Global Temperature Report: January 2019 – up, due in part to Australia

The global average bulk-layer atmospheric temperature anomaly rose by +0.12 °C (0.22 °F) in January to +0.37°C

Global climate trend since Dec. 1 1978: +0.13 C per decade

January Temperatures (preliminary)

Global composite temp.: +0.37 C (+0.67 °F) above seasonal average
Northern Hemisphere.: +0.32 C (+0.58 °F) above seasonal average
Southern Hemisphere.: +0.42 C (+0.76 °F) above seasonal average
Tropics.: +0.37 C (+0.67 °F) above seasonal average

December Temperatures (final)

Global composite temp.: +0.25 C (+0.45 °F) above seasonal average
Northern Hemisphere.: +0.32 C (+0.58 °F) above seasonal average
Southern Hemisphere.: +0.19 C (+0.34 °F) above seasonal average
Tropics.: +0.32 C (+0.58 °F) above seasonal average

Notes on data released February 1, 2019 (v6.0)

The global average bulk-layer atmospheric temperature anomaly rose by +0.12 °C (0.22 °F) in January to +0.37°C (+0.67°F) led by a warming in the SH oceanic areas of 0.31°C (0.56 °F) over last month. Once again we mention that indications remain that a warm El Niño will become fully developed in the tropical Pacific Ocean. However, the magnitude of the indicators is still not strong enough to cross the threshold of the formal definition of El Niño. If it does occur, it is likely to be a modest event.

The month’s coldest seasonally-adjusted temperature departure from average was located over the Ionian Sea (-2.7 °C, -4.9 °F) and the warmest over the North Atlantic southeast of Greenland (+3.2 °C, +5.8 °F).

The monthly map for January 2019 shows the usual situation of alternating hot and cold regions in the subtropical and higher latitudes. This time, the cold regions are found in eastern Canada, Europe (from the Barents Sea southward to the Mediterranean Sea), India, western Pacific Ocean and broad regions around Antarctica northward. The warm spots are roughly in between these, landing in western North America, the North Atlantic Ocean, the Middle East, Eastern China, and several areas over the Pacific and South Atlantic oceans with a particularly significant hot spot over southeastern Australia (summertime) for the second month in a row.

Spoiler Alert (Repeated until accomplished – no estimate yet):

Well, the time is once again approaching when new changes are required for the currently operating satellites as their performance changes with age. NOAA-18 has been operating for 13 years and is now past its time frame for accurate diurnal adjustments based on initial drifting, meaning the adjustments are adding spurious warming to the time series. On the other hand, NOAA-19 has also drifted so far that it too is introducing an error, but given its direction of drift, these errors are of the opposite sign. The two satellites are almost compensating for each other, but not to our satisfaction. In addition, the current non-drifting satellite operated by the Europeans, MetOP-B, has not yet been adjusted or “neutralized” for it’s seasonal peculiarities related to the diurnal cycle. While these MetOP-B peculiarities do not affect the long-term global trend, they do introduce error within a particular year in specific locations over land. So, all in all, we anticipate generating new adjustments for NOAA-18 and NOAA-19 to account for their behavior of late and shall also modify MetOP-B to account for it’s unique seasonal cycle. This will be part of a coordinated plan to eventually merge NOAA’s new microwave sensor (ATMS) carried on Suomi NPP and the new NOAA series JPSS. We are hoping that NOAA-19 will be the last spacecraft for which drifting adjustments will be required as the newer satellites (MetOP, NPP, JPSS) have on-board propulsion to keep them in stable orbits. With so many new items to test and then incorporate, we are waiting until we are confident that these adjustments/additions are appropriately stable before moving to the next version. In the meantime, we shall continue to produce v6.0.

As part of an ongoing joint project between UAH, NOAA and NASA, Christy and Dr. Roy Spencer, an ESSC principal scientist, use data gathered by advanced microwave sounding units on NOAA, NASA and European satellites to get accurate temperature readings for almost all regions of the Earth. This includes remote desert, ocean and rain forest areas where reliable climate data are not otherwise available.

The satellite-based instruments measure the temperature of the atmosphere from the surface up to an altitude of about eight kilometers above sea level. Once the monthly temperature data are collected and processed, they are placed in a “public” computer file for immediate access by atmospheric scientists in the U.S. and abroad.

The complete version 6 lower troposphere dataset is available here:

Archived color maps of local temperature anomalies are available on-line at:

Neither Christy nor Spencer receives any research support or funding from oil, coal or industrial companies or organizations, or from any private or special interest groups. All of their climate research funding comes from federal and state grants or contracts.



