Few tidal records are available pre-1900 for the Pacific Ocean. We improve data coverage by recovering historical tabulations and digitizing analog tide rolls from Astoria, Oregon for 1853-1876. Nearly 13,500 overlapping images of tides from 1855-1870 were digitized at a 6 minute resolution using a line-finding algorithm. Available hourly and high/low tabulations were also digitized, as were nearby hourly records from 1933-1943. Uncertainty was assessed by evaluating manual staff measurements, historical documents, and leveling surveys. Results suggest that uncertainty in mean sea level varies from ± 0.07m (early 1850s) to ± 0.03m (1867-1876) and is driven primarily by datum and benchmark uncertainty, rather than measurement precision, data reduction procedures, or hydrodynamic changes. We also corrected an up-to 0.05m error in the 1925-1960 tidal datum at Astoria. Harmonic analysis shows that major tidal constituents increased by up to 7% between 1855 and 2018. Mean tidal range increased by 0.1m (5%), with more change occurring in July (0.17m larger than winter (0.07m larger). By contrast, sea level increased most in winter, and least in spring/summer. Tidally-based estimates of river discharge suggest that these observations are caused by a ~50% reduction in peak spring discharge and a 30-60% increase in winter discharge. No evidence of altered upwelling is found. Overall, Astoria relative sea level (RSL) increased by 0.06m ± 0.04m since the 1858-1876 epoch, or, after accounting for vertical land motion, 0.11 ± 0.09m. Consistent with GNSS measurements, RSL has dropped near the estuary mouth since 1905, indicating a strong tectonic influence.