Since we have no idea how the corrections were done, the raw data is the only valid data we have.

You might want to get the methods published. Then we can go on from there.

Science is open Open up the data. Open up the methods.

http://powerandcontrol.blogspot.com/2007/08/problems-in-asia.html

Hansen is making corrections based on data.

Wouldn’t it be better to find the actual errors and base the corrections on that?

Hansen bases his corrections on assumptions about the quality of the data. If those assumptions are wrong (it appears to be since his methods did not ferret out the currently discussed error) then his corrections are of small or possibly even negative worth.

Nope, although these actual measurements are highly uncertain. The mass difference means that it’s completely expected that water would rise less in temperature than air.

(Paul can correct this if I go astray)

The density of air at 15°C is 1.225 kg/m3, while water is about 1000 kg/m3.

specific heat:

Air = 29 J/mol/K
Water (gas) = 37
Water (liquid) = 75
Water (solid) = 38

Assuming that I have that right, 75 Joules added to Air would cause its temperature to increase by 3 degrees, while the same 75 Joules added to water would only raise it by one degree.

However, the mass of the atmosphere is 5.3 x 10^18 kg. That’s a lot, but it’s only .378% of the oceans, weighing in at 1.4 x 10^21 kg. The oceans are only 6% of the mass of the crust, and the crust is only .374% of the earths’ mass.

The Sun is heating all these components simultaneously. (AGWers pretend that Sun only heats atmosphere). So, assuming for the moment that all components were at equilibrium, and the Sun became more active. Air temps would increase the most, oceans much less, crust much less, core least. However, it’s energy that matters, so a slight rise in ocean temperatures will deliver a lot of energy to the atmosphere.

A “law of large numbers” is one of several theorems expressing the idea that as the number of trials of a random process increases, the percentage difference between the expected and actual values goes to zero.

Or in other words, since we’re combining things with math, we get math effects. Or that continually averaging imprecise information makes that information more precise.

So the only reason we get down to that .01 is by the sheer volume of the measurements being combined, giving a smoothing over time, not a native accuracy to that level, no…

What is being done in climate is trying to extract a very small signal from a very large noise. Except that since there is no decrypted text to refer to there is no way to be sure all the noise has been removed.

It is ludicrous to talk about .01 C differences in a signal + noise that varies by 50 deg C.

Small differences statistically extracted might be valid if the system was stable and repeated measurements over time were collected.

However the system is not stable and is influenced by multiple chaotic oscillators.

And that is even before we get into the difficulties of actual measurement.

In general: If the system wasn’t designed to do what we’re doing with it, I’d logically think that using it for that purpose would yield results that are inherently incorrect. Which is why the temperatures get corrected, one would think. Claiming that this (the “global average based upon these stations”) means anything nearly as well as it’s being made out to be in the first place is, I think, dishonest.

But that’s not the subject here, what value the data has! It’s that if we’re collecting it, and the stations are not meant for it, but it’s all we have. So whatever it is that means, why not restrict things to the best stations?

Which ones are the best? Gather the data on the state of the stations, grade them, and see what things look like with the Class 1 or Class 1 and 2 stations only. We should have already been able to do that, and that is the entire point!

Okay, so what does that photo tell us about this site? First we have known possible contaminations, that one would imagine are actually probable contaminations. The phone pole, the material covering the phone pole, the light, the power to the light. Then there are the ones we don’t know about; insects and birds (bats too perhaps). Then are the obvious issues that break the standards; proximity to the light, the height of the sensor, and the equipment itself.

I can semi-answer that ‘errs in lights=0 matters?’ question right now. In this case, yes, it certainly appears so. In other cases, maybe, maybe not. Obviously it’s a near certainty that in some cases, and maybe in most cases, the answer will be “yes”. (If that’s really even up for debate that the obvious answer already is “yes”.) What we can’t answer, yet, is what percentage. And once we can, we will be able to determine what the extent of the “matters” is for any station that’s kept in the analysis. Does that invalidate the network? I don’t think so. (And again, what use the information is, and what it means, that’s a separate issue.)

And wait there’s more! If we ignore the question if these global means of these separete systems actually have any meaning at all, much less as a whole, how’s this for a question about the data itself.

Since the trend on the surface using land temps only (well, the material under the equipment and how it mixes with air 5 feet (or more) up) is a trend of .07 C per decade over 125 years, how that’s possible when when the SST readings have only risen a little over .01 C per decade over 150 years. (I’m surprised it’s a 7:1 ratio, but maybe I shouldn’t be.)

Regardless, measuring air at 25 feet and 5 feet are vastly different things, and adjustments shouldn’t be needed, so maybe nothing should be surprising. That’s why the sites get audited! And, I believe, why so many folks are upset it’s getting done. So all the bickering ignores the reasoning behind this, I think. Finding sites that don’t meet standards, and later, hopefully, explain or remove the inconsistencies in all the records.

2) Of course. I didn’t mean to infer, imply, or otherwise allude to any such insanity.

3) If the photo showed an MMTS hooked up to a blast furnace, along with operator notes ‘moved here in 82′, do you really need more information to say this site should be removed from the analysis? Without wading through Hansen it is tough to say the exact weighting any individual site ends up with. I, personally, would chuck this one faster than some of the more sensational photos. Height is an acknowledged and accepted influencing factor.

I do see the difference between ‘Temperature Data’ and ‘Temperature Trend Data’, and figuring out trend contamination is trickier. But the very first step is ‘achieve census of extant sites’. Until that’s done, you’re aiming to motivate people to participate to actually get the flipping data you want to evaluate. Four or five pictures of a site by anyone, of any skill level is more than NOAA has done. Even when the sensor is on NOAA property.

Is it earthshaking news that Port Angeles isn’t home to a well maintained site? No. Is it one more step towards a census? Yes.

fine – looking at only the ‘properly sited’ stations gives an independent analysis of the data. Nothing wrong with that – I don’t think anyone is objecting to that. Have at it – are yo committing to doing that analysis, and publishing in the literature where the results can be carefully examined?

But:
1. Designing an alternative analysis DOES NOT INVALIDATE THE CURRENT ANALYSIS. Hansen et al chose to do QA and corrections based on analyzing the data and comparing stations. There were good reasons to do that, and some weaknesses. Finding some badly sited stations from among those he used, which is so far all that is being done, DOES NOT IN ITSELF INVALIDATE THAT APPROACH – much as the screaming here about the stations would indicate otherwise.

2. Even if only well sited stations are retained in the analysis, a similar examination of the data must be done. You indicate this in your post above, but it bears emphasizing. Just as bad siting does not in itself tell us the trend data are irrecoverably contaminated (or contaminated at all), good siting does not tell us there is no trend contamination, or whether there are time of day issues, whether a thermometer was replaced possibly into a different location within the shield – this can have significant offset effects – whether local irrigation was commenced or discontinued – and on and on.

3. Posting the photos on the front page, with superimposed declining uncorrected temp data for the ‘well sited’ station, and increasing uncorrected temp data for the ‘badly sited’ station, without noting that the trends are indistinguishable for those stations in the corrected data, and especially without prominently noting that the preliminary photo-only data is UTTERLY USELESS for deciding the validity of the extant analysis, is at best intellectually misleading. At best.