Laserfiche WebLink
The "apparent" portion of apparent survival estimates also merits discussion. Such <br />estimates of survival may reflect the fact that estimates of S are the joint probability of an <br />individual surviving, and remaining, in the reach so that it is available for sampling. Thus, if a <br />fish moved from a sampled reach to an unsampled one and remained there, by definition, that fish <br />represents a mortality. We did not sample about 12% of critical habitat occupied by Colorado <br />pikeminnow in the Green River Basin. Those areas and others such as Lodore Canyon of the <br />Green River just upstream of the Yampa River have the potential to harbor additional Colorado <br />pikeminnow (Bestgen and Crist 2000; Kitcheyan and Montagne 2005). Although large numbers <br />of Colorado pikeminnow have been captured in such reaches, particularly Lodore Canyon, use of <br />those canyon areas occurred mostly in seasons other than when spring-time abundance estimation <br />sampling was conducted. Based on telemetry and tag recapture data, most Colorado pikeminnow <br />that moved to that area in portions of 2001, and all of 2002 and 2003, did so in summer. Most <br />fish were from reaches such as Island Park, other downstream portions of the middle Green River <br />reach, or even the Yampa River (Kitcheyan and Montagne 2005; K. Bestgen, unpublished <br />collection data RIP study 115), locations which were well-sampled during abundance estimates. <br />A few fish in 2001 moved into Lodore Canyon before summer in April and May. The few <br />Colorado pikeminnow captured in locations such as Whirlpool Canyon also appear to have <br />originated in areas included in abundance estimation sampling (K. Bestgen, unpublished data). <br />Supplemental electrofishing conducted during this study in Split Mountain Canyon produced only <br />a single Colorado pikeminnow, indicating that few fish resided there. <br />These movement and tag-recapture data for Colorado pikeminnow in canyons have two <br />main implications. First, apparent survival estimates may be relatively close to true survival <br />because fidelity to the original capture reaches was high. This is true because even though some <br />51