Does a wet summer in D.C. lead to a snowy winter?

Lately, some of our readers have been asking if there’s a link between a wet summer and snowfall the following winter. It’s always tempting to connect the dots in weather, especially when extreme events happen (think Snowmaggedon winter of 2009-10, the June 2012 derecho, or the all-time warmest summer of 2010). Washingtonians just experienced the fourth-wettest June on record and, with 10.97” recorded for the season already, a wetter-than-normal summer (12” is normal) is all but assured. We have an excellent chance of logging our wettest summer since 2006, when a little over 18.5” fell.

Does all this rain mean anything for next winter? In short, no. A snowy winter does not a wet summer make. But my investigation did uncover another possible indicator for snowy Washington winters. It turns out that you, dear reader, have been looking for snow in all the wrong places.

For starters, there is virtually no correlation between the cumulative June-July-August (JJA) rainfall amount and total snowfall the next winter. The wettest D.C. summers – 30 seasons with at least 15 inches total – have been followed by above-normal snowfall 10 times and below-normal snowfall another 15 times (five of the wettest summers occurred prior to 1888, before which authentic snowfall records are not available).

Each of the summer seasonal rainfall and winter seasonal snowfall totals are plotted here since 1888. There is no apparent correlation (r-squared of -0.02) between the two. Chart by Rick Grow.
Each of the summer seasonal rainfall and winter seasonal snowfall totals are plotted here since 1888. There is no apparent correlation (r-squared of -0.02) between the two. Chart by Rick Grow.

1905 and 1906, the top-two wettest summers of all time (27.05” and 24.22” of rainfall, respectively), were followed by 28.3” and 25.7” of winter snow, respectively. However, five of the six all-time wettest summers starting with the eighth wettest JJA period in 1955 (19.39”) and ending with the 13th wettest JJA period in 2003 (18.28”) resulted in below-normal winter snowfall.

June-July-August (JJA) rainfall total plotted against winter snowfall total for the 30 wettest D.C. summers. Winter seasonal snowfall shows no clear trend, fluctuating up and down throughout the data sample. Chart by Rick Grow.
June-July-August (JJA) rainfall total plotted against winter snowfall total for the 30 wettest D.C. summers. Winter seasonal snowfall shows no clear trend, fluctuating up and down throughout the data sample. Chart by Rick Grow.

Though in recent times our snowiest winters have followed dry summers, the historical record shows this hasn’t always been true. The summers of 2009, 2002, 1995, 1986 and 1982 all recorded drier-than-normal rainfall amounts, and were followed by these respective anomalously high winter snowfall totals: 56.1”, 40.4”, 46”, 31.1” and 27.6”. In the more distant past, the summers of 1963, 1945, 1933, 1917 and 1891 experienced above-normal rainfall; exceptional seasonal snow totals – 33.6”, 24.5”, 30.7”, 36.4” and 41.7” – also followed these summers.

The statistically insignificant correlation between summer rainfall and winter snowfall helps explain why snowier winters nowadays tend to follow dry summers, as the long-term trend shows JJA periods becoming less wet. Drier summers in D.C. are simply more common and, therefore, the snowier winters that do develop have a higher chance of occurring after a dry JJA.

Since the relationship between rainfall and snowfall yielded nothing meaningful, I decided to focus on the temperature-snowfall relationship. It turns out that no statistical significance exists between the average summer temperature in Washington and total snowfall the following winter, either. A hot summer is almost equally as likely to result in a snowy winter as a cold summer.

Something did capture my interest though in plotting the trend in average summer temperature and winter snowfall over time (since 1888). The mean JJA temperature has been increasing, while the total seasonal snowfall has been decreasing. These observations confirm other data that show a warming and less snowy D.C. Note, though, that the warming trend in summer temperatures has negligible impact on winter snowfall.

Plot of the mean June-July-August (JJA) temperatures and winter snowfall totals since 1888. The linear trendlines for both variables show opposite trends, as summer temps have been warming and winter snow amounts have been dwindling. Chart by Rick Grow.
Plot of the mean June-July-August (JJA) temperatures and winter snowfall totals since 1888. The linear trendlines for both variables show opposite trends, as summer temps have been warming and winter snow amounts have been dwindling. Chart by Rick Grow.

What could be the one true link to D.C. snow, then? Move ahead a few months to autumn. October, a month known for idyllic Washington landscapes, holds the key.

The District has had 47 winters with 20” or more of snowfall. Of these 47 winters, 36 were preceded by a cold October! The average monthly temperature for all 47 Octobers combined is 1.5 degrees colder than the 1981-2010 climatological normal temperature. For reference, both October 2002 (40.4″ of snow the following winter) and 2009 (a record 56.1″ the following winter) were about 1 degree colder than normal, and October 1987 (25″ the following winter) was about 5 degrees colder than normal.

Snowiest_D.C._Winters_&_Prior_October_Temperature
Graph showing the 47 snowiest D.C. winters (seasonal totals ≥ 20”) and the average temperature anomaly recorded during the prior October. The blue bars display the snow totals and the red bars (at bottom) display the October temperature anomaly (positive and negative values representing the distance from the normal temperature). The overwhelming majority of anomalies are negative, reflecting a colder-than-normal October. Chart by Rick Grow.

To be sure, not every cold October will lead to a snowy winter, and not every warm October will preclude a snowy winter. The outcome really depends on the type of pattern that eventually develops for the winter; blockbuster snowfall seasons in D.C. are driven considerably by “blocking” in the atmosphere (high pressure and warmth locked in over Greenland and surrounding areas across the high latitudes).


The map at left shows cold October temperatures (compared to normal) prevail over the D.C. area and eastern third of the U.S. prior to the snowiest Washington winters. Only the October temperature anomalies prior to the snowiest winters since 1949 are reflected in the map given plotting data limitations. At right, I show the November-March composite 500 mb anomaly pattern for the same subset of snowiest winters. Note the blocking high pressure (yellow, orange, and red colors) extending from Alaska to Greenland. This type of blocking pattern is actually a better indicator of winter snowfall potential, but cannot be reliably predicted until after October. Both composite maps provided by NOAA/ESRL Physical Sciences Division.

October temperatures might just be the new bellwether for D.C. winter weather, just as October polling serves as a key indicator for Washington elections in November. Let the debate begin.

Comments
Show Comments
Most Read Local

Heat Tracker

35
90-degree days year-to-date
Yearly Average
36
Record Most
67 (1980,2010)
Record Fewest
7 (1886,1905)
Last Year
24

At a Glance

Mon.

partlycloudyday
-- /96°

Tue.

thunderstorms
77° /93°

Wed.

partlycloudyday
75° /90°

Thu.

thunderstorms
70° /86°
Drop 50%

Fri.

thunderstorms
71° /87°
Drop 50%

Sat.

thunderstorms
71° /83°
Drop 20%
National Airport
Dulles
BWI

Right Now

D.C. Area Almanac

 
National
Dulles
BWI
Avg. High
 
Avg. Low
 
Rec. High
 
Rec. Low
 
Sunrise
 
Sunset
°F
 
 
°C
Next Story
Brian McNoldy · July 9, 2013