Temperature Trends

2012 Hottest Year on Record in U.S.—Surface and Satellite Temperature Measurements Agree


Earlier this month, I reported the satellite temperature trend data from  University of Alabama in Huntsville climatologists John Christy and Roy Spencer who reported for the 48 contiguous U.S. states that 2012 was the warmest year in the 34-year satellite record:

The annual average temperature over the conterminous 48 states in 2012 was 0.555 C (about 0.99 degrees F) warmer than seasonal norms.

Yesterday, the National Oceanic and Atmospheric Administration released their data for 2012 and also reported that 2012 was the warmest year on record for the lower 48 states:

2012 marked the warmest year on record for the contiguous United States with the year consisting of a record warm spring, second warmest summer, fourth warmest winter and a warmer-than-average autumn. The average temperature for 2012 was 55.3°F, 3.2°F above the 20th century average, and 1.0°F above 1998, the previous warmest year.

Most of the difference between the UAH data (+1 degree F) and NOAA data (+3.2 degree F) can be accounted for by differences in their baselines, 34 years versus the 20th century. All temperature records do show an upward trend in global average temperatures. See below:

Looking at the above five different global temperature datasets through 2010 from the Goddard Institute for Space Science (GISS), NOAA's National Climate Data Center (NCDC), University of East Anglia's Hadley Center (HadCrut), Remote Sensing Systems (RSS) and University of Alabama in Huntsville (UAH), one finds the following per decade trends:

GISS + 0.176 degree C

NCDC + 0.171 degree C

HadCrut + 0.169 degree C

RSS + 0.163 degree C

UAH + 0.141 degree C

The difference in the rate of increase between the highest and the lowest datasets is +.035 degrees per decade. Another way to cumulate the difference, assuming a linear rate of increase for the next 100 years, is that global average temperatures will have increased above the current average by between 1.76 to 1.41 degree C by 2113.

NEXT: Dozens Injured in NYC Commuter Ferry Crash

Editor's Note: We invite comments and request that they be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of Reason.com or Reason Foundation. We reserve the right to delete any comment for any reason at any time. Report abuses.

Please to post comments

92 responses to “2012 Hottest Year on Record in U.S.—Surface and Satellite Temperature Measurements Agree

  1. The difference in the rate of increase between the highest and the lowest datasets is +.035 degrees per decade.


    1. The reason why they are banging about the lower 48 states, is that they need some outlier to keep the panic going.

      You slice and dice a chaoticly equilibriated system properly and you will easily find some portion of it is at a max or min over any arbitrary time range.

      Couple this with the deliberate conflation of any warming at all with CAGW, and you’ve got most of the climactic pronouncements hitting the press.

      1. “+.035 degrees”


      2. t: On the other hand, the UAH dataset (like all of the others) finds that land areas in the northern hemisphere is consistently warming faster than average global temperatures, +0.23 degree C per decade. And the USA 48 is warming at a rate of +0.24 degree C per decade. Another simple linear projection suggests that average temperatures would increase in the lower 48 by +2.4 degrees C by 2113 (about 4.3 degrees F). Using UAH data that would suggest that the USA48 would be around 5.3 degrees F hotter than the 34 year satellite average. That’s not nothing.

        1. Not surprising:

          1) Land areas have far more variable temperatures because oceans act as giant heat-reservoirs.

          2) n observations where n 1 will inevitably result in observations that are above average.

          Moreover, using linear calculations to make predictions about a system that has complex, poorly understood negative feedback systems is idiotic.

          1. “n observations where n 1 will inevitably result in observations that are above average.” should read “n observations where n is greater than 1 will inevitably result in observations that are above average.

          2. t: I supply “idiotic” linear trends for illustrative purposes only; they are more conservative than climate model projections.

            Temperatures are certainly “more variable” on continental land masses than ocean temperatures, but the data show that their average temperatures aren’t just more variable; the data show that northern hemisphere land ares temperatures are going up faster than those over the oceans. Which again, I though I hesitate to note it, the climate models predict will occur.

