First, the number of sites for which MOS forecasts are available has been increased to include many mesonet stations. These new stations are in Alaska, Hawaii, Puerto Rico, Canada, and the CONUS, the current number of sites being in excess of 11,000 over the CONUS grid alone. We will show the extent of these stations and indicate the impact they will have on the products. The increased number of stations will also appear in the matrices of forecasts available in AWIPS and over the SBN.
Second, the areas of coverage will be increased for the CONUS and Alaska products. The CONUS is extended farther north into Canada and the Alaska area has been extended to the east into Canada. Both of these extensions are to fulfill requirements to support the NWS weather and hydrologic forecasts downstream of the headwaters of rivers in Canada. These areas of coverage will be shown.
Finally, the BCDG analysis system used for gridding the station-based MOS forecasts has undergone many enhancements. For instance, in rugged terrain, the expected change with terrain elevation of the element being analyzed has to be estimated. Previously over the CONUS, this estimate was based only on the data being analyzed. Now, the change with elevation is estimated based on a combination of surface observations and upper air forecasts. Variable radii of influence are used to accommodate the variation in data density. Less weight is given to mesonet stations than for METAR stations to account for the difference in error characteristics. Interelement consistency is enhanced by not using a station in a dewpoint analysis judged to be in error for the temperature analysis for the same projection; the same process is used for wind. A ray-tracing smoother is used over the ocean to keep the temperature, dewpoint, and wind over the open ocean from unduly affecting bays and estuaries. A variable smoother is used over land that does not smooth the gridpoints very near the data points, but smooths heavily between data points. Examples of the analysis of the MOS forecasts will be shown over the expanded areas.
These improvements are being made in partial support of the National Blend of Models, for which GMOS is an important contributor and will be available in the National Digital Guidance Database.