1. Attachments are working again! Check out this thread for more details and to report any other bugs.

Uncertainty

Discussion in 'Environmental Discussion' started by tochatihu, Apr 7, 2014.

  1. tochatihu

    tochatihu Senior Member

    Joined:
    Apr 10, 2004
    8,972
    3,501
    0
    Location:
    Kunming Yunnan China
    Vehicle:
    2001 Prius
    This is perhaps the worst possible topic, but it is active. I offer 3 more new, from AGU.

    "Extreme weather events may not be tied to loss of Arctic sea ice"
    DOI: 10.1002/2014EO140018

    "Economic losses due to catastrophes"
    DOI: 10.1002/2014EO140006
    "Worldwide economic loss due to catastrophic events added up to US$140 billion in 2013, with insured losses adding up to $45 billion, according to a report by the insurance provider Swiss Re. Though these numbers are down from $196 billion in economic losses and $81 billion in insurance losses in 2012, Swiss Re reports an upward trend in losses."

    "Historical and future learning about climate sensitivity"
    DOI: 10.1002/2014GL059484
    "Equilibrium climate sensitivity measures the long-term response of surface temperature to changes in atmospheric CO2. The range of climate sensitivities in the Intergovernmental Panel on Climate Change Fifth Assessment Report is unchanged from that published almost 30 years earlier in the Charney Report. We conduct perfect model experiments using an energy balance model to study the rate at which uncertainties might be reduced by observation of global temperature and ocean heat uptake. We find that a climate sensitivity of 1.5°C can be statistically distinguished from 3°C by 2030, 3°C from 4.5°C by 2040, and 4.5°C from 6°C by 2065. Learning rates are slowest in the scenarios of greatest concern (high sensitivities), due to a longer ocean response time, which may have bearing on wait-and-see versus precautionary mitigation policies. Learning rates are optimistic in presuming the availability of whole ocean heat data but pessimistic by using simple aggregated metrics and model physics."

    In the last the emphasis on perfect was mine because I don't know what that means in this context and it is a very special word. Hehe. If you want to know, read it. Anyway, I am hearing that we simply can't constrain climate sensitivity until we understand the oceans better. That seems an excellent goal and IMHO worth whatever billions the oceanographers say it will cost.

    Nah. Let's identify an ocean expert in the US congress to provide oversight. From Oklahoma perhaps? Or California's Orange County which touches the sea. As this is proper snark, I don't name names.
     
  2. austingreen

    austingreen Senior Member

    Joined:
    Nov 3, 2009
    13,525
    4,057
    0
    Location:
    Austin, TX, USA
    Vehicle:
    2018 Tesla Model 3
    Model:
    N/A

    There is a lot here, and I can't read the papers without paying, so some of this may be wrong.... when you actually read them.

    1) The first 2 on extreme weather put a lot of evidence against the grey put forth in IPCC 2 - ghg cause extreme weather. That is a key contention of apocalyptic climate change. Both papers seem to point to evidence that this is wrong. Ofcourse like the zombie apocalypse, its impossible no matter how high the evidence is against it there will be believers. For those that like to follow the science though we do know that to use the scientific method we need to be more specific than ACC bad, we need to define what extreme weather is then look at the statistics for those particulars, like huricanes, which should be considered separate than tornados (both evidence against more frequent extreme ones), and then use heat waves (evidence for) and not just say bad is bad.

    The 3rd is more interesting. They seem to be saying in 15 years we will have enough data to say if the sensitivity is less than 3 or if its greater than 2 or both ;-) My instinct is it is between 2 and 3 which seems to agree with the ice ages;). It will take 50 years to prove statistically that it is less than 4.5, but we have papers which say if sensitivity is a constant that over 4.5 would have caused more ice ages. So we should be able to reject these out of hand if the theory is its a constant. That begs the question of tipping points and speed. If we define sensitivity as the fast response, and tipping points happening if the combination of temperature and ghg rise above a certain level, then after we reach that point its just a matter of time. If you believe this theory with Hansen, then we are already far beyond his tipping point of 350 ppm carbon dioxide and its just a matter of time, less than 100 years, before the seas swallow NYC and many other low lying areas.

    What should we do in faces of this uncertainty? Most don't believe that you should tax the hell out of people to build nukes and solar. Why not? It won't do much of anything if the taxes are low enough to pass like the congresses cap and tax plan, and if they are high enough its simply too expensive for the middle class. This is all political anyway.

    But in the US we can definitely do more. One thing left and right seem to agree on is transportation needs to be less oil intense. One thing that economists agree on is this takes decades and fuel taxes are likely the most efficient way to do it. Unfortunately politicians seem to demagogue the issue. Hanson and the forbes editors seem to agree on a couple of things. Its more efficient to tax fuel, than regulate with cafe standards, and this should be used to reduce other taxes not just spent by the politicians. Unfortunately schemes like the congresses and californias had the politiicians spending all the money, instead of reducing other taxes. Indepenants, republicans, and democrats all don't seem to like new fuel taxes to grow government.

    In the electricity sector a combination of the policies of texas and germany would seem like an affordable way to transition. Unfortunately most states are not energy savvy. The established big utilities have a lot of hooks into government and the regulators do things to favor those utilities. The disasterous california deregulation was written with a great deal of input from PG&E and SCE, and when it cauesed blackouts the state (tax payers and rate payers) paid for their mistakes. Given that better fracking regulations along with grid improvements, feed in tarrif for solar and wind (much lower for wind than solar) and a coal tax, and removal of grandfathering in 10 years of all coal plants may work best.

    None of these ideas cost much if sensitivity is less than 2, but do make it much cheaper if sensitivities are higher. Think of it as insurance instead of the kind of wealth transfer that does not materially change ghg that is californias AB32.
     
  3. wxman

    wxman Active Member

    Joined:
    Nov 25, 2008
    619
    224
    0
    Location:
    Tennessee
    Vehicle:
    Other Non-Hybrid
    Model:
    N/A
    Regarding polar amplification and "extreme weather" events, it's been my observation (casual, not rigorous study) that upper flow typically weakens and deamplifies in mid to late summer, which is when the arctic-to-tropical temp gradient is at its minimum, not amplify (i.e., become "wavier"). It would thus be expected that extreme wx events like severe wx to decrease in intensity, not increase, and be the default position.

    Still don't understand why some mets in academia anticipate upper flow to weaken and amplify in a warming-the-arctic-relatively-more-than-the-tropics-and-mid-latitudes AGW scenario.
     
    austingreen likes this.
  4. tochatihu

    tochatihu Senior Member

    Joined:
    Apr 10, 2004
    8,972
    3,501
    0
    Location:
    Kunming Yunnan China
    Vehicle:
    2001 Prius
    Wxman, it appears that you are using seasonal temperature swings as analogues to annual or longer directional temperature changes. This is also done in many studies of T sensitivity of soil respiration (I wrote one of them). In the latter case, we are beginning to find out that it is an imperfect analogy. I don't know how well it 'holds up' in the climate system.
     
  5. tochatihu

    tochatihu Senior Member

    Joined:
    Apr 10, 2004
    8,972
    3,501
    0
    Location:
    Kunming Yunnan China
    Vehicle:
    2001 Prius
    If temperature averages get a bit higher, then the same size (or number) of extremes may not matter much. If sea level averages get a bit higher, then the same size (or number) of storm surge events may matter quite a lot. This is because your coastal defense is either good enough, or gets overtopped. So we should probably be more nuanced about extremes vs. the baseline.
     
  6. tochatihu

    tochatihu Senior Member

    Joined:
    Apr 10, 2004
    8,972
    3,501
    0
    Location:
    Kunming Yunnan China
    Vehicle:
    2001 Prius
    Often when you look at 'paleo', uncertainty really gets big. Scientifically, I think this is a very good study:

    Foster GL Rohling EJ 2013. Relationship between sea level and climate forcing by CO2 on geological timescales. PNAS 110: 1209-1214. Doi: 10.1073/pnas.1216073110

    FR13 sea level.jpg

    Over 40 million years, CO2 levels like now are associated with about 14 meters higher sea level, if you don't look at the error bars. Now this will resemble predictions such as Hanson and Gore, though I don't know if they get their ideas from here or somewhere else.

    Or, if you look at the error bars, you pretty much have to say it could be anywhere within quite a large range. The authors are also careful to say that large SL changes take at least 1000s of years to play out.
     
    bwilson4web likes this.
  7. tochatihu

    tochatihu Senior Member

    Joined:
    Apr 10, 2004
    8,972
    3,501
    0
    Location:
    Kunming Yunnan China
    Vehicle:
    2001 Prius
    Further, once you get above 900 ppm CO2, FR13 indicates that you are stuck with very high sea level (eventually). At that level, the error bars no longer let us 'hide'. I don't think we are going to 900 this century, and not the next either (if our heirs are clever). But it does suggest that there are some really truly planetary boundaries that need to be respected. It simply remains difficult to define them. As they remain undefined, it only makes sense to approach them with a great deal of caution.

    Let us not fail to notice also that such high CO2 30 million years ago was something that ocean ecosystems (fish, coral, stuff like that) managed to get through somehow. There I go again, looking for the bright side.
     
  8. austingreen

    austingreen Senior Member

    Joined:
    Nov 3, 2009
    13,525
    4,057
    0
    Location:
    Austin, TX, USA
    Vehicle:
    2018 Tesla Model 3
    Model:
    N/A

    We would definitely consider the cold and snow that was average during the little ice age, as extreme today. The climate has changed to be much warmer. On storm surges again comparing it to other coast lines, we have been building more in harms way, requiring less and less of a storm surge to cause damage. That is why the dutch have built up the dikes much higher since that time.

    There has been a directional shift in climate since the little ice age. If we consider mean as the temperatures in the last 20 years and sea levels in that period then it is highly questionable if we have gone "extreme". certainly this mean is different than 300 years ago.
     
  9. wxman

    wxman Active Member

    Joined:
    Nov 25, 2008
    619
    224
    0
    Location:
    Tennessee
    Vehicle:
    Other Non-Hybrid
    Model:
    N/A
    Doug,

    I'm using seasonal (mid-summer) upper flow configuration as a surrogate for what I would expect the trend of flow configuration to be during other seasons as a result of polar amplification. I was not able to read the full paper of your reference (abstract only - full paper paywalled). However, that first reference appears to challenge the idea that polar amplification (warming polar regions/loss of Arctic ice) is resulting/will result in increased atmospheric blocking, INCREASING amplification (jet stream "waviness"), and thus, more "extreme wx". This is also what I am questioning, as my experience suggests that decreasing poles-to-tropics temp gradient typically results in deamplifying the upper flow configuration rather than amplifying it. Weak flow, resulting from the decreased poles-to-tropics thermal gradient during summer, generally has broad, low amplitude features, if not weak zonal flow. At least that's my observation after several decades of following these sorts of things.

    Pattern recognition is a fundamental skill in forecast meteorology. Things may not necessarily translate to other seasons, I suppose, but I remain skeptical that polar amplification will result in higher amplitude upper flow, more atmospheric blocking, and "extreme wx". That said, I am not comfortable will unrestrained GHG emissions as I agree that we're conducting a vast empirical study of climate response to continuous atmospheric increases in GHG.
     
    austingreen likes this.
  10. tochatihu

    tochatihu Senior Member

    Joined:
    Apr 10, 2004
    8,972
    3,501
    0
    Location:
    Kunming Yunnan China
    Vehicle:
    2001 Prius
  11. bwilson4web

    bwilson4web BMW i3 and Model 3

    Joined:
    Nov 25, 2005
    27,064
    15,372
    0
    Location:
    Huntsville AL
    Vehicle:
    2017 Prius Prime
    Model:
    Prime Plus
    Anyone who generates a log-scale graph combining so much data . . . I'll have to get that one. Given the phase change effects of snow/ice, it makes a lot of sense suggesting there are heat-flow, tipping points, the 'knee in the curve.' Granted, beggars can't be choosers, I'm curious about that data-gap. There are systems that have two, semi-stable modes, and I'm wondering if that is what in semi-conductors is called a band-gap . . . a place where things switch rapidly from one state to another.

