[Martin Taylor 2010.05.30.22.12]
[Martin Lewitt 28 May 2010 0534 MDT]
I saw a presentation at the International Conference on Climate Change describing a proposed governor on the climate system to explain the long term stability of the earth's temperature. ...
The question is whether there are control systems involved. Rick suggested that the stability is more of the nature of the stability of a spring, and suggested ways of looking to see whether control was involved, using "The Test" for the controlled variable. I have no real problem with what Rick suggests, but "The Test" as he describes it necessarily starts with a guess about what variable might be controlled. In a system with possibly thousands of controlled variables (or none), it would be very hard to use "The Test" as Rick describes it.
One of the characteristics of a system such as a mass hanging from a spring, which resists attempts to disturb its height is that the energy of restoration is derived directly from the disturbance. One of the characteristics of control is that the energy of restoration is derived quite independently of the disturbance. In the case of climate, what is a disturbance? Does its effect on your variable of interest cause less change than you had reason to expect? Is the reduction of effect due to some influence whose energy source is different from the immediate effect of the disturbance?
In the case of the climate, these questions are difficult, because there are only two energy sources on the earth, the Sun and nuclear fission (one day we may add nuclear fusion, but not yet). The energy available from nuclear fission is negligible, so we need concern ourselves only with the energy from the Sun. The energy from fossil fuel derived from the Sun's radiation hundreds of millions of years ago, but it is nevertheless derived from the Sun. So, since all the energy in the possible disturbances and all the energy used to counter them (if any) derives from the Sun, could we say that there is any possibility of control systems existing in the network of feedback processes that influence the climate? I would say "yes, there is that possibility", because of the time differential between the various times when the Sun's energy arrived and was stored. The disturbance of increasing CO2 may well be energized by the Sun, but it is energized by the Sun of hundreds of millions of years ago, and if it is countered by anything, that thing is energized by the Sun of now (or of the more recent past). I think it is legitimate to consider energy derived from the same source at different times as being from two different sources.
Now we ask questions 2 and 4 of Rick's list taken from Phil Runkel's book:
2.Predict effect of disturbance if the variable is not under control
4.Measure the actual effects of the disturbances.
In the case of CO2 disturbance, it seems that the effect on global temperature is appreciably less than would be predicted, so there is at least the possibility that there is control. But there is also the possibility of spring-like resistance, using negative feedback without control.
This leads to another distinctive property of control, as opposed to a simple negative feedback loop: asymmetry.
A control loop has a distinctive asymmetry, in that the power supplied to the input side of the control loop is less than the power supplied by the output, usually by many order of magnitude. A generalized negative feedback loop has no such asymmetry, but it can stabilize the values of several variables within the loop. Only a control loop will specifically stabilize one specific value within it, namely what we call the perceptual value. The specific stabilization of this one variable is due to the asymmetry.
Usually the perceptual value is a function of some external variables, so that specific function of those variables is also stabilized, which means that if the perceptual value is a function of only one external variable, that external variable is as closely stabilized by control as is the perceptual variable.
So, in connection with the climate, we have to ask whether the reduced effect of the CO2 disturbance as compared with the naively calculated effect is due to simple negative feedback such as "increased temperature leads to increased evaporation from the oceans leads to more cloud cover leads to less sun energy reaching the gound" or to control. If control, as Rick points out, one has to determine what is sensed and stabilized. My own inclination (purely intuitive and with no foundation in data) is to think that the reason temperature is not increasing as fast as would be expected simply from the physically computable greenhouse effect of increased CO2 is simple negative feedback loops, not control, though I have no doubt that the side-effects of individual control loops have a very strong influence on the climate, both by augmenting the disturbance and in mitigating its effects.
It is those control loops contributing to the disturbance that must be altered if the world is to avoid the fast approaching disaster. And that is a political problem, not a technical one, unless we can develop the technology of PCT to the point where it can influence politics.
Martin
···
On 2010/05/28 7:34 AM, Martin Lewitt wrote: