The typical barrier jet extended 50 km out from the coast with a roughly log-normal distribution. There was little seasonal variation in jet width. The spatial variation was also small except in the strait between Kodiak Island and the mainland where the flow was constricted by land on both sides. A small fraction of the barrier jets were observed to detach from the coast for part of their length, a behavior not explained by existing theory. The typical separation was 10 km with values above 20 km being rare. Thus, most jets hug the coast making the near shore waters no safer for vessels seeking shelter.
Maximum wind speed for both barrier jets and hybrids ranges from 15 to in excess of 25 m/s with a median of 20 m/s. The barrier jets exhibit wind speeds of 2 to 3 times that observed on the synoptic scale further off shore. Thus, barrier jets pose both a hazard and a forecasting challenge in the Gulf of Alaska. The structural climatology for hybrid cases is similar to that for barrier jets so they pose similar problems.
The most intense barrier jets occur where barrier jets are most frequent, compounding the problem for vessels operating in this region. This climatologically favored region starts just to the west of Yakutat Bay and extends westward toward Prince William Sound, encompassing the approaches to the oil port of Valdez and the fishing port of Cordova.
These findings support the hypothesis that barrier jets are caused by hydrostatic blocking of onshore flow. They occur preferentially in regions with a climatology of onshore flow and in seasons of maximum wind speed. Likewise, their occurrence is closely tied to the elevation of the near shore terrain. The finding that "near shore" is best defined as within 100 km is a bit surprising, suggesting that the cold air damming and resultant barrier jet may be caused by mountains some distance inland and still extend past the coast.
The findings on hybrids suggest that drainage of cold air from the continental interior through coastal gaps plays a role in at least one fourth of the events (i.e. hybrids make up over one fourth of the total of hybrids plus barrier jets). Mesoscale modeling will be required to determine the extent to whether less obvious drainage flows contribute to the cases classified here as barrier jets.