The cloud aerosol effects seems to have huge uncertainty ranges (also aerosol effects in general). Is there any hope of narrowing these down in current ongoing research?
[Response: Some. But these are difficult and one of the key uncertainties is that we don’t have a good global source of information for what kinds of aerosol are in the air. This would have been one of the key datasets from the unsuccessful GLORY mission. – gavin]
The new AR5 table would be more readable if the second column had shown not just “resulting A drivers” but “net change (+ or -) in A drivers”. For example:
NOX emissions – +Nitrate+, -CH4, +O3tro
Then I would understand that CH4 negative radiative forcing in that row is actually due to -CH4 change.
As the table stands, I just glance at that row I scratch my head: methane has cooling effect, what’s going on? I must go back and figure out how aircraft NOX emissions react with atmosphere; then I finally realise it results in net decrease of CH4, that’s why negative RF.
Other wordings in the table could be changed accordingly, e.g. “Albedo Change due to Land Use” could just be “Albedo Increase due to Land Use”.
That simple change would greatly increase the table readability.
There are some pretty strong feelings about the validity of the AR5 reduction in aerosol forcing.
Is it possible that people are overlooking the larger role BC has (in terms of reducing the net aerosol cooling), or is it your sense that it may indeed be the case that the AR5 has taken too narrow (versus e.g. doi:10.1029/2012GL051870) a view of the plausible range of aerosol dimming?
Someone may want to check the math. Unless there are substanial rounding errors (highly unlikely), the total is a little off.
Not just the aerosol forcing, but the high level of confidence seems a bit awry. What happened to the albedo change due to glacial melt? Is that including in the land change?
[Response: Errors are not symmetric and so the net effect is not a simple addition. This was also the case in AR4. I have no idea why you think albedo change due to glacier melt is an important factor over the last century – for that to be true, you’d need to have a really large fraction of current glacier area disappear. – gavin]
Thanks, it is always good to see concise histories of progress in science via graphical representations, and the latest one uses colors better and really offers much information in one chart. However, Joe’s chart was simpler… :-)
The non-peer reviewed references cited on denier websites come from the group working on both actual and potential effects and comprehensive mitigation strategies, Working Group II. It’s hardly surprising that they include some non-peer reviewed references among the peer reviewed ones. This is a topic which blends science, economics, and politics.
For those still on the global warming denial train, let’s see your list of non peer reviewed references for Working Group I, Chapter 2, the part of the report that says global warming is real, and mostly caused by us. I’ll quite concede that some of the potential effects and mitigation strategies are speculative, will others concede those two points?
To give people a more objective look at what Working Group I, Chapter 2 actually considered, here are the first five references, unselected for rhetorical purposes. And the website that lists all of the hundreds of references, overwhelmingly (entirely?) peer reviewed. I invite people click the link, and get an idea of the MOUNTAIN of peer reviewed data that’s behind the statement of the scientific community that global warming IS real, and mostly caused by us.
Abdul-Razzak, H., and S.J. Ghan, 2002: A parametrization of aerosol activation: 3. Sectional representation. J. Geophys. Res., 107(D3), 4026, doi:10.1029/2001JD000483.
Abel, S.J., E.J. Highwood, J.M. Haywood, and M.A. Stringer, 2005: The direct radiative effect of biomass burning aerosol over southern Africa. Atmos. Chem. Phys. Discuss., 5, 1165–1211.
Abel, S.J., et al., 2003: Evolution of biomass burning aerosol properties from an agricultural fire in southern Africa. Geophys. Res. Lett., 30(15), 1783, doi:10.1029/2003GL017342.
Ackerman, A.S, M.P. Kirkpatrick, D.E. Stevens, and O.B. Toon, 2004: The impact of humidity above stratiform clouds on indirect aerosol climate forcing. Nature, 432, 1014–1017.
Ackerman, A.S., et al., 2000a: Reduction of tropical cloudiness by soot. Science, 288, 1042–1047
Thanks for bringing out this shift in the way that radiative forcing by gas is being treated in the WG1-AR5. But I think that it should be noted that this is also starting to make the analysis quite a bit more model dependent because for some gases it folds in a lot of atmospheric chemistry. In particular, bringing in the indirect feedback effects on OH caused by CH4, CO, etc goes into an area where models still differ and the extent to which they can reproduce observations is pretty debatable.
For example, when CH4 suddenly started going up in 2006, particularly across the southern hemisphere where it had been remarkably stable for six years, the first explanation was that this was due to a reduction in OH. But why had CH4 also suddenly stopped increasing in 2000? Was that an increase in OH? Multi-decadal increases in CH4, HCFCs and HFCs have occurred along with a downward trend in CO across most of the clean air sites in the southern hemisphere, which suggests that the reduced species are not decreasing OH, as is being assumed in the indirect effect for the CH4 global warming potential.
At the AGU meeting in San Francisco in December there will be a Union session on atmospheric methane and some of us are putting together a summary of why we think changes in the chemistry are still not well understood. We may get to include something on new evidence from 14CO data which now seems to show that past fluctuations in CH4 growth rate were related to changes in the spatial distribution of OH rather than in its global average, and while some had anticipated that this might happen about ten years ago, it is not yet coming out of the chemistry models.
So I think that this radiative forcing figure in the AR5 has a subtle form of model based attribution folded in, and for which the level of confidence is definitely lower than it is for the direct effects of each gas. Also while it’s a great figure for scientists to consider, I don’t think that many policymakers will understand this level of detail.
[Response: Hi Martin, Thanks for your comments. It is inevitable that moving closer to the source of any change requires more modelling, and so the uncertainties will have to increasingly include the multi-model spread – as constrained by observations as possible. However, instead of a reason not to do it, I see that as the price that needs to be paid in becoming more relevant. Moving from RF to ERF also has a model component in it. But even in previous incarnations the CH4 effect on OH was implicitly included in the tropospheric ozone and sulphate column. The switch to an emission-based presentation is not a change in physics, just a change in accounting. As to whether policy-makers appreciate this more or less than previous presentations, we would have to ask them. My interactions suggest that they are actually hungry for even more policy-specific, emission-based analyses (a la UNEP report on black carbon and tropospheric ozone), but perhaps your experience is different. It might be worth making a distinction here between policy-analysts/policy-makers and politicians. The former, not the latter, are the target here. – gavin]
[Response: AR4 adopted a hybrid scheme in its discussion of present-day radiative forcing and its discussion of feedbacks on emissions. In the radiative forcing bar chart, things like ozone and stratospheric water were given their own separate RF category wherever they came from. This makes sense if you’re trying to describe the present situation, since you can actually measure the constituents, and simply stating their effects without saying what causes the ozone or stratospheric WV concentration to be what it is avoids some model uncertainty. However, in the discussion of emissions effects (notably global warming potential) the feedbacks are included in the effective radiative efficiency. That also makes sense, since you need to relate RF to emissions in order to predict the future. (I’d ignore GWP, but the radiative efficiency is still a relevant quantity). There is an argument for switching to a more emissions-based approach in AR5, but as Gavin and Martin note, this incurs model uncertainty. I think AR5 made a bit of a mishmash of the transition to emissions-based accounting, though, because they included aggregate effects of methane, but they did not present the black carbon results in a way that shows the aggregate effect of all emissions from BC rich sources, which (according to Bond et al’s JGR assessment) is likely to be a small net cooling when co-emissins are taken into account. The intent of the new AR5 RF diagram is to make abatement priorities easier to read off the graph, but it doesn’t really achieve that. For that matter, it’s not possible to infer abatement priorities without taking into account lifetime; GWP was meant to do that, but as we now know it’s a hopelessly broken concept. –raypierre]
Yes, that’s an important point to emphasize. Allocation of RF to the emissions rather than to the atmospheric concentrations of individual substances brings the relationship between emissions and concentrations into the calculation, and this relationship is based partly on models (e.g. of the carbon cycle, of OH chemistry, etc.) with inherent uncertainties. This is presumably why the uncertainty range of the CO2 forcing is wider in AR5 than in the AR4 figure.
