February U.S. Temperatures

March 19th, 2014 at 2:02 am by under Bill's Blog, Weather

February 2014 divisional temperature ranks  Click on the map to enlarge.  This map from NOAA  is the divisional rankings for Feb. 2014 temperature.  Look closely at the colors and ranks.   Grand Rapids was 9.1° colder than average in Feb.  From the GRR NWS:  “AN AVERAGE DEPARTURE OF 9 DEGREES FOR AN ENTIRE MONTH IS VERY SIGNIFICANT. POLAR AIR FREQUENTLY STREAMED IN FROM CANADA AND RESULTED IN THE 11TH COLDEST FEBRUARY ON RECORD IN GRAND RAPIDS AND MUSKEGON.”   Yet, on the map, we get only a “below average”, not a “much below average”.   Other departures:  LAN -9.2, Kalamazoo -10.0, Holland -10.8, JXN -9.3, BTL -9.4, DTW -8.7, APN -9.5, HTL -9.4, TVC. -8.5, FWA -10.7 (5th coldest),  Fargo ND -10 below normal. Grand Forks -11.6. Aberdeen SD -11 below norm. Rapid City SD -8.6 below.  Great Falls MT -12.6.

Compare that to the “MUCH above” area in the SW:  Los Angeles +2.6, San Diego +3.3, Las Vegas +4.1, Kingman AZ +3.9, Phoenix +4.7, Yuma +4.4, Sacramento +3.0, San Francisco +2.0.

They’ve managed to draw the “Much Above” area bigger than the “Much Below” area, despite the fact that actual monthly temperatures would suggest otherwise.  Also, look how close the light blue is to the white average color.

Lower Michigan should be in the dark blue.  This is a 120-year record, so to be in the top 10% you would have to have a top 12 coldest winter.   Alpena and Houghton Lake had the 2nd coldest winter, Flint was 5th, Grand Rapids and Muskegon were 11th coldest.  I can’t find a single reporting station that wasn’t in the top 15.

If anyone has the time…check around and see if you can find other February departures from average.  If you were objectively drawing a map of February temperature, how might your map differ in color from this map.

Also:  Dulles Airport in Washington D.C. has had 16 inches of snow so far this month, setting a new record for March….and bombing the Yellow River in China.


32 Responses to “February U.S. Temperatures”

  1. TomKap (Michigan St. & Fuller) Grand Rapids says:

    Let’s move on to APRIL and more SUNSHINE!!!! IT’S COMING SOON! HOORAY! :)

  2. Wswplz says:

    Bill wouldn’t this be because many of these moderate / warm climates don’t have much fluxuation in there temperatures therefore not having as significant +/ – variations from there average. For example San Diego’s normal temp for the month of February 57.9 however observed was 61.2 ( 21014) and the warmest average was 63.2 ( 1980) so that’s about a 5.3 deg departure from normal. Maybe they could scale the map different to account for this. Maybe I’m missing the boat, it’s late lol or early.


    1. Wswplz says:

      * 2014 rather 5.6 deg departure from record average * Last month Feb departure + 3.3

    2. Wswplz says:

      I think it’s a little misleading, that’s why I posted there other map with the shades of blue.

    3. Cort S. says:

      You’re onto something which is also important, Wswplz, but that’s not how the map was created (but props anyway for a good suggestion!). I have addressed this below.

  3. Travis Ulberg (West of Martin, Allegan County) says:

    Joe Bastardi was all over this on his Twitter. I had to do a couple double checks to tell the difference in the shades of blue. Talk about molding the stats to your liking.

    1. Cort S. says:

      Yeah, the colors kind of suck, but it’s easy to change those:


      What’s hard to do is find color shades that satisfy everyone! Are there any professional cartographers or artists out there who can suggest alternative color schemes?

      Also, has anybody else considered the fact that depending on what monitor you use (or what angle your monitor is tilted at if you have an LCD screen), the shades will look different?

      1. Travis Ulberg (West of Martin, Allegan County) says:

        That map is much better. They just need to tweak their categories to resemble temps compared to average, and not by rankings. I kind of suspect they made it that way on purpose though.

        1. Cort S. says:

          Nothing wrong with rankings. It’s just one of many ways to look at data. The mistake is using a map of rankings in a discussion about departures. The website I posted below allows you to make maps for rankings and maps for departures. But I wish it would also give us an option to make a map for standardized anomalies, that is, what the standard deviation (sigma) of our temperature departures are.

      2. Dikehopper says:

        You are quite right, Cort. Different monitors give different shades of colors.

