Advancing environmental DNA (eDNA) beyond conventional species detection applications to support wildlife and fisheries management has garnered great interest. The management of spawning Pacific salmon (Oncorhynchus spp.) could especially benefit from highly scalable eDNA tools for estimating local salmon abundance due to their extensive freshwater habitat distributions. Past studies have shown great promise for using salmonid eDNA to estimate abundance or biomass, though this predictive ability has rarely been evaluated beyond individual systems. We evaluated the potential for broadscale applicability of eDNA to estimate salmonid spawner abundance across a suite of disparate stream and river systems representing a range of different environmental conditions. We collected eDNA, spawner counts, and environmental variables at 39 salmon spawning grounds and/or migratory routes that encompassed a variety of freshwater habitats. We demonstrate that eDNA concentrations, when corrected for discharge, can be used to predict live salmon abundance across streams, whereas dead salmon were not well predicted. We also show that in the absence of equipment-intensive discharge measurements, point measurements of flow velocity and channel width may be used as a sufficient discharge proxy. Furthermore, discharge-corrected salmonid eDNA concentrations were not significantly influenced by water temperature or turbidity, but were positively related to bacterial eDNA concentrations. Collectively, these findings support the use of eDNA as a versatile tool for enumerating salmonids across systems and for integration of eDNA into salmonid monitoring programs. This work more broadly represents an important advancement of eDNA for applications beyond species detection and towards estimating abundance across lotic systems.