Michaels’ controversial testimony: Part III

by Judith Curry

Pat Michaels’ frequent collaborator, Chip Knappenberger, has written an extended post on MasterResource that is a follow up to Michaels’ testimony and Climate Etc. previous two threads.  Chip has given me permission to repost in entirety at Climate Etc., and Chip will be available for discussion.

Divvying Up the Warming

by Chip Knappenberger
December 15, 2010

In a MasterResource article a few months back, I walked everyone through a series of recent scientific findings and described how they cast new light on how the total amount of observed global warming to date could be divvied upon among various causes. I ultimately concluded that the high confidence that the IPCC (and later echoed by the EPA) placed on the statement that “Most of the observed increase in global average temperatures since the mid-20th century is very likely due to the observed increase in anthropogenic greenhouse gas concentrations” was misplaced.

This line of reasoning was recently incorporated into statements made by Dr. Patrick Michaels when testifying before the U.S. House of Representatives, Committee on Science and Technology, Subcommittee on Energy and Environment.

During the questions and answers portion of the hearing, one of the other panelists, Dr. Benjamin Santer, quickly objected and claimed that Pat was “wrong” because he didn’t take into account the cooling influence of aerosols when determining how much observed warming should be assigned to greenhouse gases.

A day or so following the testimony, Judith Curry hosted a discussion on her blog site Climate Etc. to further examine Michaels’ logic. In her remarks introducing the thread, she too suggested that Pat was “obliged” to include sulfates in the calculation. When I stepped in to offer additional explanation, RealClimate’s Gavin Schmidt commented that he hoped I was “kidding,” and John Nielsen-Gammon of Texas A&M commentedthat my explanation was “nonsense.”

So with all these erudite folks claiming that Pat Michaels and I are wrong, I figured I ought to take another look into the logic behind our conclusions.

Our Logic

First let’s get a couple of things out of the way up front. The argument about whether or not the inclusion of sulfates is required to arrive at a logically correct conclusion has nothing whatsoever to do with the veracity and/or applicability of the scientific papers from which I’ve drawn some numbers (see my earlier post for details about these findings). I am not suggesting that there isn’t plenty of room to argue that aspect of things, just that such a discussion does not impinge on the discussion of our logic. So I’ll set aside discussion of those issues in order to focus on the topic at hand.

I’ll state the following things simply as given (if it helps I’ll add this disclaimer “The following are for illustrative purposes only”):

1) The observed warming from 1950-2009 is 0.7°C
2) 0.2°C of that is due to a warm bias in the measurements
3) This leaves 0.5°C of actual warming
4) The anthropogenic increase in well-mixed greenhouse gases is responsible for +3.0 W/m2 of extra climate forcing (positive climate forcing imparts a warming pressure on global temperatures)
5) Sulfate aerosols are responsible for -1.5W/m2 of climate forcing (note this is a negative forcing which imparts a cooling pressure)
6) Black carbon (a.k.a. “soot”) aerosols are responsible for 1 W/m2 of added forcing (warming)
7) Stratospheric water vapor changes are responsible for 0.5W/m2 of added forcing (warming)

I’ll stress again, I use these numbers to make the math cleaner. They are in the ball park, but not intended to represent the exact or even best-guess values. So please set aside any heartburn about them (also note that natural variability is not considered here, but most definitely should be in a more formal evaluation).

Also, the IPCC was unaware of numbers 2, 6, and 7 at the time it made its “most of the observed warming” statement quoted above.

In my MasterResource article and in Pat’s testimony, the logic as to how to assign various amounts of observed warming to various factors is as follows:

The 0.5°C of actual warming [in (3)] was caused by three factors, GHGs, black carbon, and stratospheric H2O. In percentage terms, GHG contributed 67%, black carbon contributed 22% and stratospheric H2O contributed 11 percent. This is calculated by taking the positive forcing from each factor and dividing it by the sum of the positive forcings. Thus for GHG, you get 3.0/(3.0+1.0+0.5)=0.67, or 67%. To determine how much of the actual temperature change that GHGs were responsible for, we multiply 0.5°C by 67% and get 0.34°C—which, we pointed out was about 50% of the “observed” warming of 0.7°C (listed in (1)). Thus, the IPCC statement rests on thin ice.

To this, Santer/Schmidt/Nielsen-Gammon/Curry cried “foul!” claiming that we have committed a sin of omission by not factoring in the negative climate forcing contributed by sulfate aerosols.

Their Logic

When calculating the percentage contribution from each climate forcing agent, they maintain that we should have divided each element’s contribution not by the sum of the positive forcing, but by the sum of all forcings (including negative ones). So instead of dividing by (3.0+1.0+0.5)=4.5 like we did, we should have used (3.0+1.0+0.5-1.5)=3.0.

And, they argue, that had we done that, we would have found out that GHGs contribute 100% of the warming, black carbon 33%, stratospheric H2O 17% and sulfates -50%. Therefore, even taking into account measurement errors (listed in (2) above) GHGs still contribute more than half of the observed warming. And the IPCC is right and Michaels and I are wrong (see John Nielsen-Gammon’s explanation at Climate Etc. for further evidence of this line of reasoning).

It is now my turn to reply, “Nonsense!”

Pieces Greater than the Whole?

You can’t divide a physical quantity into pieces that together are greater than the whole. Which is precisely where the latter logic leads you.

The flaw in the Santer/Schmidt/Nielsen-Gammon/Curry logic is in confusing “potential” warming with “observed” or “actual” warming.

I completely agree that using the numbers above, GHGs contribute 0.5°C*100%=0.5°C of potential warming, black carbon contributes 0.5*33%=0.17°C of potential warming, and stratospheric H2O contributes 0.5*17%=0.08°C of potential warming and that of this 0.75°C of potential warming, sulfates offset 0.25°C of it, leaving 0.5°C ofobserved warming.

But, in employing potential warming to divvy up observed warming is mixing apples and oranges, and leads to results that don’t make practical sense.

Take for instance this hypothetical situation:

GHG = 2 W/m2
Black Carbon = 2 W/m2
Stratospheric H2O = 2 W/m2
Sulfates = -4 W/m2
Observed Warming = 0.5°C

If this were the situation, you would arrive at the answer that GHGs are responsible for 100% of the warming AND black carbon is responsible for 100% of the warming AND stratospheric H2O is responsible for 100% of the warming. This result allows you to assign any and all warming to whatever your favorite postive forcing element is. Certainly this is creative, but it is not practical.

Doing things my way, you get the logical result that each positive forcing element contributes 33.3% to the warming and is equally responsible for what has been observed.

So there you have it, two different ways of describing the observed warming.

I’ll leave it up to you all to decide which one is the more reasonable.

The IPCC’s Take

But before I go, I’ll leave with one additional thing to consider.

Here are two successive statements in the IPCC’s Summary for Policymakers of its Fourth Assessment Report (p. 10).

The first you will recognize:

“Most of the observed increase in global average temperatures since the mid-20th century is very likely due to the observed increase in anthropogenic greenhouse gas concentrations.”

And here is the separate bulleted statement that comes next:

“It is likely that increases in greenhouse gas concentrations alone would have caused more warming than observed because volcanic and anthropogenic aerosols have offset some warming that would have otherwise taken place.”

Clearly the IPCC recognizes the difference between observed warming (as reflected in the first statement) and potential warming (as reflected in the second statement).

Bottom Line

The latter IPCC statement will remain true so long as the value of the net climate forcing remains less than that of the GHG forcing (which would take a significant scientific finding to overturn). I think this is the gist of what Santer/Schmidt/Nielsen-Gammon/Curry are saying, and that such a major development has not occurred.

The former statement, however, and the one whose veracity is being assessed by Pat Michaels and I (and the one highlighted by the IPCC and the EPA), is more readily falsifiable through even relatively minor tweaks to our understanding of the various influences on the climate system. And Pat and I are claiming that such minor tweaks have occurred and have led to a falsification of the IPCC statement—or at the very least, demonstration that it is undeserving of the level of confidence placed upon it by the IPCC.

I welcome on-topic comments.


1cknappenberger { 12.15.10 at 11:10 am }

Maybe this more practical analogy will make things clearer:

Three apple growers deliver apples to an apple seller. Grower A delivers 300 lbs of apples, Grower B delivers 100 lbs of apples and Grower C delivers 50 lbs of apples. On the way to market, the apple sellers cart overturns, and of the 450 lbs of apples that he is taking to sell, 150 lbs are lost. He is able to sell the remaining 300 lbs for $1.67/lb taking in a total of $500.

The next day, the apples growers come to collect their share. Grower A says, since the apples sold for $1.67/lb and I contributed 300 lbs, you owe me $500. Grower B, multiplying $1.67/lb by his 100 lb contribution, asks for $167 and Grower C, using the same math, asks for $84. The apple seller responds that each claim is legitimate, but, that as a result of the losses, that he only has a total of $500 to distribute. Thus, the only equitable way to distribute the proceeds is to do so based upon the percentage contribution that each grower made to the total amount of apples that were originally intended to be sold. In this case, Grower A contributed 67%, Grower B contributed 22% and Grower C contributed 11%. Thus the seller paid out $335, $110, and $55 to Growers A, B, and C, respectively (to the dismay of Grower A, who continued to insist that he should receive the entire amount).

The percentage pay-out to each Grower is the same no matter how great the losses are, although the actual dollar amount that each receives will change depending on the total sales.

Thus, in determining the percentage of the pay-out, the size of the loss does not enter into the calculation.