        Plus, I use a TV for a monitor and it gives me problems with being able to recognize some hyperlinks. That is, if the color of a hyperlink is too light, it doesn’t show up very well. Easy to miss.

  4. mart (gr) says:

    they have managed to make Florida even more phalac

  5. Cort S. says:

    Here is a great website where you can play with the climate data maps yourself:


    Based on their definitions, you are “Much Below Normal” if your rank is in the top 12 coldest, and you are “Much Above Normal” if your rank is in the top 12 warmest. Ranks within the top 12 coldest or warmest would put you in approximately the most extreme 10% of years on either side of the spectrum, since the maps deal with 118 years of data.

    There’s some apples and oranges in the blog post: The map shown is of temperature rankings, but the numbers discussed in the text are temperature departures. Here is a map of temperature departures: http://i.imgur.com/NGR3pqq.jpg

    You may notice the temperature departures in the map I posted are not the same as what Bill listed. There’s some oranges and grapefuits: This is likely because the map uses a 1901-2000 average but the numbers in the blog post used a 1981-2010 average. If I’m mistaken, let me know.

    1. Cort S. says:

      What other important statistical information would I like to see offered by the map-making website I posted? Some information about variance! What is the standard deviation of each division’s temperature departures? This is important because variance is not the same everywhere: a 10 degree temperature departure in (for example) Grand Rapids is less significant than a 10 degree temperature departure in (for example) Seattle. This is great information that is not discussed very much in weather circles! A map of temperature departures will have no information about each division’s variance, and a map of temperature rankings will have some hints about variance but won’t come close to truly representing it.

    2. Travis Ulberg (West of Martin, Allegan County) says:

      Since 3-5° above average in the SW = “much above”, and 8-10° below average in lower Michigan only = “below”, wouldn’t their definition of “much above”, and “much below” entail that in the past we have had more extreme cold compared to average than extreme warmth compared to average?

      1. Cort S. says:

        First, let’s be careful with the apples and oranges and grapefruits.

        The map of “much above” and “much below” comes from temperature rankings… that is, on a scale of 1-118 (1 being coldest and 118 being warmest), where does this year’s February fall amongst all the other years’ Februaries?

        Here are the rankings: [apples]

        Top 10-15 coldest for Michigan, and top 2-9 warmest in the SW US.

        Now here are the temperature departures:

        Bill’s numbers, based on 1981-2010 averages (I presume): [oranges]
        About -9° in Michigan and +3° in the SW US.

        NCDC’s map‘s numbers, based on 1901-2000 averages: [grapefruits]
        About -6° in Michigan and +5° in the SW US.

        There are two things that I think I could infer from this information:

        (1) 1901-2000 averages may be colder than 1981-2010 averages. (But is this equally true for Michigan vs. the SW US? It might not be.)

        (2) The February temperatures from year to year in Michigan may have a greater variance than in the SW US. This means that fewer degrees departure from normal in the SW US may lead to a more extreme ranking than it would in Michigan.

        1. Cort S. says:

          Correction! There’s actually 120 years used for rankings now (1895-2014), not 118. That leads to a minor change in my numbers: Still ranked top 10-15 coldest February in Michigan, but about 4-11 warmest in the SW US. (Based on my earlier estimates, but really it’s still 2-11 warmest in the SW US because I missed southeast Arizona in my first estimate.)

        2. Bill Steffen says:

          In 1895, Arizona, New Mexico and Utah weren’t states. If I remember right Arizona and New Mexico weren’t full states until 1912.

        3. Cort S. says:

          While we’re talking history, the United States weather bureau was established in 1870 by Ulysses S. Grant, and Utah, Arizona, and New Mexico were territories of the United States in 1895.

        4. Bill Steffen says:

          If Fort Wayne had their 5th coldest Feb. ever, why isn’t NE Indiana a darker blue?

        5. Cort S. says:

          I don’t know the methodology behind how divisional anomalies are calculated for the sake of ranking in a 120-year period, but my (and your) ignorance of the methodology does not mean that this information cannot be found out. If proper science is being done, the data set and methodology will be published somewhere or at least be open to access. This rule goes for all scientists who want to do science (which includes making one’s own maps).

          Here are some other things to consider for the sake of a nuanced academic discussion:

          (1) Additional stations beside Fort Wayne are used to calculate divisional anomalies.

          (2) The period of record for the Fort Wayne climate station goes back to 1912.

          (3) We can probably find individual cities in the SW US that also exceed the ranks for their division. But why should we? This isn’t the root of the issue.

