Sandy, the Europeans, and the Americans

Hurricane Sandy just before landfall, 1:35 p.m. EDT, 29 October 2012
(From NASA)

Hurricane Sandy came ashore on the evening of Monday, 29 October, just south of Atlantic City, NJ.  It was definitely a unique, “storm of the age” kind of event as it transitioned from a tropical system while it was offshore to a non-tropical (or “extratropical,” as meteorologists like to say) super storm as it crossed the coast and moved inland.  The combination of tropical energy (in the form of moisture and heat) and winter storm energy (in the form of jet stream winds and a sharp temperature gradient between very cold air to its west and warm, tropical air to its east) led to the lowest sea level (barometric) pressure ever recorded off the northeastern U.S.  Just before landfall, a hurricane hunter aircraft measured a sea level pressure of 940 mb (27.76″) in the center of Sandy and when it came ashore a weather observing site near Atlantic City measured a pressure just above 945 mb (27.91″).  Both values broke the previous record of 946 mb (27.94″) measured at the Bellport Coast Guard Station on Long Island during the Great New England Hurricane of September 1938.  The surface map showing contours of sea level pressure looked like this just as Sandy was crossing the coast and transitioning to an extratropical super storm on the evening of Monday, 29 October:

GFS Analysis of Sea Level Pressure and 1000-500 mb thickness, 8 p.m. EDT, 29 Oct. 2012 (0000 UTC, 30 Oct 2012)

Even more remarkable than the storm itself was the accuracy of the weather model forecasts of the storm as much as 8 1/2 days in advance.  The forecasts showed how much progress has been made in weather forecasting as computers have become more and more powerful.  The model from the European Center for Medium Range Weather Forecasts (ECMWF) gave a surprisingly accurate depiction for the location and intensity of Sandy at landfall from both its morning and evening runs on Sunday, October 21.  That’s 8 days before the storm hit New Jersey and a couple days before it had even become very well organized in the Caribbean.  Truly incredible!  Here is what the model predicted on Sunday evening, 21 October:

ECMWF 192-hour forecast of MSLP and 1000-500 mb thickness from 8 p.m. EDT, 21 Oct. 2012 (0000 UTC, 22 Oct 2012)

Unfortunately, most of the stories about this excellent forecast have focused on how much better the European model forecast was with Sandy compared to the output from the primary American global weather forecast model, the Global Forecast System (or GFS), at the same time.  For instance, see these stories:

The forecast from the GFS model was noticeably poorer as evidenced by this map showing its output from the same time as the ECMWF model run on October 21:

GFS 192-hour forecast of Sea Level Pressure and 1000-500 mb thickness from 8 p.m. EDT, 21 Oct. 2012 (0000 UTC, 22 Oct 2012)

While the GFS did develop the storm, it clearly was taking it out to sea into the central Atlantic and entirely missed the intensity of the cold air diving into the eastern United States and its interaction with the storm.  The GFS needed a few more days before it finally started bringing the storm closer to the coast and had it deepen when encountering the cold trough in the eastern U.S.  It definitely was the poorer model for this storm.

On average, the ECMWF model beats the GFS, as it has for decades, because the Europeans have focused their energies and resources on improving one thing: medium-range weather forecasting.  They have a a higher resolution model with a better scheme for ingesting initial observations and satellite data which runs on a more powerful computer system than what the U.S. has.  The U.S. agency containing the National Weather Service and responsible for developing weather models, NOAA, has limited resources and must allocate them across a broad spectrum of needs to protect life and property in the U.S.—the needs range from short-term forecasting of severe weather events, such as tornadoes, to the long-term issues such as summertime drought or wintertime cold.  And, while the ECMWF model is better on average than the GFS, it is far from perfect.  Just a month before Sandy, the ECMWF model was consistently trying to bring another storm, hurricane Nadine, into southern Europe as a destructive extratropical storm.  The GFS, on the other hand, generally kept Nadine out at sea with a looping track south and southwest of the Azores.  The GFS was the much better model in this case as the storm got nowhere near southern Europe and threatened the Azores twice.

Back to Sandy, for a moment—I would hate for us to get mired in a controversy about ECMWF vs. GFS and miss the incredible skill shown by the ECMWF model more than eight days in advance.  Billions of dollars will be saved or losses prevented if such forecast skill becomes commonplace.  This forecast gives us an example of what is possible as we continue to advance the science of meteorological modeling.  Now if we can just find the money to continue funding the research in this all-important area…


About notjustweather

Dave Linder: Meteorologist, computer scientist, writer... and one who enjoys a well-timed quick quip.

Posted on November 8, 2012, in Weather and tagged , , , , , , , . Bookmark the permalink. Leave a comment.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: