WunderBlog Archive » Category 6™
Category 6 has moved! See the latest from Dr. Jeff Masters and Bob Henson here.This winter's forecast: NOAA vs. the woolly bears
The tropics are quiet again today, so let's follow up on yesterday's discussion about the long range forecast for the coming United States winter. Those of you outside the U.S. will probably be more interested in what the International Research Institute for Climate Prediction has to say for your country, and I encourage you to check out their excellent web site for their seasonal forecasts. Interestingly, they forecast that virtually the entire world will have average to much above average temperatures during the December-February period, with only two tiny pockets of slightly below-average temperatures in Australia and central Asia. This should keep the oceans at the near-record high temperatures for near year's hurricane season, helping fuel another round of more intense than usual hurricanes. In addition, El Nino is expected to remain in a near-neutral phase the next 6-9 months (same as for this year's hurricane season), which should result in a higher than usual number of tropical storms and hurricanes for 2006. Still, I don't think we'll see anything like this year's level of activity, which was a once-in-a-lifetime hurricane season. Dr. Bill Gray's team at Colorado State University issues their first forecast for the 2006 hurricane season on Tuesday, December 6, and we'll talk more about next year's season then.
As we discussed yesterday, the official woolly bear caterpillar forecast for the U.S. mid-Atlantic coast area was for a warmer than average winter. The official NOAA 90-day forecast for the upcoming winter, issued on October 20 by the Climate Prediction Center (CPC), disagrees, calling for equal chances of an above average or below average winter over the eastern half of the country, but a higher than average chance of a warmer than average winter over the western half of the country. How well has the official forecast done in recent years? NOAA rates its forecasts using the Heidke skill score, which is a measure of how well a forecast did relative to a randomly selected forecast. A score of 0 means that the forecast did no better than what would be expected by chance. A score of 100 depicts a "perfect" forecast and a score of -50 depicts the "worst possible" forecast. For the 90-day temperature forecasts issued 1.5 months in advance done in January through May of this year, the Heidke skill score was greater than zero for two of the forecasts, less than zero two of the forecasts, and about zero the other forecast. The Heidke skill score for 90-day temperature forecasts issued 1.5 months in advance has averaged 8 the past ten years (see Figure 1.) So, while there is some skill in forecasting what the winter will be like 1.5 months in advance, this skill is not much better than flipping a coin or relying on woolly bear caterpillars. Let's look at some examples from forecasts for previous winters issued at about this time of year. The 90-day forecast done in mid-October of 1999 for the winter of 2000 was awesome, with a Heidke skill score of 50. However, the 90-day forecast done in mid-October of 2000 for the winter of 2001 was horrible, with a Heidke skill score of -15.
Figure 1. Skill of the official 90-day forecasts issued 1.5 months in advance by NOAA's Climate Prediction Center. Note that the average skill is positive, but has remained flat the past ten years, indicating that our skill in making long-range forecasts has not improved.
Why do seasonal forecasts do so poorly? It's partially because our physical understanding of what controls the climate is so poor. It's also in large part due to the fact that the long-term weather is chaotic and fundamentally unpredictable by nature, and no amount of physical understanding will help us. So, pick your forecast: woolly bear, coin flip, NOAA--the three techniques have similar levels of accuracy. Or you can check out the predictions of Psychic Helane, who wrote me to say she had correctly forecast the impact of Hurricane Wilma on South Florida over one month in advance. Her winter prediction calls for "the north especially St. Paul, Minnesota and Upstate NY will see its worst winter in memory." If I were an energy futures trader, I would wait until I saw a longer track record for Helane's weather predictions. She is also calling for a "huge hurricane in Alabama" this month, which is a virtual impossibility, given that sea surface temperatures in the Gulf of Mexico are now less than 80F (26.5C), which is too cold to support a major hurricane. Go with the official NOAA forecasts, which a have a proven track record of some having a least a little skill compared to chance.
Jeff Masters
As we discussed yesterday, the official woolly bear caterpillar forecast for the U.S. mid-Atlantic coast area was for a warmer than average winter. The official NOAA 90-day forecast for the upcoming winter, issued on October 20 by the Climate Prediction Center (CPC), disagrees, calling for equal chances of an above average or below average winter over the eastern half of the country, but a higher than average chance of a warmer than average winter over the western half of the country. How well has the official forecast done in recent years? NOAA rates its forecasts using the Heidke skill score, which is a measure of how well a forecast did relative to a randomly selected forecast. A score of 0 means that the forecast did no better than what would be expected by chance. A score of 100 depicts a "perfect" forecast and a score of -50 depicts the "worst possible" forecast. For the 90-day temperature forecasts issued 1.5 months in advance done in January through May of this year, the Heidke skill score was greater than zero for two of the forecasts, less than zero two of the forecasts, and about zero the other forecast. The Heidke skill score for 90-day temperature forecasts issued 1.5 months in advance has averaged 8 the past ten years (see Figure 1.) So, while there is some skill in forecasting what the winter will be like 1.5 months in advance, this skill is not much better than flipping a coin or relying on woolly bear caterpillars. Let's look at some examples from forecasts for previous winters issued at about this time of year. The 90-day forecast done in mid-October of 1999 for the winter of 2000 was awesome, with a Heidke skill score of 50. However, the 90-day forecast done in mid-October of 2000 for the winter of 2001 was horrible, with a Heidke skill score of -15.
Figure 1. Skill of the official 90-day forecasts issued 1.5 months in advance by NOAA's Climate Prediction Center. Note that the average skill is positive, but has remained flat the past ten years, indicating that our skill in making long-range forecasts has not improved.
Why do seasonal forecasts do so poorly? It's partially because our physical understanding of what controls the climate is so poor. It's also in large part due to the fact that the long-term weather is chaotic and fundamentally unpredictable by nature, and no amount of physical understanding will help us. So, pick your forecast: woolly bear, coin flip, NOAA--the three techniques have similar levels of accuracy. Or you can check out the predictions of Psychic Helane, who wrote me to say she had correctly forecast the impact of Hurricane Wilma on South Florida over one month in advance. Her winter prediction calls for "the north especially St. Paul, Minnesota and Upstate NY will see its worst winter in memory." If I were an energy futures trader, I would wait until I saw a longer track record for Helane's weather predictions. She is also calling for a "huge hurricane in Alabama" this month, which is a virtual impossibility, given that sea surface temperatures in the Gulf of Mexico are now less than 80F (26.5C), which is too cold to support a major hurricane. Go with the official NOAA forecasts, which a have a proven track record of some having a least a little skill compared to chance.
Jeff Masters
The views of the author are his/her own and do not necessarily represent the position of The Weather Company or its parent, IBM.