As of this week, meteorological winter has officially begun in the Northern Hemisphere. For many in the central and eastern United States, an atypically cold and snowy November has provided a jump-start to the winter weather. As we did around this time last year, we at MetStat wanted to provide some thoughts on what the upcoming season has in store across the country, including the regions that we believe might be at the greatest risk for flooding and above-normal precipitation. What follows is our 2018-2019 winter outlook for the Lower 48, with a particular emphasis on regions expected to be at risk of intense rain events.
Similar to last year, this year’s forecast is facilitated in part by the presence of a robust ENSO (El Nino-Southern Oscillation). ENSO itself is a naturally-occurring, short-term climate cycle that alters global weather patterns by changing the structure of sea surface temperatures in the eastern tropical Pacific Ocean. The effects it has are global, although the strength and significance of those effects varies widely depending on the exact region of interest. Across the continental United States, a positive ENSO state (colloquially “El Nino”) generally results in a tier of wetter-than-average weather stretching from California through the south-central and southeastern U.S., with drier conditions across the mid-Mississippi River valley and lower Great Lakes states (Figure 1). (A more complete overview of ENSO and the typical characteristics of an event may be found at climate.gov.) This is particularly relevant to this winter’s forecast since El Nino conditions are currently developing in the east Pacific, with a weak-to-moderate El Nino event expected to be active through the winter and possibly into spring.
While ENSO has perhaps the most robust influence on North American weather of any of the well-documented atmospheric indices, incorporating other indices into our assessment can offer a more complete picture of the atmospheric state. The Pacific Decadal Oscillation provides one such metric, as it encapsulates the mean sea surface temperature anomaly (difference from normal) pattern across the northern Pacific Ocean (details at https://climatedataguide.ucar.edu/climate-data/pacific-decadal-oscillation-pdo-definition-and-indices). The PDO is a longer-term pattern that generally remains predominantly in either a cool or warm phase for 20 years or more, but can have short-term deviations from its longer term pattern. Like ENSO, it has effects on the strength and location of storms and moisture sources that form across the eastern Pacific Ocean and move across North America. The map in figure 2 shows the correlations between states of the PDO (left) and ENSO (right) indices and observed precipitation changes across the U.S. (Figure 2).
Note that the two patterns are pretty similar, with +PDO and +ENSO events both causing somewhat drier conditions in the Northwest and Ohio Valley with wetter conditions across the Southwest and Southeast. The reasons for this similarity are many and complex, but in short it’s because the two patterns often work together to modulate weather around the world. As in last year’s extreme precipitation winter outlook, we will identify years that had similar ENSO conditions to those expected this year and pare down the list by considering years that also have similar PDO and North Atlantic Oscillation (NAO) conditions. This is known as analogue forecasting, and it the case of our outlook last year, it worked quite well!
First, as mentioned briefly above, the eastern tropical Pacific Ocean appears to be drifting towards El Nino conditions, as illustrated in Figure 3. The stripe of relatively warm water sitting on the equator from around 90°W through 150°W is the most tell-tale sign. This follows the weak-to-moderate La Nina (cool phase) that we had last year, which was in turn the 2nd straight year of La Nina conditions. Therefore, we’ve identified other recent El Ninos that closely fit this year’s pattern, initially identifying twelve seasons. The observed mean temperature and precipitation patterns from each of those August through October seasons was then analyzed, and the most similar five years to what we experienced this fall were selected. The states of each of the indices discussed for those seasons is summarized in the table below.
