Natural global warming

[swansont]

It's a misapplication, you take advantage of my loose and imprecise prose. To accept your argument, one could never back out any identified error in any data set on the basis that there may be unknown compensating errors that remain.

 

No, that's not true. Science doesn't set out to show that a particular thing is true, so it doesn't assume something to be true unless shown false. Quite the opposite, in fact. It assumes the effect isn't there unless there is evidence to show that it is (null hypothesis). If you can't show that this effect is present over the entire data range in question, you can't claim that it is.

[cypress]

No, that's not true. Science doesn't set out to show that a particular thing is true, so it doesn't assume something to be true unless shown false. Quite the opposite, in fact. It assumes the effect isn't there unless there is evidence to show that it is (null hypothesis). If you can't show that this effect is present over the entire data range in question, you can't claim that it is.

 

Though Knappenberger did not use your words, this is the point he was making. He assumed no effect (the null hypothesis) where there was no data available. I only need to show it is in effect from 1980 through 2009, which is the only time period an adjustment is made.

[Cap'n Refsmmat]

Though Knappenberger did not use your words, this is the point he was making. He assumed no effect (the null hypothesis) where there was no data available. I only need to show it is in effect from 1980 through 2009, which is the only time period an adjustment is made.

 

Then he should at least admit the sizable margin of error. It has been established that stratospheric water vapor can cause significant contributions for warming, so it is likely that the 1950 data is in need of adjustments. Knappenberger should add the margin of error to his results if he wishes to be intellectually honest.

 

Given the size of the effect on the 2009 data, I'm sure the margin of error is significant.

[swansont]

Though Knappenberger did not use your words, this is the point he was making. He assumed no effect (the null hypothesis) where there was no data available. I only need to show it is in effect from 1980 through 2009, which is the only time period an adjustment is made.

 

As I pointed out before, Knappenberger's goal differs from yours.

[cypress]

Then he should at least admit the sizable margin of error. It has been established that stratospheric water vapor can cause significant contributions for warming, so it is likely that the 1950 data is in need of adjustments. Knappenberger should add the margin of error to his results if he wishes to be intellectually honest.

 

One cannot add a margin of error until it is calculated and it cannot be calculated without data.

 

Given the size of the effect on the 2009 data, I'm sure the margin of error is significant.

 

I don't see how you can be sure of something when there is no data.

 

As I pointed out before, Knappenberger's goal differs from yours.

 

 

I find that interesting because only the timeframe is different. Kappenberger is interested in the time between 1950 and 2009, while I am interested in the mid 1800's to 2009, but the purpose is the goal is the same.

[swansont]

 

I don't see how you can be sure of something when there is no data.

 

Precisely.

 

 

I find that interesting because only the timeframe is different.

 

Again, precisely.

[Cap'n Refsmmat]

I don't see how you can be sure of something when there is no data.

Likewise.

[cypress]

Likewise.

 

Agreed, therefore the reasonable approach to any time period where no data is available is to make no adjustments, no estimates about margins of error, and no predictions as to what effects might have occurred. This is what I have been suggesting all along.

[Cap'n Refsmmat]

Agreed, therefore the reasonable approach to any time period where no data is available is to make no adjustments, no estimates about margins of error, and no predictions as to what effects might have occurred. This is what I have been suggesting all along.

 

You've also been suggesting that we draw conclusions (what amount of warming is caused by GHGs vs. stratospheric water vapor) from incomplete data. The time period with no data (1950-1980) can cause significant changes in the final conclusion, so the missing data is important. The only reasonable approach to any time period where data is incomplete is to make no adjustments, no estimates about margins of error, and no predictions as to what effects might have occurred. This is what I have been suggesting all along.

[cypress]

You've also been suggesting that we draw conclusions (what amount of warming is caused by GHGs vs. stratospheric water vapor) from incomplete data. The time period with no data (1950-1980) can cause significant changes in the final conclusion, so the missing data is important. The only reasonable approach to any time period where data is incomplete is to make no adjustments, no estimates about margins of error, and no predictions as to what effects might have occurred. This is what I have been suggesting all along.

 

For the time period between 1980 and 2009, we have sufficient data to draw conclusions and an adjustment during that time is justified. Lack of data for other time periods does not diminish this reality. One does not refrain from making an adjustment for a known effect because of the mere possibility that unknown effects might somehow someway counter the adjustment. I am unable to find support for your approach in any scientific or statistical texbooks, can you offer a manual that supports your approach?

