Network News

X My Profile
View More Activity
The new Washington
Post Weather website
Jump to CWG's
Latest Full Forecast
Outside now? Radar, temps
and more: Weather Wall
Follow us on Twitter (@capitalweather) and become a fan on Facebook
Posted at 11:00 AM ET, 01/19/2010

Making sense of the late-January thaw

By Jason Samenow

* Colder & stormy late week: Full Forecast | Comments for CWG? *

Average daily temperature for January, for 138 years of raw data. The graph shows temperatures increase around Jan 19-22, and also around Jan 5-7. Graphic by Steve Zubrick, National Weather Service.

After five weeks of colder than normal temperatures, we've flipped to a warm pattern over the past week. I've heard many refer to this warm-up as a "January thaw." And I think that's a fair enough description. After all, for a period of 11 days, the average daily high was in the mid-30s and now 4 of the last 5 days have reached 50 or higher.

It's not at all uncharacteristic for temperatures to swing like this in the month of January. In fact, the commonality of January temperature spikes during the second half of the month following cold waves has motivated researchers to investigate whether the January thaw can be seen in the data and whether it's a real phenomenon.

Steve Zubrick, the Science Operations Officer at the National Weather Service Office in Sterling, Va., plotted the long-term data for D.C. and Baltimore (spanning over 130 years -- as shown above), and found some limited evidence for "thaw" like increases in temperatures early and late (around right now) in the month. But do we have any legitimate scientific explanation for the fact the warm-up seems to show up (on average) around the same time each year, or has it simply been recurring around this time by chance?

Zubrick doesn't think the coincidental timing is indicative of any real scientific phenomenon: "It's in the data, but is has no significance related to some known weather driver."

Zubrick referred me to a scientific paper published in 2002 entitled "Is the January thaw a statistical phantom?" that concluded about the January thaw:

...the observed warm spells are well within the limits of what might be expected to occur by chance alone in a stationary climate during any random period...


"... no dynamical basis, or even a plausible physical mechanism, has been advanced in the literature to explain why a warming in northeast U.S. temperatures should occur during this particular narrow time window."

In other words, the repeated thaw occurrence around this time is probably a statistical phantom or as Zubrick put it "a human un-scientific perception rather than a scientific reality."

By Jason Samenow  | January 19, 2010; 11:00 AM ET
Categories:  Local Climate  
Save & Share:  Send E-mail   Facebook   Twitter   Digg   Yahoo Buzz   StumbleUpon   Technorati   Google Buzz   Previous: Forecast: Warming fades after today
Next: PM Update: Mild air to trend cooler


When I was young, it was said that the period between Dec. 3 and Dec. 14 was mild more often than not--though this does not appear to agree with the fact that our first snow in Washington seems to occur on Dec. 5 four years out of five nowadays!

In his 1955 book "Our American Weather" Dr. George H.T. Kimble advanced the opinion that warm spells and cold waves appeared to have a preference for certain dates of the year; he cited the "January thaw" and "Indian summer" as notable examples of this phenomenon. Twenty-seven years later, David M. Ludlum noted similar trends in his "American Weather Book". My guess is that seasonal climatic rhythms have a periodicity that tends to repeat somewhat in the manner of a trigonometric function.

It appears that Dr. Kimble knew about several patterns such as the eleven-year sunspot cycle, while being relatively unaware of others we know about today--the El Nino/Southern Oscillation, the North Atlantic, Arctic, Antarctic, Pacific Decadal and other regional oscillations, the Madden/Julian Oscillation, the Milankovich cycle and others. It is the interaction among these various short-term and long-term atmospheric cycles which drive the world's weather machine. Are we contributing to global warming, ourselves?
Based on the principles of chaos theory, the answer is likely yes. If a butterfly flapping along the Amazon can influence the weather at the Arctic Circle, a car engine running in Chantilly, Arlington, or Gaithersburg can exert greater influence...and nothing Andrew Freeman might argue, or Mr. Q. might say to the contrary can change that!

