What is the Temperature?

It seems like a simple question:  What is the temperature.  Well, we know now that surface temperature measurement is really hard, since its hard to get good geographic coverage when oceans cover 3/4 of the world and biases are a huge problem when most of the measurement points we had in the year 1900 have been engulfed by cities and their urban heat islands.

But John Goetz brings us a new answer to the question, what is the temperature?  Answer:  Whatever the GISS wants it to be, and they seem to change their minds a lot.  He only has the last 2-1/2 years of GISS data but finds an astounding amount of variation in the data over these couple of years.  Excerpt:

On average 20% of the historical record was modified 16 times in the last 2 1/2 years. The largest single jump was 0.27 C. This occurred between the Oct 13, 2006 and Jan 15, 2007 records when Aug 2006 changed from an anomoly of +0.43C to +0.70C, a change of nearly 68%.

I was surprised at how much of the pre-Y2K temperature record changed! My personal favorite change was between the August 16, 2007 file and the March 29, 2008 file. Suddenly, in the later file, the J-D annual temperature for 1880 could now be calculated. In all previous versions the temperature could not be determined.

  • FYI; Just seen the latest on the sunspot cycles. Doesn’t matter what all the AGW crowd say, or how they massage their figures. I think the lowered solar output is about to trump their presumed ‘ace’.

    Prediction; 2008 & 2009 are going to be much colder than average.

  • Scientist

    Deniers spend a lot of time bleating on about how the surface record is full of errors. Now you appear to be suggesting that improving the temperature record is somehow dodgy. So, would you prefer that the record is left static and never improved, or that analysis is updated and improved whenever possible?

  • Adirian

    Scientist – because you can’t “improve” a faulty record with an algorithm. Basic principle of information: Garbage in, garbage out. At best the data is unchanged, and at worst, its accuracy is damaged. No algorithm can fill in missing holes or weed out bad data, operating only ON that data.

    The reason for this is that all the information that is available, when operating on a record, is already in the record. The information is already captured. Any operation on that information generates a subset of the information contained within it, because you can’t generate a superset from a subset. (Actually, you can, but the odds against it are phenomenal, and you’d have no way to verify that it is the “correct” superset. From an informational perspective, then, it’s functionally impossible.)

    The correct approach to take would be to generate new information to operate on the old information with – say, comparing ice core samples with historical temperature records – which is not the case here. (As the GISS is just filling in holes by interpolation, and thus increasing statistical noise.) (And that particular example wouldn’t work, because ice core samples suffer from osmotic contamination.)

  • TCO

    Could you try to ONCE think critically about something that our side puts forward? Sheesh. you really make me sick with your dedication to the day one hypothesis.

  • Tony Edwards

    Scientist and TCO, it’s not a matter of “improving” data, or even something put forward by a side, simply that, once a recording is made, be it of temperature, air pressure, wind speed or whatever, that is the recorded figure. If adjustments have to be made initially, then so be it. But once that has been done, that’s it! Going back several years using some algorithm or other and changing the past should not happen.
    I worked, many years ago, on Concorde wind tunnel tests and if we had a reading of something that looked anomalous, we didn’t estimate what it should be, we did the test again. And once done, that was it. How well do you think it would have flown if we had estimated values from subsequent measurements?

  • John Goetz

    Scientist and TCO,

    The problem I have described in my post is not about “improving the record.” The actual past record never changed. An artifact of the algorithm selected by GISS to calculate global average temperature is that it will perturb past results. It would be not unlike seeing Tiger Wood’s capture the #1 world ranking for 2008 in golf and having that drop his 2000 ranking from first to second.