Less active Atlantic hurricane season foreseen by new model

A major new player in the seasonal Atlantic hurricane season forecast game is here--the UK Met Office, which issued its first Atlantic hurricane season forecast yesterday. The UK Met Office is the United Kingdom's version of our National Weather Service. Their 2007 Atlantic hurricane season forecast calls for ten named storms for the remainder of the season--12 for the entire season, when one includes Andrea and Barry. They make no forecasts for number of hurricanes or intense hurricanes, nor where the storms may strike. The UK Met Office forecast of ten storms for July through November is below the average of 12.4 for the active hurricane period that began in 1995, and well below the predictions of the other major seasonal forecast teams.

July-November 2007 Atlantic hurricane season forecasts (adjusted for the occurrence of Tropical Storms Andrea and Barry, where appropriate):

UK Met Office (June 19): 10 named storms.

Colorado State University (CSU) Phil Klotzbach/Dr. Bill Gray forecast (May 31): 16 named storms.

NOAA's forecast (May 22): 11-15 named storms.

Tropical Storm Risk, Inc. (TSR) forecast (June 4): 14.7 named storms.

How reliable are the UK Met Office forecasts?
This is the first year that the UK Met Office has issued a forecast of hurricane season activity, so we don't have any previous years to evaluate their forecasts. The results of their experimental forecasts issued for the 1987-2002 seasons are scheduled to be published later this year in the peer-reviewed journal Geophysical Research Letters. The UK Met Office claims that their forecast out-performed the forecasts made for the 2005 and 2006 Atlantic hurricane season issued by the other major seasonal forecast groups. I have high hopes for the UK Met Office forecast, since it is based on a promising new method--running a dynamical seasonal prediction computer model of the global atmosphere-ocean system. The Dr. Bill Gray/CSU forecast is based on statistical patterns of hurricane activity observed from past years. These statistical techniques do not work very well when the atmosphere behaves in ways it has not behaved in the past. The UK Met Office forecast avoids this problem by using a global computer forecast model--the GloSea model (short for GLObal SEAsonal model). GloSea is based on the HadCM3 model--one of the leading climate models used to formulate the influential UN Intergovernmental Panel on Climate Change (IPCC) report. GloSea subdivides the atmosphere into a 3-dimensional grid 3.75° in longitude, 2.5° in latitude (277.5 km), and 19 levels in the vertical. This atmospheric model is coupled to an ocean model of even higher resolution. The initial state of the atmosphere and ocean as of June 1, 2007 were fed into the model, and the mathematical equations governing the motions of the atmosphere and ocean were solved at each grid point every few minutes, progressing out in time until the end of November (yes, this takes a colossal amount of computer power!) It's well-known that slight errors in specifying the initial state of the atmosphere can cause large errors in the forecast. This "sensitivity to initial conditions" is taken into account by making many model runs, each with a slight variation in the starting conditions which reflect the uncertainty in the initial state. This generates an "ensemble" of forecasts and the final forecast is created by analyzing all the member forecasts of this ensemble. Forty ensemble members were generated for this year's UK Met Office forecast. The researchers counted how many tropical storms formed during the six months the model ran to arrive at their forecast of ten named storms for the remainder of this hurricane season. Of course, the exact timing and location of these ten storms are bound to differ from what the model predicts, since one cannot make accurate forecasts of this nature so far in advance.

The grid used by GloSea is fine enough to see hurricanes form, but is too coarse to properly handle important features of these storms. This lack of resolution results in the model not generating the right number of storms. This discrepancy is corrected by looking back at time for the years 1987-2002, and coming up with correction factors (i.e., "fudge" factors) that give a reasonable forecast. This year's GloSea forecast shows a cooling trend in the tropical Atlantic sea surface temperature (SST) compared to what we've seen in recent years, and is a major reason why the UK Met Office forecast is so much lower than the other seasonal Atlantic forecasts. I believe that the GloSea model has high enough resolution to do as good a job as the other seasonal hurricane forecasts this year, but it's hard to make an informed judgment until their research results are published. The GloSea forecast is based on sound science, though, and does call into question whether or not the other seasonal forecasts are forecasting unrealistically high levels of hurricane activity in the Atlantic this year. I think that is probably the case, and a better forecast can be made by averaging together the four models into a consensus forecast. Consensus forecasts are difficult to beat, and the consensus of the CSU, NOAA, TSR, and UK Met Office forecasts yields a prediction of 13 more named storms this year, for a total of 15.

The future of seasonal hurricane forecasts
The future of seasonal hurricane forecasts using global dynamical computer models is bright. A group using the European Center for Medium Range Weather Forecasting (ECWMF) model is also experimenting with some promising techniques using that model. Models like the GloSea and ECMWF will only get better as increased computer power and better understanding of the atmosphere are incorporated, necessitating less use of "fudge" factors based on historical hurricane patterns. If human-caused climate change amplifies in coming decades, statistical seasonal hurricane forecasts like the CSU's may be limited in how much they can be improved, since the atmosphere may move into new patterns very unlike what we've seen in the past 100 years. It is my expectation that ten years from now, seasonal hurricane forecasts based on global computer models such as the UK Met Office's GloSea will regularly out-perform the statistical forecasts issued by CSU.

Bill Proenza to appear on Thursday's Barometer Bob show
Thursday night June 21, new NHC director Bill Proenza will be the guest on the Barometer Bob Show. You can listen at barometerbobshow.com, or dial in via their toll-free number 1-866-931-8437 (1-866-WE1THER). If you want to ask him a question, you can do so using their Storm Chat web page.

Tropical Outlook
A cold front pushing off the East Coast of the U.S. on Thursday could trigger formation of a tropical depression by Friday or Saturday over the Gulf Stream waters between Daytona Beach, Florida and Cape Hatteras, North Carolina. Several of the computer models are forecasting that a weak low pressure system will form by Friday in this region. However, there will be a lot a wind shear close by, which may make any storm storm that does form subtropical or non-tropical. It is uncertain where such a storm might move, since steering currents will be weak. I'll have an update on this situation by Friday at the latest.

Jeff Masters