[Cap'n Refsmmat]

For the time period between 1980 and 2009, we have sufficient data to draw conclusions and an adjustment during that time is justified. Lack of data for other time periods does not diminish this reality. One does not refrain from making an adjustment for a known effect because of the mere possibility that unknown effects might somehow someway counter the adjustment. I am unable to find support for your approach in any scientific or statistical texbooks, can you offer a manual that supports your approach?

 

You still don't seem to understand my approach, so I can see why you're having trouble. I am not advocating refraining from making adjustments. I believe I explained my reasoning in posts 44, 47, 53, and 57.

 

You cannot draw conclusions about the amount of warming caused by GHGs between 1950 and 2009 if you have data only to draw conclusions between 1980 and 2009. You can make adjustments, but your data is incomplete, and thus your figures for 1950 are uncertain.

 

Since there is no data about 1950, there is no knowing whether the figure for GHG contributions to climate is accurate for that year. Experience shows us that stratospheric water vapor needs accounting for, and there is no data for 1950's stratospheric water vapor. Thus, the 1950 data is untrustworthy, and we cannot draw conclusions with it and the 2009 data.

[cypress]

I believe I understand your approach. My current goal is accounting for as much of the upward trend in apparent global temperature from 1950 through 2009 to sources other than CO2 from human sources (the primary effluent AGW proponents are advocating must be controlled, and the only one I question) and leaving unaccounted, everything for which I either lack data or for which appears to be a known consequence of human sourced CO2. This is the topic of Knappenberger's article.

 

You seem to be confusing my goal with one of drawing conclusions about the total actual effect (known and unknown) of GHG's between 1950 and 2009.

 

When one focuses on the goal at hand and applies standard practice for experimental statistics, one should conclude that the data provided supports accounting for the warming indicated by the data from 1980 through 2009 as not related to CO2 GHG's (scientific justification for not ascribing it to CO2 was provided in post 48) and not making any accounting for the period from 1950 through 1980 where there is no data. If you continue to take exception to this approach to the goal I have stated, please provide a reference that indicates this approach is wrong and a different approach is correct because I have looked and cannot find any such reference.

[swansont]

That may be because "you shouldn't move the goalposts" probably isn't explicitly expressed in a textbook. I thought your goal was to account for "natural causes for all but 0.2-0.4 degrees of warming" between 1800 and 2003.

[cypress]

That may be because "you shouldn't move the goalposts" probably isn't explicitly expressed in a textbook. I thought your goal was to account for "natural causes for all but 0.2-0.4 degrees of warming" between 1800 and 2003.

 

Take a victory lap if you must swansont. In the future I will choose my words much more carefully. Skeptics primarily take issue with claims about CO2 specifically (as I do) and I am following the skeptics arguments and the approaches laid out in the literature which was the driver for my original post. I will continue to focus on explaining apparent warming using the arguments available. Perhaps you might be, but I am not too concerned if a particular cause of apparent increase in temperature is due to an error in reporting or a natural effect so long as it is not attributed to human sourced CO2. In the end I intend to break out the natural from the other as the discussion proceeds. If you view this as moving the goal post so be it.

[JohnB]

You've also been suggesting that we draw conclusions (what amount of warming is caused by GHGs vs. stratospheric water vapor) from incomplete data. The time period with no data (1950-1980) can cause significant changes in the final conclusion, so the missing data is important. The only reasonable approach to any time period where data is incomplete is to make no adjustments, no estimates about margins of error, and no predictions as to what effects might have occurred.

Cap'n, you might want to consider that statement in the light of its implications.

 

If you believe it to be true, then you must agree that no conclusions can be drawn about the various proportions of forcings during the 1950-1980 period. The attribution of warming to predominanty CO2 is in a large part based on the attribution proportions of the 1950-1980 period and can be considered a projection based on that period.......

[Cap'n Refsmmat]

Cap'n, you might want to consider that statement in the light of its implications.

 

If you believe it to be true, then you must agree that no conclusions can be drawn about the various proportions of forcings during the 1950-1980 period. The attribution of warming to predominanty CO2 is in a large part based on the attribution proportions of the 1950-1980 period and can be considered a projection based on that period.......

 

I say no conclusions may be drawn because drawing comparisons between adjusted and unadjusted data is incorrect. We can't draw conclusions about the effects of water vapor in 1950 if we don't know how much there was, of course.

 

On the other hand, we know how much CO2 there was, and we can predict its effects via models validated on more recent data where more forcings are taken into account.

 

Of course, some of these models may need updating with results from more recent stratospheric water vapor data. I believe that is what Solomon suggested in the article. I do not contest this.

