7960 - Laserfiche WebLink

Home Browse Search

Modde et al. Flaming Gorge Reservoir Wyoming Colorado Utah WYOMING UTAH ~,..~, COLORADO River Duchesne ~1 ~ ~mC River t White River Yampa River spawning site `35 Kilometers Escalante (Jensen) spawning site Figure 1. Map of upper Colorado River basin indicat- ing two documented spawning sites of the razorback sucker, Xyrauchen texanus. yet juvenile (Gutermuth et al. 1995) and subadult (U.S. Fish and Wildlife Service, unpublished data) razorback suckers have recently been collected in the upper Colo- rado River Basin. Larvae suspected to be razorback sucker have been collected below the Escalante site by Tyus (1987) in 1986, and confirmed collections were made in 1992, 1993, and 1994 (Robert Muth, personal communication). The previous inability to fmd juvenile razorback sucker in the Green River has been cited as evidence that recruitment is lacking (Wick et al. 1982). Similarly, despite successful reproduction in reservoirs of the lower Colorado River Basin (Mueller 1989), juve- nile razorback sucker have rarely been collected. Lack of survival through the early life stages has been attributed to low food availability, transport from the system, and predation (Hinckley et al. 1991). Although it has been 31 years since the closure of Flaming Gorge Dam, razorback sucker adults are com- monly collected at spawning sites in the middle Green River. In the upper Colorado River sub-basin, the razor- back sucker has declined rapidly and is now rarely col- lected (Modde et al. 1995). Because a razorback sucker Status of Razorback Sucker 111 may live 30 to 40 years (Hinckley 1983), -the ability of the middle Green River population to persist is in ques- tion. We examined 13 years of capture data and evalu- ated the status of growth, survival and factors affecting recruitment of the razorback sucker in the middle Green River population. Methods We analyzed all razorback sucker capture data collected between 1975 and 1992 in the Green River by the U.S. Fish and Wildlife Service, Colorado River Fish Project (CRFP). These data are inclusive of and additive to data reported by Tyus (1987), Lanigan and Tyus (1989), and Tyus and Karp (1990). Fish were collected by elec- troshocking primarily in the spring months of April through June at sites fish have been known to aggregate between the spawning and post-spawning periods. All fish captured were marked with either dorsally attached Carlin tags (1975-1989) or internally inserted passive in- tegrated transponder [PIT] tags (1990-1992). Length Distribution, Growth, and Survival Length-frequency data were grouped into 3-year inter- vals, with the last group including the last four years of the study period. Individual growth rates were esti- mated by dividing the difference in the total length at first and last capture of an individual by the number of days between the two events. Mean growth of fish per year was determined by multiplying the average daily growth by 365 days. Testing for a trend in mean length of fish per year was done using the weighted least- squares linear regression model (SAS 1987). Because spawning and post-spawning aggregation areas were re- peatedly sampled, fish captured on multiple dates were not considered recaptures unless they were collected at least two months after the previous capture. We used Cormack Jolly-Seber (CJS) models (Lebreton et al. 1992) to estimate survival rates from capture-recap- ture data. The program RELEASE was used to compute goodness of fit (Burnham et al. 1987) of the basic, time- specific CJS model. The program SURGE (Lebreton et al. 1992) was then used to fmd the best model of a set of four models reasonable to consider. Analysis methods, including information-theoretic model selection (AIC, Akaike 1973), and the specifics of the models used herein are discussed in Lebreton et al. (1992), Burnham and Anderson (1992), and Anderson et al. (1994). The CJS models are based on capture probability (p) and apparent annual survival probability (v~). In general, ~ = S(F)(1 - L ), where F (fidelity) is the probability that a fish returns (on an annual basis) to the river reach being sampled and L is annual probability of tag loss. If fidelity is 1, then only tag loss biases annual survival rate Conservation Biology Volume 10, No. 1, February 1996