"So, that’s why I’m tending to go with the cold and snow pattern with 25 to 35 inches for the season."
This all started for me in 1997.
A very strong El Nino was developing in the tropical Pacific. There was an article in a magazine correlating strong El Nino’s with warm winters. So I went on the air and predicted a very warm winter with very little snow.
All three winter months (December, January, and February) made the Top-10 warmest lists in Philadelphia and we had less than one inch of snow during the whole winter!
It worked out so well that viewers expected me to do the same thing the next year. The whole thing started -- pardon the pun -- snowballing and before too long the other stations in town were making winter forecasts (often, coincidently? the same night NBC 10's airs).
Seasonal forecasting has come a long way in the past 15 years. But, no one can predict the winter as well as we do with short-term forecasting.
We try to be 90 percent accurate in the forecast for tomorrow but can only hope to do better than 50 percent in the long run. That would indicate “skill” in such forecasts.
In the past 11 years, I’m probably in the 60 to 70 percent accuracy range depending on how strictly it’s judged. It’s very hard to quantify accuracy of such forecasts, partly because I’ve gotten more specific over the years.
Others remain more generic, so how can they be compared?
I just hope that overall the long-range forecast has been of help to those who care about what is likely to happen in the most crucial season of the year. Last year turned out well, accurately predicting very little snow and above average temperatures. Some of the details were off, but the general trend was right on the mark.
The pressure exists now to put out a forecast every November. Frankly, there have been some years when I wish I could just pass but that doesn’t occur in the real world.
This year is far from a sure bet but there are strong signs of certain things.
Here it goes:
The first thing I look at is trends. While the atmosphere varies greatly each year and no two seasons are alike, there are tendencies that show up when we look to the past.
For example, we had only 6.3 inches of snow last winter. What happens in the winter after one with less than 10 inches of snow? The tendency is for the season to be snowier than average.
In fact, some of our snowiest winters occurred after a small snow season.
- Our record winter of 1995-'96 was preceded by a winter with 9.8 inches.
- The year before our 2002-3 season of 46.3 inches we only had 4 inches!
On average, we get 22 inches after a low snow season above the long-term average of 19.3 inches.
Over the past 20 years, the average is more than 25 inches.
A second trend is what has happened in October and November. Cold October and November’s tend to have colder and snowier winters than average.
Tony Gigi of the National Weather Service has done much of the research in this area. This October was a bit below average with temperatures -- after a mild start, November may turn out to be below average, too.
Another trend is the frequent coastal storms we’ve seen in the past couple of months. The atmosphere is obviously favoring a coastal storm track, which can be favorable for big winter storms if there is enough cold air.
Now we get to the overall conditions in the oceans and atmosphere, leading off with La Nina or El Nino.
Last year was at least a moderate La Nino, which often leads to mild winters.
This year, we are much closer to a neutral East Pacific. Looking back in history shows neutral years tend to be colder and snowy here.
Here’s the comparison from last year to now: 2007 vs. 2008
Just as the El Nino/La Nina is an oscillation in the ocean (called ENSO), there are others in the atmosphere that can be detected.
-
The Pacific Decadal Oscillation (PDO) which is related to the ocean temperature pattern in the Pacific.
-
The Quasi-Biennial Oscillation (QBO), where high winds switch from easterly to westerly every couple of years or so.
-
And the better known North Atlantic Oscillation (NAO), which is related to ocean temperatures in the North Atlantic.
The QBO stage, in particular, is the opposite of what we saw last year and the PDO is also trending toward the opposite stage as last year. The NAO is still so variable that it is basically unpredictable months in advance.
When we look at all the indexes, we try to find previous years with similar conditions. They are called "analogs." The "analogs" have been very helpful in previous winter forecasts.
This time, the closest years to this one are 1959-60, 1961-62 and 1966-67. The winter of 1961-62 was the closest analog I found.
Let’s see what happened in those winters.
Amazingly, all three had very similar weather -- basically cold and snowy. Furthermore:
-
Each November had some snow
-
December had significant snow
-
January was the relatively warmest month compared to average with very little snow
-
February had significant amounts of snow
-
March had quite a bit of cold and snow.
So, that’s why I’m tending to go with the cold and snow pattern with 25 to 35 inches for the season.
The snow season should start early and end late compared to average around here. There also will be quite a number of days with temperatures below 20 degrees.
Interestingly, the National Weather Service’s Climate Prediction Center is predicting another very mild winter, so my forecast goes against that. That’s because it’s not a guarantee, folks -- it’s a forecast. Please keep that in mind if I fall on my face.
Comments (55)
Post a Comment