Fred,
I return to resonance and these pesky quasi periodic variations. I am afraid that I may not be answering directly but more explaining a viewpoint that looks at the issue olbiquely and then asks how the existence of a resonance would effect how we look at some of your points.
If I manage to explain myself adequately it should be clear that in my view these wobbles are not antithetic to the standard theory, they could even be essential, and that we should not go looking for short term data to explain wobbles that have their origins in the prior couple of decades. A resonance is a sort of temporal teleconnection. A ghost of the past that haunts us.
Back to the wobbles and there causes.
The following seem possible:
1) They are the result of periodic forcings
2) They coincidentally result from other forcings
3) They are an apparence due to resonance and noise
4) They are purely happenstantial.
The implications differ.
Happenstance is a limiting case. The wobbles being just random fluctuations inform us as to the relative strength of the response at the frequency of interest. Their is a tacet implication that longer periods should have at least as strong a response hence a large fluctuation that happened to be seen with a 60 years period would support high sensitivites beyond 60 years.
The presence of a resonance would indicate that we should not take the strength of the response at the centre of the band as being indicative of the response at either longer or shorter periods. In the case of a resonance with a 60 year period it would suggest that the uncertainty in a ~30 year trend (in band) would be larger than otherwise expected but those for 60 years (out of band) would be much as expected. The band width being dependent on the sharpness of the resonance. So we may have the situation whereby we should express extra caution with trends over certain periods e.g. ~30 year and perhaps ~10 year (half of 20). The practical implications of which is that the IPCC WGI were wise to use a 50 year period and would be foolish to use a 30 year period.
A resonance around the 60 year period driven by noise would show up in the records as a spectral peak in standard of wavelet analysis but would lack constancy in amplitude, frequency and phase. In particular phase would not be predictive much beyond a cycle, or less or more depending on the sharpness of the resonance. Cycles would appear and fizzle out only to appear again. This lack of predictive power beyond the near term means that any prediction should be quickly blended into just a greater uncertainty. I think that ENSO, PDO, AMO etc are only predictive out to a minor fraction of their period. A whole period hence one might be uncertain not just to the amplitutde but the sign.
The existence of perceptable periodic forcings could be of enormous practical benefit. I think that a regular beat in paleo-data would cheer a lot of people up as it could help solve time calibration issues.
A precise periodic rocking by around 0.2ºC pk-pk should have some implications not just for the temperature. The AMO is associated with changes in the weather patterns and particularly precipitation. I believe that a period of 64 years in millenial and semi-millenial Nile data records has been detected and judged significant but I have no idea if it is periodic. If it was, it is too far from 60 years to be the same signal. Weere there a truly periodic signal in that data, which benefits from good record taking in that we know which years are being refered to, I can only think that it would have been spotted as this data is highly researched.
The existence of periods of around 60, 20, and 10 years is highly suggestive but you do not have to move those figures from their round values by much before the evocation is lost. The certainty in the 60 year figure from the temperature record is not good, perhaps +/- 10 years (50yr,70yr), this is a direct consequence of the uncertainty in frequency for the 160 year record (being 1/160 per annum). As I read it the paleo-AMO analysis only supports a peak in the 60 year region. I have seen nothing yet to satisfy me that we have a periodic forcing.
The suspected forcings e.g. GISS forcings covering both natural and other forcings do have a spectral peak in the 60 year band. This peak is in the right frequency band and the phase is plausibly correct relative to the temperature response.
I think it is this point were the fuss happens. These are fluctuations that are held to to be forced by some people but not by others.
However the predicted amplitude may be inadequate to explain the historic variance. This could be because there is also a period forcing which just happens to be in phase right now. Or it is just happenstance, or there really is a resonance.
If there is a resonance it implies that any quasi 60 year signal in the forcing would be disproportionately magnified in the temperature response. For instance the 1930s would have been a little warmer than otherwise, the 1960s a little cooler, and things would have perhaps been a little inflated in the 1990s. This is suggestive but far from definite.
A resonance would imply that the current “lull” could be “forced” that is that it is overwhelmingly due to the forcings, and that we should expect a “set back” at the end of any 30 year period of significantly divergent but mostly monotonic warming. To put it another way, that 30 year trends get exaggerated but 60 year trends do not.
It also keeps open the possibility that the relationship between forcing and response is linear. I do not mean a scaling but that additivity holds.
In summary:
The existence of a quasi 60 year damped resonance might explain some of the variance noted in the historic and paleo data as being just the amplification of happenstantial variations in forcings in the 60 year band.
It would imply that we need slightly different statistical arguments when we consider trends in and out of the resonance band.
It would imply that we should expect “corrections” to occur about 30 years out from current actions.
It might imply that the divergence between the models and the historic record is due to a failure to capture this resonance, or that it capures it but from different initial conditions, that we could try to correct for analytically.
That the “lull” was both predictable and the result of warming.
That there is an implication that we should perhaps be a little cautious in believing that any particular 30 year slope had well characterised error bars as an additional amplification needs to be applied to these error bars.
That there are two ways of looking at issues like this. On one hand we can rubbish the whole concept or we can winnow wheat from chaff.
Alex