Skip to content

Global Temperature Report – August 2014

October 3, 2014


By Paul Homewood  


I’m fed with waiting for HADCRUT data, so will exclude them, particularly as RSS data is already out for September.

[If the HADCRUT numbers are out, I cannot access them as I get a bad gateway. They are also not on Woodfortrees yet. If anyone can find them elsewhere, please shout!]


August 2014 0.19 0.20 0.70 0.75
Change from last month -0.16 -0.10 +0.17 +0.10
12 month running average 0.23 0.26   0.65 0.68
Average 2004-13 0.23 0.19 0.59 0.59
12 month average – 1981-2010 Baseline 0.13 0.26 0.25 0.24








There has been reference at various times this year to “record temperatures”. In reality though, temperatures are only slightly up on last year on GISS (YTD of 0.64C v 0.61C for 2013), while UAH YTD is exactly the same as they showed for 2013.

RSS figures are already out for September, virtually unchanged from August at 0.21C, which will bring the average anomaly so far this year a bit lower still.

Meanwhile both satellite datasets show YTD anomalies well below those of 2010, 0.16C and 0.21C for UAH and RSS respectively. 



And finally

I am sometimes asked to use centred running averages, although I am not quite sure what they add. Indeed, I was once accused of leaving out the chunk at the end of the graph when I used centring!

I would welcome any views on this, and will change if it is felt appropriate.



UAH and RSS are the two satellite datasets, that measure temperatures in the lower troposphere, from the surface up to about 8000 metres. The NCDC, HADCRUT and GISS datasets measure surface temperatures.

All temperatures are presented as anomalies, i.e the difference, measured against a baseline, that is different across all four sets. (This means that the anomalies are not directly comparable between sets)

The baselines used are:

RSS – 1979-98

UAH – 1981-2010

HADCRUT – 1961-90

GISS – 1951-80

NCDC – 1901-2000 


HADCRUT is maintained the UK Met Office’s Hadley Centre in conjunction with the Climatic Research Unit at the University of East Anglia.

GISS is run by the Goddard Institute for Space Studies at NASA.

NCDC is the National Climatic Data Center, part of NOAA.

UAH is the University of Alabama in Huntsville, and their dataset is part of an ongoing joint operation with NOAA and NASA.

RSS is a scientific research company, Remote Sensing Systems.


1) RSS

2) UAH




  1. October 3, 2014 5:01 pm

    “I am sometimes asked to use centred running averages, although I am not quite sure what they add. Indeed, I was once accused of leaving out the chunk at the end of the graph when I used centring!

    I would welcome any views on this, and will change if it is felt appropriate.”

    Don’t worry, even the MO have trouble with that.
    They get around it by using the same data over and over again.
    I prefer to use non-centered data, although you have to remember the plot applies the end of the data period.
    You can’t (AFAIK) use centered data if you get Excel to calculate the MA, at least not on my version.

  2. October 3, 2014 5:36 pm

    of all data sets I think this one
    from RSS comes the closest to the truth. It shows a drop of almost 0.1K since 1998.
    However, my own data set says it is at least -0.2K since 2000.
    The problem with most data sets is the question: how did they balance it? Note my conditions for a reasonable balancing act:

    • October 3, 2014 6:29 pm

      I don’t think any of the datasets come close the the truth.
      They each have their deficiencies.
      RSS doesn’t include most of Antarctica.

  3. October 3, 2014 8:11 pm

    I could not find a station in Antarctica going back a decent number of years (30-40) to give me a trend but we know that the ice there is increasing, meaning it must get cooler there.
    I also looked at Alaska and found these results:

    The average of 10 weather stations in Alaska was -0.055K/annum since 1998.
    That is -0.9K since 1998…..
    So we are globally cooling from the top latitudes down. You see less change at the lower latitudes as more condensation energy comes free here (due to the growing temp. differential between equator and poles).

  4. R James permalink
    October 3, 2014 10:15 pm

    Perhaps Hadley is still bust homogonising the Hadcrut data so it comes out right. They seem to take longer every month.

    • October 3, 2014 10:25 pm

      The Hadley data is there!

      See my links above.

      It’s just that the MO are terribly organised and amateurish..

  5. October 3, 2014 11:41 pm

    Thanks, Paul. Good hunting/gathering of data.

  6. David permalink
    October 4, 2014 9:53 am

    For the record, HadCRUT4 was +0.67 (above 1961-90 avg). Like NOAA and NASA, HadCRUT4 makes Aug 2014 the warmest August on record.

    RSS and UAH, being lower troposphere measurements, haven’t started to fully reflect the increase in surface temperatures as yet, though UAH is still running pretty high (joint 3rd warmest Jan-Sep period on record).

    I can’t replicate the charted “12 month average – 1981-2010 Baseline” data. I get:

    RSS: 0.14
    UAH: 0.24
    HadCRUT4: 0.27
    GISS: 0.27
    NCDC: 0.27

    Re centred averages: I prefer them as they centre the average of the running mean with changes in the actual data. I just think it looks better for presentation purposes. Using running data without centring tends to offset the smoothed mean forward from the actual data, giving it a slightly odd look. Just my view.

  7. October 4, 2014 9:58 am

    Surprise, surprise, HC4 v3 shows a (relatively) large increase over v2 in the years between 1998 and 2012.
    Enough to nudge the trend ever so slightly in favour of warming, of course.