53 thoughts on “Global Temperature Report: January 2019 – up, due in part to Australia

  1. ” While these MetOP-B peculiarities do not affect the long-term global trend, they do introduce error within a particular year in specific locations over land.”….

    I’m not following that…..if they introduce an error in temperature….within a year
    …how does that not effect the trend?

    • because the errors are only seasonal, not interannual. In other words it’s the seasonal cycle that is not well-characterized yet.

  2. The data sets are not updated yet. Maybe in a few more days.
    I do like having the data sets come in.
    So far TLT has done a whole lot of nothing since Jan, 2018. That trend (or non-trend, if you must) is continuing. I would have expected rather more cooling coming off the El Nino event of 2016, but it just has not happened. Temps just keep hanging up there, and with an unusually small Std. Dev. Most interesting.

  3. NOAA have finally update the the US-CRN temperature data, after the shutdown.
    This is a reliable, high quality temperature recording network in pristine locations, starting 2005.

    For the yearly average jan-dec, 2018 was cooler than 2006. 2012, 2015, 2016, 2017
    and about the same a 2005 and 2007.

      • I’ve done some reading up on calculating and propagating error in population sampling, which is essentially what’s being done with Earth temperature measurements, and I’ve come to an acceptance that the methodology is not wrong. I’ve visited the websites of many math and physics departments of universities, as well as individual blogs about using statistics with populations.

        What I’m still not sure of is if the results really mean anything. Take for example Dr. Cowtan’s trend computer: does a trend of 0.073 +/- 0.008 C / decade really have any significance at all to the biosphere? That’s 0.73 +/- 0.08 C per century, which I have a difficult time thinking to be dangerous. His significant digits are all over the place, too. There are three sigdigs (just learned that word today) in his trend and error, five, seven and even eight in others.

        I ran some statistics on the NOAA GHCRN data set, using the 991 stations marked as “GSN” (which also had WMO IDs), and calculated the mean, sd, and error in the mean for the year 2014. There were 224,812 records, a mean of 17.8C, and an sd of 15.9. That’s a pretty big standard error; however when it’s divided by the square root of N, the error in the mean becomes 0.034C. This is where I become unsure.

        There are three sigdigs in 17.8; there are three in 0.034. However, I don’t think that stating the mean as 17.8 +/- 0.034 is correct; adding two more digits to the mean implies unwarranted accuracy. The reading I’ve done says that error should be expressed with no more sigdigs than exist in the measurements, which in this case would be three.

        All of the temp data I’ve seen has been in tenths of a degree. There is no way there are enough existing measurements N that could push the error in the mean to six significant digits. Simple algebra tells you that you’d have to divide 15.4 by 100000 to reach that point, and that requires a population N of ten billion.

        Dr. Cowtan cannot be using proper methods to express error in his calculations. The paper he uses for methods is the same — it calculates simple least squares out to three and four places. It’s this disregard for proper error tracking that creates a lot of my skepticism with the scientific community today. When they play fast and loose with the numbers in the edges, it just doesn’t look good.

        • “Dr. Cowtan cannot be using proper methods to express error in his calculations.”

          Is it not a bit too easy to pretend that without bringing an own scientific contradiction on the desk?

          Kevin Cowtan after all has linked the source from which he derives his calculations (the method section of a paper written by Grant Foster and Stefan Harmstorf).

          • It’s not that his methods are wrong, it’s that he carries too many significant digits into his results, implying a precision that doesn’t exist.

            I went to the GISTEMP website and downloaded the data he represents in his graphs. There are 139 records of mean global anomalies. I’m sure he didn’t generate his trend line manually, as I did so that I could see all of the steps, but the line I generated from the data fell right on top of the line I let Excel generate for the data, so I can presume that my math was correct. I did use Excel for the math steps.

            I found a nice webpage at
            that explains the steps in calculating a least squares regression, so I followed its steps.

            First, you collect your X,Y points. I used 0 for the start of X and the 1880 anomaly for my first Y. For each X,Y pair you calculate X*Y and X*X, and then sum all of the columns. The final result is a table of X, Y, X*Y, X*X, and at the bottom, sumX, sumY, sumXY and sumXX.

            The anomaly data is in decimal format with two significant digits. Each year has a monthly anomaly, averaged to get the annual mean anomaly. That has a built-in error right there. The standard deviation for the first year’s mean was 0.2 C. The error in the mean is the standard deviation / sqrt N, so it’s 0.2/sqrt 12 = 0.06 C. Obviously the calculator has a lot more numbers in the answer, but the monthly anomalies have only two significant digits, and so the mean and error shouldn’t have more. The column XY is a measurement times a constant, so there is no error. Column XX is multiplying a constant by itself, so there is no error. The summation columns’ error is the standard deviation for the values in those columns, and worked out to 0.02 for Y, and 0.06 for XY.

