Support the Arctic Sea Ice Forum and Blog

Author Topic: Quantifying underestimates of long-term upper-ocean warming  (Read 43804 times)

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #50 on: November 11, 2014, 01:29:04 AM »
I try to calculate how many nuclear bombs per second is that 7.1 10y22 Joule.
Wikipedia says little boy was : 67 TJ   http://en.wikipedia.org/wiki/Little_Boy
Per year : 7.1 10y22 / 35 = 2,028571429×10²¹ Joule
Per day : /360 = 5,634920635×10¹⁸ Joule
Per hour : /24 = 2,347883598×10¹⁷ Joule
Per secondes : /3600 = 6,521898883×10¹³ Joule
Per nuclear bombe : /67.10y12 = 0,973417744 Nuclear Bombs/seconde

That's not 4 nuclear bombs per seconds used on the Neven blog, skeptical and other... is there underestimation somewhere ?

Quote
the durack et.al estimates are from 1970 to 2004 (34 years)  and estimate an increase in OHC for the southern hemisphere of 2.2 to 7.1 * 10^22 joules

Laurent,

The additional heat accumulation found in Durack et. al. is on top of the 4 bombs per second estimation.  In addition, the extra heat accumulation was observed over a 34 year period, not a single year.

The amount of heat added to the earth's biosphere each second is measured on the TOA analysis.  My estimations of TOA show that before Durack et. al TOA was about .7 Watts per meter squared (according to Nuccitelli et. al. and Hansen & Soto (2010))  after Durack et. al, the total heat accumulation is between 10% and 30% higher so current TOA is increased accordingly.



full image here:  http://oi58.tinypic.com/2ex60ip.jpg


However, if my observations are correct and the cause of the increased heat accumulation in the southern oceans is due to underestimations of the current aerosol forcing, then this means that the amount of total forcing (less aerosol effects) is much higher.

At this point, it should be noted that the anthropogenic aerosols will eventually go away and when they do, even using the potentially underestimated values, the amount of total incoming radiation from climate forcing will DOUBLE in intensity. (from .7-.9 to 1.5-2.0) 

If there is a significant underestimation it could go even higher (1.8-2.6)

Please note Top of Atmosphere is different from total anthropogenic forcing (as shown in the graph) as TOA subtracts the value of energy being lost as longwave (heat) to space.


Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #51 on: November 11, 2014, 01:31:16 AM »
If my understanding of Durack et. al is correct and the large uncertainty of Aerosol forcing cloud effects are severely underestimated (negative), then this cloud cover effect from persistent contrail induced cirrus will be a likely contributor to the southern hemisphere energy gains.

If this is true then the forcing component of GHG (less aerosols) is much larger and ECS is 10-30% higher.



If the warming impact of contrails is contributing to the increase in heating in the southern hemisphere, then that would imply ECS is lower.
Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #52 on: November 11, 2014, 01:31:51 AM »
I hesitate to speak for him, but I assume that jai means that the contrail effect is to keep the Northern Hemisphere cooler than it would otherwise be. So they are artificially masking warming that would otherwise be happening. Hence the claim that ECS is higher than generally proposed.
Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #53 on: November 11, 2014, 07:48:23 PM »
As already demonstrated the heat flux between the hemispheres is quite large enough to account for significant differences in the location between a change in heat input, and change in ocean heat storage.  Also consider that the heat budget may not have balanced on a hemisphere basis with the incorrect figures prior to Durack.  Durack shows a SH heat component equal to what is predicted in the models, and higher in the NH.  So Durack shows an increase in the SH heat component, but the extra heat in comparison to the models is in the NH, which is the opposite of what would be expected with a stronger aerosol cooling effect and no heat transport between hemispheres.

Note that the surface SSTs in the SH are well sampled since the satellite era.  It is the deeper ocean heat content that is in question, and the location of the extra heat probably has more to do with the location of extra mixing from the surface to the deeper ocean beyond what is predicted in the models than it does to the geographic location of the source of the extra heat.

From my reading of the Durack et al. paper,

1. The "extra heat" in comparison to the observations is in the Southern Hemisphere.

2. The "extra heat" in comparison to the models is in both the Northern and the Southern Hemisphere.  The paper suggests that both hemispheres contribute about the same amount of "extra heat", relatively, compared to the models.

See also Figure 5 of Durack et al., shown below.  The grey rectangles show the CMIP multi-model means, and the black vertical lines indicate the one standard deviation spread.  Note that the Southern Hemisphere contains about 60% of the world's oceans;  So Durack et al. suggest that the oceans in the Northern and the Southern hemispheres have been absorbing heat at a roughly equal rate (per unit area), which is in good agreement with the CMIP model results.


Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #54 on: November 11, 2014, 07:48:52 PM »
steven:

the caption for the image (5) that you posted says:

Quote
The NH and adjusted SH estimates are summed to yield global estimates (upper inset). Uncertainty estimates show the range of adjusted values obtained using the one standard deviation spread of model-simulated ratios

the little bits on the bottom right above the white line? those are the adjustments.
Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #55 on: November 11, 2014, 07:49:24 PM »
steven:

the caption for the image (5) that you posted says:

Quote
The NH and adjusted SH estimates are summed to yield global estimates (upper inset). Uncertainty estimates show the range of adjusted values obtained using the one standard deviation spread of model-simulated ratios

the little bits on the bottom right above the white line? those are the adjustments.

The little bits above the white lines in the figure are the Durack et al. adjustments to the observations in previous decades (which were data-sparse in the Southern Hemisphere).

However, this is irrelevant to my point.  The discussion was about how the findings of Durack et al. compare to the models (CMIP3 and CMIP5).

Durack et al. find that the Northern Hemisphere was responsible for roughly 40% of the total global ocean heat uptake (between 0 and 700 meters depth), and the Southern Hemisphere roughly 60%.  As I said, this is in good agreement with the CMIP models.

So Durack et al. suggest that the hemispheric distribution of ocean heat uptake in the CMIP models is (broadly) correct.  The paper does suggest that there was an incorrect infilling of data-sparse regions in the Southern Hemisphere, but that was a problem of the observations and has nothing to do with the CMIP models.


I have migrated all Durack et. al. discussion topics to the Durack et. al thread.

I would appreciate all future discussion on this topic take place there:

https://forum.arctic-sea-ice.net/index.php/topic,1011.0.html

Apparently you forgot to copy my yesterday's comment to that thread, as well as the comments preceding it.  If you copy those comments to that thread, then I will post any further responses at that location.
Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #56 on: November 11, 2014, 08:02:11 PM »


Just to be clear that the adjustments made are made solely on the southern hemisphere and that these adjustments increase the global heat accumulation significantly above the CMIP5 and CMIP3 Multi model means.

(really that last bit was shown above in the figure 5 bar graphs)

I get what you are saying steven, the MMMs all suggested a higher OCH absorption ratio from the southern hemisphere than the observational data showed. 

The proportional ratios now fit the MMMs ratios though the scale is a bit larger (10-30%) of total heat accumulation.

I guess the real question with regard to my observation of the underestimation of NH aerosol values is if the current aerosol forcing values are derived from models, from observations or from models and constrained by observations?  If they relied on the models then this would indicate a significant adjustment is needed, likely adjusting the NH system down as well as the SH system up.
Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

AbruptSLR

  • Multi-year ice
  • Posts: 17535
    • View Profile
  • Liked: 741
  • Likes Given: 236
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #57 on: November 13, 2014, 12:05:30 AM »
The linked research indicates that ocean surface warming across multiple oceans has contributed to the recent faux atmospheric surface temperature "hiatus", and that multiple mechanisms are at play including the recent trends toward positive SAM values in the Southern Ocean and with reduced winds over the Agulhas Return Current:

S. S. Drijfhout, A. T. Blaker, S. A. Josey, A. J. G. Nurser, B. Sinha and M. A. Balmaseda, (2014), "Surface warming hiatus caused by increased heat uptake across multiple ocean basins" Geophysical Research Letters, DOI: 10.1002/2014GL061456

http://onlinelibrary.wiley.com/doi/10.1002/2014GL061456/abstract

Abstract: "The first decade of the twenty-first century was characterised by a hiatus in global surface warming. Using ocean model hindcasts and reanalyses we show that heat uptake between the 1990s and 2000s increased by 0.7 ± 0.3Wm−2. Approximately 30% of the increase is associated with colder sea surface temperatures in the eastern Pacific. Other basins contribute via reduced heat loss to the atmosphere, in particular the Southern and subtropical Indian Oceans (30%), and the subpolar North Atlantic (40%). A different mechanism is important at longer timescales (1960s-present) over which the Southern Annular Mode trended upwards. In this period, increased ocean heat uptake has largely arisen from reduced heat loss associated with reduced winds over the Agulhas Return Current and southward displacement of Southern Ocean westerlies."

