Tuesday, November 27, 2007 ... Français/Deutsch/Español/Česky/Japanese/Related posts from blogosphere

Scafetta & West: Climate phenomenology

In this weekly dose of the peer-reviewed skeptical literature about the climate, we look into Journal of Geophysical Research:

N. Scafetta & B.J. West (click): Phenomenological reconstructions of the solar signature in the Northern Hemisphere surface temperature records since 1600
The use purely statistical tools applied to the correlation between the sun spot number and reconstructed temperatures to argue that most likely, the total solar irradiance had a low secular (=trend-like, non-periodic) variability but it was accompanied by a high preindustrial secular climate variability between 1610 and 2005.

The Sun is a major player

That means that small changes of the Sun can lead to significant changes of the climate. With the preferred magnitude of variability as given by Moberg et al. 2005, one of their conclusions is that that Sun is responsible for roughly 50% of the observed 20th century warming. Not too surprisingly, if they choose Mann 2003 instead of Moberg 2005, the calculated effect of the Sun becomes negligible.




One of the inevitable assumptions of their best reconstruction is that the relaxation time differs for different kinds of forcings and even though it is at most 12 years in all cases, it is somewhat longer for the solar forcing: whether these seemingly "universal" quantities such as the time constant may depend on the kind of the forcing is discussed below.

The previous Scafetta & West 2006 paper was linked to in the context of CERN's CLOUD experiment.

Amplifying solar output variations

Also, they treat galactic cosmic rays as well as some ultraviolet radiation related e.g. to the ozone layer as feedbacks that amplify the solar forcing. Note that the increasing number of sun spots has raised the total solar output by 1 W/m2 (according to Wang et al. 2005) or 3 W/m2 (according to Lean 2000) between the Maunder (sun-spot-free) minimum 1650-1700 and today when the output is slightly above 1366 W/m2. This bare change of the output is very small - it would lead to a rather small fraction of a degree of warming - but it may be amplified by the feedbacks above.

Links to other analyses:
Climate Audit
Real Climate
Steve McIntyre says that in a personal discussion, Nicola Scafetta argued that different forcings influence the climate differently - the response of the climate is non-universal - which is a subtlety not allowed in most existing models. Well, it is a somewhat awkward assumption but nevertheless, it is likely that if different forcings influence different regions of the atmosphere (different altitude and/or latitude) or if they act during different parts of the day or the year, they can lead to very different feedbacks. And different feedbacks of course can have different relaxation times (and lags). So a phenomenological analysis of such possible fudge factors could be helpful.

Can different effects be simply added?

Scafetta and West also argue that different effects - such as solar activity and the greenhouse effect - shouldn't be assumed to be independent because there can be strong "non-linear" interactions between then. They also process various papers in literature and decide that several of them - such as Hegerl et al. - are unphysical because they imply a negative, cooling effect of an increased solar variability.

Concerning the non-linearity, they offer a very entertaining but still potentially realistic example of a feedback: 200-year-long solar-induced droughts may have led to the collapse of Mayan civilization. This collapse led to refosteration, reduction of CO2 in the atmosphere, and thus cooling. ;-) You see that the solar and greenhouse forcings may contribute to each other. Well, such mixed, natural-anthropogenic events may occur but even if they do occur, the question here is whether such effects are strong enough and regular enough to be considered as parts of a useful and predictive model.

I completely agree with Larry that this kind of awkward conspiracy-like theories is often produced by the alarmists and they should be criticized both in the case of alarmist papers as well as skeptical papers which is what I do.

Phenomenology vs theory: climate edition

Even though I wrote that various non-linear and feedback-specific subtleties may exist, I tend to agree with the (somewhat chaotic) Rasmus Benestad's article claiming that these phenomenological reconstructions would become more trustworthy only when a more complete and realistic physical model becomes available. I would probably prefer an inaccurate model - that can be getting better - with some "linear" assumptions over non-quantitative clichés that everything is complex combined with ad hoc, unconvincing mechanisms such as those involving Mayan civilization. In science, it is often crucial to neglect certain effects but one must correctly choose the effects that can be neglected and the essential ones that can't.

On the other hand, I would agree that a statistical analysis such as Scafetta & West might in principle be enough to obtain strong circumstantial evidence that simple models where the effect of feedbacks is universal and where feedbacks are essentially independent - and where the Sun is more or less ignored - are unacceptable: data are sometimes good for falsification of oversimplified models; life is tough. ;-)

Incidentally, Real Climate didn't offer their readers the text of the actual article that they criticized: they only criticized it. These methods remind me of the communist propaganda.

Add to del.icio.us Digg this Add to reddit

snail feedback (0) :