This article documents a systematic bias in surface wind directions between the TAO buoy measurements at 0°, 170°W and the ECMWF analysis and forecasts. This bias was of the order 10° and persisted from November 2008 to January 2010, which was consistent with a post-recovery calibration drift in the anemometer vane. Unfortunately, the calibration drift was too time-variant to be used to correct the data so the quality flag for this deployment was adjusted to reflect low data quality. The primary purpose of this paper is to inform users in the modelling and remote-sensing community about this systematic, persistent wind directional bias, which will allow users to make an educated decision on using the data and be aware of its potential impact to their downstream product quality. The uncovering of this bias and its source demonstrates the importance of continuous scientific oversight and effective user-data provider communication in stewarding scientific data. It also suggests the need for improvement in the ability of buoy data quality control procedures of the TAO and ECMWF systems to detect future wind directional systematic biases such as the one described here.