Prev  1  2  3  4  5  Next

Comments 26 to 50 out of 121:

26. muoncounter at 02:51 AM on 5 December, 2010
    #24: "I worry that our "clathrate gun" and associated ice age relics might be cocked and loaded, so to speak. " It is, as with many other questions of climate change, a question of rate of change. Thus we do not know if the loaded gun has birdshot or a deer slug. Archer 2007 is an excellent summary of methane hydrate and their climate change potential. The hydrate reservoir is so large that if 10% of the methane were released to the atmosphere within a few years, it would have an impact on the Earth’s radiation budget equivalent to a factor of 10 increase in atmospheric CO2. ... Fortunately, most of the hydrate reservoir seems insolated from the climate of the Earth’s surface, so that any melting response will take place on time scales of millennia or longer. Acoustic images of real-time methane releases as in this example are dramatic evidence that such melting is indeed occurring, albeit in isolated places. As summer Arctic sea ice continues to dwindle in the coming few years, 'science experiments' such as this will no doubt become more frequent and widespread. In my days in the offshore O&G exploration, hydrates were a well-known drilling hazard; punch a hole in one and you cause it to go unstable very quickly. These guys are going looking for them. Combine that plan with another series of avoidable mistakes such as those leading to the BP disaster and you have given your loaded gun to a bunch of drunk teenagers. Here is a long, but quite thorough 2008 Scripps Institute video on the subject.
27. Ned at 06:36 AM on 5 December, 2010
    It probably goes without saying, so I didn't bother saying this in my comment above. But obviously the consequences of a Venusian style runaway warming are so completely unacceptable, that even a very small probability of that outcome needs to be taken seriously. So I guess I'd characterize my position as "this is very unlikely to happen, but we should be investing a lot more in understanding the relevant processes (clathrates, etc.) just in case".
28. Leland Palmer at 14:42 PM on 5 December, 2010
    Hi muoncounter- I've read one of Archer's papers, but it's been a few months and I'll reread it. I'd feel more confident in Archer's stuff if he hadn't written several joint papers with ExxonMobil chief scientist Kheshgi: ExxonMobil Contributed Papers on Climate Science

    17. Archer, D., Kheshgi, H., and Maier-Reimer, E. 1997. Multiple Timescales for the Neutralization of Fossil Fuel CO2, Geophysical Research Letters, 24: 405. 19. Archer, D., Kheshgi, H., and Maier-Reimer, E., 1998. The dynamics of fossil fuel CO2 neutralization by marine CaCO3, Global Biogeochemical Cycles, 12:259-276. 35. Kheshgi, H. S. and Archer, D. 2004. A non-linear convolution model for the evasion of CO2 injected into the deep ocean. Journal of Geophysical Research,109, C02007, doi:10.1029/2002JC001489. 13. Kheshgi, H. S., and D. Archer, 1999: Modeling the Evasion of CO2 Injected into the Deep Ocean, in Greenhouse Gas Control Technologies, edited by B. Eliasson, P. Riemer and A. Wokaun, pp. 287-292, Pergamon.

    I can see why an earth scientist might collaborate with ExxonMobil, or it's chief scientist. They undoubtedly have a monumental knowledge of geology, and an immense treasure trove of geological information. Having said that, though, Archer's estimate of the total amount of methane hydrates is on the low end of current estimates. It's a really important subject, and I'll get my information about it from sources with no known connection to ExxonMobil.
29. muoncounter at 01:14 AM on 6 December, 2010
    #28: "Archer's estimate of the total amount of methane hydrates is on the low end of current estimates." As I said above, its the rate of release that's critical. Since clathrates are so well-distributed around the world's oceans, their volume is quite significant. But a methane release from an Arctic source may occur independently of one in the Gulf of Mexico. "information about it from sources with no known connection to ExxonMobil." Fair point. I note that Maier-Reimer was with the Max Planck Institute when those papers were written. Kheshgi also co-authored a paper with Bert Bolin, who I believe was Chair of the IPCC.
30. Leland Palmer at 09:10 AM on 6 December, 2010
    Here's a quote from a 2008 paper, posted on the National Energy Technology Lab website, that uses the Tough +/Hydrate computer code for simulating methane hydrate dissociation. MODELING OF OCEANIC GAS HYDRATE INSTABILITY AND METHANE RELEASE IN RESPONSE TO CLIMATE CHANGE

