9710 - Laserfiche WebLink

Home Browse Search

408 100 -- 98 ~ is c <I.l '0 96 l;::: ID c 0 :e 94 .! <I.l Cl 92 90 CONNOLLY ET AL. (178) (143) (166) (122) (85)(72) (53) (59) (41) (23) (17) (13) (36) (17) (4) (16) Low High Downstream Low High Upstream _ Rattlesnake Protocol 1 H':Phrl Rattlesnake Protocol 2 _ Beaver Protocol 1 ~ Beaver Protoool 2 FIGURE 4.-Efficiency of detection of PIT-tagged fish (mean + SE) under two protocols for selecting fish passage events from data recorded by a three-array, six -antenna system at low and high flows in Rattlesnake and Beaver creeks. The nwnber of fish detection events is given in parentheses above each bar. Protocol 1 eliminated a fish passage event if the same fish was detected at any antenna within the previous 12 h or after the first passage event; protocol 2 extended this time interval to I month. minimal difference in detection efficiency was observed (Figure 4). Therefore, we adopted protocol 1 , which had the benefit of increasing the available sample size. Because we did not know the number of PIT -tagged fish that passed the interrogation system, we used an indirect method for determining estimates of detection efficiency. We used a three-array detection probability model (Appendix 1) in the USER program (Lady et a1. 2003~ to calculate the efficiency of detection of upstream- and downstream-moving fish at low and high flow for the 3 X 2 systems at Rattlesnake and Beaver creeks. The standard error and variance of this estimate were determined by the Delta method (Seber 1982;7-9; Appendix 2). Using the accepted fish passage events identified previously, we determined the detection efficiency of systems with lower numbers of antennas: three arrays with one antenna each (3 X I), two arrays with two antennas each (2 X 2), and two arrays with one antenna each (2 Xl). Because they were the original arrays at the Rattlesnake Creek site, we used the B (middle) and C (most downstream) arrays for the 2 X 2 and 2 X 1 systems (Figure 2). To determine whether to use the river-right or river-left antennas for the 3 X 1 and 2 X 1 systems, we calculated the percentage of detections fTom downstream passage events during low-flow periods that were recorded for each antenna and then used the dominant antenna from each array (Table 2), which proved to be the antennas associated with the thalweg where definitively present. To determine efficiencies of the reduced antenna systems, we used the detection events declared usable by protocol I, as with the 3 X 2 systems. We combined detection data. for cutthroat trout (in Rattlesnake Creek only), brook trout (in Beaver Creek only), and rainbow trout including steelhead (in both streams) for om efficiency calculations. Although other species were PIT tagged in each of the watersheds, we either did not detect these other species again (e.g., westslope cutthroat trout in Beaver Creek), or in a few cases,eliminated the detection events from seldom- seen species from our analysis. We did not believe that