Options

Chaos considered harmful ?

An event going on called the Rotman Institute Conference on Knowledge and Model in Climate Change http://www.rotman.uwo.ca/videos/

Very interesting reading some of the tweets:

Steve Easterbrook @SMEasterbrook : Fleming: The butterfly effect is misnamed. Lorenz knew the perturbation would have to be really big. Better label: Mothra effect #Rotman2014

Gavin Schmidt ‏@ClimateOfGavin : @smeasterbrook not sure I actually agree with this though. In GCMs smallest possible changes have same effect.

It goes on from there. Read the tweet conversation here: http://twitter.com/SMEasterbrook/status/525657433780535296

My take is that this speaker Jim Fleming suggested the idea that the chaotic models of climate as originally proposed by Edward Lorenz are not as chaotic as people think. Easterbrook interpreted that by stating that a butterfly was to weak a forcing to be able to change anything, and something more akin to Mothra (a gigantic SciFi moth) was needed to change the trajectory on climate.

I think that there are probably a couple of scales that we need to consider. Events such as hurricanes are likely unpredictable, but they are really inconsequential when compared to the largely deterministic trajectories of phenomena such as ENSO. Same with CO2, as that is a Godzilla of a forcing.

Gavin Schmidt on now!

Comments

  • 1.

    Gavin Schmidt said this during his talk with regard to natural variability and variance in climate models

    "No climate model can be true."

    "No physical model of the real world can be true."

    (Paraphrased from memory)

    Then he rhetorically asked:

    "Why do climate modelers pursue this endless task of increasing complexity, instead of simple energy balance models?"

    The answer he gave is because simple models do not include all the detailed factors.

    " El Ninos and La Ninos are random. ... ENSO can not be predicted more than 6 months in advance."

    Which indicates just how challenging the El Nino prediction project is.

    During the Q&A he answered a question with this statement:

    "If you were to try to 'game' the system, you would fail miserably"

    He also said that energy balance models are much more stable over time in comparison to GCM's, which evolve rapidly over generations.

    Also that credibility of models is very important in that they should be able to hindcast to prove their value.

    There were a grand total of 12 viewers of the streaming video at the end !

    Lots of red meat for Azimuthers to chew on.

    Comment Source:Gavin Schmidt said this during his talk with regard to natural variability and variance in climate models >*"No climate model can be true."* >*"No physical model of the real world can be true."* (Paraphrased from memory) Then he rhetorically asked: >*"Why do climate modelers pursue this endless task of increasing complexity, instead of simple energy balance models?"* The answer he gave is because simple models do not include all the detailed factors. >*" El Ninos and La Ninos are random. ... ENSO can not be predicted more than 6 months in advance."* Which indicates just how challenging the El Nino prediction project is. During the Q&A he answered a question with this statement: >*"If you were to try to 'game' the system, you would fail miserably"* He also said that energy balance models are much more stable over time in comparison to GCM's, which evolve rapidly over generations. Also that credibility of models is very important in that they should be able to hindcast to prove their value. There were a grand total of 12 viewers of the streaming video at the end ! Lots of red meat for Azimuthers to chew on.
  • 2.
    edited October 2014

    I'm glad I looked at the tweet log as I hadn't come across this ENSO paper.

    Wittenberg et al., ENSO Modulation: Is It Decadally Predictable? (2014)

    The abstract says:

    These 40-member reforecast ensembles display potential predictability of the ENSO trajectory, extending up to several years ahead. However, no decadal-scale predictability of ENSO behavior is found. This indicates that multidecadal epochs of extreme ENSO behavior can arise not only intrinsically but also delicately and entirely at random. Previous work had shown that CM2.1 generates strong, reasonably realistic, decadally predictable high-latitude climate signals, as well as tropical and extratropical decadal signals that interact with ENSO. However, those slow variations appear not to lend significant decadal predictability to this model’s ENSO behavior, at least in the absence of external forcings.

    What on earth does "delicately" mean? Are they saying both that high-latitude phenomena interact with ENSO and have detectable signals but have no predictive power?

    It would be interesting to know what Gavin Schmidt made of the Ludescher et al. paper as he's just tweeted that ENSO is not predictable more than 6 months in advance.