|Year||El Nino-Southern Oscillation (ENSO)||Pacific Decadal Oscillation (PDO)||North Atlantic Oscillation (NAO)|
|2018||Weak-Moderate El Nino, follows 2016-2018 La Nina||Neutral, strongly positive last few years||Generally positive, possibly trending downward|
|1977-1978||Weak 2nd-year El Nino||Neutral||Neutral, trended negative|
|1986-1987||Moderate El Nino, followed 2-3 year cool phase||Positive||Moderately positive|
|2002-2003||Moderate-strong El Nino, followed ~3-year cool phase||Negative||Moderately negative|
|2006-2007||Weak El Nino, followed weak La Nina||Negative||Strongly negative but trended upward through winter|
|2009-2010||Strong El Nino after 2-year La Nina||Negative||Neutral, trended negative|
The precipitation anomalies for the five selected years’ August-October seasons are compared to August-October precipitation anomalies for this year in Figure 4. Clearly, similar patterns exist, with wetter-than-average conditions prominent across the country’s heartland, and drier-than-normal conditions across Florida and parts of the West Coast. Some differences are also evident, such as the dryness that was observed across the central Rockies and northern Plains this year that is not present in the average of the five analogue years, but the general patterns match quite well.
Now that we’ve established that the five analogue years produced broadly similar fall precipitation across the Lower 48 to what we saw in 2018, the question becomes: what was winter precipitation like for those five years? As revealed in Figure 5, the pattern is generally similar to what you’d expect from the “typical” El Nino year illustrated in Figure 1. However, given that other factors serve to modulate the El Nino pattern in any given year, it should come as no surprise that there’s not an exact match. For example, the northern Plains from Colorado through the Dakotas and Montana were, on the average, a fair bit wetter than normal. In contrast, virtually the entire Midwest experienced notably dry conditions, perhaps representing a northwestward shift of the Ohio Valley dry region portrayed in the “typical” pattern. However, we also see the belt of wetter-than-normal conditions expected from El Nino stretching from southern California through the Gulf Coast and along the southeastern Atlantic seaboard.
As mentioned earlier, winter got off to a roaring start for many in the eastern U.S., with outbreaks of cold and snow prominent across the Midwest, Northeast, and South during the month of November. Given that our analogue matching window was Aug-Oct, this gives us the opportunity to test our hypothesized winter pattern by examining how precipitation anomalies have looked over more recent days. Figure 6 shows the percentage of normal precipitation that was observed from November 5th through yesterday, December 4th, across the Lower 48. Clearly it’s not a perfect match to our forecast pattern: the dryness across Texas and New Mexico as well as the somewhat wetter-than-normal Midwest and eastern Ohio Valley don’t quite fit our expected pattern. However, the dry Pacific Northwest, the wet Southeast, and the wet High Plains (eastern Colorado through the Dakotas and Montana) all fit our expected pattern nicely.
As El Nino ramps up this winter, we expect those dry pockets in the desert Southwest and western Texas to start to fill in, and the storm track will likely continue to shift southeast, leaving the Midwest and Ohio Valley drier. In fact, there’s already some evidence of this in current forecast models, which indicate several opportunities for precipitation across the southern tier of the U.S. in the next 10 days or so (Figure 7). Meanwhile, the upper Midwest and northern Plains look to dry out, at least temporarily.
So what does this mean for extreme precipitation this winter? In general, the locations where the storm tracks are most likely to be persistent is where the greatest risk of heavy rain and flooding will be. This year, that means events stretching from Texas through the southeastern U.S. (and notice that Figure 7 indicates that very heavy rain may be in the cards as early as next week for parts of Texas). Big storms may also be possible across southern California and the desert southwest, which will pose a particular problem where heavy rain coincides with recent wildfire burn scars (as we’ve already seen near the massive Camp Fire burn area in November). The major metropolitan areas of the mid-Atlantic and Northeast may also be threatened by multiple heavy precipitation events this year, although whether they come in the form of rain or snow will go a long ways towards dictating what, if any, flood risk materializes. Our final area to watch is in the central and northern High Plains, but winter flood risks are expected to be minimal here: these are fairly dry areas, so above-average precipitation still typically does not manifest itself in big floods. Most of the precipitation is also likely to fall as snow, dampening the immediate impacts of a heavy precipitation event.
In short: we’ll be watching for some big snows along the Mid-Atlantic coast and High Plains, with some heavy rain events across the South and Southwest. Watch this space and our Facebook page for updates on major events as they happen.