Surface winds observed at Charleston, South Carolina, and at an offshore buoy located some 300 km to the east of Charleston are analyzed to determine operational transformations for estimating the offshore sinds from the coastal winds. Both the statistics describing the wind field at each station along with those describing the correlation between stations are found to be seasonally modulated. Ensemble averaged (by season) transfer functions computed by using the first 3 years of available data are applied to a fourth independent data year for assessing their predictive utility. The transfer functions and predictability are both seasonally modulated and the transformation matrix is nondiagonal (each offshore wind velocity component depends upon both coastal components).