MXD Tree Removal Service (Choppin Wood off A Hockey Stick)
Posted by Jeff Id on October 5, 2008
We found out last week that Shweingruber MXD series had some very unusual things done to it. Series were cut off at 1960 and new fake data was pasted on the ends. The fake data of course correlated well to temperature and over 90% of the series passed correlation. The comments regarding the data warned that data after 1960, temperature records were mathematically incorporated in the series.
The values after 1960 are a combination of information from high-frequency MXD variations and low-frequency instrumental temperature variations. We recommend, therefore, that the post-1960 values be deleted or ignored in any analysis that might be biased by the inclusion of this observed temperature information, such as the calibration of these data to form a climate reconstruction, or comparision of these data with instrumental climate observations for the purpose of assessing the ability of these data to represent temperature variability.
I got this gorgeous quote from Climate Audit which was copied from the original data site. What they are saying is don’t correlate the data because you may accept too much of it, leading to criticisms. However, the added temp info didn’t really help too much.
The line to the right shows the difference between the actual series data and the data used. The blue line to the right of the yellow DOES NOT EXIST because it is faked in using a BS statistical process. The purple line is the actual data with the information to the right incorporating temperature to try and help it out.
You can see that the “real” data wouldn’t have a good correlation with rising temperatures. This is the case despite the fact that actual temperatures were used to try and reintroduce low frequency variation into the data. Therefore the original (not as messed with) data must be even worse. What we can say is that very few if any of these series would correlate according to M08 standards with the real data in place.
So what happens when you take them out. The graphs below represent reconstructions of the M08 reconstructions. I back calculated the coefficients required to make one of the final NH graphs. The result is therefore an approximate of the actual values and must be interpreted as preliminary results until the software can be reproduced.
The Schweingruber MXD series represented about 20% of the graph weight in the period they were used as shown in my weight graph below.
The light blue section in the top right side of this graph depicts an estimate of the overall strength in this series. But the question is how does it change the reconstruction.
This graph shows 3 series, the purple line is with all data reconstruction, the blue line represents the data without Shweingruber and the Yellow line is the difference (after – before).
When you consider that M08 is using these series to predict temperature to a tenth of a degree in history, amount of change is not insignificant. While cutting off data series and pasting fake data back in should be enough to stop any reviewer from accepting the paper, the addition of these series is not the smoking gun by themselves. I actually see it as a symptom of the methodology in this science where data can be preferentially scrapped through statistical rubbish, it can also be added in. It’s like opening Pandora’s box.
By the way we had 484 series, these MXD represent 95 series of 110 original series so now we are down to 389. The Luterbacher series are actually temperature and not proxies so we are down those 71 series to 318 of the original 1209 series.
The next graph is the same as above with the M08 northern hemisphere reconstruction laid on top It looks different than his paper because it doesn’t have 30 year smoothing. I really like the flat spot in the last 80 years.
This last 80 years should have been a big signal to the guys who made this graph that maybe the data ain’t quite right. They are after all sorting all the series to match a rising temp particularly in the last 80 years.
One last graph showing the percent change in alleged temperature.
Read this graph carefully as temps close to zero can show a large percent change. I think it demonstrates pretty well that this group of proxies was not insignificant.