Within 100 km of the North American coastline, ocean tide loading effects are ~1 order of magnitude larger than the present-day precision of absolute gravity (1 μGal) and continuous, differential Global Positioning System (GPS) measurements (3--8 mm). In relation to the precision of these measurements, ocean loading effects on gravity tend to be larger than the effect of loading on GPS measurements. Even at sites hundreds of kilometers from the coast, ocean tide loading can affect superconducting gravimeters at a level of 1 order of magnitude higher than the precision with which tidal constituents amplitudes can be resolved (0.1 μGal). In order to reduce contamination of high-precision geodetic measurements by ocean tide loading, a numerical representation of the nearshore ocean tide load constituents, M2, S2, N2, O1, K1, has been developed using available ocean tide, numerical, modeling results for Canadian waters. The nearshore ocean tides are represented by a grid of cells, ranging in size from 0.125¿ to 0.5¿, in which the amplitude and phase lag of tidal constituents are constant and which are bounded, in general, by lines of constant latitude and longitude. The detailed nearshore tidal representation was originally designed to supplement the global 1¿¿1¿ model of Schwiderski, but computations have also been carried out using the FES95.2, global model. Ocean tide loading computations that include the more detailed coastal ocean tide representations yield 10--20% changes in tidal correction values. A significant reduction in the variance of absolute gravity residuals is observed at Vancouver Island absolute gravity stations when the more detailed coastal model is used, but a considerable, unexplained residual remains at Holbert at the northwestern tip of Vancouver Island. Computations based on the FES95.2 model yield slightly smaller gravity residuals at west coast sites but only at the measurement precision level. Ocean loading effects on daily GPS positions in the Western Canada Deformation Array are adequately predicted by the Schwiderski global ocean tide loading model. Data from the Canadian Superconducting Gravimeter Installation, Gatineau, Quebec, for M2 and O1 were compared with ocean tide load modelling results for the Schwiderski and FES95.2 global ocean tide models enhanced by the nearshore model. With the enhancement, the two models agree with each other to within the expected computational uncertainty of 0.1 μGal for both M2 and O1. Both models predict an ocean loading effect at O1 within 0.1 μGal of the observed effect; however, the predicted loading effect for M2 is still 0.2 μGal smaller than the observed effect, which suggests that there is still room for improvement in semidiurnal ocean tide models for the east coast of North America. The enhanced ocean tide loading models will become more important in the future when gravity and position are measured at subcentimeter precision over periods shorter than 24 hours. ¿ 1998 American Geophysical Union