The purpose of this study was to assess and fine tune aircraft icing forecast algorithms for predicting probable icing and icing intensity conditions in complex terrain below 10,000 feet (MSL). The area of interest was the vicinity of Mt. Washington, New Hampshire and various observations were made and data were gathered during the period of December 2000 through April 2001. In addition to conventional surface and upper air observations that were available for the study, the Cold Regions Research and Engineering Laboratory (CRREL) collected Rosemount ice detector and rotating multi-cylinder ice measurements from the summit of Mt. Washington (KMWN) to provide in situ accretion and intensity data. During this period, 00Z and 12Z ETA model forecasts were also archived to provide additional data for nowcasts and forecasts. Air Force Weather Agency (AFWA) and Aviation Weather Center (AWC) icing forecasts for the region were also archived for comparison purposes.

The initial algorithm used was based on the decision-tree procedures outlined in the former Air Weather Service Pamphlet (AWSP) 105-56. Separate manual nowcasts were made using this algorithm based on either the closest GYX radiosonde observation or ETA model initialization fields interpolated to the KMWN area. Preliminary findings clearly show an improvement in icing nowcast results using the interpolated ETA input versus the straight radiosonde input. Using the decision-tree algorithm, icing forecasts for 12, 24, 36, and 48 hours were also made based on the corresponding ETA forecast fields. Initial results indicate that icing occurs at the summit more frequently than either the nowcasts or forecasts indicate, probably due to terrain-induced adiabatic lifting taking place in the region. This creates a cooler and moister environment that is not being taken into account. Modification of the algorithm to take this factor into account is the current focus of this research. Results and forecast comparisons will be presented at the symposium.