Wind tunnel experiments were performed to quantify the drag reduction on a wind turbine airfoil partially or fully covered with riblets. A full-scale 2.5 MW wind turbine airfoil section, typical for the near tip, was placed in the free stream flow of the wind tunnel at the Saint Anthony Falls Laboratory, University of Minnesota. Various sizes and geometries of experimental riblets were provided by 3M Company and tested at angles of attack ranging from 0 degrees < alpha < 10 degrees (0.25 <= C-L <= 1.14) and at a Reynolds number of Re = 2.2 x 10(6). Mean drag was measured via wake survey (momentum deficit) and with a sensitive force balance. Lift was measured directly from the force balance. Tests included the cases of complete and partial riblet coverage on the wing. Results indicated that riblets could provide an overall reduction of skin friction drag, and that the amount of the decrease varied with riblet height and geometry. Partial riblet coverage appears in some cases more efficient than its complete coverage counterpart. The percentage of drag the riblets reduced varied greatly and in some cases the riblets were even detrimental to the airfoil. The most efficient riblet for a completely covered airfoil was found to be the V-groove shape of 100 mu m height. It produced a reduction of roughly 6% in the operational range expected in a turbine airfoil. On the other hand, the most efficient riblet size for partial coverage was also a V-groove shape and seemed to shift slightly to a smaller peak height of 80 mu m. This configuration produced a reduction of roughly 4% in the range of angle of attack that is typical for operation in the field. The average non-dimensional square root of the groove cross-section, l(+), defined in terms of the drag coefficient at design angle of attack for the optimum riblet configuration in the fully coverage case was found to be l(+) = 10, which is very close to the optimum value found for planar surfaces. Based on our results we propose a formulation for the optimum riblet size in airfoil considering the mean drag coefficient and chord length Reynolds number. Even though the optimum full coverage case showed better performance that the partial case, the additional drag reduction benefit may be offset by the additional application cost. (C) 2012 Elsevier Ltd. All rights reserved.