DART PIT Tag Adult Fallback Adjustment Rates

The following algorithm is used to determine a successful ascension at a dam:

1. PIT-tag coil detections are sorted in chronological order for each tag code.
2. Consecutive detections spaced farther than 6 hours apart are determined to be separate attempts of ascension.
3. For each attempt, if the last two detections are in an upriver direction (i.e. the last coil the tag was detected on was upriver of the next unique previous coil), an ascension is determined to be successful. Same coil detections are ignored for this determination.
4. Single coil detections are ignored.
5. Once a fish has been determined to have successfully ascended at a ladder, any following successful ascensions are categorized as re-ascents.

The proportion of unique fish is expressed as the ratio of number of unique PIT-tagged fish passing over the ladder(s) divided by the total number of estimated ascents. A fish may ascend the dam one or more times, contributing to the total ascent count. The fallback rate is the complement to the proportion of unique fish ascending. For example, a fallback rate of 0.02 means 2% of the total adult counts resulted from counting previously observed fish two or more times.

This expression of fallback is a retrospective analysis. In other words, of the ascents recorded to date, how many of these ascents were of fish previously counted. A seasonal ladder count times the proportion of unique fish would provide an estimate of the actual abundance of fish ascending the ladder, corrected for recounts due to fallbacks.

The algorithm to determine ascension counts is based on radio-telemetry research at the Columbia dams performed by Brian Burke, Northwest Fisheries Science Center, NOAA, 2003-2004.

note: The estimates given for the BO4 site at Bonneville Dam are a combined BO2/BO3/BO4 system estimate. Since all fish detected at Cascades Island ladder (BO2) and Washington Shore ladder (BO3) must exit through the Washington Shore Vertical Slots detectors (BO4) and since nearly 100% of the fish passing BO4 are detected the estimate for BO4 can be taken as representative for the sytem as a whole. This does not apply to the 2004 migration year since the BO4 system did not go online until Feb. 2005.