9710 - Laserfiche WebLink

Home Browse Search

PERFORMANCE OF PIT TAG INTERROGATION SYSTEMS 411 efticiency varied with flow level (ANOV A, flow X design interaction: P = 0.004), prompting us to run separate tests by flow level. These tests showed that the 3 X 2 design had higher detection efficiencies for upstream-moving fish than the 2 X I design and that the 2 X 2 and 3X I designs did as well as the 3 X 2 in high flow, whereas only the 2 X 2 did as well as the 3 X 2 design in low flow (AND V As and Tukey's tests: P < 0.05). For the 2 X 1 system, the precision of detection efticiency for upstream-moving fish during high flow was much poorer (Rattlesnake Creek, CV = 55%; Beaver Creek, CV = 79%) compared with any other design we tested (all other configurations, CV < 9%). For downstream-moving fish, all four designs had detection efficiencies with good precision (CV < 8%). The observed differences in detection efficiencies for combinations of flow level and direction of fish were complex, but proved to be important to consider if faced with limitations in number of antenna~ or arrays that can be placed at a given site. Discussion The high PIT tag detection efficiencies of 96% to almost 100% that we achieved for PIT-tagged salmonids passing our 3 X 2 interrogation systems in Rattlesnake and Beaver creeks can largely be attributed to a redundancy of arrays that maintained detection fields over most of, but not all, the stream width and water colunm.Stream stage height was a factor, our systems doing better in low flow than high flow, but the difference was limited to a few percentage points that may be biologically meaningless to many applications, depending on the number of fish moving through the system and the value of each detection evenl to the study being conducted. Stream stage height is probably a major factor in the potential for fish to escape detection. What constitutes. high flow will be site dependent. We used tag-detection range to determine the division between low and high flow. The number of fish passage events recorded during high flow was somewhat low, so we did not break flow level into additional categories. However, we did not detect fish when stage height exceeded 1.94 m in Rattlesnake Creek and 2.03 m in Beaver Creek. We do not know whether this was a result of the interrogation system becoming less efficient, whether fish had a decreased tendency to move at high flows, or both. Because the distinction between low and high flow was based on water depth and the read distance of the weakest PIT tag used in the watershed (BE-type in Rattlesnake Creek, ST-type in Beaver Creek), this probably introduced a bias into detectability. Not only did the newer tags offer increased read range. but they also increa~ed the chance that they would be detected at _ Upstflitam c:::J f>t.Nmgtrc"'" 100 l 00 ~ " " 'il ~ 96 " 0 "" ~ .... ~ S2 00 Low 1-119" _C!eOl< Low High Ral:tft~&nake ClOOk FIGURE 6.-Efficiency of detection of PIT-tagged fLsh moving upstream or downstream (mean + SE) past PIT tag interrogation systems consisting of three alT'dYs and six antennas in Ratdesnake and Beaver creeks under two flow conditions: low (SO.l4 mOfs in Rattlesnake Creek and S0.57 m3fs in Beaver Creek) and high (>0.14m3fs and >0.57 m3fs). a wider range of orienl'ation to an antenna's interroga- tion field and to stronger EMF-interfering noise levels. These differences in tag models could have differen- tially contributed to an underestimate or overestimate of detection efficiency at low- and high-flow levels (Horton et al. 2007). New models of PIT tags are likely to be available in the future and readily adopted by users, especially when older models are pha<ied out of production and become unavailable. Based on the need to eliminate bias of estimates for detection efficiency. researchers and managers may need to anticipate these changes in their study designs. The arrays with hybrid antennas clearly outper- formed those with pass-by antennas for detecting PIT -tagged fish moving downstream during high flow, but the opposite was true for detecting fish moving upstream in high flow. No distinction between antenna types was evident for detection of fish moving during low flow. These findings may be important to researchers faced with a choice among the type of antennas and number of arrays to use because of, for example, lack of funds or limitations imposed by the site. This choice would probably be more effective if based on the configuration that will probably perform best for the fish behavior that it is most desired to track and for the stream characteristics during the period of interest. By using a mix of antenna types, but limiting the use of the more flow-dependent hybrid type to a single array, our systems appeared to have been a good combination for maintaining high detection efticiency during both low and high flow for downstream- and upstream-moving fish. The hybrid-type antenna~ actively moved up and down in the water column and did so regularly during