    Comment Source:I'm glad I looked at the tweet log as I hadn't come across this ENSO paper. Wittenberg et al., [ENSO Modulation: Is It Decadally Predictable? (2014)](http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-13-00577.1) The abstract says: > These 40-member reforecast ensembles display potential predictability of the ENSO trajectory, extending up to several years ahead. However, no decadal-scale predictability of ENSO behavior is found. This indicates that multidecadal epochs of extreme ENSO behavior can arise not only intrinsically but also delicately and entirely at random. Previous work had shown that CM2.1 generates strong, reasonably realistic, decadally predictable high-latitude climate signals, as well as tropical and extratropical decadal signals that interact with ENSO. However, those slow variations appear not to lend significant decadal predictability to this model’s ENSO behavior, at least in the absence of external forcings. What on earth does "delicately" mean? Are they saying both that high-latitude phenomena interact with ENSO and have detectable signals but have no predictive power? It would be interesting to know what Gavin Schmidt made of the Ludescher et al. paper as he's just tweeted that ENSO is not predictable more than 6 months in advance.
  • 3.

    Jim, I think "delicate" may mean sensitive to initial conditions and parameters, likely in the sense of a butterfly flapping its wings -- which is what they are debating in the twitter-storm.

    I am still amazed by how much the QBO controls the peaks of ENSO, which makes it a forced stimulus. In contrast, Wittenberg et al think it is all intrinsic, and so presumably emergent.

    This is definitely a major difference in modeling of ENSO.

    Comment Source:Jim, I think "delicate" may mean sensitive to initial conditions and parameters, likely in the sense of a butterfly flapping its wings -- which is what they are debating in the twitter-storm. I am still amazed by how much the QBO controls the peaks of ENSO, which makes it a forced stimulus. In contrast, Wittenberg et al think it is all intrinsic, and so presumably emergent. This is definitely a major difference in modeling of ENSO.
  • 4.

    Jim emailed you the English text, no images found nor pdf.

    Basically I cannot follow the paper, I assume he is running a simulator simulating a model. I also cannot see how the results of the paper could be duplicated. I cannot infer any of the conclusions.

    If the cornerstone of scientific investigation is ability to duplicate a result from a set of experiments (or data), no science was endeavoured.

    Dara

    Comment Source:Jim emailed you the English text, no images found nor pdf. Basically I cannot follow the paper, I assume he is running a simulator simulating a model. I also cannot see how the results of the paper could be duplicated. I cannot infer any of the conclusions. If the cornerstone of scientific investigation is ability to duplicate a **result** from a set of experiments (or data), no science was endeavoured. Dara
  • 5.

    Thanks Dara,

    Now I'm just short of the supplementary info for the 2 AMU papers Nathan cited.

    Comment Source:Thanks Dara, Now I'm just short of the supplementary info for the 2 AMU papers Nathan cited.
  • 6.

    Based on what we are trying to do with the El Nino project, if someone says that a behavior occurs "entirely at random" it means that they have essentially punted on trying to predict anything with other than a broad probability measure.

    So we go with a causal and more-or-less deterministic mechanism and see how far it will take us. It may work, it may not, but it is worth the risk.

    Comment Source:Based on what we are trying to do with the El Nino project, if someone says that a behavior occurs *"entirely at random"* it means that they have essentially punted on trying to predict anything with other than a broad probability measure. So we go with a causal and more-or-less deterministic mechanism and see how far it will take us. It may work, it may not, but it is worth the risk.
  • 7.
    edited October 2014

    Based on what we are trying to do with the El Nino project, if someone says that a behavior occurs “entirely at random” it means that they have essentially punted on trying to predict anything with other than a broad probability measure

    I disagree Paul, people say those things when they do not know how to solve a complex problem. His comment: entirely at random is what my mother says when I visit her in Toronto and she complains about the weather changing so fast! In other words that is a layman's term.

    Paul we are wasting our time reading bogus paper after bogus paper filled with non-technical verbiage. What you are doing is sound, code algorithms plots numerical analysis... and hopefully our fearless leader soon has more time to give us some theoretical framework to start formulating and solving some problems that relate to planetary climate.

    Dara

    Comment Source:>Based on what we are trying to do with the El Nino project, if someone says that a behavior occurs “entirely at random” it means that they have essentially punted on trying to predict anything with other than a broad probability measure I disagree Paul, people say those things when they do not know how to solve a complex problem. His comment: **entirely at random** is what my mother says when I visit her in Toronto and she complains about the weather changing so fast! In other words that is a layman's term. Paul we are wasting our time reading bogus paper after bogus paper filled with non-technical verbiage. What you are doing is sound, code algorithms plots numerical analysis... and hopefully our fearless leader soon has more time to give us some theoretical framework to start formulating and solving some problems that relate to planetary climate. Dara
  • 8.

