Poseidon Adventure

A paper is getting a lot of visibility on delusionist websites these days, Suggestive correlations between the brightness of Neptune, solar variability, and Earth’s temperature (Hammel & Lockwood 2007, Geophysical Research Letters, Vol. 34, pg. L08203). The visibility given this paper by delusionists is part of the general trend to claim that global warming is happening on Mars, Neptune, Jupiter, some of the moons of the giant planets, Halley’s comet, the Apollo theater, and of course the seventh circle of hell — so it must be due to the sun, and so too, must be the global warming on earth.

It’s my opinion that the paper by Hammel & Lockwood should not have been published. I have two reasons for this opinion. First, the “suggestive correlations” aren’t correlations at all. Hammel & Lockwood quote correlation coefficients between total solar irradiance (TSI) and earth’s global average temperature of r=0.91, and between Neptune’s brightness and earth’s temperature of |r|>0.9. Much of the reason for the high correlation coefficient is that their source for TSI data is the estimate of Foukal (2002, A Comparison of Variable Solar Total and Ultraviolet Irradiance Outputs in the 20th Century, Geophysical Research Letters, vol. 29, pg. 2089). According to the Foukal reconstruction, TSI showed a sizeable increase from about 1982 to 1990, of 0.017%, which amounts to an increase of fully 2.3 Watts per square meter.

However, the authors themselves admit that “Unfortunately, none of these correlations is statistically significant.” In other words, the only fact which the authors establish is that the graphs look similar, but when you run the numbers the “suggestive correlations” don’t pass muster. The same can be said about the graphs of earth’s temperature, and of worldwide beer sales.

We all know that “correlation does not mean causation” — just because two variables are correlated, doesn’t mean that one is the cause of the other. But Hammel & Lockwood, although never explicitly stating it, clearly imply that there is a causal relationship between TSI and earth’s temperature, as well as Neptune’s brightness. But if “correlation does not mean causation,” then what does the lack of correlation mean?

My other reason for thinking this paper should never have seen the light of day, is that it uses data for solar irradiance that is out of date, and has been surpassed by other data sets. They use TSI data from Foukal; I don’t have a problem with the Foukal paper, in its day it was an honest attempt to estimate TSI. But that was five years ago; the Hammel & Lockwood paper was published this year, and they could easily have found far better TSI data.

The “gold standard” these days for TSI before the satellite era is Lean (2000, Evolution of the Sun’s Spectral Irradiance Since the Maunder Minimum, Geophysical Research Letters, Vol. 27, pg. 2425), and the Lean data are freely available from the internet, having been updated as recently as 2004. If the Hammel & Lockwood paper were from 2002, I could understand considering Foukal to be a credible reconstruction. But we have recently benefited from satellite measurements of TSI, and lo and behold, the rise in TSI from about 1982 to 1990 suggested by Foukal is not supported by the satellite data. But the reconstruction of Lean is in outstanding accord with satellite measurements.

The satellite measurements have made it abundantly clear that the TSI reconstruction of Lean might actually be right — but the reconstruction of Foukal (2002) is certainly wrong. There is nothing like a 2.3 W/m^2 increase in TSI from 1982 to 1990; the only significant change is the variations due to the solar cycle, with full amplitude of only 1 W/m^2, and the average value of TSI shows no real change at all.

Furthermore, the Foukal 2002 reconstruction only goes barely past 1990. Why on earth would Hammel & Lockwood use data that only go a little beyond 1990, when Lean’s reconstruction goes beyond 2000, and easily downloadable satellite data go well beyond that? Did they not bother to look for more up-to-date data? Are they unaware of satellite measurements of TSI? I’m not even working in this field, and I’ve been aware of more current, and more correct, TSI data for over a year. Certainly Foukal is aware of satellite measurements, and his most current research makes use of these data. And even if you’re just a blogger like me, you don’t need anything more than google to find better, more up-to-date data than used by Hammel & Lockwood.

It seems to me that there are two possibilities. One, Hammel & Lockwood are astoundingly naive when it comes to TSI measurements, and too lazy to bother using google. Two, Hammel & Lockwood want to suggest that global warming on earth is due to the sun, so they deliberately avoided using the data which contradict this idea — despite the fact that it’s clearly superior to the data they chose.

If their purpose was to provide more fodder for global warming delusionists, then “mission accomplished.” If their purpose was to shed light on possible relationships between solar irradiance, earth’s temperature, and Neptune’s brightness, then they have failed utterly.