In the same way, the negative forcing from sulfate aerosols does not enter into the calculation of how to apportion the observed global warming among the positive forcing agents.


2John N-G { 12.15.10 at 11:29 am }

Chip – You’ve noted the strangeness with the IPCC statement, but your formulation is at least as strange. Take, for instance, you hypothetical situation. Black carbon and sulfates are both aerosols. So your hypothetical is equivalent to:

GHG = 2 W/m2
Stratospheric H2O = 2 W/m2
Aerosols = -2 W/m2
Observed Warming = 0.5°C

By your accounting, you get the “logical” result that GHG are responsible for 50% of the observed warming, even though your earlier accounting with exactly the same hypothetical forcings (just lumped differently) came up with GHG = 33% of the observed warming.

So we have two possible interpretations of the IPCC statement. Both interpretations are unsatisfactory. The IPCC statement is arguably wrong if interpreted in one of the two unsatisfactory ways.

My bottom-line position is not that the IPCC is right and you are wrong. It is that the IPCC statement itself is not a useful way of framing the issue. It simply does not make sense to talk about a proportion of the warming attributable to a particular factor when there are both positive and negative factors.

I hope that, next time around, the IPCC avoids this confusion by simply providing their estimates of the amount of GHG forcing compared to the total amount of forcing, and their estimates of the net effects of external forcing compared to internal variability. Then we wouldn’t need to have this argument.

108 responses to “Michaels’ controversial testimony: Part III

  1. John N-G:

    Thanks for your comments.

    Taking your suggestion to the extreme, I guess we could just lump everything together under the label “anthropogenic forcing” and then the IPCC could have attached some level of confidence to some statement about how much “observed” warming was from changes to net anthropogenic forcing. In that case, our discussion would mostly be limited to the magnitude of natural contributions and how close the “observed” warming was to the “true” warming. But as the IPCC introduced a breakdown of the anthropogenic forcing agents (i.e. GHGs), then it seems like looking at the roles of the individual forcing elements (as best as we can) is fair game to bring into the discussion. So looking at BC separate from cooling aerosols seems legitimate approach.

    I agree with you that the wording of the IPCC has led to these discussions, but the IPCC apparently liked it a lot (as it is highlighted in the SPM) and it is widely repeated (for example, by the EPA). As such, it seems that it is worthy of examination (heck, Ben Santer said that it was, in fact, one the “central conclusions” IPCC findings).

    So, here we are!


  2. Item 7 – contribution of changes in stratospheric water vapor – is false, and I believe Pat Michaels should have known it to be false. Those changes actually introduced a net slight cooling effect. I made this point in both the first and second installment of this issue. To repeat:

    “In my various comments above, I indicated that Michaels misinterpreted the contributions of GHGs to the post-1950 warming, and that they do in fact appear to contribute more than half. This is the case even if we accept his claim that Solomon et al demonstrated a warming contribution from changes in stratospheric water vapor, which if true, would reduce the GHG contribution somewhat. However, it turns out that Solomon’s work suggests a slight net cooling rather than warming from changes in stratospheric water vapor, which would make the GHG contribution larger. This can be seen in Fig. 3c in Stratospheric Water Vapor

    • Fred,

      Try as I might, I can’t see how the Solomon results on stratospheric water vapor indicate a cooling during any period other than the 2000s. From 1980 (the beginning of her data) to the end of her record, her results indicate a warming in addition to that caused by well mixed greenhouse gases (see the blue line in her Figure 3).

      This is how I incorporated her results (as originally described here):

      Now consider the results of a paper published just a few weeks ago in Science magazine by Susan Solomon and colleagues. They report that variations in the water vapor content in the lower stratosphere (apparently largely unrelated to GHG changes) have a large influence on the rate of global temperature change for periods of a decade or more. In fact, since 1980 (the start of the data analyzed), an overall increase in stratospheric water vapor content as been responsible for perhaps 15% of the overall temperature increase.

      It is impossible from Solomon et al.’s analysis to know what went on prior to 1980, so, for lack of any other guidance, I’ll assume that no changes took place (or, that the net change was zero) from 1950 to 1980. I’ll then back the 15% warming influence from stratospheric water vapor changes since 1980…

      Seems reasonable, right?


      • Somehow the link to my original analysis at MasterResource got messed up, it should be:



      • Solomon’s work shows a small net cooling – see the figure. One can also get an idea from the abstract. Start with 100 percent, add 30 percent warming to get 130 percent. Then subtract 25 percent to bring the total down to 97 percent. You can’t selectively cherry pick the warming and ignore the subsequent cooling.

        Pat Michaels should have known this if he read Solomon’s work himself. It invalidates a substantial part of his argument.

      • In addition to Figure 3C, it’s also possible to see from Figure 3A that the trend since 1980 without regard to stratospheric water is more or less in the middle of the shaded area of the range given when the earlier warming from stratospheric water changes is followed by the subsequent cooling. In the case of this figure, it’s not clear whether there is any significant deviation – i.e., a miniscule net warming or cooling – but clearly the claims of a signicant net warming are invalid.

      • By “miniscule”, I mean that the middle of the range with water included deviates from the mean value without water by what from the figures looks like less than 0.1 W/m^2. Admittedly from Figure 3A, the difference could be in either direction, but it is too small to make a meaningful difference in estimating the role of anthropogenic greenhouse gases.

      • In your comment, Chip, you have also misinterpreted the blue lines in the figure. They are not the stratospheric temperature or water vapor concentration trend lines but rather the upper limits of the estimated trends, with the lower limit given by the red line. The trend is in the middle, and fails to show a net warming from the combined effects of earlier warming and later cooling.

      • Fred,

        How the results of Solomon et al. should be applied is a valuable discussion. Nowhere have I claimed that my application/interpretation is the only one possible. I did try to be reasonable. In my reading of the Solomon et al. paper, it seems like the blue line (a .5ppm increase per decade from 1980 to 2000) was supportable from observations (this why it was included by Solomon in the first place). To my interpretation, the forcing change indicated by the blue line is 0.14W/m2 on top of the ~0.85 from the GHGs and aerosols (the dotted black line), or just over 15% (.14/.85) (I used the 15% number in adjusting the temperature trend downward). Additionally, the climate model used by Solomon et al. was reported to have a climate sensitivity that was a bit less than average. As a result she suggested that that “effects of the stratospheric water vapor changes on the warming trends considered here could be greater by about 80% or smaller by about 40%.” I interpreted this as meaning that the average climate model response to the strat H2O changes would probably be a bit larger (i.e. the average of 80% greater and 40% smaller was a little but greater) (but I stuck with 15%). So, I thought my application/interpretation was not an extreme reading of Solomon et al.
        Discussion the finer points of the applications of the findings I relied upon is definitely required…but it is a different issue than the logic behind how to divvy up the positive influences on the observed warming. If I am wrong about the latter, then there is no real reason to get into the details of the latter, because my premise would be wrong, almost by default.

        I think once we get the appropriate logic squared away (and if it comes down in my favor), then the time would be ripe to discuss how the results of Solomon et al., Ramanathan and Carmichael, Thompson et al., and any other applicable findings could/should be applied.

        Thanks for your well-thought comments,


      • Hrhm. You’re both wrong.

        Fred, your ‘abstract math’ is wrong. The abstract talks about percentage changes in the trend. The trend becomes larger by 30% in the 1990s and smaller by 25% in the 2000s. You start with 100% in the 1990s and add 30% to get 130%. Then you start with a different 100% in the 2000s and subtract 25% to get 75%. But then there’s still the 1980s.

        Fred’s right that Figs. 3a and 3b are presented as ranges of forcing and temperature change, with the blue line representing the upper bound. The figures do not present the authors’ best guess for the 1980-2000 period but instead show a range of possibilities.

        Yet, in Fig. 3c, they use the upper bound value to show the effect of an increasing water vapor trend from 1980-2000 (blue bars). They’re not clear in Fig. 3c that this is an upper bound hypothetical, and that the lower bound is given by the black and red bars.

        But the net effect of water vapor on temperature from 1980-present is shown in Fig. 3b. Extracting those numbers (which Solomon et al. never do) says that the change in present temperature induced by the incorporation of stratospheric water vapor changes from 1980-present is between -0.03 degrees and +0.07 degrees. So Fred, I don’t see how this can be interpreted as a small net cooling unless you take the lower bound, which is just as bad as Chip taking the upper bound.

        Chip, you win the battle but lose the war. One of Solomon et al.’s sources for the increase in stratospheric water vapor from 1980-2000 is Forster and Shine (1999). Their abstract reads:

        The observed cooling of the lower stratosphere over the last two decades has been attributed, in previous studies, largely to a combination of stratospheric ozone loss and carbon dioxide increase, and as such it is meant to provide one of the best pieces of evidence for an anthropogenic cause to climate change. This study shows how increases in stratospheric water vapour, inferred from available observations, may be capable of causing as much of the observed cooling as ozone loss does; as the reasons for the stratospheric water vapour increase are neither fully understood nor well characterized, it shows that it remains uncertain whether the cooling of the lower stratosphere can yet be fully attributable to human influences. In addition, the changes in stratospheric water vapour may have contributed, since 1980, a radiative forcing which enhances that due to carbon dioxide alone by 40%.

        This paper was published well before IPCC AR4, and Chapter 2 (the chapter on radiative forcing) references it. So the IPCC knew about the possible radiative effect of stratospheric water vapor changes since 1980 when it made its statement. Aside from a new calculation of that forcing, the new thing that Solomon et al. bring to the table is the drop in radiative forcing around 2000.

