MSimon wrote:tomclarke,
Just so you don't feel neglected given your lovely equations. And assuming that my previous points are of negligible value (look up Svensmark cosmic rays) please give an equation that explains zero warming (as if you can model it "globally" with your T^4 black body with temperature ranges from -40 C or lower to +30C at the same point in time) for the last 14 years.
That is not the usual "statistically" but straight line zero.
So, as I understand from your clarification what you mean is that in a temp signal that is very noisy, the value now is same as 14 years ago. And you think that is unlikely given a positive linear trend with (lots of) noise superimposed?
I thought you were a better engineer than that but if you want the math I'll give it to you. Let me know. Somehow I think you are capable of doing it yourself.
Why do 95% of the models not "predict" (and that is another little question - how many model runs do the "predictions" represent? - what runs were not included - if any?) zero for that length of time? Why do the runs we see run hot compared to the "real" "global" temperature - what ever that is?
The last 15 years have been running
cooler than trend, just as the previous 10 years were running
hotter than trend.
Why do models not track these decadal changes? It is a good question and one that requires an understanding of what the models can and can't do:
(1) they can't predict chaotic external forcings (solar, ENSO, volcano). In this case solar and ENSO are all negative ove rthe period.
(2) They can't predict long-term changes in ocean currents that affect mixing. OK maybe they should od this, but as of now they do not do it, though some recent work is incorporating this and I guess in 5 years time they will.
Given noise you'd expect that 5% of time the actual temp would be such that only 5% of models track it.
Where do you get your 95% figure from? I hope not from Roy Spencer's biassed and deceitful graph?
[img ]
http://wattsupwiththat.files.wordpress. ... s-obs1.jpg[ /img]
< This image screws up page formatting - please find one that doesn't. 1,000 pixels wide is max -
the moderator>
Apologies for that - now fixed I hope.
We can go through why that is wrongly baselined if you like, and why he is a liar (he surely must know better...).
And given the error bars in the data (and you know a small error in T becomes 4X as great re: radiation [for small changes - roughly]) how is any "prediction" less than about 3C in changes significant? So errors in data alone (not to mention models) is multiplied by 4. Well that is going to throw any model seriously out of whack. It can't balance.
I don't understand this as a criticism of the models. We'd need more care. However I think your more general point is that the temperature record is so flakey, and the unmodelled noise so large, that models cannot be validated solely from temperature predictions.
I agree. Where do you get the idea anyone thinks they are? Or that temperature predictions are the only way to validate the skill of models?
Well let us look further. Non-equilibrium thermodynamics. Which is in fact the current state of the earth and every state previous. There is no balance. And then the flows. Navier-Stokes. Dependent on initial conditions. Very dependent on initial conditions. Now let us leave the problems with T and try to figure out what the surface roughness and variable flows - locally and on larger area scales?
Now you can show your pretty numbers (and I assume they are correct - I haven't checked) but your model is a toy block compared even to a toy fire truck which is in no significant way like a real fire truck racing to a fire.
As far as the laws of mathematics refer to reality, they are not certain; and as far as they are certain, they do not refer to reality. Albert Einstein
You see reality is not simple. The black body is not of a uniform temperature and it is of course not black.
So let me say that your fiction will behave exactly as you predict. Now what about reality? It is complicated.
I refer you to above. Of course my T^4 model was simplistic. But I was replying to an even more simplistic claim and as an OOM estimate my comparison does well. Well enough to show that the 1W/m^2 does not lead to a substantial temp increase.
Actually your newer (and fictional) graph validates this. The claim is that we are entering a novel phase of solar radiation at -6W/m^2 from current. The blue line (entirely fictional) looks alarming but the real solar radiation has never yet tracked the alarming bit of it. I'm quite prepared to wait and see whether this atypical TSI trace comes to pass. I have no idea. You have no idea. But it looks so different from the past that it must be an outside chance.
About the best we can say is that in general night is cooler than day and summer is warmer than winter. Generally. Something warmed the "planet" for a while. And for some reason that something appears to have stopped. For a while. And generally the models have not predicted the stop.
That may be the best you can say but I notice some lack of attention to details. It is possible that careful attention to all of the evidence constrains things a bit better than that. But you will not know whether that is the case if you give up at first base.
And that brings us to models. Why so many? In settled science you only need one. F=ma. Of course that is only true in the simplest cases. Air resistance. And other types of friction and stiction will complicate things. Reality diverges.
Now we get to the real point. Climate science is not pretty and nice as the subjects you (and I) know well. It has large uncertainties. It also has many different lines of evidence all of which constrain hypotheses. Its a big mess. A bit like medicine. Not my type of stuff. But I don't therefore think it is impossible to work with.
It would be fair for you to admit incompetence. But not to assume no good work can be done just because it is not your sort of work.
Close enough for engineering work? E=IR is only good in limited cases. If you put 10 W into a 100mW rated resistor for long enough reality will diverge significantly. Once you let the smoke out the resistor no longer works the way it used to.
So yes. Very good. Nice model. And quite useful in some VERY LIMITED circumstances.
I think this is just rhetoric, which I understand, but don't agree with unless its scope is suitably limited.
Now explain the Maunder and Dalton minimums. There is currently no accepted explanation. Which is rather a big hole in the theories. Which says they can't be trusted over long time scales. How long is long? Well for weather you recalibrate every six hours. To keep the divergences small enough to be kinda useful. Climate? Which is the average of weather. With recalibrations ever 5 years or so? Summer will be hotter than winter. Generally. If Krakatoa doesn't blow. And you live some where other than near the equator.
The past temperature record cannot be 100% accurately modelled. Why should it be? You know well there are many uncertainties and noise sources. You also know that global temperature record is not well constrained over those periods so the evidence is weak, but I will if you like give you a decent explanation of the MWP/LIA temperature change that also validates ECS around 3C. Quality of evidence - fairly weak but not negligible.
Why should the uncertainty of past temperatures and over many different forcings prevent modelling of the effect of CO2 where the forcing is tightly constrained? We need to know only what is the forcing, and what are the feedbacks. All the other noise on top of the CO2 induced change is certainly unfortunate if you want to do weather forecasting, but not if you just want to no how much does CO2 at a given level perturb temperature from its holocene interglacial value.
Working this out does not necessarily require correlation of CO2 forcing with temperature. It can be done with any known forcing.
So, as IPCC make clear, it is is flakey, the possible feedbacks are badly constrained and so therefore is ECS. But it is not impossible. Further, the uncertainty does not mean we should assume CO2 has no effect. It means we should just not know what is the effect. except within quite large limits. There is an implicit assumption from many people that uncertainty => CO2 has zero effect which is contrary to known facts (CO2 forcing, etc).
The CEP of the models is no where close enough to justify destroying civilization. Such as it is.
There are two separate issues here: (1) you think GCM models clear enough to determine ECS. (2) You reckon is ECS is as IPCC claim (in range 1.5 - 4.5 C/doubling) civilisation will be destroyed. I can't agree with either but for purpose of clarity can I suggest that we leave out all the emotive civilisation destroyed stuff (from ecological or economical alarmists respectively) and look dispassionately at what we think are the constraints on ECS?
You have not said what are your 10% bounds constraints on ECS? Come on, tell me? Because if you can't put your own bounds on it you should not criticise others. If you really think there are no bounds, then you should reckon any major change in GHG content is possibly a big deal and operate on precautionary principle because the costs of GHG reduction are much much less than the costs of adjustment to a significantly higher temp world.
As I've said above the strongest evidence for ECS does not come from GCM climate models, nor does it depend on them, though they are no doubt wondrous things.
And I have not even covered sampling. Or gridding. Or parameterization.
No you have not, but climate scientists have looked at all these issues.
Do you know what decent climate scientists say is the strongest evidence for ECS around 3?
Not GCMs (with all those issues you have raised as bug bears, but I bet will refuse to quantify). It is simple models evaluating feedbacks by comparing temperature change with known non-CO2 forcings. Over several different timescales. If you know overall feedbacks you can calculate temp change from CO2 forcing.
Your position is fundamentally inconsistent I believe. Given evidence (CO2 doubling, known CO2 focrcing, known water vapour and lapse rate positive feedback) leading to likely significant temp change from CO2 increase, but lots of uncertainty, you claim:
(1) everything is uncertain
(2) you know the effect of CO2 is so small that a rapid not known for 1M years change in CO2 poses little risk.
I'd respect one or other statement as a possible synthesis of the evidence (though probably wrong). But both together are just inconsistent and deserve little respect.
