Climate Science Glossary

Term Lookup

Enter a term in the search box to find its definition.

Settings

Use the controls in the far right panel to increase or decrease the number of terms automatically displayed (or to completely turn that feature off).

Term Lookup

Term:

Settings

Beginner Intermediate Advanced No Definitions Definition Life:

All IPCC definitions taken from Climate Change 2007: The Physical Science Basis. Working Group I Contribution to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Annex I, Glossary, pp. 941-954. Cambridge University Press.

Recent Comments

Comments 34351 to 34400:

Russ R. at 05:37 AM on 17 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Rob Honeycutt @ 26:

You misquoted me by leaving out the quantity.

""But what we actually do know about ANT directly is far more uncertain. We only know GHG to a low level of certainty..."
Hm. Actually, I believe that is wrong...

What I actually wrote:

"We only know GHG to a low level of certainty (±0.4°C),"

Which is exactly the value in the IPCC's chart above, and is a low level of certainty relative to the ±0.2°C claimed for total ANT.

Still believe I'm wrong?

dhf at 05:27 AM on 17 September 2014

97 hours of consensus: caricatures and quotes from 97 scientists

@ Tom Curtis at 10:58 AM on 16 September, 2014

It is not my claim that calibrations can be performed by laymen. One point I tried to make was that properly presented test result can be easy to comprehend even if the model is complex and that properly presented data are strong arguments.

Rob Honeycutt at 04:43 AM on 17 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Russ R... What I'd suggest here is that, in order to work, the argument you're putting forth here would require a yet unknown mechanism that un-explains the high level of certainty that we have about greenhouse gases, and adds a yet-so-far completely unknown forcing to take its place.

Rob Honeycutt at 04:16 AM on 17 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Russ R. said @24... "But what we actually do know about ANT directly is far more uncertain. We only know GHG to a low level of certainty..."

Hm. Actually, I believe that is wrong. We have a high to very high level of scientific understand about the radiative forcing of anthropogenic greenhouse gases:

Russ R. at 03:37 AM on 17 September 2014

The 97% v the 3% – just how much global warming are humans causing?

KR,

Your comment relates to the concept that OBS - (NAT + INT) = ANT. (In other words, anything we can't explain must be anthropogenic.)

Yes, observed warming has very little uncertainty, and the warming ascribed to NAT and INT is claimed to be quite certain (±0.1°C), but is unfortunately limited to the extent of our knowledge of those processes (which is turning out to be more uncertain than previously thought). It's necessarily a case where we simply don't know how much we don't know.

But what we actually do know about ANT directly is far more uncertain. We only know GHG to a low level of certainty (±0.4°C), and know virtually nothing about OA except to say that it is equal to ANT - GHG.

But to say that we know ANT to a high level of certainty (±0.2ºC) is only possible if we assume that we know everything about NAT and INT with high certainty. Which is the same as assuming a high level of certainty of ANT.

And we can't simply assume that the uncertainty of the components (GHG & OA) is negatively correlated because the uncertainty of ANT is assumed to be low.

Circular reasoning works because circular work.

dhf at 03:27 AM on 17 September 2014

97 hours of consensus: caricatures and quotes from 97 scientists

@ Tom Curtis at 10:58 AM on 16 September, 2014

You are right, my premise were false, my assertion was wrong, it is not correct to say that: "Very many are trained through education and profession to recognize and disregard arguments containing logical fallacies."

Thank you for your response to my false assertion.

Paul Barry at 02:29 AM on 17 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Anyone got more information about the 3%?

I know a lot of careful work has gone into nailing down the 97% number and those who have been working on this are deserving of praise.

But that statistic on its own isn't enough.

3% doesn't seem that low to those who imagine that good science is only done by a minority of scientists - a few Gallileos!

Many will see this 3% as a crack in the edifice of Climate science: a chink of light in the darkness that is the future as foretold by those gloomy lefty boffin-heads!

3% isn't so tiny. Sure doesn't 1% of population of USA own 35% of all the wealth. Maybe in the same way 3% of climate scientists have 50% of the knowledge.

What is needed, I think, is an analysis showing that this minority is diminishing, becoming obsolete, that their work has been shown to be flawed, that their findings have been refuted and that many of them have been discredited by their continued use of already refuted arguments...etc.

What is needed is the extinction of that kind of stupid optimism that allows people to continue in vast numbers to entertain the notion that there is a chance - an outside chance - that maybe - just maybe - this climate change stuff is nonsense. Because that is what people do. For the reckless, the miserable and the desperate, 3% isn't such a bad gamble! It's a light in the darkness!

