These run-of-the-river fish were collected, tagged, and released at Rock Island Dam on the mid- Columbia River. Summer/fall chinook on the mid-Columbia arise from natural populations spawning in the mainstem Columbia and tributaries and from several hatcheries. They are considered to be "ocean-type" chinook, meaning that they migrate to sea as subyearlings shortly after they emerge from eggs in the spring. In contrast, spring or "stream-type" chinook migrate as yearlings after over-wintering in a stream. Although the taggers were not able to distinguish individual fish as being either spring or summer/fall, two temporally distinct distributions passed the dam each year. We followed the precedent of Chapman, et al. (1994) and identified the later peak as summer/fall chinook. Release groups over 7 consecutive days were lumped together to form cohorts.
The arrival distribution at McNary dam is very protracted, with a few individuals arriving through October, November and December. We decide to eliminate from the analysis fish arriving after October 7 (Julian day 280) for the following reasons: 1) it is likely that these late arrivals will either over-winter in the river (and thus would complete their migration as yearlings) or will die before reaching the estuary; 2) we are interested in the behavior of the active migrants; 3) mitigation efforts are likely to be directed at the 90-95% of the population that arrive during the summer months; and 4) the late arrivals strongly influence average travel times. Approximately 5% of the fish were censored. In the future, we will conduct similar analyses with later cutoff dates.
We applied six models of varying complexity to the data:
is the slope of the length-migration rate relationship and 
is the average length of cohort i
MIN after the
first year. (see the discussion below for further details)
Model details are provided by Zabel, et al. (1996).
The plots below show observed average travel times versus modeled average travel times to McNary Dam for all five models.
corresponds to observations at McNary Dam.Fall chinook have a more complex migratory behavior than spring chinook. Fall chinook tend to feed and grow as they migrate downstream, unlike spring chinook, which attain a larger size prior to migration and feed very little as they move downstream. As a result, it is much more difficult to model the travel times of fall chinook. In addition, their travel time distributions tend to be protracted, as evidenced by the large value for 
2 compared to spring chinook and steelhead.
The results of this last model and the importance of fish length in the other models demonstrates that for fall chinook, fish condition is an important factor in determining travel times. Fall chinook appear to have a more flexible life history during outmigration (as compared to spring chinook). Individuals chose between downstream migration and delaying migration to feed. More complex models than we have presented might be required to describe the observed behavior.
The results of the "year-effects" model demonstrate year-to-year differences in summer/fall chinook behavior. This variability may be attributed to several possible causes: 1) differential composition (e.g., proportion of hatchery versus wild fish) of the stocks passing Rock Island Dam; or 2) differential river conditions (e.g., food availability).
A debate exists as to the importance of river flow for fall chinook travel times. Giorgi, et al. (1994) concluded that flow is not an important factor in determining travel times while Berggren and Filardo (1993) arrived at the opposite conclusion. Our results (models 2 and 8) indicate a slight flow effect based on the
FLOW parameter and its relatively small standard error. With model 2, fish used approximately 35% of the river velocity, and this model accounted for only 25% of the variability present in the simple constant migration rate model.
Berggren, T.J., and M.J. Filardo. 1993. An analysis of variables influencing the migration of juvenile salmonids in the Columbia River basin. N. Amer. J. Fish. Manag. 13: 48-63.
Chapman, D., A. Giorgi, T. Hilman, D. Deppert, M. Erho, S. Hays, C. Peven, B. Suzumoto, and R. Klinge. 1994. Status of Summer/Fall Chinook Salmon in the Mid-Columbia River. Don Chapman Consultants, Boise, Idaho.
Giorgi, A.E., D.R. Miller, and B.P. Sanford. 1994. Migratory characteristics of juvenile ocean-type chinook salmon, Oncorhynchus tshawytscha, in John Day Reservoir on the Columbia River.
1 has units 1/date.