            1. Ron, the Climate Models predict increased water vapor in the atmosphere amplifying the warming caused directly by increased CO2. The observations don’t match that prediction. They predict warming in the troposphere. The dreaded tropospheric hot-spot has yet to manifest itself.

              The regional forecasts are wrong – and have serious consequences, eg the flooding in Australia because dams were operated at high water levels based on forecasts of a drought that never came.

              IIRC there are 18 feedback systems identified by scientists who study the climate, of which most are rated by the IPCC as being poorly understood.

              Honestly, I am pretty fed up with the whole charade. The climate models are wrong. Period. They are worthless as predictors of future climate. Period.

        2. How much of that dataset has been influenced by the relocation or abandonment of data stations, Ron? It’s pretty well established that many data collection sites have been moved to areas where their data sets aren’t consistent with the stations they’re being compared to in the past.

          1. Sattelite data doesn’t depend on station location. You guys are going to have to come up with new rationalizations.

            1. Right, it just depends on how the program takes the data collected and processes it…and that’s done by the people whose methodology is held as secret as a Freemason’s handshake.

              Satellite data is 100% massaged.

              1. Huh? The methodology is right there in their publications. The RSS group even published a paper in 1998 outlining some problems with UAH’s methodology which biased it toward cooler temperatures (they didn’t take into account the decay of the satellite’s orbit, which lowered the area it was measuring irradiance from and thus the temperature estimate was lower).

                It is extremely complex, yes. The fact that you and others do not bother to find out what the methodology is doesn’t mean it’s held secret.

    2. I’m not a global warming alarmist, but I think you’re misunderstanding the quote. 0.035 is the difference between the measurements, implying that they agree reasonably well. You don’t have one group predicting 1 deg/decade and another predicting 5 deg/decade.

      Now, if you want to laugh at the actual increases of ~0.16 deg/decade then be my guest.

      1. Do you honestly think they can measure global mean temp to 3 sig figs? It’s hard to measure a room’s mean temp with that accuracy.

        Maybe those numbers were cooked up by Bailey and are not part of anything “official”. Either way, they are laughable.

        The table should look more like this if the author wants to inspire my confidence:

        GISS + 0.2 degree C

        NCDC + 0.2 degree C

        HadCrut + 0.2 degree C

        RSS + 0.2 degree C

        UAH + 0.1 degree C

        1. The importance of this can’t be emphasized enough. One of the #1 warning signs of a shady scientist is the use of phony levels of precision.

          1. When I took Physics (III?) in college, the professor would mark anyone down who went beyond sig. figures in their answer.

            It kept you on your toes.

            1. Did the professor mark people down for claiming that the mean of a data set had the same variance as individual elements of the data set?

              1. My unit ops professor would have blasted these guys, too. Round to the measurable digit, even in deviation. If your standard deviation is less than half your smallest measurable order of magnitude, state that in a note on the chart.

                1. Significant figures are good rules of thumb for calculations using unrelated measurements, and provide a conservative estimate of the precision of the computation. But that’s all they are, rules of thumb. In the case of the mean of independent observations, the sigfig technique is going to vastly overestimate the uncertainty.

          2. The precision issue brings up another interesting question:

            Unquestionably, if you take N observations of some thing, you can use statistical techniques to arrive at conclusions that are very precise.

            For example, in 6th grade science class, we stuck a thermometer in a beaker filled with water that was sitting on a hot plate and plotted the temperature as a function of time, measured the slope to calculate how rapidly it was warming, etc. In that case, our thermometers had an error band of 2 degrees, but we were getting precisions of .2 because we took 15 observations or so.

            However, that would *not* be the case if I had taken 15 beakers on 15 different hot plates and took three temperature measurements and tried to produce a rate at which hotplates heat a beaker’s worth of water, especially if the thickness of the beaker walls, their geometry and the wattage of the hot-plates was variable.

            Simply munging the numbers together and taking a standard deviation would be inappropriate in that case since the observations would not be measuring the same phenomena but rather a number of related but different phenomena.

            1. However, that would *not* be the case if I had taken 15 beakers on 15 different hot plates and took three temperature measurements and tried to produce a rate at which hotplates heat a beaker’s worth of water, especially if the thickness of the beaker walls, their geometry and the wattage of the hot-plates was variable.