Extinguish it!

Anyone working on this?

KR at 02:13 AM on 17 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Russ - See my comment here [which you have not responded to]. The sum of forcings is strongly constrained by other data such as OHC, hence component forcings with larger uncertainties are by necessity negatively correlated. If one is underestimated, other(s) (under the constraints on totals) must be overestimated.

Russ R. at 01:54 AM on 17 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Kevin C.

Thank you, I get it... the variance of whole is less than the variance of the parts when the covariance is negative.

To accept an uncertainty of ANT of only ±0.2ºC, when the uncertainties of GHG and OA are ±0.4°C and ±0.35°C respectively, you must necessarily believe that the covariance is negative.

Got any actual evidence to make any type of valid estimate of the covariance of the error terms for GHG and OA? Or shall we just assume that it's negative?

Leto at 00:10 AM on 17 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Russ @18, why does your quoted section begin halfway through a sentence?

Kevin C at 00:08 AM on 17 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Russ:

var(x+y) = var(x) + var(y) + 2 cov(x,y)

var(x-y) = var(x) + var(y) - 2 cov(x,y)

Covariance can be positive or negative. So the variance of either the sum or the difference between two variables can be smaller than the variance of either.

Take the trivial case y=-x. Then x+y=0 by identify. Since cov(x,y) = cov(x,-x)=-var(x), var(x+y)=0, which is necessarily true.

Russ R. at 22:36 PM on 16 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Thank you Bob Loblaw and Tom Curtis,

"the combined error would be +/-0.53 C for the "likely" range. That would allow a 17.3% chance of 50% or less anthropogenic contribution"

This is exactly what I was getting at, and it's a whole lot different from the PDF shown above which states that "the probability of the human contribution being less than 50% is almost nil" or in Gavin Schmidt's words "p < 0.0001".

The point I'm making (and I don't actually care if you don't like the roundabout way I'm making it), is that the PDF shown above uses an uncertainty estimate (±0.2ºC) that is far too low for anthropogenic warming because it's not acutally derived from any calculation of the components anthropogenic warming (i.e. ANT = GHG + OA).

Instead, the uncertainties show that the derivation is actually OA = ANT-GHG, (meaning OA is a plug figure). And ANT is simply assumed to be the difference between OBS and (NAT + INT), where NAT and INT themselves are only estimated "to the best of our knowledge". (As an aside, I find it interesting that the amount of variability being assigned to NAT and INT appears to be rising in an apparent attempt to explain away the recent lack of warming).

To summarize, my argument here is that the uncertainty figure to use for anthro warming must be derived from what we can actually attribute to the components of anthro warming (GHG + OA), not what we assume must be anthro because we don't know how else to explain it. I think that's called "begging the question".

Moderator Response:

[JH] Please keep it civil.

michael sweet at 20:10 PM on 16 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Bob,

from 10; It is my understanding that the correlation is negative. The Anthro contribution is well measured. When it is split into the Greeenhouse and other contributions, there is a large uncertainty about aerosols. This shows up in both distributions. They must add to the Anthro contribution so if aerosols have a large negative contribution the greenhouse must be bigger. If aerosols have a small contribution the greenhouse must be smaller. The aerosol cannot be large negative and the greenhouse small positive at the same time or they would not add together properly. This makes the error smaller for net anthro.

jwalsh at 15:44 PM on 16 September 2014

The 97% v the 3% – just how much global warming are humans causing?

" it would be nice if you in fact let the IPCC explain, rather than cutting them of in mid explanation. "

Of course I didn't post the whole passage. I assumed you all knew where to find it! :)

Yes, it goes on into further detail. But I fundamentally stand by my orginal assessment. The 10.5 graph was primarily derivative of a very small group of papers discussing model outputs. Therefore, I think a statement like "The green bar shows the amount of warming caused by human greenhouse gas emissions during that time." is potentially misleading. The green bar is derived from climate model outputs. And that may indeed be dead accurate, but if you think so, then you think they underlying models are. And that's a subject of debate. The best way to gauge a model is by predictive ability, either forwards or backwards. And even with strong correlation it's not possible to eliminate coincidence as a possibility, except over increasingly long terms.

My take was that the 10.5 graph was [b]never intended[/b] to display the full picture of attribution, or theories behind it. And the considerable text in the section supports that view, rather than disproves it.

And I don't think "while Judith Curry from Georgia Tech represented the opinions of 2–4% of climate experts that we could be responsible for less than half of that warming." is well supported by evidence, or the IPCC here. In the same section, they go on to say that..

