To develop a habitat restoration strategy for the 8,270‐km2 Skagit River basin, we estimated changes in smolt production of coho salmon Oncorhynchus kisutch since European settlement began in the basin, based on changes in summer and winter rearing habitat areas. We assessed changes in coho salmon smolt production by habitat type and by cause of habitat alteration. We estimated that the coho salmon smolt production capacity of summer habitats in the Skagit River basin has been reduced from 1.28 million smolts to 0.98 million smolts (–24%) and that the production capacity of winter habitats has been reduced from 1.77 million to 1.17 million smolts (–34%). The largest proportion of summer non‐main‐stem habitat losses has occurred in side‐channel sloughs (41%), followed by losses in small tributaries (31%) and distributary sloughs (29%). The largest loss of winter habitats has occurred in side‐channel sloughs (52%), followed by losses in distributary sloughs (37%) and small tributaries (11%). By type of impact, hydromodification (diking, ditching, dredging) associated with agricultural and urban lands accounts for 73% of summer habitat losses and 91% of winter habitat losses. Blocking culverts on small tributaries account for 13% of the decrease in summer habitat and 6% of the decrease in winter habitat. Forestry activities account for 9% of summer habitat losses and 3% of winter habitat losses. Limitations of the analysis and implications for developing a habitat restoration strategy are discussed.

Consideration of fundamental channel processes, together with map-based and field investigations, indicates that stream channel type influences salmonid spawning distributions across entire channel networks and salmonid abundance within channel reaches. Our analysis suggests that salmonid spawning patterns in mountain drainage basins of the Pacific Northwest are adapted to, among other things, the timing and depth of channel bed mobility. We hypothesize that because the bed of pool-riffle and plane-bed reaches scours to a variable fraction of the thickness of alluvium, survival to emergence is favored by either burying eggs below the annual scour depth or avoiding egg burial during times of likely bed mobility. Conversely, annual mobility of all available spawning gravel in steeper step-pool and cascade channels favors either adaptations that avoid egg burial during times of likely bed mobility or selection of protected microhabitats. Consistent with these expectations, we find that salmonid spawning distributions track channel slope distributions in several west-slope Pacific Northwest watersheds, implying that spatial differences in channel processes influence community structure in these rainfall-dominated drainage basins. More detailed field surveys confirm that different channel types host differential use by spawning salmonids and reveal finer-scale influences of pool spacing on salmonid abundance.

We describe six habitat types for large rivers (>100 m bank‐full width), including pools, riffles, and glides in midchannel and bank edges, bar edges, and backwaters along channel margins. Midchannel units were deeper and faster than edge units on average. Among edge habitat types, backwater units had the lowest velocities and contained complex cover consisting mainly of wood accumulations and aquatic plants. Banks and bars had similar velocity distributions, but banks had more complex cover such as rootwads and debris jams. Because sampling of juvenile salmonids was ineffective in the midchannel units (electrofishing capture efficiency was low, and the units were too deep and fast to snorkel), we focused our sampling efforts on juvenile salmonid use of edge habitats during winter, spring, and late summer. Densities of juvenile Chinook salmon Oncorhynchus tshawytscha and coho salmon O. kisutch were highest in bank and backwater units in winter, whereas age‐0 and age‐1 or older steelhead densities were highest in bank units in winter. In summer, only coho salmon densities were significantly different among edge unit types, densities being highest in banks and backwaters. Microhabitat selection (velocity, depth, and cover type) by juvenile salmonids mirrored that in small streams, most fish occupying areas with a velocity less than 15 cm/s and wood cover. Among ocean‐type salmon, Chinook and chum salmon fry were captured in large numbers in all edge units and exhibited only slightly higher densities in low‐velocity areas (<15 cm/s).

No abstract

We predicted 22 years of return rates for wild Chinook salmon Oncorhynchus tshawytscha as a function of environmental conditions experienced during residency in freshwater, tidal delta, bay, and ocean habitats as well as as an indicator of density dependence (based on egg production) across life stages. The best predictors of return rate included the magnitude of floods experienced during incubation, a principal components factor describing environmental conditions during bay residency, a similar factor describing conditions experienced during the third ocean year, and an estimate of egg production. Our models explained up to 90% of the variation in return rate and had a very high forecasting precision, yet environmental conditions experienced during ocean residency explained only 5% of the variation. Our results suggest that returns of wild Chinook salmon can be predicted with high precision by incorporating habitat residency and that freshwater and nearshore environmental conditions strongly influence the survival of Skagit River Chinook