    Bob Wilson
     
  12. tochatihu

    tochatihu Senior Member

    Joined:
    Apr 10, 2004
    8,972
    3,501
    0
    Location:
    Kunming Yunnan China
    Vehicle:
    2001 Prius
    Bob, I posted that paleo SL graph and the reference mostly because it is 'floating around' the web without attribution. Using non-linear axes is justified when data span a wide range. OK here for me because CO2 effects on energy balance are logarithmic. I am not a fan of log-log plots when they aspire to show close correlations between noisy, short-range data.

    You use the word 'rapidly', which means one thing for semiconductors but very different in geology. There it roughly means a million years or less.

    On a geological timescale, earth climate appears to have two quasistable attractors - lots of ice and none at all. At present, we are closer to 'lots'. If we examine >500 million years the proxies are quite vague. FR13 looked at 50 million years- so the proxies are more effective but still with galloping error bars.

    Other reasons to limit attention to this 'short' time is that we are past the K-T extinction, which is quite a muddle in terms of asteroid impact and other causes. Photosynthesis and decomposition had evolved to the current condition (pretty much). This matters, along with how fast silicate and carbonate minerals get pushed up to where they can interact with the atmosphere (by tectonics, which is not a constant factor).

    I think we can learn from 50my paleo is that CO2 is very responsive to T. It also appears to cause T changes, but convincingly sorting those out requires very tight date controls.

    The only odd thing about recent past is that humans have added a new (paleo) source of CO2 to the mix. This appears to upset the cart. Geology and biology will continue to do what they do, but the physics of CO2 cannot be set aside. We know that it matters, but the earth system being such a beast, the uncertainty is such that we cannot yet quite say how much and how rapidly the beast can be redirected. The only thing we can say for absolute certain, I reckon, is that it will not be easy to transition to the next ice age with CO2 above 400 and rising. Yet one can read exactly that here at PC. What a thing!

    All that aside, it seems that earth-system scientists are facing up to uncertainty. Not all of them, to be sure, one can read that we are already doomed without looking too hard. Well, are we? Maybe, but perhaps one should not exclude the possibility that biology is rather effective. Even now it is sucking down half of that newly emitted fossil C. With a bit of a nudge from sentient beings, it might do even a bit more. Put that together with a nudge towards energy efficiency, and we just might get through this. Worth a try.
     
  13. bwilson4web

    bwilson4web BMW i3 and Model 3

    Joined:
    Nov 25, 2005
    27,064
    15,372
    0
    Location:
    Huntsville AL
    Vehicle:
    2017 Prius Prime
    Model:
    Prime Plus
    Doug,

    You've reminded me about what we knew of biology just 100 years before I was born, December 1949. In 1849, we didn't have germ theory or a century of what has become modern medicine. So it is humbling to contemplate how much more is likely be know in another 100 years . . . if we don't blanket the earth with the plutonium layer . . . for the next sapient order/family/species to discover. <grins>

    Still, I'm looking forward to getting the paper, assuming it is the source of the graph.

    FYI, with 50,000 miles on my wife's 2010 Prius, I'm planning in May to do another series of mph vs MPG graphs. Might as well do the 160,000 mile, 2003 Prius too.