"We conclude, consistent with Hegerl et al. (2007b), that more than half of the observed increase in GMST from 1951 to 2010 is very likely due to the observed anthropogenic increase in GHG concentrations"

Very Likely, in IPCC parlance, is 90-100%. And that's if you agree with their conclusions there. And not every climate scientist does. But I certainly wouldn't want anyone to take my word for it. Verheggen et al. 2014 asked a number of climate scientists to provide a figure for attribution and roughly 2/3 rds reported above 50% anthropogenic. The remainder either less or uncertain. Setting aside method criticisms for the paper itself (close enough for this purpose), how does one reconcile this to the 2-4% estimate? For that matter, where does 2-4% come from? Not from any study I have read. Were too many climate scientists unaware of the CMIP5 and other model(s) results?

Moderator Response:

[JH] Please keep it civil.

Tom Curtis at 14:43 PM on 16 September 2014

The 97% v the 3% – just how much global warming are humans causing?

jwalsh @13, it would be nice if you in fact let the IPCC explain, rather than cutting them of in mid explanation.

To start with, as shown in Fig 10.4 below, the models are used to determine relative contribution but are scaled to match actual temperature increases. Thus if a model shows an anthropogenic temperature increase of 0.8 C, and a total increase of 0.7 C, then the anthropogenic increase is scaled by 0.65/0.7 to determine the anthropogenic contribution. Thus any tendency to over estimate the temperature trend is eliminated as a factor in determining attribution. All that remains is the relative responsiveness to particular forcings. With respect to that, it is well known that the combined natural forcings from 1951-2010 are slightly negative, or neutral at best.

Further, as the IPCC says:

"We moderate our likelihood assessment and report likely ranges rather than the very likely ranges directly implied by these studies in order to account for residual sources of uncertainty including sensitivity to EOF truncation and analysis period (e.g., Ribes and Terray, 2013)."

That is, they multiplied uncertainty by a factor of 1.36, thus substantially expanding the uncertainty range to account for any additional uncertainty relating to the methods used. The models, note, only over estimate recent temperature trends by 18%, half the expansion of the uncertainty range - and that overestimation has been eliminated from the attribution by scaling in any event.

Finaly, they go on to say:

"The assessment is supported additionally by a complementary analysis in which the parameters of an Earth System Model of Intermediate Complexity (EMIC) were constrained using observations of near-surface temperature and ocean heat content, as well as prior information on the magnitudes of forcings, and which concluded that GHGs have caused 0.6°C to 1.1°C (5 to 95% uncertainty) warming since the mid-20th century (Huber and Knutti, 2011); an analysis by Wigley and Santer (2013), who used an energy balance model and RF and climate sensitivity estimates from AR4, and they concluded that there was about a 93% chance that GHGs caused a warming greater than observed over the 1950–2005 period; and earlier detection and attribution studies assessed in the AR4 (Hegerl et al., 2007b)."

Thus while assessment of the relative contributions of Greenhouse and OA are restricted to the studies shown in Fig 10.4, the assessement of Anthro is not. Indeed, they go on for an additional two paragraphs citing other factors which influenced the determination of the Anthro contribution as portrayed in fig 10.5, and still further discussion calling on a host of further studies relating to the role of Internal variability (which unlike forcings, which will scale alike, thus preserving relative contribution, may have been an issue). They conclude that discussion by saying,

"To summarize, recent studies using spatial features of observed temperature variations to separate AMO variability from externally forced changes find that detection of external influence on global temperatures is not compromised by accounting for AMO-congruent variability (high confidence). There remains some uncertainty about how much decadal variability of GMST that is attributed to AMO in some studies is actually related to forcing, notably from aerosols. There is agreement among studies that the contribution of the AMO to global warming since 1951 is very small (considerably less than 0.1°C; see also Figure 10.6) and given that observed warming since 1951 is very large compared to climate model estimates of internal variability (Section 10.3.1.1.2), which are assessed to be adequate at global scale (Section 9.5.3.1), we conclude that it is virtually certain that internal variability alone cannot account for the observed global warming since 1951."

Consequently, your attempt to paint the attributions shown in Fig 10.5 as exclusively model bases seriously misrepresents the IPCC - a misrespresentation sustained by quotation out of context. Further, the attempt to portray an disagreement between model and observed trends ignores directly relevant factors by which the IPCC took that discrepancy into account.

One Planet Only Forever at 13:52 PM on 16 September 2014

2014 Arctic Sea Ice Extent Prediction

franklefkin,

5.35 will be very close, but is likely to be a little high.

The NSIDC interactive chart (here) shows the following Arctic 2014 ice extents (for the area with 15% or more sea ice):

September 1: 5.579 million
September 12: 5.140 million
September 13: 5.108 million
September 14: 5.086 million

So, with the ice extent likely to continue to decline a few days past the middle of September, the September average is very likely to be lower than 5.35. It may even be lower than 5.2.

Also, the September 14th extent is lower than the minimum 2013 extent (5.101). So 2014 is already the 5th lowest minimum extent in the record. But it is unlikely to reach 4th place (4.626 - September 22, 2010).

scaddenp at 13:51 PM on 16 September 2014

The 97% v the 3% – just how much global warming are humans causing?

jwalsh - I think you are making a number of incorrect assumptions here. Firstly, as to why attribution is done with a model, then please see here. Whether ensemble model mean over or underpredicted surface temperatures is a different question from attribution. For instance, see Gavin Schmidt's response here.

jwalsh at 12:37 PM on 16 September 2014

The 97% v the 3% – just how much global warming are humans causing?

I am uncertain why there is the focus on Figure 10.5 from IPCC AR5. Not even the IPCC, years ago, considered that representative of all areas of climate science. 10.5 is derived from 10.4 which is derived from.... Well may as well let the IPCC explain.

"The results of multiple regression analyses of observed temperature
changes onto the simulated responses to GHG, other anthropogenic and natural forcings are shown in Figure 10.4 (Gillett et al., 2013;
Jones et al., 2013; Ribes and Terray, 2013)."

The papers referenced (3 in total) are based on climate models, and observations from them. The extent to which anyone (including the IPCC) considers that a true picture of attribution, relies on the extent of belief in the accuracy of the model ensembles. And everyone knows there are problems there. Even the author of one of those papers authored another, "Overestimated global warming over the past 20 years" acknowledging observed model discrepencies. Numerous theories abound about why the models were inaccurate, but there is almost universal agreement that they were inaccurate. I have no idea why. Not my field. In fact, I think it's safe to say that nobody knows yet. Models, as expected, are going to continually be refined and get better and better.

sauerj at 12:07 PM on 16 September 2014

Thousands of ‘Nameless Short-Lived Lakes’

Request: Could someone post two sets of curves here? This would be a good reference visual aid for sharing purposes.

Set 1) Rate of SLR (Greenland contribution) vs year (2000 thru ~2120) in mm/yr; A family of 3 curves with doubling periods of 5, 10 and 12 years.

Set 2) Accumulated SLR (integration of above) (GL contrib) vs year (same). With same 3 doubling period curves. ... thanks in advance, sauerj

Moderator Response:

[ps] sentence fixed as requested.

Tom Curtis at 11:57 AM on 16 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Dana, small typo in the Gaurdian version (not included in the excerpt above):

"Recent research led my by Michael Mann has..."

Tom Curtis at 11:27 AM on 16 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Just out of general interest, if Russ R's supposition of independent errors for Greenhouse and OA were correct (and allow me to underline again, it is not), then the combined error would be +/-0.53 C for the "likely" range. That would allow a 17.3% chance of 50% or less anthropogenic contribution, but also a 17.3% chance of a 150% or more contribution. If that were the case, then there is a better than 1 in six chance that when natural factors equalize out in the long term (which is more or less inevitable, and certainly so for internal variations), we will be facing 50% greater than IPCC expected warming. That is, there will be a one in six chance that strategies to keep warming under 2 C would result in a rise in GMST of 3 C. As harm increases approximately exponentially with increased temperature, that makes the task of tackling global warming more urgent, not less.

Fortunately we are not facing that bad new situation. Also unfortunately, Russ R will not recognize that as a bad news situation because for him only the low side of the uncertainty equation is meaningfull (=usefull rhetorically).

Bob Loblaw at 11:22 AM on 16 September 2014

The 97% v the 3% – just how much global warming are humans causing?

An addendum: when referring to 100% correlation above, I mean 100% positive. If the correlation is negative (i.e., one error being positive means the other error is negative, i.e., covariance is negative), then the error in A+B will be less than the error in either A or B alone.

Bob Loblaw at 11:18 AM on 16 September 2014

The 97% v the 3% – just how much global warming are humans causing?

RussR:

Your addition of ±0.4°C and ±0.35°C to get ±0.75°C is only correct if the two variables are 100% correlated in their fluctuations. This is highly unlikely.

If the two errors are independent (0% correlation), then the correct way to add the errors is the root-sum-square approach:

(0.4)² + (0.35)² = 0.53²

...so the error in the sum (or difference) is ±0.53°C

For correlations between 0 and 100%, you use the covariance to adjust between 0.53 and 0.75.

Basic propagation of uncertainty in mathematics. I'm surprised you're not familiar with it.

KR at 11:08 AM on 16 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Russ R. - Because the sum forcings are more tightly constrained than the individual components. If natural forcings are in reality higher than estimated, then anthro forcings are lower, and vice versa.

This is similar to the case with estimates of top of atmosphere (TOA) imbalances, individual components of evaporative, convective, etc. in the Earth energy budget. Uncertainties on each component may be+/-10 W/m², while the total Earth imbalance over the last 50 years from ocean heat content (OHC) is +/- a fraction of a Watt. Because the values for the total are established from different evidence than the values for any one of the components.

Tom Curtis at 10:58 AM on 16 September 2014

97 hours of consensus: caricatures and quotes from 97 scientists

dhf @51, you have been caught out arguing from a clearly false premise, whose only support has been your mere assertion. It would be nice of you to acknowledge that your premise was in fact false, and has now been comprehensively refuted. Merely picking an isolated sentence as a bridge to an urelated rhetorical sally is evasive and shows further discussion with you to probably be unfruitful.

For what it is worth, while testing calibration is a useful test on theories, it often involves subtleties that require expert knowledge to understand, and hence it is not true that layman can check calibrations without detailed research. In particular, the nearest analog of "calibration" in climate science is validation of models, which is a very complex subject. In particular, it involves assessing the interplay of probabilities of multiple factors which are not independent; the ability to assess the significance of scale issues; and the ability to assess the relevance of timing of short term variable events such as ENSO (which because chaotic, can be reproduced statistically in models, but not with regard to exact timing and magnitude of particular events).

The denier strategy has been to simplify that complexity by focusing on just one variable (GMST) over a cherry picked interval (typically 1998 forward) and to declare the "instruments" to have failed because GMST almost drops out of the uncertainty range. They frequently claim that it has dropped out of the uncertainty range by treating 1998 (which almost exceeded the uncertainty range) as a mean value thereby offseting the temperatures relative to model predictions.

Further, nearly all of the purported model/observation discrepancy on GMST vanishes if models are contrained (either by selection or by constrained responses in the ENSO region) to match actual ENSO fluctuations. That is easilly demonstrable by a number of means, the simplest of which is simply taking trends of La Nina, neutral, and El Nino years seperately; and hence should not be subject to dispute. AGW deniers, however, simply ignore it as a factor in an approach best described as dishonest.

These approachs are not scientific. They are pseudo-scientific.

They are also not on topic on this thread.

What is on topic is your apparent claim that these "calibrations" can be easilly performed by laymen, but apparently have escaped the experts notice. Sorry, I see no point in debating such absurdities.

Tom Curtis at 10:35 AM on 16 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Russ R @6, I have already explained that in my second point @2. That you choose to ignore it is no reason for me to repeat myself (or to waste time on an elaboration that will be equally ignored by you if it does not serve your rhetorical interests).

Russ R. at 08:51 AM on 16 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Hello Tom Curtis,

If you agree that:

GHG =0.9±0.4°C
OA = -0.25±0.35°C
ANT = GHG + OA

Could you please explain how the uncertainty of ANT can be lower than that of either of its components?

Tom Curtis at 08:47 AM on 16 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Thankyou Leto. I did indeed misplace a decimal point, and that should be to the nearest 0.05 degrees.

Leto at 08:29 AM on 16 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Tom,

I don't think you mean "to the nearest half degree" (i.e. 0.5 degrees) - perhaps to the nearest half graph-interval, where the graph-intervals (gap between axis ticks) are 0.1 degrees.

dhf at 08:05 AM on 16 September 2014

97 hours of consensus: caricatures and quotes from 97 scientists

@ Tom Curtis at 06:43 AM on 12 September, 2014

"Therefore we no longer include classics scholars among the people we relly on to understand physics. Instead we relly on scientists."

If we rephrase this argument you say that: "We rely on scientists to understand physics." I will say that there is much more to it than that. For important matters we are not likely to rely on scientists. We will likely require to see independent data demonstrating that the result of applying the physics is ok. When the product is a measurement result or a prediction, the responsible will likely require to see data demonstrating the uncertainty of the prediction and that it is without systematic errors. The application of physics rapidly becomes extremely complex also to the scientists applying it. This complexity implies that nobody can be sure that end product is ok without testing it. That is why testing and experimenting is an extremely important part of physics.

Imagine constructing a complex instrument, like a new electronic weighting scale or a much more complicated instrument. To build the instrument you depend on hardware components, electronics, software, primary measurements, correctly set parameters, physical models etc. Many scientists are required to design the instrument, develop the models, build the instrument, develop the software and test it. The instrument is based on physics, but still, there is an enormous amount of possible errors which can be made when designing and building the instrument. Hence you do not know the uncertainty of the instrument or if there is a significant systematic error in the instrument before you perform a calibration. A calibration is performed by letting the instrument perform measurement on a number of test weights traceable to international standards. The central information on the calibration certificate consist of the measurement result from the weighting scale, the mass of the test weights and the uncertainty of the mass of the test weights.
Even if no single person has all the competence that is required to design and build the instrument, the calibration certificate is very easy to understand. The calibration certificate can be completely understood by a person without any of the competencies required to design and build the instrument. Hence the capabillities of the instrument can be completely understood without having to rely on the scientists required to design and build it. There are also other ways to demonstrate that a complex construction works for its intended use. Even if I have no clue how it is made, I can still tell if a television works. There is also more to it, because, even the scientists who built the complex instrument cannot be sure about the uncertainty of the instrument and that it is without systematic errors without also using the same kind of data a layperson can use to comprehend the capabilities of the instrument. Hence I think it is more valuable to say that we rely on other scientists, text book´s and data tools to understand physics and construct a product or a model, we rely on testing to se if it works. A properly performed test and properly presented test results can be easy to comprehend. To summarize I would say that We do not rely on scientists in physics, we rely on data from calibration, testing and experiments. In that respect, I would also like to quote William Edwards Deming:
“In God we trust; all others bring data.”

Tom Curtis at 07:48 AM on 16 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Further, to Russ R @1, I did not note Russ R's absurd treatment of error where he merely added the error bounds of Greenhouse and OA. That would be mathematically in error even if the uncertainties are independent, which they are not as explained above.

Tom Curtis at 07:46 AM on 16 September 2014

The 97% v the 3% – just how much global warming are humans causing?

Russ R @1, here are the relative contributions as determined from Fig 10.5 by pixel count:

(Press on figure for larger version).

I note first that the Anthropogenic contribution (Anthro) has a far more tightly constrained uncertainty than does either the Greenhouse gass contribution (Greenhouse) or the Other Anthropogenic contribution (OA). Further, the Anthro contribution has a mean (and median) contribution of 107.7% of the total increase in temperature. Therefore you are using the less certain data in preference to the more certain data to make your case, which is the wrong way to go about it.

Second, the uncertainty of Anthro is substantially less than the uncertainty of the other four (including Natural forcings (Nat), and Internal Variability (Intern Var), which have an uncertainty about that of Anthro, but centered on, or very near zero) because their uncertainties are not independent. If we estimate Greenhouse to be higher, it follows that OA, Nat, or Intern Var, or some combination of them are lower, and vice versa. Further, estimates of the relative contribution of Intern Var are inversely depedent on the estimated strength of all forced responses together.

Because of this interdependence, you cannot simply add mean or median values for Greenhouse and OA to determine the mean or median value for Anthro.

That is particularly the case because, third, the uncertainty of Greenhouse (almost definitely) and of OA (probably) are slightly right biased. This shows up in the pixel counts of Fig 10.5, but is so small a bias (1 or 2 pixels at 150% magnification of the large version of the image on the IPCC site) that I (and probabily Real Climate) ignored it for mathematical simplicity. The bias in both cases, however, is a right skew meaning that the simple addition of their means will underestimate the total anthro contribution.

Further, that the Anthro, Greenhouse and OA contributions are normally distributed is just a simplifying assumption made by me and Real Climate. Often in these cases the PDF is log normal with a right skew, rather than normal. If that is indeed the intent of the IPCC, then that would further bias the results of simple addition. Of course, a log normal distribution would also greatly reduce the probability of a low (50-75%) Anthro contribution and represents a conservative assumption for the AGW case.

Fourth, and most simply, you neglect the probability of rounding errors. If the IPCC has rounded figures to the nearest half degree (likely given that all values amount to half degree figures in the table), then the difference may be due to rounding error alone. In that case the mean (or median) Anthro contribution may be as little as 100%, but may be as much as 116%, with the additional uncertainty due to rounding error. That is an additional level of uncertainty not accounted for in my graph (nor, I suspect, in Real Climate's graph), but it will only negligibly increase the probability of low (50-75%) attributions, while equally increasing the probability of high (140-165%) attributions.

