Winter. You either love it or hate it or… you’re just fine with it taking its time. For meteorologists and climate watchers it’s another major phase of the year’s seasons and we’re constantly on the lookout for signs of pattern consistency this time of year. Despite an earlier than normal snowfall this fall and an unusually early cold snap, signs of the impending switch to winter are anything but consistent or imminent.
While every year is different, statistics / averages do help us understand how the consistently wintry weather sets in across Minnesota. One such good average is the date of the first one inch snow cover. This isn’t when does the first snowfall occur, but rather when does snow cover exist for at least a day at 1 inch or more.
As we know, once snow is on the ground, if it’s late enough in the season, that usually locks winter in across Minnesota pretty consistently. There are exceptions of course, but you never quite recover mild weather again until the sun overwhelms us again late in winter or early spring. This is because the sun angle is at its weakest mid November through mid January plus November into December is our cloudiest time of the year. Snow exacerbates this cooling process by reflecting away what little solar insolation we get.
As one might expect, the warmer the November, on average, the later the snow pack develops (again, there are some random exceptions). It is no perfectly fit line but there is a clear trend. On average, November 24th is about when this date occurs.
The overwhelming consensus in the long range computer models is to continue the mild trend we’ve seen this fall. October is likely to end up about 2 to 3 degrees above normal, which is similar to what September was. Guess what? Pretty much all the models are forecasting similar anomalies for November.
The range in these four models is from +1 to +4 degrees (F) above normal, but ALL are forecasting warmer than normal temperatures. If we extrapolate that to the chart above, solely looking at climate statistics, we come up with December 5 to 15 for the first 1” snow cover, coming 10 days to possibly 2 or 3 weeks later than the historical average. Now does this mean that will happen? No, these are statistical correlations. This October is a great example of how we can average 2 or 3 degrees above normal but also have an anomalous cold outbreak mid month that created an unusually early snowfall.
In long range forecasting, looking at statistics and models both are your best bets and especially looking at overall large scale patterns. One such pattern meteorologists look at a lot in the cold season (fall through early spring) is the Arctic Oscillation (or AO index). This a great, quick, view of an overall hemispheric pattern. We can analyze the models from continent to continent through many days or weeks out to get the same picture, but this gives us a quick look at differences between the arctic and the mid latitudes.
Here's the basic idea:
The AO index measures the difference from normal (anomalies) between the arctic and mid latitude and subtropic pressure patterns. This matters because it can tell us how strong or weak the jet stream is that dominates the balance between cold and mild and storm tracks in this region.
When the AO is in a ‘positive phase,’ this means the air pressure is lower than normal across the arctic. Cold air is contained, and the Jetstream blows strongly west to east around the northern hemisphere. This keeps the overall storm track farther south and places at mid latitudes (like Minnesota) are milder than normal or near average.
When the arctic oscillation (AO) turns to a negative phase, air pressure over the arctic is higher than normal and indicates a weaker pattern from normal that allows a very wobbly jet stream and protrusions of cold air deep into mid latitudes, such as over Minnesota, the northeast and Europe. This also pushes major storm tracks south and create some major snow storms for areas east of the Rockies. If you’re a winter lover, you want to see negative AO phases.
The AO is a great overview of these complex blocking patterns and can measure disruptions to the polar vortex which create the same conditions just outlined above (i.e. extreme cold blasts farther south than they normally and stormy patterns on the edges).
For example, we’ve already had some briefly negative AO phases this fall that led to our first cool snap in late September. Remember we went from 92 degrees to the first frost for many parts of Minnesota? The recent mid October cold snap was another negative AO phase that brought an earlier than average snowfall and pretty cold air for mid October for a few days.
Of course now you’re impatiently wondering… well, what’s the AO index look like in the future? Well, in the shorter term (into first 10 days of November) the AO index stays positive or neutral (i.e. near normal or mild conditions). Here’s a look at the European model ensembles (many different runs of tweaks of the European model):
The Canadian and American models look very similar.
In the longer range, things become more chaotic, as you would expect. Forecast possibilities diverge all over the place the farther in time we get. That’s why I started this article with saying that we’re looking for consistency in forecast patterns which is lacking so far. As of the latest computer model runs, both the long range European and long range American models see change brewing later in November:
This might look messy, and it is. The individual bars are all the different ranges of possibilities within the model (literally all over the place). The blue line however is the control, or specific model run. The green line is the mean, literally the average of all the possibilities. This can help filter out the noise to look at a consensus. They differ on mid November, but you can see a trend for later in November of becoming neutral to negative, which could bode well for colder, stormier weather. I have to preface this by saying, every couple days these solutions have looked different so this is hardly ‘locked in.’
Indeed, we see a similar pattern if we look at he long range snowfall possibilities. DISCLAIMER: This is mere meteorological entertainment. No model can forecast specific snowfall accurately past a week.
But here it is, the meteorological version of Netflix. There is something brewing in the pattern. Will we get 1 or 15 inches of snow is anyone’s guess, but the probability of seeing snowier and colder conditions definitely increases later in November.
You might be saying well, ‘no kidding Sherlock.’ Remember, the averages back up the end of November for all this too, but our same models forecasting a shift are also forecasting overall warmer than normal conditions of November but with a distinct, end of month change.
As usual, it’s wait and see for anything past 7 days, but hopefully you understand a little bit more about how the ‘meteorological sausage is made’ when it comes to winter weather forecasting.
Comments