# Loehle has corrected his results. Shouldn't Spencer?

Update:
Back in 2007, Craig Loehle published an amateurish paleoclimate reconstruction paper. Loehle used only previously published non-treering data and arrived at his average by, well, simply averaging all 18 datasets year by year. This paper contained precisely zero equations, zero tables, and three...
show more
Back in 2007, Craig Loehle published an amateurish paleoclimate reconstruction paper. Loehle used only previously published non-treering data and arrived at his average by, well, simply averaging all 18 datasets year by year. This paper contained precisely zero equations, zero tables, and three graphs.

http://www.earthandocean.robertellison.c...

Loehle's paper was immediately jumped on by Gavin Schmidt (among others), who pointed out a pleothera of issues, especially statistical.

http://www.realclimate.org/index.php/arc...

Shortly after that, Loehle apparently recruited a statistician, J. Huston McCulloch, and circulated a "correction" to his original paper. The correction was three times longer than the original, and contained 5 equations, one table, and 46 graphs.

http://www.econ.ohio-state.edu/jhm/AGW/L...

One of the things Loehle did in his co

http://www.earthandocean.robertellison.c...

Loehle's paper was immediately jumped on by Gavin Schmidt (among others), who pointed out a pleothera of issues, especially statistical.

http://www.realclimate.org/index.php/arc...

Shortly after that, Loehle apparently recruited a statistician, J. Huston McCulloch, and circulated a "correction" to his original paper. The correction was three times longer than the original, and contained 5 equations, one table, and 46 graphs.

http://www.econ.ohio-state.edu/jhm/AGW/L...

One of the things Loehle did in his co

Update 2:
One of the things Loehle did in his correction was to weight each proxy series by the covariance of the series. This has the effect of weighting those proxies with the greatest errors (compared to the raw mean) by the least amount, while giving the greatest weight to those proxies with the smallest errors compared...
show more
One of the things Loehle did in his correction was to weight each proxy series by the covariance of the series. This has the effect of weighting those proxies with the greatest errors (compared to the raw mean) by the least amount, while giving the greatest weight to those proxies with the smallest errors compared to the raw mean. This is a rather standard procedure in studies of this type. It has the effect of sharply reducing the variability in the results, since you're turning down the volume on the noise, while turning up the volume on the signal.

Question:

Since Loehle went to the trouble of putting out a corrected graph, why is Roy Spencer still providing Loehle's uncorrected, overly-noisy graph of raw averages? And, has Jim Z. actually read Loehle's correction? Since he seems to be citing it.

http://www.drroyspencer.com/global-warmi...

Question:

Since Loehle went to the trouble of putting out a corrected graph, why is Roy Spencer still providing Loehle's uncorrected, overly-noisy graph of raw averages? And, has Jim Z. actually read Loehle's correction? Since he seems to be citing it.

http://www.drroyspencer.com/global-warmi...

Update 3:
Edit for Jim Z:

I'll take that as a "no". And sheesh, how paranoid do you have to be to think I'm Dana?

I'll take that as a "no". And sheesh, how paranoid do you have to be to think I'm Dana?

Update 4:
EDIT for Eric C: In the raw average graph, the difference between the top of the MWP and the bottom of the LIA is about 1.1° C. But in the weighted graph, the difference is about .7° C. This implies that at least a third of the variation in the raw graph is noise. And since Jim Z. cites this graph as...
show more
EDIT for Eric C: In the raw average graph, the difference between the top of the MWP and the bottom of the LIA is about 1.1° C. But in the weighted graph, the difference is about .7° C. This implies that at least a third of the variation in the raw graph is noise. And since Jim Z. cites this graph as "evidence" of natural climate variability, that's significant.

2 following

7 answers
7

Are you sure you want to delete this answer?