9565 - Laserfiche WebLink

Home Browse Search

STATUS OF GREEN RIVER COLORADO PIKEMINNOW Colorado pikeminnow abundance estimates to a large degree, based on correlation analysis. Osmundson (2002) found poor agreement between abundance estimates and CPUE indices in the Colorado River, perhaps because of the lower precision of the abundance estimates. If one assumes that the CPUE and abundance estimates closely reflect each other, the Green River Colorado pikeminnow population in 2003 may have been reduced to a level at or near that in 1991. . Survival Rates The apparent survival for the average-length Colo- rado pikeminnow for 2000-2003 (0.65) was lower than that estimated for 1991-1999 (0.82). The more recent rate was also lower than the survival rates of adult Colorado pikeminnow in the Colorado River from 1991 to 1994. There, survival rates for adult fish were estimated as 0.85 and 0.86 using two different analytical techniques (Osmundson et al. 1997; Os- mundson and Burnham 1998). The lower recent survival estimate at least partially explains the decline in the abundance of recruit and adult Colorado pikeminnow in the Green River basin. Apparent survival rates reflect the fact that estimates of S are the joint probability of an individual fish's surviving and remaining in the reach so that it is available for recapture. In other words, estimates from these models cannot distinguish a fish that died in the study area from one that survived and moved to an unsamPled reach. Because the ISMP sampled only 23% of Colorado pikeminnow habitat, the estimated survival rate for 1991-1999 may be conservative (low). The survival rate obtained from 2000-2003 probably approaches true survival because sampling was spa- tially extensive and intensive, with the implication that survival rate differences among the two sampling periods are also conservative. This was supported even when relatively large numbers of Colorado pike- minnow were captured during our study and at other times in the 32-km-Iong Lodore Canyon of the Green River, one of the reaches that we did not sample (Bestgen and Crist 2000; Kitcheyan and Montagne 2005; K.R.B., unpublished collection data). Movement into and out of that reach occurred between annual abundance estimation sampling periods, and nearly all radio-tagged fish moved back to the reaches sampled for abundance estimation before that sampling began (Kitcheyan and Montagne 2005). Comparison of the shape of the length-dependent survival rate curve for 1991-1999 with that for 2000- 2003 suggests different relationships in the two periods. The 1991-1999 relationship shows that survival increased as a function of TL, eventually 1373 approaching an asymptote of 1. The asymptote was an artifact of the logistic relationship, as we do not reasonably expect 100% survival of larger fish. A key point is that a quadratic term added to the relationship had a positive (albeit nonsignificant) slope, which indicated that the survival of larger and older age- classes continued to increase. The addition of polynomial terms in the analysis of 2000-2003 data showed very high survival of very small Colorado pikeminnow, which was an artifact of the trajectory of the predictive model. We truncated the relationship at the lowest point for fish up to 350 mm because it seemed reasonable that the survival of smaller-bodied Colorado pikeminnow would continue to decline, as in the 1991-1999 period. More importantly, the relationship that described the survival of larger-bodied fish for the 2000-2003 period peaked at about 580 mm, after which the survival rate declined. This was the result of the negative quadratic term in the survival rate relationships. We fully investigated the phenomenon that larger fish had poor apparent survival in this study, but various model analyses consistently produced survival estimates for large fish that approached zero. The relationship was also nearly identical when refit without the four largest fish captures (2900 mm), which suggests that the few large individuals were not the cause of the negative quadratic term. Inspection of capture history records for the 18 largest Colorado pikeminnow captured (800 mm or larger) showed that none were recaptured among years. Because only recaptures across years contribute to survival analyses, this was deemed the cause for the declining survival relationships for larger fish for 2000-2003. We considered but did not find support for the hypothesis that the lower survival rates of Colorado pikeminnow were due to the effects of handling (including tagging) or electrofishing in the Green River basin in the period 2000-2003. The number of Colorado pikeminnow mortalities that we observed as a direct result of capture during this study was small (6 over the entire study), so high levels of electrofishing- or handling-induced direct mortality were not evident. Colorado pikeminnow recaptured in multiple passes within a year generally appeared to be in good condition, with no external signs of damage (e.g., bruising), handling-induced disease (e.g., fungus), or other ill effects. We have no data with which to address the effects of delayed mortality, except that electrofishing- captured fish implanted with radio transmitters for use in telemetry studies generally had high survival (Tyus 1990; Kitcheyan and Montagne 2005). In support of our observations of healthy fish being captured and recaptured in the field, independent