A Tricky Recipe:
The winter outlook in the Midwest 
| By Tony Schumacher |
|
In the 1960s and 1970s, Upper Midwest winters averaged a few degrees colder than the long-term normals, while in the last 20 years it's been a few degrees warmer than normal. In fact, the last three winters have been some of the warmest
on record this century for the entire U.S. This has left many skiers in our area exasperated, feeling like the "good old winters" they once knew will never happen again. You can be sure the meteorologists at the Climate Prediction Center wonder about the state of our winters as well, although they take the nostalgia out of it, and just look at the data and facts.
The Climate Prediction Center is the part of the National Weather Service that
deals with long-range weather outlooks. Extended seasonal forecasts are nothing new. In fact, our third president, Thomas Jefferson, dabbled with them in the late 1700s. In the late 1800s the federal government assigned the Army Signal Corps with the mission of defining the climate of regions of the country being opened for farming. In 1890 the U.S. Department of Agriculture created the Weather Bureau Climate and
Crop Services, which published a weekly bulletin on such things as the weather patterns.
Winter 2000-2001 Outlook
Here's the good news for winter enthusiasts: The Climate Prediction Center indicates temperatures could average 4 to 7 degrees colder this winter (than the past three) in an area from New England through the
Great Lakes into the Northern Rockies. More frequent cold air outbreaks should hit the upper Midwest, thus increasing lake effect snows in comparison to recent years. This seems reasonable given that a foot of lake effect snow hit parts of far northern Wisconsin and Upper Michigan already on October 6-8 of this year. Precipitation levels as a whole in inland areas are expected to be near normal in the Midwest.
Warmer-than-normal temperatures are anticipated south a line from Washington State to Colorado to North Carolina. Above normal precipitation is foreseen from Texas to Alabama northeast into the Mid-Atlantic States. This stormy swath should develop in the area where a well-developed polar jet stream, diving into the southern U.S., interacts with the subtropical jet stream. Perhaps some nasty snowstorms will roll up the Appalachians towards the big cities, like Washington, D.C.
How in the world do forecasters come up with seasonal outlooks in the first place? After all, it's hard enough to get the forecast right three or four days ahead. First, keep in mind that long-range outlooks don't attempt to say it will rain or snow or be cold on a particular day; they basically define trends and averages over the course of the season. Even in a forecast "normal winter," there will be wild warm and cold spells,
periods of no snow and periods of heavy snow. To be sure, there is nothing easy or simple about looking out several months. In the following section I will attempt to briefly describe the tools the Climate Prediction Center uses to develop a long-range outlook. I will warn you, the terminology and concepts are quite heavy. Don't get mad at me if it's boring; I was encouraged by editor Greg Marr to tackle this subject (nothing personal, Greg).
Tools Used to Make Long-Range Forecasts
The Climate Prediction Center uses about a half-dozen computer models for its long-range predictions. Each one emphasizes different variables of the earth-ocean-atmosphere system. The models crunch huge
amounts of historic and recent weather data. It's then up to the team of human experts to decide which model solutions will receive the most weight in their final outlook. Here is a breakdown of those models:
The coupled model is forced primarily by atmospheric interactions with tropical Pacific sea surface temperature and recently has incorporated soil moisture. This model is particularly useful when El Niņo or
La Niņa conditions are present or are expected to develop in the forecast period.
The canonical correlation analysis linearly predicts the evolutions of weather patterns based on global sea surface temperatures, surface-to-10,000-foot altitude temperatures, and U.S. surface temperatures and precipitation from the past year. It emphasizes El Niņo/La Niņa to a lesser extent than the coupled model.
The composite analysis provides guidance for the U.S. based on historical effects of moderate or strong El Niņo/La Niņa events on various regions. For example, if Wisconsin has experienced warmer-than-normal winters in eight out of 10 strong El Niņo events this century, you can guess what it would predict for an upcoming winter where a strong El Niņo was expected.
The optimal climate normals predicts weather patterns on the basis of persistence of the observed average
anomalies for a given season during the last 10 years for temperature, and the last 15 years for precipitation. Global warming trends are taken into consideration by this model.
The constructed analog on soil moisture is driven by historical data on soil moisture conditions and associated weather effects in the U.S. since 1932. The theory here is that dry soil heats up faster than wet
soil. This model is ignored in winter predictions because with the low sun angle, the wet vs. dry land heating differential is relatively small.
The screening multiple linear regression model uses the same predictor fields as the canonical correlation analysis model except it is applied to single stations rather than regional patterns. It does better with the local effects like urban heat islands, terrain features and adjacent water bodies.
Can you trust long-range weather outlooks?
Accuracy of long-range outlooks (30 to 120 days out) is highest for late winter with a secondary peak for late summer. It is lowest for late spring and late fall. For all the Climate Prediction Center models,
precipitation forecasts are generally less skillful than for temperature. However, when strong El Niņo or La Niņa conditions are present, precipitation skill can be as high as temperature skill for the winter season in many parts of the U.S. Strong La Niņa conditions imply the possibility of moderate precipitation skill for parts of the summer season as well.
The long-range outlooks for the past four winters have been 65 to 90 percent accurate in terms of
temperatures. Precipitation accuracy has been more in the 30 percent to 70 percent range for the same period. Taken through the year as a whole, as you might expect, there are some nasty undulations in long-range accuracy, going down to near 0 percent at times. Obviously we still have a ways to go on figuring out the atmosphere.
If you would like more information on long-range weather outlooks, including some neat charts and maps to
the data I discussed in this article, go to the following website:
HYPERLINK http://www.cpc.ncep.noaa.gov/
In the climate highlights area, click on winter outlook. Try to enjoy whatever winter throws our way but I hope by the time you're reading this, you'll have already used your cross country skis!
| 
