Bottom pressure variability at synoptic periods (4--12 days) is investigated using observed data from the Labrador shelf and a barotropic numerical model. The model uses realistic bathymetry with 0.25¿ ¿0.25¿ resolution and is driven by wind stress derived from European Centre for Medium Range Weather Forecasts wind data. In some runs the model is coupled to a coarse-resolution version applied to the North Atlantic in order to determine the influence of remote oceanic forcing on the shelf. At the northern end of the shelf, the model shows strong coherence with the data. Coupling to the North Atlantic model suggests that remote forcing is important there. At the southern end of the shelf the model shows weaker but still significant coherence except at the outer edge of the shelf. No effect of remote North Atlantic forcing could be detected at the southern end of the shelf. A significant improvement in the fit at all locations and periods could be found by fitting a complex arbitrary constant to the data. This result suggests that some important physical forcing, relevant to the entire shelf, has not been included in this model. A modification of the model to include the effect of stratification did not improve the agreement between model and data. ¿ American Geophysical Union 1992