WunderBlog Archive » Weather Extremes

Category 6 has moved! See the latest from Dr. Jeff Masters and Bob Henson here.

Hottest Summers, Coldest Winters for Contiguous U.S.: A Few Years Loom Large

By: Christopher C. Burt, 6:27 PM GMT on May 28, 2015

Hottest Summers, Coldest Winters for Contiguous U.S.: A Few Years Loom Large

Keeping track of all-time warmest/coldest daily maximum temperatures and all-time warmest/coldest months on record for any given site is a fairly easy task. However, very few NWS sites provide data concerning what their respective coldest climatological winters (December-February) or hottest climatological summers (June-August) have been. Researching 300 sites in the contiguous U.S. I have put together this summary for such. Below are the methods I used and some of the results, which proved quite interesting. For example, 73 of the 300 sites (almost 25%) measured their hottest summers on record during just the summers of 2010 and 2011.

Choosing the 300 Sites

In order to make an attempt to make this research of some use, I chose 300 sites across the contiguous U.S. that have long POR’s (periods of record) dating back at least 100 years or more and that are also more or less evenly spread across the U.S. geographically. The sites represent all of the country’s various climatic zones aside from high Alpine regions (where data for long POR’s is generally lacking). That being said, there are a handful of locations on the list that do not have very long POR’s but were the only sites I could find that represented large regions with very small populations. For example: Chisos Basin, Texas (in Great Bend National Park) with a POR only back to 1948. This site is in Brewster County that comprises some 6,200 sq. miles but with a population of only 9,300. Another example is Danner, Oregon with a POR back to only 1929. This site is located in Malheur County which is 9,300 sq. miles but with a population of just 30,500. These large regions simply have very few sites with climate data that span long periods of time. In any case, all but 10 of the sites have POR’s of 100 years or longer and 240 of the 300 have POR’s dating back to the 19th century. The list of all the sites, their data, and POR’s can be found in the links near the end of the blog.

Below are two maps showing the actual locations, one with the site names and one with only the locations.





Methodology

After identifying the 300 sites used in this analysis, I used several climate data sources to determine each site’s coldest winter and hottest summer of record. The most useful source was NOAA’s Online Weather Data (NOW Data) which has comprehensive statistics for thousands of locations across the U.S. For the most part, NOW data has been ‘Threadexed’ meaning that all the data available for each location includes such for all the various places within the town or city where official weather stations have been maintained. Virtually all the towns and cities in the U.S. have moved their respective weather stations from one location to another (within the city) at various times since the beginning of their POR’s. In the several cases where NOW data for a site was not ‘Threadexed’, I resorted to the climate summaries available from the Regional Climate Centers. These are outlined in the map below:



Unfortunately, the Northeast Regional Climate Center and Southern Regional Climate Center do not provide detailed historical climate summaries on their web sites. In the few cases where neither NOW Data nor the RRC’s provided the temperature data I needed it became necessary to dig into the actual published records contained in the USWB/NWS/NOAA Climatological Data by Sections and in the Department of Agriculture various state weather service monthly bulletins as well as the Monthly Weather Review for the early era of weather record-keeping. I have not included data from the Smithsonian observations (1849-1874) and only partially from the Signal Army Corps Service era (1870-1890). Much of this data is still currently recognized as ‘official’ by the NWS/ASOS/COOP sites and reflected in the NOW data collection including, for example, Lansing, Michigan, Minneapolis, Bismarck, Cheyenne, and New York City among many others. Of course, this raises the legitimate concern of comparing apples to oranges since the POR’s for all the sites vary. For instance, the winter of 1874-1875 is officially considered the coldest such on record for Minneapolis, Madison and Milwaukee in Wisconsin, and Lansing in Michigan but not for other sites in the region even though the data is available and the winter of 1874-1875 was likely their coldest as well. A more thorough analysis by another party with more time and resources at their disposal is warranted. This report is simply an initial ‘broad brush’ approach to the subject.