See also:

http://www.reportingclimatescience.com/news-stories/article/heat-uptake-by-several-oceans-drives-pause-says-study.html
“It is not the strongest or the most intelligent who will survive but those who can best manage change.”
― Leon C. Megginson

Steven

  • Grease ice
  • Posts: 553
    • View Profile
  • Liked: 129
  • Likes Given: 16
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #58 on: November 13, 2014, 05:06:33 PM »
Jai,

I get what you are saying steven, the MMMs all suggested a higher OCH absorption ratio from the southern hemisphere than the observational data showed. 

The proportional ratios now fit the MMMs ratios though the scale is a bit larger (10-30%) of total heat accumulation.

Agreed.

I guess the real question with regard to my observation of the underestimation of NH aerosol values is if the current aerosol forcing values are derived from models, from observations or from models and constrained by observations?  If they relied on the models then this would indicate a significant adjustment is needed, likely adjusting the NH system down as well as the SH system up.

I boldfaced the word "models" in the above quote.  You probably meant to write "observations" rather than "models"? 

I don't think the results in Durack et al. can be used to constrain the aerosol forcing.  The paper suggests the global ocean heat uptake in recent decades was somewhat higher than the CMIP multi-model mean.  But this doesn't necessarily imply that climate sensitivity is higher than expected.

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #59 on: November 13, 2014, 07:58:55 PM »
Steven

If the models were correct, then the adjustment would have been to adjust the NH OHC values down, not the Southern values UP.

According to Gavin Schmidt, this does raise the ECS, however, he only projected a 15% increase of potential maximum, and so probably relied on the median value.

https://twitter.com/ClimateOfGavin/status/518928639615070210

any way you look at it, the revision increases the total global heat accumulation by a significant figure, therefore the TOA estimates are to be revised higher, significantly. 

If the TOA values are revised higher, and the NH OHC values remain the same then there MUST be an upward revision to NH aerosols to keep things in balance.  If the NH Aerosols are revised (more negative) then the GHG forcing effects must also be revised upward to compensate.

do you see a problem with these assumptions? please say!

In my estimation, the increase in Aerosol forcing values is a no-brainer and that the AMO and PDO are both aerosol driven.
Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

Steven

  • Grease ice
  • Posts: 553
    • View Profile
  • Liked: 129
  • Likes Given: 16
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #60 on: November 13, 2014, 10:45:23 PM »
According to Gavin Schmidt, this does raise the ECS, however, he only projected a 15% increase of potential maximum, and so probably relied on the median value.

https://twitter.com/ClimateOfGavin/status/518928639615070210

Gavin Schmidt was referring to 1 specific paper about climate sensitivity: Lewis and Curry 2014.  It turns out that Durack et al. leads to an upward adjustment of the upper bound for climate sensitivity in the Lewis and Curry paper.  So Gavin Schmidt was referring to 1 specific paper, and he was not referring to climate sensitivity in general.  See also here.

If the TOA values are revised higher, and the NH OHC values remain the same then there MUST be an upward revision to NH aerosols to keep things in balance.

As discussed before, the hemispheric distribution of ocean heat uptake (i.e. the ratio of SH versus NH ocean heat uptake) in the Durack et al. paper is in good agreement with the CMIP models.  Moreover, Michael Hauber already pointed out that there is substantial ocean heat transport between the two hemispheres.  I really don't think the results of Durack et al. can be used to constrain the aerosol forcing.

Note that Chapter 7 of the IPCC report gives more information about aerosol forcing estimates.

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #61 on: November 14, 2014, 04:28:39 AM »
Quote
in the Durack et al. paper is in good agreement with the CMIP models.

I am sure you are referring to the NH/SH distribution ratios here, not the global values of OHC (and by derivative, TOA)

as the heat uptake coefficient (eq. 13) increases, so does equilibrium sensitivity.  With temperatures uncorrected, if the ocean heat content is revised up then the heat uptake coefficient (k) goes up and equilibrium temperature also goes up (see lines 436-454)

http://www.ecd.bnl.gov/steve/pubs/ObsDetClimSensy.pdf
Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

Steven

  • Grease ice
  • Posts: 553
    • View Profile
  • Liked: 129
  • Likes Given: 16
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #62 on: November 18, 2014, 09:05:09 PM »
as the heat uptake coefficient (eq. 13) increases, so does equilibrium sensitivity.  With temperatures uncorrected, if the ocean heat content is revised up then the heat uptake coefficient (k) goes up and equilibrium temperature also goes up (see lines 436-454)

http://www.ecd.bnl.gov/steve/pubs/ObsDetClimSensy.pdf

You linked to a 2012 paper by Stephen Schwartz.  Schwartz also authored a 2007 paper, in which he suggested that Equilibrium Climate Sensitivity is extremely low: 1.1 +/- 0.5°C.  That 2007 paper was widely criticized, e.g. here.  Apparently the method in that paper was based on several unrealistic assumptions and simplifications.

From my cursory reading of the Schwartz 2012 paper in your link, it looks like Schwartz has changed his 2007 method to some extent.  I agree that higher estimates for OHC (as suggested by Durack et al.) should increase the estimates for climate sensitivity in the Schwartz 2012 paper, but I'm not motivated enough to make the calculations.

The main conclusion in the Schwartz 2012 paper is that different sets of 20th century forcing data (which include aerosol forcing) yield very different results for his calculated climate sensitivity.  E.g. the last paragraph in his paper is: 

" The analysis presented here, although focusing on observational data, nonetheless rests heavily on the forcings over the twentieth century as calculated by several modeling groups based, ultimately, on measured or modeled changes in atmospheric composition. Of these the forcing due to anthropogenic aerosols is the source of the greatest uncertainty, and it this uncertainty that is mainly responsible for the differences in forcings over the twentieth century. Confident determination of Earth's climate sensitivities thus remains hostage to accurate determination of these forcings."

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #63 on: November 19, 2014, 03:09:51 AM »
Steven,

a more recent paper has a better estimation of forcing, heat accumulation and equilibrium climate sensitivity.

Please also realize that this sensitivity does not include the fast feedbacks that a total loss of arctic sea ice (and subsequent carbon cycle and frozen soil feedbacks) will add to future warming.

Schwartz apparently has a new paper, here:

Schwartz et. al has some very interesting observations.  Did the AR5 suffer a fatal error in allowing bad-faith actors within their consensus model to  intentionally move ECS to the downside?



(open access)
http://onlinelibrary.wiley.com/doi/10.1002/2014EF000273/pdf

Earth’s Climate Sensitivity: Apparent Inconsistencies in Recent
Assessments
Stephen E. Schwartz1, Robert J. Charlson2, Ralph Kahn3 and Henning Rodhe4

Quote
1. What degree of confidence can be placed in the large reduction in the magnitude of
negative aerosol forcing and resultant increase in total forcing
over the industrial period,
as assessed in AR5 versus AR4?

2. Given the increase in forcing adopted by AR5 relative to that of AR4, why is there so
little decrease in the assessment of ECS adopted by AR5
relative to that of AR4, as would
be expected from energy-balance considerations?

3. Why, especially in AR5, is there such a great difference between the likely range of ECS
given in the assessment, 1.5 to 4.5 K/(3.7 W m-2), and that inferred from the likely range
of forcing over the industrial period, together with observed increase in GMST and
planetary heating rate, 1.2 to 2.9 K/(3.7 W m-2)?

4. Why are the values of F – N in the CMIP5 model calculations of climate change over the
twentieth century systematically lower than the range of this quantity determined as the
AR5-assessed likely range
of forcing minus the observed planetary heating rate (Figure
1)?


without regard to the results of the current paper, do you agree then that the increase in the rate of heat accumulation necessarily produces a higher ECS?

It seems that an underestimation of aerosol forcing will also push ECS higher.  do you agree?

Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

Steven

  • Grease ice
  • Posts: 553
    • View Profile
  • Liked: 129
  • Likes Given: 16
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #64 on: November 22, 2014, 10:00:10 PM »
Quote
1. What degree of confidence can be placed in the large reduction in the magnitude of
negative aerosol forcing and resultant increase in total forcing
over the industrial period,
as assessed in AR5 versus AR4?

The current anthropogenic radiative forcing as assessed in AR5 vs. AR4 changed not only due to the modified aerosol forcing estimates, but also due to continued growth in greenhouse gas forcing over the past several years. 

E.g., Fig. 8.16 of the IPCC AR5 report is shown below.  The blue lines correspond to aerosol forcing, red is greenhouse gas, and black is total anthropogenic forcing, for the year 2011.  The green line segments show the corresponding estimates for the year 2005, as assessed in the AR4 report:



Regarding the modified estimates of aerosol forcing in AR5 vs. AR4, Chapter 7 of the AR5 report says:

Quote
...This range [for the Effective Radiative Forcing due to aerosols] was obtained from expert judgement guided by climate models that include aerosol effects on mixed-phase and convective clouds in addition to liquid clouds, satellite studies and models that allow cloud-scale responses. This forcing can be much larger regionally but the global mean value is consistent with several new lines of evidence suggesting less negative estimates for the ERF due to aerosol-cloud interactions than in AR4. {7.4, 7.5.3, 7.5.4, Figure 7.19}

Regarding the Schwartz et al. 2014 paper to which you linked above: from my reading of that paper, I don't think that Figure 1 of that paper is an apples-to-apples comparison.  The pink circles in that Figure correspond to the CMIP5 models.  However the equilibrium climate sensitivity (ECS) and forcing (F) associated to these models, as used by Schwartz et al. in the Figure, turn out to be different from the "usual" definitions.  I would take that Schwartz et al. Figure with a grain of salt...
 
« Last Edit: November 23, 2014, 04:20:34 PM by Steven »

AbruptSLR

  • Multi-year ice
  • Posts: 17535
    • View Profile
  • Liked: 741
  • Likes Given: 236
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #65 on: November 23, 2014, 03:02:42 PM »
Personally, I do not find Steven's revelations very comforting.  When the RCP scenarios were being developed it was recognized that:
(a) The average anthropogenic forcing was increasing (as we have followed a BAU pathway in the years since the RCP scenarios were developed), and in order to counter the influence of this trend, the developers of the RCP scenarios decided that by the intelligent application of regulations controlling emissions that they could decrease the uncertainty range (narrowed confidence levels) on these emissions, as can be seen by the first attached figure comparing the averages and ranges of the SRES and the RCP scenarios (it is reasonable to ignore RCP 3/2.6 as this has such a low probability of occurrence).  While it may be the case that intelligent regulations have been applied in developed economies like the USA and the EU; in the developing world and China this has not been the case.  Therefore the reduced uncertainty ranges (narrowed confidence levels) of the RCP scenarios should be taken with a grain of salt, while the increases in the average forcing very clearly have occurred.
(b)  The fact that James Hansen challenged the magnitude of the negative forcing of the aerosols assumed in AR4, resulted in the increase of this magnitude in AR5.  Thus it is now very clear that as China is very determined to clean-up the air pollution in their country, this will result in a decrease in negative forcing in the next decade (or less), which will result in an acceleration of global warming.
“It is not the strongest or the most intelligent who will survive but those who can best manage change.”
― Leon C. Megginson

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #66 on: November 23, 2014, 07:24:04 PM »
The current anthropogenic radiative forcing as assessed in AR5 vs. AR4 changed not only due to the modified aerosol forcing estimates, but also due to continued growth in greenhouse gas forcing over the past several years. 
Steve, thanks for your response 


"this is not the forcing that you are looking for"

(Sorry, had to add that in. . .trying to include a bit more humor now and then!)

from the paper, Schwartz stated your point succinctly. . .

Quote
The change in forcing is due largely to a reduction in the magnitude of the negative
aerosol forcing between AR4 and AR5, -1.2 W m-2 in AR4 vs. -0.9 W m-2 in AR5; there is as
well an increase in forcing by greenhouse gases (GHGs), from 1.66 to 1.82 W m-2,

so far your response is in agreement with the paper.

-------------------------------

E.g., Fig. 8.16 of the IPCC AR5 report is shown below.  [. . .truncated]

This is good, love the image but, again, it only asserts what Schwartz has said in the paper. . .

Regarding the modified estimates of aerosol forcing in AR5 vs. AR4, Chapter 7 of the AR5 report says:

Quote
...This range [for the Effective Radiative Forcing due to aerosols] was obtained from expert judgement guided by climate models that include aerosol effects on mixed-phase and convective clouds in addition to liquid clouds, satellite studies and models that allow cloud-scale responses. This forcing can be much larger regionally but the global mean value is consistent with several new lines of evidence suggesting less negative estimates for the ERF due to aerosol-cloud interactions than in AR4. {7.4, 7.5.3, 7.5.4, Figure 7.19}

yes, indeed, they did that.  They have a justification for the reduction in aerosol forcing parameters, the real question is, does this make sense?  Does it fit a 'true-up" constraint using observed heat accumulation and GHG forcing parameters?

Image 1 from Schwartz that I showed (the magenta circles) shows that this statement:

Quote
This range [for the Effective Radiative Forcing due to aerosols] was obtained from expert judgement guided by climate models that include aerosol effects

is highly dubious.  what "expert judgments" were used that showed such a high deviation from the actual modeled values (as shown in Schwartz' figure)?

Schwartz says that the AR5 revisions to (F-N) deviate completely from the models, even more so than the AR4 values.  This says that there are some very big problems with either the models or the "experts" polled by the IPCC for their bias of the results.  The actual modeled (F-N) values imply a 3.0-5.0 ECS and this bias deviation from the models and used by the IPCC point to 1.2-1.6 ECS.  Then the the IPCC produces a synthesis range of estimate values somewhere in the low average of the two. 

I am suggesting that the "expert judgements" incorporate a significant overestimation in the aerosol forcing parameter (coupled with an underestimate of GHG forcing), perhaps intentionally so.

Really, this is the point of Schwartz' paper. . .he is basically saying "WTF???"  you just deviated even MORE from the models than before? Who are your "experts"???

Regarding the Schwartz et al. 2014 paper to which you linked above: from my reading of that paper, I don't think that Figure 1 of that paper is an apples-to-apples comparison.  The pink circles in that Figure correspond to the CMIP5 models.  However the equilibrium climate sensitivity (ECS) and forcing (F) associated to these models, as used by Schwartz et al. in the Figure, turn out to be different from the "usual" definitions.  I would take that Schwartz et al. Figure with a grain of salt...

hmmm, I disagree.  Do you have proof? have you reviewed Taylor et al (2012) and Forster et al (2013) to compare inputs?  I find your assertion highly suspect (sorry) compared to the documentation in Schwartz et al (2014)

to whit

Quote
Also shown in Figure 1 are values of F - N and ECS of 23 coupled atmosphere-ocean general circulation models (GCMs) that participated in the CMIP5 model intercomparison (Taylor et al., 2012) conducted in conjunction with the IPCC AR5 Assessment, as inferred by Forster et al. (2013) from twentieth century climate runs and 4 × CO2 experiments. The GCMs all employed forcings over the twentieth century (up to 2005) that resulted in values of F - N that were less, to substantially less, than those given by the AR5 and Otto et al. assessments.

these appear to be apples to apples to me.
« Last Edit: November 23, 2014, 07:34:43 PM by jai mitchell »
Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

Steven

  • Grease ice
  • Posts: 553
    • View Profile
  • Liked: 129
  • Likes Given: 16
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #67 on: November 23, 2014, 11:10:23 PM »
Jai,

from the paper, Schwartz stated your point succinctly. . .

Quote
The change in forcing is due largely to a reduction in the magnitude of the negative
aerosol forcing between AR4 and AR5, -1.2 W m-2 in AR4 vs. -0.9 W m-2 in AR5; there is as
well an increase in forcing by greenhouse gases (GHGs), from 1.66 to 1.82 W m-2,

This is an error in the Schwartz et al. 2014 paper.  The boldface numbers in the above quote are not the total greenhouse gas forcing, but only the CO2 forcing.  E.g., see Figure 8.16 of the AR5 report, which I already posted in my last post.  That figure shows that the best estimate for total greenhouse gas forcing in AR4 is about 2.9 W m-2, and in AR5 it's about 3.2 W m-2. 

