ECCO state estimation results from 10 years during the World Ocean Circulation Experiment are used to assess the quality of surface flux adjustments made to the initial NCEP re-analysis-1 products. During the state estimation procedure, surface fluxes are adjusted together with initial temperature and salinity conditions so that the model simulation becomes consistent with ocean observations. Independent estimates of the adjustments from bulk formula and regional field observations are also employed to evaluate the results. Buoyancy flux adjustments are found to be within the crude prior error bars on these fields. Outside the boundary current regions, they are consistent with known large-scale deficiencies in the NCEP products. Wind stress adjustments are also everywhere within the prior error bars, but exhibit regional small-scale features that reflect ocean model failures to resolve intense boundary currents. On large scales, the inferred adjustments to NCEP wind stress fields are consistent with inferences made from satellite wind stress measurements. Further improvements in the surface flux estimates obtained through state estimation procedures are anticipated as the estimation procedure becomes more complete by including the use of improved prior error covariance information, and as the ocean model becomes more skillful, for example, in simulating boundary currents by increasing its resolution.