Stream: Hottest Yeah Evah

Stream: The Hottest Yeah Evah! Really? Or a yet another example of activism masquerading as science?

Assume for a moment, as the press with triumphant glee is reporting, that 2016 was the hottest year evah! Believe the claim for the sake of argument. Swallow the idea, for at least the next minute, the media and government really do have your best interests at heart and are reporting the truth, the whole truth, and nothing but the truth about the world’s temperature.

How much hotter than previous years was 2016? Bare your wrist and blow a huh on it from about half a foot away. Don’t blow—stay with me here: this is a genuine scientific experiment—but utter a soft ugh so that your breath wafts over your wrist gently. Feel that increase in heat? Well, that boost to your skin was much hotter than the increase supposed to have happened to the atmosphere in 2016.

Here’s a better experiment. You are likely reading this article sitting down. Sense the temperature around your face: it might help to think about your cheeks. Now stand up. Take a second mental reading. Feel the difference? That same tenth or a so change in degree, which was probably imperceptible to you, is about the same as the change in temperature scientists say they measured over the entire globe, including over the salty seas from last year to this.

Yes. Climatologists gathered measurements from buoys at sea, from thousands of thermometers at airports and other locations, from balloons, even, and then took their average—sort of. That number was then declared as the Official Temperature of Earth for 2016.

The “sort of” is important. Because the places and methods of measurement used in 2016 were not exactly the same as those used in 2015; and those used in 2015 were not the same as those used in 2014; and so on. And those used in, for instance, 1914 are completely different than in 2014. A century ago, mercury-in-glass thermometers were in a different class than the digital complexities in use today. Too, 100 years ago the places of measurement were few in number. Vast areas of the globe went unmeasured. And at places which were the same, well, thermometers out in the woods in 1914 now have a cities grown up around them. Even in modern times, thermometers break and are serviced. Buoys corrode. And so on. Things change….

[Don’t miss the exciting conclusion!]

Go there to read the rest.

Update That point is that NASA’s (or NOAA’s) way of calculating uncertainty about the 0.07 degree increase is wrong, as is detailed in the links at Stream (which come back to here to the technical articles about parametric versus predictive uncertainty, remembering all probability is conditional, and so forth). The hubris of thinking we can measure the surface temperature of the earth to the hundredth degree!

Update See Bob’s comment below. Main story. My take.

Update Below it was suggested that if I had a better method to account for the uncertainty in temperature reconstructions that I should publish it, that climatologists would love it and me, and that the world would be a better place. Well, I have published it. See Chapter 10 in Uncertainty: The Soul of Modeling, Probability & Statistics. See also the links at Stream which point back to my pieces on the BEST reconstructions, etc., which (at the time) were noticed but then—somehow—forgotten. I will now sit back and await the accolades.

Categories: Statistics

26 replies »

  1. What does NASA have to say about all this? Check out

    were you can learn that

    “Earth’s 2016 surface temperatures were the warmest since modern recordkeeping began in 1880, according to independent analyses by NASA and the National Oceanic and Atmospheric Administration (NOAA).

    Globally-averaged temperatures in 2016 were 1.78 degrees Fahrenheit (0.99 degrees Celsius) warmer than the mid-20th century mean. This makes 2016 the third year in a row to set a new record for global average surface temperatures.

    The 2016 temperatures continue a long-term warming trend, according to analyses by scientists at NASA’s Goddard Institute for Space Studies (GISS) in New York. NOAA scientists concur with the finding that 2016 was the warmest year on record based on separate, independent analyses of the data.

    Because weather station locations and measurement practices change over time, there are uncertainties in the interpretation of specific year-to-year global mean temperature differences. However, even taking this into account, NASA estimates 2016 was the warmest year with greater than 95 percent certainty.”

    Yes, rather than vague handwaving about “uncertainty”, they’ve done the actual calculations. Think they have it wrong? Do some work yourself:

    “The full 2016 surface temperature data set and the complete methodology used to make the temperature calculation are available at:

  2. Does NASA/GISS believe that attaching their imprimatur to a press release implying that we have had a planet wide data collection system in place since 1880 that allows the ‘Annual Temperature of the Earth’ for 1880 and all subsequent years to be calculated with sufficient precision and accuracy to allow the annual temperatures to be placed in rank order so that anomalies of 0.07 degrees are scientifically and statistically justified will actually make us believe it?

    Didn’t work for me; YMMV.

    And that doesn’t even take into account the fact that the data on which the claim is made has been ‘adjusted’, repeatedly, over the years. Or that comparing ‘old’ adjusted data with ‘new’ adjusted data shows that the more recent the adjustment, the colder the past. Or so I am told by people posting old/new data, superimposed, as evidence.

