- JoNova - https://joannenova.com.au -

Zombie Hockey stick dies again

Just when you think it’s too dead to kill: along comes a new paper in a top ranking statistics journal by McShane and Wyner. It’s worth taking stock. It’s a damning paper:

…we conclude unequivocally that the evidence for a ”long-handled” hockey stick (where the shaft of the hockey stick extends to the year 1000 AD) is lacking in the data.



But in the big scheme of things the Hockey Stick Graph was already dead.

Each one of these points is enough to cast grave doubts on the Hockey Stick.

  1. The Hockey Stick uses the wrong type of proxy – tree rings. Trees grow faster when it’s warmer, and when it’s wetter, or when the tree next-door falls down and a herd of manure-making cows move in. Almost all other types of proxies disagree (like ice cores, ocean sediments, corals, and stalagmites). Over 6000 boreholes, hundreds of studies, as well as recorded history show the world was warmer 1,000 years ago. (See here for the refs.)
  2. Even among tree rings, the Hockey Stick uses the wrong type of tree – Bristlecone pines – which appear to grow faster as CO2 rises, regardless of the temperature.
  3. It uses the wrong type of averaging. The Principal Component Analysis (PCA) was centered over the last 150 years instead of the entire millenia. McIntyre and McKitrick showed that this would produce a hockey stick even if it were fed random numbers instead of tree ring data.
  4. The data is massively incomplete, spatially autocorrelated, the signal is weak, and the number of covariates greatly outnumbers the independent observations.
  5. The data has been calibrated with a short period of temperature records that are themselves substantially processed with smoothing, adjustments, discontinuities and imputation of missing data, all of which may introduce errors.
  6. Assuming that tree rings are not so bad, that bristlecones are not misleading, and that the calibration data is not in error, McShane shows that during the last 150 years the most random of “fake” data (white noise and brownian motion) has more predictive ability than the proxy data, and that uncertainties are huge, and neither real (nor fake data) has any meaning over the last 1000 years.

McShane Wyner 2010 The Hockey Stick Graph reanalyzed


It is not clear that the proxies currently used to predict temperature are even predictive of it at the scale of several decades let alone over many centuries.

As I said back in December, one of the landscape shifting forces about ClimateGate was that it suddenly motivated skeptics, imbuing them with conviction and energy, because it triggered off the universal warning lights that the freeloaders and parasites were at work, playing on our good intentions. Prior to this many smart incisive people were busy and otherwise occupied. Now these movers and shakers are being pulled into the debate, and the climate establishment can no longer get away with their schtick.

Because there were no Climate Science au Naturale Institutes that had an interest in busting the CAGW hypothesis, inept, inadequate work which supported the fashionable theory was allowed to stand for years. The institutes whose funding depended on the Big Scare Campaign did not hire expert statisticians to try to check their own work, and expert statisticians, for the most part, had other things to do than to check papers in an obscure branch of science. Determined volunteers like Steven McIntyre moved in, exposing major flaws, and now finally the world of expert statistics is waking up to the fact that there are nice openings thank-you-very-much for posting papers in top journals, just be redo-ing the work of climate scientists.

Who knows exactly what motivated McShane and Wyner, or when they started investigating, but the arrival of other science related experts into this debate is not a day too soon.

For ten years mediocre scientists have been able to get away with poor work, and have been given the red-carpet (and the odd Nobel Peace Prize). That time is over.

McShane:
“The panel found that the statistical tools that CRU scientists employed were not always the most cutting-edge, or most appropriate. “We cannot help remarking that it is very surprising that research in an area that depends so heavily on statistical methods has not been carried out in close collaboration with professional statisticians,” reads the inquiry’s conclusions.

So it goes, McIntyre, McKitrick,and now McShane: Perhaps one day someone will figure out why the Scottish Y chromosomal inheritance is so adept and determined at statistically destroying government funded science. (No doubt Monckton-the-Scot has a view on that.)

McShane, B.B. and  Wyner, A. J.: A Statistical Analysis Of Multiple Temperature Proxies: Are Reconstructions Of Surface Temperatures Over The Last 1000 Years Reliable? Submitted to the Annals of Applied Statistics

7.1 out of 10 based on 8 ratings