When tourists see my Alaska Weather Calendar a certain percentage chuckle and ask “is is accurate?” What they’re assuming is that I’m forecasting the weather for the entire next year (almost a two year forecast since I print the calendars the spring previous to the calendar year). They liken it to the Old Farmer’s Almanac.
It is hard enough to predict the weather three days in advance and usually fruitless to make a specific forecast for seven days out. So am I going way out on a limb in taking a stab at the coming winter? Will I be like the cartoon characters who cut off the limb while standing on it?
Well, first it is important to note that a seasonal forecast differs from the 3-7 day variety that is commonly exchanged over morning coffee and ridiculed at lunch (sometimes for good reason, sometimes due to a misinterpretation). The difference is the level of time-specific detail required in the short term forecast (which is usually more than the science can support after about day three or four.) So we’re dealing in general statements about a long period of time, usually with a low confidence factor assigned. These are routinely made by the Climate Prediction Center and here are the temperature and precipitation outlooks for December through February (as usual, click on image for larger version):
Most of Alaska is predicted to be below normal in temperate, but the precipitation forecast calls for “equal chances” of above or below normal, which to say there is not enough certainty to say one way or the other. Note that there is no snow forecast…since snow depends on precipitation and temperature, the variables are too tricky. I’ll take a shot later, after a little on what might help one in this situation. Also note that the forecast only goes through February, which even southerners know does not nearly cover winter in Alaska.
I’m not sure about all that goes into these particular forecasts, but one thing is sure: long range forecasting has benefited in recent years from understanding links between the weather and semi-regular cycles in the ocean-atmosphere system. Benefited does not mean perfected. There is still plenty of uncertainly. But let’s review what is fairly common knowledge:
El Nino-La Nina, also called the Southern Oscillation, or ENSO
The southern oscillation is measured as temperature and wind patterns in the tropical Pacific, but when these elements flop back and forth (or somewhere in the middle) there are connections to the weather over a far wider area, perhaps half the globe. Here’s an easy-to-read graph showing the cycles from 1950 to 2010. When the graph is far up in the red the ENSO is in the warm phase or El Nino. Significant dips into the blue are La Nina Periods.
During El Nino winters, Most of Alaska is warmer and often wetter. Strong jets of warm air from way south in the Pacific visit more often, bringing episodes of above-freezing temperatures and rain, even in the midst of the winter. [As I write this Anchorage, Fairbanks and most points between and even further north are just getting through a terrible freezing rain event that would seem more consistent with an El Nino instead of the strong La Nina we seem to be headlong into. This reinforces that there are no absolutes, only generalizations.] La Nina generally means colder winters, often drier. For northern areas such as Fairbanks and north, a cold winter pretty much equates with less precipitation including less snow. For Southeast Alaska, colder usually also correlates with less precipitation, but because SE gets a more even mix of rain and snow in the winter, and a colder winter shifts more of the precipitation to snow. Its not hard to get more snow out of less precipitation; remember an inch of precipitation means an inch if it’s rain but 10-30 inches if it’s snow. As I mentioned in passing, the ENSO is shifting decidedly negative, so a moderate to strong La Nina is expected through the winter. You can read more about it at the Climate Prediction Center.
Pacific Decadal Oscillation or PDO
Since the ENSO cycle was discovered many years ago, many more ocean-climate cycles have come to light, including an important one for Alaska, the PDO. The PDO also is keyed to Pacific Ocean temperatures, but over a much larger tropical and extra-tropical area. The time scale is also different. It tends to switch phases roughly every 20-30 years, rather than 1-3 for the ENSO. However, the two are related, not surprisingly. After all they both have to do with Pacific sea surface temperatures.
The Great Pacific Climate Shift
Take another look at the above graph of the ENSO. Near the middle, in about 1977, there is a undeniable shift from more cool cycles to more warm cycles. This marks the change of the PDO from cold to warm, and was so profound that it has been called the The Great Pacific Climate Shift. This table from the Climate Diagnostics Center clearly show the most recent PDO shifts and relates them to the ENSO:
And this graph from the Alaska Climate Research Center at UAF shows the dramatic effect on Alaska’s temperatures:
This graph is nothing short of stunning from a climate statistic standpoint. After 1977 there is only the one big stalactite into below-average temperatures (in 1999), and that correlates to a slip of both the ENSO and PDO into the cool phase. Hold that thought about the cool 2008.
So What?
It’s not too surprising that a warmer Pacific means a warmer Alaska and vice-versa. But what forecasting value does it have? Much. First, the sea changes much more slowly than the weather patterns, so the current state of the ocean can help project weather regimes and averages much further foreword than traditional weather forecast techniques and technology. Please remember the difference between “weather regimes and averages” and a “traditional weather forecast” as discussed in the third paragraph of this post. So let’s look ahead.
