Across the United States, meteorological fall is coming to an end. Trees in the northern states have already shed their leaves, and those in the south are rapidly following suit. Naturally, this is often the time of year that people start to ask the question, “what will winter be like this year?” Each winter season seems to take on its own character, and the atmosphere certainly does have a tendency to settle into a particular pattern for much of the winter season. While we here at MetStat have historically not offered seasonal forecasts, we thought we would try our hand at it this year in the interest of identifying regions of the country that we believe to be most at risk for unusual extreme precipitation events this winter.
Numerous factors play a role in shaping the winter pattern that eventually materializes, and some are more predictable than others. Global sea surface temperature patterns are known to play a role in altering the position and strength of the jet streams, thereby affecting where storms and precipitation have the greatest and most frequent impacts. The best example of this for North America is the El Niño-Southern Oscillation (ENSO) index, which describes the tendency for surface waters in the eastern tropical Pacific Ocean to be either colder or warmer than normal, depending on the phase of the oscillation. These patterns of warmer and colder tropical waters affect where clusters of thunderstorms in the tropics are enhanced and suppressed, which in turn alters the course of the jet stream in the subtropics. This alteration of the subtropical jet stream has the most direct impact on those of us in North America, as illustrated in the diagram below (Figure 1).
During El Niño winters, as we had during 2014-2015 and 2015-2016, the subtropical jet stream is very active, and typically leads to the southern United States experiencing wetter than normal conditions. La Niña winters, in contrast, tend to have a very weak subtropical jet stream, which leaves much of the southern U.S. warm and dry, but with the potential for cold air outbreaks across the eastern U.S. It also tends to usher in a consistent train of wet Pacific storms throughout the Pacific Northwest states and British Columbia. Oftentimes, a secondary wet region develops in the Ohio Valley due to the frequent development of mid-latitude cyclones in the central and southeastern U.S. (Figure 2). The Climate Prediction Center recently declared that we have entered a La Niña state, and the pattern for much of October and November has strongly resembled the “prototypical La Niña” shown in Figure 2.
While the ENSO state is arguably the single largest factor in setting the North American winter weather pattern, other indices that assess atmospheric conditions can also be useful for putting together a long-term outlook. The North Atlantic Oscillation (NAO) is one such index, and it describes the deviation from normal pressures between the Greenland/Iceland vicinity and far western Europe. However, it also has implications for U.S. weather, particularly along the east coast. By definition, such indices distill very complex atmospheric patterns into simple, easy-to-comprehend numbers (e.g., the NAO index was +0.71 for October 2017). While these one-dimensional numbers are certainly an oversimplification of hugely variable atmospheric conditions, they give us a simple way to compare one time period to another. By identifying past years that had fall and early winter patterns similar to what we’ve observed and are expecting this year, we can draw some conclusions about how this winter’s pattern is likely to evolve. This process is known as analogue forecasting.
To choose our analogue years, we must first identify current conditions. Figure 3 illustrates the sea surface temperature pattern over this past October in terms of a difference from normal (1971-2000 average). The circled area, showing a long, skinny area of cooler-than-normal waters, is indicative of classic La Niña conditions. The small anomalies in the northeast Pacific Ocean are indicative of a near-neutral Pacific Decadal Oscillation, which is another important sea surface temperature pattern. Therefore, we have focused our analogue year search on other La Niña years, with a preference for La Niñas that are in their second year (as ours currently is), and filtered our selected years by similar PDO and NAO conditions. This gave us a list of 10 potential years to consider, which were then pared down by comparing the pattern of precipitation anomalies observed in August, September, and October of those years to the Aug-Oct precipitation anomalies this year (Figure 4). The five most similar years were as follows (chronological order): 1971-1972, 1981-1982, 1996-1997, 1999-2000, and 2008-2009. Characteristics of these years as they pertain to the important patterns discussed previously are summarized in Table 1.
Table 1. The general state of several atmospheric indices for the chosen analogue years, as compared to 2017 so far.
|Year||El Niño-Southern Oscillation (ENSO)||Pacific Decadal Oscillation (PDO)||North Atlantic Oscillation (NAO)|
|2017||2nd year weak La Niña||Strongly positive last few years, fading to neutral||Near-neutral|
|1971-1972||2nd year moderate La Niña||Predominantly negative||Strongly negative to positive|
|1981-1982||2nd year cool-neutral||Strongly positive weakening into fall 1981||Predominantly negative|
|1996-1997||Just below La Niña threshold after weak-moderate Niña in 1995-96||Strongly positive weakening to neutral||Predominantly negative|
|1999-2000||2nd year strong La Niña||Strongly negative||Predominantly positive|
|2008-2009||2nd year weak La Niña following strong 2007-2008 Niña||Strongly negative||Predominantly negative|
As you can see, considerable similarities exist in the general pattern. An area of dryness was present in 2017 from California through the desert southwest and parts of the Rocky Mountains, with a similar area from southern California through Nevada and Idaho in the composite. A corridor of wetter-than-normal conditions stretches through the Plains from Texas to Minnesota as well, with patchy dryness in middle Mississippi and lower Ohio River valleys. Given the similarities in the late summer/early fall weather patterns, we might expect similar conditions to those observed in the analogue years this winter. So what did those five winters bring in terms of precipitation across the U.S.? Figure 5 has the answer:
Notice something about this pattern? It looks very similar to the “average” La Niña conditions depicted in Figure 2, with wetter-than-normal conditions in the Pacific Northwest, Northern Rockies, and the Ohio Valley/Great Lakes region, with drier-than-normal conditions across much of the southern tier of states from southern California through Florida. Given that the first 2/3rds of November so far in 2017 have borne a strong resemblance to this pattern (Figure 6), we believe it is a reasonable expectation that the winter of 2017-2018 will end up quite similar to this composite. Therefore, the two areas we believe to be most at risk for extreme precipitation are:
- The Pacific Northwest and the Northern Rockies, including Washington, Oregon, Idaho, and the mountains of Montana and Wyoming; and
- The Great Lakes states, including Wisconsin, Michigan, Illinois, Indiana, Ohio, and the lake-effect snow prone regions of Pennsylvania and New York.
The La Niña pattern means lots of strong coastal storms will be possible for the Pacific Northwest, thus the elevated heavy rain risk there; it also means lots of northwest flow across the upper Midwest and Great Lakes states, leading to the potential for prodigious lake-effect snows. La Niñas are also associated with more low pressure systems crossing the Ohio Valley, meaning higher potential for widespread heavy rain and snow exists. Sometimes such storms can even take more of a coastal track, resulting in “Nor’easters” that produce heavy rain and/or snow up much of the coastal northeast. The northern Plains states also have the potential to be wetter-than-normal, as shown in Figure 5, although La Niña conditions most often result in frequent light snow events for this area rather than a few extremely large events.
We hope you enjoyed reading our thoughts on extreme precipitation potential this winter! Be sure to check back with us for analysis of extreme storms as they happen.