El Nino/La Nina Defined and Ski Areas Favored by El Nino (as of 2016)
The El Nino/Southern Oscillation (ENSO) has been studied extensively for its impacts upon climate. The shorthand definition of El Nino is an abnormal warming of the normally cold ocean waters of the Humboldt Current off the coast of Peru. A more formal measure used by many meteorologists is The Multivariate ENSO Index , which tracks both sea and air temperatures along with wind, pressure and cloud cover across the tropical Pacific. The data is combined and normalized to a monthly value where +1 = one standard deviation above normal (El Nino condition) and -1 = one standard deviation below normal (La Nina condition). I'll refer from now on to the MEI index, which is graphed below since 1950:
As most of us have observed, skilled weather forecasting has become very accurate in the short term, very useful for storm-chasing 3-5 days out, but more speculative beyond a week. Beyond 2 weeks I find it more useful to rely on historical records. El Nino/La Nina is the exception to this rule of thumb because the condition tends to persist for several months at a time. I have examined the monthly MEI table, which is used in the graph above and is updated with about a 1.5 month lag. With over 700 monthly values since 1950, each month's MEI index can be correlated with the index 1, 2, 3, etc. months in the future to demonstrate how likely an El Nino or La Nina condition might persist. Results are shown here:
Notice that there is a very mild tendency for El Nino/La Nina to reverse itself 15-30 months in the future (think of a bathtub sloshing one side to the other), and only after 3 years do all effects of prior conditions disappear. When I first wrote this article on 11/21/07 the last posted value of MEI was -1.117 for SEP/OCT 2007, and I expressed strong confidence that the observed La Nina condition would be in effect through the end of 2007, and some confidence that it would persist for the entire 2007-08 ski season. It turned out that La Nina remained strong (2007-08 was 5th highest La Nina in the past 50 years) through FEB/MAR, then dissipated over the next 3 months to a neutral value of +0.050 as of MAY/JUN 2008. Similarly the El Nino of 2009-10 was strong through FEB/MAR (the 7th highest El Nino in the past 50 years), but MEI turned negative by MAY/JUN. By AUG/SEP 2010 MEI registered -1.99, the strongest La Nina reading since 1955. The La Nina of 2010-11 remained strong through MAR/APR (the 2nd highest La Nina in the past 50 years) before weakening abruptly in APR/MAY 2011. La Nina strengthened to a moderate level for fall 2011 and gradually weakened starting JAN/FEB 2012. The MEI Index did not have a sustained significant El Nino or La Nina stretch for the next 3 ski seasons. The 2015-16 El Nino exceeded +2.0 for all but one month from MAY/JUN 2015 through MAR/APR 2016 and was the 3rd highest in the past 50 years.
I had noticed anecdotally that the larger El Nino and La Nina episodes tend to break up most often during the Northern spring. So tested by each calendar month the correlations 1 month and 6 months into the future. The one month correlations are all over 90% but they are highest between July and January. The 6 month correlations are more revealing. A JUL/AUG or AUG/SEP reading, available in early September or October, has an 82% correlation with what we'll have during the middle of the upcoming Northern Hemisphere ski season. For South America this has much less predictive value, as El Nino/La Nina during their ski season has less than 20% correlation with its status 6 months prior.
While we have established that El Nino/La Nina are persistent weather events, their effects upon ski area snowfall are less clear-cut. I have correlated the monthly MEI table with all of the monthly snowfall data I have collected through 2016, and the list of ski areas with statistically significant snowfall sensitivity to El Nino/La Nina is much shorter than most people think.
The monthly correlations are not large enough to have much predictive value. But by combining 6 consecutive months together to form seasonal data, the correlations for some areas get into the 50% range. This fits with observed experience that in big El Nino or La Nina years the expected effects occur from time to time but not consistently. So I considered the seasonal correlations to be the main criteria in classifying areas. Not all areas provide complete November to April data, and I like to have 20+ seasons to draw conclusions.
For areas without enough complete seasons I looked at the monthly correlations, but also at the seasonal ones for nearby areas with many complete seasons. For example Sugar Bowl is likely to be affected similarly to nearby Donner Summit and Alpine Meadows. It is important to realize that season correlations based upon 22-60 data points have much more uncertainty than the El Nino/La Nina persistence correlations based upon over 700 data points.