For the record, ignoring any rounding errors, and assuming a normal distribution, the lower bound of the 99% confidence interval of Anthro is approximately 78%, and of the 99.9% confidence interval it is 62%. Therefore merely shifting the confidence interval 8% lower will leave a lower bound of the 99.9% confidence interval 57%. That would be an absurd response to the additional error from rounding, but even it does not result in substantial probabilities of less than 50% anthro contribution.

Russ R. at 06:27 AM on 16 September 2014

The 97% v the 3% – just how much global warming are humans causing?

The probability density function shown relies on an estimate that warming due to anthropogenic forcings was 0.7±0.2ºC (5-95%), while observed warming was 0.65ºC, (citing IPCC AR5 Figure 10.5).

However, the actual chart in Figure 10.5 displays something rather different.

By the IPCC's definitions, combined anthropogenic forcings (ANT) equals greenhouse gas warming (GHG) plus other anthropogenic forcings (OA), or ANT = GHG + OA

And (reading off the scale) they show GHG =+0.9±0.4°C and OA = -0.25±0.35°C.

So, by simple math, ANT = 0.65±0.75°C.

So, the PDF would would be centered around 100% (not 110%) of the observed warming with (5-95%) uncertainty of ± 115%.

How would that change the shape of the PDF, and what would be the likelihood of ANT/Observed being less than 50%?

Phil at 20:52 PM on 15 September 2014

2014 SkS Weekly News Roundup #37B

@1 From Peru asks "What do you think" ?

Given your quote, there are three seperate statements made by Modi:

"On climate change, the Prime Minister said exploitation of nature was a crime.
“At best, you have the right to milk nature. You can milk a cow, but you can’t kill the cow.”
“Climate change? Is this terminology correct? The reality is this that in our family, some people are old... They say this time the weather is colder. And, people’s ability to bear cold becomes less.

In Statement 3 he suggests that the temperature record is not accurate, i.e. he questions whether the planet has warmed at all

Statement 2 seems to imply that we can continue BAU if the planet hasn't warmed (because we are only "milking" the cow)

Statement 1 seems to imply that because we are only "milking the cow" then BAU is not exploitation and so not a crime.

Thats how I read it.

From Peru at 15:21 PM on 15 September 2014

2014 SkS Weekly News Roundup #37B

My point was that the Guardian article attacked Modi suggesting he has turned into a climate skeptic, just few years after expressing strong concerns about global warming and supporting a "renewable energy revolution" in his country, like giving solar panels to every family in the 1000 billion people federation.

That seemed a lot strange to me, and in effect, I saw abolutely no hint of anti-environmental statements in the original source (The Hindu).

I have one concern about the Guardian article: Modi is a right-wing politician, so he strongly supports private investments instead of a state-dominated economy. One could agree with that or not, but one should praise the right policies no matter of what political end of the spectrum is involved.

In the Anglo-American world the political right seems to be highly anti-environmental, pro-pollution, but elsewere in the world that it is not the always the case:

Germany, governed by a centre-right coalition has been leading Europe in the Energy transition away from fossil and nuclear fuels
In Latin America some of the worst polluters are the petro-states like Venezuela (that have the world top tar sand reserves) and Argentina (that is developing a giant shale oil and shale gas in Patagonia), both left-wing, and in the case of Venezuela, we are talking about an authoritharian stalinist regime.

I hope my concern is wrong.

N.B.: I belong to both the left-wing and the libertarian end of the spectrum (google "political compass" where is shown the 2-dimensional political spectrum)

Moderator Response:

[JH] Your points are well taken. Only time will tell if Modi is a champion for tackling climate change or not.

chriskoz at 14:17 PM on 15 September 2014

2014 SkS Weekly News Roundup #37B

From Peru@1,

I find quoted Modi’s arguments non-sequitur: is he commenting about the human causality of AGW or about climate change impacts on humans or the old age impact on an aging human individual? Those 3 observations are independent issues and each should be commented on separately. By mixing those 3 in a single comment, Modi is making a logical error and the resulting statement is a fallacious bubbling rather than a reasonable opinion/assessment. Unless the true sense of his statement was lost in translation but that possibility is low in case of Modi who speaks fluent English and must have pronounced his comment in English rather than in Hindu or his other native tongue.

noa at 22:54 PM on 14 September 2014

Thousands of ‘Nameless Short-Lived Lakes’

MA Rodger @ 12.

