The WeatherOptics Winter Forecast goes in-depth regarding the various environmental and modular variables that play a role in how winter will play out for different regions. If you’re only here to see our final takeaways, read the first few paragraphs. If you’re interested in our full analysis, read our full discussion:
What To Expect This Winter – In Short:
Seasonal forecasts are products of statistical interpretations. Short observational periods and a wide array of pattern forcing mechanisms makes long-range forecasts impossible to predict with certainty under most conditions. The winter of 2018-2019 will likely feature El Niño, one of the most profound and most understood forcing seasonal forcing mechanisms. In its weak or moderate phase, El Niño will contribute to a wet winter in Northern California, the Southeast, and parts of Texas. At the same time, it will contribute to drier than normal precipitation in parts of the Northwest. El Niño may also contribute to a warm winter for the Plains and Upper Great Lakes. Correlation is too weak elsewhere and for other factors to include possible impacts from El Niño.
Very warm SSTs near Alaska will contribute to dry weather in the Northwest while decreasing the southeastward penetration of storms from Canada until the second half of winter. The Midwest and the Northeast can therefore expect a larger proportion of storms to originate along the Rocky Mountains or Gulf of Mexico until the second half of winter. Warmer SSTs along the East Coast will enhance any storms that develop there, keeping the Eastern Seaboard wet throughout the winter. Big Northeast snowstorms will be most probable in the second half of winter, when cold shots may finally penetrate as far south as Florida. Some of the storms leading the cold will cross (or originate out of) the Rocky Mountains, where snow will likely be more frequent during the second half of winter.
What To Expect This Winter – Full Analysis:
A brewing El Nino in the equatorial Pacific has made the rounds across social media this fall as a likely indicator for a cold and snowy winter in the Northeast. For snow lovers, the scientific literature on El Nino yields some unfortunate news. But that doesn’t necessarily mean winter will be absent (at all). After a mild December, the worst of winter’s cold and snow will arrive during February in the Northeast. Many of the storms that will eventually bring snow to the Northeast will originate on the lee side of the Rocky Mountains or the Gulf of Mexico. Consequentially the Southeast will generally face a consistently wet winter, with increasing cold shots in February along with the Northeast. Meanwhile, El Nino will likely maintain mild and wet conditions throughout the West Coast in January and February, with frequent thaws reaching the Central US during this period.
ENSO (El Nino Southern Oscillation) will be the most influential factor on the upcoming winter season in North America. ENSO’s warm phase is defined as El Nino, when sea surface temperate (SST) anomalies reach or exceed +0.5°C in the equatorial Pacific for at least three months. Only monthly mean SSTs in the Niño3.4 region outlined in the figure below provided by NOAA are considered in NOAA’s Oceanic Niño Index, as studies have shown this region to be most representative of ENSO conditions.
Below is a gif depicting the weekly SST anomalies in the equatorial pacific between August 8 and October 24. Notably, the anomalies have increased in the Niño 3.4 region since early October.
NOAA has issued a 70-75% chance of an El Niño developing over the next few months based off of these trends and among a consensus among numerical climate models that these anomalies will persist well into 2019.
Another seasonal climate system that is heavily relied on in this winter forecast is the North American Multi-Model Ensemble (NMME). The NMME is a coupled atmosphere-ocean seasonal forecasting system comprised of climate models from NOAA and Canada’s CMC. Studies have shown that multi-model ensembles improve predictability of weather and climate and increase the ability to characterize forecast uncertainty. The NMME is run once per month, using initial conditions (IC) from the previous month. Familiar climate models like NCEP’s Climate Forecast System (CFS), the NOAA Geophysical Fluid Dynamics Laboratory Climate Model (GFDL CM) as well as Environment Canada’s Climate Model (CMC CM) are themselves ensembles and are incorporated in the NMME.
The plot below provided by the Climate Prediction Center (CPC) NMME website depicts the consensus among the various climate ensembles that comprise the NMME. It essentially shows that SST anomalies observed over the last month in the Niño 3.4 region will increase through January, with El Niño conditions persisting into May.