Posted by: Bombo47jea | January 19, 2010 1:02 PM | Report abuse

The conclusion in the paper seems intuitively right to me. You have to have a putative mechanism (some complex chaotic phenomenon, perhaps damped reverberation from the "shock" of the onset of winter? seems pretty far fetched) for this to be anything but the chance happening of how our 130-year data has played out. And obviously they compared the size of this bump to the general ongoing variance to decide that it wasn't significant.

You could probably pick any city in the country and look at their 130-year average trace and find some intriguing bumps and valleys. All noise, against the background cycle of the seasons.

[No, wait! That thaw happens right after the 20th. Clearly it is the temporary warming of political differences after the inauguration of a new president that's driving it! And in the off-election years, nature just sometimes does it out of habit... yeah I'll go with that.]

Posted by: B2O2 | January 19, 2010 1:16 PM | Report abuse

i find that graph absolutely fascinating. i would have expected a MUCH smoother curve. i'm sure you could "flatten" it out some, but those two warm bumps seem "real" to me...

1)there's a pretty good correlation btwn wash and baltimore. i wonder how far, geographically, that holds? like, does frederick or harrisburg or philly or ny or boston show those "bumps".

2)if the curve were smoothed, where would it bottom out? i.e. what's the coldest day of the year?

Posted by: walter-in-fallschurch | January 19, 2010 1:33 PM | Report abuse

That graph does not seem to agree with noaa statistics for BWI as rendered at The noaa stats show the average daily high sinking to 41 on 5 January and not climbing again until reaching 42 on 30 January. The average lows sink from 24 to 23 on the 11th and rise from 23 to 24 on the 27th. I don't see any evidence of a spike in early or late January.

Posted by: batebud | January 19, 2010 2:29 PM | Report abuse

Great synopsis. I love reading about the statistics behind the January thaw! It is often the source of much speculation. Just like big snowstorms statistically happening in DC on weekends vs. weekdays.

Posted by: Camden-CapitalWeatherGang | January 19, 2010 4:42 PM | Report abuse


I think the data you're referring to is smoothed somehow. I'll check with NWS and verify.

Posted by: Jason-CapitalWeatherGang | January 19, 2010 5:56 PM | Report abuse

I remember even as a kid in the Mid-to-Late 1980's, there would be days around here in January when we'd come home from school and have to put on shorts (It would be in the 60's to 70's (Fahrenheit) ). At the same time, we would also have Winters where you wouldn't even see the grass on the lawn until late March or April (From December onward).

Also, the latter Winter periods of bitter cold and snowfall should never be ruled out (Obviously history has taught us this). I remember my grandmother recalling that one of the biggest winter storms they ever faced in Southern Virginia, was EASTER weekend!

The bottom line is that living in the very epicenter of the US East Coast (Within the Mid-Atlantic), always provides for rather unpredictable and exciting weather patterns. Folks down South take their warmth for granted, while people up North take the snow for granted. We on the other hand have a craps table upon which we exist, and at times we win, while sometimes we don't (It all depends upon your preference as well!).

Posted by: TheAnalyst | January 19, 2010 11:44 PM | Report abuse


Jared Klein, Climate Program leader at the local NWS office,sent me an email stating the following:

"The webpage that the commenter referred to is the 1971-2000 normals for BWI. The National Climatic Data Center calculates the 1971-2000 daily (max, min, avg) temperature normals using a interpolation method called the cubic spline fit in order to smooth the temperature data without using the daily data. By definition of a cubic spline-fit, the average of all daily (max, min, avg) temp values for the month equals the 1971-2000 monthly normal temperature. You can refer to for more details on the cubic spline method and go to *A. Spline-Fit Daily Normals *under the *Daily Station Normals *section."

Posted by: Jason-CapitalWeatherGang | January 21, 2010 3:09 AM | Report abuse

The comments to this entry are closed.

RSS Feed
Subscribe to The Post

© 2012 The Washington Post Company