            The formula for the slope is Slope(b) = (N∑XY – (∑X)(∑Y)) / (N∑X2 – (∑X)2)
            and for the intercept it is Intercept(a) = (∑Y – b(∑X)) / N
            and you get your trend line by plotting that equation.

            Keeping track of the errors in these formulae, there are N (no error), sum(XY) (0.06), sum(X) (no error), sum(Y) (0.4), so for that result the errors intermingle as sqrt(0.06^2 + 0.4^2) = 0.4 C.

            The intercept equation has sum(Y)(0.4), b(0.4), sum(X)(none), and N(none). They all add up to sqrt(0.4^2 + 0.4^2) = 0.6.

            Run all that through the a + bx formula and you have sqrt(0.4^2 + 0.6^2) = 0.7 C.

            That’s a pretty fair-sized error, but I think I did all of the calculations correctly. I wish I was more knowledgeable about posting equations (I haven’t got LaTex to work for me) and images than I am. If anyone wants more clarification I’ll be glad to do so.

            The point I’m trying to make is that it’s not the methods that are necessarily wrong, but that the errors in the calculations are not being tended to properly, which results in a false sense of accuracy in the measurements. I don’t think the world is known to that degree.

  4. “Global Temperature Report: January 2019 – up, due in part to Australia …”.
    IPCC telepathists say the planet would be cooling without human GHG emissions:
    I just want to say how proud I feel, proud that we Australians are contributing to keeping global cold-ing at bay mainly through our fossil fuel exports though we could be doing more locally if layer upon layer of government would get out of the way.

  5. The high Australian temperatures are important as they drive the ‘in-filling’ of temperatures for the huge surrounding oceans which have no thermometers. This multiplies the Australian impact on the reported global temperatures.

    This of course is one of the reasons the CAGW crowd put so much effort into keeping Australia in their camp.

    • These measurements are from the satellites which have no infilling. They look at the whole globe (minus a wee bit at the top and bottom)

      • Thanks for the clarification, you are correct for the satellite data. I should have been more specific. All the other commonly quoted surface temperature data-sets, which are the ones that show the most warming, do use infilling and so are heavily influenced by the Australian data.

    • Which is also better than raw instrument siting data which is very heavily East Coast dominated because of more sites as it has more population and more towns.

      • And most instruments are situated at airports, airfields and aerodromes. In a recent article 15 of the hottest places on Earth were in Australia during our summer. 10 were either at an airport, airfield or aerodrome.

        I am reliably “told” this is BoM best practice.

  6. As long as TLT anomalies keep bouncing above ~0.25C, showing no persistence below that level, the warm phase of multi-decadal and longer cycles continues unabated.

    • rbabcock

      Do you think Roy Spencer & alii should round their data up to integer multiples of a degree extra for you?

      I think this data is used by many people who would accumulate errors when basing their own work on rounded data.

  7. I thought that there were no seasons in the tropics. What does it mean – a seasonal average?

    • Go down to the islands of the West Indies during the rainy season.
      A) You will find out all about the seasons there.
      B) You will never need a reservation at any restaurant.
      C) You will discover all the indoor activities the island has to offer.

      Been There, Done That. Once was enough.

      • OK, winter season in the North, summer season in the South, dry season in the West Indies, monsoon season in Townsville, QLD.

        • It’s moving south. Mackay getting drenched right now. At least it’s moving instead of sitting, like it did over Townsville.

      • What do you mean for a couple of days in February, it was cold for practically the whole of January.
        But the reason is that the Satellites are not measuring the Surface Temp, they are measuring the lower atmosphere, which is the warmth leaving the planet to space.

  8. “The global average bulk-layer atmospheric temperature anomaly rose by +0.12 °C (0.22 °F) in January to +0.37°C … Global climate trend since Dec. 1 1978: +0.13 C per decade”
    Lets not tell Javier that he has just proved that global warming stopped in 2017 and has been
    decelerating since 1994.

    • Poor, poor Percy. Your math skills are so poor that you don’t even realize that you can have a one month increase in temperature without disputing Javier’s point.

      Or is it that you have no interest in reality, you’ve been paid to push an agenda and you will push it, no matter how badly you have to embarrass yourself.

      • Mark,
        Look at the global 40 year trend noted above as 0.13 C per decade. Or even the graph. There is no sign that global warming is over. There is a decline over the last couple of years from the 2016 EL Niño peak but that still means that 2018 was the 4th warmest on record. All temperature records
        show that the world is warming.

        • If you look at even longer trends, it’s obvious there never was any global warming to begin with.
          100 year, 1000 year, 10,000 year.

