9633 - Laserfiche WebLink

Home Browse Search

Colorado Pikeminnow Distribution 463 methods used during the study period were those of the Interagency Standardized Monitoring Program (lSMP) , wherein the first two backwaters larger than 30 m2 in area and deeper than 0.30 m encountered in each 8-km reach were sampled with two nonoverlapping seine hauls (McAda et al. 1997). The backwaters sampled in each year were not necessarily the same. Abundance was re- ported as the catch per unit effort (CPUE), reported in the units of number of age-O fish per area seined. We assembled the ISMP data (McAda 1993; Valdez and Cowdell 1999), assigned backwaters to the geomorphic reaches that we had defined, and calculated CPUE for each backwater and each reach. Sampling between 1990 and 1995 did not occur in reaches A, B, or C. The mean CPUE and standard error of the CPUE for each reach were based on the total number of backwa- ters sampled per reach. Mean CPUE is routinely used by researchers in the upper Colorado River basin as an in- dex of the abundance of age-O pike minnow in backwa- ters (McAda et al. 1997). The standard error of the arithmetic mean CPUE was large because of the large variability in the number of fish collected in sampled backwaters, the high proportion of seine hauls with no fish, and the small number of backwaters sampled. We used the arithmetic mean rather than geometric mean CPUE reported elsewhere (0. S. Fish and Wildlife Service 1987), because the geometric mean reduces variance, and we were interested in the relationship between the variance of field data and error of model predictions. Results The six years of field data span a range of flow and drift conditions. Runoff was relatively large in 1993 and 1995 and was least in 1990 and 1992. Larvae began entering the Green River earlier in 1995, in relation to the time of flood recession, than in other years, and the discharge was about 250 m3s -1 at the time of drift, al- most three times more than in other years (Figure 4). Runoff was nearly as large in 1993, but larvae entered the Green River after recession to base flow. The esti- mated number of drifting larvae in 1991, 1992, 1993, and 1995 was between 500,000 and 700,000. In 1994, there was an order of magnitude fewer drifting larvae than in average years, and there were about twice as many in 1990 as in average years. Thus, 1990 was a year of low discharge and a large number of drifting larvae, 1992 was a year oflow discharge with an average number of drifting larvae, 1991 had average discharge and av- erage drift, 1994 had average discharge and very low drift, and 1993 and 1995 had high discharge and average drift. The CPUE of age-O pike minnow in backwaters in September varied from year to year and from reach to reach. The prevailing view of fishery scientists is that the Uinta Basin, especially reach H, provides critically im- portant backwater nursery areas, and there is some support for this view. Reach H in 1993 had the largest CPUE in any reach in any year and also had higher CPUE in 1992 than elsewhere (Figure 5). However, it is difficult to detect longitudinal patterns in the distribu- tion of age-O fish in other years, because the reach standard errors are very large. There is uncertainty in linking backwater CPUE of age-O fish with the estimated number of larvae that entered the study area in the same year. CPUE for the entire study area was least in 1994 and largest in 1991. The value for 1994 is consistent with the low estimates of the number of larvae entering the study area in that year, but the high CPUE value for 1991 contrasts with the estimate that twice as many larvae entered the study area in 1990 (Figure 4). Other sampling in 1990 and 1991 does not resolve this apparent inconsistency. Val- dez and Cowdell (1999) sampled 281 age-O pike minnow in fall 1990 and 526 in fall 1991 in reach H, indicating that 1991 backwater populations in that reach were larger than in 1990. However, sampling of this cohort the next year as juvenile (age 1) fish suggests the op- posite because 41 juveniles were collected in fall 1991 and none in fall 1992 (McAda, unpublished data). Thus, estimates of the age-O population in backwaters are imprecise, are sufficiently large to mask most longi- tudinal trends in the distribution of the population, and are not necessarily linked to the number of larvae that drift downstream in the same year. The uncertainties in linking larval drift and occupancy in backwaters with dam operations are therefore many. A dynamic model offers an attractive tool with which to evaluate how well the field sampling program describes the drift process, some of the factors that control backwater. populations, and the relationship of these processes to dam opera- tions. Dynamic Model of Drift and Transport into Backwaters Overview One value of model building is to provide a template with which to organize available information about complex systems and to assess information gaps and