Shell Oil made a misguided and poorly informed decision to move a huge drilling platform (the Kulluk) from Dutch Harbor Alaska to Seattle starting December 21st. As described in the Seattle Times and elsewhere the problems grew from broken tow lines and faulty engines on December 26th, to the eventual grounding the Kulluk on an island just south of Kodiak island on December 31st.
|The Kalluk beached south of Kodiak Island|
Lets take a look at the surface charts during a few points in this ill-favored trip.
First, chart of 10am on the 26th, when wind conditions and large seas caused them to have lots of problems with tow lines. A 960 hPa low is found southeast of Kodiak; keep in mind this is about the same central pressure of the Columbus Day Storm, perhaps the greatest storm to hit Seattle during the past century. Such storms are a dime a dozen in the Gulf!
Then 10 AM on the 28th. Two major cyclones..one with 962 hPa and the other around 967 hPa to the southwest.
Were these storms big surprises as claimed by Shell? Let me show you some forecast model output and you decide! The University of Washington forecasts over the Pacific are virtually identical to that of the main U.S. global model, the model used heavily by the National Weather Service to forecast conditions over the Pacific. I have easy access to the graphics, so lets look at them. The tow began on 21 December, so lets examine the forecast started at 4 PM Seattle Time (3 PM Kodiak Time) valid at 4 PM on December 26th. A strong 967 hPa low south of the Aleutians. Not bad five days ahead of time...and of course the subsequent forecast got even better. Clear warning that the tow was a bad idea.
So there they were on December 26th and in trouble. You would think they would check the forecast to see whether they should head to port. So lets looks review the 5 day forecast for 31 December at 4 AM. ANOTHER strong storm...even deeper... with Kodiak Island in the area of strong pressure differences and thus strong winds.
I should note here that forecasts over the oceans have gotten much more accurate during the past decade or so, the key reason being the huge amount of satellite data we get over the oceans these days. In the old days we used to get a lot of observations in the middle of storms as storms got caught due to the poor forecasts. Today they avoid the storm!
Ok, that is one forecast. State-of-the-art forecasters use ensembles (the output of many model runs) that can give us uncertainty information and a measure of the confidence of the forecasts. One of the most popular ensembles collections used by U.S. forecasters is NAEFS...the North American ensemble system that combines the ensembles of both the U.S. and Canadian ensemble system. Here is the output from this system for the runs started at 4 PM on 26 December for the town of Kodiak. Look at the wind speed panel (note the wind speeds are given in km per hour, e.g., 40 km per hour is 25 mph). The bracket shows the range of speeds from the various ensemble members and the middle bar is the median). Clearly, some of the solutions are for very wind conditions.