Comments on “Prediction of Vessel Icing for Near-Freezing Sea Temperatures”

Abstract

The operational NOAA categorical vessel icing algorithm is evaluated with regard to advances in understanding of the icing process and forecasting experience. When sea temperatures are <2-3degC above the saltwater freezing point there is the likelihood of supercooling of the spray during its trajectory and extreme ice accretion on topside structures. The NOAA algorithm shows excellent results when compared to a new cold-water dataset from the Labrador Sea (mean sea temperature of -1.3degC), even thohgn the algoritme was developed from an Alaskan dataset with a mean sea tempearature of 3.6degC. A rederived algorithm from the combined dataset is nearly identical to the operational algorithm. The influence of sea temperature in the NOAA model is consistent with the supercooling hypothesis and an additional iceing category of extreme is recommended for the algorithm. Severe icing isn the Bering Sea, Gulf of Alaska, and Sea of Japan is primarily caused by extreme cold-air advection, while low sea temperatures contribute to severe iceing in the Labrador Sea, Denmark Strait, and Barents Sea. Indirect verification showd that NOAA provided excellent forecasts to over 140 fishing vessels in Alaskan waters during late January 1989, the worst iceing episode of the decade. This case suggests that current-generation atmospheric models are capable of providing reliable 36-h forecasts of cold-air advection, and thus indicating regioins of heavy icing. A wave height/wind speed threshold for the onset of topside icing is 4 m s-1 for a 15m vessel, 10m s-1 for 40m trawler and 14m s-1 for a 100m vessel, developed from seakeepning theory. These wind speeds are exceeded 83%, 47%, and 15%, respectively, during February in the Bering Sea.