Blocked From Real Climate and Tamino
Posted by Jeff Id on October 12, 2008
Yesterday, I tried to call attention to my latest post at Real Climate and request an explanation of the fact that CPS methods and r sorting can create any pattern you want from the proxies in Mann08 including sine waves and negative hockey sticks. I have been snipped before from Real Climate so I saved a copy of my post before I sent it.
I have just completed a post which demonstrates how you can pull any pattern you want from the M08 proxy set using CPS methods. Positive hockey sticks, negative ones, historic trends of any shape even sine waves.
I would like someone to review and explain why I am wrong in my conclusions that these methods are faulty.
I also implanted the CPS northern hemisphere result in red noise and used CPS to go find it. I found that the method distorts non-calibration range temperature scale. The red noise was not calibrated to proxy noise but I have used hundreds of different noise levels to the same effect.
Your site deleted my first post on this subject, leading me to spend hundreds of hours in study and even create an entire blog just to react to what seems like bad mathematics.
Now I posted this on his thread “Progress in reconstructing climate in recent millennia” which is a relatively old thread without much activity so my post was on topic. I wasn’t surprised not to get a reply.
I posted another comment on Open thread #6 at Tamino, hoping to get some of that open mind discussion group going on it. Tamino has in the past been impatient with me for asking Ian Jolliffe if he was the nature reviewer but he always let me post. This time though, when my first comment didn’t make the cut I assumed that it was my mistake so I tried again. The second comment didn’t make the cut for his Open thread either.
I didn’t save the posts but they simply asked his group who was still discussing weather PCA would induce the hockey sticks as pointed out by Climate Audit something like
- The r sort in CPS can produce any pattern you want from the latest proxy set and I had a higher correlation percentage to a negative slope on the non-infilled data than Mann08 achieved after painting in an upslope on 90% of the data series. And I provided a link.-
BLOCKED! From an open thread. Twice.