It seems depending on who you talk to, climate sensitivity is either underestimated or overestimated. In this case, a model suggests forcing is underestimated. One thing is clear, science does not yet know for certain what the true climate sensitivity to CO2 forcings is.
There is a new Paper from Tanaka et al (download here PDF) that describes how forcing uncertainty may be underestimated. Like the story of Sisyphus, an atmospheric system with negative feedbacks will roll heat back down the hill. With positive feedbacks, it gets easier to heatup the further uphill you go. The question is, which is it?
Insufficient Forcing Uncertainty Underestimates the Risk of High Climate Sensitivity
ABSTRACT
Uncertainty in climate sensitivity is a fundamental problem for projections of the future climate. Equilibrium climate sensitivity is defined as the asymptotic response of global-mean surface air temperature to a doubling of the atmospheric CO2 concentration from the preindustrial level (≈ 280 ppm). In spite of various efforts to estimate its value, climate sensitivity is still not well constrained. Here we show that the probability of high climate sensitivity is higher than previously thought because uncertainty in historical radiative forcing has not been sufficiently considered. The greater the uncertainty that is considered for radiative forcing, the more difficult it is to rule out high climate sensitivity, although low climate sensitivity (< 2°C) remains unlikely. We call for further research on how best to represent forcing uncertainty.
CONCLUDING REMARKS
Our ACC2 inversion approach has indicated that by including more uncertainty in
radiative forcing, the probability of high climate sensitivity becomes higher, although low climate sensitivity (< 2°C) remains very unlikely. Thus in order to quantify the uncertainty in high climate sensitivity, it is of paramount importance to represent forcing uncertainty correctly, neither as restrictive as in the forcing scaling approach (as in previous studies) nor as free as in the missing forcing approach. Estimating the autocorrelation structure of missing forcing is still an issue in the missing forcing approach. We qualitatively demonstrate the importance of forcing uncertainty in estimating climate sensitivity – however, the question is still open as to how to appropriately represent the forcing uncertainty.
h/t and thanks to Leif Svalgaard
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
M. Simon says:
What you are missing is the simple fact that an infinite series can converge to a finite value. Let’s take a very specific example: Let’s suppose that each degree of warming due to whatever cause has the first-order effect of causing an increase in water vapor that then produces another 1/2 degree of warming. So, that means that if you put enough CO2 into the atmosphere to cause 1 C of warming by itself, the additional water vapor then produces 1/2 C of warming, then the additional water vapor that this 1/2 C of warming produces causes an additional 1/4 C of warming and so forth. What you have is the infinite geometric series 1 + 1/2 + 1/4 + 1/8 + 1/16 + …, which converges to 2, meaning that the positive feedback amplifies the warming by a factor of 2.
To be honest, I am not familiar enough with the electronic design field to know exactly how they use the term. I just know that a few engineers here have stated that it is used differently and that “positive feedback” in their lingo refers to an unstable system.
It is well-understood in climate science that the issue of sensitivity to initial conditions prevents you from predicting the DETAILS of the climate over the long term. In fact, the models are often run in ensembles with slightly perturbed initial conditions. However, even when the models produce different DETAILS (e.g., when a particular El Nino or La Nina event occurs), they produce the same general result in response to a forcing over a long enough period of time (where a long enough period of time essentially means that the temperature change due to the forcing dominates over the temperature fluctuations due to natural variability such as ENSO).
First of all, the prediction to 2020 is a prediction of one recent paper (which really predicts a pause in the warming, not a cooling). It is not “the latest IPCC estimate”.
Second of all, it is indeed true that the generally-accepted way of running the models does not allow the prediction of short term trends. One can compare the statistics of the models with the statistics of the actual climate to see how well the models capture natural variability (and, I believe the answer is pretty well, although not perfectly). And, for example, one can see how common a decadal or so period with a negative temperature trend is in models run with steadily-increasing GHG forcing (the answer being that it is fairly common, see http://www.cdc.noaa.gov/csi/images/GRL2009_ClimateWarming.pdf ). However, which actual specific realization the climate follows is very sensitive to the initial conditions.
There is recent work, such as a paper from the Hadley group and another by Keenlyside et al, that attempt to make actual short-term climate predictions (i.e., predictions on the scale of several years to a decade or so) but these are still pretty experimental and there is considerable debate about whether they have been done correctly. These attempts basically work on the idea that, while there is extreme sensitivity to initial conditions, the timescales for the oceans are slow enough that it may still be possible to get a reasonable prediction for the oceans over these decadal timescales if one starts with a good initialization of the ocean’s current state and that the ocean will then tend to strongly influence the general characteristics of the climate. However, this is quite tricky to do in practice.
