Living with, and forecasting for Snow Eaters


(written about Anchorage, AK, with invaluable assistance from Charles Bell at the National Weather Service in Las Vegas, NV)

Though it is more accurately termed a “Foehn Wind”, perhaps the most commonly used term for a warming downslope wind is a “Chinook Wind” or “Snow Eater”. Sourdough Alaskans know the term “Chinook” quite well. Considering there are 14 mountain ranges, and six distinct climates in a state that is one fifth the size of the continental U.S., it makes sense. The problem is that most Alaskans don’t really know what they are.

Around Anchorage, the most noticeable Chinooks are ones that come from the south and run into either the Chugach, or Kenai mountain ranges. They bring more noticeable warmth than the northerlies coming over the Alaska Range that occur primarily in the winter. Typically, the southerly winds are driven by a “Pineapple Express” where the jet stream essentially becomes oriented straight up from Hawaii into Southcentral Alaska. Thusly, most Alaskans have the notion that all Chinooks are southerlies. It was my realization of this fact that inspired my topic for this paper.

The realization came from days of talking about a Pineapple Express on the air and giving it the lion share of the credit for the warming trend that the town was currently experiencing. Several viewers called or wrote in, some of them mincing no words in telling me that my “Washington weather terms (me being a former Seattle-ite) don’t belong in Alaska weather” and that I was incorrect about my assessment of the situation. My challenge was to find a way to explain on the air that they were half right, but they were half wrong as well. Thankfully for me, my on-air presentation served to only increase my “weather cred” with the sourdoughs.

A discussion of Chinooks in Alaska requires that I point out again that Alaska has essentially 14 mountain ranges and at least six distinct climates in a state that is one fifth the size of the continental United States. The point is redundant, but important to remember when evaluating the challenge at hand. There are two dominant weather drivers: 1) Low pressure systems created off the Aleutian Islands (created in winter by the warmer ocean contrasted by colder land temperatures, known commonly as the “Aleutian Low”) that track east and eventually pull across Prince William Sound into Southcentral Alaska, bringing with them mP air; and 2) A semi-permanent high pressure region that sets up over the central Interior that carries cP air that can be dry and scorching in summer (with high forest fire danger), and bitterly cold in winter (with brutal cold snaps that can last for weeks).

In order for a Chinook wind to occur there are several factors that must be present. First, the wind must be flowing in a predominantly perpendicular flow to the mountain range at 20kts or greater. This gives sufficient “juice” to the wind so that it has the strength to rise above the barrier, dumping whatever moisture it has in the process. Once it completes that process, it then descends down the lee side much drier, and much warmer (thanks largely to compression heating) and often times much stronger than the wind speeds at the peak of the barrier. Secondly, there must be an inversion at, or just above the mountain range. Though the maximum height above the barrier that the inversion can be is still undetermined, a properly placed inversion allows for a perpendicular component that is somewhat constant or decreasing with height, allowing a kind of “empty space” that allows the winds to be drawn downward for lack of a lifting mechanism. Finally, wind speeds 2km above the barrier must be less than 1.6x the wind at the peaks of the barrier. If the winds are higher, the shear will act as a lifting mechanism and will typically cause the winds to become trapped lee waves instead, posing more of an aviation hazard than a warming event. A wind flow on the lee side of the range that is flowing parallel to the barrier can also serve as a helper in the process, but it is not required for the event to happen.

The perpendicular flow in Anchorage can be northerly due to clockwise high pressure in the interior - thus intersecting the east to west oriented Alaska range (best known as the location of Mt. McKinley) or Talkeetna mountains. It can also be southerly or southeasterly thanks to low pressure off the Gulf of Alaska that’s center of rotation is usually located just south of Prince William Sound, thus intersecting either the Kenai or Chugach mountain ranges.

Southerly Chinooks (typically fueled by a Pineapple Express) can make actual and forecasted temperatures differ by 15 degrees or higher, quite easily. One strange effect that is not completely understood is that Chinooks from the south can essentially “splash” into town. Much like if you were to turn a hose on and point it directly at the ground. That is, one part of town warms up to puddles and snow melt, while another location within a few blocks downstream of the winds continues to stay below freezing. This obviously presents forecasting challenges when temperatures within a city of just 270,000 can vary by twenty degrees or more, with essentially no resources to pin point specific hot spots. Another interesting, but thoroughly explainable effect of such events in Anchorage is that due to terrain barriers (mountains and inlets), there regularly can be light winds in Anchorage and hurricane level gusts literally just a few miles out of town.

In speaking with NWS Meteorologist Charles Bell, a former Anchorage resident who now works at the NWS Las Vegas office (whose recent research includes the “T-Rex” or Terrain-induced Rotor Experiment conducted in 2006 in Owens Valley, CA and others locations), I learned that high resolution models have been, and are being developed that are having a dramatic effect on downsloping wind forecasting. The COAMPS model provides the highest sensitivity, but current computing space, and speeds in particular, are limiting it dramatically. The 4km WRF is the most operationally useful model, but also requires solid pattern recognition on behalf of the user and must initialize well in order to be useful.

Use of the models has brought forth several key lessons: 1) Mountain wave and downslope events are much more common and localized than previously known. 2) There seems to be a favored time for such events (typically 0000z and 1200z for Charles’ research region); and 3) Strong events occur frequently ahead of, or during passage of a trough when the stability at the ridge increases.

Thankfully, such research has yielded several positive conclusions as well. First, research and computer processing advances are making the accuracy of forecasts improve greatly. Secondly, lead times have improved. Finally, forecasts of wind speed values and expected windy locations are also improving.

The biggest challenge at this time for forecasting Chinook events is not so much science, as it is technology. Where the COAMPS model is far more complete, the calculations of it can take hours and sometimes render it essentially useless for anything other than hindsight. However, with the recent developments of dual processors and higher computation speeds, the dawn of accurate Chinook event forecasting seems to be rising now.

Meso-forecasting of Chinook events may pose a challenge, but it is certainly a challenge that is being not only accepted by researchers, but is being handled by computer forecast models with greater accuracy than any of us would have expected even just a few years ago as well. The future of Chinook forecasting is not on a downslope, but it is heating up.