Global Temperature Update: Nov 2010


The most recent update of the global temperature is done.  This is the blended data set that I initially proposed here.  I will have some follow up articles as I compare the previous hot year of 1998 to 2010.

As this data is new to me I am still playing with it.  I keep finding things that I find interesting.  If I come across anything that I think will be interesting to others, you can be sure I will put together an article about it.  🙂

Here is the overall temperature set:

The Inconvenient Skeptic

The Blended Global Temperature

One thing I have been looking at is the behavior of each set in comparison to the others.  The satellite sets are usually in good agreement which is a good indicator.  Any difference is likely due to the difference in global coverage that they have.

The Inconvenient Skeptic

The Individual Sets

The single most interesting thing is the CRU data set has been higher than the Hadley set on an annual basis since 1985.  In the monthly data since 1985 there is a cumulative difference of 50 °C between the Hadley and CRU set.  Since 1997 there have only been 5 months where the Hadley has higher than the CRU.  Prior to 1985 they were much more similar.  Clearly there is a systematic difference with what is going on with the CRU data.  More on this in the future as well.

I will make my official projection in a later article, but I am predicting that 1998 will retain the hottest year of the instrument era.  If you want to look at the data itself, here it is.

Global Blended Temperature: Nov 2010

based on:

CRU – The original global set.  It covers from 1850-current.  It is a station set.

Hadley – A different global set that also covers from 1850-current.  Also a station set.

UAH – This is a satellite measurement of wavelength. The analysis of the satellite data is done by the University of Alabama in Huntsville.

RSS – Also a satellite measurement of wavelength.  The analysis of the satellite data is done by a private company named Remote Sensing Systems.

Posted in Anomaly by inconvenientskeptic on November 16th, 2010 at 6:19 am.

10 comments

This post has 10 comments

  1. Robert Nov 16th 2010

    Question?
    Why not include any temperature record that does not have 1998 as the warmest on record?

    Considering that the met office concluded that hadley was at the “lower end” of the warming, I find it hard to understand why that is the only group whose data you include? Where is NOAA’s or NASA’s ?

    http://www.metoffice.gov.uk/corporate/pressoffice/2009/pr20091218b.html

    In fact there are so many station reconstructions and not all show 1998 as the warmest on record

    http://www.realclimate.org/wp-content/uploads/globalT0.png

    So I hate to say it, but the fact that you “cherry picked” only records that showed 1998 as the warmest is pretty blatantly obvious. Using the land-only and Land+ocean datasets both from Hadley almost gives the impression you included 2 station datasets when really they come from the same organization.

    Say what you like, but until you show ALL the data, you will forever be a cherry-picker. Speaking of which, nice graph of one region on the planet.

  2. inconvenientskeptic Nov 16th 2010

    I picked 4 sets because that is plenty. I picked the 2 of the 3 most common for station. I didn’t look at the data for any of them before choosing. I simply wanted 2 station and 2 satellite. CRU is the most commonly used one, Hadley is also common. Either way I show all the data.

  3. I think you need the GISS and NOAA sets also. Wood for Trees calculates their own composite of HadCRUT, GISS, UAH, and RSS. They keep all data up to date.

  4. robert Nov 16th 2010

    Well you see in science arbitrarily deciding how much is “plenty” doesn’t quite make it past peer review… especially when its clear that the individual stood clear of the 2 series which show the most warming.

    Look I just showed you a link that said that hadley is likely underpredicting warming and yet you’re okay with that?

    Another thing, you didn’t pick 2/3 most common. That’s a lie. The 3 most common are hadcrutv3, Giss Global Land and Ocean Index and NCDC/NOAA.

    You picked 1 out of the 3 not 2.

    If you wanted 2 station and 2 satellite then why didn’t you pick 2 station ones from different sources. You do realize that hadley and cru are the same except one is land+ocean and one is land only right? That’s not two unique datasets… You HAVE to include another group than just the UAE Cru ones. I just think its awfully convenient of you to neglect the two that show the greatest warming and to double up on the one that shows the least…

  5. inconvenientskeptic Nov 16th 2010

    Robert,

    Most of the data is related. Much like RSS and UAH are from the same source, but different analysis.

    I am however open to feedback. I will take a look at switching out the Hadley for the NCDC. Would that be an appropriate move? I am not particular about the ones I use, but I do believe that a merged set is better than the easy option of picking one of them.

    One purpose I feel strongly about is getting more working together. Feel free to provide any other feedback you have.

    Also I prefer to keep things from the same circular arguments. That is not very helpful for learning. Most people learn more when challenged. I have learned a lot from the posts that people leave. 🙂

  6. Not sure what Robert’s big deal is. The 2 satellite sets come from the same source with slightly different processing and adjustments. All the land or land+ocean sets come from the same sources with different processing and adjustments made to them. Without actual justification (and of course there would always be disagreements) that the warmer NCDC or GISS sets are somehow better, it’s rather moot. So many of the individual station adjustments in NCDC and GISS are ridiculous and seem to be done with the purpose of increasing warming.

    Variance and uncertainty is too often ignored (for example the NCDC website doesn’t even provide error estimates or confidence intervals for their trends.) Averaging sets hides the uncertainty – a complete set should have the multple data points for each year (or month). I like this graph because plotting all the sets gives an idea of uncertainty: http://climate4you.com/images/AllCompared%20GlobalMonthlyTempSince1979.gif

  7. More on averaging: When calculating averages and then pretending that the averaged data is the source data, calculations of statistics produce misleadingly reduced variance (and thus misleading estimates of significance). Global averaged temperature sets are actually averages of stations with large in-station variations. Most of the variation is lost in the wash when performing averaging.
    One of the CRU emails stated: “Thus, if people simply looked at several records they would get the impression that temperature variations were large, ~1.5°C. Imagine their surprise when they see that the ensemble averages you publish have much smaller amplitude.” (See: http://www.appinsys.com/GlobalWarming/UnprecedentedWarming.htm )

  8. As I understand it, satellite measurements are of the lower troposphere and station measurements are surface temperatures, at least the air temperature at a metre and a bit above the surface. doesn’t blending them together give you a temperature of nothing real at all? I know you’re using anomalies but wouldn’t we expect variations in surface temperature and variations in tropospheric temperature to be different, have different time constants and be significantly different in their anomalies too?

  9. inconvenientskeptic Nov 24th 2010

    Rich,

    There are even more items than that. Much of the anomaly data is sea surface temperature (SST). So much of the anomaly data already contains different types of data in the global anomaly.

    The biggest advantage of the satellite data is coverage. The station data can be used for interpolation over area’s up to 1200km away. A single data point can be be used to represent millions of square kilometers.

    I am not proposing that this is perfect, but in measuring the global temperature, it is better to include more rather than less data. This accomplishes that.

  10. inconvenientskeptic Nov 24th 2010

    Alan,

    Seeing the range like that is useful and better perspective is always a good thing.

    As for the variance, there is less global variance than a particular station. If it is warmer at one place it is generally cooler somewhere else. Larger events like the ENSO can cause large area’s to be warmer or cooler for a while.

    So the global variation is less than what most stations will see. The polar areas have the highest variation and the topics have the lowest variation.

    As I have shown here, there is variation within each year. So even a single number for a year is very misleading.

Web Design & Dev by

Mazal Simantov Digital Creativity