the Air Vent

Because the world needs another opinion

Ten Unprecedented Points

Posted by Jeff Id on August 28, 2009

1 – Historic temperature reconstrucitons are critical for determining how much energy to put into the all important CO2 mitigation effort. Advocates like to redirect this point when a reconstruction fails quality control.

2 – There are no verified temperature proxies. NOT one proxy for thermometers has been verified to be associated with temperature in an engineering style test. Ice cores, speleothum, tree ring widths, tree ring latewood density, boreholes, pollen, sediments, mollusk shells or historic records. They are un-calibrated and completely unverified and several may be completely unrelated.

3 – Not one reconstruction uses an egineering style verification of method to determine if the math is extracting the signal it’s supposed to. Instead odd and unique methods are created and implemented completely without verification in paleoclimatology.

4 – Historic records show temperatures were warmer in the recent past than today. LINK

5 – Many reconstructions are verified using r correlation values. These only verify the correlation to the most recent actual thermometer based temperature records. Since there are non-stationary variances created by correlation based sorting of proxies, the calibration range has different scale than the historic reconstruction temperature range. This means correlation can only tell if the shape is similar, not whether the scale is similar.

6 – Some papers have estimated that the historic pre-calibration range temperature variance is less than half of the calibration range due to the math used rather than the data. My own work estimates Mann08 historic sigal is 30 percent of actual – we can multiply everything pre-1850 by 3 and get a more reasonable approximation of historic temps.

7 – The point in #6 assumes proxies are perfect and makes no correction for reductions in variance due to some proxies being unrelated to temp. As they are 100% unverified, there is a high degree of certainty that some if not all proxies are actually not temperature related. Still like most, I do believe most have a temp component to them under the massive noise.

8 – Borehole proxy temperature data is taken from actual measurements of the rock temps. The temperature of the rock is assumed to be created from air temperatures over thousands of years. The fact that they are water filled and part of active hydrological systems which carry heat away WITH WATER is completely ignored. Therefore, to use a borehole for historic temp you must first accept that flowing underground water has no effect on the temp data for thousands of years. I suspect it’s a leap of faith that even a suicide bomber would have trouble with.

9 – Proxy linearity to temperature is assumed, again without validation. For instance tree ring widths are first matched to temp in recent times and the whole pattern of growth for the tree is assumed to be linear with temperature for thousands of years. No verification of any kind is made.

10 – Few of the above points are actively disputed in peer reviewed literature. Proxy folk are creating paper after paper with the same methods that caused unprecedented temps, ice, hurricanes and whatever else the authors need to be unprecedented. In blogland AGW advocates refute claims against the hockey sticks by saying so many other studies got the same result. We the non-sheep are stuck saying, um….. they used the same kind of math.

All these papers saying ‘whatever’ factor is unprecedented, and completely without verification.

Crazy world.

24 Responses to “Ten Unprecedented Points”

  1. Tony Hansen said

    Say it ain’t so ,Jeff. Say it ain’t so.
    … I just wish we could.

  2. Good point, same assumptions, same mathematical techniques, similar outcomes.

  3. Geoff Sherrington said

    That’s quite a vent.

    In olden times we would first establish a cause-effect relationship, then deduce the maths of the relationship, then look for any extraneous factors, then go to work on applications.

    I have watched with some terror the medical world in particular doing meta statistical studies with often no idea if any extraneous factors are involved or whther the separate studies are capable of being combined. There are some lovely cases of reverse causation. Then they make drugs and I’m a bit sensitive about this because my wife of 45 years has had a couple of quite severe reactions including Lupus, to new drugs that have not been tested with statistical rigour.

    This comes on top of people calling themselves
    climate “scientists” as if you can hang up a shingle with a new discipline etched onto it and expect people to believe that it contains the accumulated wisdom of years of research.

    I can even spell “uniformitarianism”, which is a bit like “stationarity”.

  4. Greg Smith said

    Jeff

    I have followed your rise on the “sphere” with admiration for your consistency and scientific approach to data management. As a geologist, like many of my kind, I am sceptical without being able to do all the maths. Modelling and statistical analysis are black arts like reservoir engineering or seismic processing, motifs that I am familiar with. However it saddens me that those who promote the scientific method such as yourself or Anthony Watts, Lindzen et al are from a generation that actually believed in the scientific approach to theory truth testing (ie with actual measurements). I fear that we may have won the battle in a losing war now that the politicians have committed publically to positions.
    In my home state, Australia, we will have an ETS by November to satisfy the ego of our PM to be a
    “world leader”

  5. Paul said

    Jeff,

    All valid points. To expand a bit on a few of them.

    #8: The deeper you go in the ground the warmer it gets. Deep mines are quite hot. Heat coming up from the interior also messes up the borehole temperature profile.

