Nicola Scafetta published another paper today, confirming the period dependency of climate sensitivity. (I would have loved to write this, but he attributes the original idea to a book chapter by Wigley in 1988, so its not original anyway.)
In his words, climate sensitivity is frequency dependent:
However, the multiple linear regression analysis is not optimal because the parameters ki and Ï„i might be time-dependent and, in such a case, keeping them constant would yield serious systematic errors in the evaluation of the parameters ki . Moreover, climate models predict that the climate sensitivity to cyclical forcing increases at lower frequencies because of the strong frequency-dependent damping eï¬€ect of ocean thermal inertia [Wigley, 1988; Foukal et al., 2004].
When the signal is properly decomposed, solar forcing is significantly stronger at longer periods of oscillation:
In fact, the actual climate response to cyclical forcing is stronger at lower frequencies because the damping eï¬€ect of the ocean inertia is weaker at lower frequencies [Wigley 1988, table 1]. This frequency dependence arises because the system is typically not in thermodynamic equilibrium. The ratio Z8 /Z7 = 1.55 Â±0.55 is consistent with that between the damping factors for 20 and 10 year periodicities Î·20 /Î·10 â‰ˆ 1.45 indicated by the Wigleyâ€™s model [1988, table 1]. Wigleyâ€™s model also predicts a response-lag of 2.5-2.8 years for a 20 year periodicity.
He concludes with the omitted factors in climate models, without which it is not possible to determine the magnitude of natural variation, above which the AGW signal could rise.
As Lean  noted, the models might be inadequate: (1) in their parameterizations of climate feedbacks and atmosphere-ocean coupling; (2) in their neglect of indirect response by the stratosphere and of possible additional climate eï¬€ects linked to solar magnetic ï¬eld, UV radiation, solar ï¬‚ares and cosmic ray intensity modulations; (3) there might be other possible natural ampliï¬cation mechanisms deriving from internal modes of climate variability which are not included in the models. All the above mechanisms would be automatically considered and indirectly included in our phenomenological approach.