The mean anomaly is forecast to peak at about +1°C, just above the threshold for a “moderate” El Niño. El Niño conditions are classified as “weak” if the SST anomaly is between 0.5°C and 0.9°C, “moderate” between 1.0°C and 1.4°C, “strong” between 1.5°C and 1.9°C, and “very strong” if greater than 2.0°C. Each classification favors different seasonal impacts across North America. The observed impacts on temperature and precipitation resulting from every observed El Niño event since 1950 are presented below, courtesy of NOAA.
Only four moderate El Niño events have been observed since 1950. Thus, such a sample size is too small to draw conclusions. However, the El Niño may very well remain categorized as weak, for which twelve events have been observed. Seven of these events corresponded to warmer than normal winter temperatures in the West, with eight corresponding to colder than normal temperatures in the East. Very few of these episodes were met with a significant precipitation response, but only half of the weak episodes corresponded to wetter than normal precipitation anywhere in California, a response typically associated with El Niño. The precipitation response is notably more significant during strong El Niño seasons.
Regardless of any interpretations, a sample size of twelve is too small to make statistically significant inferences based off these observations alone. Many studies on El Niño’s impacts on temperature and precipitation use a computational technique known as bootstrapping to resample the observations into much larger sample sizes. A 2012 collaborative study between NOAA Pacific Marine Environmental Laboratory and the University of Washington performed a bootstrapping technique on traditional El Niño years and those identified by anomalies in outgoing long wave radiation. The study found that very few of the traditional El Niño events since 1979 produced statistically significant weather anomalies in North America (Chiodi and Harrison). Only the four cases identified by the outgoing long-wave radiation behavior produced the statistically significant anomalies shown below.
The referenced study only considers nine El Niño events since 1979, so it does not contain the entire range of observations in its bootstrapping sample. Nevertheless, this study strongly suggests that when El Niño is important, its strongest impacts are on precipitation in the Southeast and Gulf Cost and temperature in the north-central US. Impacts from the upcoming El Niño could very well produce a wetter than normal winter in the southeast and Texas, and a warmer than normal winter in the north-central US. Elsewhere, El Niño is not a relatively helpful indicator.
What about those strong precipitation signals in California during strong El Niño events? A 2017 publication from the American Geophysical Union with contributions from NOAA and research universities across the US suggests those anomalies will only occur if the anomalous SSTs resulting in El Niño-like conditions occurs in the eastern Pacific. The anomalies so far are based in the central-Pacific, although various climate models suggest that may change. But given the current state of the Pacific SST anomalies, the impact of this weak El Niño in California will be slim to none. A wetter than normal winter is still possible, but it would likely be the result of other factors.
That ENSO is the most significant teleconnection driving global weather but impacts from El Niño (its warm phase) are so ambiguous for much of the US means one should look elsewhere for insight about the winter. Other commonly referenced teleconnections like the North Atlantic Oscillation (NAO) and the Arctic Oscillation (AO) have poor predictability beyond 14 days such that other observations and climate models are the best remaining sources of insight available.
Climate models naturally have a lot of variability. The dichotomy alone between the CFS and the GFDL for December, just one month from this writing, is quite remarkable. The plots presented below from the CPC NMME website are enough to induce many sleepless nights for a long-range forecaster. The former favors a warm December for the entire continental US. The latter favors a cold December for almost the entire lower 48, only keeping Alaska warmer than usual.
Fortunately, the NMME clears some of this uncertainty by averaging the forecasts for all of its component members. Skill plots with NMME component forecasts for historical observations from 1981-2010 set as a reference indicate that the NMME generates more accurate seasonal forecasts than its components.
Evidently, the GFDL climate model is an outlier. The NMME forecasts a warmer than normal winter for most of the US as the following plot demonstrates. Not shown here, the NMME actually intensifies the warm anomaly in the north-central US, as expected during El Niño per Chiodi and Harrison. It simultaneously reduces the anomaly to neutral in the East by January, suggesting that any remaining warmth will be offset by cold shots. Individual members of the NMME generally follow a similar trend. It can therefore be said with reasonable confidence that the East will end colder than it begins.