        In other words, with respect to what was known by the IPCC when it made its statement, Solomon et al. adds a negative forcing, not a positive one.

      • John N-G,

        Thanks for the quote from Forster and Shine (1999). Indeed, I had not tracked things back that far. And yes, it seems like the IPCC would (should) have taken this into account already.

        But, my initial reason for including the results of Solomon et al. was that when I checked the IPCC’s table of forcings (Figure SPM.2) the forcing from stratospheric water vapor was attributed to CH4, and according to Solomon et al.:

        Methane oxidation increases stratospheric water vapor, but its contributions are small near the tropopause (29), the region of greatest impact for radiative forcing as shown in Fig. 2. This explains why studies in which methane oxidation is the only adopted source of increasing stratospheric water provide considerably smaller radiative forcings than those shown here.

        I interpreted this as meaning that the Solomon et al. results indicated a new (or greater) positive stratospheric water vapor forcing than considered by the IPCC.

        Perhaps I was mistaken in this interpretation? Or did the IPCC underestimate its impact?



      • Speculating without reading, the IPCC may have regarded it as a transient phenomenon rather than a forcing.

      • John N-G.,

        You are correct.

        In further reading the IPCC AR4 Chapter 2 section on stratospheric water vapor, I guess I should consider that the IPCC has already incorporated strat H2O into natural variability, and thus already into its consideration of influences on the observed warming other than GHGs.

        By that reasoning, I should remove the consideration of Solomon et al. from my current analysis. (Perhaps saving it for an investigation into producing a tally of the influence of the components of natural variability

        Thanks for helping me work through this.

        Does anyone disagree?


      • Actually, I now disagree!

        On second thought, perhaps I was a bit too hasty in agreeing to remove the impact of stratospheric H20 trends on observed warming trends…after all, it is an influence that is possibly apart from GHGs. I should just distinguish it from other forms of natural variability which are not readily assigned to a positive forcing value (e.g., PDO).

        So, even though the general results of Solomon et al. are not new, strat H2O still represents a quantifiable positive forcing that is (perhaps) not an anthropogenic GHG. And thus I can include it among the elements among which to divvy the observed warming.

        So I guess that gets us we back to discussing just how much of the observed warming since the mid-20th century to assign to strat H2O.


      • Chip – see my comment below, written before your comment above. I agree with you that the 1980-2000 warming can be included as part of the total temperature rise, even if the IPCC knew about it. Where I still disagree is with the inference of net non-anthropogenic warming of any significance, once the post-2000 cooling is also factored in. As I pointed out below, the balance between the two is very close to zero, and could be a very tiny net negative or positive cooling – far too small to make much a difference in assessing the anthropogenic fraction, and far smaller than what was implied in Pat Michaels’ testimony. Beyond this, though, you have to consider that some small fraction of stratospheric water changes are anthropogenic in nature from methane oxidation, and so even the warming interval presumably included a contribution from anthropogenic greenhouse gases. I conclude that on balance, it is likely that non-anthropogenic stratospheric water changes averaged over both intervals are at least as likely, if not more likely, to have had a small net cooling effect, but whether that is true or not, the emphasis should be on the word “small” rather than the direction of the effect.

      • Atmospheric methane rose about 130 ppb between 1980 and 2000, but of course, some of that rise (probably the lesser fraction) may have been non-anthropogenic.

      • Fred, John, Chip,
        Considering we are discussing about 60 years worth of warming with 30 years of data, 20 of which is very uncertain, do you think it is even reasonable to use this particular paper or the previous one for this purpose? The data really only suggests inter-decal variability superimposed upon other forcings anyway, right?

      • Here is an additional reference. It suggests that if all stratospheric water vapor changes can be considered as forcings (i.e., none as feedbacks), and non-anthropogenic in nature, the contribution to surface temperature rise since 1950 would have been about 0.15 C. That value is reduced if some of the change is a feedback response to increased surface warming, and the non-anthropogenic contribution is also reduced to the extent that anthropogenic methane contributed. The link is
        Forster and Shine-2002

        I derived that estimate from figure 3. The cooling after 2000 is not included in the analysis.

      • gryposaurus, Fred, John N-G.,

        I am coming around to the idea that over the longer run (e.g. since 1950) the influence of random strat H2O variability is probably small or at least difficult to currently quantify—and thus is probably left well-enough alone (at least from the standpoint of my analysis) until there is new insight.

        It seems like that is the general feeling of this (sub)discussion group. Thanks everyone for seeing this through.


      • Chip – Ooo, I won the argument for 61 minutes! Actually, this justifies the time I’ve spent on this topic…there’s no point in arguing with someone who isn’t willing to entertain the possibility that they’re wrong.

        I agree with you that strat H2O should fall into the same category as PDO. But the way I look at it, the IPCC had this big blob of possible sources of natural variability, with some opinion of how much natural variability was contributing to the observed warming based. Attributing the natural variability to one or more causes doesn’t affect the size of the overall big blob and shouldn’t change the allocation/proportion/attribution of the anthropogenic forcings.

        Hypothetical conversation:

        IPCC: “We think that somewhere around 0.0 to 0.2C of warming since mid-century was due to natural variability.”
        Forster or Solomon or Knappenberger: “You know, as much as 0.06C of that might have been due to variations of stratospheric water vapor.”
        IPCC: “OK, thanks.”

      • The arguments for Stratospheric water vapour(swv) are well entrenched in the literature,and the problematic issue of the extrapolation for doubled co2 is that the forcing is nearly equivalent to co2 forcing
        eg Joshi et al 2010

        in the radiative forcing associated with doubling CO2 from pre-industrial concentrations (in HadCM3) is 3.75Wm^−2. If the extra downward LW effect associated with SWV in the LEP2 experiment is 2.8Wm^−2, this will almost double the total radiative forcing

        Further pg 7166

        …A scenario that should be considered is whether the high temperature response in LEP2 might occur in reality because of a real change in convective entrainment or other processes that significantly increase SWV in a warmer climate. There
        has indeed been an increasing trend in stratospheric humidity over the latter half of the 20th century, which is thought to be climatically significant (Forster and Shine, 2002; Solomon et al., 2010). However, the trend is noisy (e.g.: Rosenlof et al., 2001), has many possible causes not related to climate warming (e.g.: Scaife et al., 2003; Joshi and Shine, 2003),
        and at present is hard to attribute (Fueglistaler and Haynes, 2005). In addition, stratospheric water vapour values decreased from the late 1990s to the early 2000s and there is no evidence of a positive trend since the year 2000 (Randel et al., 2006).
        Since LEP2 exhibits a radiative effect from the change in SWV that is about 80% of the CO2 forcing, one might expect that the radiative forcing associated with observed SWV changes since pre-industrial times should be a significant fraction of the 1.6Wm−2 associated with CO2 since 1860,if the real world behaved like LEP. Forster and Shine (2002)
        estimated a value of only 0.29Wm−2 for stratospheric water forcing in the 20th century, and this was based on the peak trend, which has now lessened.

        Indeed the wmo expert assessment (in press) finds

        Chemistry-climate models predict increases of stratospheric water vapor, but confidence in these predictions is low. Confidence is low since these same models: 1) have a poor representation of the seasonal cycle in tropical tropopause temperatures
        (which control global stratospheric water vapor abundances) and 2) cannot reproduce past changes in stratospheric water vapor abundances.

        Tropical lower stratospheric water vapor amounts decreased by roughly 0.5 parts per million (ppm) around 2000 and remained low through 2009. This followed an apparent but uncertain increase in stratospheric water vapor amounts from 1980–2000.
        The mechanisms driving long-term changes in stratospheric water vapor are not well understood.

        The bottom line eg Joshi et al is quite succinct

        This analysis has again shown that changes to minor constituents in the stratosphere can have profound effects on the evolution of the surface climate in models. Any future metrics of model behaviour should take account of potential biases
        arising from this region of the atmosphere, especially if
        the stratosphere is poorly resolved as is the case in HadSM3.


      • John – Thanks. When it comes to details, all three of us have been somewhat imprecise. The 3o percent increase, according to the text (but not the abstract) represents an upper bound, not a mean, which is closer to 15 percent per decade from 1980-2000. The 25 percent decrement is per decade since 2000. The net effect is very close to zero, so I should not have declared it a net cooling. It might deviate in either direction, depending on the statistics. If the scatter of points at the high end is greater than for the lower values (often the case), the mean might well be below the halfway point between upper and lower bounds, and would be consistent with a tiny net cooling. If the mean is exactly in the middle, it would suggest a tiny net warming. To me, the more important point is that the effect is much smaller than implied by Pat Michaels’ testimony and would have very little relevance to the anthropogenic contribution.

      • Fred – I agree with your last point, but not your 30/2=15. The abstract lists a 30% (upper bound) change in the rate of temperature rise for the 1990s alone. By comparing the first black bar in Fig 3c with the first blue bar, I get a 28% change (upper bound) in the rate of temperature rise for the 1980s.

        Oversimplified example. If the slope of straight line A is larger than the slope of straight line B by 30% over half its length, it’s larger by 30% everywhere. You’re thinking of the slope in terms of “rise” rather than “rise over run”.

      • I was trying to say that the 30 percent represented the slope of the upper bound of the trend line, where the zero slope is defined not as the horizontal but rather the slope of the trend in the absence of water vapor, and the ~15 percent represented the slope of the mean trend line vs the no-water trend. (Did I confuse everyone? In terms of rise over run, I was saying that the run is the no-water trend line.) I agree that the slope would be the same in each of the two decades.

