One of my favorite bits of irony is that the primary defenders of using surface temperature measurement over space-based satellite measurements is … the Goddard Institute of Space Studies at NASA, and James Hansen (its director and friend-of-Al) in particular. If find it amazing that people still want to use the GISS surface temperature numbers in preference to satellite figures, despite their proven biases and lack of consistent coverage. But the GISS numbers give a higher number for warming (since they are biased upwards both by measurement biases and GISS-added adjustment factors) and that is what is important to global warming alarmists. Its the "fake but accurate" meme brought to the realm of science.
But, since we do have to keep reminding people of the problems in surface temperature measurement, here is a study by Ren et al in 2008:
What was done
Noting that "a major divergence of views exists in the international climatological community on whether the urbanization effect still remains in the current global and regional average surface air temperature series," the authors employed a dataset obtained from 282 meteorological stations, including all of the ordinary and national basic and reference weather stations of north China, in order to determine the urbanization effect on surface air temperature trends of that part of the country over the period 1961-2000, dividing the stations into the following categories based on city size expressed in millions of people: rural (<0.05), small city (0.01-0.10), medium city (0.10-0.50), large city (0.50-1.00) and metropolis (>1.00).What was learned
Ren et al. report that mean annual surface air temperature trends for the various station groups of north China over the 1961-2000 period — in degrees C per decade — were 0.18 (rural), 0.25 (small city), 0.28 (medium city), 0.34 (large city), 0.26 (metropolis), and 0.29 (national), which makes the urban-induced component of the warming trend equal to 0.07 (small city), 0.10 (medium city), 0.16 (large city), 0.08 (metropolis), and 0.11 (national), all of which results are significant at the 0.01 level.
GISS numbers give a higher number for warming – no they don’t. As always, you can avoid making basic and embarrassing errors by looking at the data.
A link to the Ren et al paper, rather than simply a link to some biased commentary on it, would be useful.
scien-whatever you like to pretend to be-
tamino has been repeatedly exposed for statistical fraud/errors in his attempts to support the “hockey stick” and other bad data by mckitrick and others so you’re going to need to show us how he got these numbers.
the graph you show says “synchronized”. that could mean pretty much anything. care to link to his methodology?
i’m dying to see how he gets this chart:
http://www.junkscience.com/MSU_Temps/GISSglobal.html
which shows 2 new temperature highs since 1998
to look like this one:
http://www.junkscience.com/MSU_Temps/MSUvsRSS.html
in which neither of the two global satellite temperature datasets shows a reading within 3 tenths of a degree of the 1998 high since then.
just changing the baseline used to set departure from anomaly should not change the relative positions of data points to one another.
so why are 2001 and 2006 now cooler than 1998 by a large margin in the synchronized graph when they are warmer in the reported data?
Are you blind? Did you look at the graph? why are 2001 and 2006 now cooler than 1998 by a large margin in the synchronized graph, you ask. They are not. Hint – look at the red line, look out for peaks since 1998, and then refer to the key to find out what the orange line is.
sorry, mistyped the second to last line.
it should read
“just changing the baseline used to determine anomaly should…”
given how tangled and overlapping those lines are, perhaps you can link to the data?
and the methodology he used to “synchronize”?
because a look at the raw graphs i posted shows quite a divergence since 1998. if you are going to claim it’s not there, you need to back it up and show us the numbers, not a barely legible chart that was produced by a low quality researcher with unknown methodology.
that’s hardly the way to accuse others of not checking data…
There was a typo in my post – ‘orange’ should have read ‘red’.
The raw data is easily obtained – google GISTEMP, HadCRUt3, ‘RSS temperatures’ and ‘UAH temperatures’. The methodology is simple – select a reference period common to all datasets, substract their mean anomaly during the reference period from all anomaly values.
Your confusion seems to result from a belief that all the peaks in all the datasets should have the same amplitude. El Niño is a tropical phenomenon; GISTEMP includes polar regions while other datasets don’t; therefore the effect of El Niños is smaller in the GISTEMP data than in others. You can clearly see this in Tamino’s graph. Your two separate graphs show us nothing (and of course, you didn’t bother to make links). Eyeballing is not a good substitute for measurement.
“The methodology is simple – select a reference period common to all datasets, substract their mean anomaly during the reference period from all anomaly values”
So wait, if one dataset has a consistently higher anomaly than the others, this would remove the effect. Correct? If dataset 1 = dataset 2 + 0.5C – their “Synchronized” values will be identical.
