|
|
Post by icefisher on Nov 8, 2009 18:00:00 GMT
Briffa didn't consider khad, but if he had the results wouldn't have significantly changed. How it would have changed can be seen clearly in the upper right box from Briffa. McIntyres analysis of including Khad is the lavender line. The black line is when Briffa includes Khad and others. And the blue line is a modified version of the original. As you can see the blue line is way way higher than either the version McIntyre came up or the reworked version by Briffa both of which are very close. It should also be noted that only the blue data set shows a significant difference from the MWP and essentially no difference from the MWP until after dropping to an unacceptably low number of trees as you will see next. Below are the standard deviations of the largest of the data sets in this reworked chronology. The "all" black line data. Still as you can see the std dev is all over the place in the area of the hockey stick as a result of the number of trees dropping below 10 at 1990 and below 5 at 1995. By Briffa's own published standards this method of anaylsis should have at least 15 trees. But the standard of even 15, IMO, should be understood to be generally an indicator only for scientific debate, not national policy. I have never done a sample less than 30 and that was only for relatively unimportant financial information purposes. Briffa compounded this error by using his position as IPCC lead in his area to eliminate challenges to his work, which by his own stardards were inadequate on top of which he kept the archives unpublished so nobody would know just how inadequate it was. The excuse that it wasn't his data to share simply does not hold water as that would not have prevented him from reporting the numbers of trees by year in his study. |
|
|
|
Post by sigurdur on Nov 8, 2009 18:57:28 GMT
I am perplexed. Why even debate Briffa's paper. It is garbage. Anyone with a smattering of stats understands this.
|
|
|
|
Post by socold on Nov 8, 2009 20:52:13 GMT
Briffa didn't consider khad, but if he had the results wouldn't have significantly changed. How it would have changed can be seen clearly in the upper right box from Briffa. McIntyres analysis of including Khad is the lavender line. The black line is when Briffa includes Khad and others. And the blue line is a modified version of the original. As you can see the blue line is way way higher than either the version McIntyre came up or the reworked version by Briffa both of which are very close. In the final figure he compares the previously published records with the new one that include Khad:  Whether the number of trees are acceptable or unacceptable is independent of Khad, it's a different argument, something which McIntyre has aptly described as "moving on". It does hold water seeing as people were demanding the data from him and accusing him of hiding it. Again this sounds like "moving on", the original argument was that he was hiding the data which has "moved on" to an argument that he was holding back the number of trees used by year. |
|
|
|
Post by icefisher on Nov 8, 2009 21:16:42 GMT
It does hold water seeing as people were demanding the data from him and accusing him of hiding it. Again this sounds like "moving on", the original argument was that he was hiding the data which has "moved on" to an argument that he was holding back the number of trees used by year. He both hid the data and hid the fact the data was inadequate. Not hiding the data would have taken care of both problems. Reporting the details of how many samples he had would have only fixed half the problem but would have been sufficient to discard his results and prevent the misuse of it. For example assume for a moment he had a sample of sufficient size. Reporting that fact would not have relieved him of the responsibility to provide sufficient information so others could replicate his work. He is still doing that by not providing crossreferencing between the underlying data and the final assemblages. He obviously has the information necessary but continues to obfusicate. This man is a bad man!!! |
|