  3. The observed warming corresponds to the net forcing. Thus in your basic case 0,5K corresponds to 3 W/m^2. This is true both for the total effect and to the GHG warming alone as both are 3 W/m^2. This is significant. Semantic arguments on percantage shares are irrelevant.

  4. If you compare GHG influence with observed then you compare 3 to 3 or 0.5 to 0.5 (with your numbers). The other possibility is to compare GHG effect to supposed sum of warming factors. Then 3/4.5 = 67% or 0.5/0.75 = 67%, but this is not a comparison to observed warming but to hypothetical total warming of 0.75K.

    • But Pekka, there are influences other than GHGs that have been warming the climate. Of the 0.5C of observed warming, how much did those other elements, e.g., BC and strat H2O contribute?


      • Chip,
        According to your numbers GHG contribute +0.5K, sulfate aerosols -0.25K, black carbon +0.33K and stratospheric water +0.17K. Thus all warming contributions add to +0.75K (or the thing I called hypothetical total warming) which is reduced to +0.5K by the single cooling contribution.

        This is your stated model. If you want to compare to observed warming, you cannot avoid dividing by 0.5K and getting percentages on that basis.

      • Pekka – that makes no sense. If we are considering factors that contribute to warming, we considerer only those that warm. The total WARMING is 0.75K, so that would be used for the total warming. This way the warming factor percentages total 100%, and we can see the relative contribution of each. Doing it the other way seems illogical and possibly politically motivated.

      • You may compare to the total, but then you do not get shares of observed warming but shares of the total of all factors that contribute warming.

        In addition this may be ill defined, as some factors may include both positive and negative subfactors. If all aerosols are combined then sulfates and black carbon add up to a smaller total and the percentages change.

        Whether you like it or not, the most consistent choice seems to be the comparison of each component separately to the net warming – or not calculating any percentages at all.

      • If you want to derive numbers that represent the relative effect on temperature where some are positive and some negative as we have here, then take the absolute value of each and sum. This gives a total of 1.25 and % for GHG of 40% influence on temp.

      • Pekka,

        I agree with your numbers for potential warming.

        At issue, though, is how to divide up the amount of observed warming.

        These are two different things. And, I argue, that the IPCC recognizes this is the case. If the IPCC had strictly adhered to your logic, then its statement that

        It is likely that increases in greenhouse gas concentrations alone would have caused more warming than observed because volcanic and anthropogenic aerosols have offset some warming that would otherwise have taken place.

        would have sufficed and been their highlighted finding. But, instead, they decided to offer up something else, and draw more attention to this:

        Most of the observed increase in global average temperature since the mid-20th century is very likely due to the observed increase in anthropogenic greenhouse gas concentrations. …

        “Most” is obviously less than “more…than observed.” So, to me at least, they are recognizing a difference between contributions to potential warming (in the first quote) and contributions to observed warming (in the second quote). Otherwise, there is no need for the second quote.

        Clearly, they hold the idea that something other than GHGs has also contributed some part of the observed warming. I am arguing that new findings indicate that the non-GHG contribution to the IPCC’s “observed” warming is larger than the IPCC recognized.


  5. “also note that natural variability is not considered here, but most definitely should be in a more formal evaluation”. Quite so; unless you know natural variability, all the rest is just guesswork.

  6. If it is acknowledged as a given that 40% of the alleged change is bias, how do you know additioanl noise from imprecision and data base merging is not another 20% or 40%?
    And is temperature even the most significant metric to determine dangerous climate change? What about heat content of oceans? What about the manifestations of cliamte- weather?
    As to carbon balck vs. other aerosols, it seems that carbon black is in a category by itself. Shoud it be lumped together or kept seperate?

    • Hunter,

      In theory, the IPCC took into account all the things that it knew about (like weather variability) when assigning confidence (“very likely”) and percentages (“most”) to the GHG contribution to “observed” warming (which was in reference to the global surface temperatures).

      Many of the things that I have looked at, the IPCC did not know about, as they have been published subsequent to the AR4. In light of these new findings, it seems reasonable to re-assess the original IPCC statement.

      As you point out, there are other perhaps other considerations as well.

      And, yes, I think that BC should be separated out from the cooling aerosols.


  7. I think it’s worth noting that if we consider Chip’s illustrative numbers as correct, the statement “Most of the observed increase in global average temperatures is due to the observed increase in anthropogenic greenhouse gas concentrations” is correct even under Chip’s interpretation of the math. With biases corrected, the observed increase is 0.5C, and anthropogenic greenhouse gases provided 2/3 of the positive forcing.

    Chip’s argument is that, based on his illustrative numbers, the statement, while correct now, was incorrect two years ago. So if you like a good semantic argument, proceed.

    Since I like a good semantic argument, I will proceed. I’ll even use Chip’s illustrative numbers as well as his interpretation of how to do the math. Even given all this, I will argue that the IPCC statement was correct two years ago.

    The statement was: “Most of the observed increase in global average temperatures since the mid-20th century is very likely due to the observed increase in anthropogenic greenhouse gas concentrations”.

    Since then, Chip says, we’ve found a +0.2C trend bias, plus a 1.0 W/m2 black carbon positive forcing, plus a 0.5 W/m2 stratospheric water vapor forcing. These three things, Chip says, were just enough to knock the AGG contribution to below the 50% line.

    What were the odds of that? Imagine, three major new climate factors discovered in two years! And all three reduce the apparent contribution of AGG! The chances of all three having an effect of the same sign is 1 in 8. And furthermore, the magnitudes of the factors were just barely large enough! I suppose the chances of that happening were maybe 1 in 2. So collectively, the odds of enough things coming along to demonstrate that AGG was not responsible for most of the observed warming at the time was 1/8 x 1/2, or 1/16. So the IPCC statement is correct as long as they include the 6% chance of newer findings all lining up with the right sign and the right magnitude.

    What does “very likely” mean? Greater than 90%.

    So even using Chip’s illustrative numbers and Chip’s logic, the IPCC statement would be correct when it was made.

    Notwithstanding the fact that, using Chip’s illustrative numbers and Chip’s logic, the IPCC statement would be correct if made today.

    • John N-G,

      How is chance involved in the three new scientific findings I cited? I can see how there is a chance that Ramanathan and Carmichael’s value for positive BC forcing is too high (but isn’t there a chance that it is too low as well?), and there is a chance that Solomon et al.’s strat H2O is related to GHG increases, or that it showed a different trend prior to 1980, or that its warming influence since 1980 is less than I assumed, and there is a chance that the factors identified by Thompson et al. will have little influence on the trend since 1950, etc. Surely, different assumptions than those that I made will lead to different results…some leading to more non-GHG warming, some leading to less. But this is not chance as you applied it.

      I agree that there is some semantics involved, especially when it comes to the IPCC’s “observed” warming. In fact, to apply the results of Thompson et al., I have to draw a distinction between “apparent” (i.e. observed) warming and “actual” (i.e. true, but perhaps unknowable) warming (but note that I am not the first to draw this distinction).

      Bottom line, is that I think, in IPCC semantics, that “more likely than not, > 50% probability”, or even better “about as likely as not, 33 to 66% probability)” is a more appropriate descriptor (given what we know now…which may change tomorrow…by chance, I guess. :^)


  8. I believe I understand Chips argument. I’m sure others will correct me if I’m wrong.

    GHG’s can only do “their work” if radiation from the sun arrives at the surface. However sulphates scatter some of that radiation so ‘that some’ never reaches the surface, i.e. GHG warming never reaches it’s potential.
    Just like the apples that never reached the market to be exchanged for cash.
    It makes perfectly good sense to me.

  9. Chip, how much is normal variation a contributor in this? You explicity did not consider it, but would be interested to know if random variation would override some or most of the forcing.

    Here is an example of what I mean. Take 10 different faced dice and roll them at different frequencies. The number of dice rolled would be randomly generated, where the seed to the random number comes from the previous outcome of rolled dice. Some times 2 dice get thrown, sometimes 7, occasional 10. Being different faces (some with 6, some with 8, some with 12) would influence the total count of all the dice numbers added up. Over a long period of time there would be a heartbeat like pattern to the total addition of the faces. There might even be a short term trend to an increase in the count, but then drop back down.

    No forcing happening, just random fluctuations like waves on a pond intersecting. Several crests come together to produce a high crest than normal (a warm spell), several troughs come together and make a deep trough (a cold spell).

    Linking this to the climate and what we have is an endless number of pendulum swings of different frequencies, from a day long up to a century or more. All swinging “left” then back “right”, interacting to produce wide variation in the climate. No forcing doing this part, just normal random variation of pendulum swings.

    • jrwakemen,

      Sure, natural variability has an influence on the value of the observed trend. I assumed that the IPCC already had taken this into account when ascribing “most” of the observed warming to GHGs, thus I did not further explore it.


      • So no one really knows how much such random swings account for what we are seeing. Wouldn’t that be a bit of a priority to find out? Is it even possible to find out?

      • Based on the phrasing in the AR4, and on what I have read elsewhere, I assume only volcanic activity was the only “known” natural variance. It is safe to assume natural variability as one of the things that was not known when AR4 was written, as it is still unknown today. Please correct me if I am wrong.

