurred in all but one of the flannelmouth (!;) 2004 The Fisheries Society of the British Isles, Journal of Fish Biology 2004, 65, 1643-1652 (No claim to original US government works)