Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

The positive feedback factor of CO2

  1. Mar 23, 2007 #1
    In the other thread we / errm I discussed that theories cannot be proven but they can be falsified.

    Actually, without positive feedback, more CO2 is merely a mild friendly climate alleviator at the most

    So the issue here is the positive feedback factor of CO2 on the termination of ice ages. It is exhaustively discussed that in the ice cores we see the oxygen isotope (d18O) spikes rise before the CO2 does (d18O is assumed to be temperatures). As the CO2 is lagging 600+/- 400 years, normally this would refute the idea that CO2 causes temperature rise, instead temperature rise seems to rise the CO2.

    But there is always a remedy, positive feedback. So some trigger (earth -milankovitch wobbles are assumed to trigger a first faint warming, which induces CO2 increase with some delay, probably because of warming oceans and as soon as CO2 rises, positive feedback kicks in, increasing the warming, which increases the CO2 etc etc a strong positive feedback loop. Case solved.

    However, if that is a hypothesis, where are the studies that test it? I haven't seen one. Why are the warmers not pointing at any scientific substantiation when talking about positive feedback?

    because it's not true?

    So I used this high resolution graph of Antarctica's EPICA Dome C ice cores during the last glacial temination between 20,000 and 10,000 years ago, to demonstrate that there is no positive feedback because the typical behavior of positive feedback is not seen.

    Data here: Stenni et al 2001 for the d18O and monnin et al 2004 for the CO2

    http://home.wanadoo.nl/bijkerk/epica5.GIF [Broken]

    more to follow

    Attached Files:

    Last edited by a moderator: May 2, 2017
  2. jcsd
  3. Mar 24, 2007 #2
    Is there Correlation of Planetary Temperature & CO2 Levels!!!

    Certainty or Uncertainty Concerning whether CO2 Climatic Forcing is Real?:

    http://www.geosc.psu.edu/~kump/KumpCommentary.pdf [Broken]

    The following is an excerpt from the above linked paper that notes there have been periods in the Earth's history when the planetary temperature has been high and CO2 levels have been low and periods when the planetary temperature has been low when CO2 levels have been high. The paper is curiously titled "Reducing the Uncertainty of CO2 level and Climatic Change".

    I have not heard any discussion in the press concerning the lack of correlation between CO2 levels and planetary temperature, in the geological past. The discussion has been centered on the 20th century where there has been rising global temperatures and rising CO2 levels.

    Also I have not heard any mention in the press that the sun is at its highest activity level in 8000 years or any mention of the specific mechanism that high solar activity can cause the planetary temperature to be high.

    As I have noted in this forum (with links to the relevant papers) there is satellite data and data from observation of the earthshine on the moon, to support the assertion that the majority of the 20th century warming is due to the extraordinary 20th century increased in solar activity, which reduces low level planetary cloud cover and hence causes the planet to warm, not due to an increase in CO2 levels. Also as noted, the 20th century period of very, very, high solar activity appears to be ending. Based on proxy climatic data, the sun will now start a very low activity stage and the planet will severely cool. i.e. This is a recurring phenomenon, not a weird event, which has never happened before.

    In my opinion, the true problem (i.e. Why we are not talking about climate change, global cooling.) appears to not be not scientific, but rather political. The recent very strong support to eliminate global warming is an indirect method to protect the environment and reduce obvious egregious consumption.

    Climate change - global cooling not warming - is based on the facts, a very, very important problem which the majority appears to not be aware of.
    Last edited by a moderator: May 2, 2017
  4. Mar 24, 2007 #3
    Thank you for your contribution, William, I concur in general.

    On the topic of strong positive feedback of CO2 during the glacial transitions, I made a very rudimentary basic excel model of a strong positive feedback here

    http://home.wanadoo.nl/bijkerk/positive-feedback.xls [Broken]

    So let's have a look at the results of one run (attached).

    So, the thin black line is simulating the natural forcing or the natural climate variability down the ages. I shaped it with an initial increase followed by a retreat and then resuming the increase, all with a random riple to let it look more natural. Then we add the CO2 reacting linear with the temperatue on a certain delay. This value is added to the black line to get the thin blue line as the total forcing of natural + positive CO2 feedback. The fat red line is the assumed temperature response on the total forcing, closing the positive feedback loop.