              Depends on what you’re trying to measure. If you were measuring the average temperature change of the beaker-hot plate combinations, then that would be a perfectly valid technique. This wouldn’t be a good way to measure the specific heat of water (which is what I assume you were trying to do) but it would produce a precise estimate of some variable.

              Now, if you replaced the water in all the beakers with antifreeze and did the same experiment, that would produce a meaningful comparison of the specific heat of antifreeze vs. water. And that’s closer to what the temperature studies are doing.

        2. The mean of a bunch of measurements is more precise than any of the measurements themselves. Since these studies include a shitload of temperature measurements, they can indeed get that degree of precision for the means.

      2. He’s just being glib as usual, and Bailey is enabling him by using an unusual measure of the spreadedness of the measurements (difference between max and min). Real statisticians use standard deviation, which is 0.0137 in this case. Assuming the error is normally distributed, that means there’s a 97.5% chance that the true value of the increase is higher than 0.1367 degrees.

        1. Is there a justification for assuming that the error between these data sets is normally distributed? Their difference is likely due to systematic errors, not noise.

          If each one of the data sets is based a statistical analysis of many individual temperature measurements, then each set should have it’s own standard deviation. And if the mean value of any data set doesn’t fall within the standard deviation of every other data set then someone needs to do some equipment recalibration.

  2. When I’m 113, I’ll probably like it a bit warmer.

  3. Huh.

  4. IIRC, these data-sets are not independent of each other, but built upon raw temperature observations held in common. The stations each data set uses have significant overlaps, and the calculations used to arrive at the average are also similar, resulting that largely the same inputs into the same calculations are producing similar outputs.

    One of the telling bits of information that modern climatology is dominated by terrible scientists is the common contention that the process of feeding similar inputs into similar algorithms that produce similar results is somehow a validation of those algorithms or results.

    1. I wouldn’t say that climatology is dominated by terrible scientists. I’d say that the stakes are higher, and that some people are justifiably more critical of the claims.

      I’ve seen plenty of unproven predictions from modeling in my field, but it’s not really a big deal if one’s experimental UV spectrum only extends down to 80 nm instead of 40 nm. You use the experiment to fix the model and continue on.

      Testing is much more difficult in climate science. It’d be nice if the scientists were willing to admit that.

      1. That would be nice. It would also be nice if would-be policy makers were willing to admit that they have no idea how any level of warming would effect humans 50 years from now.

      2. If the field isn’t dominated by bad scientists, the dominant figures decided to stay quiet when confronted by Michael Mann’s incompetent statistical analyses why?

        Face it, climatology was founded by institutions that were concerned about man’s influence on climate, and as a result are doing exaclty what Feynman warned against on that Youtube clip I post every few weeks here, thinking that the theory is so beautiful that it must be right so the observations falsifying it may safely be ignored.

        1. You don’t get to deploy Richard Feynman in your ill-informed corporate-funded anti-science propaganda-fueled opinion on climate science. Mann was demonized but never discredited, which is far more than can be said for the climate change denial movement. No matter how often and convincingly that movement is shown to be a sham, it never gives up on its claims–which, of course, are absurd and outlandish.

          1. I have to say, this is the most beautifully pure form of trollery I have seen in many a week.

            I must admit that my mouth was open in admiration at the blistering illogic of the first sentence alone.

            1. Mine remains pursed in annoyance at the stunning persistence of people’s ability to believe in a massive global conspiracy of all the world’s scientific bodies and governments to invent a crisis for vague and ill-defined purposes instead of just trusting what current science says like you do with respect to all other scientific subjects.

              1. Other sciences don’t have $70 trillion worth of ill-conceived, thinly-veiled redistributionary policies on the line.

              2. Oh, their purposes aren’t vague and ill-defined. They’re after more money, pure and simple.

                And if you think there’s a consensus among “all the world’s scientific bodies and governments” (your words), then you’re fucking retarded.

              3. How many times does this have to be said to you before you get it – consensus is not science. It’s consensus.

                Fucking leftists love to prattle on about how they respect the scientific method (unless economics, drugs, guns or medicine is the subject), and almost none of them seem to have any fucking clue what the scientific method actually is.

