Scafetta and Willson: “the ACRIM composite as the most likely and precise representation of 35 years of TSI monitoring”

Nicola Scafetta and Richard Willson have a new paper in press which contains the most thorough analysis yet of the intercomparison of the empirical ACRIM and modeled PMOD TSI series. It’s a comprehensive yet readable paper of high interest to all diligent climate researchers interested in determining the relative strengths of various climate drivers. It is also an important historical document for philosophers of science investigating the shift from observation based empirical solar science to model based  dogma underpinning preconceptions of the power of trace gases to control Earth’s surface temperature. The IPCC and Team Wassup’s Leif Svalgaard are not going to like it, and will therefore try to ignore it, thus further underminng their credibility.

ACRIM total solar irradiance satellite composite validation versus TSI proxy models
Nicola Scafetta & Richard C. Willson

From the paper:

PMOD TSI composite (Fröhlich and Lean 1998; Fröhlich 2004, 2006, 2012) is essentially a theoretical model originally designed to agree with Lean’s TSI proxy model (Fröhlich and Lean 1998). It relies on postulated but experimentally unverified drifts in the ERB record during the ACRIM Gap,and other alterations of the published ERB and ACRIM results, that are not recognized by their original experimental teams and have not been verified by the PMOD by original computations using ERB or ACRIM1 data.

Our findings support the reliability of the ACRIM composite as the most likely and precise representation of 35 years of TSI monitoring by satellite experiments. The only caveat is that the ERB record prior to 1980 may require some correction for degradation, but available evidence indicates that it would be much less than used in the PMOD composite.

We argued that the ACRIM composite most closely represents true TSI because the very corrections of the published TSI data made by Fröhlich to construct the PMOD composite are not supported by a direct comparison between ERBE and ERB records in the proximity of September/October 1989.

Direct comparison of ERB and ERBE during 1989 showed that Fröhlich’s postulated Sep/29/1989 step function increase of ∼0.47 W/m2in ERB sensitivity, which coincided with a power down event, did not occur. The KBS07 proxy model does not support Fröhlich’s ERB ‘glitch’ hypothesis either. A divergence between the two satellite records did occur in November 1989; but this is more than one month later and clearly not associated with the ERB end-of-September power down event.

We have demonstrated that the update of Lean’s TSI proxy model (Kopp and Lean 2011), used originally to validate PMOD’s lack of trending from 1980 to 2000 (Fröh- lich and Lean 1998), has inadequate predictive capability to properly reconstruct the TSI decadal trending. Lean’s model predicted an upward trend between the TSI minima in 1996 and 2008 while both ACRIM and PMOD present a downward trend.

This demonstrates that Lean’sproxy model cannot reconstruct TSI decadal trending with a precision smaller than ±0.5 W/m2 the same order as the difference between PMOD and ACRIM TSI composites. Thus, Lean’s TSI proxy model is not useful as a guide to correct satellite measurements. The WSKF06 TSI proxy models contradict the primary PMOD rationale by the following findings:

(1) there was a TSI peak in late 1978 and early 1979 as recorded by ERB(although some early mission degradation of the instrument may have been uncompensated for);

(2) the ACRIM1 published record is more stable than ERB during 1980– 1984 and should be preferred for constructing a TSI composite during this period;

(3) ERB did not experience either the end-September 1989 step function drift in sensitivity or the upward linear drift claimed by Fröhlich during 1990–1992.5. The latter result is also evident in the upgraded SATIRE model (Ball et al. 2012).

Thus, if ERB requires some correction during the ACRIM Gap, our results suggest that Fröhlich overestimated those corrections by at least a factor of two due to the fact that at least one of the two hypotheses (the ERB glitch in Sep/29/1989 or the ERB drift from Oct/1989 to 1992) are not confirmed by our cross-analysis. The ERB-ERBE divergence during the ACRIM Gap most likely resulted from uncorrected degradation of ERBE in its first exposure to short wavelength fluxes driven by enhanced solar activity during the 1989–1993 solar maximum or other events.

Consequently PMOD should be shifted upward by about 0.5 W/m2after 1992 which produces a 1980–2000 TSI upward trending similar to that observed in the ACRIM composite. Our results demonstrated that the validity of TSI proxy models should not be overestimated since they frequently produce conflicting results and contradictory features.

Although the TSI proxy models proposed by the Solanki’s science team (KBS07 and WSKF06) appear to reproduce the lack of a trend during solar cycles 21–22 in the PMOD TSI composite, they contradict one or more of the hypotheses advocated by PMOD to alter the originally published TSI used in constructing the PMOD TSI composite. Thus some of the arguments used to promote the PMOD composite (Fröhlich and Lean 1998; Wenzler et al. 2006; Krivova et al. 2007; Wenzler et al. 2009) are little more than speculations and coincidences.

Moreover, the TSI models proposed by Lean and by Solanki’s science team differ significantly from the TSI model proposed by Hoyt and Schatten (1993) who constructed a TSI record since 1700 using five alternative solar irradiance proxy indexes—sunspot cycle amplitude, sunspot cycle length, solar equatorial rotation rate, fraction of penumbral spots, and the decay rate of the sunspot cycle.