Stream restoration is a common conservation strategy for salmonids, but evidence that it has improved watershed scale habitat conditions or fish population abundance, survival, or spatial distribution is rare. We employed an intensively monitored watershed approach to measure the responses of both the habitat and Coho Salmon Oncorhynchus kisutch to a culvert replacement (2002) and a three-phase large woody debris (LWD) addition (2007, 2009, and 2016) in Little Anderson Creek (LA), a small stream in western Washington. Following the replacement of a fish barrier culvert with a channel-spanning bridge near the mouth of the creek, average smolt abundance for Coho Salmon more than tripled (910 vs. 275), demonstrating a significant increase in production capacity. In evaluating the habitat responses to the LWD placement, a series of mixed effects models indicated a modest, reach-scale increase in LWD density and a statistically divergent trend in spawning gravel and pool frequency, whereby both metrics increased at a slightly lower rate in treated sites than in reference sites. However, substantial interannual habitat variation appeared be driven by regional processes rather than treatment effects, as habitat change among years was consistent between the treatment and reference watersheds. Although we observed more Coho Salmon smolts after the 2009 LWD placement (364 fish), this difference was not statistically significant (P = 0.22) and neither parr abundance nor parr-to-smolt survival increased. Also, we did not detect a change in density-dependent growth dynamics or the spatial distribution of redds. Our results suggest that increasing accessibility through barrier removal can provide an immediate boost to freshwater fish production, but measurably improving population performance by adding LWD requires greater treatment magnitudes or more time for response than was examined in our study. Our ability to detect a fish response also may have been limited by low population abundance, which appeared to be below carrying capacity in most years following the LWD treatment.

We used PIT tags implanted in juvenile Oncorhynchus mykiss to monitor movement into and out of two coastal Washington State rivers, East Twin River and West Twin River. Movement patterns revealed at least 18 life histories of steelhead O. mykiss with variations in age and seasonal migration of juveniles, juvenile use of the ocean prior to migration, years spent in the ocean, season of adult return, and iteroparity. While most migrants left the river in their first fall or winter, we did not detect any returning adults from these age-0 migrants. Adults were only produced from age-1 and older migrants, of which most were age-2 spring migrants that returned after two summers in the ocean. Our results indicated a positive relationship between fish length at tagging and the probability of being detected as amigrant, while the probability of a migrant leaving at age 1 and older decreased with increasing length at tagging among fish that were detected as migrants. We hypothesize that fish attaining a large enough size early in life to survive over the winter but not big enough to trigger migration at age 0 were more likely to remain in the river to become age-1 migrants, which were more likely to produce a returning adult steelhead. We also found evidence that density-dependent growth may influence juvenile steelhead migration patterns and production of migrants as evidenced by increasing contributing-adult steelhead escapement being negatively related to average cohort body size, probabilities of fish being detected as migrants, and production of age-1 and older migrants. We anticipate that the findings of this study can be used to inform the development of steelhead recovery strategies for East Twin and West Twin rivers, which have experienced recent declines in adult returns much like other North Pacific Ocean stocks.

Recent improvements in tagging technology allow for the examination of the migration of individual fish, the detection of previously unidentified life histories, and the detailed examination of factors affecting growth, migration, and survival. Using passive integrated transponder tags and instream readers installed near tidewater, we examined the migration, growth, and survival of 18,642 juvenile coho salmon Oncorhynchus kisutch in two small western Washington rivers from 2005 to 2009. In most years, more than 50% of the juvenile coho salmon from a given brood year migrated to sea between 1 October and 31 December (fall migrants). These fall migrants were significantly smaller at tagging than fish that migrated between 1 January and 30 June (spring migrants) but were similar in size to fish that were never detected after tagging and assumed to have died. Annual coho salmon survival estimates from tagging to out-migration ranged from 31% to 40% for fall and spring migrants combined but from 5% to 15% for spring migrants only. The best fitting regression models indicated that survival differed by river and year and was negatively correlated with tagging location (river kilometer) and positively correlated with fish length: larger fish and those tagged lower in the watershed were more likely to survive. The number of days juvenile coho salmon spent in freshwater before migrating to sea was positively correlated with tagging location, fish length (mm), and habitat depth (m) and negatively with density (coho salmon/m(2)). Our results suggest that fall or early winter migration is a common life history for juvenile coho salmon that is driven in part by fish size and location in the watershed. The exclusion of fall migrants may lead to underestimates of the total number of migrants and parr-to-smolt survival.