    #9: Tree ring linearity. Take a look at the picture of Michael Mann on his website holding a polished section of a tree. The progression of thick and thin rings varies as you go around the circumference of the tree. A core taken from the north would give a different series of thick and thin then one from the southeast. As my local tree service guy told me, it depends on how much water the roots to the north and southeast got. Tree rings aren’t thermometers, period.

  6. Jeff Id said

    #4 Thanks for the comments, it’s scary what’s happening to Austrailia, like looking into the future. What is an ETS?

    #5 Paul, that’s what makes some of the tree selections so nuts. Strip bark? Come on, if they wanted to choose a tree which would have minimal historic signal the best way would be to look for one with a great deal of variance circumferentially, i.e. lumpy or damaged trunks, and sample randomly. Then you sort them for upslopes in temperature range throwing out the bad samples and done! Instant unprecedented temperature.

    The fact that sane people are even discussing the possibility that these reconstructions may have validity is beyond my tiny little brain’s understanding.

  7. DeWitt Payne said

    ETS = Emissions Trading Scheme (Scam?)

    And it’s worked so well in Europe!

  8. Jeff Id said

    Ah, thanks.

  9. Raven said

    There is another aspect to these proxies that people overlook: pre-analysis sampling.

    i.e. if a climate scientist goes into the field and collects samples from 10 trees they will only use those samples that ‘look good’ in further analyses. The bad samples are ignored and no effort is put into finding out why the bad samples are bad. If you are lucky there will be some handwaving about non-climatic influences.

    CA has documented this problem with some high profile tree ring series but it would not surprise me to find out that all proxies series suffer from sample selection bias. This, of cource, further undermines the value of these proxies.

  10. Jeff Id said

    #10, I’m suspicious of the latest Mann hurricane paper for this. It demonstrates a hockey stick shock and recovery yet there are only 8 proxies. I haven’t seen if any magic was applied to the proxies yet so it could come from the very heavy math processing of 8 time series but it also could be a mathematical pre-sort which is not discussed in the paper.

  11. Mark T said

    If I’m not mistaken, the 8 proxies only intermittently record hurricanes over the known record, making their use for past records even more tenuous. I could be wrong, but I thought that had already been pointed out.

    Mark

  12. John Nicklin said

    As a biologist, I could tell you that tree ring width is a function of warm/cold, wet/dry, nutrient +/-, and the angle at which the tree grows from the horizontal. If you wanted to use tree rings to compare growing seasons, that would be fine, but to use them as defacto thermometers is a bit of a stretch.

    The beach sedement proxies are not a new thing. I read a paper a couple of years ago describing beach sedement stratification around New Orleans (I beleive) and a few other places that showed evidence of previous storms, but without the conclusions that Mann can up with.

    Proxies may tell us some things about past conditions, but when spliced onto instrument records, the story changes.

    Thanks for the 10 points Jeff, spot on as usual.

  13. Mark T said

    but when spliced onto instrument records, the story changes.

    Yeah, to fraud.

    Mark

  14. Jeff Id said

    #12, Thanks John, the perspective of a biologist in the debate on tree rings is really the key.

    I asked Craig Loehle once if he knew of any other factors in tree wood which might respond to temp or even temp and moisture. Something intracellular for instance. Capillary size, some protein density or something an aeronautical engineer couldn’t hope to discover.

    If we have moisture, and temp with two equations and two unknowns it would potentially help resolve the signal. That would be pretty cool. Temps do affect plants so the signal is in there, but the linearity and separation of multiple variables. Some kind of verification of the signal would be nice.

  15. edaniel said

    There is at least one other critical issue. The Earth’s systems are not now in equilibrium, have never been in equilibrium, and never will be in equilibrium. Equilibrium refers to both internal to each subsystem and between subsystems. This consideration means that the Earth’s systems comprise a heterogeneous thermodynamic system.

    Can the quantity being measured, and measured in only parts of some of the subsystems, reflect any changes in the energy balance of the composite system? Far from equilibrium, as the state of the composite system presently is, the quantity being measured reflects, firstly, the redistribution of the energy content of the subsystems; both within individual and between subsystems.

    Relative to comparisons of the measured quantity with ‘projections’ by models, does the physical quantity being measured correspond to the ‘modeled’ quantity that is expected to exhibit the increasing energy content of the composite system. That increasing energy content being due primarily to the effects of CO2 emissions into the atmosphere.

  16. Francis said

    “NOT one proxy for thermometers has been verified to be associated with temperature in an engineering style test.”