    Bob Wilson
     
  14. austingreen

    austingreen Senior Member

    Joined:
    Nov 3, 2009
    13,525
    4,057
    0
    Location:
    Austin, TX, USA
    Vehicle:
    2018 Tesla Model 3
    Model:
    N/A
    And here is where the general questions of uncertainty lie. If the the sea level rises 2 feet in 100 years, which seems to be around the mean prediction, then there is certainly time for people to move and adjust and build sea walls. If on the other hand the seas rise 20 feet in 100 years then we get widespread suffering.

    Note, nowhere does he say how soon in the movie, but it is implied to be rapid. Numerous papers have been published in the last decade to say that there just is no mechanism to rise more than 2 meters in the next centcury, and that is unlikely giving rise to this good miami study of 0.3-1 meter (1 to 3.3 feet). But still there is a great deal of uncertainty on that we have the younger dryas that switched the north atlantic conveyer, and that may slow sea level rise down.

    One thing we can be certain from the ice cores, man or no man, during the last 2 interglacials sea levels were much higher, so this seems a likely scenario that man is speeding up.


    We seem to be in a rapid period of extinction today, but this has more to do with land use (man destroying habitat) than climate change. We do know that many of the migratory paths have been blocked by fences and development, so animal populations not only have the natural challenges they did in the past, there are new man made ones, that climate change simply makes worse.
    I'm sure we will have much more data in a couple of decades about what is happening. That is very small in geological time. It would seem to be a good idea to purchase a little insurance and prepare and decrease. The army corps of engineers in the 1950s proposed a sea wall to protect NYC, that would have protected from sandy. Note no one needed to invoke man made climate change to know that this low lying area, with so many people was at risk to storm surges, given the huricanes that had hit the area in the 1930s and 1940s. People decided it was not worth the money, but of course the damage from sandy cost a great deal more in simple property damage. The lives affected were never in the calculation.
     
  15. mojo

    mojo Senior Member

    Joined:
    Sep 28, 2006
    4,519
    390
    0
    Location:
    San Francisco
    Vehicle:
    2012 Prius v wagon
    Model:
    Three
    Your statement is rather uncertain.
    Geological_Timescale.jpg
     
  16. austingreen

    austingreen Senior Member

    Joined:
    Nov 3, 2009
    13,525
    4,057
    0
    Location:
    Austin, TX, USA
    Vehicle:
    2018 Tesla Model 3
    Model:
    N/A
    Gee mojo,
    I can't even try to pretend I know what I'm looking at there. At a minimum we should be using T not changes in temperature, and have radiation. If we do that absolutely we have a great deal of uncertainty on fast ghg response.

    At a minimum we should be looking at changes of radiation in milankovitch cycles where the earth changes in distance to the sun, creating ice ages and interglacials. Over millions of years in your graph mojo the suns radiation itself has changed a great deal, not just the distance.

    Now we can make a simple model of just our guessed changes of solar radiation, the known distance changes in milankovitch cycles, forcings in terms of ghg equivance, and find sensitivity. Alley has done that and come up with 2.8, a number that has to be within range of error, but there is a great deal of uncertainty to each of the proxies. We can do some sanity checking though. sensitivity without any feedback is around !, and everything that we know about feedback is that it is positive in relation to carbon dioxide (for example more carbon dioxide, higher temperatures, more water vapor, higher temperatures). Most set this lower limit at 1, but some dispute and claim negative feedback is possible, although they can't point out these mechanisms well (more ghg -> hotter ? more volcanic eruptions -> more blocking particulates->colder). There is much we don't know, 1 seems a sane choice.

    Even if out temperature proxies are wrong, we do have a great deal more confidence on the timing of ice ages. If sensitivity was over 4.7 we get more ice ages, so that is out, leaving 1 to 4.7 as our sanity checks unless major science is wrong.

    Closing the uncertainty gap means understanding natural variation, and that means more data. If we model ENSO and other mechanisms better we can close that gap, and do not need to go back as far in the paleo record. The problem is the climat models have been back tested, and don't seem predictive. Still we can throw in numbers like 1.5 for sensitivity and find they underestimate warming in the last 30 years, or numbers like 4 and find they overestimate warming.
     
