1) Six of the ten snowiest snowy winter seasons since 1950 occurred during an El Nino event, when sea surface temperatures in the eastern tropical Pacific were warmer than usual.
2) 16 of the 17 seasons when D.C. recorded over 20 inches of snow happened when the Arctic Oscillation (AO) was negative meaning pressures to our north were above normal. Under these conditions, the location of the storm track and penetration of Arctic air was farther south than normal.
So, the obvious conclusion is that a combination of El Nino and a negative AO offers the best chance of a snowy winter.
But not El Nino events are created equal and not all El Nino events produce big snows. Two of the least snowy winter seasons on record occurred during El Nino. And what about those years when El Nino conditions are borderline or weak, like this year?
Related link: Capital Weather Gang’s winter outlook
Today we’ll take a holistic look at El Nino and borderline El Nino aka “warm neutral”events to try and crack this nut. But here’s the punchline: based on limited data, there’s no precedent for a blockbuster snowy winter given predicted conditions this year. But the record offers promise for at least a little snow.
Let’s begin by examining the table below of all El Nino and borderline events since 1950. The years that met El Nino criteria are listed in red while the four that were borderline (at the bottom), like this year, are listed in black. The table also includes the seasonal snowfall during that winter (with above normal amounts listed in red and below normal in blue), the Arctic Oscillation (negative years in blue, positive in red) and the number of 5.5 inches or greater snowstorms during each season. In that latter column, the number of 10” or greater events is listed in brackets. The two snowfall amounts with an asterisk indicate that a significant snowstorm in November skewed their totals and pushed them above normal.
Here are some key points from this chart:
* For El Nino years (not including borderline cases in black), the average snowfall of 19.9 inches is well above the average for all winters (15.4 inches - this includes La Nina, etc.).
* For El Nino years, the median snowfall was 15.25 inches which is near but slightly below the average for all winters.
* Heavier than normal snowfall was reported in 50% of the El Nino winters. That 50% probability was higher than the 32% probability of above normal seasonal snowfall for all of the 62 winters considered (since 1950).
* During most, but not all , of those above normal seasons, the AO averaged in negative territory.
Strength of the El Nino matters
Weak and borderline events - most similar to this year
If we look at a combination of just the weak El Nino (sea surface temperatures anomaly of 1.0 degree C or less, per second column above) and borderline cases since 1950, we see the snowfall variability is constrained some. Considering the 13 such cases, the snowfall varies in those winters between 4.9 and 25 inches. 69 percent of the time snowfall was in the 8-12.5 inch range and 17 percent of the time range was 18-25 inches. Remember we’re expecting weak to borderline El Nino conditions this winter. So using the past as a guide, it wouldn’t argue for exceptionally high or low snowfall amounts. I see similar probabilities for this season unless the AO becomes strongly negative early in December which would raise the probabilities for an above normal year.
Moderate to strong events - not like this year
The impact of a moderate to strong El Nino is not always the same. In some cases, it supports very snowy conditions. Looking back at the table above, all of the El Nino winters in which over 25 inches of snow were reported occurred during moderate to strong El Ninos (with a temperature anomaly above 1 degree C). The active southern stream jet and the southern storm track in those years apparently worked their magic.
Below is a composite pressure map (at 18,000 feet) for the 10 above normal El Nino seasons. Note the above normal pressures across Greenland and Canada, the below normal pressures across the southern U.S. into the Atlantic Ocean as well those well off the west coast of the U.S. The graphic below is very similar to the one shown last week for the 10 snowiest winters since 1950.
However, two least snowy El Nino seasons were associated with the two strongest El Nino events during the past 50 years. These El Nino events produced so much warm air that we got rain rather than snow. The AO phase didn’t even matter in these years: one occurred when the AO was strongly positive while the other happened with a negative AO. While a negative AO raises the probability of getting snow and having a snowy season, there are other factors that can also influence whether a season is snowy or not.
El Nino and below normal snow cases: How did they happen?
Six of 10 El Nino years with below average snowfall years occurred with a positive AO. The surprise is that 4 of the relative low snow years happened with a negative AO.
The composite (all the seasons averaged) pressure pattern of these two groupings - El Nino years with either a positive or negative AO - are shown below.
The panel on the left is the positive AO grouping. It has a classic positive AO look but also has low pressure along the West Coast and high pressure over the U.S. That forecasts storms to take a track from the Southwest U.S. into the Northern Plains or Great Lakes region. Our temperatures during those 6 years averaged above normal.
The grouping on the right is a composite of the negative AO cases. It has a snowy look, but didn’t produce snow. With a negative AO firmly in place, temperatures averaged below normal but the snow never came. It includes the 1976-1977 season which was cold enough for the Potomac to freeze over and for people to skate in it. That year the polar jet was so far south that storms mostly passed well to our south and east.
Subtle differences in the overall pattern can dictate whether it will be a snowy one or not. If I were to see a forecast anomaly pattern on 8-10 day model forecast similar to the grouping on the right, I would probably start raising the alarm for a period of higher than normal probability of getting measureable snow. However, not all negative AO patterns produce snow so I would probably end up disappointing some readers. That’s part of what makes long range forecasting so difficult.
A closer look at this coming winter
This winter, the sea surface temperature anomaly in the Pacifc is likely to be in the 0.2 to 0.7 degree C range. Looking at those borderline El Nino cases alone in the top chart, above normal snowfall occurred in 3 of the 7 winter seasons. However, that percentage may be misleading as significant early November snowstorms impacted two of those years. Both those winters would have ended up as below normal without the November storms. Only one season experienced above normal snowfall during the December-March period.
What are the chances of a big snow storm during a borderline El Nino year? The only 10-inch or greater event in that 7 winter grouping was the Veteran’s Day storm of 1987 suggesting that the chances of getting a whopper are small.
Of course, 7 years is way to small a sample size to draw any firm conclusions. However, the climatology of these borderline (warm neutral) years does not look that different from the climatology of all years combined suggesting that the probability is higher for below normal snowfall than above. A negative AO this winter would certainly raise the probability of getting a 5.5 inch or greater storm but blocking does not guarantee a snowier than normal winter even during a moderate-strong El Nino.
For snow lovers the statistics for the past 24 winters are even more discouraging as we’ve been in a cycle of having 6 below normal years for every above normal one (a 16.7% probability). Snowy seasons have definitely been in the decline. However, when it has snowed, it has poured. Three of the four above average seasons in that stretch had seasonal snowfall accumulating 40” or more.
Matt Ross contributed to this post.