The state of Alaska has both the highest number per capita of registered pilots and higher per capita general aviation accident rates within the United States, according to National Transportation Safety Board statistics. Many of these accidents are weather-related and stem from VFR-certified pilots traversing conditions that they or their aircraft are poorly equipped to handle. In particular, localized zones of inflight icing and turbulence tied to Alaska's complex
geophysical environment make for challenging flight conditions even for aviators with decades of experience within Alaska.
In this paper we address the utility, for forecasts of aviation impact variables over Alaska, of employing mesoscale model data as input to commonly used algorithms for inflight icing. Specifically, we will use gridded output data from Penn State/NCAR Mesoscale Model (MM5) forecasts over the Alaska region as input to several commonly used icing algorithms, including the Stovepipe algorithm (Bernstein, 1997) and the Schultz and Politovich (1992)
algorithm among others, as well as direct forecasts of cloud liquid and supercooled water available from MM5.
Our tests currently focus on a documented event where inflight icing was present over an extensive area of Alaska during early December 1992. Our evaluation will focus on the 1) accuracy of areal coverage, 2) ability to reproduce trends from PIREPs and 3) statistical measures of skill, of the resulting icing forecasts produced by the use of MM5 in tandem with the algorithms
The 8th Conference on Aviation, Range, and Aerospace Meteorology