This year’s winter forecast is going to look a lot like last year’s. That’s because last winter was a La Nina winter and this winter almost certainly will be one, (or already is depending on your point of view). And how did my forecast for last winter turn out? Here’s that story. A twist to this winter is the speculation by some that there might be some significance to a 2nd consecutive, or double-dip, La Nina.
But first, what is a La Nina, how does it affect our weather, and can it really allow one to make a five month or longer forecast? Very briefly, a La Nina is one phase of a oscillating weather pattern in the equatorial Pacific involving air pressure patterns, winds and sea water temperatures. That a weather pattern roughly 4,000 miles (6,000 km) away can affect Alaska’s weather shows the large, interconnected nature of Pacific weather and climate systems and how important ocean temperatures are to them. The tropical part of this system has been studied for decades and is termed the El Nino/Southern Oscillation, or ENSO for short. La Nina is the cool phase of this tropical system, El Nino the warm phase. It oscillates between the two famous kids on a more or less yearly basis, typically with a lull during the northern hemisphere summer and an intensification in fall and through the winter. Many years the phenomenon is weak or noncommittal…a neutral phase. Here’s a intuitive graph from the NOAA’s Earth Systems Research Lab:
For more information on the the ENSO see the links at the end of the post.
As for what it means for Alaska’s winters, the highly simplified version is that El Nino winters are more often warmer, La Nina more often colder. In the following charts I’ve plotted the average temperature from November through March vs the seasonal snowfall total for four hopefully representative locations. This kind of X-Y scatter plot can tell you how winter temperature and snowfall correlate. Each diamond represents one winter of the 62 plotted since the winter of 1949-50. From the scales you can see that winters that fall much to the upper left were cold and snowy, lower left cold & not so snowy, upper right warm & snowy, lower right warm & not so snowy. Near the center winters were close to average. In addition, and pertinent to this topic, I’ve colored the diamonds according to the ENSO phase for that winter: red for El Nina (14 of the 62), blue for La Nina (16) and gray for neutral (32). The lime green circles indicate the “double dip” La Nina winters. Keep in mind that there are quite a few ways to quantify and define El Nino/La Nina/neutral years and which criteria are used could affect this sort of analysis to some degree. My source seems to leave more winters in the neutral bin than some, possibly understating some relationships.
Starting from the north, here’s Nome and Fairbanks (please click for a better look):
Both Nome and Fairbanks show little correlation between winter snowfall and overall winter temperature. At this latitude I would have expected warmer winters to have more snow and colder ones less. Possibly if we considered the water equivalent of the snow it would show up. A little farther north I’m sure the correlation would be obvious. But that’s not the thrust of this article. It’s how the ENSO plays out for our winters. So look at the colors. Both cities show strong evidence to support the warm El Nino, cold La Nina pattern. In both cities only 2 of 16 La Nina winters were warmer than average by any significant amount (one of Nome’s was last winter) and those only moderately warmer. Some of the La Nina winters were only slightly cold but the majority were roughly 3-6 degrees F (1.7-3.3 C) colder than the 62 year average. That amount of deviation makes a big difference over a winter. Snowfall shows no conclusive pattern in these two cases. Neutral years are scattered fairly uniformly throughout the graph for both temperature and snowfall. Interestingly, most of the very snowy years in Fairbanks were neutral years. Perhaps a Goldilocks situation. By the way, the double-dip La Nina years don’t show anything conclusive. There are only four, and they are spread fairly evenly through the La Nina years. Perhaps if the analysis went went further back more could be gleaned. Also, as mentioned above, the criterion used for categorizing La Ninas is not standardized. The Climate Prediction Center calls the winter of 1950-51 a La Nina (a pretty strong one too, so I don’t know why the FSU ranking I used does not). It was, for three of the four cities (all but Nome), one of the four coldest winters of the 62, and it was a double dipper to boot.
For more southerly climes, let’s look at Anchorage and Juneau:
A major difference from the two northern stations is the inverse relationship between temperature and snowfall that is evident at Anchorage, and strong at Juneau. In other words, warmer winters have less snow, colder ones more. This relationship is expected in temperate areas. With respect to the ENSO phase, La Nina winters tended to be colder at both locations, and to a lesser degree snowier. A strong exception is the double- dip La Nina winter of 1999-2000 when in Juneau it was 3.9 degrees F (2.2 C) warmer than average with 44% of average snowfall. At the other three station that particular winter also did not follow the pattern but was not so extreme: Temperatures near average and snowfall near, to above average. Again the double-dip winters do not tell us anything definitive. One thing skiers can be pretty happy about is that there were no La Nina winters in the last 62 years that were both above average in temperature and below average in snow in Anchorage or Nome and only one in Fairbanks and two in Juneau. Study these graphs more and I’ll bet you can tease out more revealing details about our winters.
Useful for seasonal forecast?
What can be deduced from this data to make a projection for the coming winter? With a moderate to strong La Nina already under way, I think we can expect most of Alaska to have a colder than average winter. Snowfall in Southeast is likely to be higher than average but for the rest of Alaska it is indeterminate. If temperatures are not significantly colder, then near average is the likely result—a significantly warmer than usual winter has little statistical chance of happening in a La Nina year. The idea of a double dip La Nina bears further study before I’d want to say much. Let’s just say it will likely be a cold winter for most of Alaska, and I would not be surprised if if it is not much colder than the long term average in many parts.
Lest you think I’m something special for making a 5-month forecast, I’m not alone, and I’m not going against the crowd. To see more of the consensus, follow these links to the Climate Prediction Center or Juneau Forecast office’s summary. The CPC’s maps are right there on the home page and are easy to use. For more maps click on 30-day or 90-day outlooks below the home page maps. All the winter 2011-2012 maps you can find will say the same thing: confidently colder than average for the southern 2/3rds of Alaska. These are updated around the 20th of each month but I would not expect any change in the basic idea.
It is important to note that the above forecast could prove to be flat out wrong, and likely it will be for some parts of Alaska. Alaska is so big that the chances of every part of it being colder than average are pretty low. In particular, the far north and far west (not really covered in our graphs), i.e. Barrow, and the western Aleutians, seem to often be exceptions. The strong warming in Barrow recorded since about the mid 70s seems to defy trends and drivers that effect other areas. More study needs to be done, but possible reasons are decreases in extend and thickness of the ice pack, and the urban heat island effect (artificially warming the local populated area only). The latter effect has been well documented in Barrow by Dr. Ken Hinke of Univ. of Cincinnati using in situ measurements.
Even if the winter is colder than average, day-to-day and even week-to-week winter temperatures are volatile and there will be fluctuations above and below the mean. Different factors affect the weather on different time scales. For snowfall there are even more variables than for temperature. That is why we deal with averages and percentages, not absolutes. But the ENSO phase is probably the most powerful seasonal forecast tool yet discovered for a vast part of the globe. Another cycle I’ll briefly mention is the Pacific Decadal Oscillation which, due to its current cool phase (cool for Alaska), bolsters the confidence of this colder-than-average winter forecast. I covered the PDO more in last year’s winter forecast.
Please let me know what you think via the comments link below. And remember, don’t kill the messenger just because you don’t like the message.
NOAA NCEP ENSO page http://www.cpc.ncep.noaa.gov/products/precip/CWlink/MJO/enso.shtml
NOAA ESRL ENSO page http://www.esrl.noaa.gov/psd/enso/
NOAA PMEL ENSO data links http://www.pmel.noaa.gov/tao/elnino/measurements.html