    Two of the most recent studies, each accounting for the coupled contribution of organic matter decomposition and mass transport, have produced drastically different results. Klauda and Sandler [8] provide an upper estimate of 74,400 Gt of methane carbon in hydrate form (27,300 Gt along continental margins, while Buffett and Archer [9] used both compaction and advection in a 1-D methanogenesis/hydrate formation model to reach an estimate of 3,000 Gt of methane in hydrate and 2,000 Gt of gaseous methane existing in a stable state under current climate conditions.

    This paper seems to show increased levels of hydrate release for shallow hydrate deposits with less than a one degree C temperature increase. There are chemical reactions that oxidize the methane or transform it into bicarbonate that I was not aware of until recently. Still, the bigger the reservoir, the smaller the percentage that has to dissociate to cause the climate serious or catastrophic harm.
31. Leland Palmer at 13:43 PM on 9 December, 2010
    Here are a couple other very interesting papers on methane hydrates, that Skeptical Science readers likely haven't seen yet: NETL Methane Hydrates Page

    The results generated through this project have lead to LBNL and LANL researchers publishing four papers in the peer-reviewed literature. (For more information, see the methane hydrate bibliography document.) The first paper, published in the Journal of Geophysical Research (Vol. 13, C12023, 2008) assessed the stability of three types of hydrate deposits and the dynamic behavior of these deposits under the influence of moderate ocean temperature increases. The results indicated that deep-ocean hydrates are stable under the influence of moderate increases in ocean temperature; however, shallow deposits can be very unstable and release significant quantities of methane under the influence of as little as 1 degree C of seafloor temperature increase. A second paper, published in Geophysical Research Letters (Vol, 36, L23612, 2009)here presented the first results of the 2-D slope-scale modeling, demonstrating that shallow hydrates in sloping systems may, alone, generate significant methane and lead to the formation of gas plumes at the seafloor. The results were consistent with the observation of methane venting along the upper limit of a receding GHSZ [Gas Hydrate Stability Zone- LP] off Spitsbergen. The third paper, published in Geophysical Research Letters (Vol, 37, L12607, 2010) and the fourth paper, in final revision for the Journal of Geophysical Research, present the first results of forward-coupled methane release, water column chemistry, and transport via ocean currents using a 1o version of the POP code. These establish a new paradigm for understanding the response of the oceans to methane release on a large scale. In particular, the work highlights the importance of resource limitations. Large and concentrated methane plumes may deplete the surrounding water of oxygen and other trace nutrients, reducing the ability of methanotrophs to consume the methane and increasing the chance of release into the atmosphere. This is in sharp contrast to previous assumptions of “99% consumption” of methane for all release scenarios.