    I disagree Paul, people say those things when they do not know how to solve a complex problem.

    Dara, That is close to what I meant when I said "punted". In the game of football, when you punt you are giving up. To punt on first down, you are really giving up on solving the problem without even trying. To be charitable, and w/o reading the paper, they probably at least tried.

    Comment Source:> I disagree Paul, people say those things when they do not know how to solve a complex problem. Dara, That is close to what I meant when I said "punted". In the game of football, when you punt you are giving up. To punt on first down, you are *really* giving up on solving the problem without even trying. To be charitable, and w/o reading the paper, they probably at least tried.
  • 9.

    Paul sorry heh heh heh.

    I do not know what is the value for publishing a paper when nothing new is done, no positive results and whatever the author says cannot be duplicated nor verified.

    These are superstitions about climate

    Dara

    Comment Source:Paul sorry heh heh heh. I do not know what is the value for publishing a paper when nothing new is done, no positive results and whatever the author says cannot be duplicated nor verified. These are **superstitions** about climate Dara
  • 10.

    I call these types of arguments with nothing to really back them up just-so stories. One can elevate them to a hypothesis, but unless someone has a way to verify them, they are kind of worthless.

    There is a discussion going on at WUWT today, started by meteorologist Joe Bastardi. Everyone is chipping in with their own opinion on whether the Kelvin wave is fast enough, or whether the warm water is upwelling, or whether the winds are strong enough, etc, to generate an El Nino.

    Yet, unless someone has a mathematical model that ties the pieces together, it is just armchair quarterbacking. The grizzled weather guys, such as Bastardi, have some intuition based on how often they study the data, but --- because they can't express their insight in terms of an algorithm --- it is kind of a useless exercise. They take a risk in predicting an El Nino, because they know that if they get it right, their consulting services will become much more valuable the next time around.

    Consider that the famed hurricane predictor, William Gray, was one of the first to make the connection between QBO and ENSO. Yet, after all this time, he never came up with any type of algorithm. Gray is now a global warming skeptic. Neat if we can use the QBO as part of an algorithm to predict ENSO. It gives one extra motivation :)

    Comment Source:I call these types of arguments with nothing to really back them up [just-so stories](http://en.wikipedia.org/wiki/Just-so_story). One can elevate them to a hypothesis, but unless someone has a way to verify them, they are kind of worthless. There is a discussion going on at [WUWT](http://wattsupwiththat.com/2014/10/26/yes-virginia-and-everyone-else-there-is-an-el-nio-coming/) today, started by meteorologist Joe Bastardi. Everyone is chipping in with their own opinion on whether the Kelvin wave is fast enough, or whether the warm water is upwelling, or whether the winds are strong enough, etc, to generate an El Nino. Yet, unless someone has a mathematical model that ties the pieces together, it is just armchair quarterbacking. The grizzled weather guys, such as Bastardi, have some intuition based on how often they study the data, but --- because they can't express their insight in terms of an algorithm --- it is kind of a useless exercise. They take a risk in predicting an El Nino, because they know that if they get it right, their consulting services will become much more valuable the next time around. Consider that the famed hurricane predictor, [William Gray](http://en.wikipedia.org/wiki/William_M._Gray), was one of the first to make the connection between QBO and ENSO. Yet, after all this time, he never came up with any type of algorithm. Gray is now a global warming skeptic. Neat if we can use the QBO as part of an algorithm to predict ENSO. It gives one extra motivation :)
  • 11.

    With the advent of GPM and TRMM satellites networks and the vast volumetric data collected in real-time from atmosphere and upper atmosphere, these guys are all already laid off persona non grata, the real deal will be the algorithms, software and parallelization and we all agree to coronate John as our king to lead us to sound theories and mathematical framework so we could develop code against.

    Dara

    Comment Source:With the advent of GPM and TRMM satellites networks and the vast volumetric data collected in real-time from atmosphere and upper atmosphere, these guys are all already laid off **persona non grata**, the real deal will be the algorithms, software and parallelization and we all agree to coronate John as our king to lead us to sound theories and mathematical framework so we could develop code against. Dara
Sign In or Register to comment.