        Out of curiosity I looked at what the numbers would be if we chose the max-min variability from the early part of the recorded temperature measurements; the time before AGV. A quick look at the temperature record from 1892-1942 would show roughly a 0,48C warming (max), but from 1898-1948 the “natural” warming would be only 0,08C (min). (As above in the calculations Chip made, only for having some numbers to play with, I looked at a 5 year average graph on Wikipedia, so the numbers are all but accurate).
        If we assume the natural warming from 1950-2000 is the same as it was from 1892-1942 (max), then natural warming would account for 96% of the warming? (0,48C/0,5C)
        If we assume the natural warming from 1950-2000 is the same as it was from 1898-1948 (min), then natural warming would only account for 16% of the warming? (0,08C/0,5C)
        So if we assume the above, what would be warming contribution% of CO2 in the max and min natural variability cases? How should natural variablity be handled here?

  10. I agree with Chip’s logic on this. Unless there is some reason to do so, you cannot say that the negative forcings only cancel the effect of black carbon and stratospheric H2O but do not affect the forcing of CO2. The impact of the negative forcings should proportionally cancel a portion of each of the positive forcings. In other words, a positive forcings percentage contribution to observed warming will always be its percentage of the total positive forcing regardless of the affects of negative forcings (unless there is some physical reason why the negative forcing would preferentially cancel out one positive forcing over the others).

    Nonetheless, as noted by John N-G, in Chip’s example, 67% (which fits at least one definition of most, i.e., “the majority”) of the warming could be attibuted to GHGs. Whether we should have a 90%+ confidence in this, I do not know.

  11. Sorry, but as this is an internal debate among those who accept the IPCC logic, and the science as settled except for small details, but who still disagree on how best to express it proportionately, it reads to me like AGW angels dancing on the head of a pin. Mind you the dance is quite elegant.

  12. Just a heads-up.
    Originally this thread appeared on Judith’s front page. I just came back and find it is no longer appears there. I tried two different browsers with no joy. It is still accessible from the recent comments list.

    • Brandon Shollenberger

      I don’ know why you experienced that, but this thread appears on the front page for me. Maybe it was just a temporary issue.

  13. A similar energy balance analysis, but then based on numbers from Ramanathan and Feng (2009) would look as follows:

    Global average surface temperatures have increased by about 0.75 degrees Celsius since the beginning of the industrial revolution, of which ~0.6 °C is attributable to human activities. The total radiative forcing by greenhouse gases is around 3 W/m2, with which we have ‘committed’ the planet to warm up by 2.4 °C (1.6-3.6 °C), according to a climate sensitivity of 3 °C (2-4.5 °C) per doubling of CO2. The observed amount of warming thus far has been less than this, because part of the excess energy is stored in the oceans (amounting to ~0.5 °C), and the remainder (~1.3 °C) has been masked by the net cooling effect of anthropogenic aerosols.

    I.e. Earth has warmed approximately as much as expected, though uncertainties in some of these numbers (esp sensivity and aerosol forcing) are rather large.

    I highy doubt Chip’s numbers for items 2,3, 6 and 7. What sources did you use for these strong forcings from BC and strat water vapor?

    • Bart,

      My forcing from BC comes from Ramanathan and Carmichael (2008).

      And it looks like we may have solved the strat H2O issue in a discussion above (and that the IPCC has already considered it as natural variability).


      • Chip,

        Thanks for the ref (which I really should have looked up before firing off my question…)

        In that paper they write:
        “At the TOA, the ABC [atmospheric browen cloud, i.e. aerosol] (that is, BC + non-BC) forcing of –1.4 W m–2
        (sum of TOA values in Figs 2c,d), which includes a –1 W m–2
        indirect forcing, may have masked as much as 50% (±25%) of
        the global forcing due to GHGs.”

        I.e. you take their BC forcing, but not their net forcing, which would lead to the numbers I just used (based on a later paper by the same first author, Ramanathan). There seems to be enough wiggle room / uncertainty to use a broad variety of numbers, which underscores the conclusion that within that broad range of plausible values, there is not disconnect between the observed and theoretically calculated warmiong over the 20th century. Or in other words, the 20th century does not impose a strong constraint on the climate sensitivity. Stronger constraints are found from periods from the past (e.g. glacial-interglacial) or better yet, combining several constraints (e.g. Annan and Hargreaves, 2006).

  14. We have been told, time and time again that ice is retreating, yet you do not include any temperature change due to albedo. I search the thread for albedo and it is not there. That has to be wrong. If ice is retreating, the albedo is decreasing and that is causing warming. You don’t have it right without albedo.

    • Alex,

      Thanks for your comment.

      Certainly changes in albedo influence the earth surface temperature. But, presumably the ongoing changes in albedo are caused by the increases in positive climate forcings that we have been discussing, and not random variability, and thus are being accounted for. The exception to this may be that black carbon deposition onto snow and ice may further hasten the albedo decline thus warming the earth to a greater amount than I have currently assigned to BC.


  15. “(to the dismay of Grower A, who continued to insist that he should receive the entire amount).”

    Thanks Chris, that made me laugh out loud.

    Another forcing that seems to have been left out in the cold is the sunshine which made the apples ripe enough to take to market. The increase in the level of solar activity in the C20th reflected in the increase in ocean heat content needs to be factored in. The drop in OHC since 2003 shows the forcing on the ocean was solar in origin, not from increased co2, which is still increasing as the ocean cools.

    If the co2 hypothesisers still believe it was co2 which warmed the ocean, and it is natural variability which has caused it to cool for the last seven years, they need to tell us what natural variable they believe is responsible, and how much it contributed to the warming in its positive phase, since it obviously strong enough to overcome the mighty warming power of co2.

    • Tallbloke

      It’s not only the drop in ocean heat content since 2003 which suggests that the sun rather than CO2 is what heats the oceans. Comparing the sea surface temperature record with almost any solar reconstruction shows that the +/- 0.5C of observed sea surface warming between 1910 and 1950 correlates with a period of rapidly increasing solar irradiance. There’s no way this warming could have been caused by CO2, or at least none that I can see.

  16. I think John-NG’s argument settles it. At least forcing due to CO2 and total net forcing are physical quantities that notionally can be measured. And you can express their ratio as a percentage if you want. If that carries an implication that the total consisted only of positive quantities, that can be corrected. But we’re used to that. If I was describing my sources of income, some might be nett, like share trading. If I said that salary was 80 % of total, I wouldn’t mean a total including total sales of shares (but not purchase costs).

    But the sum of the positive components of forcings is arbitrary. By reclassifying you can get any figure you want. And that can’t be rescued by using clarifying words.

    Chip’s reasoning surfaces in a more extreme form in the sceptic argument that fossil fuel burning is insignificant because it is only 3% of “total” CO2 emissions. To get that, they add to the total the emission part of big cyclic processes like photosynthesis/respiration and ocean exchanges. And with ocean exchange, if you got down to a finer level, you could probably make the denominator as large as you like.

    • Nick,

      The temperature change of the earth is a physical quantity that can notionally be measured.

      So, too, is the pie on the table in front of me. When my wife pulls up a chair and the dogs gather round, I know I am not going to get to eat the whole thing…even though I could if they weren’t there. Thus, even though I could eat 100% of the pie (and probably even then some), ultimately, I only get to enjoy some portion of it that is less than 100%.

      And BTW, this argument is not at all related to the 3% of emissions issue (which of course is an improper notion when explaining why atmospheric CO2 elvels are increasing).


  17. My approach does not try to remove all the factors. I agree with Chip’s points 1-3. Then, I argue that the PDO type oscillation (which I believe to be solar forced, perhaps by cosmic rays/clouds and perhaps merely by kicking) on a 60 yr period lines up nicely to account for perhaps .2 to .3 deg C of the period warming from 1980 to 2000. The instrument bias plus the 60 yr cycle then add to .4 to .5 out of .7 deg–whatever else was going on. This only leaves .2 to .3 deg C to be explained by the other forcings. Not so alarming. I have a paper submitted about this (with the usual headaches…).

    • Keep pushing Craig. The turn of the solar forced ocean will come. Nir Shaviv got the ‘using the oceans as a calorimeter’ paper published in JGR.

    • Dr. Loehle,

      Based on Tsonis’ work your 0.2 to 0.3 C for other (anthropogenic) forcing is probably in the ballpark (likely on the high side, but not by much). I doubt that solar influence has much to do with the PDO though. I recommend you avoid older TSI reconstructions references in your paper.


  18. <i<"This only leaves .2 to .3 deg C to be explained by the other forcings. Not so alarming. "

    Disregarding for the moment whether it’s 0.2 to 0.3 deg or 0.4 to 0.5 deg, and also disregarding that this response to forcing has distance to go before reaching equilibrium, there is still no reason to assume that the trend will continue to be “not so alarming”. The trend is the resultant of both positive and negative anthropogenic forcings, but their balance can’t continue at the same level. As Stephen Schwartz et al put it in an article in NatureReportsClimateChange:

    “The century-long lifetime of atmospheric CO2 and the anticipated future decline in atmospheric aerosols mean that greenhouse gases will inevitably emerge as the dominant forcing of climate change, and in the absence of a draconian reduction in emissions, this forcing will be large.”

    As you probably know, Schwartz is among the climate scientists who tend to estimate climate sensitivity towards the low end of the range, but his point is that an unmasking effect of aerosol reduction can have significant consequences even if sensitivity is below the value of the average estimate.

    • Fred says: “there is still no reason to assume that the trend will continue to be “not so alarming”.” I am sorry but it is up to model proponents to show that it IS alarming. If half of the warming observed or more is NOT anthropogenic, then the climate sensitivity and feedbacks may be less than claimed. Furthermore, this “sulfate aerosol” negative forcing is a fishy number. Data to back it up are very weak, and the different models use different levels of this forcing, as noted by Schiermeier [2010].
      Schiermeier, Q. (2010), The real holes in climate science, Nature, 463, 284-287.