See also Figure 8.15 in AR5:




[...] what "expert judgments" were used [...] ?

In my last post, I provided references to the relevant parts of Chapter 7 of the IPCC report: {Sections 7.4, 7.5.3, 7.5.4, Figure 7.19}.  There's more information there.


Regarding the Schwartz et al. 2014 paper to which you linked above: from my reading of that paper, I don't think that Figure 1 of that paper is an apples-to-apples comparison.  The pink circles in that Figure correspond to the CMIP5 models.  However the equilibrium climate sensitivity (ECS) and forcing (F) associated to these models, as used by Schwartz et al. in the Figure, turn out to be different from the "usual" definitions.  I would take that Schwartz et al. Figure with a grain of salt...

hmmm, I disagree.  Do you have proof? have you reviewed Taylor et al (2012) and Forster et al (2013) to compare inputs?  I find your assertion highly suspect (sorry) compared to the documentation in Schwartz et al (2014)

Take a look at the supplementary information of the Schwartz et al. 2014 paper, more precisely Table S2.

According to that table, the two CMIP5 models with the highest ECS values are the

CSIRO-Mk3-6-0 model   (ECS:  5.87°C)
HadGEM2-ES model       (ECS:  5.78°C) 

However, I looked it up in Table 1 of the Forster et al. 2013 paper, which is the source of the Schwartz et al. numbers.  It turns out that the "true" ECS associated to the latter 2 models is 4.08°C and 4.59°C respectively.  What Schwartz et al. call "ECS" is based on an alternative, simplified formula which also uses some model output from the Forster et al. paper.  So Schwartz et al. are not using the ECS as used in the Forster et al. paper, or as used in the AR5 report for that matter, e.g. see Fig. 9.42 of the AR5 report (horizontal axis):



Moreover, the forcing (F) associated to the CMIP5 models, as used by Schwartz et al., turns out to be "adjusted forcing", as discussed in the Forster et al. 2013 paper.  This cannot be directly compared to the assessment of anthropogenic forcing as used in the AR5 report. 

Moreover, for the estimates of forcing, Schwartz et al. use model output for the year 2003, whereas the AR5 report assesses the forcing for the year 2011.  Throughout the paper, Schwartz et al. are comparing data from different years or periods.  It's not apples-to-apples.

In my opinion, the Schwartz et al. 2014 paper should be taken with a grain of salt.  I don't think this is very surprising, given Schwartz' earlier work on this subject; upthread I've already mentioned the absurdly low estimate for ECS in the Schwartz 2007 paper.
 
 
« Last Edit: November 24, 2014, 12:12:14 AM by Steven »

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #68 on: November 24, 2014, 08:26:15 AM »
Steven,

your response is excellent, I am still working through it, What I have found so far is that Schwartz' methodology appears to be a way to "true-up" multiple models with varying ancillary feedback parameters.  He references 3 papers that use this same methodology to cross check different models, and this methodology appears to be robust.

In Forster data (table 1) the individual GCMs held a .31 Watt/Meter^2 standard deviation for net feedbacks.  That is a huge difference and it is this feedback that caused the models to have 'adjusted' ECS values.

I believe that this is the difficulty in performing comparisons between models so Schwartz is pulling out the underlying data as a way to perform cross-comparison. 

It is a good catch and I understand what you are saying.  I think that perhaps the issue isn't as much the absolute value of ECS (with or without feedbacks) used in the different papers (and the IPCC), but rather to challenge what it is that has made the IPCC values so much lower than the model values for the underlying framework of analysis?

one note - when you said

Quote
Moreover, for the estimates of forcing, Schwartz et al. use model output for the year 2003, whereas the AR5 report assesses the forcing for the year 2011.

I don't think you are reading this right, Schwartz used a range value as did the IPCC

from the paper:

Quote
For forcing we use the time periods and values given in the several assessments

« Last Edit: November 24, 2014, 08:35:44 AM by jai mitchell »
Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

Steven

  • Grease ice
  • Posts: 553
    • View Profile
  • Liked: 129
  • Likes Given: 16
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #69 on: November 25, 2014, 12:50:51 AM »
...Schwartz [...] references 3 papers that use this same methodology to cross check different models, and this methodology appears to be robust.

(Edit: I've rewritten the next few paragraphs). 

There seems to be a problem with the way in which Schwartz et al. apply this methodology.

Schwartz' formula for the "ECS" of the CMIP5 models is: ECS = 3.7 / λ, where λ is the climate feedback parameter associated to the model.  The numeric values for this parameter λ are taken from Forster et al. 2013 (it's called alpha there, and determined from abrupt 4xCO2 simulations).

The problem seems to be that this formula uses a fixed numerator 3.7.  The idea is that a doubling of atmospheric CO2 corresponds to a radiative forcing of about 3.7 W m-2.  However, in this context, I think Schwartz et al. should have used a model-dependent numerator; in fact they should have used the "Adjusted Forcing" data in the Forster et al. paper, Table 1, which can be markedly different from 3.7.


Quote
Moreover, for the estimates of forcing, Schwartz et al. use model output for the year 2003, whereas the AR5 report assesses the forcing for the year 2011.

I don't think you are reading this right, Schwartz used a range value as did the IPCC

In the above quote I was referring to Schwartz' forcing (F) data for the CMIP5 models, i.e. the pink circles in Figure 1 of the paper.  As explained in the supplementary information of the paper, the source of those forcing data is Forster et al. 2013, Table 2, first column.  These forcing data are based on model output for the year 2003.  Also, it's Adjusted Forcing (AF) which, as discussed by Forster et al., is usually lower than the true Radiative Forcing.

 
« Last Edit: November 25, 2014, 11:50:43 AM by Steven »

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #70 on: November 25, 2014, 07:05:30 PM »
Steven,

I believe that you have a fundamental error in your analysis.  The adjusted ECS value in Forster is a function of the 3.7 2XCO2 forcing and λ values  in fact, it is exactly the following (adjustedECS / 3.7 Wm^-2) = λ

So it turns out that Schwartz actually did use the adjusted ECS in his calculation.   

If I use the values of table 1 in Forster and divide the first model's ECS by the adjusted forcing for 2XCO2 I get the following (3.83K / 2.98 Wm^-2) = 1.28 K/Wm^-2 (table here:  http://eprints.whiterose.ac.uk/76111/22/JGRA_PFjgrd50174%5B1%5D_with_coversheet.pdf )

This is the value that Schwartz used in his determination of a model normalized ECS value using 3.7wm^-2 per 2XCO2.  (1.28 * 3.7 = 4.74 )

Note that 4.74 K is the ECS value for the first model in Schwartz' supplementary table.  http://onlinelibrary.wiley.com/store/10.1002/2014EF000273/asset/supinfo/eft253-sup-0002-TableS2.pdf?v=1&s=fa3d2d2b500d722456c9799bdd78c4a305558aad


If I used your equation then I would get a value of 3.7 as ECS for every model since you are basically substituting out the λ.   (normalized ECS = Adjusted ECS/ λ  and Adjusted ECS = λ * 3.7 ).

In addition, if the math errors of Schwartz was really so egregious and simple as that, he would not have passed peer review. 

Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

AbruptSLR

  • Multi-year ice
  • Posts: 17535
    • View Profile
  • Liked: 741
  • Likes Given: 236
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #71 on: November 26, 2014, 12:43:38 AM »
If anyone has the patience to read through the long dialogue on climate sensitivity at the following website, it addresses many of the points being discussed in this thread:

http://www.climatedialogue.org/climate-sensitivity-and-transient-climate-response/
“It is not the strongest or the most intelligent who will survive but those who can best manage change.”
― Leon C. Megginson

Steven

  • Grease ice
  • Posts: 553
    • View Profile
  • Liked: 129
  • Likes Given: 16
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #72 on: November 26, 2014, 12:36:36 PM »
Jai,

The adjusted ECS value in Forster is a function of ...

"adjusted ECS"?  That phrase is not used in the Forster et al. 2013 paper.  There is only 1 set of ECS values in that paper (Table 1), and they are exactly the same as the ECS values mentioned in Chapter 9 of the IPCC report.  The ECS values calculated in the Schwartz et al. paper, on the other hand, are aberrant.


...in fact, it is exactly the following (adjustedECS / 3.7 Wm^-2) = λ

Hmm, where does that formula come from?  This is not correct.


