American Physical Society and Monckton at odds over paper

peer reviewer says: “another reason I think about heat capacity issues is that they allow cumulative energy transfers, essentialy slow flows. take the .o25 watts for burning things which doesnt seem to be much but in ten years its .25 watts and in 40 years its 1 watt/m^2. so its heat added to the system. Ask yourself: if you were to forget c02 but integrate heat evolved since 1750 what temperature would you get?”
Your reasoning here is confused. Watts is a unit of power, i.e., energy per unit time…so, no, 10 years of 0.025 W/m^2 is not 0.25 W/m^2. It is still 0.025 W/m^2. Now, it is true that the total cumulative energy (the power integrated vs time) would add up…but that is also true of all the other energy sources, in particular, the sun…or the radiative forcing from added CO2. So, it makes sense to just compare the powers for the different sources. And, the point is that the power per unit area we directly generate by burning fossil fuels and such is about 2 orders of magnitude down from the radiative forcing that we are producing by increasing CO2 levels and thereby trapping more of the sun’s energy. Of course, eventually, if we start using enough energy…e.g., 50 or 100X what we use today, we would not to start worrying about this heat input. But, it is not an immediate concern.

Duae Quartunciae (12:47:58 ) :
Turning energy into Watts, i.e. power, makes the same mistake. Power is not conserved. It is energy that is conserved.

You’ll need to be familiar with the SI units to work with the almost written analysis of this stuff. Heat is a form of energy; and Joules are the SI energy unit. One calorie is 4.18 J.
Power is energy per unit time, and is measured in Watts (one Joule per second).
The conservation of energy implies a balance of power. This is really really basic stuff. Because energy is conserved, and because the Earth is not generating its own internal energy, the total rate at which energy comes in is equal to the rate at which it goes out. (You can manage heat sinks as well by giving the flux of energy into the reservoirs as part of the equation; or you can look at equilibrium conditions, in which case reservoirs don’t matter.) The flux of energy is a rate of energy per unit time; which is a power value.
Almost everything about this problem uses power. The Sun gives a certain amount of energy per unit time. That’s power. The Earth emits heat radiation all the time; a certain amount of energy per unit time. That’s power. The effect of a change in the Sun is a change in the rate at which energy is received, which is power. The effect of a change in greenhouse effects alters the rates at which energy is absorbed and emitted. Power. You still use conservation of energy; this is the basis of the “energy balance” equations — which basically say that total power in and out together is (at equilibrium) zero.
This isn’t about disputing for and against AGW, by the way. It’s just about getting a bit of basic physics to even engage the subject effectively at all.

If you are a physicist, then you should be able to see that “power is conserved” is a strawman. No-one uses a conservation of power; they use a BALANCE of power, which corresponds to conservation of energy. If there is any imbalance in power, then there is a transfer of energy, which has to be stored somewhere, or else transfered again to make up the balance in power.
If energy is being stored into some reservoir, then you can use the capacity of that reservoir to figure how long the imbalance can persist (since power by time is energy), or you can make the flux into the reservoir explicit in a balance equation; and bingo, we’re right back at balance of power again.
Your speaking of “errors”, in this context, is …. odd; to put it as gently as I can manage. Especially if you are a physicist. The power treatment is perfectly correct, and pretty much essential when you are speaking of rates of warming or cooling, or looking at the temperature of anything the Sun shines upon.
Models, and measurements, continue to be refined and mutually tested. None of the actual claimed “errors” described here actually stand up as being anything outside of measurement uncertainties; and certainly no actual simple “error” has been shown in physical models of planetary thermodynamics, which could be attributed to mix up of power and energy.

links
http://www.nasa.gov/centers/goddard/images/content/93391main_AngularMomentum.jpg
http://www.nasa.gov/centers/goddard/news/topstory/2003/0210rotation.html
http://www.iers.org

Duae Quartunciae (06:02:50) :
“If you are a physicist, then you should be able to see that “power is conserved” is a strawman. No-one uses a conservation of power; they use a BALANCE of power, which corresponds to conservation of energy. If there is any imbalance in power, then there is a transfer of energy, which has to be stored somewhere, or else transfered again to make up the balance in power.”
Juggling like?
Maybe no one uses conservation of power, but it sounds and smells very much like it when reading the various “forcing” set ups: translating other forms of energy into power. That is where oversights happen: ignoring PDO and ENSO and what not acronyms of the great convection currents of the earth, ignoring precipitation and evaporation, ignoring the heat outlets from magma at the ocean floor, the crucial contributon to albedo from clouds and land management, the heat capacities of everything,…Turning everything into energy per second when the time scales of the earth are decades and the time scales of the solar system centuries and aeons.
Lets put it a different way. It is as if, when faced with a perturbative expansion of the solution of a differential equation system , the first terms are used ( averages) whereas the higher terms can be highly outbalancing and influential. This method works for a few steps, because most solutions have a first order term in the expansion and the divergence of the higher terms can be absorbed in the constant. That is why weather prediction from these complicated models for a few days is fine. Once the variable steps become many, the higher terms come in with a vengeance, and that is why the models cannot tell the weather for next month.
Let alone for next century.
Monks at some decadent time had a blessing on fridays, and christend meat “fish” so they could have it at the table.
That is what has happened by taking weather models and calling them climate models.