    • David permalink
      October 4, 2014 1:51 pm

      Still doesn’t show as much warming as UAH, though. HadCRUT4 v3 shows +0.05C/dec warming 1998-2012; UAH v5.6 shows +0.06C over the same period.

  8. October 4, 2014 1:23 pm

    ….and there is no warming since 2000.
    There is only [global] cooling.
    Look at my 3 graphs at the end of each table:

  9. October 5, 2014 7:03 pm

    hadcrut4’s new version 4.3 shows an increase in GSTA compared to version 4.1 by 0.001C/month since 1990. Again, warming the last 20-25yrs… Oh yes, the difference is less compared to v4.2, but each new version subsequently adds just enough. Can’t wait to see how v4.4 or 4.5 etc compare to 4.1… if it ain’t warming, we’ll make it warm.

  10. October 5, 2014 7:10 pm

    HadCrut4, v4.3 shows the following linear temporal trends (slopes; C/yr):

    last 5yrs: 0.002 (NS)
    last 10yrs: -0.000 (NS)
    last 15 yrs: 0.007 (NS)
    last 20yrs: 0.011 (SS)
    since 1976 low: 0.014 (SS)

    Hence, no statistical significant warming over at least the last 15yrs according to HadCrut4, version 4.3, which is in line with the other databases, of which RSS shows no warming over the last 18yrs and 1 month.

    Looking ahead, I am afraid the coming La Nina next year, after the current “la nada” will further the continued lack of GW.

  11. Dumbo permalink
    October 6, 2014 2:01 pm

    For the life of me, I can’t figure out how one calculates (or should calculate :)) the 6th line of the table.
    You only confuse me further,

    Maybe I’m just dumb. Help!

    • October 6, 2014 4:56 pm

      As an example, RSS anomaly during 1981-2010 was 0.1C, i.e. it was 0.1C warmer than 1978-98 that RSS use as their baseline.

      The latest 12 months averages 0.23C against their 1978-98 baseline, which equals 0.13C against 1981-2010.

      I like to present all datasets against the same baseline so as to make direct comparisons easier

      • Dumbo permalink
        October 7, 2014 1:10 pm

        Tks. I’ve seen the light. The mistake I made was two-fold:
        1) I attempted to reconcile line 6 with the information in the table alone.
        2) I misread the minus sign (that between “average” and “12”) in “12 month average – 1981-2010 Baseline” as a dash.

      • David permalink
        October 7, 2014 3:26 pm


        Still can’t figure out how you’re getting 0.25 for GISS and 0.24 for NOAA.

        I think it’s a rounding issue for RSS, but to end August I get 0.14 rather than 0.13. However, for GISS to Aug I get 0.27, not 0.25 and for NOAA I get 0.26 rather than 0.24.

        The GISS and NOAA data move around a bit (#cough#), but I make the current 1981-2010 average in GISS 0.40 and in NOAA it’s 0.42. Is that what you get?

      • October 7, 2014 4:58 pm

        I cant check till I get back home, David, as all my spreadsheets are there.

      • David permalink
        October 7, 2014 6:12 pm

        No problem Paul. Take a look at it if you get a chance once you get back from (your latest) holiday.

        Wish I had a job like yours… 😉

      • October 7, 2014 8:55 pm

        It’s a hard life!

  12. David permalink
    October 6, 2014 3:51 pm


    Over the past 18 years and 1 month the other main satellite data producer, UAH, shows warming of 0.10C/dec. This is faster then all the surface data sets, though it’s in much better agreement with the surface data than it is with RSS.

    Interestingly, the rate of warming over the past 18 years and 1 month in UAH (from Sept. 1996) is faster than it was during the 17 years and 8 months leading up to Sept. 1996. Most of the warming in the UAH record has occurred over the period when RSS says there was no warming.

    Yet people seem to think satellite data are more reliable than surface.

  13. October 6, 2014 4:22 pm

    I figured UAH has issues with calibration, (what is zero in space just outside earth?) because it should read the same as RSS, which is currently the most correct set. They all have problems due to imbalance,
    i.e. surface stations are predominantly NH
    and RSS and UAH are not seeing that we are busy cooling from the top latitudes down because you get more condensation energy coming free at the lower latitudes…..

    Why don’t you get a class of students to copy my set using different stations and employing my sampling technigue?
    It won’t take you that long to find out that we are busy globally cooling at a rate that is scary [to me]

  14. October 9, 2014 9:12 am

    The MO have finally worked out how to change their links to HC4.3 data

    • David permalink
      October 9, 2014 10:42 am

      Looking at that data, HadCRUT4 has been 0.5 C or above for 12 out of the past 14 months. If El Nino conditions develop, or if near El Nino conditions persist for the rest of the year, we could easily see 2014 challenge for 3rd or even 2nd warmest year on record.

      Anomaly average of 0.51 or above for Sept-Dec would guarantee 3rd spot ahead of 1998. 0.52 or above would tie for 2nd with 2005. 0.56 would be required for a new record, but I suspect full El Nino conditions would be required before the end of the year to achieve that.


  1. Global Temperature Report – August 2014 | The Global Warming Policy Forum (GWPF)

Comments are closed.

%d bloggers like this: