The National Park Service monitors coho (Oncorhynchus kisutch) to detect population trends within the Golden Gate National Recreation Area, Muir Woods National Monument, and Point Reyes National Seashore. Coho have a complex life cycle: early life stages inhabit park streams while adults inhabit open ocean, returning to spawn. They are semelparous and typically have a three-year lifespan. Juvenile data are particularly useful for understanding how coho are influenced by environmental factors occurring within the parks, but current protocols require relatively intensive effort. We evaluated three methods for estimating juvenile abundance: the current method, which uses a stream-wide correction factor to account for imperfect detection, and two alternatives, which incorporate environmental covariates to estimate spatially explicit detection rates and coho densities within streams. We found that the two methods incorporating environmental covariates yielded similar results, and produced significantly more precise estimates of streamwide abundance than the current method. The two alternative methods, but not the current method, allowed us to detect a climate effect on juvenile recruitment at one of the two watersheds. Our analyses demonstrated that investing in understanding environmental influences on detection and abundance at finer spatial scales can lead to a more efficient and robust watershed monitoring program.