Correlations are not necessarily the best way to analyze El Nino/La Nina. Many meteorlogists believe that only the stronger episodes have a material impact. Thus we should only examine snowfall during the months with the highest and lowest MEI readings. I chose months above +.750 for El Nino and below -.750 for La Nina. The problem here is that when I first collected this data in 2007 there had been only 3 La Nina seasons in the past 30 years, and only a few areas had data from the La Ninas of the early and mid-1970's. With the 2 strong La Ninas of 2007-08 and 2010-11 there is now more data, so I've revised my tables to show snowfall percents of normal for strong El Nino months and strong La Nina months. These columns are blank for areas with less than 22 months of data, which at a minimum would be all La Nina months December-March since 1988.
The list of ski areas favored by El Nino, along with their monthly and season correlations to the MEI index and average snowfall during strong El Nino and La Nina months, is shown below:
El Nino strongly favors only Southern California and Arizona, with milder effects extending to the southern Sierra, far southern Utah and New Mexico. In El Nino years the only big destination resort that is favored is Taos, and that in the mild category. Taos takes until nearly February to get fully covered in normal years, and skiers should be more wary during La Nina years. The data I acquired for Las Lenas in 2005 and Portillo in 2007 support the prevailing view that the high Andes are strongly favored by El Nino. As the 2010-11 La Nina strenghtened these areas received almost no snow after August 1, 2010. Advance bookings to these lower latitude South American ski areas (also the Valle Nevado group) should be avoided in La Nina years until snow is on the ground.
I have constructed graphs to illustrate the variability of the snowfall correlations to El Nino/La Nina. The one below is for selected areas favorable to El Nino. Since the 2 strongest La Nina years were 1973-74 and 1970-71, I select areas with data that goes back that far.
The horizontal axis lists the ski seasons since 1966-67 in order of strong El Nino at left to strong La Nina at right. The vertical axis is percent deviation from normal snowfall. The blue line is the sum of MEI indicies from OCT/NOV to APR/MAY, scaled to fit the graph.
The purple line shows the dramatic boost to Southern California snowfall from El Nino, with the 2 biggest snow years correponding to the 2 big El Nino of 1982-83 and 1997-98. 5 of the top 8 El Ninos produced at least 170% of normal snow. There are no guarantees even here, as the #3 and #5 seasons 2015-16 and 1986-87 were real stinkers at only 59% and 62%. Of the top 9 La Nina seasons only 2010-11 was above average in Southern California, though only one was lower than 77%.
Moving to Taos (yellow line) the effect is less dramatic. The 2 big El Ninos were 116% and 118% of average, while Taos' record 1972-73 season at 174% was in the 6th highest El Nino year. But only 2015-16 and 1991-92 of the top 7 EL Nino years was below average at Taos, and those were still 98% and 95% of average. For the top 9 La Nina years, Taos is missing data for 1970-71, 1974-75 was a good year at 138% and 2007-08 above average. 1999-2000 and 2010-11 were bad at 58% and 65%, and the other 4 were below average in the 85% range.
In the Sierra the picture is mixed. Everyone remembers the huge Sierra snow during the record El Nino of 1982-83. But the 4rd and 5th strongest El Ninos (1991-92 and 1986-87) were severe drought years at Tahoe, and 1986-87 was Mammoth's 4th worst season ever at 5%. Nonetheless 5 of the top 9 El Nino years (1982-83, 1992-93, 1994-95, 1997-98 and 2009-10) were at least 130% at Mammoth (orange line) and at Donner Summit (light blue line) and Lake Tahoe.
At Donner Summit the top 3 La Ninas and 6 of the top 9 are above average. Lake Tahoe La Nina seasons are probably assisted by colder temperatures minimizing low elevation rain. This relatively good La Nina track record is the reason that the MEI correlations are low and statistically insignificant for areas like Kirkwood, Squaw Valley and Alpine Meadows. Historically Mammoth was classified with Taos in mildly favored by El Nino category. Before 2010 the top 6 La Nina seasons were all below average, 3 of them by 30% or more. However Mammoth set a 40+ year snowfall record during the strong La Nina of 2010-11, lowering its seasonal correlation with MEI from 28.6% to 17.9%. Mammoth's near average snowfall during the strong El Nino of 2015-16 further lowered seasonal correlation to 16.3%. Bear Valley has a similar weak correlation to El Nino as Mammoth. I've chosen to leave both areas in the mildly favored by El Nino category because snowfall during strong El Nino months exceeds snowfall during strong La Nina months by 15-16%.
All content herein copyright © 1996-2016
All Rights Reserved.
No copies or reproductions may be made in whole or in part without express permission by Tony Crocker.
Prices for commercial users will be determined based upon intended use and distribution.