Progressive vector diagrams (PVDs) have been used to estimate transport in the coastal ocean from point measurements of velocity time series, although they strictly only approximate true particle trajectories in regions where the currents are spatially uniform. Currents in most coastal regions vary significantly in both space and time, making coastal transport estimates from PVDs questionable. Here, we used synoptic surface currents measured over periods of several months by HF radars in two coastal areas with distinctly different circulation features (the Gulf of Eilat, a deep, semi-enclosed basin, and the Delaware Bay mouth, a coastal estuarine outflow) to assess the time scales over which PVD paths computed from velocity time series at fixed points separate from particle trajectories computed from two-dimensional measured horizontal currents that vary in both space and time. In both study regions, PVDs and particle trajectories separate by 1 km over a mean time period of 7-10 h with no significant month-to-month variation in either region. The separation time statistics presented here should serve as a strong caution for investigators motivated to estimate transport using only point measurements.