So basically you’ve posted numbers using a methodology that is designed to get all of the graphs to line up nicely. The only way they could diverge is if there is a significant difference in the rate of change (slope) of one of the graphs versus the other – though this methodology could mask even that effect if the reference period is chosen properly. Let us hope that Tamino used a single reference period for that graph, and not some sort of sliding window.
As for the raw data, it’s quite obvious looking at the junk science graphs that GISS is higher. The MSUvsRSS graph doesn’t cross 0.6 in the new millennium, while the GISS graph spends a significant portion of it’s time above 0.6 and even cracks 0.8 a few times. I don’t pretend this is any sort of quantitative analysis. I am sure one could integrate the area between the curves and produce an average yearly difference between the two sources. It’s quite obvious the number would be positive, and significant.
The fact that there exists a statistically dubious method of subtracting values from each of the graphs that gets them to more of less line up doesn’t change the fact that one is in fact higher than the other.
but what did tamino do?
you cite his chart.
what does it say? what are the peak GISS values in 1998, 2001, and 2006? i would like to compare the magnitude of their departure from baseline to one another.
to my eye, they seem divergent from the GISS reported data, but the way to be sure is to go to the numbers. GISS raw data is easy to get, but i cannot find tamino’s values anywhere.
“Eyeballing is not a good substitute for measurement”: and measurement is no substitute for pummelling, adjusting and fudging the data, eh?
The only way they could diverge is if there is a significant difference in the rate of change (slope) of one of the graphs versus the other – yes. You realise that’s the point, yeah? The allegation is that GISS numbers give a higher number for warming – ‘warming’ is the slope. GISS numbers have the same slope as the other numbers, when comparing like with like by normalising to the same reference period.
though this methodology could mask even that effect if the reference period is chosen properly – no it couldn’t.
The MSUvsRSS graph doesn’t cross 0.6 in the new millennium, while the GISS graph spends a significant portion of it’s time above 0.6 and even cracks 0.8 a few times – that’s because they are relative to different reference periods. The GISS reference period is 1951-1980. The RSS reference period is 1979-1998. To not realise this is to be profoundly ignorant of the data.
there exists a statistically dubious method of subtracting values from each of the graphs that gets them to more of less line up doesn’t change the fact that one is in fact higher than the other. – to illustrate how stupid your point is, let me see if I can draw you an analogy. We both walk towards a point, and measure our distance from the point every five seconds. We plot it on a graph. You start five metres closer to the point than I do. Therefore, your line sits lower on the graph than mine. Does that change the actual position of the point, or cast doubt on either your or my measurements?
GISS raw data is easy to get, but i cannot find tamino’s values anywhere. – the values are GISS, HadCRUT, RSS and UAH and they are all easy to find. I told you how above.
Out of interest, and leaving whatever Tamino is up to to you to chase up (not much of an Authority to Appeal to), I thought I’d take a look at the GISTEMP numbers against UAH. (I might look at RSS another time.)
A few minutes of tinkering with the data from 1979 to 2007 shows that besides the GISTEMP series being about 0.34 C above UAH, it gives a linear regression slope of about 0.19 C/decade compared to 0.14C/decade.
So looking back at the original statement “the GISS numbers give a higher number for warming (since they are biased upwards…”, it would appear that the warming is higher 0.19 to 0.14, or about 35% higher. So the original statement would appear to be technically correct. Whether the reason for the difference is as stated is another matter, and it’s questionable whether either slope is meaningful in the presence of this sort of high-amplitude strongly autocorrelated variability (the 1-month autocorrelation being about 0.75), but those points weren’t in the objection.
I think so-called ‘Scientist’ read as far as “higher number” and jumped to the conclusion that the criticism was about GISTEMP reporting higher numbers, rather than a greater warming (i.e. steeper slope). Oops! It also happens to be true that GISTEMP reports higher numbers, although as noted by so-called this doesn’t really mean anything. Subtracting a constant gets rid of this difference (without affecting the slope in any way), although it’s maybe also worth noting that even if the original intent had been to claim GISTEMP gave higher numbers, that would still be a true statement in no way disproved by subtracting the difference.
Eyeballing the spaghetti graph does vaguely indicate a reasonable but loose correlation – r^2 about 0.7 as it happens – although with the difference varying by nearly 0.4 C in the monthly values. They are clearly related, but the differences could indeed be consistent with one of them being of poor quality, subject to slope-biasing errors. That’s not deducible from the data of course, but given the methods by which GISTEMP is constructed, I know which my money would be on. So if looking at graphs was supposed to show that GISTEMP was of comparably high quality, it failed in that too.
The difference in slope is visible in my graph, although not obvious, and it looks different to Tamino’s, but I won’t bother to enquire why.
Given that we’ve only just finished discussing GISTEMP’s shortcomings in the other post, I’m a little surprised to see it being so vigorously defended here. Oh well. I guess that amongst AGW enthusiasts it’s still the temperature anomaly series to GOTO. 🙂
I have no idea what you are getting on about with your analogy Scientist, but there exist real numbers behind these anomalies – actual absolute temperature averages – not “anomalies” calculated from an arbitrarily chosen (and apparently different for every data set) “reference period”. Those absolute measurements *do* in fact differ from each other.
Now it’s a very simple mathematical fact, if GISS = UAH + 0.2C (hypothetically, referring to the absolute temperature averages, not the anomalies relative to various reference periods), and you grind them through your methodology, the resulting data series would be identical. It’s apparently your contention that this result would contradict the fact that GISS is consistently 0.2C higer than UAH. Whatever works for you man.
Now you might be able to justify your methodology as a means of adjusting for different anomaly reference periods, but why go to the trouble? We’ve got the original data. Why not just ditch the “anomalies” and plot the absolute measurements vs. time on the same graph. I am sure you will have a condescending explanation as why doing such a thing would be stupid and profoundly ignorant.
Now, there might be good reasons to remove systematic constant differences between observed temperature data series, but I haven’t heard you provide any such explanation. Where I come from this is usually an indication of measurement error, or some sort of systematic measurement bias. “Global average temperature” is one thing (though I’ve never seen a proper scientific definition) – if one assumes that such a thing can be measured, and if you measure it, and I measure it, and there is a consistent statistically significant difference between our measurements, we are either each measuring something different, in which case our measurements can’t really be compared, or one (or both of us) are in error. You can’t seriously be suggesting that it’s scientifically valid to just “subtract a constant and forget about it”.
The way to test whether GISS is increasing faster than UAH from 1979 to
present is to take the figures from GISS,
http://data.giss.nasa.gov/gistemp/tabledata/GLB.Ts+dSST.txt
Take the figures from UAH
http://vortex.nsstc.uah.edu/public/msu/t2lt/tltglhmam_5.2
and if you don’t have the tools yourself, go to a site like
http://www.xuru.org/rt/LR.asp#Manually
And run a regression from one to the other. I ran a regression of UAH on GISS, and got the equation
UAH = 0.847 GISS -0.187.
That means the correction constant between the two was a minus 0.187 for GISS to get the average UAH temperature.
UAH temperatures are increasing only 84.7% as much as GISS temperatures.
I was too lazy to compute the confidence interval for the slope of 0.847,
– A. McIntire
I think so-called ‘Scientist’ read as far as “higher number” and jumped to the conclusion that the criticism was about GISTEMP reporting higher numbers, rather than a greater warming (i.e. steeper slope). Oops! – looks like Stevo can’t read, or didn’t notice that the two words after ‘higher number’ were ‘for warming’. Oops! And it looks like Stevo didn’t calculate any error bars for the trends he derived. Oops! And it looks like the same data set when analysed by A. McIntire gives different results. Oops!
“Global average temperature” is one thing (though I’ve never seen a proper scientific definition) – no? Read any papers by Hansen? He’s quite a well known figure in climate science.
Why not just ditch the “anomalies” and plot the absolute measurements vs. time on the same graph. I am sure you will have a condescending explanation as why doing such a thing would be stupid and profoundly ignorant. – yes, yes I do. From the GISTEMP web page: Our analysis concerns only temperature anomalies, not absolute temperature. Temperature anomalies are computed relative to the base period 1951-1980. The reason to work with anomalies, rather than absolute temperature is that absolute temperature varies markedly in short distances, while monthly or annual temperature anomalies are representative of a much larger region. Indeed, we have shown (Hansen and Lebedeff, 1987) that temperature anomalies are strongly correlated out to distances of the order of 1000 km. For a more detailed discussion, see The Elusive Absolute Surface Air Temperature.
You can’t seriously be suggesting that it’s scientifically valid to just “subtract a constant and forget about it”. – of course – in the same way that in the simple analogy I described, you’d have to subtract the distance between us from my measurement to the point, to compare both our measurements to the point. You can’t seriously be suggesting that recomputing anomalies so that they refer to a common reference period is somehow a flawed way to compare temperature data sets.
“GISS numbers give a higher number for warming – no they don’t. As always, you can avoid making basic and embarrassing errors by looking at the data.
A link to the Ren et al paper, rather than simply a link to some biased commentary on it, would be useful.
