Pielke Senior: NASA GISS Inaccurate Press Release On The Surface Temperature Trend Data

[quote]Joel Shore (16:22:06) :
I am not sure how you get your intuition. What is your intuition for what would happen to you if you were placed in a room with 0.038% of plutonium?[\quote]
Are you insane? You’re comparing X-rays to microwave. Get a life. Get an EM spectrum chart. You obviously don’t have a clue. (Sorry, mod, I can’t take much more of this ignorance.)
[quote]deech56 (16:49:32) :
Really? Most of the atmosphere (O2 and N2) is transparent to infrared radiation,[\quote]
Demonstrably untrue by anyone with any knowledge of AAS/AES, unless rephrased to identify the specific wavelength ranges you are discussing. And the ranges had better be really d**m tight.
[quote] Ric Werme (17:10:52) :
At certain important infrared wavelengths of light,…[\quote]
What is this, a coordinated attack? Please explain in simple terms, in the same manner in which you whack jobs (sorry, that may not include you personally) always do, what exactly it is that makes one frequency of the EM spectrum more important than any other one? Is it your imaginary “windows”? Guess what, they are exactly that, IMAGINARY! There are no windows in the atmosphere, there is only the vacuum beyond it. The spectral “windows” are of no more importance than the lines on the map of the world, which separate countries from each other and are also – wait for it – IMAGINARY!
I have a suggestion for you true believers. The next time the temperature in your house starts heading south, simple open the CGA valve on a very large cylinder of compressed CO2 and bask in the warmth (Seal your windows first, we wouldn’t want you to pollute the atmosphere the rest of us are trying to enjoy, would we?).

Theo Goodwin says:

Old Jim Hansen just trots out his same old assumption. He believes that he can estimate temperatures in adjacent grids that have no sensors and do so without falsifying the data.

Actually, if you read the statement that Hansen et al. have put out ( http://www.columbia.edu/%7Ejeh1/mailings/2010/20100115_Temperature2009.pdf ) , you will see that they have done considerable testing of the assumption that they make. In fact, their testing of this goes back many, many years.
And, they do NOT estimate temperatures from temperatures in adjacent grids. They estimate temperature ANOMALIES in this way. The two are very different beasts: Temperatures are not strongly correlated over distance. (In fact, if you consider a mountainous region like the top of Mt. Washington and the valley only a few miles away, you can see how a few miles can make a huge difference in surface temperature!) However, temperature ANOMALIES turn out to be correlated over quite large regions. I.e., if we have a cold month here in Rochester, it is also likely to be colder than average in, say, Buffalo and Syracuse and the Adirondack Mountains…and, to a slightly lesser degree, even more distant places like New York City, Boston, Pittsburgh, Philadelphia, and Cleveland.

Right Joel, all the work that E. M. Smith has done demonstrating the gross dishonesty in the numbers has been answered by your short refutation… Riiiiighhhht………….

Some comment on : If It’s That Warm, How Come It’s So Damned Cold?
James Hansen, Reto Ruedy, Makiko Sato, Ken Lo, link above in lead comments.
This seems to have been written during the current NH cold spell. Since conditions are so obviously not what they’ve been predicting you would think that a comment with a title like this would be their best shot. Well, I find nothing convincing here at all. In fact, they hardly seem to be trying.
They say ”There is a contradiction between the observed continued warming trend and popular perceptions about climate trends. Frequent statements include: “There has been global cooling over the past decade.” ”
The main reason for this (they say) is that HdCRUT has 1998 as their hottest year, while GISS has that as 2005. Investigating why the difference, they show that if they adopt the same gridding as HadCRUT they get the same result (their Figure 4; warmest 1998, slight cooling since then). They don’t seem to notice that what they demonstrating is that the whole of the rise that they claim comes not from the actual data but from the interpolated “data”. Remove the interpolation, remove the warming. In passing, they also demonstrate that the interpolation is in no way necessary to get a result. They can do it without, they just get the wrong answer.
In attemting to justify their methods, they say that with interpolation
“the patterns of warm and cool regions have realistic‐looking meteorological patterns”, not bothering to say what it looks like without (the actual observations are maybe “unrealistic”?).
They acknowledge that this is qualitative support, and continue: ” One way to estimate …uncertainty, or possible error, can be obtained via use of the complete time series of global surface temperature data generated by a global climate model that has been demonstrated to have REALISTIC spatial and temporal variability of surface temperature.[my caps]” No doubt this REALISTIC means the the model output agrees with the assumptions that underlie their interpolation algorithm. And yes, the find the “errors” are very small.
As far as I can see, that’s all they’ve got.

