Part of the common arguments from skeptics/deniers series
Round I : Catarina
In 1979 William Gray said the formation of a hurricane in the South Atlantic was an impossibility:
Genesis does not occur in the tropical southeast Pacific and South Atlantic because the background seasonal climatology is so unfavorable. Although short-term positive deviations of genesis potential may be as large at these locations elsewhere, they can never overcome their strongly unfavorable climatoligical background.Yet the United Kingdom's Hadley Centre for Climate Prediction and Research, a branch of the UK's Met Office, had predicted that a hurricane would form in the South Atlantic in a globally warming world. So what the models said would occur, Gray said would "never" occur. In March of 2004 such an 'impossible' hurricane formed. Her name was Catarina.
2004 went on to be a record shattering year for hurricanes. The number of intense storms were double of that predicted by William Gray. His explanation was simple: "This year did not behave like any other year we have studied." . Gray's forecasts are based upon comparison to the past. Perhaps in a warming and changing world past observations of seasonal trends won't apply anymore. On the other hand the GCM models forecast based upon our understanding of the laws of physics. This brings up a rather famous quote: "The laws of physics are eternal and cannot be changed with additional research, venture capital or majority votes."
Round II: Seasonal Forecasts
In December of 2004 Gray announced that he did not expect "anything close to the U.S. land-falling hurricane activity of 2004" for the 2005 hurricane season. Gray predicted 11 named storms, 6 hurricanes and 3 major hurricanes. Only a slightly above average year. Unfortunately 2005 was going to to shatter all records. By late may 2005, Gray had upped the forecast to fifteen named storms. While this was an increase it was still only about half of what to come. Meteo-France on the other hand, was using a new form of hurricane forecasting which employs climate models. They predicted 22 named storms. The end game resulted in 28 storms. While both forecasts were off Gray was much farther off than the climate models. Meteo-France was off by 6 while Gray was initially off by 17 and later 13 names storms. Grays margin of error was 2-3 times larger than the Meteo-France's climate models. The biggest margin of error by a major climate model I've found so far is UK's Met office. But even this model which predicted 16.2 storms was much better than Grays initial prediction and was even better than his revised prediction (albeit not by much). Looking at 2005 alone, the climate models are the best forecasters we have. Much better than Gray's intuition and gut instinct and traditional forcasting in general. Surely as our understanding of the climate engine increases these models will only get better and better.
Thomas Kuhn once wrote that anytime there is a paradigm shift in science it will be met with resistance. Accepting climate models might be doubly tough for William Gray as he spent 52 years developing a forecasting method that might now be obsolete. But given that the models are currently outperforming Gray's own predictions and are correctly predicting what Gray thought was impossible then it might be prudent for him to acknowledge that the modellers do in fact know what they are doing.
Discover, Discover Dialogue: Meteorologist William Gray, Kathy A. Svitil, 09.09.2005
SUMMARY OF 2004 ATLANTIC TROPICAL CYCLONE ACTIVITY AND VERIFICATION OF AUTHOR'S SEASONAL AND MONTHLY FORECASTS, William M. Gray and Philip J. Klotzbach
Meteorology Over the Tropical Oceans, Hurricanse: Their formation, Structure and Likely Role in the Tropical Circulation, William Gray, 1979, pp. 155-218 (quote on 181)
FORECAST OF ATLANTIC HURRICANE ACTIVITY FOR SEPTEMBER AND OCTOBER 2004 AND SEASONAL UPDATE THROUGH AUGUST, William M. Gray and Philip J. Klotzbach