    As I mentioned, positive feedback systems are bistable, all or nothing, therefore I limited temperature response to an artificial -stable- value of 15 plus random ripple.

    Now you see the essential difference between the ice core proxies and the positive feedback model, as soon as CO2 kicks in, the temperature response accelerates and continues all the way to the high stable position, seemlingly rather independent of the natural climate variability, the downward natural trend only delays the accelleration of increase in the output somewhat and that's the big difference, the reason why we don't see a strong positive feedback in the ice cores, to explain the alleged strong sensitivity of climate to changes in CO2

    Of course you can get response to natural variability back by downplaying the positive feedback factor, but that would also downplay the role of CO2 in climate with the same amount.

    I'd appreciate possible comments of engineers on the system response issues.

    Attached Files:

    Last edited by a moderator: May 2, 2017
  5. Mar 24, 2007 #4
    In the run here, the positive feedback is tuned down 35%. We see the return to response to natural variation, but with a considerable lag. The response of the "temperature" is counteracted by the positive feedback of the CO2, which we dont see in the original reaction of the proxies in the opening post where a sharp trend change is showing against the rising CO2.

    So when that CO2 was unable to counteract that as seen in this run, it simply shows that there was no positive feedback of any significance.

    The alleged temperatures fluctuated wildly but CO2 did neither cause it nor enforced it with positive feedback.

    And that falsifies anthropogenic global warming as in CO2 causing a strong greenhouse effect.

    Attached Files:

    Last edited: Mar 24, 2007
  6. Mar 25, 2007 #5
    The ice core records show co2 increasing by about 50% during interglacial warmings, therefore assuming a 3C increase per doubling the co2 amplification should be about 1.5C in all. Ie the amplification should be less than 20% of the total forcing. In the examples above the amplification is a lot higher than that. The above really falsifies anthropogenic global warming causing a greenhouse effect stronger than anyone proposes.
  7. Mar 26, 2007 #6
    If we go with the current scholar views, Ice core proxies are really comparing a global CO2 signal with a local "temperature" 10-15 degrees variation or is it really?


    Anyway, many other proxies elsewhere (sea surface temperatures, pollen etc) give some range of 0-5 degrees temp changes between cold stadials and interglacials. I'll flood you with refs if so desired, but I'm pressed at the moment. This would have to increase the positive feedback range relative to the other leading variation factors, the higher this factor, the stronger the typical positive feedback characteristics, which as said could only be a very weak factor when considering the typical system response.

    Personally I think that we are looking at something completely different but that's another matter.
  8. Mar 26, 2007 #7


    User Avatar

    Please could you summarise...

    What is meant by, and what is the nature, of this "positive feedback"?
  9. Mar 26, 2007 #8
    I'll try, I'm afraid conveying complicated procedures in a simple way is not one of my best talents.

    So we have the Pleistocene ice ages, most notably since the last million years. It is thought that several long cold glacial periods have been interrupted by shorter interglacials in irregular cycles each about 100,000 years in duration, some 80-90,000 years for the glacials and some 20-10,000 for the mild periods in between.

    This information is inferred from many geologic proxies. These are chronologic sequences of data on which conclusions are inferred. For instance the deep ice cores of Antarctica can be analysed on CO2 content and water isotope ratios. And after lots of complicated processing these kind of graphs are produced:


    It's clearly visible that the isotope 'temperature' and the 'atmospheric' CO2 are very closely related and here you're looking at the pre-Hockeystick convincing evidence that CO2 causes global warming.

    There was a nagging problem however, air in the air bubbles is younger than the surrounding ice because the fresh snowy ice remains open for the air to circulate in it. Trapping of the air happens some 80 meters down the ice due to the increasing pressure, but that's trapping of air of today. Now there have been libraries written how to tackle this problem and it's still ongoing but meanwhile it became clear that the CO2 spikes were in fact lagging the isotope (temp?) spikes. Therefore, the ice cores dissapeared out of the limelight of global warming, instead the hockeystick was created.

    So if temperature rises first at the end of the glacial period, apparantly it causes the CO2 to increase. A possible scenario would be warming of the oceans, warmer oceans can hold less CO2.

    But if you advocate that more CO2 causes increased warming and more warming causes increased CO2 then you have a positive feedback system, the both elements stimulating each other to much higher levels than without positive feedback and this is where the explanation of the global warmers stops, without any closer study.

