Jan P Perlwitz:
I feel much more comfortable with “slowdown”. I would add “statistically not significant” as long as statistical significance of such a slowdown hasn’t been established.
Yeah I agree—and the statistical significance question is a tough one to crack. I’m going to jump on the soap box for a minute, sorry.
The problem is you can’t simply pick an interval (e.g., 1998-2014) then use statistics that are designed for an arbitrary (as opposed to selected) interval to compute the statistic significance.
Secondly our ability to measure temperature is much better than our ability to infer the secular trend uncorrupted by short-period unforced variability.
Because the short-period variability contains autocorrelation this leads to the practical reality that you can generate a wide range of features that we can recognize visually in the data, but the feature can amount to little more than a particular phase relationship among the various components of the unforced variability.
Thus, we can all clearly see a slowdown in trend for say 1998-2014 and the slowdown is real. And yes this trend does translate into a change in other physical variables—remember even this is due to unforced variability, unforced variability is not a measurement artifact, so the laws of physics negotiate how that slowdown in trend occurs. If it’s due to natural variability, energy has to be conserved etc.
But what we can’t easily infer is how frequently this sort of slowdown for this period occurs.
To this point, I haven’t seen anything that convinces me we’ve really reached true significance, where the result can’t be explained in terms of natural variability.
Certainly there is nothing here to suggest to me that there is a problem with the overarching theory of AGW. Possibly it suggests a somewhat smaller number for CO2 climate sensitivity, but a larger number for natural variability.
Even then, because natural variability is even harder to model than CO2 climate sensitivity, this isn’t good news, because less predictability means more uncontrolled risk not less.