Anna, think we have moved away from the notion that there’s something wrong with using “power” which is fixed by using “energy”, and that we agree no-one is using an oxymoron like conservation of power. That’s good!
But hey… let’s proceed a bit futher. Even if we don’t agree, some of the differences may be clarified.
Weather is certainly chaotic, but that does not mean temperatures in a weather model diverge without limit. It means that after a comparatively short period of time, a weather model has no idea of what weather will be within the given distribution of weather.
At a given location, you can obtain a probabilistic distribution for different weather variables (temperature, wind speed, humidity, precipitation, anything you like), for a given season. If you want to know the weather next year, a weather model is no good to you. The best you can do is the probability distribution as an idea of what to expect.
A climate model is in someways similar to a weather model; but you can think of it also as a way to see how the probability distribution changes. It’s actually aiming at something quite different from a weather model, even though there are similarities in the way they run. What’s really different are the kinds of boundary conditions applied, but that’s getting a bit detailed.
Suffice to say… the temperature trend from a climate model is not actually a prediction for the future temperature; since you still have all the usual chaotic variation of weather, which cannot be predicted years in advance. But what you can predict is the change in distributions. This can be both the changes in the mean and in the variance of the distributions of temperatures; but most particularly the mean. The temperature trend from a climate model is actually telling you how much the mean of the weather distribution is likely to shift; which is quite a different thing from predicting the actual weather itself… the day by day specific temperatures. In general, climate models also suggest changes in the envelope or distribution for other variables, like precipitation or humidity.
Now I am sure you are skeptical about all that; and I do not presume to actually persuade you that climate models are useful. But as an actual argument, the analogy from lack of prediction of weather to lack of prediction of climate is invalid; because they aren’t the same, and perturbations don’t diverge in the same way for the weather itself, and the distribution within which weather varies. At least; the presumption that they will diverge in the same way is not actually backed up with any argument other than an analogy, the validity of which must be assumed since it is not actually argued.
I quite agree that we don’t predict weather months or years in advance.
And by the way. Anyone who works with a climate model is bound to be something of an expert in thermodynamics.
Fred, your question is a bit open ended, and simply presumes the points in dispute. I don’t think we know everything, and I don’t think there are any big issues with models conflicting with data. If you have some more specific issue, I may be happy to comment. Make the point as hard as you like, but not as open ended as you like, since any reply has to fit into a fairly small space! Have mercy.
Peer review; that’s an interesting idea about Earth’s rotation. Unfortunately, it’s not enough to know the rotation speed; you also will need to know moment of inertia; and that varies as well, as mass shifts within the Earth. Since the variations in length of a day are in about the ninth significant figure, you would need to know the change in moment to at least that accuracy as well. But there’s one thing you can estimate! The biggest consistent trend in varying rotation speed is actually a gradual slowdown as angular momentum is transferred to the moon. Most of the energy involved is actually dissipated as tidal drag; which is know is about 3 TeraWatts. Since solar input as a bit over 120 PentaWatts; even tiny variations in the solar constant, or albedo, will swamp changes in tidal drag. So in fact, the role of changes in Earth’s rotational speed is quite negligible in the energy budgets that bear upon climate.

Hi old construction worker,
The amplification factor is not something you can measure at a point in time using observations, but a kind of diagnostic of climate sensitivity. Since climate sensitivity cannot be measured directly or inferred from any short term trend, the question of “amplification number for the last 8 years” is ill-formed.
Best estimates of climate sensitivity continue to be somewhere in the range 2 to 4.5 K per CO2 doubling; or (as I prefer to give it in SI units) somewhere in the range 0.5 to 1.2 K/(W/m^2). This is an amplification of somewhere from 1.7 to 3.9
Best estimate is still around about 2.5; but you can see that the uncertainties associated with this are large. The quoted extract you provide is a good statement of the state of play; although 4 is a bit too large for the forcing from doubling of CO2. 3.7 would be better.
You don’t indicate what particular “observations” of Spencer you mean. Names don’t have any weight; and PhDs are dime a dozen. It’s the observations themselves that matter, whatever they are.
Spencer’s most important actual observations are associated with satellite measures. Some of this was taken very seriously indeed a few years ago, when the data seemed to indicate a lack of tropospheric warming. It’s since been recognized by all concerned that Spencer and Christy of UAH made some significant errors in their analysis. Spencer and Christy and a few others continue to argue for a weaker version of their original strong conclusions; but the rest of the field has mostly moved on. Tropospheric warming continues to be actively investigated; and there are right now a couple of new papers coming out dealing with wind shear. This is going to help narrow the uncertainties, and give even more strong support to the reality of tropospheric warming.
As well as observations, Spencer also has his own opinions. They are rather idiosyncratic and have pretty much no impact on other scientists; just as the overwhelming contrary opinion of almost all his professional peers has pretty much no impact on Spencer. Which is fine; it’s healthy to have a couple of mavericks doing real work in the relevant science.

