Observations made on the northern Portugal mid-shelf between May 13 and June 15, 2002 were used to characterise the near-surface velocity during one upwelling season. It was found that in the surface mixed layer, the 'tidal current' was diurnal, but the tidal elevation was semi-diurnal. Both the residual current and the major axes of all tidal constituents were nearly perpendicular to the isobaths and the tidal current ellipses rotated clockwise; the major axis of the major tidal ellipse was about 3 cm s-1. The extremely strong diurnal current in the surface layer was probably due to diurnal heating, cooling, and wind mixing that induced diurnal oscillations, including the diurnal oscillation of wind stress. This is a case different from the results measured in the other layers in this area. The near-inertial spectral peaks occurred with periods ranging from 1 047 min to 1 170 min, the longest periods being observed in deeper layers, and the shortest in the surface layer. Weak inertial events appeared during strong upwelling events, while strong inertial events appeared during downwelling or weak subinertial events. The near-inertial currents were out of phase between 5 m and 35 m layers for almost the entire measurement period, but such relationship was very weak during periods of irregular weak wind. Strong persistent southerly wind blew from May 12 to 17 and forced a significant water transport onshore and established a strong barotropic poleward jet with a surface speed exceeding 20 cm s-1. The subinertial current was related to wind variation, especially in the middle layers of 15 m and 35 m, the maximum correlation between alongshore current and alongshore wind was about 0.5 at the 5 m layer and 0.8 at the 35 m layer. The alongshore current reacted more rapidly than the cross-shore current. The strongest correlation was found at a time lag of 20 h in the upper layer and of 30 h in the deeper layer. The wind-driven surface velocity obtained from the PWP model had maximum amplitude of about 7 cm s-1, corresponding to a wind stress at 0.1 Pa, and the horizontal velocity shear due to thermal wind balance had the order of 3 cm s-1. So the local wind and thermal wind would only explain a part of the strong surface velocity variations.