      • Richard S Courtney

        Craig Loehle:

        You say;
        “Furthermore, this “sulfate aerosol” negative forcing is a fishy number. Data to back it up are very weak, and the different models use different levels of this forcing, as noted by Schiermeier [2010].
        Schiermeier, Q. (2010), The real holes in climate science, Nature, 463, 284-287.”

        “Fishy”? Oh, it is much more interesting than merely ‘fishy’.

        In a thread on this blog, Fred Moulten and I recently discussed how this “number’s” piscene nature is used in climate models.

        I commend reading that discussion to all who are interested in numerical piscatology. It is at


    • As Stephen Schwartz et al put it in an article in NatureReportsClimateChange:
      “The century-long lifetime of atmospheric CO2

      In view of the favorite denier statistic that anthropogenic CO2 is only 3% of the carbon cycle (true enough), is it possible to make any connection at all between the “lifetime” of atmospheric CO2 (however defined) and its mixing ratio (currently 0.39‰), other than in the steady-state business-as-usual scenario where we continue to double anthropogenic CO2 every three decades or so as we seem to have been doing for more than a century?

  19. Found in just two years
    Three unknown factors working.
    Perhaps there are more?

  20. What’s up Doc (Curry)?

    I agree with you for what it is worth. With 3.0, 1.0, 0.5 and -1.5 the total forcing is 6. GHG would be 50 percent (GHG forcing / Total forcing) of the total forcing. Of positive forcing, GHG would contribute 3/4.5 or 67% and aerosols would contribute 100% of the negative forcing.

    What you did was correctly determine GHG’s contribution to the positive forcing (warming). Whatever your detractors was calculating was more than vague, it was incorrect.

    Nice blog by the way. Thanks for the chuckles.


  21. I think the public would understand the raw data here, i.e. a histogram of each warming component separately that includes positive and negative bars, and these add up to a total. It is nonsense to try to turn such a histogram into percentages of one arbitrary column compared to a sum of some other chosen columns. This is not at all helpful, and only confuses the matter. Why not just the raw histogram?
    This would show the CO2 bar actually higher than the total bar, and would give a correct impression of its effect.

    Taking a total, and dividing the warming parts into it is almost exactly like the old fallacy where you say you work 33% of a day which is 120 days a year, but then you subtract 104 weekend days leaving only 16 days per year, and you subtract holidays, and find you hardly work at all.

  22. One quibble I have is the distinction you make between “actual” and “observed” warming due to the supposd warm bias of 0.2% in the measurements, so the 67% contribution of GHG’s to the actual warming becomes a < 50% contribution to the observed warming.
    As you say your figures are for illustrative purposes only I won't argue the point of whether this "warm bias" actually exists, but assuming it does and we have good reason to believe the 0.2C figure is accurate then we should surely make our assessments based on the "actual" figure not the "observed" figure.
    I mean if a figure is known to be incorrect due to measurement error then how can you use it to make meaningful calculations?

  23. due to the supposd warm bias of 0.2%

    should of course read 0.2C

  24. Chip,
    In the AGW debate, motion is not considered a factor.
    Just absorbsion or deflection in molecules of CO2 or H2O. H2O changes with temperature into a crystaline form if the temperture in the atmopshere is below freezing. H2O vapour can be in a small patch or cover many miles. It can be light in density or heavy in density when ready to release precipitation. This has an effect of how much solar radiation actually hits the planets surface.
    Now motion (not considered) is the planet turning and the deflection and angles of solar radiation going through the atmosphere.
    Temperatures make a poor diagnosis of planet warming and cooling when considering it is 4.5 billions years old and has faced far worse challenges than humans. The shape of our globe is not factored in as well. As far as current science is concerned, we may as well be measuring temperatures off a map.

    • Chip,
      Atmospheric pressure build-up has never been look at as well. As far as science is concerned, just the greenhouse theory is ONLY considered.

      • Chip,
        Also not considered is the speed of systems moving on the planet surface. I am finding that these systems are slowing down generating massive precipitation dumps. Also the winds are slowing too which effects the break-up of cloud formations.

  25. The average ‘global temperature’ for the 40 years from 1950 to 1989 is 1.5° below the average for the 11 yrs. from 1990 to 2000.
    In 1990-91, inclusion of data from ~60% of the reporting stations was (permanently) discontinued.
    The coincidence beggars belief.

    • Just to be explicit, as you will see if you inspect the linked graph, the temperatures were trending down, if anything, ’50 – 89. Then there was a step jump of 2°, with a drop to the upper end of the previous range, surges to new heights in ’98, and then starts back down.

      Visually there appears also to be a strong negative correlation year-to-year to the number of stations, which shows up as that 1.5° difference for a drop from about 14,000 to about 6,000. So the 57% cut in stations pays off at about a 1.3°C rise in average temperature per halving of station count. We can thus expect that “warming trend” by the time the stations are cut to 3,000. The current slope is a cut of about 100/yr., with the current count around 5,500. So we should expect GISS et al. to report (project) a 1.3°C rise within 25 yrs (starting from 2001), for a rate of 5.2°C/century. Using the log system from 2025 ff., you start with 3,000 stations, multiply by 2^-3, and get 375 stations by 2100 to achieve this 5.2° warming!

      But I’m sure they’ll be very good stations with very wide coverage. ;)

      But it could be worked backwards, too, of course. To hold the temp rise to a nice officially approved 1.3°C, just stabilize the number of stations at 3,000 in 2025. Much cheaper than gutting the world energy economy.

      There! All fixed!

      Which is very close to what they do in fact project. To

      • Oops, that broken sentence at the end is a rewrite edit error. S/B after “warming” and the “To” dropped.

  26. Richard S Courtney

    This argument is about words, their meaning, and how they should be interpreted. And, as Chip Knappenberger says, the words being considered are two staements from the AR4; i.e.

    “Most of the observed increase in global average temperatures since the mid-20th century is very likely due to the observed increase in anthropogenic greenhouse gas concentrations.”


    “It is likely that increases in greenhouse gas concentrations alone would have caused more warming than observed because volcanic and anthropogenic aerosols have offset some warming that would have otherwise taken place.”

    Much of the discussion has been concerning the emphasis to be placed on one or the other of these statements. But I submit that the word of most importance in this disccusion is the first word in the first statement; viz. “Most”.

    Clearly “most” has to mean ‘more than half’. And ‘most’ is applied to “of the observed increase in global average temperatures since the mid-20th century”. So, the first statement is an assertion concerning ‘more than half’ of the ‘observed increase’ over the most recent ~60 years.

    That assertion is that the ‘more than half’ is “very likely” to have been “due to the observed increase in anthropogenic greenhouse gas concentrations”.

    It seems that the word “due” is intended to mean “caused by”.

    The second statement provides emphasis to the consideration of the ‘observed increase’ because it suggests there could have been even more of a rise than was observed if the total increase had not been offset (by e.g. volcanic and anthropogenic aerosols).

    Hence, the only relevant considerations are
    (a) the magnitude of the observed rise
    (b) factors other than anthropogenic greenhouse gas concentrations that may have contributed to the rise.

    If those other factors could have contributed more than half the rise then the first statement is wrong. In that case it would not be “very likely” that ‘more than half’ of the ‘observed rise’ was “due to the observed increase in anthropogenic greenhouse gas concentrations”: at best it could only be said to be likely.

    And if those other factors probably have contributed more than half the rise then the first statement is very wrong. In that case it would not be “very likely” that ‘more than half’ of the ‘observed rise’ was “due to the observed increase in anthropogenic greenhouse gas concentrations”: at best it could only be said to be possible.

    Importantly, if those other factors were shown to have contributed more than half the rise then the first statement is a falsehood. In that case it would not be “very likely” that ‘more than half’ of the ‘observed rise’ was “due to the observed increase in anthropogenic greenhouse gas concentrations”: it would be true to say that some, but less than half, of the ‘observed rise’ was probably “due to the observed increase in anthropogenic greenhouse gas concentrations”.

    Michaels and Knappenberger provide an analysis which accepts the magnitudes of the asserted “anthropogenic greenhouse gas concentrations”. Both these assertions could be disputed. There is great uncertainty concerning the anthropogenic proportion of the increase to atmospheric greenhouse gas concentrations and the climate sensitivity to greenhouse gas concentrations. Hence, the analysis by Michaels and Knappenberger is conservative.

    Their analysis considers the contributions of two factors that the IPCC were “unaware of” when the IPCC made its statements. These two factors provide individual warmings of 1 W/m^2 and 0.5 W/m^2. So, these individual warmings sum to 1.5 W/m^2 which is equivalent to about half of the ~3.0 W/m^2 warming due “anthropogenic greenhouse gas concentrations” is 3.0 W/m^2.

    Hence, addition of thisnew information requires the IPCC statement to alter the word “most” to ‘more than a quarter’.

    The amended IPCC statement is thus corrected by new knowledge to be:

    “More than a quarter of the observed increase in global average temperatures since the mid-20th century is very likely due to the observed increase in anthropogenic greenhouse gas concentrations.”

    Furthermore, Michaels and Knappenberger say additional new knowledge shows the ‘observed warming’ to be in error. They say the IPCC reported the ‘observed warming’ was 0.7 deg.C but says 0.2 deg.C of this is an error “due to a warm bias in the measurements” that was not known at the time of the IPCCs statement. Hence, they say, this accounts for (0.2/0.7) = 0.3 of the ‘observed warming’.

