Most objective analysis methods, such as the so called Cressman, are quite good for situations where the data points are somewhat random and uniform over the grid. However, these criteria are many times not satisfied. This is especially true for ocean areas around Alaska portrayed in the National Weather Services's National Digital Guidance Database (NDGD). A few buoys exist in the Gulf of Alaska, mostly along the seacoast, and only one in the entire Bering Sea. Usually, observations over land are not of much use over the open ocean, so the analysis method has to rely on a good first guess, bogus values, careful use of radii of influence tailored to the data distribution, and judicious smoothing.

The Meteorological Development Laboratory is producing MOS guidance for the Alaska area, and is using an enhanced version of the Cressman analysis method called BCDG to grid the forecasts for the NDGD. MOS forecasts exist at points where observations are available, so the gridding difficulties are the same as for the analysis of observations. This paper will detail how an analysis of MOS wind speed is made for the Alaska area, including a data-preserving smoothing method.