[cypress]

I say no conclusions may be drawn because drawing comparisons between adjusted and unadjusted data is incorrect. We can't draw conclusions about the effects of water vapor in 1950 if we don't know how much there was, of course.

 

Attributing warming to known causes when possible is not making a comparison between adjusted and unadjusted data. Your complaint does not apply.

[Cap'n Refsmmat]

Attributing warming to known causes when possible is not making a comparison between adjusted and unadjusted data. Your complaint does not apply.

 

Comparing warming caused by GHGs between 1950 and 2009 is indeed making a comparison between adjusted and unadjusted data, when only a portion of the data has been adjusted.

 

Again, I repeat my example: Suppose we're focusing on warming from 1950-the present. Suppose the temperature in 1950 was depressed due to non-anthropogenic forcings. For example, suppose aliens spent all of 1950 lobbing gigantic chunks of ice into the Atlantic ocean. If we wish to determine anthropogenic climate change over Earth's history, we must account for non-anthropogenic forcings. If we account for non-anthropogenic forcings and determine that the temperature in 1950 would have been cooler, the anthropogenic warming since then would be larger.

 

Thus, corrected data for 1950 is important.

 

edit: and for what it's worth, my hopes for the reputation system involved rewarding for helpful contributions and removing points for unhelpful ones, not docking points from those you disagree with.

[cypress]

I don't know how to say this any plainer. I am not accounting for all warming or even warming of a particular cause. What I am doing is not covered by your example. I am extracting as much of the apparent warming attributed to known causes other than CO2 into a bucket and leaving warming that is unattributed plus warming that is attributed to CO2 into the other bucket.

[Cap'n Refsmmat]

Suppose we're focusing on warming from 1950-the present. Suppose the temperature in 1950 was depressed due to non-anthropogenic forcings. For example, suppose aliens spent all of 1950 lobbing gigantic chunks of ice into the Atlantic ocean. If we wish to determine anthropogenic climate change over Earth's history, we must account for non-anthropogenic forcings. If we account for non-anthropogenic forcings and determine that the temperature in 1950 would have been cooler, the anthropogenic warming since then would be larger.

 

If we began collecting data on giant alien ice chunks starting in 1980, and knew the impact of them on our climate in 2009, we could adjust 2009 figures for other climate forcings accordingly. However, we knew there were ice chunks in 1950, but we didn't know how many or what impact they had at all. So we remove them from the figures (or "put them in a separate bucket"), but we know that the data from 1950 has some influence, positive or negative, from alien ice chunks of doom.

 

Thus, we know there is some unknown factor in our 1950 data that should be put into a separate bucket, but we cannot. We did, however, remove that factor from 2009's bucket.

 

When we then compare temperatures between 1950 and 2009, one will have unknown goo in the bucket, and the other will have had that unknown goo extracted. Right?

[cypress]

I understand your example and do not see any particular reason to take issue with it because, as I have now said several times, your example and your method is not what I have proposed. I do not use the 1950's bucket or the 2009's bucket. I do not remove anything from the 2009 bucket. I leave them untouched.

 

Instead, I have a single bucket of unattributed cumulative apparent warming from 1950-2009 and at the beginning of my analysis, it is filled to the 0.702 degree C hash mark. From that bucket, I attributed 0.150 C to a measurement error in surface sea temperatures measured by ship-borne sensors and removed that amount from the bucket and placed it in another bucket labeled attributed warming from 1950-2009. Next I attributed 0.066 C to changes in upper atmospheric water vapor driven by tropical moisture that correlates to El Nino Pacific Ocean oscillations of surface sea temperatures, and I place that amount in the attributed bucket. The amount remaining in the unattributed apparent warming bucket is some unknown amount of warming (or cooling) due to any changes in upper atmospheric water vapor from 1950-1980 plus any error in what was attributed between 1980 and 2009, plus all other unattributed causes.

[Cap'n Refsmmat]

My point is that cumulative apparent warming between any two years is a comparison of the warming of each of those years; that is, one must compare the warmth of 1950 to that of 2009. If, for some reason, the warmth of 1950 were depressed for some unaccounted-for non-anthropogenic reason, our comparison is wrong.

