“The study team analyzed several million daily high and low temperature readings taken over the span of six decades at about 1,800 weather stations across the country, thereby ensuring ample data for statistically significant results.”

Comment:
Measures of statistical significance rest upon untenable (in this context) assumptions of randomness, which rip & tear at the body of truth.

There is a deeply-rooted epidemic in our statistical schooling.
This is epistemological.

You lost me right there (…as I pause to remember the various & countless devout & reverent tones in which I have heard that religious phrase ["suppose further that..."] preached & sung …and as I recall naively-uncritical students obediently taking notes, during the indoctrination they received to blindly swamp & undermine the world with unrealistic modeling assumptions).

Keep in mind that Meehl has to:
a) survive the swarm of alarmist-sharks in his pool.
b) maintain funding for his solar / terrestrial hydro-cycle research.

Meehl seems sophisticated enough to maintain the required “cordial” relations by exercising (administratively) “appropriate restraint” & careful timing of announcements regarding his solar-terrestrial findings. He may be the picture of a sufficiently-non-threatening inside-player wielding actual power (as opposed to the screaming-protester-type, throwing rocks at the ivory tower from outside). We may not know for many years what his true game is… and we may never know. If he is really smooth, we may just be left to read between the lines in his solar-terrestrial papers.

As for these abstract simulation-studies based on nonsense-assumptions:

Academics are very fond of them for their “cute effect” (which is one proven way to attract academic attention, so long as you do not resort to it too often). It is worth noting that academics don’t always take “cute” research seriously (in physical terms), but they like it anyway because it stimulates computational ideas [and that is hard stuff, even when CDOs (convenient dramatic oversimplifications) are made]. Spending a decade around math & stats departments was helpful in learning the “tricks of the trade” – lots of coy smiles from clever players I assure you …& lots of fooled grad students who don’t recognize academic gaming right in their faces – you can build an elaborate world with a strategic false assumption – and the troops are conditioned to accept all assumptions without question …and the smart players sit through presentations with a big grin on their faces, not blowing their colleagues’ cover.

IMO the only valid critique can be summed up in two words: “so what?”

> 1/n suggests pink noise to me. Similar to 1/f noise in electronics.

Not the same. Suppose we’re tracking low temperature records (hey,
high temperature records get too much attention! Suppose further that
the temperature is a random process.

In the first year, n=1, and the probablility of setting a record is 1/1. This is sort of by definition, but it’s the only special case.

In the second year, n=2, and probablility of setting a record is 1/2. In the 10th year, 1/10, etc.

By looking at the ratio of high records to low records, the random stream would continue with a 1::1 ratio indefinitely, and hence a warming or cooling trend will stand out as a change in the ratio.

Also, it’s one heck of a good way to hide the really extreme decades, the general loss of station coverage and probably a couple other things that some people might want to hide.