Effects of large wood restoration on coho salmon in a northern California watershed: a before-after-control-impact experiment

Substantial time, money, and effort are invested in river and stream restoration projects to aid in the recovery of imperiled salmonid populations, but there is little evidence that these efforts have had lasting positive impacts on juvenile fish growth and survival. To assess the effectiveness of large woody debris (LWD) restoration, which is one of the most common restoration practices, I evaluated the growth and survival response of endangered Central California Coast coho salmon (Oncorhynchus kisutch) in a paired watershed before-after impact-control (BACI) study. To determine if LWD supplementation influenced coho salmon growth and survival, two neighboring, similar watersheds in Northern California were selected to conduct long-term monitoring of both fish and habitat metrics. Fish and habitat monitoring consisted of summer and fall electrofishing surveys, juvenile outmigrant trapping, passive integrated transponder (PIT) array detections, and summer and winter habitat surveys. After three years of pre-treatment monitoring in both watersheds, Pudding Creek (the ‘experimental’ watershed) was supplemented with 1,365 cubic meters of LWD throughout 80% of the mainstem anadromous spawning habitat in 2015. Post-treatment monitoring then continued in both watersheds until 2020. Though wood density increased more in the experimental watershed (31%) compared to the control watershed (13%) following wood treatment, there was no winter slow water habitat response, meaning the limiting factor for coho populations in these watersheds was not addressed. I used generalized linear mixed effects models with year as a random effect to evaluate summer and winter growth response to wood supplementation. I found that summer and winter growth were positively associated with wood, but the experimental watershed had consistently higher winter growth compared to the control. The wood treatment response did not align with the biological response; wood density increased more in the experimental watershed, but growth did not increase more in the experimental compared to the control watershed. To estimate winter survival rates, I used a Cormack-Jolly-Seber (CJS) model. I found that winter survival increased through time in the control while it stayed level in the experimental watershed. This thesis illuminates the utility of having a paired watershed study design with habitat and biological response analysis in tandem. The results from this experiment lead to a variety of questions and concerns relating to the treatment design and how treatment is paired with the study design. This thesis provides a foundation for long-term monitoring to understand the effects of restoration efforts for a species at the southern extent of its range. This is particularly important given the at-risk status of these salmonid populations and the additional threats these fish face from a changing climate