    Bob Ludwick

  3. PLOT THE DATA for yourself. Download it, import it, create charts with it. This is for all the folks who think I am a nut. I am. I did download data from BerkeleyEarth, import it and plot it. I plotted it as a whole. I plotted it by station. I made a nice little interface to let me look at stations on the map and click on it so I could see the chart of the data for that station.

    Processing the data from Berkeley Earth is part of the learning process. As you click though the difference reporting stations, you find huge gaps in reporting. Some stations are completely updated. Some barely have a few years of data.

    Click through all of the stations and you get wide varieties of almost flat lines as defined by the sinusoid that is seasonal temperatures.

    Be sure to wander though the eastern part of Russia. The temperatures there are mind numbing. There is enough empty space in eastern Russia to house all of us, but I don’t see it getting out of the deepest of freezes any time soon.

    Plotting the data helped me. It didn’t help me believe in the consensus. It only helped me become more steely eyed. Anyone who thinks they can make me change my beliefs on climate change better be able to tell me what temperature, enthalpy, psychometrics, and humidity are. They better be able to tell me what an average is and how many different ways they can do an average themselves. I will joyfully joke through the difference between the simple averages (mode, mean, and median) and why they are useful tools to have on hand and how they can bend you over a table and turn your day into a nightmare.

    The more expert the person is, the more patience they are going to have to have.

    At heart, I might just be a snotty teen telling off my elders, but I expect my elders to demonstrate that they are my elders and more experienced and better at doing the job. =

  4. Lee: Too bad 95% probability is just a term with no mathematical or objective definition whatsoever. They can be 95% positive about anything with no justification other than “I said so”. Quite scary that people just dive in and believe.
    As for temperature adjustments, there appear to be as many sets thereof as there are of people hating on Trump. You can pick and choose till you get the numbers you love and make you feel all warm and fuzzy.

    As for averaging, this is how it was explained to me:
    Anomalies are taken BEFORE averaging.
    Start with any 30 year period and average EACH station to obtain a “normal”.
    For each station, subtract it’s normal from the recorded temperature. That is the anomaly.
    Average the anomalies using an area-weighted formula.

    No GAT is EVER calculated and does not exist. The base period is irrelevent to the calculations. Anomalies are taken BEFORE averaging.

    Would anyone care to take a stab at explaining or rebutting this? My impression was averaging the anomaly was based on the GAT. It appears this is not the case. There is no GAT ever. That would explain why there are virtually no estimates thereof, short of one in 1997 on NASA (62.45° F) and in the AR4 (58.1° F). Beyond that, I couldn’t find any numbers. Now it seems there is not GAT, which would explain why I can’t find the values but not why people keep referring to it……How to spin yourself silly with statistics and models…..

  5. I looked at that article. Unfortunately for you, it seems to make it clear that the adjustments exaggerate ‘global’ warming’ {which is nonsense in the first place]. In addition, there is no indication in that article’s plots of a proper set of error bands nor a proper error analysis and propagation.

    What ‘warming’ there is derives from the globe’s cities warming, in part. The cities are not the globe. Temperatures are things attached to a defined sample of matter. For gases not in a bottle, this is tricky as all get out. Do not elide over the kinetic theory of gases. Also do not elide over the fact that radiant ‘color’ or ‘brightness’ temperatures are not the same thing as the thermodynamic temperature. They may coincide, to a greater or lesser degree. Also, old Sol is a variable star.

  6. Quick brief: watching post-inauguration ceremonies on Fox News (where else?) and Brit Hume just announced that the “Global Warming” section on the White House web site has just been removed. Portent of things to come?

  7. Briggs, I saw that link elsewhere. And it is indeed hypocritical of left/liberals to bemoan a supposed lack of civility by Trump supporters in the transfer of power. Also, I’m not unhappy (which is to say, I’m quite pleased) with the deletion of the “Global Warming” bit from the White House web site

  8. Briggs, I didn’t read the comments below the video. I had read that this was what people were serenading George W. Bush in 2009 when he left on the heliocopter. Applied to Obama’s departure–Hooray!

  9. If anyone could actually show (rather than merely assert) that “NASA’s (or NOAA’s) way of calculating uncertainty” is incorrect, he would have made a significant contribution to climate science, worthy of publication in a journal that scientists actually read. I would love to find out that, despite basic physics and overwhelming data, there were real reasons to doubt the conclusions of >99% of climate scientists.

  10. @ Lee Phillips

    Hi Lee,

    Followed your link at 12:05 PM and found it very interesting.

