To AKRONOS Main Page
To the top of  'Global Warming': An Official PseudoScience


4.  A modeller's myth of oceanic warming: how the new (pseudo)science is secreted

Let us start by considering the oceans.  The latest weapon in the arsenal of 'global warming' advocates is a study by Syd Levitus' group at NOAA's NODC (National Oceanographic and Atmospheric Administration's National Oceanographic Data Center), which supposedly established the 'fact' that 'the oceans are warming', irrespective of its cause.  Levitus et al had previously built an extensive ocean database where yearly data from 1948 to 1998 was logged after an 'objective grid analysis' [9].  In the 2000 study, they admit that of the data they employed for values of heat content from sea surface to a depth of 300 meters, the most reliable only dates from the period following 1975, when real-time reports of data logs were initiated [10].  The determinations are made only in terms of heat content, likely so that standard errors appear smaller by attribution to supersets of data sets, rather than to actual sets of raw data.  No actual salinity data is employed.  In the same 2000 study, the entire set of results presented for depths from sea surface down to 3000 meters is a simulation, an estimate based on arbitrary 5-year running composites, which, in the authors' own words, was made "necessary due to the lack of deep-ocean observations" [10].  By any measure of the scientific method, this could never be mistaken for scientific results, for empirical data.  It could only be considered for what it is, an hypothesis that 'massages' data, even fabricates it, in order to construe a rationale that justifies supposition of the starting hypothesis.  The simple truth is that no scientist  can suppose the hypothesis proven by virtue of the 'massage'! And since the hypothesis remains to be proven, it is not yet proven, and should never have been published as a 'finding'!  That such could happen must be credited to the corrupt nature of our social institutions  - from the 'ecology shows' to the UN-affiliated NGOs, from the oil-industry to regulatory mechanisms - and the spread of this corruption to science itself. 

Could it be that only science that is incomplete, and remains axiomatic, is susceptible to official appropriation? Could it be that the greater part of Official Science is corruptible science? That only science which is susceptible to falsification, and strays so far from the scientific method that it settles for systematic inconsistency, can be official? 

With no real science to back them up, Levitus et al claim that from 1955 to 1995 the 'evidence' demonstrates a net warming of 0.06 deg C down to a depth of 3,000m, and a mean increase of 0.31 deg C down to a depth of 300m depth.  Taking at face value their interpretation of the treated data - that during the years 1985 to 1998 there has been a net warming trend of seawater (they actually claim in the paper's discussion that the trend began in the 1970's for surface sea water, though none of the present trends shown in their data - in the Atlantic, Indian and Pacific oceans - began in the 1970's...) - what can one conclude?  One can conclude what one should by looking at their own data (Fig. 1 of their paper) - namely, that the data for all three oceans shows annual and quasi-periodic, supra-annual oscillations in cooling and warming.  The latter oscillations appear to be decadal, or even hemidecadal, likely also with patterns that are longer term than decades.  Little can be said about these patterns - and thus also about long-term trends predicated on the moving averages of these oscillations - since Levitus et al failed to make any correlation with periodic variations in solar radiation, such as the sunspot cycle, etc.  But the patterns are there, showing that long-term variations in ocean temperature, and oscillations of its thermal content, alternate over the period of the study.  

In fact, conventional climatology accepts that between 1920 and 1940 there was a bidecadal period of surface ocean water cooling, followed by a decadal period of warming, and then - from 1947 to 1967-1970 - another period of cooling.  And Levitus et al's data itself, on all three ocean panels, shows that while some warming occurred between 1967 to 1977 or so, there was equally a cooling from 1977 to 1984 or so.  They passed by these oscillations in silence, and instead concentrated on showing what appear to be impressive 'trends' on the simulated data for the greater depth of 3000m (their Fig. 4).  Any decent reviewer should have asked for the physical rationales for constructing such conveniently smooth composites.  The rationale for a statistical construct cannot be merely the invocation of a statistical technique.  Nowhere do Levitus et al provide a physical and chemical reasoning for their constructs.  But what is more important, they fail to test for simple physical correlations - such as correlation of temperature variations against cloud cover, salinity, or, more importantly still, solar periodicities (eg the Gleissberg, 22- and 11-year cycles of sunspot activity; the long and short solar orbital periods on the plane of the ecliptic; the variations in the solar constant, in particular, those affecting atmospheric UV photon production; the variation in speed and composition of solar winds).  But aside from such tremendous shortcomings or effective cop-outs, there is a simple fact that should be staring every scientist in the face: if the data can be trusted, it presents the record of a series of oscillations, and any mean taken at any arbitrary point in time cannot be extrapolated linearly into the future (or into the past, for that matter), since it is part of an oscillation or an oscillatory train of variations.  A sample of 50 years suffices to identify a small set of quasi-periodic oscillations, but the mean value of the oscillation about the mean, either way (for warming or cooling), even if it is said to be an anomaly ("the anomaly in heat content"), cannot be taken - with any scientific legitimacy - as a basis for a projection that increases the "anomaly" indefinitely at the same rate of warming (or cooling), because of some out-of-the-blue postulate of a "positive feedback mechanism"...

The real 'anomaly' is the nonsense of the reasoning that extrapolates a global trend from a periodic oscillation.  Even the overall mean temperature of reference is a relative mean, dependent upon the size of the sample, which is merely the same 50 year period... 