    So, the news from these papers appears to me to be bad. The rate of release from hydrate deposits is limited by the endothermic nature of hydrate dissociation, and by fluid flow limitations, according to the second paper mentioned above, though. So - a crucial point - whether these deposits will lead to runaway warming may be very dependent indeed on their total quantity. An order of magnitude difference in estimates of their total quantity, with one of the estimates coming from Archer, who writes papers with ExxonMobil chief scientist Kheshgi, is just unacceptable.
32. muoncounter at 13:16 PM on 13 December, 2010
    #31: "The rate of release from hydrate deposits is limited by ..." Nothing limited about these numbers: Gas escape features off New Zealand: Evidence of massive release of methane from hydrates Multibeam swath bathymetry data ... show gas release features over a region of at least 20,000 km^2. Gas escape features, interpreted to be caused by gas hydrate dissociation, include an estimated a) 10 features, 8–11 km in diameter ... If the methane from a single event at one 8–11 km scale pockmark reached the atmosphere, it would be equivalent to ∼3% of the current annual global methane released from natural sources ... Full paper, with graphics, available here.
33. Leland Palmer at 08:41 AM on 2 January, 2011
    Thanks, muoncounter, it's a very interesting paper. Glaciations tend to start gradually, but end abruptly. Various people, referenced in the your paper and the Wikipedia article on the clathrate gun hypothesis, suspect that lower sea levels during those glaciations can cause pressure induced release of methane from hydrates, rapidly increasing atmospheric methane levels leading to an increase in temperatures and higher sea levels. If true, this does demonstrate one thing: large scale methane releases from oceanic hydrates can reach the atmosphere. This is supported by the carbon isotope data from events like the PETM and the End Permian- although direct intrusion of magma from the Siberian Traps volcanism into methane hydrate deposits may have been necessary to trigger the End Permian mass extinction. These carbon isotope data suggest that two to five or more trillion tons of isotopically light methane from the hydrates entered the active carbon cycle during those events. Can our greenhouse driven pulse of heat, which is working its way inexorably downward into the oceans, stimulate a similar rapid release of methane?
34. ribwoods at 17:43 PM on 17 March, 2011
    I think the distinction between gain < 1 (no runaway even though feedback is positive) and gain > 1 should be explicitly stated early in the article. Climatology's use of "positive" (gain could be either > 1 or < 1) in regard to feedback is not the same as electronics's use of "positive feedback" (gain strictly > 1), IIUC. _That_ is why "positive feedback" has such a strong connotation of "runaway" for many newcomers to climate science. I think you need to jump on this early and explicitly in any discussion of "runaway", in order to cut down on misunderstandings. Don't bury the significance of _gain_ halfway down the article at the end of a paragraph ("and call the gain factor g").
35. Leland Palmer at 15:05 PM on 22 May, 2011
    Here's another recent paper, which uses a state of the art atmospheric chemistry model to predict much stronger positive feedback from indirect atmospheric chemistry effects of large methane releases, than from the methane itself. They are talking about several hundred percent increases in stratospheric water vapor, for example, increased methane lifetime of roughly 100 percent for very large releases, and large increases of tropospheric ozone. The hydroxyl radical, by their modeling, decreases in the troposphere, where it is needed to oxidize methane, and increases in the stratosphere. The positive feedback factor that they calculate (eta) ranges from 1.5 for small releases, up to 2.9 for large ones. Strong atmospheric chemistry feedback to climate warming from Arctic methane emissions

    Here we apply a “state of the art” atmospheric chemistry transport model to show that large emissions of CH4 would likely have an unexpectedly large impact on the chemical compositioof the atmosphere and on radiative forcing (RF). The indirect contribution to RF of additional methane emission is particularly important. It is shown that if global methane emissions were to increase by factors of 2.5 and 5.2 above current emissions, the indirect contributions to RF would be about 250% and 400%, respectively, of the RF that can be attributed to directly emitted methane alone.