    So can I assume that after reading the history of the temperature collection system in place in 1880, how the system has evolved since 1880, the procedures and record keeping used to collect and archive the data over subsequent 136 years, and how Berkley Earth and others have kriged, infilled, adjusted, corrected, and averaged the data you are convinced that the resulting table of annual temperatures of the earth is precise and accurate enough to allow the 136 annual temperatures to be placed in rank order and to provide scientific and statistical justification for the announcement that an anomaly of 0.07 degrees higher than the previously calculated peak anomaly constitutes a record?

    I am willing to bet a significant portion of my pitiful fortune that two INDEPENDENT teams of climate experts could not deploy data collection systems in my county, collect data for a year, process the data, and have the resulting two annual temperatures of my county agree within a few hundredths of a degree. I would be especially shocked if they were able to use the temperature sensors available in 1880 and achieve results that matched with two decimal point precision.

    But I am not a climate scientist, so my skepticism may be completely unjustified.

    Maybe our host can weigh in as to whether a table placing the annual temperatures of the earth since 1880 in rank order, with two decimal point resolution, can be justified scientifically and statistically?

  11. Oh, Lee, don’t hold back. We’re not snowflakes.

    Exaggerating percentages is a dead giveaway you aren’t using evidence, you’re making things up. Only people dead set on being right use a statistic of >99%. That’s even above the faked 97% consensus.

    Again, religious beliefs are NOT affected by reality and you have a religious belief going. It’s faith based. There are many explanations of why the stats are improper, etc, but none of the true believers ever cares. They BELIEVE.

  12. I am asking again—how is the average used in these statistics figured and anomalies calculated? Does anyone know?

  13. Sheri Nobody knows, that have taken the original data processed and reprocessed again and again. Their record keeping is so bad they don’t know if what they called the original data was that or someone else adjusted or projected data. When I say projected data as I think you well know there are two types projected data. First, nearly half of the “data” where we actually measure temperature is infilled because somewhere in the process they lost the original measure of simply failed to take the measurement. Temperature reading taken in one of their grid boxes (which are 1500 miles across and 1500 miles wide,) is supposed to represent the temperatures in the whole box and neighboring station in that box are suppose to be reflective of each other, they may be in different climate zone and altitude, lastly they also make such projection too the most part without a consideration of topography, The second projected data comes from the grid boxes that have no reading since we do have a device there to measure said temperature in those boxes, So they project the temperature from adjacent boxes and then they have the gall, to call those manufacture temperature “data”. That created “data” covers over 80% of the earth. Yet the useful idiots believe the adult bovine fetal matter the so called “climate scientist” put out, funny one such useful it dares to comment even on this blog.

    He then wonders why those of us if why we would have a problem with such a bucket of spit manufactured numbers, that not relevant to anything. I say not relevant to anything since not only do they use manufactured data. They left out humidity in their calculations and to know how much energy a given air mass possesses one needs to have humidity in the calculation. That humidity data is something the so called “climate scientist” leave out of their calculations. And yet the useful idiots call us the anti-science people.

    I could live with the “climate scientist” making the claim 2016 was the warmest year ever, if they just simply would be honest about and say according to our calculation we believe it was the warmest year ever, but there is a good chance our measurement method and data collection may have miss lead us. Oh by the way here are the raw unaltered numbers, do with them as you may. It will be a cold day in hell when that happens.

  14. Ya’ know, Briggs, Trump has yet to fill 96% of his appointments, let alone all the agency jobs, and they’re having a hard time finding people, and among the administration’s first acts was to delete references to climate change from the WH website. I’d bet they’d love to have you on board. Just sayin’…


  15. Correction to experiment:
    The change in altitude of your head when standing up against sitting down will be, say, 3 ft. Given the atmosphere dry adiabatic lapse rate of 3oC/1000 ft, that would be about 9 mili-oC, not 70 mili-oC (i.e. about 8 times less than the 0.07oC stated).

    Yes, it probably isn’t the “hottest year ever” and it wouldn’t matter even if it were, but you can’t correct error with error 😉

    Best Wishes,

  16. Mark: Thank you for responding. I did some more research and NASA does indeed not calculate a GAT—they use the method I described. Basically, averaging the average of the averages. There are two other methods that I have yet to research. The use of the the term “global average temperature” is apparently indeed smoke and mirrors in the case of NASA. I realize a huge proportion of the data is fabricated. (Sorry, folks, extrapolated and interpolated are still fabricated even if math and science are claimed to be involved. There is NO measurement involved.) One of my contentions is that we have far too little data to address the issue and we must throw out the theory until actual MEASURED data, consistent over time, is available. That’s SCIENCE.