    Now positive feedback systems are very well studied, your computer contains billiards of them, each memory bit is a positive feedback system, knowing only two stable conditions at the extreme ends, on or off, 1 or 0.

    So if we know how positive feedback works then we can observe the proxy data and see if it behaves as that kind of system response, knowing how output and input of systems relate. However I have never seen anybody considering that in geologic proxies.

    Now it has been 30+ years ago when I struggled with system response on the academy with Laplace, Fourier, S-domains and what not, but I still recognise from the high resultion ice core data that there cannot be a positive feedback at work there, resuming:

    1: As the CO2 starts to rise 600 years after the isotopes at about 17,000 years ago, in a positive feedback, there is an enhanced of whatever cause of the warming, hence the warming should accelerate. It does not, so an enhanced working of CO2 greenhouse effect cannot be seen.

    2: There is a strong abrupt reaction to natural variability halfway, which is impossible with strong positive feedback, where the feedback takes over the (natural) system input and steers it all the way directly to the system limit, like the status "one" in the computer memory bits.

    With the much weaker positive feedback in the last XLS run we still see a considerable delay and a smooth leveling off transition of the signal output, as the CO2 fights against the natural input, wanting to continue up where the variability pulls it down. We see it clearly in the rudimentary simulation but this effect is completely lacking in the ice cores. This is possibly the strongest clue that there is very little feedback going on if any.

    So, we can safely conclude that we see no strong positive feedback. Instead the CO2 curve just show passive response to temperature

    So if the rising CO2 was unable to boost the "temperatures" during the glacial transitions, why would we think it does it now?

    Now, if we dare to think Popperian fasification philosophy, we cannot but conclude that CO2 cannot be a strong climate factor, not then, hence not now.

    I hope it helps.
    Last edited: Mar 26, 2007
  10. Mar 28, 2007 #9
    Perhaps co2 could be a strong climate factor today simply because there hasn't been much warming. Perhaps important is a better word than strong though in this case. Certainly co2 could be a weak climate factor on geological scales but recent warming is <1C so there doesn't seem to be a contradiction.
  11. May 27, 2007 #10
    Well finally I got the critique I was looking for, allbeit in another thread, burried under a deluding deluge of fallacies,

    So you are qualified to disqualify me? What difference would that make to reality. Are only the taylors qualified to judge the new clothes of the emperor? Isn't the little kid not qualified to exclaim that the emperor wears no clothes?

    Where is the analysis that would lead to such a conclusion? How can you be biased with numbers?

    Why would that be erroneous? What is the problem?

    All it shows is that the process in the ice core is not showing positive feedback behavior.

    Now I'm completely lost. Reasoning has nothing to do with ad hominems. Ad hominems are red herrings about non relevant information pertaining a person, intended to degrade the credibility. I've looked twice, there is no irrelevant information about persons in this thread, although the quoted post here appears to be very revealing about it's author.

    I think it would be required to substantiate your claims with objective reasoning or be exposed for what these claims really signify.

    Please show me where I do that. I'm no aware of ever accusing governments of conspiracy, victim of a positive feedback lfear mongering loop yes, but not conspiracy. So the allegation is another strawman.

    That's the fallacy of the restricted choice or the ignorance fallacy, either A or B or C and the ignorance fallacy: it can't be A or B therefore it must be C. I don't know of anything else.

    But again, debunking theories is not about offering competing ideas, it's about showing that the idea is false, which is still standing here, without substantiated claims showing otherwise

    Anyway: peer reviewed, no CO2:

    The hyperbole. There is no such thing as an experiment. The variation in CO2 is nowhere near the maximums of the paleo climate of the past, even when there were ice ages in the distant past, CO2 has been higher.
    Last edited: May 27, 2007
  12. May 27, 2007 #11
    Let us begin here the study I believe you are looking for in the OP.

    http://icebubbles.ucsd.edu/Publications/CaillonTermIII.pdf [Broken]

    Nothing you have provided me here has convinced me that their conclusions are in error.
    Last edited by a moderator: May 2, 2017
  13. May 27, 2007 #12
    Nothing in the world will ever convince you, not even a new ice age, that's for sure. However, you just show the first of a series of white swans, which lead to the all-swans-are-white hypothesis which I killed with a black swan.

    I asked you to substantiate your claims and demonstrate why I did not kill the positive feedback hypothesis, not to produce white swans. Have a look at the 4AR WG1 chap 6, lots of white swans. I know them. In fact Caillon et al is my first reference in demonstrating how the positive feedback hypothesis was develloped.
  14. May 27, 2007 #13
    Your interpretation of what the data should look like is not a black swan, disproving the white swan theory. Actually I found your use of the white swan theory as an analogy to be a stretch.

    You are choosing an area where their is still a lot of unknowns and attempting to twist the unknowns it into a proof of the falsification of a widely accepted theory.

    I see that no engineers commented on your positive feedback model.

    I don't know much about it but from a layman's perspective you are saying that the feedback response is polarized + -, 0 1, on off, etc. Why should I assume that feedback is as you say it should be?

    Why should I assume that because you can demonstrate exceptions that the rule is invalid?

    Is there more going on than we know? Absolutely. But as far as the CO2 amplifying and enhancing the warming during the interglacials, it is the only explanation, at present that explains it. The science behind the theory is sound. If your little positive feedback model were the black swan that debunked the IPCC, I think it would get more attention than a thread on the Earth Forum on PF.
    Last edited by a moderator: May 27, 2007
  15. May 28, 2007 #14

    No there is a claim that temperature differences between interglacials and ice ages are boosted by CO2 positive feedback. It's repeated about seven times in the last chapter 6 of WG1 part of the 4AR. That claim proofs to be false by the lack of persistance typical for positive feedback, especially with such a lag.

    Curiously indeed. What can you make of that? There are a few engineers around here renowned for never missing an opportunity for beating up crackpots. Hmmm, why didn't that happen?

    Now is this from the same person who said:

    So my guess is that the post where that came from, is a copy-paste of a reply on a email requesting to crunch this attack. The offensive maverick style is typically Gavin. Could also be Mike.

    I think I tried to explain that feedback is either positive, negative or zero with large differences in system response. By the way, I should explain that over three decades ago on the Mil Acadamy my second study choice was system response and feedback, signal processing. Level about equivalent to Bsc, but that's a long time ago. Halfway that period I needed some refreshment for a test pilot course. I admit to be rusty on the calculus but I have some proficiency in building simulation models (not prediction models). That's actually surprisingly easy considering the head acheing complex math.

    The essential part is that it demonstrates that the rising CO2 did not provide warming feedback. Now, skip the last unneccesary word: It demonstrates that the rising CO2 did not provide warming. Actually more precise: It demonstrates that increasing CO2 values had no effects on the Antarctic isotopes.

    If it didn't do it then, why should it do it now. If water didn't boil at 200F in the ice ages, why would it do so now? Physical rules have no exceptions otherwise it wouldn't be a physical rule.

    No it does not explain anything, as soon as you bring the different specialities together, what you get is a mess.

    if you study Thomas Kuhn you would understand why that is not so:

    Last edited: May 28, 2007
  16. May 29, 2007 #15


    User Avatar
    Gold Member

    I'm an engineer, electrical. I'm not sure I follow the definitions of the system or the forcing function you are guys are trying to model. I'll just state that to the degree that pieces of the system can be described as linear and time-invariant then in that region any introduction of positive feedback will introduce an exponential term something like
    [tex]e^{at}, a>0[/tex]
    [tex]sin(\omega t) e^{at}[/tex]
    into the system output depending on the damping and bandwidth of the original system and the positive feedback (sub) system. If the system is non-linear, well, most anything can happen in non-linear as long as energy is conserved but as an engineer Id just keep looking for some trick that would allow me to approximate the system as linear.
    Last edited: May 29, 2007
  17. May 29, 2007 #16
    I found your interpretation biased because you are using your own models reaction to bias your conclusion. I find it amateurish because you are basing your conclusion on a small sample of data. CO2 and 18O, relating one to the other as if they existed in a vacuum and not as part of a complex climate environment.

    14,000 years ago when we see a spike in CO2 and a drop in 18O.

    What else was happening 14,000 years ago that might have had an impact on climate aside from CO2?

    Whether or not the mwp was triggered by Antarctica or Greenland the effects of such events are important factors when deciphering and interpreting the paleo record. Not only does all the fresh water in the oceans change circulation, but it also alters the isotopic ratio of the oceans.

    Your interpretation does not consider these factors.

    Your simple model does not contain a carbon cycle.

    Climate models used by climate scientists that do not contain a carbon cycle treat CO2 as a forcing, and not a feedback.

    The fact that the real world does not match the expectations of your model, IMO is more a statement about your model than the positive feedback of CO2.

    The scientists on the NERC thread seemed to have trouble understanding your question about CO2. My take was they couldn't believe you would be questioning something so fundamentally well understood and universally accepted. Without CO2 the earth is a snowball. The effect is well and thoroughly understood, measured, and quantified.

    If you want to be have your ideas considered seriously you need to explain the warming without CO2. Telling scientists that they are making fundamental mistakes without a plausible counter explanation for what is observed is not going to convince anyone.

    My take is you just want to become famous by exposing a false scientific assumption. I am just speculating here based on the subjects of most of your threads. You have threads and posts about a conspiracy to supress CO2 measurements, accusing scientists of not knowing BC from BP, challenging the interpretation isotope proxies, 10,000 year old cities buried under glaciers, to challenging the basic science underlying the greenhouse effect.

    Good luck on your search for your black swan. but I don't think you have found one here.
    Last edited by a moderator: Apr 22, 2017
  18. May 29, 2007 #17


    User Avatar
    Gold Member

    Eh? I thought it was well known that most of the delta between a pure black body earth and current temps was due to water vapor? Is it all CO2?
  19. May 30, 2007 #18
    You have very peculiar ideas. There is something as curiousity.

    About the feedback model. You must realize that it's merely a simulation of a standard positive feedback process, any process. regardless if it is nature, electricity, mechanical, the same principles apply. once more: One of the principles of a strong positive feedback mechanism, as required here is the bi-stability in the extreme values, either high or low, on or off. Nothin in between. Once the system departs from one stable extreme, the positive feedback becomes increasingly the overwhelming primary steering factor, regardless if it is mechanical, electric or natural. This means that halfway the system is highly insensitive to natural inputs. So, whatever happens halfway should be irrelevant, the system should barely react to that. But yet it does. it reverses trend for some reason, despite the overwhelming primary controlling positive CO2 feedback. Which reason is irrelevant, whatever happens.

    The point is that this reason must be strong, much much stronger than the positive feedback, successfully shutting down the feedback effects for some 800 years. That doesn't really make sense. Alternately, it is simply showing that there is no strong positive feedback. That's it, regardless what the complexity of the systems.

    I'll get back to meltwater pulse 1A which you describe in the other thread
  20. Mar 19, 2008 #19
    I think the main reason this is believed, is that the temperature difference between glacial minima and maxima can't be explained with just the effects of milankovich cycles+ice albedo+increased water vapour.

    To make this calculation you would take data from periods where the temperature was stable and not from periods with rapid warming where production of meltwater, ice albedo changes, changes in sea currents etc. could easily produce variations in the temperature that have nothing to do with co2
  21. Mar 19, 2008 #20
    This is simply not true. If the positive feedback is small enough it will result in amplification of the forcing.
  22. Mar 20, 2008 #21


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    This is correct. Positive feedback can do anything.

    If you have a system with an "open loop" amplification A, which means that

    (out) = A (in-raw)

    then positive feedback means that one adds part of (out) to the input:

    (in-raw) = (in) + b x (out)

    where b gives you the strength of the feedback.

    (out) = A ( (in) + b x (out) )

    You can solve this for the output:

    (out) = A / (1 - A x b) (in)

    So it seems that the system WITH feedback has an overall amplification of:

    A_overall = A / (1 - A x b)

    For negative feedback, the sign of b is reversed, and we see that for NEGATIVE feedback, we have that the overall amplification is a fraction of the "open loop" amplification. In fact, if the open loop amplification A is big enough, then A_overall ~ -1/b and our amplifier is now purely determined by the negative feedback, something which electronicians like a lot.

    For positive feedback, it depends on whether A x b is bigger than 1 or not. If A x b is smaller than 1, then we've just INCREASED the overall amplification. The thing acts as a stronger amplifier than in "open loop". If A x b comes close to 1, we get HUGE amplifications, and when it crosses 1, we get an instable behaviour.
  23. Mar 20, 2008 #22
    Note that I said:

    For a bi-stable flip flop mechanism all you need is the total gain (both process and feedback) in the closed loop to be >1.
  24. Mar 20, 2008 #23


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Yes, that's correct. But the loop gain (my b x A) might be smaller than 1 ; in that case you simply get amplification. And if you allow for a frequency-dependent loop gain, then you can do whatever you want.