 

Anyway, these are accounting details; we still have not covered the other issue: whether stratospheric water vapor is a feedback of any other known anthropogenic forcing. You have previously quoted a number of researchers speculating that changes in stratospheric vapor were caused by global sea-surface temperature patterns, but did not cover the possibility that GHGs and other gases have an impact on these patterns. For example, a 2001 article about stratospheric water vapor has this to say:

 

There are two driving forces behind the change in stratospheric moisture. Increasing emissions of methane are transformed into water in the stratosphere by chemical reactions. This can account for about a third of the observed increase in moisture there. In addition, there is a greater transport of water from the lower atmosphere, which happens for several reasons. First of all, more water may be available in the lower atmosphere to be carried up. Warmer air holds more water vapor than colder air, so global warming will make the lower atmosphere wetter. Another possibility is that air is carried up more rapidly into the stratosphere. Climate models indicate that greenhouse gases such as carbon dioxide and methane may enhance the transport of air from the lower atmosphere up into the stratosphere. Additionally, the coldest temperature through which the air passes could change, which would alter the amount of water that freezes out along the way.

 

http://www.giss.nasa.gov/research/briefs/shindell_05/

 

 

These are all anthropogenic sources. It would thus seem to me that at least some of the water vapor changes are anthropogenic, and so some of the climate changes you want to place in the "attributed" bucket are in fact anthropogenic.

 

Incidentally, in post #29 you acknowledged that the measurement error doesn't matter in long-run warming calculations, such as warming from 1800-present. Now we refer to warming from 1950, in which case it's applicable -- but your original claim was to attribute warming from 1800 to the present (see the OP). Shall we avoid changing our references and agree on the whole of modern warming, 1800-present? I believe that's what you intended to discuss originally.

 

 

 

[cypress]

My point is that cumulative apparent warming between any two years is a comparison of the warming of each of those years; that is, one must compare the warmth of 1950 to that of 2009. If, for some reason, the warmth of 1950 were depressed for some unaccounted-for non-anthropogenic reason, our comparison is wrong.

 

Anyway, these are accounting details; we still have not covered the other issue: whether stratospheric water vapor is a feedback of any other known anthropogenic forcing. You have previously quoted a number of researchers speculating that changes in stratospheric vapor were caused by global sea-surface temperature patterns, but did not cover the possibility that GHGs and other gases have an impact on these patterns. For example, a 2001 article about stratospheric water vapor has this to say:

 

http://www.giss.nasa.gov/research/briefs/shindell_05/

 

These are all anthropogenic sources. It would thus seem to me that at least some of the water vapor changes are anthropogenic, and so some of the climate changes you want to place in the "attributed" bucket are in fact anthropogenic.

 

The research by Solomon et al, address the methane component. The comments by Solomon and the other authors plus the other researchers (see post 48) all consistently rule out changes in stratospheric water vapor by transport other than changes due to temperature drivers between the critical layer addressed by the research and changes in sea surface temperature in the tropical regions. Observed changes in sea temperature in the tropical region is almost entirely driven by ENSO. Furthermore the effect correlates well with the ENSO index.

 

I would be inclined to accept your argument if you were able to attribute ENSO to anthropogenic causes, but this does not seem to be the case.

 

Incidentally, in post #29 you acknowledged that the measurement error doesn't matter in long-run warming calculations, such as warming from 1800-present. Now we refer to warming from 1950, in which case it's applicable -- but your original claim was to attribute warming from 1800 to the present (see the OP). Shall we avoid changing our references and agree on the whole of modern warming, 1800-present? I believe that's what you intended to discuss originally.

 

I am currently addressing the years between 1950 and 2009 based on the articles and arguments of skeptics. I include it only to make it easier to follow the skeptic's argument. I will address the mid 1800's through 1950 afterwards and combine them at that time. I described this approach previously. I also noted that the attribution you refer to would be backed out at that time since, as swansont first noted, it is only significant in the intermediate.

[iNow]

<cough> Obfuscation </cough>

[cypress]

Knappenberger next attributes 25% of the warming in toto to black carbon deposition. He describes it this way:

 

"... results from Ramanathan and Carmichael published in Nature Geoscience last year. These researchers reviewed the scientific understanding of how black carbon aerosols (aka soot) warm the earth’s climate. Black carbon is not a GHG. Black carbon warms the earth by directly absorbing reflected solar radiation and also by darkening the surface of snow and ice when it is deposited there (and enhances melting). "

 

I will come back to this when I address warming from the mid 1800's to 1950 but for now I will use Knappenberger's illustration and numbers for 1950 through 2009. Warming attributed to carbon black during this time frame is 0.120 C. The graphic Knappenberger chose to illustrate this is below.

 

 

Of the 0.702 C apparent warming that occurred from 1950-2009, 0.366 is yet unattributed but 0.336 is attributed to causes other than anthropogenic GHG's.