Posted by: Scientist | May 13, 2008 at 10:31 AM “
Why not admit that Steve was right? UAH temperatures are increasing at
only 84.7% the rate of GISS
When Watts is finished, we will likely find that the failure of the climate scientists to demonstrate basic competence in taking accurate temperatures will be a much larger component of “warming” trends than the urban heat island effects. I.e. urban heat will be a serious factor, but taking the temperatures on top of the building will have caused even more problems than the fact that the building was downtown.
Is it too much to ask “climate scientists” to demonstrate basic competence in recording the temperature before we entrust them to tell us what the temperatures will be in 50 years?
I combined and scaled the graphs from JunkScience’s
http://www.junkscience.com/MSU_Temps/GISSglobal.html
http://www.junkscience.com/MSU_Temps/MSUvsRSS.html
here: http://www.flickr.com/photos/toptechwriter/2491608357/sizes/o/
Looks like GISS is consistently warmer to me.
It’s pretty simple. The GISS slope is greater than the UAH slope. As pointed out by SteveO and Alan McIntyre. I get 0.167 and 0.136 degrees/decade (it depends exactly where you start and stop).
So anyone who says “GISS numbers have the same slope as the other numbers” is either innumerate or a liar.
And look at the March 2008 numbers: UAH: 0.09, RSS: 0.08, GISS: 0.67! Way off.
“Read any papers by Hansen?” Yeah, it’s quite revealing. If you compare the old ones and the newer ones you can see how he’s gone back and fiddled the older data, see the ‘Zen and the art… ‘ thread.
The Ren et al article is at
http://ams.allenpress.com/perlserv/?request=get-pdf&doi=10.1175%2F2007JCLI1348.1
but you can only get the abstract.
UAH temperatures are increasing at only 84.7% the rate of GISS – and what are the error values on the trends? Are they, statistically speaking, indistinguishable, or are you arguing that there is a statistically significant difference between then? And why are you arguing that Stevo is right when he got a difference answer to you from the same data? Are you saying you were wrong?
Looks like GISS is consistently warmer to me. – that’s because despite all the comments about it, you’ve made the basic error of not normalising to the same reference period. GISS temperatures are measured relative to a colder reference period. Why would it be any surprise at all that the anomaly values are higher than those computed relative to a warmer period?
i asked you for tamino’s numbers and methodology, not the GISS numbers. it is HIS graph, the one that you cite, into which i wish to inquire. you cite it as evidence, yet can’t provide the data or the methodology by which it was created.
telling me where the GISS raw data is helps not at all. it has no bearing on what tamino has purported to show.
looks to me like you cited “evidence” for which you have no evidence, or at the very least, which you never bothered to check.
again, hardly the way to accuse others of not checking their facts…
Are you too dense to understand that Tamino’s chart simply displayed the GISS, HadCRUt, UAH and RSS numbers? Did you not understand when I explained the methodology?
i understand what you claim was done and the results, but i want to see the actual numbers or his specific methodology laid out in detail.
i don’t think you have them/it and have never seen them/it and are just making assumptions based on a visual read of some very tangled overlapping lines. you claim that visual inspection of his lines is insufficient, but have no data.
your process seems to leave a great deal to be desired.
and now you resort to snippy ad hominem attacks which seem unwarranted as i have simply made a repeated request for some basic data you have cited.
“Temperature anomalies are computed relative to the base period 1951-1980.” Doesn’t this coincide with the cooling period from 1944-77 roughly? So, if you take a period where things are on a mild increase and compare them to a period where things were decreasing, wouldn’t you get a result saying things were increasing by more than they might actually were? Divergent slopes would create the appearance of of a greater increase.
According the analyses that I have plodded through, it would be incorrect to say that the GISS global temperature estimates are rising consistently faster than other estimates in recent years. That being said, there are several points that should be made. The guess that GISS made for the Arctic Ocean probably made it an outlier for 2007 and maybe 2002. Also, there has been virtually no increase in GISS temperature estimate since 1980 except for the 1994 to 1994 period. Yes, its temperature estimate in 2007 is higher than its esimate for 1980, but it has not been a consistent increase — the trend basically was quite flat to 1994, then a surge until 1998, and it has been flat to a slight decrease since then. (Not only GISS, but other estimates also share this trend.) BUT MOST IMPORTANATLY about GISS: it seems to have a relentless pursuit in making older temperatures — such as the 1930s — cooler so that it looks like we are so much warmer now. As Hansen has said, about half the the time GISS adjusts urban temperatures to show more of a long term warming trend because it feels that other factors are more important that UHI. In summary, GISS does not seem to be relentlessly pushing its current anomalies up as much as it seems to be pushing old temperatures down.
morganovich, I’ve told you several times now where to get the data. What didn’t you understand?