The Results

What immediately becomes clear is that just a handful of extremely cold winters and hot summers account for the vast majority of the records. For instance, the winter of 1948-1949 was the coldest climatological winter (December-February) on record for virtually every site in Washington, Oregon, California, Nevada, Arizona, and Idaho. The winter of 1917-1918 remains the coldest for almost all sites in New England, New York, New Jersey, and points south to and including Washington D.C. The back-to-back winters of 1976-77, 77-78, and 78-79 accounted for almost all the record cold winters in the Ohio Valley, Midwest, and Central Plains. These five winters alone account for two-thirds (199 out of 300) of all the coldest winters for all the sites used in my summary. Conversely, the climatological summers (June-August) of 1934, 1936, 1988, 2010, and 2011 account for 130 out of 300 of the hottest summers for all sites. The other interesting result of the study is the apparent ‘clustering’ of some of the coldest winters and hottest summers. The summers of 1934 and 1936 and 2010 and 2011 stand out, as do the winters on 1976-1977 through 1978-1979. It is also interesting that in the anomalous years (weather-wise) of 1936 and 1949 many sites across the U.S. endured both coldest winters and hottest summers. In fact, in 1936, a few individual sites measured BOTH their coldest winter and their hottest summer that same year (Glasgow, Montana, Aberdeen, South Dakota, and Grand Forks, North Dakota).





Above are graphics indicating the number of sites (out of 300) recording their coldest winters and hottest summers arranged by year (top graph) and by decade (bottom graph). In the annual graph the winters and summers that had at least 10 sites with record seasons are highlighted. The data for 2015 includes the winter of 2014-2015 (none of the sites recorded their coldest winter) but, obviously, there is no data yet for the summer of 2015. In 2014 two sites recorded their hottest summer on record (Portland, Oregon and Yakima, Washington) and during the winter of 2013-2014 three sites their coldest winter (Eau Claire, Wisconsin, and Ironwood and Marquette in Michigan).

It is hard to draw any macro-climatic conclusions from these events since the contiguous U.S. only comprises 2% of the Earth’s surface, but the results are interesting nevertheless. The other trend that becomes obvious is that it has become more common to observe ‘hottest summers on record’ versus ‘coldest winters on record’ over the past 25 years or so. This trend is confirmed by the NOAA trend analysis graphics (represented below) that include the entire GHCN database.





Average temperatures for climatologcial summer (top) table) and climatological winter (bottom table) for the U.S. since 1895. This data is for the entire contiguous U.S. and based on all reporting sites in the GHCN database. It is interesting to note that the top three warmest summers on record, 1936, 2010, and 2011 also equate with years with the most sites with warmest summers in my database. This also holds true for the coldest winter of 1978-1979. Note that in the NOAA data the winter of 1948-1949 does not stand out the way it does in my graphic. That is because all the sites that endured their coldest winter on record were confined to the western quarter of the nation. The Midwest and East had a fairly normal winter that season (1948-1949). Tables from NOAA.

The Data

For those interested in the actual data for each of the 300 individual sites used in my project please find the complete data set here. In this link I also included data for Hawaii and Alaska, although these records are not used in my analysis for obvious reasons (Hawaii and Alaska are beyond the geographical parameters of this project) so just a FYI. I have also sorted the data by year for the coldest winters here and hottest summers here. In addition, for those interested, find here a chronological list of the POR’s for each site. The dates indicate the beginning of temperature records for the site indicated. However, in many cases there are months and even a few years missing since the beginning of the record for some sites. NOTE: The links provided above will load as Excel files.

Next Step

Hopefully, someday someone will produce a more professional survey of what I have done here and refine this initial attempt. Homogenizing the site POR’s would be an important first step.

KUDOS: Thanks to Mark Stroud of Moon Street Cartography for the maps and graphics reproduced above.

SOURCES USED:

NOW Data (NOAA Online Weather Data). This can be found on all the NWS weather web sites for each respective city/region.

U.S. Regional Climate Centers: Historical Climate Summaries U.S. Weather Bureau

Climatological Data by Sections U.S. Weather Bureau

Climatological Data (by states) NOAA

Climatology of the United States: Bulletin Q, U.S. Department of Agriculture, Weather Bureau, 1906

Monthly Weather Review, U.S. Department of Agriculture


Christopher C. Burt
Weather Historian

Extreme Weather Cold Heat

The views of the author are his/her own and do not necessarily represent the position of The Weather Company or its parent, IBM.

Reader Comments

Quite an undertaking, Christopher. Thanks for sharing your methods and data sources in addition to your interesting results.
It would also be interesting to plot the top 5 or top 10 for each station. That might average out the anomalous years.

Of course, at some point you're just looking at average temperature, so...
Since it took me 440 hours to just research the number 1 hottest summers and coldest winters (although that includes time spent updating ALL the WU city extremes database), researching the top 5 or 10 would take a year at least (if I spent all my time 40 hours a week doing so). One has to look at every single December, January, February, June, July, and August average temperature for all 300 sites to figure that out!