    They then use this to calculate a further reduction to the proportion of the ‘observed rise’ that is “due to the observed increase in anthropogenic greenhouse gas concentrations”. But it is not clear from their explanations that this is valid. Assuming all the forcings are accounted, then the effect of each forcing contributes to the ‘observed rise’ in proportion to its magnitude. And the magnitude of the ‘observed rise’ does not affect this.

    The second of the IPCC statements is not relevant to the analysis. If it were then several other considerations should also be included; e.g. the anthropogenic proportion of the increase to atmospheric greenhouse gas concentrations and the climate sensitivity to greenhouse gas concentrations.

    In summation, I consider that Michaels and Knappenberger make a good case but they overstate it. They have proven that new knowledge disproves the statement in the AR4 that says;

    “Most of the observed increase in global average temperatures since the mid-20th century is very likely due to the observed increase in anthropogenic greenhouse gas concentrations.”

    And, if the IPCC analysis is agreed, they have proven that the statement should now be;

    “More than a quarter of the observed increase in global average temperatures since the mid-20th century is very likely due to the observed increase in anthropogenic greenhouse gas concentrations.”


  27. Nullius in Verba

    What if you got a negative total? Or zero?
    GHGs +2 C
    Black carbon +1 C
    AMO/PDO -4 C

    Is it still true to say that GHGs are responsible for 2/3 of the -1 C change? Or do you now have to say that ocean oscillations are responsible for 400% of it?

    GHGs +2 C
    Black carbon +1 C
    AMO/PDO -3 C

    What does it mean to say that GHGs are responsible for 2/3 of no change at all?

    The only way I can see to ascribe “responsibility” is as Jim suggested above: take the absolute values.

    Then in the first case, GHGs are responsible for 2/(2+1+4) = 2/7 of the result, and in the second case, ocean oscillations are responsible for 3/(2+1+3) = 1/2 of the result (which because the two halves are in opposite directions, cancels out).

    There is a difficulty though. When ascribing percentage blame for the rise in CO2 levels to human industry, we have the same problem of different parts of the oceans adding and subtracting very large amounts (plus the biosphere generally, of course, but I’ll simplify). What factors do you choose to match up and hence cancel out? Without man, it would not have gone up, but without the release of CO2 from tropical oceans, absorption by the polar oceans would have sucked a great deal of it out. So which is more “responsible” for the result?

    All of the above assumes they combine additively. But what if some feedback effect, say, halves the output? What if you have two of them, in combination? Do you calculate their contribution one after the other, or take logarithms, or what?

    Is it even a meaningful question?

    • NIV – At the risk of belaboring a point made several times above, the conflict is between the IPCC’s claim that most of the observed warming was due to anthropogenic GHGs and Pat Michaels’ testimony disputing that claim. For most of us trying to assess that particular argument, the issue is not what the true value of the warming was (i.e., whether it was overstated or correctly stated), but rather what percentage of that true value, whatever it was, is attributable to GHGs.

      It turns out that based on the evidence presented in this post and the ensuing comments by Chip Knappenberger and others, the IPCC assertion is supported. More than half of the warming influence came from the GHGs, and this is independent of the magnitude of cooling influences.

      The math is as follows: GHGs = 3 W/m^2, black carbon = 1 W/m^2, so that 75 percent is due to the GHGs. Not cited here, but included in Michaels’ testimony is an adjustment that he claims removes warm biases in land temperature estimates, and this would reduce the 75 percent figure somewhat. However, temperature trend data mainly track changes in sea surface temperature (the oceans comprise more than 70 percent of global area), and so the adjustment would not contradict the main IPCC point.

      • Nullius in Verba


        Yes, but my point was not what the true value of the warming was, but what it even means to talk about a “percentage attributable”.

        My numbers were totally made up, simply to illustrate the ideas when pushed outside the range of the actual data, to see if the concepts stood up and made sense in general. I don’t think they do.

        At the least, I don’t think there is a single, obvious, unambiguous interpretation of the phrase.

      • I agree the semantics lead to ambiguity. I’m more concerned with dividing up the real world warming influences. We can address some of them – GHGs and black carbon – but not the natural climate variations. Between GHGs and black carbon, Raymond Pierrehumbert has pointed out that because of their vastly differing lifetimes (BC measured in weeks or months), any failure to address CO2 has implications for centuries, whereas putting off efforts directed at black carbon would incur climate costs that coud be remedied quickly.

      • Between GHGs and black carbon, Raymond Pierrehumbert has pointed out that because of their vastly differing lifetimes (BC measured in weeks or months), any failure to address CO2 has implications for centuries.

        Every now and I then I take issue with the dogma, and this is one of those occasions. Deniers love to point out that anthropogenic CO2 is only 3% of the carbon cycle. This is irrelevant assuming business as usual as “we” all know (all of us except the deniers). However it suddenly becomes very relevant with any significant carbon mitigation.

        The reason is that nature draws down some 50-60% of what we emit in the way of CO2. This has been going on for over a century. Basically nature is some 30 years behind us in sweeping up after us.

        As long as we continue to add an exponentially growing amount of CO2 (“business as usual”) this situation will remain unchanged. But if we were, say this year, to halt further increases of carbon emissions and hold them steady at some 8 GtC/yr (30 Gt of CO2), nature would start to narrow that 30 year lead we’ve held over her all this time, and we’d see the 400 ppmv level of CO2 gradually start to decline. Exactly to what I don’t know, but it would be some equilibrium level between preindustrial 280 and current 390.

        RP’s argument (and David Archer’s?) about CO2 staying in the atmosphere “for centuries” after mitigation is just as fallacious as the deniers’ argument about it being a harmless amount during business-as-usual.

        Nature is doing what Klaus Lackner wants to do, only 10,000 times more effectively than anything he’s been able to demonstrate so far. Money spent on Lackner-type solutions would be better spent on carbon mitigation.

      • There’s another subtlety that hasn’t been discussed. Most black carbon comes from anthropogenic sources – fossil fuel combustion and biomass burning – so that if we ask what percentage is anthropogenic rather than merely due to GHGs, the value rises to a higher level than cited above.

        Here’s the subtlety – black carbon is rarely emitted in pure form, but rather as a mixture of aerosols with different warming and cooling properties. The characterization of these is subject to some uncertainties, but is known well enough to assert that the actual emissions that include black carbon as a component exert much less net warming effect than the 1 W/m^2 value cited. They may in fact mediate a net cooling efffect, although less than the effect of typical industrial effluents. In that sense, singling out the black carbon contribution from the mixture tends to give a somewhat misleading impression. It implies that we can account for warming without needing to acknowledge that much of the aerosol-induced warming is accompanied by its own cooling components. In other words, it would be misleading to claim that we can make claims about anthropogenic warming but with no reliable information to support the conclusion that there is a significant countervailing aerosol cooling. This is one case where we can say that a component of anthropogenic warming is significantly opposed by a component of anthropogenic cooling within the same mixtures.

      • Fred,

        Aren’t the cooling components of carbonaceous aerosols already included in the negative forcing from aerosol tally?

        And, while discussing the subtleties of BC, the snow darkening effect shouldn’t be left out of the discussion.


      • Chip – I believe they are included. My point was addressed to claims made here and elsewhere (not by you) that we really have no idea whether aerosols have a significant cooling effect, and that they may not. I’m trying to point out that when an aerosol mix clearly contains components that made it less of a “warmer” than the effects of its warming component alone, negative forcing by aerosols can’t be dismissed as simply an invented effect designed to explain away less observed warming than would be expected from GHGs and black carbon alone..

      • Fred,

        Surely if there were firmly established warming biases in the land temperature estimates then we would have to adjust for these by lowering our figure for total observed warning, so the the proportion of warming due to GHGs would be calculated as a % of a lower amount and it wouldn’t reduce the 75% figure.

      • My use of the word “biases” may have been misleading. I believe McKitrick and Michaels argued that some of the observed temperature change was real but due to socioeconomic variables, including UHI effects, rather than GHGs, black carbon, etc.

      • Even so, you’re right in that those authors implied that the measured temperatures represented local effects that while real, overstated the mean value of land temperature increases. That’s a good point.

      • Continuing the point: An example of a bias is a measurement error due to inappropriate proximity of measuring devices to urban heat. An example of a true effect is a land use change from socioeconomic development that favors warming – e.g., changes that incur warming by reducing regional albedo or water content. The authors suggest that both real and spurious warming unrelated to the “usual suspects” may have affected land temperature trends.

      • andrew,

        You are correct that a warm bias in the observed temperature record would not influence the percentage contribution of GHGs to the actual warming, but it certainly would reduce its percentage contribution to the observed warming. There potentially is a difference between what we have measured and what has really happened.


      • Indeed! See the naive but suggestive graph and comments I posted above; cutting reporting stations by half seems to pay off with about a 1.3°C absolute jump in “trend”, virtually instantaneously.

      • Of course, it’s quite possible that the cut stations might have to be skillfully chosen, rather than counting on random elimination and simple coarsening to do the trick. But I’m sure the ‘keepers are up to the job!

      • The obvious ones to go for, based on past successes, are the ones with high “‘tudes” (altitudes and latitudes). Then go for the lonely isolated ones far from the loving attention of meteorologists at airports and data centers. If there are any of those left; Anthony Watts seems to be having a hard time locating many.

  28. Michael Larkin

    Speculation on a previous thread, semantics on the current one. My scepticism may yet be replaced by cynicism.