If I used your equation then I would get a value of 3.7 as ECS for every model since you are basically substituting out the λ.   (normalized ECS = Adjusted ECS/ λ  and Adjusted ECS = λ * 3.7 ).

You have to read it more carefully.  I was not referring to "adjusted ECS", but to Adjusted Forcing.

What I was saying is that the correct formula is: "ECS = AF / λ", where AF stands for the Adjusted Forcing values in Forster et al. Table 1.  (λ is the climate feedback parameter, which is called alpha in Forster et al.) 

Schwartz et al. use a different formula: "ECS = 3.7 / λ".  So their calculated ECS values are different from those in Forster et al. (and hence different from those in Chapter 9 of the IPCC report).  It's quite weird that the numerator of Schwartz' formula is a constant 3.7, rather than the model-calculated Adjusted Forcing.  Also, note that the multimodel mean of the AF values in Forster Table 1 is 3.44 W m-2, and not 3.7.


In addition, if the math errors of Schwartz was really so egregious and simple as that, he would not have passed peer review.

It's not a question of "simple math errors".  It's a question of precise definitions and time periods of the data that Schwartz et al. are using.  This requires reading in detail the Forster et al. paper, which is the main source of Schwartz' data. 

The fact that a paper is published in some journal doesn't necessarily mean that everything is correct.  Schwartz et al. are making extraordinary claims.  I already pointed out several problems with the paper.  Figure 1 of the paper is not an apples-to-apples comparison.
 
 
« Last Edit: November 26, 2014, 06:49:03 PM by Steven »

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #73 on: November 26, 2014, 10:01:56 PM »
Sorry I used ECS when I meant to use forcing.  Lets try this one last time shall we?

Each model has different 2XCO2 forcing values.  To find out how they compare directly to each other one must find out how each one responds to a single Watt per meter squared of forcing.

To find this value all one needs to do is divide each model's ECS value by each model's adjusted forcing value and you get the amount of warming that occurs for each model under 1 watt per meter squared of forcing.

this term is called Seq in Schwartz.

Then, to show how the different models compare to each other under an IDENTICAL forcing scenario. Schwartz simply multiplies each model's Seq value by the IPCC's equilibrium forcing value of 3.7 Watts per meter squared.

This is how he compared the models.  This is why it makes sense.  If you try doing it your way (please do and post your math!) you will find that each model produces identical values for ECS.

As Schwartz et al stated:

Quote
we present equilibrium
climate sensitivity in the unit K/(3.7 W m-2) evaluated numerically as ECS = 3.7 Seq. The
unit K/(3.7 W m-2) preserves the familiar numerical values associated with climate system
response to doubled CO2 while removing ambiguity due to differences in values of this
forcing used in the several studies examined.

WRT the differences in dates, the GCMs in Forster et al all have an end date of 2005.  While the IPCC evaluated to 2011.

If anything, this last bit proves that increased lobbying and external media pressures on the "hiatus" in 2012 led to increased conservatism within the IPCC results.  This conservatism shows a (potentially) significantly underestimated ECS value and further exacerbates continued BAU emissions within an environment of complete policy failure (if higher ECS values were considered).
Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

Steven

  • Grease ice
  • Posts: 553
    • View Profile
  • Liked: 129
  • Likes Given: 16
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #74 on: November 27, 2014, 12:39:35 AM »
...If you try doing it your way (please do and post your math!) you will find that each model produces identical values for ECS.

Huh?  I already posted the math.  Here is what I wrote:

...the correct formula is: "ECS = AF / λ", where AF stands for the Adjusted Forcing values in Forster et al. Table 1.  (λ is the climate feedback parameter, which is called alpha in Forster et al.) 

The data in Forster et al. Table 1 satisfy this relationship.  E.g. for the first model (ACCESS1-0) in that table:

ECS = 3.83 K    (K stands for kelvin)
AF = 2.98  W m-2
λ = 0.78 W m-2 K-1  (this is called alpha in the table)

These 3 numbers indeed satisfy the relationship in the above quote: 3.83 = 2.98 / 0.78  (ignoring rounding issues).

Now how does Schwartz calculate the "ECS" for that same model (ACCESS1-0)?  Here is his calculation:

"ECS" = 3.7 / λ = 3.7 / 0.78 = 4.74  (with the same units as above)

So Schwartz calculates "ECS" = 4.74°C, whereas Forster's value was ECS = 3.83°C.  So what is Schwartz saying here?  That Forster is wrong?  And, consequently, that the ECS values in Chapter 9 of the IPCC report are wrong?  And apparently nobody spotted that major error, except Schwartz et al.?  EDITED TO ADD: Clearly these last few sentences were sarcastic.  In reality, I think it's quite obvious that Schwartz' formula is just a rough approximation for the "true" ECS values, which are the ones calculated in Forster's paper.
« Last Edit: November 27, 2014, 01:55:31 AM by Steven »

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #75 on: November 27, 2014, 05:23:14 AM »
...If you try doing it your way (please do and post your math!) you will find that each model produces identical values for ECS.

Huh?  I already posted the math.  Here is what I wrote:

...the correct formula is: "ECS = AF / λ", where AF stands for the Adjusted Forcing values in Forster et al. Table 1.  (λ is the climate feedback parameter, which is called alpha in Forster et al.) 

The data in Forster et al. Table 1 satisfy this relationship.  E.g. for the first model (ACCESS1-0) in that table:

ECS = 3.83 K    (K stands for kelvin)
AF = 2.98  W m-2
λ = 0.78 W m-2 K-1  (this is called alpha in the table)

These 3 numbers indeed satisfy the relationship in the above quote: 3.83 = 2.98 / 0.78  (ignoring rounding issues).

Now how does Schwartz calculate the "ECS" for that same model (ACCESS1-0)?  Here is his calculation:

"ECS" = 3.7 / λ = 3.7 / 0.78 = 4.74  (with the same units as above)

So Schwartz calculates "ECS" = 4.74°C, whereas Forster's value was ECS = 3.83°C.  So what is Schwartz saying here?  That Forster is wrong?  And, consequently, that the ECS values in Chapter 9 of the IPCC report are wrong?  And apparently nobody spotted that major error, except Schwartz et al.?  EDITED TO ADD: Clearly these last few sentences were sarcastic.  In reality, I think it's quite obvious that Schwartz' formula is just a rough approximation for the "true" ECS values, which are the ones calculated in Forster's paper.

It is clear that you are not thinking about this clearly,  The point of Schwartz was to provide direct comparison between models.  simply using the individual model outputs found in forster for adjusted forcing and ECS doesn't do that.

To compare apples to apples, as you say, you must first normalize the models with their different warming rates for the same amount of forcing.  in doing that then we can project what the model outputs would be if they were all acting on a single, normalized, value of 2XCO2 forcing.

This is what Schwartz did.

you are basically arguing that he should have simply reproduced forster's work, by comparing the model outputs of ECS only.  This is not the point of the work.  The point is to provide apples to apples comparisons between GCMs and the IPCC.

It is clear that the GCMs output a much higher ECS than the current IPCC range. 



Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

Michael Hauber

  • Grease ice
  • Posts: 841
    • View Profile
  • Liked: 46
  • Likes Given: 12
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #76 on: November 28, 2014, 02:14:42 AM »
Two observations on Schwartz:

1) His assessment of ECS in the models relies on energy balance calculations.  The standard ECS of a model is not normally calculated in this way, but measured by running the model with 2xCo2 and seeing what the warming is.  Gregory et al in 2012 perform a similar calculation and discuss the issues involved.  Overall Gregory finds that sensitivity diagnosed from energy balance considerations is quite similar to that diagnosed from a doubling experiment.   When comparing Gregory values for individual models some are the same as Schwartz (eg CNRM-CM5 is 3.25 in both), but others are quite different - CSIRO-Mk3-6-0 is 4.08 in Gregory, and 5.87 in Schwarz.

2) Schwartz argues that the range for ECS inferred from the likely range of forcing is 1.2 to 2.9, which is lower than the IPCC range of 1.5 to 4.5.  The final line for the abstract is that explanations for the observed discrepancies may involve (among other things) an overestimated climate sensitivity.  It is a misuse of a supplementary table from Schwartz et al to argue that IPCC have underestimated climate sensitivity in contradiction to the abstract which clearly states that IPCC may have overestimated sensitivity.



Climate change:  Prepare for the worst, hope for the best, expect the middle.

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #77 on: November 30, 2014, 02:38:35 PM »
Two observations on Schwartz:

1) His assessment of ECS in the models relies on energy balance calculations.  The standard ECS of a model is not normally calculated in this way, but measured by running the model with 2xCo2 and seeing what the warming is.  Gregory et al in 2012 perform a similar calculation and discuss the issues involved.  Overall Gregory finds that sensitivity diagnosed from energy balance considerations is quite similar to that diagnosed from a doubling experiment.   When comparing Gregory values for individual models some are the same as Schwartz (eg CNRM-CM5 is 3.25 in both), but others are quite different - CSIRO-Mk3-6-0 is 4.08 in Gregory, and 5.87 in Schwarz.

2) Schwartz argues that the range for ECS inferred from the likely range of forcing is 1.2 to 2.9, which is lower than the IPCC range of 1.5 to 4.5.  The final line for the abstract is that explanations for the observed discrepancies may involve (among other things) an overestimated climate sensitivity.  It is a misuse of a supplementary table from Schwartz et al to argue that IPCC have underestimated climate sensitivity in contradiction to the abstract which clearly states that IPCC may have overestimated sensitivity.

he is saying that the forcing used by the IPCC would imply a range of 1.2 to 2.9  he is not saying that this is correct, in fact he implies that the reduced expected impact of aerosols is the underlying inconsistency between the IPCC values of ECS and a correlation to (F-N) as well as an overall reality check in comparison between the IPCC ECS values and the normalized GCM values.

It is clear that the massive underestimation of aerosol forcing has skewed their projections (of ECS lower) to untenable values.
Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

Michael Hauber

  • Grease ice
  • Posts: 841
    • View Profile
  • Liked: 46
  • Likes Given: 12
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #78 on: December 01, 2014, 04:44:03 AM »

he is saying that the forcing used by the IPCC would imply a range of 1.2 to 2.9  he is not saying that this is correct, in fact he implies that the reduced expected impact of aerosols is the underlying inconsistency between the IPCC values of ECS and a correlation to (F-N) as well as an overall reality check in comparison between the IPCC ECS values and the normalized GCM values.

It is clear that the massive underestimation of aerosol forcing has skewed their projections (of ECS lower) to untenable values.

Schwartz notes an inconsistency when he calculates climate sensitivity based on IPCC forcings.  He then lists several explanations, which can be simplified as climate sensitivity is lower, or aerosol forcing is higher, or something else.  That is if climate sensitivity is not lower (as the calculations suggest), then either aerosol forcing is higher, or something else.

Schwarz makes not specific claim, and provides no evidence that aerosol forcing should be increased (in comparison to any of the other options he lists).  Schwarz definitely does not state that climate sensitivity should be increased - but if anything states that climate sensitivity should be decreased.
Climate change:  Prepare for the worst, hope for the best, expect the middle.

wili

  • Young ice
  • Posts: 2589
    • View Profile
  • Liked: 163
  • Likes Given: 301
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #79 on: December 01, 2014, 05:05:05 AM »
Wow, you two seem to be looking at the same article and coming to opposite conclusions. Perhaps if specific passages were quoted or at least pointed to that you think support your interpretation, others could weigh in, or perhaps you'd figure how where one or both of you erred in your interp.
"A force de chercher de bonnes raisons, on en trouve; on les dit; et après on y tient, non pas tant parce qu'elles sont bonnes que pour ne pas se démentir." Choderlos de Laclos "You struggle to come up with some valid reasons, then cling to them, not because they're good, but just to not back down."

Steven

  • Grease ice
  • Posts: 553
    • View Profile
  • Liked: 129
  • Likes Given: 16
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #80 on: December 01, 2014, 09:51:40 PM »
2) Schwartz argues that the range for ECS inferred from the likely range of forcing is 1.2 to 2.9, which is lower than the IPCC range of 1.5 to 4.5.  The final line for the abstract is that explanations for the observed discrepancies may involve (among other things) an overestimated climate sensitivity.  It is a misuse of a supplementary table from Schwartz et al to argue that IPCC have underestimated climate sensitivity in contradiction to the abstract which clearly states that IPCC may have overestimated sensitivity.

Agreed.  I wasn't suggesting that IPCC have underestimated (or overestimated) climate sensitivity, but rather that Schwartz' methodology is different from what is normally used.

Perhaps I should return to the "big picture": Schwartz et al. Figure 1.  The purple circles in that Figure suggest a correlation between ECS and F-N, in the context of CMIP5 models.  But this seems to be (at least partly) due to Schwartz' methodology.  If anyone knows another paper which suggests the existence of such a correlation (statistically significant?) in the context of CMIP5 models, then I would be interested.

Michael Hauber

  • Grease ice
  • Posts: 841
    • View Profile
  • Liked: 46
  • Likes Given: 12
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #81 on: December 02, 2014, 04:35:29 AM »
Wow, you two seem to be looking at the same article and coming to opposite conclusions. Perhaps if specific passages were quoted or at least pointed to that you think support your interpretation, others could weigh in, or perhaps you'd figure how where one or both of you erred in your interp.

The end of the abstact:

Quote
Importantly, the likely range of ECS to doubled CO2 given in AR5, 1.5 to 4.5 K/(3.7 W m-2) exceeds the range inferred from the assessed likely range of forcing, 1.2 to 2.9 K/(3.7 W m-2), where 3.7 W m-2 denotes the forcing for doubled CO2. Such differences underscore the need to identify their causes and reduce the underlying uncertainties. Explanations might involve underestimated negative aerosol forcing, overestimated total forcing, overestimated climate sensitivity, poorly constrained ocean heating, limitations of the energy balance model, or a combination of effects.
Climate change:  Prepare for the worst, hope for the best, expect the middle.

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #82 on: December 02, 2014, 10:17:32 PM »

Agreed.  I wasn't suggesting that IPCC have underestimated (or overestimated) climate sensitivity, but rather that Schwartz' methodology is different from what is normally used.

Perhaps I should return to the "big picture": Schwartz et al. Figure 1.  The purple circles in that Figure suggest a correlation between ECS and F-N, in the context of CMIP5 models.  But this seems to be (at least partly) due to Schwartz' methodology.  If anyone knows another paper which suggests the existence of such a correlation (statistically significant?) in the context of CMIP5 models, then I would be interested.

Steven,

Several papers using this methodology are referenced within Schwartz.

Quote
Equation 2, which is widely used in interpretation of
observed planetary temperature change (AR5, §10.8.1, p. 920; Knutti and Hegerl, 2008;
Hansen et al., 2011) and which found to hold to good accuracy in analyses of climate model
responses (e.g. Andrews et al., 2012; Forster et al., 2013)
and
Quote
The black diagonal line
in Figure 1 (slope -1 on log-log plot) shows the relationship between ECS and F - N
given by Eq (2). The observed increase in GMST and planetary heating rate together with
Eq (2) constrain the values of ECS and forcing; if the value of one is known or assumed, then
the value of the other is determined.
image again for reference.




Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #83 on: December 02, 2014, 10:37:31 PM »
Wow, you two seem to be looking at the same article and coming to opposite conclusions. Perhaps if specific passages were quoted or at least pointed to that you think support your interpretation, others could weigh in, or perhaps you'd figure how where one or both of you erred in your interp.

The end of the abstact:

Quote
Importantly, the likely range of ECS to doubled CO2 given in AR5, 1.5 to 4.5 K/(3.7 W m-2) exceeds the range inferred from the assessed likely range of forcing, 1.2 to 2.9 K/(3.7 W m-2), where 3.7 W m-2 denotes the forcing for doubled CO2. Such differences underscore the need to identify their causes and reduce the underlying uncertainties. Explanations might involve underestimated negative aerosol forcing, overestimated total forcing, overestimated climate sensitivity, poorly constrained ocean heating, limitations of the energy balance model, or a combination of effects.

(underestimated negative forcing) --> lowers the F-N value to be more in line with higher ECS
(overestimated total forcing) --> lowers 2XCO2 forcing value to be closer in line with range
(poorly constrained ocean heating) --> same as underestimated neg aerosol forcing
(limitations of energy balance model) --> may be either lowers forcing or raises F-N value
(combination) --> Same as above
(Overestimated Climate Sensitivity) --->will bring the IPCC ECS back in line with their stated (F-N) values, however this will then conflict even more with the GCMs and therefore is unlikely.  I see this point almost as tongue in cheek, given the evidence presented.