By the way, I should just add that I don’t really understand what your excursion into Lorenz chaos had to do with my original point to Steve Keohane that his estimate of a trend was way off because of the various ways in which he had cherrypicked the data. Even the UAH folks themselves report a trend over the entire satellite record since 1979 of ~0.13 C per decade. I think the RSS folks and the HadCRUT and NASA GISS surface data support a somewhat larger trend of something like 0.17 C per decade over that time period.
Joel Shore (13:31:15) :
“What you are missing is the simple fact that an infinite series can converge to a finite value. Let’s take a very specific example: Let’s suppose that each degree of warming due to whatever cause has the first-order effect of causing an increase in water vapor that then produces another 1/2 degree of warming. So, that means that if you put enough CO2 into the atmosphere to cause 1 C of warming by itself, the additional water vapor then produces 1/2 C of warming, then the additional water vapor that this 1/2 C of warming produces causes an additional 1/4 C of warming and so forth. What you have is the infinite geometric series 1 + 1/2 + 1/4 + 1/8 + 1/16 + …, which converges to 2, meaning that the positive feedback amplifies the warming by a factor of 2.”
Unless, of course, it rains.
M. Simon (11:02:12) : “Some other forcing (internal variability?) has been aliased for CO2 forcing.”
Hydrological cycle?
Joel Shore (13:31:15) :
Rising sea levels would mean more ocean to influence, and slow the process, no?
Tom in thunderstorm cooled Florida (14:53:35) :
An inconvenient truth.
Let’s have a look at some of the numbers that are used for the TempC response per watt/m^2.
First, a simplified version of Trenberth’s and the IPCC’s Earth Radiation budget.
http://en.wikipedia.org/wiki/File:Greenhouse_Effect.svg
Incoming solar radiation – 235 watt/m^2 – responsible for raising the Earth’s temperature from 4 kelvin to 255 kelvin or 1.07C /watt/m^2.
Greenhouse effect – 324 watt/m^2 – responsible for the 33C greenhouse effect or just 0.1C /watt/m^2.
How about doubled CO2/GHGs according to the IPCC – 4 watt/m^2 – which will translate into a 3.0C rise in temperatures or 0.75C /watt/m^2
In Gavin’s solar paper above – we have the estimated numbers for the solar cycle influence ranging from 0.05C /watt/m^2 to 0.40C /watt/m^2 – but greenhouse gases give numbers anywhere from 0.31C /watt/m^2 to 0.86C /watt/m^2.
So, we have “climate science” math that uses any figure from 0.05C /watt/^2 to 1.07C /watt/m^2.
Seems a little unusual to me (unless you already the know the number you want to get and the intermediate steps can be camoflaged in wierd terms like the Forcing is estimated to be X watts per metre squared and the Z value is 0.xxK/[watt M-2].
Bill Illis (17:30:19) :
Those numbers indicate that CO2 was responsible for 1.9C of the 5.0C change in temperatures during the ice ages.
So the majority of the temperature change, 3.1C, is due to other factors.
How do we know that the other factors are not in fact responsible for 4.0C and CO2 was only responsible for 1.0C (or the same 1.5C per doubling).
I understood that the log temp vs CO2 was due to the CO2 absoption bands being nearly saturated.
If this is the case then an Ice age with very low levels of water vapour in the air would desaturate some of these absoption bands making the sensitivity to CO2 non-logarithmic.
Looking at the CO2 CH4 temp from various ice core data seems to show that CO2 rises from 200ppm simultaneously with temperature at the end of the ice age:
http://img11.imageshack.us/img11/6826/iceage040kkq1.jpg
Entry to the ice age seems to be co-incident with CH4 reduction.
So during the ice age CO2 levels are low H2O levels are low CH4 levels are low which gives a much less saturated IR window. Which gives greater sensitivity to GHGs until levels equivalent to todays are attained.
As I and many others have said gouing back more than 9M years and talking of CO2 levels and temperature is irrelevant. The worl was too different.