Quoting 2. JazzChi:

It would also be interesting to plot the top 5 or top 10 for each station. That might average out the anomalous years.

Of course, at some point you're just looking at average temperature, so...
Personally, showing the trend across 120 years of data tells alot about the situation,
yearly differences aside, I have noticed more of a warming trend in the winter, Fall held out until Christmas week this year, when it usually rolls in around Thanksgiving

This is local however, I remember a couple of years ago where a cold high pressure system set up near Alamosa, where it was dropping to -45F at night.(location was actually colder than Alamosa by quite a bit) My folks propane furnace wouldn't run (scary dangerous to even try, liquid propane was pooling in the pipes) and they were burning through a cord of wood a week in a 1400SQFT superinsulated house that January, I remember getting up at 2am to restoke the fire, which we usually let die off overnight.

Aside from that bit, I haven't seen temps like that since the late '90s here.
What a fantastic entry Mr. Burt! I'll be studying it a good while. This is the best entry you've done.
Congratulations Christopher, on some excellent research.

One thought though (to add to your massive work load) and that is, do you know of any particular correlated reasons for the abnormalities of the 1930s and the 1970s and were they different to the causes of the latest 2010 trend?
Were those years of high (or low) El Nino/La Nina activity?
Were there any large volcanic eruptions or sun cycle peaks that influenced the weather - or any other major events that match up?
Or were they merely part and parcel of greater ongoing cycles?
Always a delight to see your work and graphically depicted in a manner to condense many hours of hard work into something visually enlightening. The coldest winters, when grouped by decade seemed to be starting a pattern, ie 1910-19, then 1940 -49, then 1970-79. Thought a simplistic observation, it almost suggested a 30 year period between clusters and an increasing trend culminating with 1970-79. But from there we see no similar cluster 30 years later, ie 2000-2009. And perhaps that is no surprise as a warming global climate would make a cluster of coldest climatological winters difficult. The sobering thought about 2010-2015 is that we are only halfway through this decade and already at 93.
Chris, this is an excellent piece of research. And I greatly appreciate you including a link to the actual data you used; I plan on borrowing it perform some analyses of my own. That you labored so many hours to do this is greatly appreciated.

I am wondering, however, why the Chisos Basin station in BBNP is the station of record; it sits well over a mile above sea level, and the basin generally experiences far cooler temperatures than most anywhere else in the entire park. I realize the important part is maintaining a long-term record wherever that station may be. But one has to wonder whether some very hot temperature readings are being completely overlooked due to siting. (I know, because the Big Bend one of my favorite places in the country, and I've been a number of times. In addition to its rugged beauty, remote location, and almost complete lack of people, It's got the darkest skies in the US, ranking a solid 1 on the Bortle Scale. In fact, it's the only place I've ever been in the States where the Zodiacal light was visible all night long. Passing clouds appear not as faint wisps of white, but as black voids drifting against the stars. But I digress...)
Thanks, Chris. The graphics are especially compelling.
You reminded me that my father, who was a NWS (Weather Bureau, actually) meteorologist from the early 30s through the 50s, espoused a theory that extremes follow extremes. And your annual plot shows why he would have come to that belief.
Yes, choosing Chisos Basin was a tough call. I wanted to use Presidio which has a POR beginning in 1928, but for some reason, the NOW data for Presidio ends in January 2012. Its coldest winter was 1963-1964 (47.7° avg) and its hottest summer 2011 (like most of Texas) with an 91.9° average. But since it's 'official' record seems to have come to end it will be hard to track going forward. Frankly, I don't understand this situation since the site still seems to report weather data on a daily basis. BTW, I have the incorrect figure for Chisos Basin's coldest winter of 2009-2010: it should read 44.3° not 40.7° (that was its single coldest month set in January 1992) but the years are still correct. I was surprised that 1963-1964 was milder than 2009-2010! I could find no other sites in the region with long POR's (Marfa is just 1958-2006) and nada for Alpine.

1
Quoting 8. Neapolitan:

Chris, this is an excellent piece of research. And I greatly appreciate you including a link to the actual data you used; I plan on borrowing it perform some analyses of my own. That you labored so many hours to do this is greatly appreciated.

