Rushing to Judgment (Global Warming Questioned

That's fine, just lets me work on collecting Solar & Climate related datasets into a spreadsheet along with studies for developement of the Solar/Other factors side of the debate ;O)

Sometimes its difficult to find just the right chart, or locate one remembered, on the internet to demonstrate one's point. Hoping someone else has worked along the lines one wants to develop an argument is not always effective. I'm working to remedy that situation by collecting datasets and sources from which to verify/test and graphically present results under discussion.

This is just an alert that I am prepared to commence our discussion again. A new paper has been published (and I was able to obtain a copy) that is very relevant to the discussion of solar forcing of climate. You will enjoy some of the author's conclusions. You won't agree with all of their conclusions, one in particular.

I hope to post a short summary of the paper tomorrow.

In the same issue, there was a second paper that was also quite interesting, related to this issue. I'll summarize it, too, but I think that most of the discussion will be focused on the first paper.

Future reference global warming farce bump...

Do you have a hyperlink for the paper or have you posted the document online so we can review it with you?

I can't post it online, and when I checked for it online, you have to be a subscriber to view it. It's published in Journal of Climate. I'll provide the URL for the abstracts:

Do Models Underestimate the Solar Contribution to Recent Climate Change?

Volcanic and Solar Forcing of Climate Change during the Preindustrial Era

p>Do Models Underestimate the Solar Contribution to Recent Climate Change?

Interesting Abstract statement:

"It is found that current climate models underestimate the observed climate response to solar forcing over the twentieth century as a whole, indicating that the climate system has a greater sensitivity to solar forcing than do models. The results from this research show that increases in solar irradiance are likely to have had a greater influence on global-mean temperatures in the first half of the twentieth century than the combined effects of changes in anthropogenic forcings. Nevertheless the results confirm previous analyses showing that greenhouse gas increases explain most of the global warming observed in the second half of the twentieth century.

Though it would be nice to see the results expressed in numbers perhaps you may provide us with that information.

As a quick and dirty linear regression shows, over 70% of the change in temperature since 1850's has been directly correlated with change in solar activity. Less than 15% can be attributed to increasing CO2 (natural plus anthropogenic) concentrations. The substantive remainder variability being short term cyclical fluctions (e.g. elNino etc) and spordic volcanic events.

Of interest is that we passed through a peak in solar activity over the last 50 years in which little has changed in overall activity related to the >80yr Gleisberg cycle.

50 years of unusual solar activity
October 21, 2003
http://physics.about.com/b/a/036554.htm

A moderate (0.0266^oC/decade) linear trend in temperature appears to be detectable which could be corelated with an apparent exponential increase in CO2 concentration. The CO2 correlated trend in temperature is linear because the radiative forcing forcing of CO2 is a logrithmic function of CO2 concentration.

The following global temperature reconstruction is composed of the sum of the relative contributions of Solar & CO₂ concentration as components of temperature.

The Solar Component(S) is the the solution of a linear regression of Solar Activity as measured by Lean '98 for (1956-1977) scaled and appended to the composite ACRIM Satellite data series (1978-2000) of total solar irradiance Frohlich and Lean '98 vs global instumental land & ocean temperatures, Jones et.al '01.

T_s = 0.2685*S-366.95;
Stderror 0.17^oC,
Correlation (R) 0.722

The CO₂ component is the linear regression solution of the natural log of CO₂ concentration from Law Dome ice core data serie(1865-1978) scaled and appended to Mauna Loa Atmospheric CO2 record (1979-2000) vs the residual of the global intrumental temperature series minus the Solar Component above.

T_c=0.6318*ln(CO₂)-3.6324;
Stderror 0.17^oC,
Correlation (R) 0.25

Global Temperature Anomaly, ^oC
Instrumental Global Temperature(T), Jones et al. '01 (black solid line)
Reconstructed temperature (T_s + T_c) from linear regression components(red solid line)

CO₂ + Solar Temperature Anomaly Reconstruction, ^oC
CO₂ contribution to temperature (blue area)
Solar contribution to temperature anomaly (red area)

Note the relative contributions of Solar as compared to CO2 the last 50 years. Solar appears to be peaking out in its Gleisberg cycle while CO2 related factors appear to have established a small linear trend in temperature, at least so long as the exponential rise in concentration is maintained (not a very tenable long term projection).

Of more concern long term is a potential secular trend underlying Solar Activity that has been extracted from the ACRIM 1,2, & 3 satellite data:

Researcher Finds Solar Trend That Can Warm Climate
http://www.earthinstitute.columbia.edu/news/2003/story03-20-03.html

In this study, Willson, who is also Principal Investigator of the ACRIM experiments, compiled a TSI record of over 24 years by carefully piecing together the overlapping records. In order to construct a long-term dataset, Willson needed to bridge a two-year gap (1989-1991) between ACRIM1 and ACRIM2. Both the Nimbus7/ERB and ERBS measurements overlapped the ACRIM ‘gap.’ Using Nimbus7/ERB results produced a 0.05 percent per decade upward trend between solar minima, while ERBS results produced no trend. Until this study, the cause of this difference, and hence the validity of the TSI trend, was uncertain. Now, Willson has identified specific errors in the ERBS data responsible for the difference. The accurate long-term dataset therefore shows a significant positive trend (.05 percent per decade) in TSI between the solar minima of solar cycles 21 to 23 (1978 to present).

That "0.05 percent per decade" may not sound like much but over a centuries time that amounts to (.005*1367*0.7/4) 1.2wm-2 at 0.85oC/w climate sensitivity the IPCC modelers are telling everyone is out their, that is a full 1^oC to look forward to due to the sun alone with around 0.3^oC from exponential increase in CO2 concentrations.

Sure looks like that global warming of at least 1.0^oC + short term variations no matter what and an extra 0.3^oC if we don't manage to run out of exponentially increasing amounts of fossil fuel to burn over the next 100 years.

But wait +1.3 ^oC? Hmmm that puts us back some few thousand years ago.

Somehow I fail to be terribly worried about being able to raise grapes in greenland again. Especially looking at the patterns of past climate cycles of the Quaternary in which we find ourselves and that our progenitors had to go through:

Summary for discussion tomorrow; sorry for the delay.

If you would, a list of the primary datasets, Solar & GHG's including how he handled the effects of watervapor and varying cloud cover, he used in the studies and the period of time that was actually analyzed would be useful.

Also it is important to know the methodology used in the statistical analyis, was a full Principal Component Analysis using selected indicies used or are the papers merely comments on simple linear regressions?

I've been working on a more rigorous application of PCA analysis of the lean + satellite data and the CO2 concentrations used in the linear regressions above. The results are intriguing and don't support a strong role for forcing from anthropogenic CO2 at all. The PCA shows a strong influence of solar activity right on through 2001 with only a pause in increase of Solar activity during the 1950-70 time frame correlating with the Jones et al. instumental temperature record, and a nil role from the CO2 residual component that is not correlated with solar activity, (e.g. anthopogenic, and volcanic CO2). I'm currently looking for a volcanic index to help in the extraction of the anthopogenic siganl from the CO2 concentrations.

I'll do my best on the Stott, Jones, and Mitchell paper. Feel free to ask questions; this is an important paper, and consideration of it will go a long way toward resolving some of the many questions, accompanied by data, that you have posted. (By the way, I think you REALLY need to get a copy of this paper for yourself. Here's the email of the corresponding author: "peter.stott@metoffice.com" Ask for a reprint or the PDF emailed to you as an attachment.)

In any case; Introduction summarizs work to date, quite well. Discusses solar irradiance measurements. Lean estimates 2 W m-2 increase in solar irradiance between 1900-1950, largest sustained period of increase since Maunder Minimum. Cites all reconstructions (Lean; Hoyt and Schatten (HS); Solanki and Fligge). Discusses amplification mechanisms for solar forcing. Discusses Tett et al. 2002; "no evidence found that the model systematically underestimated the observed climatic response". In all cases, the "model" is the Hadley Centre Coupled Ocean-Atmosphere GCM (HadCM3).

Outlines method: they amplified solar and volcanic forcing and made a single simulation for each forcing.

Section 2 is entitled "Natural contributions to twentieth -century temperature change". Several simulations will be considered: GHG (4 runs, historical chgs in well-mixed GHGs). ANTHRO (4 runs, GHGs + sulfur emissions + effects on clouds + ozone chgs). NATURAL (4 runs, with chgs in stratospheric aerosol due to volcanoes and changes in solar irradiance, based on Lean 1995a. ALL (4 runs with include forcings in both ANTHRO and NATURAL).

This is followed by math and methodology. I can't really summarize. One result is that the warming signal from reduction in volcanic aerosol in the first half of the century is 0.1 K, cooling due to Agung + El Chichon + Pinatubo is 0.25 K, with estimated GHG warming of 1K over the 20th century.

Next: looking at possible underestimation of response to solar or volcanic forcing. Made three further HADCM3 simulations with enhanced forcings to get a clear signal of climate response. Simulations are:

10xLBB: Forced essentially by Lean et al. reconstruction.

10XHS: Forced by Hoyt and Schatten reconstruction (enhanced in the same way as 10xLBB).

5xVOL (forced by volcanic aerosol changes)

10xHS shows two distinct warming periods with cooler periods in 60s and 70s; 10xLBB shows more gradual/consistent warming trend through century. 5xVOL shows expected influences of Krakatoa, low volcanism 1920-1960, then Agung/Chichon/Pinatubo. Notes differences between HS and Lean reconstructions.

Next section is a "test for linearity". Next section is "decadal-mean attribution analysis", mainly an evaluation of scaling factors.

Next section is the "5-yr mean attribution analysis". Conclusion is that with whatever reconstruction is used, "the large-scale temperature response to changes in solar output appears to be underestimated by the model". This is consistent with other detection studies.

Next section is "Reconstructed anthropogenic and natural contributions to observed temperature changes". General warming trend through century due to GHGs; sulfate cooling important mid-century; volcanoes "relatively minor role" but detectable.

Analysis indicates that solar forcing is "likely to be proportionately more important in the first half than the second half of the twentieth century". In first half, solar warming is 0.29 K century, GHGs 0.27 K century.

In first half, HS reconstruction indicates 60% of warming due to solar, Lean 40%. Second half, GHG warming is 2.75x solar for HS, 6.35x solar for Lean. Over entire century, solar accounts for 16% of warming based on Lean, 36% of warming based on HS. [Those are important results.]

Next part discusses scaling factor differences, simulation comparisons. Interesting statement: "... it seems likely that our methodology erroneously overestimates the solar component and underestimates the greenhouse component of observed warming, as a result of the degeneracy between the patterns of response to these two forcings."

Very vital start to the next paragraph: "Amplifying the solar signal, in combination with the anthropogenic and volcanic signals, produces an improved fit to the observed large-scale temperature evolution during the twentieth century. Global warming observed over the past three decades is well reproduced in the ANTHRO ensemble alone, but the addition of an enhanced solar contribution improves the fit to early century warming." Then discusses some observed regional discrepancies.

Summary and discussion section: Most important conclusions have already been covered above. Noteworthy sentences:

1. "Our results imply that solar forcing had a greater impact on near-surface temperatures than simulated by HADCM3, and that previous attribution analyses may have underestimated the potential contribution of solar forcing to twentieth- century global warming." (Climatic processes could amplify surface temperature response by 1.34x-4.21x for LBB, 0.70x to 3.32x for HS).

Summarizing next paragraph, not quoting: Even with enhanced climate response to solar forcing, warming over last 50 years caused by increasing GHGs. GHGs probably caused more warming than observed, as some cooling was due to sulfate aerosols. Warming from solar forcing est. 16-36% of warming due to GHGs.

Next section prognosticates the future a bit; I'll skip that for now; I can summarize it subsequently if you're interested.

Final paragraph, interesting sentence: "Although our study indicates that there could be an enhanced global-scale temperature response to solar forcing, convincing evidence for a mechanism remains elusive." Mentions UV-ozone connection and possible changes to planetary waves and Hadley circulation; also mentions cosmic ray-cloud (or electrical!?) connections.

That should get you thinking; I should be able to continue next week with obvious interruptions expected.

I'll be going out for a good portion of the rest of today so I might not be able to respond quickly after this reply. To let you know where I am going with this the following is a general description of the methodology I am using to extract the solar related components of instrumental temperature for comparison with non-solar related CO2 and residuals such as el-nino and lesser effects.

There are a couple of things that are coming out of my own Principal Component Analysis, removing cross correlations between terms accumulates the direct and indirect forcings under each principal term.

The essence of the method aggregates the total effects of "Principal Components" into singular terms of a linear series.

Basically one starts with an component having the highest correlation with the subject signal, extract that signal from all other components under investigation, then goes to the component having the next lower correlation doing the same in turn. The object is to develop a series of varibles that are minimally correlated between each other from which an index may be developed to reflect the subject data series to be simulated.

Principle Component Analysis
http://goanna.cs.rmit.edu.au/~santhas/research/paper2/node9.html

Principle component analysis is one of the simplest multi-variate methods that can be useful in analyzing data sets involving high-dimension feature vectors. The object of the analysis is to take n-dimensional vectors (represented by n variables, X₁, X₂, ... X_n) and find combinations of these to produce indices Z₁, Z₂, ... Z_n that are uncorrelated. The lack of correlation is a useful property because it means that the indices are measuring different ``Hypothetical dimensions" of data. However, the indices are also ordered so that, Z₁ displays the largest amount of variation, Z₂ displays the second largest amount and so on. When doing a principle components analysis, there is always the hope that the variances of most of the indices will be so low as to be negligible. In that case the variation in the data set can be adequately described by the first few Z's with the variances that are not negligible. The best results are obtained when the original variables are highly correlated. If that is the case, then it is quite conceivable that 20 or 30 dimensions can be adequately represented by two or three principle components.
The i^th principle component is defined as;
Z_i = a_i1X₁ + a_i2X₂ + ... + a_inX_n
where a_ij are the coefficients of the i^th eigen vector of the corresponding correlation matrix.

While I do not have a matrix oriented spreadsheet available to me, it is not too difficult to adapt standard linear regression techiques to accomplish the same results when only looking at small number of independant variables for the index under study.

The magnitudes of the results I am achieving are giving somewhat stronger results for solar influence and lesser role for non-solar related CO2 concentrations than the studies you summarize above.

I suspect the basic difference being the PCA technique does not attempt to emulate a multiplication of the a particular level of forcing i.e. 2wm^-2 of say Solar Activity by an atmospheric component such as water vapor. PCA merely fits the independant component to the observed data thus incorporating any "climate senstitivity" to component forcing into the resulting coefficient.

Since the method removes any cross correlation from succeeding components to create independant variables, each term becomes unique and truly independant of all others each with primary forcing and related multiplies incorporated into the term coefficient.

The particlular series I am working with essentially works out to look something like:

T = aS + bX +R

Where

S = the solar component index,
X = CO₂ component index, [ln(CO2)] less any Solar correlation.
R = Uncorrelated residual from the linear regression of S & X with respect to the Subject temperature series T.

First pass indicates the order of the terms to be Solar first, CO2 takes the secondary position as a consequence of lower correlation with temperature as well as Solar activity having a positive strong causitive relation on CO2 concentration both from direct measurement and theoretical considerations. There is simply no avenue for earth based CO2 to have an effect on the measures of Solar activity (principally observations of sunspot count etc.)

Initial pass puts the 11 yr averaged Solar index alone in primary position with a correlation of >0.76 with respect to the Jones '98 instumental series.

The logrithm of raw CO2 concentration alone has a lesser correlation with respect to the Jones '98 instumental series.

CO2 is cross correlated with the Solar index by >0.82 requiring the removal of the solar related component from it before using it as a term in the index series.

That just means the the raw CO2 component has a strong dependency on solar factors. Not a tremendous surprise as one source of CO2 is from solar heating of the oceans causing an immediate rise in atmospheric CO2 concetrations as solar activity increases. Thus the Solar faction of CO2 concentrations must be removed from the raw CO2 term to meet the conditions of uncorrelated variables that define the Principal Component Analysis.

The CO2 term ends up having only geophysical(mainly volcanic) and biomass decay components(delayed solar relationships) and anthropogenic(fossil fuel consumption etc.) related concentrations while immediate solar related concentrations are removed.

Anyway, when I complete the necessary speadsheet and can pull out a numeric & graphical result to display the index , I'll be sure to pass it on to you ;O)

Looking over the summary you provided, my primary remark remains the same as I make concerming the inadequacy of the GCM's in general. They are not designed to adequately model the processes of solar variations and thus overstate the contributions of GHGs.

The study is fine, its primary result being to test the response of a Model to a set of solar forcings.

In using the Lean data to test the sensitivity of a "Model" to solar forcings, Lean is one of the most concervative of the solar forcings that can be used. The Lean data is scaled on the assumption that the Maunder Mininmum was the result of no more than a 0.25% change in solar activity. Thus the 2wm^-2 variation applied as forcing in the GCM. However, other studies estimate the variation in solar forcing to be substantially larger than that Lean(no pun intended ;o) 0.25% estimate.

ftp://ftp.ngdc.noaa.gov/paleo/climate_forcing/solar_variability/bard_irradiance.txt

2. Data from Figure 3. Reconstructed Solar Irradiance Scaled against Maunder Minimum Total Solar Irradiance reductions of

0.25% (Lean et al. 1995),
0.40% (Zhang et al. 1994, Solanki and Fligge, 1998)
0.55% (Cliver et al. 1998), and
0.65% (Reid 1997)

Using Zang, Solanki & Fligge, Cliver, or Reid the Solar forcing introduced into the GCM tested would have caused a substantially greater Solar response greatly reducing the CO2 contributions necessary to reflect historical instrumental record.

That is what PCA and Linear Regressions of Solar activity are showing and what a model that focuses on greenhouse effect with built in presumptions favoring GHG processes cannot adequately reflect.

A model can only reflect the apriori postulates of its programmers. If any physical processes are not adequately characterized in a model, all outputs are in question.

Essentially what the Stott, Jones, and Mitchell paper appears to achieve is test a particular climate model against a couple of the more conservative solar irradiance series available and declares it inadequate as regard solar forcing responses.

Conclusion is that with whatever reconstruction is used, "the large-scale temperature response to changes in solar output appears to be underestimated by the model". This is consistent with other detection studies."

That being the case, no conclusions can be drawn from that model as regards the accuracy of how it reflects real world processes. The model tested in the study must be viewed as inaccurate as a consequence of inadequate treatment of solar forcings. Thus one must conclude, because the GCM underestimates the response to solar forcings,the GCM must give excessive weight to minority GHG's to achieve an apparent fit to instrumental measurements.

In my view, I cannot accept the GCM's configured by the UN/IPCC teams as being much more than over characterised polynomials and incapable of anything more than interpolation between known global temperature data points. The individual components making up those interpolaters do not accurately reflect the response of true physical processes and thus are incapable of making accurate projections outside the range of the dataset it is fitted to.

The GCMs appear to make the mistake of adding terms to what has become little more than an ultra fancy polynomial regression forced to fit historical data by adding compensating arbitrary terms rather than emulating measured physical processes with known inputs. Works great as an interpolator between known data points, but don't expect anything reflecting reality outside the range of the fitted dataset.

I was hoping to have time to reply to you on this thread today, but I had more to do than expected. I have a lot of "consolidation" thoughts in my head at the moment, and I want to provide a well-composed response. Given what I was able to shove aside today, I may have a chance to do that next week. Hope your Christmas was merry and I hope your New Year will be happy (but you should hear back from me before then).

Greenland Ice Cap Is Melting, Raising Sea Level
Source: The Associated Press
Published: Jul 20, 2000 - 04:05 PM Author: By Paul Recer
Posted on 07/20/2000 14:37:50 PDT by Ms. AntiFeminazi
http://www.freerepublic.com/forum/a3977712e1941.htm