Niche Modeling

11
Aug
Comment on McLean et al Submitted
Posted by David Stockwell in All, Climate, Programming, Reviews, Theory

Table of contents for SOI

Here is the abstract for our comment submitted to Geophysical Research Letters today. Bob Tisdale is acknowledged as the source of the idea in the first paragraph. Lets see how it goes. If you would like a copy, contact me via the form above.

Update: Now available from arXiv

Comment on “Influence of the Southern Oscillation on tropospheric temperature” by J. D. McLean, C. R. de Freitas, and R. M. Carter

David R.B. Stockwell and Anthony Cox

Abstract

We demonstrate an alternative correlation between the El Nino Southern Oscillation (ENSO) and global temperature variation to that shown by McLean et al. [2009]. We show 52% of the variation in RATPAC-A tropospheric temperature (and 59% of HadCRUT3) is explained by a novel cumulative Southern Oscillation Index (cSOI) term in a simple linear regression model and 65% of RATPAC-A variation (67% of HadCRUT3) when volcanic and solar effect terms are included. We review evidence from physical and statistical research in support of the hypothesis that accumulation of the effects of ENSO can produce natural multi-decadal warming trends. Although it is not possible to reliably determine the relative contribution of anthropogenic forcing and SOI accumulation from multiple regression models due to collinearity, these results suggest a residual accumulation of around 5 ± 1% and up to 9 ± 2% of ENSO-events has contributed to the global temperature trend.

Next in series

Tagged as:
Published by David Stockwell in: All Climate Programming Reviews Theory
If you like this blog please take a second from your precious time and subscribe to my rss feed!

Andrew

From the paper:

"However, our estimated sensitivity
of global mean temperature to anthropogenic forcing
when cSOI is included is 0:19 0:02%K=W=m2"

This just happens to be very similar to the value found by Douglass and Knox 2005:

http://www.pas.rochester.edu/~douglass/papers/2...

and also that found by Lindzen and Choi 2009:

http://www.leif.org/EOS/2009GL039628-pip.pdf

Nick Stokes

David,
I read your paper. My first comment is that it is absolutely necessary to read the paper to understand the abstract. You should rewrite the abstract if you get a chance.

The second comment is that the abstract doesn't convey more than a small part of your findings. The key results are in Table 1. And they show that by far the best explanatory variable is AF, the anthropogenic forcing. 69% for RATPAC and 66% for Hadcrut3. SOI explains nothing, and cSOI is weaker than AF. But cSOI is subject to the criticism that it looks like a variable just made up for the purpose. SOI is the difference in barometric pressure between Darwin and Tahiti, and there's no clear physical meaning for a running sum of these.

The multi-variable models don't seem to show anything significant. You can always get some increase with R2 by adding any variable at all to the model.

I'm not sure that yet another paper showing that GTTA is best explained by anthropogenic forcing will appeal to GRL, but good luck!

davids99us

Nick, You are right the a correlation with anthropogenic forcing is uninteresting. The point is that cSOI explains almost as much. This is against a background of Foster thinking that SOI can only explain 7% and is insignificant to trend. cSOI demonstrates this is not true, and supplies the parameters under which it is not true, a small 5% accumulation of the heating effect of ENSO.

The literature on SOI is extensive. It is the best proxy for trade wind strength, temperature difference, etc. It has more physical relevance to temperature than a tree ring width, for example.

The purpose of the multi-variable model is to show the relative contribution. Given the coefficient for AF is in the right ballpark, you are getting about half the variance being explained by cSOI, and half by AF, a large part as McLean claim.

In a nutshell, the increase in temperature since 1960 is restricted to a period between 1976 and 1998 where we had a series of strong El Nino and very weak La Ninas. Its the El Ninos that are responsible for a large part of the warming we observe since then.

All the index does is accumulate over that period, its flat elsewhere. There are physical theories for it, the AR4 acknowledges it, but no-one has really quantified it properly.

Nick Stokes

David,
firstly I think you should be commended for computing the AF data for comparison. One big fault with papers like McLean et al is that they just quote an R2 figure and say "isn't that big!". It is important to have something to relate it to.
But the weakness of your cSOI argument is summed up by the cliche "correlation is not causation". The fact is that correlation is valuable in affirming a causative link, provided you have made that link independently. There is a well-developed theory that says that AGW (AF) could cause warming, so the correlation is significant. You haven't provided any such mechanism for cSOI. It's true that one can say that El Nino's cause temproary warming etc..., but that's an argument for SOI correlation, which fails. If you want to advance cSOI, then you would have to explain why that is plausibly causative, while SOI isn't.

davids99us

Nick that is absolutely true. Correlation is necessary but not sufficient for causation. But there is only so much I am going to do for a comment, particularly it is to try to clear up the SOI correlation issue, and not the mechanisms. There does not seem to be any objection to El Nino heating, its just assumed that La Nino cools to the same degree, which some have contested. Its all in the papers cited. The simple minded mechanism is that El Nino gets locked on the heat cycle through local feedback loops, and some of that heat gets transferred globally via meridional flow and Hadley circulation, etc, and cant get away so easily (the asymmetry issue). I don't know and I don't know if anyone else knows for sure either.

Dr Jones

Under global warming evaporation increases (about) linearly in temperature but water holding capacity increases exponentially in temperature. This relationship forces a slow down in the Walker circulation which will lead to a tendency for more negative SOI values with warmer global temperatures. See Vol 441|4 May 2006|doi:10.1038/nature04744.

Climate has long ago moved from the blind application of statistics. Your 9% mixes up correlation with causation (though even so is correct in showing that global warming is not caused by ENSO).

davids99us

"Climate has long ago moved from the blind application of statistics."

Phenomena still need to be shown to be statistically significant, or
else they could be just coincidences. If understanding of the basic
physics was enough, we wouldn't need experiments and observational
evidence.

davids99us

From the paper you mention by Vecci et al
http://www.gfdl.noaa.gov/bibliography/related_f...
on the moderation of change in sea level pressure. Couldn't the trend have
simply been caused by the frequency of the El Ninos over the period? Now we
are getting fewer El Ninos the trend is reverting, independent of GHG
concentrations.

davids99us

Dr Jones, thanks for dropping by. I presume you are saying there is an
interation between global warming and ENSO, such that GW causes more ENSO
conditions, which is certaintly possible and speculated. The assumption
that there is no interaction is one of the assumptions I explicitly mention
in the paper.

There a few issues I would appreciate your comment on that are of interest
to me. The first is that some climate models show an intensification of
ENSO with GW, but others don't. So the models are mixed on this issue.
This was remarked on by Susan Solomon in the AR4 I think.

The second is that the period of more pronounced El Nino events seems to
have stopped.

If there really is an interaction, leading to increased El Nino's due to GW,
then we would expect the period of high El Nino to have continued on from
1998. Instead, we sem to be seeing a reversion, independent of GHGs.

Doesn't this contradict your view?

Dr Jones

You mix changes in variability and changes means. The slowing down of the Walker circulation is an observed fact and can happen regardless of whether you get more, less or the same number of El Nino/La Nina events. A shift towards a weakened Walker circulation is required by basic physics otherwise the tropical atmosphere will quickly rain itself out. The slow down leads to a shift towards a more negative mean SOI which has clearly been seen seen over the last century (http://www.agu.org/journals/gl/gl0718/2007GL030...).

I have shown you just one paper - there are many more - in the peer reviewed literature which deal with this very issue.

davids99us

Here is a quote from <a
href="http://ams.allenpress.com/perlserv/?request=get-abstract&issn=1520-0442&volume=009&issue=05&page=1020">Long-Term
Changes in the Equatorial Pacific Trade Winds by Allan J. Clarke and
Anna Lebedev used a a primary reference for the methodology in the
Vecchi paper.

A proxy record of eastern Pacific sea surface temperature from
coral suggests that such long-term (decade and longer) weakening and
strengthening of the Pacific equatorial trades has occurred before
major anthropogenic greenhouse gas release and at least back to 1600
AD.

What is the evidence that the weakening exists (and is caused by
GHG's)? It would be interesting to see an updated &DeltaSLP, because
the Figure 3 in Vecchi is less than convincing. Weakening appears due
to the strong El Nino's between 1976-98.

Bob Tisdale

Thanks for the recognition, David. May I have a look at the rest of the paper, please?

My first question is why use SOI instead of one of the SST-based indexes, NINO3.4 or CTI?

davids99us

Bob, for comparison as SOI was used by McLean. I will get the paper out tomorrow.

Dr Jones

If McLean was "clever" he would have used an index which includes SST (the MEI, Nino 3.4 etc, GEOPHYSICAL RESEARCH LETTERS, VOL. 32, L07709, doi:10.1029/2005GL022350, 2005). These show a strong warming trend and would have allowed him to explain an even greater fraction of the variance.

Nick Stokes

these results suggest a residual accumulation of around 5 ± 1% and up to 9 ± 2% of ENSO-events has contributed to the global temperature trend.
Could you explain what that means please?

davids99us

Nick, It means that the coefficients in the linear regression fit have
5-9% of the increase in global temperature from the ENSO SOI term
accumulated in the cSOI term. cSOI = 0.05 Sum SOI(t)

Sign up/in

Authors

Abouts

Blogroll

Literature

Comment on McLean et al Submitted

Table of contents for SOI

archives

tag cloud

Copyright © 2008 - 2010 Niche Modeling

Britney's Online Dating Websites