    It's a very important result, IMO, which could provide a bridge from mild CO2 based warming to runaway methane and atmospheric chemistry change based greenhouse heating. It's a very different atmosphere that they are talking about, with sustained methane release rates of 4 to 13 times those of today. Stratospheric water vapor and stratospheric hydroxyl radical increase, tropospheric hydroxyl radical decreases, and tropospheric ozone increases, leading to indirect warming several times that of the warming from methane itself. It's particularly worrisome because this appears to be an honest result, resulting from a fair query of a state of the art atmospheric chemistry transport model. If this work holds up, it may help explain the strong positive feedback of past apparent methane catastrophes including the Paleocene-Eocene Thermal Maximum and the End Permian mass extinction, I think.
36. MajorKoko at 22:22 PM on 10 June, 2011
    How do we know that the system is limited purely by the nature of its positive feedback? I'm trying to understand how the possibility of negative feedback being a limiting factor has been disregarded? There is a theory that the increasing levels of water vapour in the atmosphere due to increased temperatures might greatly increase cloud cover, increasing the earth's reflectance, reducing heat absorbtion into the system. As far as I know, the question of what the impact of increased atmospheric water vapour has on cloud formation and, therefore, albedo is still very open? So I wonder whether it isn't premature to assume that the system is self-limiting due solely to the reason given above? Particularly as the methane release issue casts further doubt on the idea that the system is self-limiting, without some other unconsidered factor coming into play. Thoughts?
    Response: See Water vapor is the most powerful greenhouse gas and Clouds provide negative feedback.
37. KR at 07:30 AM on 18 September, 2011
    jpat A gain < 1.0 can still provide a several-fold amplification: Output = Forcing / ( 1 - gain ) A gain of 0.9 will result in an amplification of 10.0, as the sum will be 0.9 + 0.9^2 + 0.9^3 + ..., summing to 10x.
38. jpat at 09:14 AM on 18 September, 2011
    True but irrelevant. The candidate system must be capable of power amplification. The output variance (power) in the passive system you describe (unit forward gain, attenuated feedback) can never exceed the input variance. In the system we are considering, the output variance due to Milankovitch forcing exceeds the solar variance from same.
39. KR at 11:27 AM on 18 September, 2011
    jpat - Not power amplification, that's a misunderstanding of the system. GHG's act as a throttle controlling energy flow out of the Earth climate, all said energy coming from the sun. Input energy inevitably gets dumped to space, the question is in what thresholds are in place in the mid-point of the system, driving internal energy levels so that the throughput can occur. You seem to be treating this as a limited system, rather than an open system with energy flows. I suggest more reading on your part. I would point you to The Discovery of Global Warming as a starting point. Not a passive system - a dynamic system. That's a serious error in viewpoint.
40. jpat at 12:39 PM on 18 September, 2011
    rf -----^v---- ri | | -^v----*----|+\__| ? Vin --|-/ | _|_ |______| \ / Consider the circuit above. The op-amp is configured for unity gain. The feedback "gain" = ri/(ri + rf) < 1 and the feedback is positive. Since the forward gain is unity, the open loop gain (referred to as "gain" in #37) = feedback attenuation < 1. The claim in #37 is that this circuit will amplify. Not so much. The inverting input of the op-amp is wired to the output (Vo). The op-amp is ideal so the non-inverting voltage (v+) is equal to the inverting input voltage (v-) So Vo = v+ = v-. This means the voltage drop across rf = 0 which by Ohm's law means the current through rf (Irf) =0. But Irf = Iri. Thus the voltage drop across ri = 0. Thus Vo = Vi. The gain of this circuit is unity. Now suppose the op-amp is configured for a gain of 2 and the feedback divide ratio < .5 (again, so that gain as defined in #37 < 1). At first blush it appears capable of amplification. Now add the stability constraint that the input impedance > 0. We find that for stability, ri > rf. The open loop gain for the circuit can be shown to be 2-rf/ri which again must be <= 1 for stability. This doesn't disprove anything but rather shows that the feedback gain equation is subject to boundary conditions including conservation of energy. I really wonder whether one could devise a physical system capable of doing work using the formulation provided by KR.
41. jpat at 12:48 PM on 18 September, 2011
    I guess my attempt at input a schematic got mangled by the html processor (block tags doesn't seem to work). KH - I posted before I saw your response. I take your point but am still struggling to understand how feedback amplifies the Milankovitch cycles. The system has to do real work, melting ice, heating the oceans etc. and I thought the argument was the system does more work with feedback than it could do without. That sounds like power amplification to me but I'll come back after I read up some more.