    But even in the static case, say, if the "amplifier" open loop has a gain of say, 1 degree for 1000 ppm (just an example), then for an increase by 300 ppm you'd have an increase of 0.3 degree in open loop. Now, imagine that there is a feedback mechanism that gives you a b-factor of, say, 900 ppm per degree. Well, then the loop gain is b x A = 900 x 1/1000 = 0.9, and our closed loop amplification now becomes:

    (1/1000 degree per ppm) / (1 - 0.9) = 1/100 degree per ppm.

    So now an initial increase of 300 ppm will give you an output of 3 degrees, 10 times more.

    But there is no bistability in this model.

    I'm not saying that this is so, but I'm saying that it is not mathematically impossible.

    My personal objection would be that it needs "fine tuning": the feedback mechanism (the b) must be such that it is just a little bit below the inverse of the amplification, and if these are two totally distinct phenomena, there is a priori no reason why these numbers should be close. It doesn't exclude the possibility, but it would be strange.

    It would be strange that some atmospheric transmission gives us 1/1000 degree per ppm, and that a totally different process, oceans or I don't know what, give us 900 ppm / degree.

    Because this is CLOSE to instability, and on the other hand if it were only, say, 90 ppm/degree, the effect would not be noticable (a change by a factor of 1.1 instead of a change by a factor of 10).
  25. Mar 20, 2008 #24
    Note that for the Ice age logic bi-stability is required for the swaps from the glacials to the interglacials. It is believed that when small Earth orbital variations (Milankovitch cycles) amplify each other, they would trigger the swap to the other stable condition only, CO2 being the driving force in combination with ocean temperatures to maintain that condition for a prolongued time. (warm oceans means less capacity for gas absorbtion, means more CO2 in the atmosphere means higher temperatures for interglacials and vice versa for glacial periods).

    This bi-stability was also the conclusion from the enigmatic Dansgaard Oeschger events as can be seen here:


    Both the Cariaco basin (Caraibic sea and the Greenland Ice cores) display behavior what could be explained as a bi stable positive feedback system. Hence my initial introduction that going in and out ice ages is assumed to be a oversteered process, ocasionally swapping from low to high system limits, and that is what is refuted by the Antarctic cold reversal of ~14 K years ago.

    There are more problems with that feedback idea though, not only the one that this thread was about. See also here.
    Last edited: Mar 20, 2008
  26. Mar 20, 2008 #25


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I have been thinking a bit about this, and I see a problem, as a matter of principle.

    First of all, "strong positive feedback" by itself doesn't give rise to bi-stability. It only gives rise to exponential divergence, one way or another. To obtain bi-stability, one needs a non-linear system, of which the "feedback" is only a local linearisation. In fact, to obtain bi-stability, you need a system which has an instable region (with strong positive feedback, in the local linear approximation) AND you need two regions with stability (hence, BI-stability!). In these regions of stability, you CANNOT have strong positive feedback - rather negative feedback.

    So demonstrating that one has bi-stability would NOT be a proof of positive feedback OUTSIDE of the region in between the two former regions of stability.

    So if you have a positive feedback (instability) in between one stable situation (glacial period) and another stable situation (interglacial), which makes one "flip" quickly from one to the other once the external perturbation is large enough to leave the "islands of stability", then this is in no way a proof that one has the same instability on the OUTSIDE of this interval. Even more so: if we are now in an interglacial period, we should rather be stabilized against further heating, because otherwise there is no reason to expect that the heating that occured between the glacial and the interglacial would STOP at the interglacial: if we are truely in a system with general positive feedback, it would continue to diverge.

    That said, it is also possible that ABOVE the interglacial island of stability, there is a NEW region of instability, but this then has nothing to do with the instability in between glacial and interglacial periods, as it is necessarily separated from it by an island of stability.

    To take your electronic flip-flop analogy: between 0.5V and 4.5V (TTL), yes, there's an instability given by the positive feedback. But when reaching 0.5 or 4.5 V, the diminishing gains of the transistors (because they are squeezed off) bring this instability to an end, and even reach a stable operation. If you now artificially bring the output to 4.7V, it will not diverge again, to say, 10V or so, but will return to 4.5V after taking away the stimulus.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Similar Threads for positive feedback factor
Cloud position