      • “…. no matter how badly you have to embarrass yourself.”

        And why is it you feel the need to attack postes you disagree with on here?

        Try criticising their observations/evidence instead.
        It’s not big and it’s not clever ny friend.
        And something I have had personal contact via mail with Charles about.
        Yes it was you.

  9. Australia, punching about it’s weight, AGAIN!

    Oi Oi Oi!

    Now all we need is a few low ranked international sporting teams to tour so we can bully then and really feel good about ourselves.

    • Not soccer though -that is a second rate sport in Australia for men but suits some females. My granddaughter plays soccer and was asked to join an elite squad after her district team were premiers but luckily she decided to play at districts -too expensive on time and money to play at state level.

  10. Just some simple observations …
    1. There appears to be a sine wave pattern for the warmest areas for Jan ’19 LTL in the graphic.
    2. They all lie between 30°N to 60°N and between 30°S to 60°S.
    3. This is the same latitude for the jet stream.
    4. Can any potential teleconnections be gleaned from both the pattern and regions most affected? Are they somehow related?

  11. RSS TLT update for January 2019 is +0.67C. Transferred to the same anomaly base as UAH this is +0.53C, compared to +0.37C in UAH. This continues the long running discrepancy between the current versions of the UAH and RSS TLT data sets, in which the RSS anomaly is consistently around 50% higher per month than UAH. At least one of these two sets is wrong. RSS is in much closer agreement with the surface data.

  12. While it is still summer here in Australia, March is our “Fall”month, other than the week of high temperatures , clearly understood as coming from Asia plus the Australian desert, it has here in South Australia been a little cooler than usual.

    But the BOM who just like the once highly respected CSIRO, have a vested interest in scaring the politicians, and this shows in their figures. Perhaps if they were to start their “Hotter” figures at 1898, they would show that the worst was way back. The 1900 and the 1930 tees were far worse than todays figures.


  13. Temps for January here in west MD were near-exactly average.

    But, but, but, we only have 12 years left before we all perish! (running around in circles, wringing hands, gnashing teeth)

  14. I am not convinced that we need even monthly satellite temp readings during an interstadial warm period. That said, it would be good to have them opporating during the jagged slide downward towards a stadial tough and at the turn back up. But that is many thousands of years into the future. Wriggle watching using Uber expensive orbiting devices is too high of a cost for what return we get. Just isn’t justifiable.

    I would much rather we use climate balloons and diving ocean buoys to help us better understand natural weather mechanisms at work in short and long term weather patterns.

    • They’re weather satellites. They aren’t up there to check on global climate trends, that’s just something UAH and RSS attemt to use the data for.

  15. We had a series of high pressure blocking events in January, where highs get stuck in the Tasman Sea between the eastern Australian coast and the mountains of New Zealand. This slowed progression of weather systems and caused the air to heat up over the continent. The MSM were their usual ululating selves at this horrifying heat (which seemed pretty normal for summer here in Newcastle).

    The blocking highs appear to get stuck in the Tasman when NINO3.4 is more neutral. I suspect the latitude of the high pressure systems is affected by the relative pressures (the SOI is a measure of pressure difference) and at a certain ENSO level the highs are just at the right latitude to get stuck.

    Since ENSO isn’t a feature of CO2 the warmth over Australia in January looks to be entirely natural. (Indeed ENSO follows the ~60 year cycle, like the AMO, PDO and other indices.)

  16. How many stations are used in this report, and where can the list of them and their data be found?

    • There are no stations. Temperatures are derived from satellite measurements of upwelling microwave radiation in the atmosphere.

      • Doi! Obviously I didn’t read the article very well or I would have noticed it was about the satellite data. Sad to watch one’s own mental faculties diminish before one’s very eyes.

  17. Borrona Downs AWS recorded the highest minima and also the lowest minima in New South Wales and has been found to be faulty , its listed as being Wanaaring but is 105 kilometres west of Wanaaring.
    Last year I took photos of the Wanaaring Stevenson screen partly because someone had put rocks on top of it and partly because of where it was located and what was near it .

    As an aside a land,mark ruling in Oz today a judge has ruled against a proposed coal mine because of climate change .

  18. Yep, an extraordinarily hot January here in SE Australia (warmest I can ever recall). But of course, the local hysterics here think that 1/10th of our continent is indicative of the entire planet (Peter Hannam of Fairfax the most notable hysteric once again). Good to see the UAH map to put things in perspective – pity the hysterics never do.

  19. In my opinion, a global average temperature is a completely meaningless figure. People can live in very cold places like Siberia and in very hot places like Africa. However, nobody lives in a place called “global average” so why should we worry what its temperature is?

Comments are closed.