Also important to emphasize that the total anthropogenic RF has not become more uncertain because of this change. It is only the allocation to individual emitted substances that is affected, not the overall total.
Given that emissions-based breakdown is needed for policy purposes, it is far better that IPCC WGI do this and include an assessment of the additional uncertainties, than leave it to others to do after the fact.
For AR4, in addition to Fig. 2.21 that Gavin pointed to, it’s worth looking at footnotes to AR4 Table 2.13 to see the numbers and uncertainties in AR4, compared with the new AR5 estimates/uncertainties.
Black carbon is mainly a factor in the direct effect which hasn’t changed much in the final reckoning: -0.5W/m2 for DirectRF in AR4 and -0.45 for DirectERF/ERFari in AR5. Estimated Black carbon positive forcing is larger but that’s also the case for some reflective aerosol species. The real difference from AR4 comes from aerosol-cloud interactions.
It looks to me that their total aerosol forcing best estimate has essentially been determined by their assessment of the six satellite obs. constrained studies listed in Table 7.4. It seems to me the treatment of the six papers is somewhat questionable, with choices and interpretations tending to push the median towards less negative values. Also questionable is the imposition of a +0.2W/m2 longwave forcing on the four papers which only estimated shortwave changes. No reference is provided to support this figure. They talk about it as the low-end of a range from models but don’t cite which models. Seems a little ad-hoc and puntish.
Ultimately though, a total aerosol forcing of -0.9W/m2 is probably a plausible expert judgement (not at all to say it’s the only plausible expert judgement) estimate given the full balance of evidence accrued post-AR4. However, this is far from the final word and I wouldn’t be at all suprised to see future assessments go back down to -1.2 or even more negative.
I concur with your comment on the imposed longwave forcing “correction”. While I would have preferred a percentage-based reduction (as function of the actual forcing in each respective study/model), I consider the missing reference for their statement (“[…] modeled longwave effects which varied from +0.2 to +0.6 W/m2 in the assessed models”) a bit unfortunate. Any idea which paper they are referring to? For the same reason (missing reference) I had some trouble to make sense of the -0.85 W/m2 for Bellouin et al. 2013. Although I now think I know what they’ve done, I hope Nicolas can help (I’m waiting for his answer). Otherwise, Drew Shindell might be the right person to ask anyway.
As you said, ultimately, total aerosol ERF of -0.9 W/m2 is certainly more than plausible, so my bet wouldn’t have been much different (probably -1 W/m2). However, it’s important to note that this is only true for the current forcing, while I tend to think that it might have been (globally) higher when sulphate emissions peaked (1970s). Little discussion on the temporal evolution of the aerosol forcing in AR5 either.
My guess was that they were referring to LW estimates given in the papers listed in Table 7.4 since they were mentioned earlier in the section and included mixed-phase/ice clouds and convective schemes. However, when looking at the papers I found a range of -0.3 to +0.6 and average about zero.
Also, Wang et al. 2011 discussed the source of their +0.26W/m2 LW estimate and noted that it was mostly found in clear sky areas. It was basically a response to aerosol-induced land-surface cooling in the fixed-SST simulations used to diagnose model forcing.
To me that seems like a confusion of forcing and response.
while there is, indeed, a range of values extractable from the reference list in Table 7.4, it’s not even close to their suggested range, as you’ve pointed out. As far as I can see, the only model exercise which reports on aerosol LW effects (!) is Quaas et al. 2009. So not even a forcing estimate. In any case, “scaling” to some degree can be seen there, i.e. higher neg. SW forcing is on average counterbalanced by higher pos. LW effect. Regardless of the fact that the effective LW forcing might be stronger, I’m still having trouble to follow their argument with respect to LW.
Is there any hope that this plot will ever revert back to units of delta T as opposed to watts per square meter? I frequently refer to this chart in one form or another when I answer questions from readers about how much such and such a forcing has contributed to climate change, or “have scientists thought about forcing X as an explanation for climate change?” More often than not, there is at least one more round of follow-up questions in which they want to know “but what does that mean in terms of temperature?” It would be valuable for communicators to have this chart reproduced using units of temperature change.
You can multiply the numbers in the charts by a factor of 0.3 to get the no-feedback temperature change in degrees Celsius. Multiply by 0.75 or so to get a ballpark “real temperature change” once you get to equilibrium (though with a bit more uncertainty).
When 99% of your audience won’t understand something where a different wording would mean 100% of your audience would understand, then use the 100% language, EVEN IF IT’S NOT THE MOST CONCISE. In this particular case, where (I think) there’s a linear relationship between w/m2 and delta T within the range being considered, simply putting two Y axis labels would have satisfied everyone.
Comparing the last two graphs 2007 vs 2013 I see that for CO2 both the RF values and the error bars are more streched in both directions. This indicates less confidence.
Strangely, in the confidence column the confidence level has gone up from H to VH (very high). This does not make sense.
I’ve realised I didn’t address your question in the straightforward way it deserved, because it is quite simple: Global satellite observation-constrained estimates of the cloud albedo/first indirect effect have consistently indicated a forcing of ~ -0.4W/m2 compared to the -0.7W/m2 given in AR4.
The AR4 figure was derived from a weighting of pure model results (typically around -1.0W/m2) against a couple of observationally-constrained results. Since AR4 a few more independent estimates have been published still consistent with about -0.4, plus the passage of time has not thrown up clear contradictions to these results. Hence, this time around the observationally-constrained studies have been trusted almost to the exclusion of model results, meaning a smaller (less negative) indirect effect.
There have been a few recent papers which suggest some, if not all, of these estimates may be significantly biased (notably Penner et al. 2011 and McComiskey and Feingold 2012). AR5 mentions these but suggests the sign and extent of any bias is yet to be demonstrated.
I am always struck by the large error bars on the radiative forcing due to CO2. What is that all about?
In the AR5 figure the range is given as 1.68 +/- 0.35 (or +/- 20% of the central estimate).
In AR4 this range was 1.66 +/- 0.17 (or +/- 10% of the central estimate).
Why the large increase in uncertainty?
In AR3 it is hard to tell, because CO2 is bundled with the other well-mixed GHGs, but in the concurrent Hansen paper it is given as 1.4 +/- 0.2 (+/- 14%), which is still lower than the AR5 uncertainty range.
On the one hand this isn’t massively important – even at the lower end of the bound CO2 is still a key GHG, etc, but I would have thought that the radiative physics of CO2 were the simple bit, compared to the aerosol effects, the cloud feedback, the response of ice sheets to warming, etc, and a 20% uncertainty range is really huge, which is not helpful for making specific climate predictions.
[Response: Actually this is quite interesting. The main shift is because of the change from stratospherically-adjusted radiative forcing to ‘effective’ radiative forcing. This shift allows more very fast adjustments into the forcing numbers and (inevitably) leads to slightly more structural uncertainties. This was done for ease of calculation across a whole range of forcings and the larger error bars on the CO2 is a acceptable price. Note that the clear sky forcings for CO2 and most other GHGs are very well known, but once you are looking at the global all-sky conditions, there is some inevitable impact from the water vapour/cloud/temperature distribution and properites and that is where the bulk of the uncertainty comes in. – gavin]
Comment by Timothy (likes zebras) — 9 Oct 2013 @ 11:34 AM
Is it possible to compare the older charts to the newer ones to determine how accurate the old charts were compared to what is known now? Have the additional data led to forcing estimates becoming larger, smaller or about the same? Obviously the CO2 forcing has increased, but are current estimates of the forcing in 1990 similar to estimates of the forcing that were made in 1990? This could give another way to evaluate the accuracy of the older reports.
The solar forcing has been downgraded ,
I’m puzzled why such a small forcing has in the past be a climate driver
[Response: At different time periods it could have been larger – this is simply an assessment for the long term trend over the period 1750-2010. – gavin]
from what I understand the way the forcing is calculated
is the total Solar irradiance in W/m2 divided by the apparent disk area ,
should the troposphere layer thickness too be included ,
after all it refract the sun rays quite a lot ,
Uncertainty in aerosol effects (direct and indirect) seems to dominate the overall uncertainty in the AR5 table of forcings. Is there any chance of a replacement satellite for the failed Glory mission?
Comment by Steve Fitzpatrick — 9 Oct 2013 @ 10:08 PM
So far nobody mentioned this so I guess I will have to embarrass myself: why is there no forcing from volcanoes on some charts while the other chars have it? Even if it’s zero it should be present with error bars, I reckon.
[Response: The graphs are now radiative forcings since 1750, and over that period there is no trend to speak of in volcanic forcings. Over shorter periods, there can be important contributions though and a graph focussed on that would include them. – gavin]
It seems the bottom line is without sulphate aerosol cooling we would be on course for 2c warming already. The question is what happens when the emerging industrial counties get serious about air pollution control?
Thanks for putting this together. This is the figure I use to teach non-science policy-makers-to-be about climate change accounting, how the accounting has changed and not changed over the years, and esp. the influence of air quality controls on what the future will bring. Will China and India scrub their coal? How soon?
This concise history of how the science has changed will be invaluable in these efforts.
I’m in discussion with a couple denialists swearing the chart is all a big fraud, because the TSI difference since 1750 is more like +0.5 W/m^2, while the chart shows a solar irradiance factor of +0.05 W/m^2.
Now, I understand that the solar irradiance factor on the chart is not the same thing as the straight TSI difference. For one, there’s the geometry factor knocking it down by a factor of 4. But I’d like to be able to explain the rest. Quick and dirty, what else goes into calculating that factor?
The forcing at the top of the atmosphere (TOA) prior to stratospheric adjustment is equal to 1/4 of the TSI change, multiplied by the fraction absorbed by the Earth (about 0.7), so it’s about 0.175 * TSI change. For TSI change of 0.5 W/m2, that would be ~ .08 W/m2. There have been different estimates for TSI change over the years; it’s also possible some ‘prefer’ the higher estimates.
The forcing at the tropopause level prior to stratospheric adjustment would subtract the fraction absorbed by ozone (and anything higher up, but that’s tiny) – I think that’s a few percent. much of that gets added back after stratospheric adjustment because the stratosphere warms so that it radiates more, by as much as it’s been heated, and some fraction of that is downward.
An earlier IPCC radiative forcing chart was for tropopause-level forcing after stratospheric adjustment; I’m unclear on whether this one is done the same way.
re 39 mamoo – I have read that solar UV is more variable than solar TSI, so presumably fraction of solar forcing (a change) that is absorbed in the stratosphere would be higher than the fraction of total solar heating of the Earth that is absorbed there.
“TSI is very difficult to measure from Earth’s Surface. We have a correlation between TSI and Sun Spot # “
I was able to pick out the TSI contribution from a multivariate linear regression analysis of GISS with respect to 4 noise terms — SOI, Aerosols(volc), LOD, and TSI. This is an interactive view: http://entroplet.com/context_salt_model/navigate
The TSI temperature response should be about 0.05 C for the dP ~ 1 watt/m^2 coming out of the sun-spot fluctuations at P=1360 watts/m^2, and sure enough that is what the linear regression model picks up.
This is simple differential calculus
P ~ T^4
dT = 1/4*T/P dP = 1/4 * 288/1360 dP = 0.053 dP
… re evaluation of dT/d(OLR or forcing): (of course, if GHGs were shifted toward one end of the OLR band or the other, the zero-feedback sensivity could be farther reduced or enhanced; At peak wavelength, there is a T^5 proportionality. @220, 255, 288 K, peak wavelength is 13.2, 11.4, 10.1 microns; the last two are in the Atmospheric window, so… whatever the line-by-line results are.)
Gavin – “Note that the clear sky forcings for CO2 and most other GHGs are very well known, but once you are looking at the global all-sky conditions, there is some inevitable impact from the water vapour/cloud/temperature distribution and properites and that is where the bulk of the uncertainty comes in.”
So much of the uncertainty in the carbon dioxide forcing is for the same reason that there is so much uncertainty in the climate sensitivity. Damn clouds, eh?
Thanks for the answer!
Comment by Timothy (likes zebras) — 18 Oct 2013 @ 11:16 AM
Re 49 WebHubTelescope – okay; when I looked at your equation I was puzzled so I had to go through it myself.
and another correction (bold in following) (I’m just restating the end of my 46, with corrections):
dP = 3*OLR/T * dT; for OLR ~= 240 W/m2, that’s
2.5, 2.8, 3.3 W/m2 per K , or 0.40, 0.35, 0.31 K per W/m2, @220, 255, 288 K (zero non-Planck feedback equilibrium climate sensitivity, setting aside the matter of the stratosphere, etc.).
1 W/m2 of TSI -> ~ 0.175 W solar forcing TOA -> 0.070, 0.062, 0.053 K @three Temps (see above), not taking into account stratospheric portion, feedbacks, or time for equilibration.
(where dP is either the change in OLR required to restore equilibrium or it is the solar forcing TOA ; they would be equal, of course, when the only forcing is solar; P is OLR, but can be expressed in terms of TSI(hence the OLR,TSI*0.175 parts) (I’m not sure if I originally just forgot the 0.175 factor in the above equation or if I didn’t included it because the cancelation.))
then, manipulating the last equation:
dP/dT = 3*P/T ~= 3*(5.67E-8 W/(m2*K4) * T^4) / T = d/dT (5.67E-8 W/(m2*K4) * T^4)
(as a double check that dP/dT = 3*P/T , and of the prior equation)
dT = (1/3)*T/P * dP
= (1/3)*T/OLR * d(OLR) in general
= (1/3)*T/(TSI*0.175) * d(TSI*0.175)
= 1/3 * T/TSI * d(TSI) for solar forcing.
It is time to really think about what the graphs mean.
Circa 1750, we started burning the geologic carbon, increasing the amount of greenhouse gases in the atmosphere and warming the surface of the Earth. We also released trivial amounts of methane. These small methane releases have preoccupied modelers that should have been thinking about the big picture.
The big picture is simple. The concentration of CH4 in the atmosphere is the partial pressure of CH4 in equilibrium with a large number of reservoirs of solid methane clathrate. Each clathrate reservoir has its local equilibrium dependent on local temperature, total pressure, and the CH4 concentration in the reservoir’s local environment.
Thus, as long as there are clathrates, we can expect them to control the partial pressure of methane in the atmosphere. The legal and diplomatic fiction of a low global warming potential for methane is based on an incorrect definition and bounding of the methane system that fails to recognize the equilibrium relationship between small flows (megatonnes) of methane in the atmosphere/ biosphere and large (gigatonnes) reservoirs of methane clathrates.
Circa 1750, conditions had been stable for a relatively long period, and all clathrate reservoirs were in equilibrium with atmospheric concentrations of CH4 on the close order of 0.7 ppmv to 0.4 ppmv. As the Earth started to warm, the equilibrium point for some clathrate reservoirs moved from solid to gas, and they released methane into their environments. As the methane equlibrium moves toward gas, we are seeing atmospheric CH4 concentrations approaching 2 ppmv. This produces a radiative forcing of almost 200 ppm of CO2 ,and thus our effective green house gas concentration is now on the close order of 600 ppm CO2e.
Since clathrate/ methane gas is an equilibrium system, as methane is removed from the oceans & atmosphere by oxidation, and the partial partial pressure of methane declines, then methane clathrate in the reservoirs will dissociate to restore the equilibrium partial pressure of methane for that temperature. As the Earth continues to warm, the equilibrium will continue to move from solid clathrate to methane gas resulting in higher atmospheric partial pressures of methane. I am as sure of this as I am that the atmosphere will hold more water vapor as it warms. It is the same physics.
I have oversimplified the methane cycle. Still, the subject chart does not tell the big picture.
However, in parts of the Arctic, we are now separated from sea floor clathrates by only a hundred meters of ocean that is becoming increasingly storm mixed. It is time to see the oceans as circulating fluid, that helps bring the system to equilibrium.
Assuming drilling and military seabed activity have perforated the permafrost, and warm water circulates down in the sediments laterally between faults in a way not seen during past glacial cycles.
Assuming new clathrate doesn’t form to plug up the holes within the temperature/pressure stability zone, or doesn’t form fast enough to limit the process.
Assuming the seabed surprises us — could it be unstable like the ice caps, and for the same reason — heat moving by circulation not diffusion?
I feel lucky. I doubt it’s happening. But hey, the Russian and ex-USSR records from their petroleum operations and Arctic navy ought to give some idea. Betcha their Navy (and likely ours and several others) have hydrophone sound files going back decades that would be searchable for sounds indicating outbursts of bubbles, for example. Is anyone looking?
*UFOD: Some guy on the Internet
Question re: Relationship between forcing, CO2 concentrations and Temperatue.
Where’s the error:
CO2 forcing abatement = 5.35 ln (Concentration 1 / Concentration 2).
Keeping PPM levels to 409 versus 410 results in a reduction in forcing by 0.013 watts per square meter.
Multiply by sensitivity (of temps to forcing), and you get 0.44×0.013 = 0.057 Degrees C.
1 ppm CO2 (by volume), or 1.6 ppm by mass corresponds to 1.6×10^-6 x mass of atmosphere, or 1.6×10^-6 x 5×10^18 = 8×10^12 kg or 8 billion tons.
Thus to reduce the temperature increase by one fifteenth of a degree you have to reduce CO2 levels by 8 billion tons, or one ton per person on the Earth. Because of the logarithmic relationship, the next fifteenth of a degree will require an even greater reduction.
For the Earth, with OLR ~=240 W/m2 (I think I saw 235 W/m2 once; if 240 W/m2 is 5 W/m2 too great, dT values will be ~ 2.1 % too small)
@ 288 K (global average surface T), 255 K (effective broadband emission T), 220 K (at or near tropopause, roughly)
dT per W/m2 of OLR change (in response to SW (solar) forcing) or recovery (in response to LW forcing)
0.300 K, 0.266 K, 0.229 K
per 1 W/m2 change in TSI (~0.175 W/m2 solar forcing TOA)(one can substitute change in TSI in place of solar forcing TOA if TSI replaces OLR in the equation; the 0.175 factor drops out)
0.0525 K, 0.0465 K, 0.0401 K
per doubling of CO2 (based on 3.7 W/m2 – which is actually the tropopause level forcing after stratospheric adjustment)
1.11 K, 0.983 K, 0.848 K
Notice the sensitivity …
(this is equilibrium climate sensitivity with no non-Planck feedbacks, based on T^4 proportionality (which doesn’t generally apply to specific absorption/emission bands, and may be a little off due to horizontal and temporal T variations (I don’t think that’s a big factor, though… not entirely sure; I estimated once that the variations at the surface over a year may cause the surface to emit as if it were 1 K warmer than it is; higher highest cloud tops and greater low-level H2O tend to be in warmer places (making them look cooler from above) and upper troposphere has smaller variations but …))
… decreases with decreasing T. It would be the other way around, if the OLR were proportional to T^4, but I am using the same OLR. I used 3 temps in an attempt to find a range of possible sensitivities, since at some parts of the spectrum much OLR comes from near the surface (absent higher clouds, etc.) while at other parts it comes more from the upper troposphere or higher. However, there is also the spectral variation in n for T^n (for blackbody P(T), peak over wavelength n=5, peak over frequency n=3, peak over (log or ln)(wavelength or frequency) n=4, longest wavelengths n approaches 1, shorter wavelengths n gets larger).
Also, this excludes the effect that a portion of the forcing is realized in the stratosphere … my first approximate guess is half, or a bit less than half, of that adds to the tropopause level forcing when the stratosphere adjusts (via warming or cooling and changes in downward LW flux as a result) (the fraction depends on where in the stratosphere the warming or cooling occurs, of course, but if enough of the downward LW flux is from bands with small stratospheric optical depth, the fraction is less sensitive to that distribution).