    …Seems so obvious that I’m surprised to see it mentioned. I bet that the earliest thermometers couldn’t pass that test, either.
    I’m new to the proxy issue. And skeptical. I didn’t know that borehole temperatures in rock were being used? I don’t really believe the borehole temperatures in ice.
    I was aware of the Mann hockey stick controversy, but lost interest after the results had been replicated.

    “Still like most, I do believe most (proxies) have a temp component to them under the massive noise”

    Being NAIVE and NON-MATHEMATICAL, I wonder if MORE SAMPLING will eventually settle the matter. Every year there are more proxy papers published. That means many proxies, with more individual samples, in more regions (with or without the southern hemisphere).

  17. Mark T said

    That means many proxies, with more individual samples, in more regions (with or without the southern hemisphere).

    Nope, it doesn’t. They reuse the same series over and over, even after known flaws are pointed out. They also repeat the same mathematical tricks over and over, even after known flaws are pointed out. Sorta makes one sick, actually.

    Mark

  18. Gary P said

    One thing that the trees rings do accurately, is record the growing season. So why don’t these fine dendrochronologists stick to giving a plot of growing season vs. year?
    This would give them less chance to twist and contort the data. It would be interesting to compare growing seasons during the MWP and LIA to recent times. Of course with the CO2 plant food increasing, people might look at the data see things are getting better for the trees.

  19. Jeff Id said

    #16, Very nice post, it’s simple yet makes a very good point. My own work shows that the unverified math is equally as bad as the unverified proxies but people don’t get that. It is my own work though, and no matter how I explain it not everyone will follow. Mann has never criticized my results though.

    For the engineers.

    Those of us forced to use and design instruments can’t even imagine using an unverified source for measurement. We’ve been wrong too often to know something without verification. Any engineer who claims not to have been certain and proven wrong is going to get called out around here. We’ve all been humbled by the god of physics.

    Hahaha, The god of physics name……Perspective.

  20. j ferguson said

    We used to say that Engineering was “where youcouldn’t fool Mother Nature.” If you would bear with me, I’d like to summarize my perception of the Mannian method of proxy/temperature correlation.

    He used a time series of tree ring cores which paralleled a time series of thermometer data. Since there appeared to be little if any correlation between the rings and the temperatures due to what everyone seems to identify as “noise” he ran a routine which identified those specific ring records which did “sort of” correlate (40% of proxies). He then accepted that if 40% of the tree rings “correlated” with temperature, these ring data could be used as proxies as far back as they went – 800 years.

    If only 40% were good in recent years, how did he know which 40% were good in the pre-historic rings?

    If I haven’t misrepresented this, it wouldn’t take more than a junior high school education to wonder why something that only does 40% where you can check it could suddenly be reliable where it can’t be checked.

    It looks to me that tree rings respond to a number of influences and that it is likely impossible to separate temperature out of this very simple indicator.

    What am I missing?

  21. John F. Pittman said

    Jeff, one of the problems with using plants, in general, is that they respire in growing season with a function similar to this growth = f(Ta-Ti)^1.x where Ti is the initial temperature, Ta is a function with the maximum where the maximum is the plant’s best growth temperature such that Ta decends after Tmax is reached. The decline is not necessarily linear. So one does not necesarily get a good U or rather a bell shape (as is assumed by most biologists). The x can be a f(phosphates, nitrates, water, and even elevated CO2). However, keep in mind that unless adapted to wet root conditions, water has a maximum on growth, then it has no effect. Usually nitrates and phosphates are the limiting growth factor. But as someone pointed out, gotta wonder about bighorn sheep dumps in some areas, I imagine.

    THe long and short of it, growth measures, growth best. Anything else is a general indication. There are some very good relationships determined for matching tree-rings with precipitation in the Rockies. However, the one study I read has the same Achille’s heel as temperature studies. They select the trees as hydrometers and throw out the thermometer trees. I think you have shown this can be probelmatic. Though, I believe modern tree ring studies for precipitation use the wood density and do not have to select by post facto re-sampling.

  22. j ferguson said

    “It looks to me that tree rings respond to a number of influences and that it is likely impossible to separate temperature out of this very simple indicator.”

    Simple is the wrong word. Read “very complex but misleadingly simple indicator”

  23. j ferguson said

    Sorry, one more question. After Mann identified the 40% of the proxies that “worked” did he go back and physically or chemically analyze those rings to ascertain how they might differ from the rings which didn’t track?

    More junior high level questions.

  24. Mark T said

    Nope. Post-hoc selection criteria and then declare victory. Outside the calibration interval the verification statistics go to zero.

    Junior high level questions are appropriate because that was the level at which Mann conducted his studies. They hand out PhDs for that now, however, making him an obvious “authority.”

    Mark

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

 
Follow

Get every new post delivered to your Inbox.

Join 147 other followers

%d bloggers like this: