A blogger identified as hpx, over at the well named SaveCapitalism blog started messing around with the GHCN temperature dataset. The GHCN data is apparently 95% used in the recently implicated east Anglia university dataset by professor Phil Climategate Jones. His post was extremely interesting because it has to do with the GHCN overstatement of Antarctic Warming. It’s important to me because we need to verify that the datasets Ryan is using are of good quality for publication and reliance on previous work isn’t always enough. In this case, however, instead of temp records, I found a box of old socks.
Before getting into it, I’d like to say that hpx did a great job. He grabbed hundreds of records and their duplicates, compiled them into a database, sorted for what he needed and plotted the results. He found a trend of 4C/century in the Antarctic as calculated from the ‘adjusted’ value added data. I’ve been able to verify this value to within a couple of tenths of a degree, using some slightly different methods. Check out his post at this link.
Almost all the data we have in the CRU archive is exactly the same
> as in the Global Historical Climatology Network (GHCN) archive used
> by the NOAA National Climatic Data Center
And that is the key– almost all. If they used the GHCN Antarctic stations, there is a teenie tiny problem. Of course as with anything in science these days, the devil is in the detail.
First, each GHCN station can have multiple records. These records are taken at different times and simply recorded with the others so for each station the data might look like this one from the Antarctic.
Which is then processed and adjusted to look like this.
It’s a bit difficult to see which series are what because many of them are absolutely the exact same numbers as others in the raw format. These are absolute temperature values in Deg. C so for my analysis, I simply averaged the three series. This is different from SaveCapitalism’s analysis in determining what GHCN is exactly saying about trend.
In the raw data from GHCN there are 27 stations from which data was collected. Their positions are shown on this map.
This data is reduced in the final version to only ten adjusted stations. All the denialist blue stations have been removed, including Amundsen Scott at the South pole. The peninsula stations here show up as black.
1 – Average multiple series from same location together in Degrees C
2 – Calculate anomaly for each individual station
3 – Average all anomalies together ignoring area weighting effects
The slope in C/Decade is in blue on the graphs. So by the raw data we’re looking at 2.3C/century, this is a several times higher trend than the results from a more complete BAS dataset. For those familiar with the GHCN it’s no surprise that after the data is “homogenized” the trend leaps upward to 4C/century.
So the next step was to look at the data a little closer at the ten series.
There are like 1500 six figure people in the Antarctic and the primary GHCN global dataset has only one single series since 1993. The heroic global warming scientists at Rothera Point have been slaving away, reporting data for all of the Antarctic. That’s 16 years without anyone else bothering to take a temperature reading in the Antarctic for the GHCN or anyone from the GHCN making a phone call to any of the Antarctic towns for the data.
Oops, not so quick. There was data in the other series. In the GHCN’s unique value added process, (invented and widely sanctioned in the superior science of climatology) denialist data has been eliminated from the Raw data as we can see below.
Below is the same map as above with a different temperature scale so we can see the magnitude of the black dots. The red circle is surrounding Rothera Point station.
So as we can see, of all the stations available in the antarctic, GHCN has chosen to use a single station on the Antarctic Peninsula to represent an entire continent of the earth for the past 17 years (red circle). But it’s not just any station, it’s a special one. Rothera Point has the single highest trend of any of the adjusted station data.
In the meantime, the BAS maintains an up to date, value added dataset of their own. Using 63 stations and simple area weighted averaging we get antarctic trends of 0.05 C/decade with a temperature distribution in C/Decade that looks like this:
Ok, so for the regulars, you know I’ve maintained my calmness quite well. However, it’s not easy. I’m sick to death of advocate scientists pretending there are only minimal problems in the temperature record. Currently the ‘homogenized’ value added version of GHCN has a trend that is EIGHT times higher than actual for the ENTIRE ANTARCTIC CONTINENT. So I wonder if we can now, spend some of the ‘BILLIONS OF DOLLARS’ on cleaning up the temperature record!!! It’s no coincidence that AGW scientists aren’t demanding this be done in my opinion either.
Which of these records is used in CRU, GISS, NOAA — hell if I know (nobody else does either because at least CRU won’t say) but it’s pretty clear none of this data should be used in this condition.
It’s time the GOOD scientists demand GOOD TEMPERATURE DATA. It’s time the world embarked on a real project for gathering the true warming data rather than this kludged mess. It’s past time that the whole thing was done in an open and transparent way. The whole experience with GHCN this weekend felt like looking through a box of old socks.