Further, the levels of CO2 are obtained from a computer model (geocarb 3, i believe). Why should this be more accurate than current climate models?
check out this site for continental movements:
http://www.scotese.com/newpage13.htm
Leif Svalgaard (03:45:18) :
tallbloke (23:34:55) :
Does Leif Svalgaard agree with the characterisation of solar forcing displayed in the graph shown, and please would he explain what the red solar curve is representing in case I misunderstand it.
The red curve shows some representation of the solar cycle and does not seem too much out of whack. It has the smallest amplitudes of all the forcings, so is in line with what I would expect.
Had to add my 2 bits worth ;
the graphs seem to show percentage change as if amounts are equivalent.
As any engineer in refrigeration can tell you a small shortage or overage of energy input can make a large change in temperature over time in a balanced, constant operated system. That is why a system is over built to allow for forced control.
Solar input is by orders of magnitude the most important factor in the earths climate,and the hydrosphere is the second.
I’m not sure why vulcanism only shows negative periods and aerosols show a marked decrease over the graphs time.
Joel,
In a chaotic system if you can’t get the details right the farther you go into the future the more the divergence from reality. Because every prediction (time slice) is the initial condition for the next time slice and the system is sensitive to initial conditions.
Now if the system was a well damped amplifier with negative feedback (in the engineering sense) you would get a reversion to the mean. In a chaotic system that reversion to the mean is not assured. Which is why the terms “strange attractor” and “period doubling” are used for those kinds of system rather than amplification factor.
Now let us look at a system with positive feedback (in the electrical sense). As long as the feedback factor is less than 1 the system will be somewhat stable to component changes. A feedback factor of .1 in such a system is rather stable although gain changes and component changes will be multiplied rather than divided. As the feedback factor gets higher the gain of course increases exponentially. If you go to a feedback of .99 the gain is multiplied by 100 (roughly). At .999 (a small change) the gain is multiplied by 1,000. Things are getting twitchy. At .9999 – 10,000. Really twitch. At a feedback of 1.000 or above you get either a Schmidt Trigger or an oscillator depending on lags. The Schmidt trigger is the alarmists tripping point. However, with lags in the system the more likely result in oscillation.
BTW in electronics the positive feedback amplifier was invented by Armstrong and the common name for it (if you want to use Google) is “regenerative receiver” or “Q multiplier”.
If you look at the temp recodes for the last million years you do see tripping points. Those represent the transition from glaciation to inter-glacial periods. However, in the periods themselves negative feedback seems to dominate because the temps in the periods are relatively stable. So glaciation and inter-glacials seem to be the strange attractors.
Which says that excess warming is not our problem. Tripping into a glacial period is our problem. If CO2 is as effective as you assume (I doubt it) we should be pumping it out as fast as we can because given the historical record we are due (overdue?) for a glacial period.
What should we look for re: the tripping point? Increasingly chaotic weather (period doubling). Are we seeing that? Hard to say. The data record is not long enough.
Let me also note that lagged negative feedback (phase delayed) can also lead to oscillation if the phase is delayed by 180 deg. Below 180 deg the system gets twitchier the closer you get to 180 deg. It also becomes more sensitive to certain frequencies of input.
So positive feedback is not the only way to get a twitchy system.
But getting in deeper – the phase delay looks like a positive feedback.
Really. It is rather sad that climate “scientists” are so ignorant of electrical analogs.
So what do we have – a system being banged with 24 hour cycles, yearly cycles, solar cycles, PDO cycles, etc. If the phase delay for feedback matches any of those cycles oscillation will result. And guess what we have ENSO which is rather chaotic and the PDO which is more stable. Which says that the PDO is likely caused by some in phase lag. Where as ENSO is probably due to some not quite in phase lag.
So the deal is: positive feedback or negative feedback – lags matter.
Joel,
Re: feedback – the climate models do not match the data:
http://wattsupwiththat.com/2009/06/02/lindzens-climate-sensitivty-talk-iccc-june-2-2009/
My take:
http://powerandcontrol.blogspot.com/2009/06/who-ya-gonna-believe.html
No resort to positive feedback is require (besides that is too simplistic anyway) it could all be explained by a difference in lags (so called heat in the pipeline) between reality and the models.
It would be really great if climate “scientists” were smart enough and/or educated enough to produce an electrical analog of the climate system.
DaveE (14:45:29) :
“Allan M (14:09:24) :
You’re wasting your time with that one. Warmists don’t appreciate that positive feedback also acts on the signal fed back, I’ve had the argument too many times now.
DaveE.”
I have to admit you are right. They just evade the issue. But then it is all words (and numbers) to them. If they don’t understand it themselves, then they can’t explain it.
Bill Illis says:
While the sensitivity is expected to be fairly constant over some range, it is not expected to be constant over a range defined by taking the sun all the way from zero strength to full strength or the greenhouse effect all the way from zero strength to full strength. Hence, the numbers that you calculate are not estimates of the climate sensitivity for the current climate.
M. Simon says:
I know how chaotic systems work. However, extreme sensitivity to initial conditions does not mean that nothing can be predicted. For example, while we may not be able to predict the weather on July 4th in January, we can still predict that the climate here in Rochester will be much warmer in July than in January. Likewise, the models robustly predict the warming in response to increasing greenhouse gases when the initial conditions are perturbed, even though the year-to-year wiggles are sensitive to these initial conditions.
As for tipping points, they represent an additional factor of concern. Your hypothesis that we are unlikely to encounter one when we are applying a forcing in the warming direction is based on nothing more than hope.
The electrical analog may be useful to a certain extent and I am sure that there are plenty of climate scientists who understand the electrical analog, but analogies can only get you so far.
Joel Shore (07:43:37) : “However, extreme sensitivity to initial conditions does not mean that nothing can be predicted. For example, while we may not be able to predict the weather on July 4th in January, we can still predict that the climate here in Rochester will be much warmer in July than in January.”
One reason you can predict this is because the temperature range/variation is in the tens of degrees, probably in the neighborhood of 80 degrees F. AGWers are talking about long term predictions of a couple of degrees.
Another reason you can predict this is because you have available a history of hard data evidence to support your prediction something sorely lacking in climate models.
@Tom in thunderstorm cooled Florida (13:59:44) :
You accidently left out the minor detail that seasonal variation is cyclical and not chaotic. Joel seemed to be playing a little fast and loose with that analogy to climate models.
H.R.
I also left out the minor detail of Earth’s obliquity which is the well known and documented cause of Joel’s “successful” prediction about seasonal temperature differences.
H.R. says:
And, that is relevant how? The rise in forcing due to the rise in greenhouse gas levels is steady and not chaotic too. Chaos, i.e. extreme sensitivity to initial conditions, is a property of the SYSTEM…If the weather / climate system exhibits chaos (which it does), it will exhibit this chaos whether it is with a cyclical forcing like the seasonal variation or a steadily increasing forcing like that due to greenhouse gases. However, in both cases, there are still predictions that we can make about how the climate will respond even if the chaos prevents us from predicting some of the details.
Tom in high and dry Florida says:
Yes…And, increasing radiative forcings due to greenhouse gases are the cause of the climate changes under increasing greenhouse gas concentrations. Your point is…?
Joel Shore (13:18:13) : “Yes…And, increasing radiative forcings due to greenhouse gases are the cause of the climate changes under increasing greenhouse gas concentrations. Your point is…?”
You said previously: “For example, while we may not be able to predict the weather on July 4th in January, we can still predict that the climate here in Rochester will be much warmer in July than in January.”
My point is this, you were trying to equate being able to predict a well known seasonal temperature change at a location caused by the tilt of Earht’s axis to climate models predicting tiny temperature changes many years from now due to the unproven hypothosis of man made CO2.
May I add that you knew exactly what my point was.
Tom,
Actually, I didn’t know what your point was. I would agree that there are quantitatively greater uncertainties associated with predicting future climate change from the buildup of CO2 than there are in predicting the seasons (although the effect of CO2 is far from “an unproven hypothesis”). However, my point is that in both cases, these sorts of predictions can be made despite the fact that the system is chaotic. Chaos constrains our ability to make certain kinds of predictions but it does not eliminate our ability to make any predictions.
Hi, Joel.
I can’t add much to Tom’s latest reply to you because I was seeing things pretty much the way he was calling them. Your analogy of climate models predicting a few degrees a hundred yers out to predicting the seasons just doesn’t seem to be apt.
Lucia has some interesting comments on models in this post.
http://rankexploits.com/musings/2008/gavin-schmidt-corrects-for-enso-ipcc-projections-still-falsify/
I read discussions such as the one I just linked to at The Blackboard and I see no particular reason to abandon my skepticism of the skill of climate models. As for seasons, I’ll grudgingly go with the consensus that predicts fall following summer, which follows spring, which follows winter. I’ve got a bit of the herd mentality about seasons.