In the Northeast, mild temperatures are likely to persist into early or middle January, before a pattern reversal to colder than normal-temperatures by late January. The cold shots will eventually work their way to the Southeast. Meanwhile, the Central US, especially the Northern Plains and Upper Great Lakes, are expected to start mild, with warm air intrusion from the Gulf of Mexico becoming more significant as winter wears on.
Precipitation, especially snowfall, is a lot trickier to forecast than temperature. There is high variability among individual members of the NMME, resulting in very small mean forecast anomalies. Precipitation signals are strong in the tropics, while they are much weaker over continents. In the US, only the coasts are forecast to experience precipitation anomalies. Since it is statistically unreasonable to assume there will be no precipitation anomaly this winter across much of North America, it is necessary to employ other factors that could help favor a particular forecast from the NMME’s component members.
SSTs anomalies can help identify the most likely precipitation solution. Though the climate models are atmosphere-ocean coupled, they still struggle representing the impact of SST anomalies on weather patterns. Nowhere on Earth are the anomalies as extreme as those surrounding Alaska, especially north and south of the Bering Straight. These warm anomalies have already contributed to Alaska’s fourth warmest fall on record by promoting the development of ridging of high pressure. This high pressure has already reduced baroclinicty (the strength of temperature gradients) and diverted storm paths away from northwestern North America. These storms are essential in propagating cold air eastward across North America.
Consequentially, Alaska and the Pacific Northwest are expected to remain relatively dry throughout most of the winter. Early season snowstorms often arise from systems originating over western Canada in the Midwest and Northeast, so big snowstorms are unlikely in these regions until storms begin to tap into Gulf of Mexico moisture, a factor that will become more likely the second half of winter when El Niño intensifies.
Some storms originating from disturbances in western Canada will still reach the Northeast coast this winter, even early in the season. Brewing cold in Siberia can help break down the ridge over Alaska this winter, at least temporarily, by generating anomalously deep troughs of low pressure west of Alaska. This will enhance baroclinicity and help push the ridge east. Additionally, ridging will likely become less dominant over Alaska once sea ice forms. Ice emits more longwave radiation than liquid.
They won’t be as frequent during the first half of winter, but longwave troughs will reach the East and will have plenty of fuel for extratropical cyclones, which intensify when there is contrast between temperatures. SSTs along the entire western Atlantic basin are running higher than normal. That means the contrast between the warm ocean and the cold continent will be greater, allowing storms to develop stronger near the coast. It also means that more water will be available to those storms, so coastal storms may be wetter than normal this winter along the East Coast, especially during the first half of winter. Unfortunately, those hoping for a white Christmas along the coast ought not to hold their breath. The warm SSTs will likely prompt a storm track closer to the coast where the contrast between air masses will be greatest, at least through early January.
On the upside, if any storms do manage to tap into cold air, they could become heavy snow producers for coastal cities like New York and Boston. Storms of this scale will be more likely in late January and February, but it won’t take many of these kinds of storms to make up for lost snow from earlier in the season. A slightly above normal winter total snowfall is therefore favored for the I-95 corridor.
Conversely, early season snow is expected to be enhanced in the lower Great Lakes due to slightly warmer than normal waters in Lake Michigan, Erie, and Ontario. The warmer lakes will enhance the moisture content of lake effect snow. Since normal temperatures are cold to begin with in the Great Lakes region, heavy lake effect snow is likely even during the slightly warmer than normal start to winter.
These extrapolations from SST anomalies correspond most closely to the forecasts from NOAA GFDL’s Climate Model. Its precipitation forecasts for December through February are referenced from the CPC NMME website .
These forecasts suggest that much of the Central US and the Pacific Northwest can expect a fairly dry second half of winter, with normal precipitation expected in the Rocky Mountains. These anomalies fill in the gaps where El Niño has little correlation with winter weather patterns.