Absolutely agreed. I wanted to edit my post @10 and add the caveat that there is to little data to decide wether the trend is linear or exponential (but I needed a break from my heroic struggles with the foreign language monster).

I think Hansen himself says that in the first paper you linked to. But as you show with your projection, even the SLR from a constant acceleration would be quite substantial. Thanks for that.

I guess it is time to buy a boat, a rubber boat will have to do it for now ... "Way hay and up she rises, ..." :-)

Have a nice day

MA Rodger at 19:20 PM on 14 September 2014

Thousands of ‘Nameless Short-Lived Lakes’

noa @10.

Thanks for the info. Sad to hear that Tom & Jerry the two GRACE satellites are running out of umph.

The GRACE data at the NOAA's 2013 Greenland Report runs up to July 2013 and it looks from the graph in the link you provide to the Polar Portal that the last 4 moths after March 2013 have been pulled.

The IMBIE/ICESat/CryoSat data does show a doubling, but so indeed does the GRACE data. The numbers scaled from the NOAA graph show 2011-13 averaging ~415Gt/yr loss, while the 2003-2009 data gives a ~210Gt/yr loss. But I do not see this, a doubling in 6 years, as being inconsistent with even a constant acceleration of ~30Gt/yr/yr as illustrated in the graph I link to @6.

Not that such a linear trend is small. If sustained to 2100 we will still be receiving 420mm SLR from Greenland alone.

noa at 17:54 PM on 14 September 2014

Thousands of ‘Nameless Short-Lived Lakes’

I just remembered that the Morlighem paper was covered on SkS here, of course. ;-)

An other paper re SLR projections is this one from Spence et al. 2014 (free PDF). They seeked to improve the modeling of the consequences from the poleward shifting westerlies in the SH, which is - according to them - projected to persist under continued anthr. forcing. They find that the westerlies weaken the Antarctic Slope Front (a counter-current on the outer antarctic shelfs, if I am not mistaken) and this opens the doors for the relative warm circumpolar deep waters. From that they project a warming of the waters near the grounding lines of some outlet-glaciers up to 2 degrees till 2100. This feedback is, as they say, not included in the AR5 SLR projections, because the resolution of the CMIP5-models is not high enough to account for the changes of the Antarctic coastal waters.

(Please bear with me if I made some basic errors in my summary - I am just an interested layman).

Well, I'll better quote the abstract of the paper:

The southern hemisphere westerly winds have been strengthening and shifting poleward since the 1950s. This wind trend is projected to persist under continued anthropogenic forcing, but the impact of the changing winds on Antarctic coastal heat distribution remains poorly understood. Here we show that a poleward wind shift at the latitudes of the Antarctic Peninsula can produce an intense warming of subsurface coastal waters that exceeds 2°C at 200 – 700 m depth. The model simulated warming results from a rapid advective heat ﬂ ux induced by weakened near-shore Ekman pumping and is associated with weakened coastal currents. This analysis shows that anthropogenically induced wind changes can dramatically increase the temperature of ocean water at ice sheet grounding lines and at the base of ﬂ oating ice shelves around Antarctica, with potentially signi ﬁ cant rami ﬁ cations for global sea level rise.

noa at 17:06 PM on 14 September 2014

Thousands of ‘Nameless Short-Lived Lakes’

@6 MA Rodger

As far as I know there seem to be some issues with the GRACE data for 2013. See repley 156 here on the ASIF. Answer from the DMI Polar Portal (see here there GRACE chart):

Your question is really good, and the answer is actually quite simple: The GRACE mission is already way past the originally intended duration, but the satellites are still flying. But systems do fall out once in a while and, as an example, the 2013 summer data are unavailable due to power system problems. The 2013 summer data are thus missing from the Polar Portal GRACE figure. The linear interpolation across the summer negative peak suggests an extremely low (even no) loss summer and therefore is very misleading. We will work on a different way of representing this.

Recent GIS MIB estimates derived from CryoSat-2 published in TC (Helm 2014) do not look that promissing (within doubling rates faster than 10 years). From p. 13 table 4:

IMBIE 2003-2008: −189 ± 20 Gt

ICESat 2003-2009: −146 ± 13 Gt

CryoSat 2011-2014: −375 ± 24 Gt

Climate Science Glossary

Term Lookup

Settings

Term Lookup

Settings

Latest Posts

Recent Comments