SkepticalScience Needs to Update their Escalator

“How is it possible for anyone passing beyond Groundlevel at University anywhere not to have learn what it takes to present valid arguments that can lead up to a sound conclussion?”
Lol.

What Mosher is saying, and I agree with, is that El Niño is descriptive of a weather pattern and not a cause (as in a new heat source) for global warming. The fact that it occurs and that the surface temperatures go up globally on average are automatic.
Bob is describing the mechanism of heat transfer that occurs during this process (not new heat creation).
Both are right and both views are helpful in looking at the the weather pattern.
Mosher believes the CO2 influence is causing an upwards inevitable rise in temperatures without regard to the checks and counterbalances present in the earth’s climate stability ie that rising CO2 may lead tofeedbacks in clouds etc that restore a balance albeit with more CO2 in the atmosphere.
Hence he cannot a accept that the temperature rise has stalled and is reversing.
Hence his comment.
He did not disagree with you.
Bob as soon as I get over my aversion to credit cards on line and develop some internet skills I will buy a copy. You should send one to Mosher gratis but he may already have bought it.
Please continue your work.
Ps no El NIno yet is there ?

Robert Brown says:
“but there is no way that the addition of literally thousands of sampling stations would not reduce the error bars. In fact, reduce them by a factor of (wait for it) order \sqrt{1000} \approx 31 ”
I generally like your posts but if you add a drum roll to a statement, give it just one more proof read.
Adding whatever number of sampling points will not reduce anything by a __factor__. Simple example, you have 1000 noisy data points and you “add” another 1000. The noise is not reduced by a __factor__ of sqrt(1000) but by sqrt(2) or the gods of statistics would shit themselves.
You have to increase the number of points by a _factor_ to get the reduction of noise by the \sqrt factor.

Trenberth et al., are making it up as they go along.

You have to increase the number of points by a _factor_ to get the reduction of noise by the \sqrt factor.
I’m aware of that and I’m sorry I wasn’t sufficiently clear. I was assuming — possibly incorrectly — that there were only a handful of contributing measurements back in 1955 (as in literally — perhaps 4 or 5, maybe 10 or 20). Hence my statement about ORDER of . There are other reasons it might have been more or less than this number, as well — if they had 50 “sites” in 1955, but they consisted of 50 sites at only 5 locations, spread out within a few hundred miles of one another, the sites are hardly independent or randomly selected and might constitute only 5 “iid” samples as far as stats are concerned. Of course the same thing can be true (and is, to at least some extent) with the ARGO floats — they aren’t either uniformly distributed or randomly distributed, so even 4000 floats might not make up 4000 iid samples.
None of which should detract from the central point, which is that even if the number of “independent” sampling sites only went up by a factor of 100 (and it is my strong recollection that for most of the pre-Argo period, N was indeed never much greater than 50, with a lot of site correlation) there should be at least an order of magnitude reduction in the error bar between 1955 and the present. Its absence makes me wonder how they compute the error in what is supposedly a 3D kriging of sparse samples.
At that point we could go a step further and note that measurements of heat content can only be derived quantities from measurements of ocean temperature in depth (0 to 2000 meters according to the figure title). The entire data set including ARGO thus suffers from my incomprehension of how they can krige temperature in planetary oceans to that depth from only a 4000 — let alone 40 — sites to five or six significant digits absolute, using instrumentation that is likely not accurate to more than 3 or 4 digits on a good day with a tail wind for the first (pre-transistor) half of the record. How exactly did they measure the ocean temperature at 1 km of depth back in 1955? With a mercury thermometer (who read it)? With a thermocouple (who adjusted and normalized it, how were the results transmitted and tube-amplified, how did they handle the heat generated by the tube-driven apparatus)?
Remember, in 1955 electronics was still at the green screen radar level and kids were still listening to The Shadow on radio for electronic entertainment. It’s not that they couldn’t make precise measurements — oh, wait, yes it is. At least not without enormous expenditure. And who was funding this, way back then? So add in instrumental error to the error that ought to be reduced and I suspect you can easily find an additional order of magnitude — Ask Anthony about the error levels in MODERN WEATHER STATION apparatus using solid state electronics, and whether it suffices to measure something to 5 significant digits. Now re-ask the same question in 1955, concerning a piece of apparatus that has to be backed into a cylinder and then lowered to 200 atmospheres of pressure, from a ship that has junketed a small group of scientists to particular locations that, um, have just the right “scientific appeal”.
rgb