Yet, a paper with simulated data from scientific technobureaucrats of a government service so pleased their masters at the IPCC, the UN CLIVAR program and the US NRC, that it passed peer-review - during which, incidentally to our subject, none of these objections were raised - and was published in Science... Mar-vel-lous.

To further underline just what faith the raw data itself deserves, the oceanographer Robert E. Stevenson describes his early days at the mechanical job of data collection - by dipping a bucket into the sea water from a travelling ship, to take a temperature reading:

"Most of the thermometers were calibrated into 1/4 degrees Fahrenheit.  They came from the US Navy.  Galvanized iron buckets were preferred, mainly because they lasted longer than the wood or canvas ones.  But they had the disadvantage of cooling quickly in the winds, so that the temperature readings needed to be taken quickly.  I would guess that any bucket temperature measurement that was closer to the actual temperature by better than 0.5 deg C was an accident, or a good guess. (...) The archived data used by Levitus, and a plethora of oceanographers, were taken by me, and a whole cadre of students, postdocs, and seagoing technicians around the world." [11]

And, as if this weren't appalling enough, there are the further unaddressed problems of the limitations in the resolution of those thermometers and thermistors, the proper calibration of the latter, their drift and noise...  Science is method, or it is nothing but any odd lump sum.

Not to be outdone, the same group took the next step in another Science report, in 2001 [12].  There they categorically asserted that the Earth has increased its heat content in the atmosphere (all the way to the stratosphere), in the oceans, and in the cryosphere, and that this effect is anthropogenic.

What is the proof?  The best fit of a modelled curve to a pre-agreed curve tracking a cumulative global warming... Even a doodling grocer coming out of Cambridge or Oxford can become a scientist. 

Present-age of climatology is filled with mathematical gimmicks or 'computer games', called models, which are axiomatic and self-validating by a vast number of corrective, 'normalizing', 'fitting', 'infilling' procedures which they include in order to fit a set of overprocessed 'data' to the desired result.  The precarious epistemological situation of climatology only worsens when such models are built with 'treated' data, and the whole procedure justified by the poverty of real data.  In a model, an algorithm takes over which weighs a variable number of variables ('forcings').  The selection of such forcings is, itself, a matter of contention.  Irrespective of that, the data is further 'processed' (blanched, laundered).  In the case of the 2001 Levitus at al paper, the proof for the correlation of ocean warming and global atmospheric warming is obtained in reverse, as the fit between a modelled curve that integrates all the separate ocean 'results' of the 2000 study for ocean warming down to a depth of 3,000m, and a modelled curve for the effect of one or several 'forcings' related to atmospheric parameters (greenhouse gases and sulfate aerosols) predicted by the global warming climate model. The data of the 2000 paper was processed into 5-year running averages (the points plotted), using a 'smoothing' method (or 'Einstein-fudge factor') which is supposed to 'compensate' for the sparseness and the 'intrinsic noise' (sic) of the data. The best fit found by the Levitus group included 'forcings' representing the greenhouse gases, sulfate aerosols, volcanic aerosols, and solar variability (perturbations in solar irradiance, which is assumed to be constant otherwise).  The authors conclude that their study "is evidence" that the warming of the terrestrial surface is of anthropogenic origin and due to the increase in greenhouse gases and sulfate aerosols...

One might wonder where the data is - the real measurement data, down to 3000m into the deep ocean; or the data for correlation, or lack thereof, with solar processes; or the data that justifies the assumptions intrinsic to the 'global warming' climate model, including those predicating it upon an increase in greenhouse gases.  Where is it?  Thousands of papers have been published, yet so very few are worth citing. 

That the results of modelling, or even simulation, can be treated as if they could replace and equate to empirical data and the product of empirical methodologies, is a still more dangerous corollary of the myth of 'global warming'.  It presents us, in fact, with a 'retrograde paradigm shift' brought about by the unquestioned reign of computer modelling in climatology (or that of String Theory in physics, and so on).  It is the price paid for the new hegemony of 'sophisticated' manipulation of data, and the inscrutable rule of technobureaucratic cadres.  With the right selection, any fact can be omitted from existence and replaced by any of the imaginaries of modelling...  This formalistic turn of Official Science may well be signaling the end of science - at least the point at which Official Science is no longer able to absorb science itself, because it ceases having a handle on what is or is not science.

It is modelling which is a forcing, and not a reliable one.  As reliable as the predictions of the path and intensity of hurricane Katrina before it hit Louisiana, or the effectiveness of FEMA in helping people in the aftermath.  One cannot apply the term "science" to that which lacks data but asserts its hypothesis as proven by the mere fact that the hypothesis itself can be enunciated.  An hypothesis, no matter how 'neat', must be proven experimentally, by actual observations and measurements, before being considered "proven", or even on its way to being "proven".  Even then, it is most likely destined to be disproven later by a better and finer hypothesis.  But all this, the very method of science, has now been discarded by the apologists of 'global warming' and the high priests of Official Science.

In passing, we should remark how the language of these climate modellers is the technocratic language of the neutered and the inverted: simulations that extrapolate a model become a "control run" (!); filling in missing data becomes "smoothing" (one of the 'great evils' of modern science); statistical procedures are employed as if they were equivalent to empirical investigation and methodologies, and are referred to by such euphemisms as "optimal detection methodology"; long-term variations in temperature are called 'anomalies' with respect to operational means extracted from small sets; the "heat increase predicated on global warming", becomes the "observed heat gain".  And so on.

Next:  Where is the evidence for global atmospheric warming?
Previous:  The rationale for 'global warming': manufacture of a global consensus