We examined the potential for using riparian conifer trees (coast redwood and Douglasfir) to assess historical nutrient flux and salmon escapement (fish that escape capture by commercial fisheries in the open ocean and are able to return to their natal stream to spawn) to a pair of coastal California watersheds (Mill Creek and Waddell Creek). In each basin, periods of known annual salmon escapement were compared with data on tree-ring growth, nitrogen content [N] and stable nitrogen isotope ratios (δ15N). All tree-ring variables (lagged by +1 year) were positively correlated with Mill Creek salmon escapement, but relationships were inconsistent at Waddell Creek and highly dependent on the time lag applied (0 to 7 years). Annual ring δ15N was especially variable among individual trees growing at the same location with differences as high as 4.8‰ at spawning sites and 6.8‰ at salmon-free control sites. Linear regression-based models were used to reconstruct historical escapement for the periods 1946- 2002 and 1904-2003 in the Mill Creek and Waddell Creek watersheds, respectively. Tree-ring based reconstructions indicated considerable variation in historical escapement to both basins, and contemporary population declines which began in the 1960s (Mill Creek) and 1970s (Waddell Creek) and appeared to level off in the 1990s. Although δ15N was easily quantified in annual tree rings, stream, soil, and within-tree nitrogen dynamics greatly hinder the robust assessment of marine-derived versus autochthonous sources of nitrogen from riparian tree rings.