Natural subyearling fall Chinook salmon in the Clearwater River emerge later, grow slightly slower, and become subyearling smolts later in the year than those in the Snake River.  Natural fall Chinook salmon produced in the Clearwater River frequently exhibit the reservoir-type life history (e.g., 60-85% without accounting for winter passage) and it is assumed that these subyearlings commonly pass dams undetected during the winter while the Passive Integrated Transponder (PIT)-tag monitors are not in operation.  Ignoring detections of yearlings that occur in the spring following release, a typical “survival” estimate to the tailrace of Lower Granite Dam (LGR) made for Clearwater fish made by use of Cormack-Jolly-Seber (CJS) methods might be 16%.  This estimate is a product of the probability of migrating as a subyearling smolt and passing the dams when the PIT-tag monitors are operated (e.g., 40%) and the probability of surviving to the tailrace of Lower Granite Dam (e.g., 40%).  That is, 16% = 40% x 40%.  The joint probability estimate in this example is 24 percentage points lower than actual survival.
The Comparative Survival Study (CSS) study design calculates smolt to adult returns (SARs) for inriver migrants (i.e., those that are not transported) as the number of returning adults divided by the number of smolts estimated to have passed Lower Granite Dam.  The number of smolts estimated to have passed Lower Granite Dam is called a "Lower Granite Equivalent."  For simplicity, let's use the 16% CJS estimate of survival described above to illustrate bias in the calculation of SARs.  If we released 1000 smolts the Lower Granite Equivalent would be 160 (i.e., .16 x 1000).  If 16 adults returned, the SAR would be 10% (i.e., 16/160).  In reality, the true survival was 40%.  Therefore, the true Lower Granite Equivalent was 400 (i.e., .4 x 1000) and the true SAR was only 4%    In contrast to the inriver group, the Lower Granite Equivalent for the transport group is known with 100% certainty.  For example, if 1000 smolts are released and 400 are detected as they are routed to a barge, the Lower Granite Equivalent will be 400.  If 16 adults return the SAR for the transport group would be 4%.  So, under the CSS study design the inriver group (SAR = 10%) wins over the transport group (SAR = 4%) because the Lower Granite Equivalent for the transport group was biased low.
Note that this is example is oversimplified.  However, no one has modeled a range of "what if" scenarios to illustrate this situation for managers.  This is an important objective of our study.   The field tasks that follow will move us toward better estimating LGR equivalents.

During this study, we will assume that fish from the Clearwater will represent late migrants from the lower Hells Canyon reach of the Snake River.  This will allow us to focus our work on the Clearwater.  Migratory delay is one mechanism that contributes to a fish adopting a reservoir-type life history.  One related question is where does migratory delay occur?  The answer is significant because if fish delay near the confluence as we observed in 2006, then these fish will not benefit from management actions such as summer spill and transportation.  The timing of delay and the reinitiation of downstream movement will be of interest to managers in regard to these actions. Radio telemetry ( RT)  and acoustic telemetry (AT) tasks will identify when and where fish delay.  A related question will address the magnitude of delay in a release group and the fate of those fish.  It is currently assumed that fish that are not detected at Lower Granite are mortalities in survival estimation.  The first step is to determine the fates of fish in the lower Clearwater and upper reservoir to determine how much mortality is incurred by fish that delay in this area.  The mark-recapture data collected by RT will address this need.  The next step is to determine the fate of fish that reinitiate downstream movement in late summer and fall and determine the survival and fate of those fish.  AT will address this need by determining how many fish survive to reach the forebay of LGR and of those fish, how many will pass the dam, how many will stay in the forebay, and how many will not survive.  This is a critical piece of information needed to estimate LGR equivalents for SAR determination.  Finally, we need to relate the data we collect to potential environmental data that might influence the behaviors we see.  People always ask whether flow augmentation has something to do with the reservoir-type life history.  Putting hydraulic data together with the fish data will further our understanding of why fish delay and reinitiate downstream movement.

The objectives for the PNNL portion of Project 2002032 in 2007 are to:

1. Produce detection histories for juvenile Snake River fall Chinook salmon tagged with JSATS acoustic transmitters in the lower Clearwater River and Lower Granite Reservoir.

2. Collaborate with study team to determine passage timing, survival, and fate of  juvenile Snake River fall Chinook salmon tagged with JSATS acoustic transmitters in the lower Clearwater River and Lower Granite Reservoir.

3. Document the temperature profiles in the lower Clearwater River and Lower Granite Reservoir.

4. Provide an indication of the seasonal presence and longitudinal extent of thermal stratification within Lower Granite Reservoir.