          Let’s back the trolley up. We need to acknowledge that your primary case against the divisional ranking map you posted (besides being a crappy color scheme) is that the rankings and the definitions of “much above” and “much below” don’t intuitively match up with temperature anomalies. You made this point without proper consideration and academic discussion given to the fundamental differences between what rankings, anomalies, and standardized anomalies are. Your audience needs this stuff explained! Additional information missing from the discussion was the map’s definition of “much above/below” (which is a Top 12 rank), and no mention of what period your normals are from (are they from 1981-2010?) and how this compares to other periods for normals (including the century average used by NCDC anomaly maps) and how these temporal comparisons may also vary spatially.

          You need to first acknowledge and correct these weak points in the foundations of your argument, then you can take NOAA to task about other things! These might include: limited or unreliable data set especially in the Southwest especially early in the period of record, potentially flawed methodology(ies) in the massaging/smoothing of data, questionable application of ranking data to form the conclusions used in a certain narrative… I don’t care! Go nuts! Science is all about questioning, proposing, and testing! But make sure that you at least get the fundamentals correct first.

          This has been a peer review, by the way. :)

        6. Bill Steffen says:


          Many stations in the SW have a much shorter climate record.

          Another question, and I don’t have an answer to this, is whether this map of rankings was given wider dissemination this month than any other month, or in general because it does give the impression that the month was warmer than it actually was. I’ve seen this “trick” used by picking the 30-year mean of 1951-1980 in order to get a lower average. Even stopping the longer century average at 2000 makes the average cooler by excluding 2000-2013. The common person doesn’t also think of such things as moving weather stations to airports since 1900 and the switch to MMTS (which required electricity and caused some weather stations to be moved closer to buildings where there was electricity).

      2. SlimJim NW GR (1) says:

        Thank you Cort for your very good explanation on this. If you look at this site you can get a very good idea of how the data was used and see there was (is) no (from the top) conspiracy



    3. Cort S. says:

      Correction! There’s actually 120 years of data, not 118.

  6. Travis Ulberg (West of Martin, Allegan County) says:

    I think Wswplz hit the nail on the head with his post. This map is based on temperature relative to the record(hot/cold), and not based on the average. Which that type of analysis creates a very misleading map. 10° below average in the midwest is much more noticeable/impactful than 4° warmer than average in the SW.

    1. Cort S. says:

      No, Wswplz is incorrect (sorry Wswplz!). But he made a good guess, and he’s onto another important point which I discussed in one of my posts above. He’s talking about variance, Bill’s talking about departures (or anomalies), but that map is rankings. In my post above, I explained how that map defined “much above” and “much below”. It’s in bold.

      1. Wswplz says:

        Enjoyed the discussion, wish I could have jumped in earlier as I would have liked to clarify my position, and it’s interesting to see what I thought I was saying and what actually was interpreted.( I can see how it’s not clear) My San Diego example was merely to the point out variance as Cort so nicely explained in an earlier post. I wasn’t trying to separate record / average just the departure from average still could be ” large” without much variance at all, since the temperatures there are very stable. That was my point, as I never dove into a solution , but fell asleep. Lol.

        1. Wswplz says:

          I figured I should have left my terrible example out as it made no since the way it was presented, much to do with falling asleep and typing. Thanks Travis U for trying to give me credit even when where credit was not due, but I see how that assertion could be made. I still feel the map could be presented in a manor which is more defined. I didn’t open the map that Cort has modified as I need to update this tablet with Adobe Flash, so not sure if you just changed shading or provide more information as well. Thanks again Cort, for a wonderful explanation, I especially like the phrase you used , let’s back the trolley up, my prof at SDSU used to frequently say that. Lol

    2. Bill Steffen says:

      I think this map does a better job of showing how cold February 2014 was for much of the country: http://icecap.us/images/uploads/ncep_cfsr_noram_t2m_djf_anom_%2822%29.png

      1. Wswplz says:

        Yes I like that for true representation of just how cold it was ( conus – 1.572 ) given the combined temperature anomaly we experienced. It would however be nice to see it broken down by region as well, or drilling further down by state. I guess it’s not a big deal since I can access that information fairly easy, but if they do create a map like the one at the top of thread they then should have an additional map showing the departure from average including that variance from the control base.( average ie )

        1. Wswplz says:

          This map here is a great one as far as lay out!!


          Nice find Cort.

  7. fixxxer says:

    Lets just focus on when we start to warmup.

  8. Wswplz says:

    Has anyone accessed the NWS daily forecast page for there area as of late? Mine is all messed up…

Leave a Reply