There are two observations that stand out in the results. Statistical significance was stronger for monthly snowfall compared to seasonal snowfall. The AO and PDO showed the most
significance overall versus the other patterns, while the PNA and and AMO were the least. Both Boston and Minneapolis showed no significance whatsoever in
any of the patterns. They were also the furthest north in latitude. On a synoptic scale, the phase of each of the patterns can have a big impact. This can affect both storm track
and precipitation type, which will affect snowfall. Independently, the patterns show no effect on snowfall compared to locations futher south.
The next two paragraphs will focus on the results for city of Philadelphia, as it cleary stood out over the other cities because of high statistical significance for several patterns.
With five winter seasons to compare, the results verify nicely with the conditions that have occured to date. Despite a borderline high La Nina during the 2010/2011 winter season,
there was much above average monthly and seasonal snowfall, totaling over 44 inches. In my opnion, the dominance of the negative phase of the AO during that winter was the only
reason for the result. To backup that claim, the winter of 2009/2010 also saw a domiant negative AO of record levels, resulting in high monthly snowfalls and two historic
snowstorms. One could argue that there was moderate El-Nino that winter contributing to the result. While it likely did contribute, the AO contributed mostly to the massive
snowstorms leading to the high snowfall that winter. The results in this study are evidence for support and the montly data for the AO is shown in Table 1.
Table 1 - Arctic Oscillation for Philadelphia, PA(Monthly)
Thresholds
Low: -4.266 to -1.046
Medium: -1.036 to 0.186
High: 0.194 to 3.495
| |
|
|
Index Value |
| |
|
|
LOW |
MEDIUM |
HIGH |
TOTAL |
| Total Snowfall |
7.3 - 51.5 |
26(17.7) |
18(17.7) |
9(17.7) |
53 |
| (inches) |
2.4 - 7.0 |
17(17.7) |
16(17.7) |
20(17.7) |
53 |
|
0 - 2.2 |
17(24.7) |
26(24.7) |
31(24.7) |
74 |
| |
|
Total |
60 |
60 |
60 |
180 |
| |
|
|
|
|
|
|
| X2=12.8 |
|
P-Value= 0.012 |
Very Statistically Significant |
However, the winter of 2014/2015 saw a dominant positive AO. Snowfall was high in February, but not in December or January. There was also a record postive PDO. Since both
patterns were both very statistically significant, there was a battle for superiority. A positive AO lacks snowfall, while a postive PDO increases snowfall. The winter saw a
lack of snowfall in December and January, while February saw more snow. In fact, it was only the 10th occurance of double digit monthly snowfall with high positive AO values
since 1950 (See Table 1). One key event that backed up the results of the AO was during the January 26-27, 2015 historic blizzard that hits parts of Southern New England from
Long Island to Boston. The storm was a major bust for Philadelphia. The results showed from a statiscial standpoint that Boston(p-value=0.744)and New York
City(p-value=0.126) could see major to historic snowstorms with this type of pattern, but not Philadelphia (p-value 0.012). The monthly data clearly showed that big snowfalls
don't occur often when the AO is in the positive phase. The values of the AO during this event were greater than '1', easily putting them in the high(postive) category.
The differences between a postive and negative AO can be shown synoptically. Two snow events are shown. Figure 1 is from the December 18-19, 2009 event where there was a
strong negative AO, with 23.2 inches of snow that fell on the city. The map is from the 12Z NAM on Dec. 17, 2009. Figure 2 is from the January 26-27, 2015 event where there was a
strong postive AO and only 1.2 inches of snow fell on the city. The map is from the 00Z GFS on January 26. In both images, note the strong ridge on the West Coast of the U.S. and the deep
trough on the East Coast. This is a very good example showing Newton's Third Law of every action has an equal and opposite reaction. **Click on images to enlarge and open in a new
tab**
The strongest significance for ENSO occurred with the city of Chicago. The p-values were very statistically significant for seasonal values. This was for both the Nino 3.4 and
SOI values. A high number of seasonal snowfall totals occured for high Nino 3.4 values and low SOI values. This indicates stronger emphasis should be placed on El Nino events
over Neutral or La Nina conditions. This verified for the winter of 2014/2015 where high monthly snowfalls occurred during January and February with weak El-Nino conditions.