  17. mojo

    mojo Senior Member

    Joined:
    Sep 28, 2006
    4,519
    390
    0
    Location:
    San Francisco
    Vehicle:
    2012 Prius v wagon
    Model:
    Three
    Any 5 year could ascertain that there has been no correlation between CO2 and temp throughout Earths history.
    Nor in the present day where temps have paused for the past 17 while CO2 has continuously risen.
     
  18. tochatihu

    tochatihu Senior Member

    Joined:
    Apr 10, 2004
    8,972
    3,501
    0
    Location:
    Kunming Yunnan China
    Vehicle:
    2001 Prius
    Before I speak about that long geological record of T and CO2 I want to highly praise Scotese's web site

    Home Page

    His maps of continents through geologic time are the best you will find.

    OK. if you go back through time in search of T and CO2 you will use proxies. Each of those have some uncertainty in 'both axes'. That means in dating and also in the variate that your proxy stands in for. Neither are shown in the graph @35. Overall this limits its utility as a tool of correlation.

    If one looks at 50 million or 5 million years, things rather firm up. I agree that 5 year olds might not appreciate that. But it is not a very important point.
     
  19. austingreen

    austingreen Senior Member

    Joined:
    Nov 3, 2009
    13,525
    4,057
    0
    Location:
    Austin, TX, USA
    Vehicle:
    2018 Tesla Model 3
    Model:
    N/A
    If most of the 5 year olds don't understand, but all of the scientists that study it find correlation, perhaps you should raise your level of understanding of the relationship beyond that of a 5 year old. Here is an alley talk where he takes on the 5 year olds. It is rather long, but if you are interested, I am sure you will benefit.



    Now the cherry picking of 1998 is how you can distort with statistics. If we take natural variation, the last 17 years seem to have increasing temperatures ;-) In fact if you look at 10 yeaer, 20 year, 30 year, 10 year moving average, etc, temperature and ghg do go up together. Why pick 1998? Because ENSO made the year exceptionally warm. This is one of the criticisms of the IPCC models they may overestimate ghg sensitivity because they badly model natural variation.


    If we look beyond 800,000 years, at least according to alley these proxies for CO2 and temperatures get quite bad. If we look at 5 million yeas or longer we must also include the effect of solar radiation that changes quite a bit on the earth's orbit and the suns radiation along with the coarseness of the proxies, mainly ocean sediment cores. Ignore that and none of us will do much better than a 5 year old.

    I would be surprised if we had consensus on the co2 levels 5 million years ago, let alone 500 million, but we do have some guesses.
     
  20. tochatihu

    tochatihu Senior Member

    Joined:
    Apr 10, 2004
    8,972
    3,501
    0
    Location:
    Kunming Yunnan China
    Vehicle:
    2001 Prius
    Paleo CO2, T and SL by proxies are not what I would call guesses, but they include some uncertainty.

    If we limit our attention to 50 million years, terrestrial biology has done all its 'major innovations' (most recently C4 grasses). Tectonics is relatively stable over the period. So, if the proxies can be firmed up over that period, it is something to work with.

    The 'Scotese graph' (which in other versions does label CO2 concentrations) shows some variations that are likely to be meaningful and sorta accurate. My favorite is the CO2 crash during (coal-forming) Carboniferous, which was followed by an ice age.

    The CO2 increases at Permian Triassic boundary and during the Eocene are reasonably well established, but missing from the Scotese graph. These contribute to the apparent lack of correlation in the presentation.

    In summary, it is far from perfect, and pretty much can't be perfect for this entire swath of time. But again you will find many other more valuable things on that website.

    +++

    Besides the ENSO cycle doing what it does to air T, it causes (or is at least highly correlated) with terrestrial carbon cycling. 1998 was really bad in terms of photosynthesis, which was very strong during La Nina in 2011. I cannot report that the terrestrial carbon chasers understand that very well. But it is something that the earth system does, and we'd better try to understand it. Weird though that the 2011 La Nina did not reduce air T.