42. KR at 13:02 PM on 18 September, 2011
    jpat - A strictly electronic analogy will not work well with respect to the climate system. To some extent a transistor or op amp would correspond, but transistors are too non-linear. An operational amplifier could be used to build a corresponding circuit, with the Stephan-Boltzmann law providing the negative feedback, but not a bare op-amp. The S-B law indicates that thermal radiation to space scales as T^4 (in Kelvin), and means that excess energy accumulation in the climate, simply by heating/cooling the Earth, will rebalance the input/output levels. That's your negative feedback. There are fast energy state response elements (water vapor, clouds), slow response elements (ocean CO2, albedo from ice coverage, vegetation), and over and above this the forcings that drive those feedback items. In the ice age cycles a small amount of insolation variation (orbital changes shifting land/ocean exposure and polar insolation/albedo changes) acted as a forcing, with the temperature sensitive elements responding at various delays. Currently we are introducing a direct forcing with GHG's, and should expect to see water vapor levels, albedo, vegetation, and in fact CO2 levels from ocean solubility respond to the initial forcing with their own changes. All of these - forcings and feedbacks - act as throttles on the flow of energy from the sun into the climate and back out again, with rates dependent on the various states, gas concentrations, albedo, etc.
43. jpat at 02:05 AM on 19 September, 2011
    KR - The op-amp is only their to provide the unity forward gain and ideal summing node required to implement the transfer function you gave in #37. I went through the exercise because I am pretty sure that a passive (i.e. forward gain <=1), physical implementation cannot provide variance amplification. I think this is a general result, imposed by boundary conditions and conservation of energy. Think of a step up transformer. True the voltage measured at the secondary will be higher than that at the primary but the variance (power) remains unchanged. I do think that climatologists use the term feedback in a different way than I am used to thinking about it. In this review, they state "S is proportional to 1/(1 − f)" where S is the sensitivity and f is the "net feedbacks". A control systems guy would call f the open loop gain, as in CL(s) = G(s)/(1-G(s)H(s)) where G(s) is the forward transfer function and H(s) is the feedback transfer function and s is the Laplace operator. I deduce from the above that climatologists model the forward gain as unity (which I find odd - if there were no feedback would all of the energy go into heat flux? Would some get reflected back immediately?) which is why the op-amp is configured for unity gain in my analogy. In any case I'm still trying to formulate a transfer function model for the radiant balance equations so that I can understand how it is that the CO2 could lag the temperature through the entire glaciation cycle, be responsible for the rapid interglacial rise and yet allow the small Milankovitch forcing to turn things around. Seems implausible but I've been fooled by intuition before.
44. KR at 06:14 AM on 19 September, 2011
    jpat - The Milankovitch forcings last for millenia, which allows time for various feedbacks to take effect. Just the Milankovitch forcings alone should change the average temperature of the Earth by a total of 1-2.5C; the 6-8C swings seen over the ice age cycles are due to the amplification of the forcing change by water vapor, CO2, ice retreat/advancement, etc. The Wiki on this is actually fairly reasonable. I would suggest caution in reasoning from electronics - that would be an analogy, and while providing an analogy is a useful way to explain something, you cannot reason from an analogy back to the original system, as analogies only resemble the original complex system in part. You will inevitably get tripped up on the differences - there is no substitute for actually studying the real thing. And the real thing includes multiple time lags, chaotic/non-linear variations, and lots of different forcing inputs
45. KR at 06:17 AM on 19 September, 2011
    jpat - I strongly suggest you take a look at the CO2 lags temperature thread in this regard.
46. jpat at 11:11 AM on 19 September, 2011