rbateman (16:51:32) :
A machinist would say that the C02 concentration in the atmosphere is .38 of a thousandth.
Try opening your dial caliper to 1/3 of a thousandth. Takes some practice and due diligence to deal with the backlash.
That’s what micrometers are for. Hope it’s a good one with carbide faces. Set the four-tenths on the Vernier marks, you can eyeball it just under for .38, flick the spindle lock. If really fussy you can check the gap on an optical comparator.
And it better be a regular type. Electronic digitals are cute with the implied precision, but just try getting the correct feel for them to repeat that very last digit every time. And the mechanical digitals are only good until you drop them one time too hard and/or one time too many. 😉
Of course, machinists know that four-tenths can be critical, even one tenth can be important, especially on tight tolerances like small holes. So they likely wouldn’t say that. One good fart spread throughout the shop, now that would be a good description.

Dr. Pielke.
It’s very interesting the way certain factions of climate science leave unfinished threads and unresolved issues. One such “hole” in the science is
peterson’s postulate ( his 2003 paper on UHI) where he postulates ( his word) that temperature stations should be located in “cool parks,” That seemed a fair “postulate” that was begging for on site verification. With Surfacestations at over 1000 sites visited, perhaps we can say something definitive about this ‘cool park’ postulate.
Also, thx for the plug

Dear Anthony,
I have been following your postings for a couple of months and I am obviously the reason for the huge increase in the number of hits that you have been having. Well done you!
I am not a meteorologist or a scientist. I am not a denier or an alarmist. But I do like to reduce our collective use of resources – be it oil or coal or gas or gold.
I was particularly interested in two recent posts “Coal Creek Redux” and “Darwin Zero”. I have been looking at some temperature data from Macau (or Macao if you prefer). Don’t ask why but I was.
For those that don’t know, Macau is a very small port city on the southern coast of China (http://en.wikipedia.org/wiki/Macau but use with the usual caution about wiki). It used to be a Portuguese colony until recently and they have weather records that go back quite a long way. I have no special knowledge about the weather recording station there – it may be that it was a mercury thermometer hanging out of somebody’s window for all I know – but I doubt it. Macau is a very small place and I suspect that the weather station was in the old Portuguese fort (probably for all of its life) and it was probably read religiously from when it was set up. Anyway, cutting to the chase, the Macau meteorological service have a website which lists a hundred years of temperature data 1901-2000 (see http://www.smg.gov.mo/www/smg/centenary2/index_four.htm ). They have more data, as well, going forward and back but not as easily accessible. I thought that I’d take a look at it. It is not in an easy database format but with a bit of effort I was able to export the mean temperature data, year at a time, on a daily basis. The advantage appeared to me that with daily means, minima and maxima the temperatures were less likely to have been ‘adjusted’. I thought, well does this show any warming or cooling over the century?
Most of the analyses that I have seen elsewhere present an annual mean temperature and see what the trend is over the period. I did this for Macau. See below.
What’s up with that? Well in this case we have 55 annual points covering the 100 year period (because I have not completed grabbing the data). Each of these points represents the mean value from 365 (or 366 in a leap year) daily mean temperature values for the year in question. The trend line, using a linear regression, provides an estimate for the mean temperature in 1900 of 22.28°C and in 2000 of 22.62°C. Thus, we have an estimated warming of 0.34°C in the century from 1900 to 2000.
As you know, each time that you get a mean or average value of a variable and state it by itself, you lose some of the information that lies behind the mean. Thus a yearly mean temperature loses a lot of information about the variability within that year. I thought that getting the mean monthly temperatures might be a second estimate – and this is what the plot looks like for the year’s that I have grabbed so far. I’m sure somebody with more time could grab the years that I have not got so far (1906 to 1910, etc)
What’s this mean? If taken at face value, it looks as if the mean of the ‘mean monthly temperatures’ averaged month at a time was about 22.4°C. It also appears to mean that the trend of increase in the mean temperature is 0.0003°C per month. (We need to be a bit careful because I’ve used Microsoft Excel and Excel does some strange things when you use dates and converts them to numbers – starting at 1 January 1900). This gives an estimated potential increase of 0.39°C for the century from 1900 to 2000.
Hold on a second though, this is giving equal weight to every month in the year. Are the very hot months and the very cold months skewing the data? Another way of looking at the same data is to check what happens in each quinquennium. See below:
In this case, the slopes of the trend lines are probably not relevant since they are only taken over a short time period of five years. We can see, however, that there is no large shift in the curves from the period 1901-1905 to the period 1996-2000. We do get estimates of 22.49°C for the first quinquennium and 22.71°C for the last quinquennium of the century. This represents an estimated warming trend of 0.22°C (over 95 years) or 0.23°C over 100 years.
Another way of looking at the same data is to first get the mean values for each month – we know that there is a nearly periodic cycle of about 12 months which affects the temperature. If we take the average for each month over the 100 years then (so far) we get the following average annual cycle:
We can use these monthly means to get the deviations in each year from the monthly means and therefore de-trend the monthly data for the effect of the annual cycle.
This de-trended data, using 660 monthly data points (out of an available 1200 data points for 100 years) provides a very slight estimated warming trend of 0.03°C per century. You can see that the spread of data is roughly in the range ±5°C and there do not appear to be ‘outliers’ of large magnitude in the data set. However, it might be as well to look at the distribution of the de-trended data. See below:
The distribution of the de-trended data is shown on the histogram. Also plotted is a Normal distribution for comparison. The data looks pretty normal and the extreme values – the hottest and coldest months in the century do not look outstandingly hot or cold – the sort of values that might be expected.
Of course, we can do a further check on this – and I suggest that somebody with more time could do so. This is to grab to maximum and minimum temperature data sets on a daily basis and check that the maximum monthly temperatures and the minimum monthly temperatures always dance above and below the mean monthly temperatures respectively. Another advantage of checking the maxima and minima would be that one could see whether the maxima are getting larger (hot getting hotter) or the minima are getting smaller (cold getting colder) or vice versa. This would tell us more about why there is a warming or cooling trend – if indeed there is such a trend.
(Just as an aside, there seems to be a tendency for the modern environmental scientists to discard ‘outliers’ just because they are outliers. Clearly, there is case for do so if, for instance, somebody has misread the original handwritten record and has read ‘54’ instead of ‘34’, but I think one should only discard outliers if one can see why – and then tell the world which outliers have been discarded and why.)
Oh, I forgot to mention, Macau used to be a very small place indeed as far as population is concerned. In recent years its population has grown quite considerably. According the World Bank, Macau had a population of 172,608 in 1960 and 526,178 in 2008 (http://www.google.com/publicdata?ds=wb-wdi&met=sp_pop_totl&idim=country:MAC&q=macau+population ). I do not know what the population was in 1900, but I would guess at less than 50,000 and maybe much fewer. Should there be a sign of an UHI (urban heat island) here? I suspect not, but that may be for others to look at.
What’s the point of all this? Well I think the first lesson that it teaches is that for each temperature reading station we need to have some history. Where was the temperature read and how? Has the temperature reading or system been changed over time? These are some of the points that were being made in the post about Darwin. Is the station data from a consistent source and is it reliable? (By the way, Willis might want to look at temperature data from nearby Indonesian islands such as Bali to compare with Darwin – probably more relevant than Alice Springs.) The second lesson is that each station needs to be analysed in detail before it is added to the global database. The third lesson is that using a linear regression trend to project forward may or may not be appropriate. Are there any other trends within the data which can be decomposed and thus eliminated from the data? Is Macau relevant in any way as a station to search for global warming? Search me!
I have not completed the analysis of the Macau data – just started in fact – but you can get different answers depending upon the analysis – particularly if you have a small number of data points.
As you can see, I’ve just done what I’m suggesting that nobody should do – presenting data that is incomplete and only partially analysed!
Cheers
Bob the Builder
PS You can see that I’m an amateur and can’t submit the charts that go with this post. Let me know how and I’ll send them separately.

RE LAShaffer (18:56:01) :

What is this, a coordinated attack? Please explain in simple terms, in the same manner in which you whack jobs (sorry, that may not include you personally) always do, what exactly it is that makes one frequency of the EM spectrum more important than any other one? Is it your imaginary “windows”?

I would guess that all the “whack jobs” have their physics textbooks open to the same chapter. Just think of it as independent confirmation of results. 😉

RE Ric Werme (21:19:15) :

One problem I have with the historical contexts people like to fall back on, is that I like to think science has progressed since the days of Hershel (solar activity and climate), Tyndall, and Arrhenius. In particular, the CO2 studies neglected convection, and while the CO2 blanket slows down some of the IR transport, convection provides another means of transporting heat upward. The more CO2 slows down IR radiation, the more important other processes become, and that simply wasn’t examined. As far as I know, it may still not be well examined or understood.

Ric, you raise an interesting point, but we should look on these earlier studies as the building blocks of the later studies. They are incomplete in the same way that Watson and Crick (1953) was incomplete. (You might also note that their earliest papers had an error in the number of hydrogen bonds between C and G.) On a related note, Robert Grumbine wrote an interesting post about successive approximations. Reading the climate literature, especially the idea of climate sensitivity, is a study in the progress scientists have made in their understanding of climate.

Very nice explanation, Mr. Smith. Looking forward to Joel’s refutation.. but not expecting it. Joel sometimes even admits it when he missteps. This would be a good time to do it.