I like

Quote
One possible explanation for the apparent inconsistencies between estimated forcing and
sensitivity in the recent assessments may be that the measurements leading to best estimates
of increase in GMST and/or planetary heating rate are erroneous or have not sufficiently
sampled the planet to provide an adequate assessment.

and

Quote
Recent re-examination of modelbased
estimates of aerosol direct forcing (Samset et al., 2014) has led to a slight increase in
the estimated uncertainty over that given in AR5, but this increase (presented as a lower
bound) does little to resolve the inconsistencies highlighted here.

In other words, it looks like there is much less energy being absorbed than we thought and that the aerosol forcing being stronger negative (samset et al 2014 indicated slightly more positive black carbon component) would explain a lower F-N value and a respondent higher derived ECS.
Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

Steven

  • Grease ice
  • Posts: 553
    • View Profile
  • Liked: 129
  • Likes Given: 16
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #84 on: December 04, 2014, 11:40:34 PM »
Jai: I was referring to the correlation between F-N and ECS across the CMIP5 models.  I.e., the correlation between the horizontal and the vertical position of the purple circles in Schwartz' Fig. 1

The location of those purple circles, and the correlation between their horizontal and vertical coordinate, depend on the simple energy balance method used in the paper. 

For example, if Schwartz had used the "true" ECS values of the CMIP5 models (as used in Andrews, Forster, etc.) then the positions of the purple circles would be different.  In that case they would be roughly at the locations of the black circles shown below:


jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #85 on: December 05, 2014, 02:50:48 AM »

Steven,

I understand what you were saying, of course!

The graph is very good, I am sure that took some time to make it,

I used your graph and posted the forcing values associated with each data point.  As you can see, the temperature response between models for a given forcing value varies wildly!

This is why simply using the ECS values doesn't work, as you can see the correlated 2XCO2 forcing values vary wildly between models and each has a very different temperature response for the same forcing.

The graph that you posted is definitely not APPLES to APPLES!

Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

Steven

  • Grease ice
  • Posts: 553
    • View Profile
  • Liked: 129
  • Likes Given: 16
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #86 on: December 05, 2014, 05:38:10 PM »
The numbers that you included next to the circles in the above graph are forcing data from 2xCO2 simulations.  This is different from the forcing F used on the horizontal axis of the graph.  In fact the latter uses estimates for year 2003, obtained by running the models with historical rather than future scenarios.

In Schwartz' graph, the assessments of the current F-N (for 2003 or other recent years) are projected on some diagonal line to determine a corresponding range of ECS values.  I would think that the best metric to use in that context is the "true" ECS of the CMIP5 models, rather than Schwartz' ECS values.  Anyway, we can agree to disagree about that.

What I was looking for is other papers (different from Schwartz) about the correlation between ECS and F-N in recent climate models.  The references in Schwartz' paper don't seem to discuss that.  Schwartz refers to Kiehl 2007, but that paper seems outdated in this context.  Another relevant reference is Forster et al. 2013, which finds that the correlation between ECS and F is considerably weaker in CMIP5 models than in older models.  Forster also finds that a better correlation can be obtained by selecting a certain subset of the CMIP5 models, namely those models that are within the 90% uncertainty range of the observed temperature trend over the past 100 years.  The latter subset of models is shown in green in Forster Fig. 7:
« Last Edit: December 05, 2014, 06:55:48 PM by Steven »

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #87 on: December 05, 2014, 09:21:31 PM »
I agree that the historically observed rates of forcing, minus the observed rates of heat accumulation (the term F-N) is different from the absolute forcing value at 2XCO2 associated with the ECS value.

The reason that I introduced this paper here is because I wanted to show how an underestimation of aerosol forcing in recent models works to produce an underestimation in ECS (by reducing the N value, the term F-N goes up and the associated ECS value goes down).

again, if you want to find the references that used the Term (F-N) and its relationship to ECS, simply go to the referenced documents in Schwartz (since that is what we are talking about)

In looking for usage of the F-N correlation (constraint) to ECS have you checked the following?  The AR5 only seems to imply its use as a constraint relative to TCR. 
Quote
AR5, §10.8.1, p. 920; Knutti and Hegerl, 2008;
Hansen et al. 2011;

The reason that Schwartz needed to normalize to the 2XCO2 forcing of 3.7 is because this is what is used in the AR5.  Trying to compare all of the GCMs' ECS values to the AR5 (and really, each other in the context of F-N) independently of this normalized value is impossible.

Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

Steven

  • Grease ice
  • Posts: 553
    • View Profile
  • Liked: 129
  • Likes Given: 16
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #88 on: December 05, 2014, 10:08:49 PM »
Googling, I found a reference to an earlier submitted (to another journal) version of the Schwartz et al. 2014 paper.  Oddly, this version has five rather than four authors, including someone called L. Bengtsson (source):

Quote
Schwartz, S., L. Bengtsson, R.J. Charlson, R.A. Kahn, and H. Rodhe (2014).
Earth's Climate Sensitivity: Apparent Inconsistencies in Recent Analysis
Rem. Sens. Environ. Lett., (Submitted).

This seems to refer to Lennart Bengtsson.  He has been the subject of a "controversy" in May 2014 and has been quoted as saying that "colleagues are withdrawing from joint authorship".  E.g.:

http://scienceblogs.com/gregladen/2014/05/15/lennart-bengtsson-joins-quits-denialist-think-tank-cries-mccarthyism

The link below may also be worth reading.  Many of the details discussed by Referee one and Referee two in that link seem very similar to, and possibly relevant to, the Schwartz et al. 2014 paper:

http://ioppublishing.org/newsDetails/statement-from-iop-publishing-on-story-in-the-times

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #89 on: December 06, 2014, 11:04:49 PM »
That is very interesting, good catch, it seems that their original paper as submitted was not clear in suggesting what changes needed to be made.  The other authors seem to be very legitimate in their respective fields and their response to a critique of their 2010 paper that brought up some of these same considerations is also consistent with this latter work.

response here:  http://www.ecd.bnl.gov/steve/abstracts/ResponseKnuttiPlattner.html

Quote
2) the present uncertainty in climate sensitivity precludes determination even of the sign of the amount of future CO2 emissions that would be allowed so as not to exceed a given increase in GMST, and 3) that the only realistic way to reduce these uncertainties is to greatly reduce the uncertainty in aerosol forcing

It appears from the reviewer notes that the original comparison was strictly a (F-N) to ECS comparison which was definitely not correct.   Obviously they had to rewrite it I don't know if the GWPF denialist was booted or left voluntarily as the revisions most likely weakened his argument of lower ECS values.

Regardless of their intentions and potential insinuations, I am reading it and interpreting it as an implication of significant underestimation of the negative forcing value for Aerosols as well as the indication that the recent adjustments of observed heat accumulation (10-30% increases in Durack et. al 2014) would make the the N in F-N that much larger and that both of these uncertainties may indicate a significantly underestimated ECS value.
« Last Edit: December 07, 2014, 02:41:27 PM by jai mitchell »
Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

AbruptSLR

  • Multi-year ice
  • Posts: 17535
    • View Profile
  • Liked: 741
  • Likes Given: 236
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #90 on: February 03, 2015, 05:59:14 PM »
The linked article indicates that 67 to 98% of the heat gain in the oceans from 2006 to 2013 occurred in the Southern Hemisphere (see the red areas of the attached plot of the ocean heat gain in the 0-2000 m layer of water):

Dean Roemmich, John Church, John Gilson, Didier Monselesan, Philip Sutton & Susan Wijffels, (2015), "Unabated planetary warming and its ocean structure since 2006", Nature Climate Change, doi:10.1038/nclimate2513


http://www.nature.com/nclimate/journal/vaop/ncurrent/full/nclimate2513.html

Abstract: "Increasing heat content of the global ocean dominates the energy imbalance in the climate system. Here we show that ocean heat gain over the 0–2,000 m layer continued at a rate of 0.4–0.6 W m−2 during 2006–2013. The depth dependence and spatial structure of temperature changes are described on the basis of the Argo Program's accurate and spatially homogeneous data set, through comparison of three Argo-only analyses. Heat gain was divided equally between upper ocean, 0–500 m and 500–2,000 m components. Surface temperature and upper 100 m heat content tracked interannual El Niño/Southern Oscillation fluctuations, but were offset by opposing variability from 100–500 m. The net 0–500 m global average temperature warmed by 0.005 °C yr−1. Between 500 and 2,000 m steadier warming averaged 0.002 °C yr−1 with a broad intermediate-depth maximum between 700 and 1,400 m. Most of the heat gain (67 to 98%) occurred in the Southern Hemisphere extratropical ocean. Although this hemispheric asymmetry is consistent with inhomogeneity of radiative forcing and the greater area of the Southern Hemisphere ocean, ocean dynamics also influence regional patterns of heat gain."
“It is not the strongest or the most intelligent who will survive but those who can best manage change.”
― Leon C. Megginson