The Forecast, Please
For this winter, confidently going to be an La Nina winter and likely a strong one, I’m agreeing with pretty much everybody in expecting colder than average for most of AK, more snow in Southeast Alaska, possibly more snow in Southcentral.
Additionally I’m going to stick my neck out a little and forecast a similar pattern most years for the next 20ish years. It’s no special revelation. The latest pacific regime change, back to a cooler Alaska, has most likely already happened, although it may not be as large a change as the 1977 jump, or as sudden. The CDC table above puts it happening in 1999. But this is not totally convincing, especially looking at Alaska, so maybe it was closer to 2005. Time will clarify. So my forecast is merely for a continuation of the current regime for a period of time consistent with past regimes. Not too bold, really. But even if it holds, there is bound to be year-to-year variability. Here’s the power in this: If there was confidence that for the next 20 years in SE AK, colder, snowier winters were to be twice as common as warmer, rainier ones, (a change from the past 25-30 years when warm ones bested cold ones 3:1) then planning decisions by individuals, governments and other organizations become easier. Should people winterize their homes? Yes. Should municipalities upgrade their snow removal departments? If it has been discussed, then it might be time. Should the State Park people in Haines invest in ski trail grooming equipment as they are currently planning? I’d say go for it. [bias disclaimer: I’m an avid cross country skier]
I’d love to hear any comments, questions or complaints from skiers and non-skiers alike. Use the comment link below.
Nice description of the PDO. That Alaska temperature departure plot is pretty amazing.
So, we’re still reporting our climate normals from the 1971-2000 average. With such a strong PDO signal in Alaska, and our period for climate averages stuck in the warm phase, isn’t it about time we switched to a different system for talking about the “average” climate? Perhaps this can explain why the official “average” annual snowfall in Haines is 133 inches (a la: http://www.wrcc.dri.edu/cgi-bin/cliMONtsnf.pl?akhain). Yet the last several years here have been much snowier than “average”. It also doesn’t help that they stopped recording snow depth at the airport.
Perhaps if we were to use the records from say 1949-1977 to calculate our average snowfall in Haines, we would find that the last several winters haven’t been much above average at all!
Score one for the Haines winter recreationists! I too have my fingers crossed for a deep winter. And even better, a local backcountry hut system!
Here’s some historical notes from an old retired meteorologist that
might be of interest to you:
In late 1997, Ants Leetmaa visited the Juneau forecast office shortly
after he became Director of the Climate Prediction Center. During a
long private conversation, he made a very compelling argument that the
inflection point for the next PDO cycle would occur in 1998. This
conclusion was based on unpublished data and summarized on hand drawn
overhead slides (still have those slides stashed somewhere). Some 12
years later, there is little doubt that he was correct.
At that time, we both agreed that the 1998 PDO inflection point could
also mark the start of the next ice age. With time, I am
now more convinced that premise will turn out to be correct, too. In
fact, a more proper description would be that I am a ‘true believer’
that the next ice age is underway. And, by ice age I mean the long,
approximately 12,500 year cycle, not just a century long ‘mini’ ice age.
The transition to an Ice age, based on ice cores in Antarctica and
Greenland, usually have occurred over a 200-400 year span. But, there
are cases where the transition took place rapidly, in as little as
40-80 years (one lifetime). Perhaps, one indicator will be a series
of weaker PDO warm cycles, with the cold phase becoming more dominate.
Global warming is /_the_/ trigger for the onset of an ice age. One
prerequisite is melting of the permanent ice cover over the Arctic
Ocean. This is necessary to expose more open water for a longer
period to provide the moisture source for increased snow cover over
the land areas. From a precipitation perspective, the high latitude
Arctic is a desert. Global warming also provides increased ocean
evaporation for increasing global precipitation. An increasing
percentage of the precipitation will be in the form of snow over the
high elevations and high latitude land areas.
The U of Alaska Geophysical Institute has studied the implications of
the relationship of Arctic open water to snowfall, with at least one
PhD dissertation on the topic published in the early 1970s. For snow
cover changes. satellite data tracking the areal extent of seasonal
snow cover over North America will be the main clue. Seasonal snow
cover trends need to be observed over a period of more than one PDO cycle.
Conventional climate data is less useful. For example, the NWS uses a
rolling period (30 years) for computing average temperatures
that changes every decade. Such baseline data will always be
dominantly composed of one PDO cycle, or the other. If the current
period is within a cold cycle, and the last 30 years is mostly made up
of a warm cycle, temperature comparisons are flawed from the start. A
better way would be to only compare current temperatures to those in
matching past PDO cycles to credibly validate climate change.
Bob Kanan
Juneau