                1. You’re right. Consensus is consensus. As in, there’s a nearly unanimous consensus of experts on this issue. So what do you have in support of your claims? Conspiracy theories on rightwing websites?

                  1. Unanimous consensus is still just consensus. It isn’t science. What do you not get about this, other than everything?

      3. This isn’t a climate projection from a computer simulation. It’s raw frigging observed data.

        1. It’s raw frigging observed US data.

          Is the claim that there has been NO SIGNIFICANT INCREASE in global temperature for 16 years, true or false?

          1. That widely spread factoid is a classic case of cherry picking. Why is the range 16 years? Precisely in order to show a period where it seems there has been no warming. But the relevant scale is multiple decades. Try the past 50 years. Try wrapping your mind around the fact that eight of the hottest years on record were in the last decade, and that every decade has been hotter than the last, or that the glaciers and polar ice caps are melting and causing the sea level to rise at this very moment, or try cracking open a single reliable source on this subject rather than hanging your hat on one minor, dishonest cherry-picked factoid.

            1. T: It’s not a “factoid,” it’s a fact. Average global temperatures have been plateaued. Even Hadley Centre climatologist Phil Jones acknowledges that fact. As I noted (see link) back in 2007, Hadley Centre climatologists flat out predicted:

              Most interestingly, and to its credit, the Hadley Centre has now gone out on a risky prediction limb. The Centre has combined its weather prediction model with a climate change model to make definite forecasts about the world’s climate for the next decade. To wit: “We are now using the system to predict changes out to 2014. By the end of this period, the global average temperature is expected to have risen by around 0.3 degrees Celsius compared to 2004, and half of the years after 2009 are predicted to be hotter than the current record hot year, 1998.” Since various temperature records?surface, satellite and weather balloons?have shown a temperature trend that increases at about 0.2 degrees per decade or less, this is a truly bold prediction.

              According to UAH folks, none of the years after 2009 have been hotter globally than 1998. See their data for 12 hottest years:

              1. 1998 0.419
              2. 2010 0.394
              3. 2005 0.260
              4. 2002 0.218
              5. 2009 0.218
              6. 2007 0.202
              7. 2003 0.187
              8. 2006 0.186
              9. 2012 0.161
              10. 2011 0.130
              11. 2004 0.108
              12. 2001 0.107

              The Hadley prediction looks shakier and shakier.

              1. The Hadley prediction being wrong does not mean global temperatures have plateaued.

                1. T(L): From UAH which I cited in earlier posts:

                  While the atmosphere has warmed over the full 34-year time span, it has not warmed noticeably since the major El Ni?o of 1997-98 ? giving us about a decade and a half of generally stable temperatures.

                  Since 2002, there has been a plateau of relatively warmer temperatures with only 12 months when the global average temperature was cooler than the long-term seasonal norm. In fact, compared to the 30-year temperature baseline, the most recent five years (12/07-11/12) averaged only 0.003 C (0.173 to 0.176 above seasonal norms) warmer than the preceding five years (12/02-11/07).

                  What would you call a “plateau”?

            2. “Precisely in order to show a period where it seems there has been no warming.”

              In this case, the “cherry-picking” is appropriate. The point is that there hasn’t been any warming for 16 years and that that is troublesome fact for the alarmists.

            3. I’m sorry but this is hardly cherry picking.

              For ANY of the climate models currently used to have any validity then climate should climb roughly im proportion with CO2 concentration. Since CO2 production has continued unabated and even at accelerating levels during those 16 years.

              Since 1997 CO2 concentrations in the atmosphere have risen by nearly 10% (363.7 PPM to 393.8 PPM) and ALL of the climate models used by the IPCC show that the climate would have to warm given that fact.

              That it has not shows that the models are false and overestimating the climate sensativity to CO2, therefore overestimating the levels of future warming we can expect.

              None of which is to deny that AGW is occurring, but it does very much cast doubt on the idea that it is even a pressing forget an existential threat to hummanity at large.

