Winter Speculation: The Arctic Oscillation

Many of you powder hounds out there have heard the terms "El Niño" and "La Niña" thrown around quite a bit in the context of seasonal prediction, but these are not the only long-term atmospheric oscillations that are potentially useful for looking into an upcoming winter. Ever heard of the Arctic Oscillation?

In the 2nd installment of the series of posts giving the people what they want (speculation about the upcoming winter), we will take a brief look the Arctic Oscillation Index and its potential impact on monthly snowfall in the Wasatch.

First of all, what is the Arctic Oscillation? The AO takes into account the difference in sea-level pressure between the Arctic and the mid latitudes (i.e. the continental U.S.). When the pressure is relatively low in the Arctic and relatively high in the mid latitudes, this promotes a predominately east-west jet stream and a "bottling-up" of colder air in the high latitudes. This is a positive AO phase. When the opposite occurs (pressure is relatively high in the Arctic and relatively low in the mid latitudes), the jet stream may exhibit more "buckling" and north-south flow, allowing cold air to spill southward more readily. This is a negative AO phase, and the Central and Eastern U.S. often see significant outbreaks of cold air in this scenario.

So where does Utah fit into this? Let's take a look at the correlation of the monthly mean AO Index and monthly snowfall at the Snowbird SNOTEL site. Below is a figure plotting the AO Index versus total monthly SWE at Snowbird for the period 1991-2014. The red line is the best fit line.

The results are…well…not encouraging. Here are the R-squared values (a value of 1 indicates a perfect fit of the best fit line) for each month:

November: 0.03

December: 0.16

January: 0.01

February: 0.12

March: 0.00

April: 0.16

In other words, the AO Index has no correlation to snowfall at Snowbird in November, January, and March, and a very slight correlation the rest of the winter. It is interesting that December is the only month with a trend of greater snowfall for a lower AO Index…this makes me even more wary of drawing any conclusions from the slight “trends” observed. So, the AO is overall not of much use for predicting snowfall in the Wasatch on the monthly scale.

While medium-range weather forecasting (out to 10 days) has become increasingly skilled in the computer age, seasonal prediction is still in its infancy. Also, the Arctic Oscillation is also calculated over the entire Northern Hemisphere, and there are other climatic oscillations such as the Pacific Decadal Oscillation (PDO) that capture more of the conditions that impact western North America. So perhaps some of these other indices have a more useful correlation to Wasatch snowfall…we’ll explore this some other time. Until then, stick to the 7-day forecast and keep doing those snow dances.