    “Since there is no universally accepted definition for Earth’s average temperature, several different groups around the world use slightly different methods for tracking the global average over time, including:” This is from UCAR. It’s something you never see in the MSM. One of the methods is from NASA, one from NOAA and one from the UK Met. There is no actual GAT in NASA and possibly not the and there are several methods that yield DIFFERENT results, by more than the quoted difference itself. As Brad Tittle said, using the data yourself and finding out what really is there is the only way to know just how bad this so-called data is. So much is routinely hidden or left out in climate science, the results are little more than Madame Greia psychic crystal ball predictions. That this passes for science is terrifying.

    In college when science was science, if you lost the data or were caught “adjusting” the data so your theory was true, you FAILED. You flunked, you did not pass. Now it seems you get a PhD and a cushy office. Science has become the lie.

  17. Sheri,
    This used to be the definition of the scientific method.
    A method of procedure that has characterized natural science since the 17th century, consisting in systematic observation, measurement, and experiment, and the formulation, testing, and modification of hypotheses.
    The climate scientists don’t modify their hypotheses to fit the data, they modify the data to fit the hypotheses.

  18. “In college when science was science, if you lost the data or were caught “adjusting” the data so your theory was true, you FAILED.”

    In my (undergraduate) labs, we had overworked TA’s supervising and grading. Often the equipment was worse for wear, and our assigned lab partners less than ideal. We quickly learned that the TA’s were pretty useless – they spent the lab period working on their own classwork or thesis, and their grading was pass/fail, depending on if you got the right answer. Back in our rooms writing up the lab, the temptation proved too much for most of my fellow students, and their measured data improved dramatically with just a little aging. But I had principles, you see, and I could never bring myself to fudge the measured data. Instead, somewhere around a half-page before the final answer was presented, I would insert an additional step that I clearly labeled something along the lines of “multiplying by an appropriate fudge factor”, and correct the measurement error that way.

    In hindsight, I guess having TA’s running the labs was good training for climate science. Engineering classes tended to be different, though, as in “show your work”. We all had at least one professor relate the classic story that ends with “You build bridge – bridge fall down – no partial credit.”

  19. “Below it was suggested that if I had a better method to account for the uncertainty in temperature reconstructions that I should publish it, that climatologists would love it and me, and that the world would be a better place. Well, I have published it. ”

    Not quite. The suggestion was that if you got it right, you would have something “worthy of publication in a journal that scientists actually read”. Perhaps you haven’t gotten it right yet.

  20. While qualifying to teach naval folks how to run a nuclear reactor, one of our advisors sat us down and asked a question..

    “What does it mean when the temperature gauge shows 335F and it changed to 334F”…

    We all danced around this. The stated accuracy of the temperature for that instrument was +- 5F. What does it mean when the temperature changes?

    When we got to the end of the discussion, the answer was simple. If it changed a degree, it changed a degree. Whether the temperature was 335 or 330 didn’t really matter. What mattered was whether the plant had cooled a degree or warmed a degree. 1 degree (IIRC) was basically an inch in the pressurizer. DO NOT LET THE PRESSURIZER EMPTY or GET FULL. These are two endpoints that were very very important. Mess that up and we all get to hang around for a month filling out paperwork and figuring out how to find our next job.

    This sort of leads into the idea of “anomaly”. If the average changes by a degree then the change is likely a degree. Attempting to back calibrate all the measuring devices is problematic, so we make an assumption about the leaning on the average. The accounting nightmare that comes from that though is chilling. You have to pick something as your average point, but if you pick an average point, you are biasing your analysis. Is the bias a lot? maybe maybe not. Picking the range induces the bias. You accept that you have and you tatoo it to your forehead to make sure everyone grasps the bias you have caused. (and you make sure that everyone understand. THERE IS BIAS HERE!)

    Part of my request to PLOT the data is to get people to go get the source files and download the $()#@)@#$*@ data. OPEN the data set up and read the files. If you do this with the BEST data, you know that they are not anomalies, they are just in celsius/centigrade. 0 means “water is freezing” or maybe “water is thawing”.

    Kriging was an awesome thing to optimize the revenue for the state. It is a load of horse dung when applied to temperature data. There is a difference between applying distribution functions to static data like ore content vs non static data like weather. There is great uncertainty in the former, but it is manageable, because the goal is to get more money for the state from mining claims. I suppose the same goal is also achievable for weather data, but it is coming at it from the opposite direction…

    There is more to learn from plotting the data for yourself. Just remember what 0 means. I have this strong belief that in addition to CPR refreshers every year, all folks should also be required to revisit the definition of 1 and 0. Both number are abused daily. People shift definitions mid sentence and don’t know it.

Leave a Reply

Your email address will not be published. Required fields are marked *