7145 - Laserfiche WebLink

Home Browse Search

in the 18-mile reach (Fig. 4), though none were found in the 15-mile reach. Flannelmouth sucker (C. latipinis) had consistently low correla- tions. Correlations for roundtail chub (Gila robusta) were inconsistent: larval abundance decreased with increased peak flow in the 18-mile reach but increased in the lower Gunnison, while in the 15-mile reach there was no correlation. For some species, high correlations were not biologically meaningful because values of three of the four data points were zero, thereby falsely producing a high r value when only one or a few individu- als were captured. This was the case for white sucker (Catostomus commer- soni) and black crappie (Pomoxis nigromaculatus) in the 18-mile reach. Because sample size (4 years) was low, few of the correlations were highly significant; r values had to be 0.900 or greater to be significant at the 0.10 level, 0.950 at the 0.05 level, and 0.990 at the 0.01 level. Speckled dace, bluehead sucker, and red shiner in the 15-mile reach, speckled dace, fathead minnow, and sand shiner in the lower Gunnison, and red shiner, fathead minnow, and Colorado squawfish in the 18-mile reach were signifi- cant (P < 0.10). Larval fish samples were also collected during 1990, but identification of these fishes is not yet complete. Additional data will be collected in 1991. If the correlation coefficients are maintained as more years of data are collected, the number of significant relationships will increase; for instance, for six years of data, r values of 0.729 or greater will be significant at the 0.10 level; >.811, at the .05 level. However, even with only four years of data, catch rates of larval Colorado squawfish were highly correlated (r - .931; r2 = .867; P - .069) with annual peak flow (Fig. 5). Thus we conclude with 93% confidence that 87% of the annual variation in abundance of larval Colorado squawfish in the 18-mile reach can be explained by variation in annual peak discharge. 18