Why are 30-m winds around coastlines sometimes exactly zero?
We have identified the problem to have been caused by a code error in the following manner. 30 m winds were derived on the model’s native E-grid using the same algorithm as that used to obtain the 10 m winds (wgrib abbreviations also UGRD, VGRD, merged files record #s 293, 294). Both of these sets of winds are calculated by the Eta code at mass points, using the algorithm based on mid-layer winds at the four neighboring wind points. If the Eta bottom layer depth is greater than 30 m, the 30 m winds thus obtained are left as is. But if it is less, a linear vertical interpolation in z is performed between the lowest layer wind, and the wind of the layer above, to obtain the 30 m winds. In doing this, our code erroneously was not interpolating the winds derived for the mass point, but was doing the interpolation using values at a neighboring wind point, the one to the east or to the west of the mass point. Given that the lowest Eta layer of our RR code is thinner than 30 m, this means tha