The bottom line of all serious discussions on IPCC models is that there is a lot of stuff to discuss.
I.e. the science is not settled.
In my humble opinion it behooves also all ethical climatologists to come out and say this loud and clear, and stop politicians from misusing their research and plunging the world and the already shaky economy into chaos.
They should say : “I believe I am right and stand fully behind my analysis of the data, but I accept that more research is needed on all fronts”.
Duae Quartunciae (11:15:54) :
“At least; the presumption that they will diverge in the same way is not actually backed up with any argument other than an analogy, the validity of which must be assumed since it is not actually argued.”
Both solve the same differential equations. From what I have read climate models use 20 minute steps in boxes 1kmby150ksby150km. Is it the boxes that are different between meteorology and climate?
20 minute steps to go up to a century are way out of any first order approximation in a perturbative expansion: all those mean values set in the boxes are really that, a first order approximation. Nobody who has worked with coupled differential equations can believe that a first order approximation can be valid after thousands of steps.
Tsonis et al, I do not have the reference at hand but it is in a discussion at CA, studied with neural nets the chaotic system of coupled ocean and air currents. It is interesting that their results also predict a leveling off of the temperatures.
Real life is not a first order approximation.
Observation:
On feedbacks and forcings instead of normal thermodynamics: I feel as if I am being forced to predict an eclipse using the epicycles system. It can be done, but should one do it?

Anna, I think you continue to miss the point. Sure, climate models and weather models use the same equations and the same physics.
The difference is that weather prediction is trying to identify a time series, and climate prediction is trying to identify the envelope within which time series are constrained. In both climate and weather models, it is normal to repeat a simulation with small initial perturbations, which in both cases will diverge chaotically. The difference is in what you do with the multiple time series that you obtain.
With weather simulations, you see how far into the future the divergence starts to become significant. You then can have some confidence in using the time series up to that point as a projection of the weather.
With climate simulations, you take all your many diverging time series together, and look at the distribution for those series. Your projection is not a single time series (weather) but is rather the envelope within which the many runs are contained (climate).
The envelope does not diverge chaotically in the same way as weather, but tends to be a well defined and repeatable projection, giving a trend and a natural variance for your weather variables.
That simplifies things somewhat; and there is still plenty of room to improve the “skill” of climate models; and weather models, for that matter. But failing to grasp this distinction is at the heart of invalid objections to climate models based on confused analogies with weather prediction.

I’m not trying to describe “perturbation expansion” at all. What I am actually describing remains as stated above. It’s simply a description of the difference between what people are trying to predict when they model weather, and what they are trying to predict when they model climate. Failing to understand the difference is the source of all kinds of completely of the wall “criticisms” of modern climatology, that merely miss the whole point.
The “video game” method, as you call it, is not being used here as a way to obtain a solution to the equations at all (and hence it is doing something rather different from “perturbation expansion”; they are apples and oranges). It is being used as a way to find a range within which various solutions will lie. The more usual term for this sort of technique is a “monte carlo” method, and it is a perfectly legitimate technique with wide general applicability. It’s computationally very expensive; but for some problems it’s the way to go.
Your various analogies are still not making any direct connection with the problem at issue, except as a convenient way to express what you think climate models must be like, and how you believe the solutions of the equations on which they are based must behave.
You are effectively asserting that the trends obtained by climate models are unreliable because of higher order effects. I am effectively asserting that the higher order effects actually don’t happen to upset the trends. I am content to let our two different perspectives on this stand side by side, for readers of this blogs to see as two incompatible views.
Fortunately, those interested are also able to check out the matter further, with a bit of work, by looking at the actual matter of climate models; rather than drawing conclusions on the basis of looking at particle physics. In brief; this has become merely silly. I’m sure you think the same.
On the other hand, on a positive note. Your mention of Tsonis is a genuinely very interesting and relevant. I suspect your reference might have been Tsonis A.A. et al (2007) A new dynamical mechanism for major climate shifts., Geophysical Research Letters, Vol. 34, L13705, doi:10.1029/2007GL030288. (Or possibly related papers by the same team.) I thank you! It looks really interesting.

That is interesting. And it is certainly true that when used for integration Monte Carlo methods are another way of getting the one solution to a definite integral.
But the implication that this is the only way in which Monte Carlo methods are used, or that every other use is merely borrowing terminology from your own preferred application, is going to get you tied up in knots. Your professional experience will give you some very useful insights in your specific areas of expertise. But if you can never see any other field or application except as a some kind of shadow of your own, then your expertise can easily become a handicap.
And the truth is… both the applications we describe are borrowing terminology from gambling! 🙂
Anyhow — Good night!