I am wondering, however, why the Chisos Basin station in BBNP is the station of record; it sits well over a mile above sea level, and the basin generally experiences far cooler temperatures than most anywhere else in the entire park. I realize the important part is maintaining a long-term record wherever that station may be. But one has to wonder whether some very hot temperature readings are being completely overlooked due to siting. (I know, because the Big Bend one of my favorite places in the country, and I've been a number of times. In addition to its rugged beauty, remote location, and almost complete lack of people, It's got the darkest skies in the US, ranking a solid 1 on the Bortle Scale. In fact, it's the only place I've ever been in the States where the Zodiacal light was visible all night long. Passing clouds appear not as faint wisps of white, but as black voids drifting against the stars. But I digress...)
A tremendous piece of undoubtably tedious work! Thanks
Quoting 3. weatherhistorian:

Since it took me 440 hours to just research the number 1 hottest summers and coldest winters (although that includes time spent updating ALL the WU city extremes database), researching the top 5 or 10 would take a year at least (if I spent all my time 40 hours a week doing so). One has to look at every single December, January, February, June, July, and August average temperature for all 300 sites to figure that out!




This is what grad students are for...combing through endless streams of data looking for that one nugget.
Average number of Hottest Temp sites * 1900-1979 = 12.5
Average number of Coldest Temp sites * 1900-1979 = 33
Average number of Hottest Temp sites * 1980-2015 = 45
Average number of Coldest Temp sites * 1980-2015 = 4.5
Who can deny that the contiguous U.S.A. has gotten MUCH warmer since 1980 ??
Quoting 12. Astrometeor:



This is what grad students are for...combing through endless streams of data looking for that one nugget.
Favorite joke of a department chair at a college of engineering where I studied:

When should you award a graduate student their degree?

Just before they decide to kill you. :^/

What happened to Riah's weekly weather roundup?  Please bring them back, Ol' weather historian!
Sorry, who is "Riah"? I must have missed something here.

Quoting 16. Lemming:

What happened to Riah's weekly weather roundup?  Please bring them back, Ol' weather historian!
Quoting 17. weatherhistorian:

Sorry, who is "Riah"? I must have missed something here.


Riah used to work for Weather Underground and made quirky but entertaining weather news videos. Riah Gouvea | LinkedIn
Thanks Mr. Burt- A great project of great interest. Im looking forward to the expanded global version….
The trend lines for coldest winters versus hottest summers are most interesting, even accounting for the 2% landmass they represent...
Mr. Burt I found this interesting site which is right up your alley. Do you know of a site giving state monthly extreme maxima and minima updated to the present, or do you have a page of your own? Thanks!

Monthly State Maximum and Minimum Temperature Extremes (as of May 1, 2004)
One minor correction... You write: "This data is for the entire contiguous U.S. and based on all reporting sites in the GHCN database. It is interesting to note that the top three warmest summers on record, 1936, 2010, and 2011 also equate with years with the most sites with warmest summers in my database. "

It looks to me like it's actually 1936, 2011 and 2012 that are the three hottest on the chart. 2010 is a little cooler - still top ten warmest on the graphic, but not top three.
Interesting study; however, an analysis of extremes in any study never yields realistic statistical results. This is especially true for climatological trends - particularly when the data set used represents only 2% of all measurable data. So many such studies 'cherry pick' data collection sites and employ methods of measurement and extrapolation to achieve a result that supports a scientific theory and/or political agenda. I'm not implying that was Mr. Burt's intension. But, if we are truly concerned about climate change, we must be honest in our approach and avoid presenting findings that are easily unravelled by opposing scientific theories and/or political agendas. 95MERCURY's comment "Who can deny that the contiguous U.S.A. has gotten MUCH warmer since 1980 ??" indicates that even knowledgeable people can be easily lured into accepting that a study of extremes is irrefutable evidence of a scientific fact. I'm surprised that 95MERCURY's and other reader's comments don't reflect any curiosity over how and why temperatures can be higher now than during years when US industry was at its height - and prior to the existence of any environmental legislation. The EPA's own data shows that average vehicle fuel economy in the US has essentially double since the 1970s, and that vehicle CO2 emmisions have been effectively cut in half over the same time span. Since passage of the Clean Air Act in 1970, the EPA claims that the most common air pollutants have decreased by 50%, toxic air pollutants from industrial sources (i.e. chemical, petroleum, and paper production) have decreased by 70%, current production vehicles are 90% cleaner, and production/use of ozone depleting chemicals has virtually ceased. In that same time, every fuel-powered (electric and fossil source) industrial machine and consumer product in the US operates more efficiently than it did 30 to 40 years ago. Further, since 1970 the steel industry in the US, the single most focussed target of EPA greenhouse gas emission reduction protocols, has declined by 50%. So, here's my simple question: if climatological trends are implying a predictably steady increase in average temperatures in the US, caused as some believe by 'human activity', why have all the admitted successes of EPA greenhouse gas policy and related changes in human activity failed to reverse or even slow these grim climatological trends? Maybe it's not us. Maybe we are simply too eager to accept a cause/solution model because it's easier for politicians sell and more convenient for believers to apply in their lives. Science is a pursuit for truth, and posing challenges to accepted beliefs or proposed theories will always advance that pursuit. This may not be a popular concept to those looking for easy answers, but to deny this concept reduces our chances of accurately identifying the causes of global climate change and puts at risk any opportunity to discover a realistic solution.
Re. 22


