Final distribution

Final geographic location assignments for each fish were based on all available radiotelemetry, PIT-detection, reported harvest, and transmitter recovery data accumulated during the study [e.g., 45, 46]. Fish assigned to tributaries included those that were reported harvested in tributaries, entered hatcheries, were last detected on tributary radio or PIT antennas, or were recovered in tributary spawning ground surveys. Fish last detected in the main stem Columbia or Snake rivers included those that were reported harvested and those with unknown main stem fates. Most fish assigned unknown fate were last detected at antennas in tailraces or at dams and primarily represented presumed mortalities and apparent unreported harvest, rather than transmitter failure (based on concurrence of radio- and PIT telemetry records). The only excluded data were downstream movement records from presumed post-spawn fish (e.g., for steelhead kelts).

We used the final distribution data for two purposes. First, the proportions of fish from the full sample that entered spawning tributaries provided an indirect measure of dam passage effectiveness (i.e., the fish was able to complete migration after passing one or more dams, [20, 23]. Second, we assessed the final distribution of each subset of fish that approached–but ultimately did not pass–each of the eight dams. There are several potential reasons for non-passage, including some that are not directly related to the fishways, such as downstream predation, fisheries harvest, and the location of natal sites. Many adult salmon and steelhead detected at Columbia and Snake River dams, for example, have migrated upstream past their natal sites and must move back downstream to complete homing [e.g., 87, 88]. This natal site ‘overshoot’ behavior, as well as some harvest, should probably not be interpreted as dam passage failure.