Prior years showed mixed results. The data did show lower seasonal snowfalls during La Nina events. This verified during the 2011/2012 winter, but was a big failure during the
winter of 2010/2011. Indianapolis showed strong significance for seasonal values, but only for the SOI. The same was noted for Philadelphia and Washington, D.C.
The TNH pattern was used for monthly statistics over the Nino 3.4 due more volatile short term changes. Sea-surface temperatures can remain fairly constant for several months over
the equatoria Pacific ocean and thus the need to find any significance would be stronger on a seasonal scale. Only the city of Chicago saw significance with the TNH pattern.
And it was strong. The data favored higher snowfall during a positive phase. There was a very strong +TNH pattern during the winter of 2013/2014 and that verified
extremely well as Chicago saw one of the coldest and snowiest winters on record. Figures 3a and 3b show the 500mb height anomalies, averaged over two months. Figure 3a is averaged
for Deccember 2013 and January 2014, while Figure 3b is the average for January and February 2014. The snowfall during those three months was in the top five for Snowiest winters
on record. Table 2 shows that over half of the total number of high monthly snowfalls occured with high(positive) values of the TNH.
| Figure 3a |
Figure 3b |
 |
 |
| Image Courtesy of ESRL |
Image Courtesy of ESRL |
Table 2 - Tropical-Northern Hemisphere Pattern for Chicago, IL(Monthly)
Thresholds
Low: -3.26 to -0.41
Medium: -0.40 to 0.36
High: 0.39 to 2.14
| |
|
|
Index Value |
| |
|
|
LOW |
MEDIUM |
HIGH |
TOTAL |
| Total Snowfall |
12.9 - 41.3 |
15(17.7) |
15(17.7) |
30(17.7) |
60 |
| (inches) |
2.4 - 7.0 |
23(17.7) |
24(17.7) |
14(17.7) |
60 |
|
0 - 2.2 |
22(24.7) |
22(24.7) |
16(24.7) |
60 |
| |
|
Total |
60 |
60 |
60 |
180 |
| |
|
|
|
|
|
|
| X2=11.4 |
|
P-Value= 0.022 |
Highly Significant |
The EPO showed significance for a few cities, but only for monthly snowfall. The highest significance was seen for Washington, D.C. The data was very strong. In fact, it was
the strongest significance seen in the entire study. This was due to very few moderate or high snowfall totals during a low(negative EPO). Roughly 67% of all monthly snowfalls
during the low phase of the EPO were in the 'low' category (less than 2 inches total). The data is shown in Table 3. This was a very interesting find. None of the other major cities along the East Coast saw
this result and only Phialdephia saw mild significance.
Table 3 - Eastern Pacific Oscillation Pattern for Wasington, DC(Monthly)
Thresholds
Low: -2.60 to -0.55
Medium: -0.54 to 0.28
High: 0.36 to 3.18
| |
|
|
Index Value |
| |
|
|
LOW |
MEDIUM |
HIGH |
TOTAL |
| Total Snowfall |
6.2 - 32.1 |
11(17.7) |
17(17.7) |
21(17.7) |
60 |
| (inches) |
1.9 - 5.9 |
8(17.7) |
19(17.7) |
22(17.7) |
60 |
|
0 - 1.8 |
40(24.7) |
24(24.7) |
17(24.7) |
60 |
| |
|
Total |
45 |
50 |
81 |
180 |
| |
|
|
|
|
|
|
| X2=19.0 |
|
P-Value= 0.001 |
Very Statistically Significant |
Other cites that saw stromg significance on a monthly scale were Indianpolis and St. Louis. Of the three cities, only Indianapolis saw any signficance on a seasonal scale.