Source: ESRL Global Monitoring Division - Global Greenhouse Gas Reference Network

CO2 has continued to accumulate in the global atmosphere despite any variations that have occurred in any particular region. Some of the common air pollutants that have decreased in the US since the Clean Air Act in 1970 were aerosols that reduced solar radiation reaching (and warming) Earth's surface.

Until someone can cite an actual paper by a degreed climatologist that presents a conclusion like "climatological trends are implying a predictably steady increase in average temperatures in the US" I have to conclude that is either a mistaken idea or a made up straw man argument.
During this 2010s time period it seems like the 1930s keep popping up as a some what similar weather era. Your graph makes the perspective of the difference and similarity of those two more clear.
Dear LowerCal,

You may find this report by a few degreed climatologists useful: https://www.ipcc.ch/pdf/assessment-report/ar5/wg2/ WGIIAR5-Chap26_FINAL.pdf

Cheers
Quoting 25. KelvnHelmholtz:

Dear LowerCal,

You may find this report by a few degreed climatologists useful: https://www.ipcc.ch/pdf/assessment-report/ar5/wg2 /WGIIAR5-Chap26_FINAL.pdf

Cheers
No "steady increase" in the entire 60 pages of that report much less in section 26.2.2 where you expect to find it (if the claim was actually made).

Anyway the poster of comment 22 made the claim so it's up to that poster to cite a credible source. Otherwise that poster it leaves me to conclude that it's a mistake or a lie by that poster or their unreliable source.
XLPilotSmith has made a good point in that the U.S. has made some serious advances in air quality management, but unless I read wrong there was no mention of global industrialization, which my guess is has well compensated for the decrease in U.S. emissions.
I don't suppose you could do one for the wettest vs. driest years. The info you provided is fabulous!
Is there any data that correlates solar activity with these findings?
On the average summer and winter charts two entirely different lines could be drawn for each; 1895 thru 1970, and 1970 thru 2014. In each case the former would be flat and the latter would have all the gain. What happened around 1970 to start the rapid increase?
Yes, that would be interesting and since I researched ALL the extreme records for all these sites this will be a relatively easy task. I'll try to post the results in August.

Quoting 28. revclaus:

I don't suppose you could do one for the wettest vs. driest years. The info you provided is fabulous!
Just look at the actual data (which I provide at the end of my blog with URL's to take you there) and that will answer your question. So far as "what happened", I don't know, that is beyond my pay grade.

Quoting 30. joynickhhi:

On the average summer and winter charts two entirely different lines could be drawn for each; 1895 thru 1970, and 1970 thru 2014. In each case the former would be flat and the latter would have all the gain. What happened around 1970 to start the rapid increase?

It would be interesting to see if the crazy temp anomalies in the 1930s had something to do with solar activity.

Quoting 29. whitenoise500:

Is there any data that correlates solar activity with these findings?
What would your graphs look like if you used unadjusted original data that has not been massaged by any federal agency? NOAA black-box algorithms are not to be trusted since Climategate. It was much hotter during the Dust Bowl than today. Ask your grand parents-- if they are old enough to remember.
thank you
weatherhistorian has created a new entry.
Quoting 34. TrueSkeptic:

What would your graphs look like if you used unadjusted original data that has not been massaged by any federal agency? NOAA black-box algorithms are not to be trusted since Climategate. It was much hotter during the Dust Bowl than today. Ask your grand parents-- if they are old enough to remember.

Of course, so the ice melts because elves have raised the melting point of water and growing seasons are changing because the commies told them so :D :D
I love your blog posts, but the plethora of science denial(y) folks that seem to be attracted here is quite a slog to navigate. Good Data tends to do that. :)

weatherhistorian has created a new entry.