So I’d go with a more refined model. The no non-Planck feedback response for CO2 doubling is supposed to be a bit over 1 K, and I think my range of values above didn’t quite reach as high. The OLR response to the imposed decrease (after stratospheric adjustment) would tend to be excluded from the center of the CO2 band …
(and also, I’d expect, possibly some of the highest cloud tops, via the FAT hypothesis (I don’t have time just now to reread the link completely (I may have left off before finishing, now that I think about it) but I think it’s specifically about tropical cirrus)- but wait, this is before considering such feedbacks)
… But it’s not as if the OLR increase in response to the imposed decrease is coming entirely from the surface in the tropics.
… also exluded taking into account portion of OLR from stratosphere (solar heating there would tend to equal a net contribution to OLR from there, so the net upward flux at the tropopause would be smaller, which would increase the calculated dT values.
… but I think that’s wandering outside of the territory where such a simple relationship would hold anyway (the total LW flux upward at the tropopause will be larger than the net upward LW flux, and in the spirit of the relationship derived, the gross flux may be more appropriate). But obviously it was intended as a rough approximation.
… but the last 2 comments (of mine) are at least in some contexts moot as forcing at tropopause level and at TOA are the same after stratospheric adjustment; so OLR does have to increase (after strat.sph. adjustment) by the amount of forcing at tropopause level (excluding any SW feedbacks); change in gross upward flux at tropopause level will probably exceed that.
Aaron Lewis (#53) wrote:
“Circa 1750, we started burning the geologic carbon …”
Not in significant amounts, especially compared to the burning of wood.
Ice cores and methane reconstructions seem to show something exceptional happened in the 18th century. And it seems tracable to the Northern hemisphere. But it’s not clear it started around 1750 and in any case can not possibly be explained by the burning of minute amounts of geologic carbon.
Furthermore, your theory doesn’t seem to explain either the mid-16c. methane spike or the slow rise in atmospheric methane concentrations between high antiquity and 1000 years ago. And certain human activities which do not necessarily involve fossil fuels are known to cause methane emissions…
Comment by Anonymous Coward — 20 Oct 2013 @ 10:49 PM
Re 53 Aaron Lewis wrote that an increase of 2ppmv pf methane = 200ppmv of co2. Surely methane has a forcing of approx. 20 times the equivalent amount of co2. So 2ppmv of methane = 40ppmv of co2, not 200.
“Multiply by sensitivity (of temps to forcing), and you get 0.44×0.013 = 0.057 Degrees C.”
If climate sensitivity is ~ 3 K/doubling of CO2 ~ 3/3.7 K/(W/m2) = 0.81 K/(W/m2); using that value I get 0.0106 K (using unrounded 3/3.7); using 0.7 K/(W/m2) instead I get 0.0091 K. 0.057 K seems a bit too high, even after slow-acting feedbacks. (But I just realized that 0.0057 K is the result of 0.44 K/(W/m2) * 0.013 W/m2, so you’re decimial point was just one point off.)
“Thus to reduce the temperature increase“… “you have to reduce CO2 levels by 8 billion tons, or one ton per person on the Earth. Because of the logarithmic relationship, the next fifteenth of a degree will require an even greater reduction.”
Actually a larger reduction is required because some is going into the ocean. On the flip side, you’ll be reducing ocean acidification as well.
But the next 0.01 K will require a smaller reduction if you’re going from 409 to 408 ppm. Or did you meant it will be even harder to go from 411 ppm to 410 ppm? That of course assumes a constant airborne fraction per unit emission, which isn’t really what’s going on; some part of the rate of CO2 uptake out of the atmosphere should be an ongoing response to prior emissions, but also there’s potential for a response to climate change itself.
… ie absent climate feedback on C-cycles, a slower rate of emission would tend to reduce the peak atmospheric CO2 because the removal by the ocean, etc, would keep up a little bit more – of course, that only goes so far in the short term, since a return to the initial atmospheric CO2 level requires some very slow-acting processes. (I’m under the impression that to some extent a ‘climate fate’ is insensitive to the exact rate of release of the same given total CO2 (maybe because of the slow feedbacks?); I assume that’s only true for a limited range (relative to Earth’s time) – I mean, if we took 1 Trillion tons and spaced it out over the next 4 billion years, rather than a century or so, I’d have to think it wouldn’t be to much of a problem…)