This report describes a study of survival and migration behavior of juvenile coho salmon in the Klamath River relative to discharge at Iron Gate Dam in 2006. This was the second year of a multi-year study with the goal of determining the effects of discharge at Iron Gate Dam on survival of juvenile coho salmon downstream. The study was a collaborative effort among U.S. Geological Survey (USGS), U.S. Fish and Wildlife Service (USFWS), and the Yurok and Karuk Tribal Fisheries Departments. The goals of the study included: 1) estimating the survival of wild and hatchery juvenile coho salmon in the Klamath River downstream from Iron Gate Dam, 2) determining the effects of discharge and other covariates on their survival and migration, and 3) determining if fish from Iron Gate Hatchery could be used as surrogates for the limited source of wild fish. The major findings of the study in 2006 include:

River discharges during the 2006 study period (4 April through 21 July 2006) were among the greatest on record. Average daily discharge at Iron Gate Dam was 3,956 cubic feet per second (cfs) and ranged from 997 to 10,300 cfs. Discharge at Iron Gate Dam was positively correlated with discharges of tributaries downstream due to the above average water year and frequent occurrence of spill at Iron Gate Dam. Average daily discharge near the estuary was 25,789 cfs and ranged from 4,740 to 50,600 cfs. This study was based on hatchery fish taken directly from a tank at Iron Gate Hatchery and wild fish captured in a rotary trap on the Shasta River. Releases of both groups began on 4 April when the catch of wild fish in California Department of Fish and Game‟s Shasta River rotary trap increased, but trap catches varied throughout the study period, resulting in differences in release dates of hatchery and wild fish. A total of 211 hatchery fish were released from 4 April through 26 May. Wild and hatchery fish released on a regular schedule between 25 April and 16 May 2006 were used in comparisons of the survival and migration of hatchery (N = 120) and wild (N = 162) fish. Additional analyses were performed using hatchery fish from all dates.

The data and models did not support clear differences between survivals of hatchery and wild fish released on common dates, so estimates of reach survivals were made after pooling these data. Estimates of survival were lowest in the Iron Gate Dam to Scott River reach (0.813) and greatest in the Salmon River to Trinity River reach (1.000). The overall survival from river kilometer 309 (Iron Gate Hatchery) to river kilometer 33 was 0.653 (95% CI 0.578 to 0.729). Estimates of survival based on all hatchery fish releases were similar to those from release dates common to hatchery and wild fish and are similar to those in other river systems over similar distances. The migrations of hatchery and wild fish were different in the uppermost sections of the study area and were similar thereafter. A lag between release and migration, primarily upstream from the Scott River (river kilometer 234), was present in hatchery fish to a greater extent than in wild fish, resulting in differences in migration rates. Fish from both origins spent more time between release and the Scott River than in individual reaches downstream, and this was the only reach in which travel times of fish increased as discharge decreased. The travel times of hatchery and wild fish between sites were statistically similar downstream from Indian Creek (river kilometer 178).

There were differences and similarities in the analyses of the effects of covariates on survivals of hatchery and wild fish. The models of covariate effects based on hatchery and wild fish released on common dates indicated effects on wild fish survival that were not supported in data from hatchery fish. However, when the entire suite of hatchery fish releases were used the results of the analyses were similar to those based on wild fish. In both instances the effects of temperature and release date were primarily in the first reach, the reach fish of both origins spent most of their time within. The signs of the effects of these covariates differed among the fish origins (negative for wild and positive for hatchery fish), presumably due to differences in their migrations in the first reach. The effects of dam discharge on survivals of hatchery and wild fish were generally similar (positive relation), and the effects on hatchery, and to a lesser extent wild, fish were largely downstream from the Scott River. This is likely due to the prolonged residence of the naïve hatchery fish, and to a lesser extent, migrant wild fish between release and the Scott River. Inasmuch as the differences between hatchery and wild fish we observed were likely those of migrants vs. non-migrants, the use of hatchery fish captured as they are migrating downstream, rather than those directly from hatchery tanks (i.e., naïve), may improve similarities between hatchery and wild fish in future studies. The data and models used in 2006 do not support the use of naïve hatchery fish as surrogates for migrant wild fish in determining the effects of discharge on survival upstream from the Scott River. This conclusion is based on the different effects of covariates in this reach that were likely attributable to the differences in hatchery and wild migration behaviors in this reach.

The results of this second year of research provide insight to the migration and survival of hatchery and wild juvenile coho salmon in the Klamath River, but the results are from a single unusual water year. The results may be different during other water year types. The current information supports a positive relation between discharge at Iron Gate Dam and survival of juvenile coho salmon downstream, but additional data should be used to refine this relation. Discharge at the dam was correlated with discharges of Klamath River tributaries during this above average water year. The data and models from the 2006 study provide the first estimates of survival of these fish in the Klamath River and can be used with data from years with other water year types to examine the effects of discharge on survival. This will only be possible over a period of years in which the correlations between discharge and other factors, such as water temperature and date, are diminished. An experimental approach in which discharges are varied at Iron Gate Dam is the most direct method to determine if survivals are affected by discharge, but this may not be feasible given the limited storage capacity of the project.