    I think I have an analogy that maybe helpful to those like me who come from an engineering background and have trouble putting the feedback discussed here into a more familiar context. Consider a circuit comprised of a voltage source connected to a resistive divider. Assume the divided voltage is our output node. To this output node we connect the control port of a voltage controlled voltage source of gain b whose output is connected to the output node through a series resister. This is our feedback path. Assume all resisters are 1 ohm. If b is zero the output voltage is simply one third the input voltage. As b is increased from zero to 1, the output rises from vin/3 to vin/2. Thus we see a "gain" of 1.5 compared to the no feedback case but in both cases the forcing (vin) is attenuated. With b=0, the thevenin impedance looking into the output node is 1/3 and therefore the output power density is 4kT/3 W/Hz. With b = 1, the thevenin impedance rises to 1/2 and so too does the power density, to 2kT. This indeed we can see an increase in variance with feedback gain < 1. Finally note that b in my analogy is not the same as f in #37. In the example, b could actually vary between 0 and 3, (corresponding to 0<=f<1) before the circuit became unstable. However, to get actual gain (in the normal sense of the word, i.e. an output greater than vin), b must be greater than 2. That is, to get actual amplification requires active feedback, as expected.
47. jpat at 11:36 AM on 19 September, 2011
    "The Milankovitch forcings last for millenia, which allows time for various feedbacks to take effect...The Wiki on this is actually fairly reasonable." From the Wiki: "Each glacial period is subject to positive feedback which makes it more severe and negative feedback which mitigates and (in all cases so far) eventually ends it." And yet in this tread (see #54), chris accuses me of making an "evidence free assertion" when I tried to make that very point. Which is it? Is climate sensitivity constant or does it vary as negative feedback(s) kick in at the extremes?
48. muoncounter at 11:53 AM on 19 September, 2011
    jpat#47: "Is climate sensitivity constant or does it vary" Why does it have to vary? The sensitivity to any particular forcing may remain constant while the forcings themselves vary. For example, as more CO2 enters the atmosphere, that forcing increases with constant sensitivity. FWIW, I wonder if you've considered a circuit analogy from a simpler time: Hartley and Colpitts oscillators exploit feedback without necessarily running away.
49. scaddenp at 12:09 PM on 19 September, 2011
    I think that climate sensitivity has to vary. Consider two extremes. If you melt all the ice pack, then the albedo feedback component practically disappears. Second consider an iceball earth with just enough solar input and CO2 start a melt the ice at the tropics. At point water vapour enters the atmosphere where it wasnt before. In this scenario, climate sensitivity has to be extreme. I think this is reason (beside the measurement errors) for the wide of range of sensitivities from paleoclimate studies. Carbon-cycle modelling still has a lot of uncertainty with many different ways to replicate known data. However, I dont think any of them would give a linear or log-linear response of atmospheric CO2 to temperature.
50. jpat at 13:56 PM on 19 September, 2011
    "Why does it have to vary? The sensitivity to any particular forcing may remain constant while the forcings themselves vary." The Milankovitch forcing are very regular and of essentially constant amplitude. The sensitivity is determined by the sum of all the feedbacks so to the extent that one or more feedbacks depends on temperature or other dynamic variables in the climate, the sensitivity will change. I don't think it is controversial that the assumption of constant sensitivity is only valid over a few degrees. With regard to your other point, I think perhaps a more apt analogy would be a class of circuits called injection locked amplifiers. In and of themselves they are stable but have positive feedback. They don't oscillate because their complex poles have very low quality factor and are unable to remain on the jw axis for any period of time. However, if we inject a small periodic signal near the eigenvalue frequency, the circuit exhibits behavior very much like a phase-locked loop. Its phase trajectory tracks the input signal inside a bandwidth determined by the injection amplitude. Outside of this bandwidth, the phase trajectory variance falls as f^4. This makes them useful as filters but they are not widely employed because they have a tendency to exhibit chaotic behavior. The interesting thing about these circuits is that they behave much like oscillators in that each node in the circuit is a delayed version of the previous node (like the CO2 curve in the ice core is kind of a delayed version of the temperature). Another interesting thing is what happens when we add a constant forcing bias. The peak amplitude of the output does not change but rather the duty cycle is modified. It reaches equilibrium by modifying the symmetry of the output waveform. For instance with positive constant forcing the waveform adjusts to spends more time in the positive realm than in the negative realm but the peak amplitude remains unchanged. What got me thinking about this was a Fourier analysis I did of the Vostok ice core data. The phase noise power spectral density exhibits the same BW tracking fingerprint I saw when analyzing these injection locked amplifiers years ago. This doesn't prove anything but it is an interesting idea.

Prev  1  2  3  4  5  Next

Post a Comment

Political, off-topic or ad hominem comments will be deleted. Comments Policy...

You need to be logged in to post a comment. Login via the left margin or if you're new, register here.