              Based on the best current data we have, global warming will be a nuisance over the next 150 years or so at which point it will abate because of technological changes (or our killing large numbers of ourselves off)

        2. No, it isn’t raw. HADCRUT and GISS sure as hell ain’t raw data.

          1. B: All data are noisy and thus need to be processed; even Anthony Watts processes his data. Of course, one can legitimately question the processing. May I direct your attention to my July column, “Everyone Freaks Out About Two New Climate Studies” that looks at the Berkeley Earth Surface Temperature report and Anthony Watts’ critique of NOAA data.

            As I noted in an earlier comment, as far as I know Watts has not yet publicly addressed the flaws some critics found in his earlier analysis.

            1. In the comments on Judith Curry’s blog, he says TOB (Time of OBservation) issues are not turning out to be a problem.


        3. This isn’t a climate projection from a computer simulation. It’s raw frigging observed data.

          What, did they have thermometers hanging from the satellites to collect the “raw data”?

          1. No, they use proxies for temperature. Just like every thermometer in existence does.

            When you look at a mercury thermometer, for example, you’re not measuring temperature, you’re measuring volume.

            1. And how, pray tell, do they do that with a satellite? In your thermometer example, they’re using a defined mechanism where an actual event occurs to the mercury in a thermometer. Based on the change in volume and the amount of space it displaces, a reliable measure of temp can be determined.

              With satellites, which are not in the atmosphere they’re measuring, all the data collected is massaged, since the data is not being directly collected, but rather is “observed”. And that “observation” occurs within the parameters designed by the program, which means it’s not really raw data.

              1. For atmospheric temperatures, the satellites measure radiance through certain wavelengths, which reflect an actual event happening in that level of the atmosphere. It does take some work to convert this to temperature.

                For surface temperatures, they just use IR measurements, though this can only be done when there are no clouds above the surface.

                1. For surface temperatures, they just use IR measurements, though this can only be done when there are no clouds above the surface.

                  Which is what % of the time?

                  Sorry, but until they come up with one that measures the exact same temp of my back yard when compared to a mercury thermometer every day for 10 years, I’ll still put their margin of error for the entire world at a level too high to be considered reliable.

                  1. The margin of error for the entire world’s AVERAGE temperature is LOWER than the margin of error for your backyard temperature.

                    1. That’s fucking retarded, as they take temp readings for specific locations to determine the average. If they do so for 10k locations, and there is a margin of error for each location of 5%, which is reasonable to assume for a backyard location, do you realize how unlikely it is that they have an accurate global average?

                      Jesus Christ, this is basic methodology stuff here.

                    2. Actually no.

                      In many cases taking a large number of inaccurate measurements can very much lead to a much more accurate measurement than any individual one of those measurements.

                      The key is for there to be no systematic flaw skewing the results in one direction or another, often there is and it can be very difficult to determine and filter out such a systematic error when it occurs, but it does not invalidate the methood or the fact that it can produce very accurate measurements.

    2. You don’t have proof for this contention, and it’s ridiculous in view of the satellite data which doesn’t depend on station location.

      If I didn’t know better, I’d think you were ideologically opposed to global warming being true and looking for a pseudoscientific rationalization to dispute the data with.

      1. “I’d think you were ideologically opposed to global warming being true and looking for a pseudoscientific rationalization to dispute the data with.”

        Find anyone here who thinks there has been no global warming. I don’t think so.

        The issue being addressed with the plateau of warming is that it is counter to the model predictions.

        According to the models, it was supposed to be warming, measurably, over the last 16 years. It hasn’t, therefore the models aren’t predictive.

    3. I should amend this statement: the UAH is based entirely on satellite observations and is independent of the other datasets.

      1. That’s more of a refutation than an amendment.

    4. t: Actually, you’re right that there is a considerable overlap between various the surface datasets. But the two satellite datasets are independent of those surface thermometer data, although they process the satellite signals in slightly different ways.

      For more on temperature data battles, see my column, “” dealing with the Anthony Watts and Berkeley Earth Surface Temperature (BEST) temperature data.

      2 BTWs: (1) As far as I know Anthony Watts has not yet reprocessed his data to account for flaws identified by critics, and (2) the BEST global trend over the past 50 years is +0.174 degrees C per decade, which is good agreement with the other surface data trends.

      1. Grrr: Here’s the link to my column on temperature data wars,“Everyone Freaks Out About Two Climate Change Studies”.

  5. The reason why they are banging about the lower 48 states, is that they need some outlier to keep the panic going.

    Is that an outlier in your terducken ?

    No, I always feed my turkeys black swans.

  6. I seem to recall that when certain parts of the globe had particular cold years a few years back, and anti-AGW folks latched onto it, we were told that AGW is a global phenomenon and can’t be measured by simply looking at particular areas. Has that scientific view evolved?

    1. Ahh, yes, here we go.


    2. I hope Bailey is just posting this for shits and giggles. Admittedly, most of his audience is American, so the report is somewhat relevant to us. I don’t think he was trying to draw any AGW related conclusions from this data. In fact he has a history of posting this type of information without comment.

  7. It has to be warmer. Look at all that red on the map! Everyone knows that red means hot.

  8. assuming a linear rate of increase for the next 100 years


  9. Now we know why big Al sold Current to Al Jazeera. He’s going to buy Reason as his new propaganda machine.

  10. speaking locally, it was hot here in Michigan last summer but never truly HOT. It was worse in the 1990s/Aughts when we would routinely hit the 100 degree mark. 100 degrees + humidity = hell.

    The greater concern last year was the drought which ravaged the fruit growers.

  11. So what?

    Do we have AGW or are we concerned with AUSW? 16 years with no significant warming GLOBALLY. Why the focus on the US? What are the global numbers? Why would anyone even report on this unless to take the focus away from the global numbers that DON’T fit the talking points?

    Come on Bailey, what’s the full story?

    1. FdA: You could click on the link included above for my post reporting on global trends from last week. See also, my December post, “34 Years of Satellite Temperature Data Show Global Warming Is on a Plateau.”

      1. So why no mention of it here? You know damn well that anything pointing to higher temps anywhere will be used as evidence that the end is nigh.

  12. 2012 Hottest Year on Record in U.S. – Surface and Satellite Temperature Measurements Agree

    Well, it was an election year…


    1. Warmer, cooler, no change… Doesn’t matter…

      It’s AGW all the way down.

  13. When there is a warm summer alarm bells go off, to signify the validation of anthropogenic climate change. When there is a cool summer or winter, the alarm bell ringers caution us that “climate is not weather”. That’s an incredibly convenient way to examine data, no wonder there is such a ‘consensus’ on the issue.

    1. That’s not what this analysis is doing, though.

  14. Temperatures? Ocean acidification is all the craze now.

  15. The Russians could sure use some of our warmth. I wonder if we could export it…

  16. LOL…The RSS AMSU data shows that 2012 was the 11th warmest year GLOBALLY out of a 34-year-long record and very close to the median. That means that there is nothing much to fear. So what does GISS do? After telling us that the US is too small to matter in the distant past (2010), it uses the American data, which has been ‘adjusted’ by its gatekeepers to claim that 2012 is the warmest year. But if you look at the old data sets you still find that the US was warmer in the 1930s and if you look at the measurements from the same stations there is nothing to fear again.

    Why is it that Bailey is so naive for a ‘science’ writer? Can’t he look to the data and do a bit of reading? Perhaps an interview with Steve McIntyre might be in order.

    1. VIV: Please see links to my various columns and posts above. BTW, as I reported earlier this month, the UAH satellite data say that globally 2012 was the 9th warmest year.

  17. Whatever. All I know is that I rode my motorcycle in Jan and Feb because it was so warm and there was so little snow.


    MOAR warming, please.

    s/ Inhabitant of the Midwest USA

  18. If an average persists at a particular level, then when one area is higher, some other area must be lower to maintain the average.

  19. Thats pretty hot when you think about it man!


  20. Am I right in suspecting that Alaska and Hawaii were left off because their temperatures ruin the headline? That is the kind of thing our vaunted climate ‘scientists’ and their lackeys in the media are known for, after all.

Comments are closed.