  29. Richard S Courtney

    Michael Larkin:

    You say;
    “Speculation on a previous thread, semantics on the current one.”

    Yes. I agree.
    As I said in the first sentence of my above comment,
    “This argument is about words, their meaning, and how they should be interpreted.”
    That must be so because this argument is about the accuracy of statements made in Michael’s Testimony that pertained to a statement in the AR4.

    I explained my understanding of the pertinent statement in the AR4 and then assessed the accuracy of Michael’s statement according to that understanding.

    I find it interesting that nobody posted agreement or dispute with my explained understanding (or my assessment).

    Clearly, everybody is content that only their individual interpretations warrant acceptance and debate. But the ‘discussion’ is entirely semantic nonsense when there is no agreed meaning of the statements being discussed.


  30. Michael Larkin

    Yes, Richard.

    And the thing is, one doesn’t have to be able to follow the technicalities to the ultimate degree to detect things like speculation or semantics. Any reasonably intelligent person can detect them. The climate debate, which Dr. Curry is generously allowing to happen here (perhaps for the first time in earnest), is increasingly demonstrating to me how bizarre, almost surreal, the situation has become.

    Proponency and scepticism alike are becoming increasingly problematic for me. I feel like an uncollapsed probability function for which neither of two states may be actual possibilities.

  31. It occurs to me to ask why people don’t remove short-term events from the last couple of century’s temperature record in order to see what’s going on temperature-wise. The fast-moving short-term events tend to obscure the slower-moving long-term events. Moreover they don’t seem terribly irrelevant to long-term global warming.

    You can see this high-frequency noise in the global temperature record since 1850. It’s easily removed by using a wider moving average.

    What remains are basically two long-term signals: the 65-year Atlantic Multidecadal Oscillation, or AMO, and the global warming signal.

    The latter distorts the former so badly that it’s hard to make out its most recent zero-crossings. These have occurred at 1860 (up), 1892 (down), 1925 (up), 1957 (down), and 1990 (up). (Judith prefers 1995 for the last but that’s only because it’s climbing so fast by then as to almost completely mask the zero-crossing. To see where its real zero-crossing is one must subtract off the global warming signal.)

    The reason the period 1980-1998 climbed so quickly is that the AMO and the global warming signal were both climbing at the same time, culminating in the incredibly hot 1998. At that point the AMO started to level out and was flat by 2004. (There was a big dip during 1999-2000 and another during Oct 2007-May 2008, no idea what caused those.). The AMO is now 6 years into its 32.5 year downward swing, offsetting the global warming signal to some extent.

    However thanks to our exponentially increasing consumption of carbon-based fuels, easily seen in both the energy data and the atmospheric CO2 data, the global warming signal continues to get steeper. This makes it a challenge for the AMO to drag the temperature down as effectively as it did between 1950 and 1970 where it was just able to balance global warming, and, 65 years earlier still, where global warming put up no resistance at all against the AMO’s easily visible downswing between 1885 and 1905.

    1925 was a somewhat scaled-down version of the horrific rise in temperature centered 65 years later on 1990. 1925 rose a little more steeply than 1860 because the global warming signal had gotten a little stronger by 1925.

    • “don’t seem terribly irrelevant” –> “don’t seem terribly relevant”

    • 1925 was a somewhat scaled-down version of the horrific rise in temperature centered 65 years later on 1990

      Indeed it is,


    • The AMO is now 6 years into its 32.5 year downward swing, offsetting the global warming signal to some extent.

      The average AMO index for 2010 is going to be near a record high. Through 11 months, the average is 0.393. That’s exceeded only by 1878 at 0.448 and 1998 at 0.394. It does appear to be going down again, so the average for the full year will be a bit lower. But it will almost certainly be the third highest on record. The index for December would have to get into negative territory to lower the average to 1944’s 0.357.

      I vote for 1995 too. That’s when the loss in Arctic ice area and UAH anomaly moved into high gear. I wouldn’t be so precise in quoting the wavelength of the AMO either. It’s likely to be one of those quasi-periodic things that crop up in chaotic systems rather than a true oscillation.

      • The average AMO index for 2010 is going to be near a record high. Through 11 months, the average is 0.393. That’s exceeded only by 1878 at 0.448 and 1998 at 0.394. It does appear to be going down again, so the average for the full year will be a bit lower

        Global warming is making it harder to measure. 1878 was pretty much at the peak two cycles ago. I would model the current peak at 2005, but 1998 was extraordinarily hot and may throw off the AMO. Ditto for this year. The AMO should be going down now but global warming is going to make this really hard to see. Better to look for troughs than peaks, except these may be hard to pick out.

        I wouldn’t be so precise in quoting the wavelength of the AMO either. It’s likely to be one of those quasi-periodic things that crop up in chaotic systems rather than a true oscillation.

        I agree. In fact my sense is that there are really two cycles, one of 56 years and the other 75 years, of comparable amplitudes. They seemed to have been in phase around 1925, so even though the current peak looks high I suspect that’s an artifact of global warming and that the amplitude of the AMO oscillation was higher around 1925 than either now or around 1850. I obtained these from the HADCRUT3 global data alone, no peeking at more local data (but I will at some point). Always more than happy to be proved wrong. Milivoje Vukcevik may want to weigh in here.

        Delworth and Mann also obtained the 56-year period from Atlantic data (I did not know about their observation before I arrived at the 56-year figure, so it’s interesting I could see it in just the HADCRUT3 global data). They had the same concern as you about overfitting (unwarranted precision). They also talked about a 70-year cycle, which may be what I’m calling a 75-year cycle—they might just have rounded it down to 70 years.

        The very interesting tree-ring analysis is reasonably consistent with this 56-75 beat back to about 1680, including the vanishing of the AMO at around 1760. It would be very interesting to get other data of that vintage to see how well it matches both the tree-ring data and the 56-75 model.

  32. Chip’s analysis of the IPCC’s statement of attribution discusses only one of many problems with this statement. Everything would have been fine, if the IPCC had limited itself to saying that “the anthropogenic GHGs (AGGs) contributed more than half of the known positive FORCING driving temperature rise in the second half of the 20th century”. However, the IPCC wanted a statement about how much AGG’s contributed to 20th century TEMPERATURE rise, not to overall positive forcing. The only way I know to convert a forcing into a temperature rise is via climate sensitivity; a process that introduces additional uncertainties.

    Unfortunately, AR4 only provides us with forcing change since 1750, not forcing change for the period 1950-2000. [Why is the IPCC doing drawing conclusions about the cause of 1950-2000 temperature rise from 1750-2005 forcing and 1900-2000 climate models runs?] This is a non-trivial problem because: 1) Most of the increase in sulfur emissions was in place by 1950. 2) Sulfur emissions didn’t have a large NET increase between 1950 and 2000 (although there was a higher peak in 1970-85). 3) Sulfur emissions fell in the late 20th century. (Figure 1, Stern (2006); http://www.economics.rpi.edu/workingpapers/rpi0504.pdf). So aerosols probably play a much smaller role in Kapperburger/Michaels analysis of 1950-2000 than everyone assumes from looking at the aerosol forcing SINCE 1750 in Figure SPM2 of AR4 WG1.

    If we use the forcing data from Figure SPM2, we can calculate the expected temperature change for all anthropogenic forcings by multiplying the forcings (1.6 W/m2 with a 90% confidence interval of 0.6-2.4) times the climate sensitivity (3.0 degK/2X CO2 = 0.82 degK/W/m2, with a 90%?confidence interval of 0.4-1.2 W/m^2. I then converted the confidence intervals into standard deviations by assuming that each confidence interval represented +/-1.6 standard deviations and used Excel to do 1000 randomly chosen examples of 1.6+/-0.56 (1 Std) times 0.82+/-0.25 (1 std) to get a temperature rise of 1.46+/-0.87 (1 Std) and a 90% confidence interval of 0.07-2.32 degK.

    Now we compare these calculated temperature rises to the observed 1906-2005 temperature rise (we can’t go back to 1750 like the forcings do). P5 of the SPM of AR4 WG1 says the temperature rise is 0.76 (0.57-0.95). Only 7.7% of my 1000 multiplications gave a temperature rise 50% of observed warming. If 20th century temperature rise due to AGGs is predicted to be 0.07-2.32 degK (90% confidence), does this really tell us anything useful about what the cause of most of an observed temperature rise of 0.57-0.95 degK? No. Uncertainties in forcings make it impossible use 20th century warming to place any useful constraints on climate sensitivity. For exactly the same reason, uncertainty in climate sensitivity and forcings make is impossible to make any useful predictions how much 20th century temperature should have risen.

    The other half of the evidence for the IPCC’s attribution statement comes from Figure SPM4 and climate models. These models provide a much narrower range of predictions for 20th century warming because the IPCC’s “ensemble of opportunity” doesn’t take into account the uncertainty in the input AGG amounts and other parameters. Instead, “convergent evolution” has probably resulted in similar temperature rises for the 20th century. (After all, who would continue funding a climate model that can’t reproduce the historical temperature record?) If one used Stainforth’s ensembles of models and a range of AGG inputs, the red range of model predictions would be huge and probably would overlap the blue observations.

  33. Well done. Remarkable correlation after 1920. Even has the peak at 1880, though not to scale (should be bigger). Not sure what to make of the peak at 1860 though, some student should be assigned that for their thesis topic. ;)

    Interesting difference when you replace Michael by Horace. Evidently Horace is much better known, who’d have thought.