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #91 on: February 04, 2015, 02:35:58 AM »
That extreme low insolation zone east of the U.S. is striking.  Is it possible that the Offshore pacific blocking ridge is pushing the Jetstream so far north that the resultant "polar vortex" low pressure surge is producing a kind of "Aerosol amplification" effect streaming in this offshore region?
Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

AbruptSLR

  • Multi-year ice
  • Posts: 17535
    • View Profile
  • Liked: 741
  • Likes Given: 236
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #92 on: February 07, 2015, 12:26:44 AM »
That extreme low insolation zone east of the U.S. is striking.  Is it possible that the Offshore pacific blocking ridge is pushing the Jetstream so far north that the resultant "polar vortex" low pressure surge is producing a kind of "Aerosol amplification" effect streaming in this offshore region?

I believe that the figure shows heat gained from 0 to 2000 meters in the indicated ocean regions between 2006 and 2013, and not insolation.  So local "Aerosol amplification" probably has little to do with the plot.
“It is not the strongest or the most intelligent who will survive but those who can best manage change.”
― Leon C. Megginson

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #93 on: February 07, 2015, 11:28:54 PM »
ASLR

of course you are correct, I was making what I thought was a logical intuitive analysis.  Being that GHG driven warming is fairly equitable and that regional ocean heat increases would be due to either current changes or insolation effects from pervasive wind patterns, such as the RRR off the California Coast feeding into and feeding off of a region of unusually low upwelling and surface temperature anomaly.

This is an interesting analysis of the Roemmich paper.

http://dotearth.blogs.nytimes.com/2015/02/02/a-fresh-look-at-the-watery-side-of-earths-climate-shows-unabated-planetary-warming/

Quote
In an email exchange, Kevin Trenberth of the National Center for Atmospheric Research said he was concerned that the analysis, limited to data from the relatively sparse array of Argo devices, was missing large areas of the seas that other studies, including his own, have identified as significant. As a result, he said, “their estimates look low-balled”. Here’s more from Trenberth:


It is disappointing that they do not use our stuff (based on ocean reanalysis with a comprehensive model that inputs everything from SST, sea level, XBTs and Argo plus surface fluxes and winds) or that from Karina von Schuckmann. [Trenbert pointed me to two studies, here and here.]

From Karina (2014 Ocean Sciences p 547) : “Our findings show that the area around the Tropical Asian Archipelago (TAA) is important to closing the global sea level budget on interannual to decadal timescales, pointing out that the steric estimate from Argo is biased low, as the current mapping methods are insufficient to recover the steric signal in the TAA region.” [Here’s the von Schuckmann paper.]

Indeed, their paltry .4 to .6 average Watt per meter squared ocean insolation is significantly below the range of other work and more recent OHC data from Argo, as their more recent analysis holds larger coverage.
Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

AbruptSLR

  • Multi-year ice
  • Posts: 17535
    • View Profile
  • Liked: 741
  • Likes Given: 236
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #94 on: February 08, 2015, 12:32:55 AM »
jai,

Thanks for the links.  It seems that many/most people do not understand how dynamic the ocean, and the atmosphere-ocean, responses can be when subjected to climate change forcing.  While denialists use uncertainty to justify BAU measures; it seems obvious to me that given that positive feedback mechanisms dominate negative feedback responses, that uncertainty (particularly about ocean heat content) is justification for taking immediate action.

Best,
ASLR
“It is not the strongest or the most intelligent who will survive but those who can best manage change.”
― Leon C. Megginson

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #95 on: February 09, 2015, 09:26:01 AM »
ASLR,

The precautionary principle is invoked under a scenario where potentially catastrophic and irreversible results MAY be the result of a current scenario that is not well understood.

This principle would have been logically engaged in, say the late 1980s or even the early 1990s.

Now that we are already experiencing the deleterious effects of anthropogenic climate change, knowing that we have already locked in as much warming as has been experienced since the pre-industrial, and this without albedo feedback effects, we MUST now engage in what comes AFTER the precautionary principle.

Namely, a response to a clear and present threat that is beginning to effect all levels of human endeavor, from the national level all the way down to the individual.
Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

AbruptSLR

  • Multi-year ice
  • Posts: 17535
    • View Profile
  • Liked: 741
  • Likes Given: 236
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #96 on: February 10, 2015, 12:09:26 AM »
jai,

The American Anthropological Association agrees with you and they say so in the linked statement issued in January 2015:

http://www.aaanet.org/cmtes/commissions/CCTF/upload/AAA-Statement-on-Humanity-and-Climate-Change.pdf

Also, you might enjoy reading the linked article about the dilemma of pointing-out long-tail risks vs Alarmism:

http://climatechangenationalforum.org/tail-risk-vs-alarmism/
“It is not the strongest or the most intelligent who will survive but those who can best manage change.”
― Leon C. Megginson

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #97 on: February 10, 2015, 06:19:57 AM »
ASLR,

Thanks! that second link was on point!
Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today

AbruptSLR

  • Multi-year ice
  • Posts: 17535
    • View Profile
  • Liked: 741
  • Likes Given: 236
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #98 on: February 12, 2015, 01:31:26 AM »
jai,

Given your interest in the influence of aerosols, I thought that you might like the following linked reference:

Harriet E. Ridley et al, (2015), "Aerosol forcing of the position of the intertropical convergence zone since AD 1550", Nature Geoscience, doi:10.1038/ngeo2353

http://www.nature.com/ngeo/journal/vaop/ncurrent/full/ngeo2353.html

Abstract: "The position of the intertropical convergence zone is an important control on the distribution of low-latitude precipitation. Its position is largely controlled by hemisphere temperature contrasts. The release of aerosols by human activities may have resulted in a southward shift of the intertropical convergence zone since the early 1900s by muting the warming of the Northern Hemisphere relative to the Southern Hemisphere over this interval, but this proposed shift remains equivocal. Here we reconstruct monthly rainfall over Belize for the past 456 years from variations in the carbon isotope composition of a well-dated, monthly resolved speleothem. We identify an unprecedented drying trend since AD 1850 that indicates a southward displacement of the intertropical convergence zone. This drying coincides with increasing aerosol emissions in the Northern Hemisphere and also marks a breakdown in the relationship between Northern Hemisphere temperatures and the position of the intertropical convergence zone observed earlier in the record. We also identify nine short-lived drying events since AD 1550 each following a large volcanic eruption in the Northern Hemisphere. We conclude that anthropogenic aerosol emissions have led to a reduction of rainfall in the northern tropics during the twentieth century, and suggest that geographic changes in aerosol emissions should be considered when assessing potential future rainfall shifts in the tropics."
“It is not the strongest or the most intelligent who will survive but those who can best manage change.”
― Leon C. Megginson

jai mitchell

  • Nilas ice
  • Posts: 2042
    • View Profile
  • Liked: 82
  • Likes Given: 20
Re: Quantifying underestimates of long-term upper-ocean warming
« Reply #99 on: February 16, 2015, 12:52:32 AM »
ASLR,

THANK YOU!!!  I really appreciate all of the wonderful references you put on this website.  It is really invaluable stuff.

I have considered Chinese aerosol reductions to be the impetus of the climate shifts we are seeing in the northern hemisphere since 2011.  However, the effect on the ITCZ and the AMOC with regard to northern hemisphere vs. southern hemisphere aerosol loading is terribly understudied.  Also the effect of point source aerosol loading in east pacific vs. west pacific causing a shift in the pacific surface winds.   This effect on the PDO is unknown, however, I am 100% certain that, if all anthropogenic aerosol emissions were stopped today, within 2 weeks we would experience a .5C jump in surface temperatures and there would be complete agricultural collapse as precipitation belts and heat-wave zones shift.

We are going there only gradually, but when we do, and we will, it will not be pretty.
Haiku of Past Futures
My "burning embers"
are not tri-color bar graphs
+3C today