So, if you take a period where things are on a mild increase and compare them to a period where things were decreasing, wouldn’t you get a result saying things were increasing by more than they might actually were? – no. You would get larger anomalies, but you would get exactly the same trend. No matter what period you normalise to, it does not affect the derived warming rate.
the trend basically was quite flat to 1994, then a surge until 1998, and it has been flat to a slight decrease since then – not true.
what i’m saying is that i don’t think the tamino graph is what you (and likely he) claim it to be. therefore i want to see the values of the GISS numbers that IT charts and the methodology by which IT was derived. i know i can build my own if i want to. but i don’t. i want to see how tamino built his and to see how the relative positions of recent temperature peaks on his graph line up with the raw GISS data.
i thought of a clearer way to ask the question:
please provide a link to the actual (post synchronization) numbers used to draw the lines on tamino’s chart and the methodology by which they were computed.
What an idiotic demand. You have the numbers. You know the methodology. Plot the graph yourself. How am I supposed to provide you with numbers that someone I don’t even know has computed? Fuck off and ask Tamino if you’re too fucking stupid to work out how his graph was produced.
Here’s a better site you can use to compute confidence intervals.
http://www.changbioscience.com/calculator/scientific/cal0.htm
The 95% confidence interval is a slope of .847 +- 0.307
for UAH/GISS, a 99% confidence interval of 0.847 +- 0.415, so we can
only be about 84% certain that GISS temperatures are warming faster than UAH temperatures.
you were the one who said it was no good to eyeball the numbers when i commented that the GISS slope looked different in his graph than the raw data and questioned what tamino had done as the claimed process should not alter slope. i have a suspicion he has doctored the data and would like to find out.
then when i ask you to provide them (or even if you have seen them), you don’t have them and hide your charlatanry behind a broadside of abuse.
once more, your have shown yourself to be a hysterical, hypocritical fraud. you waste the time of others making ludicrous data requests, but cannot even manage simple ones of you own, even when you suggest them yourself.
oh, and you still owe me $100.
i have a suspicion he has doctored the data and would like to find out. – so fucking find out, you fucking idiot. You have everything you need because I’ve told you what data to get and what to do with it, you don’t need anything from tamino, you certainly don’t need to whine on at random people who have nothing to do with tamino to get ‘his’ numbers when they are not ‘his’ numbers, and the whole fucking point is that you want to check yourself whether the fucking graph is accurate. What the fuck is the problem with your tiny mind that you can’t understand this?
scientroll, it’s very simple. you linked to a chart and made some claims about how it was created. i said, that based on the apparent shape of the curve, it could not have been created that way.
you then said i needed to check the numbers.
but you can provide no numbers nor a link to methodology.
therefore, i take your “evidence” to be worthless and just more of tamino’s nonsense that he consistently fails to back up.
and your attitude sucks.
what is it with you anyway? did someone not hold you as a child or something?
I’ve provided links to the numbers. I’ve described the methodology. A 12 year old wouldn’t even need to be particularly good at maths to reproduce the graph and check it worked. Why are you not able to even understand what data you should be using? Why are you not able to follow the incredibly simple steps to check that the graph is valid?
because i don’t have the values that the graph purports to show.
so how can i check them?
you seem to think that i don’t understand where the basic data is or how to set it to a common reference period. i do. there’s nothing hard about it. i have been saying this all along. how can i get this through to you?
but how can i compare those results to those you cited until you show me what they are?
are you just being deliberately difficult here to try to get off the hook for having linked to a bunch of unsubstantiated data?
there is really no way to make this any more clear.
i want to check the tamino work you cited. to do that, i need it’s results.
but you are unable/unwilling to provide them, implying that you have no idea if they are accurate either despite your constant demands that others provide back up data and peer review etc.
you fail to adhere to your own standards. this makes it difficult to take you seriously.
but how can i compare those results to those you cited until you show me what they are? – you plot the graph, you fucking idiot. What is your problem?
and then what, eyeball it with tamino’s?
how do i compare the slope of the line i generate with his?
you don’t even make the rudiments of sense. go read your own comments above about eyeballing not being a good substitute for measurement. isn’t that exactly what you are proposing i do?
you want me to generate data then eyeball the difference to what you presented?
how do you propose i compare my data to tamino’s?
and calling others f’ing idiots when you seem unable to grasp a basic issue like needing two sets of numbers to make an accurate comparison seems like the act of either a fool or someone dangerously emotionally unbalanced. seriously, what is it with you and all this pent up anger? all i am asking for is the data you cite so i can compare it to data i derive. why is this such an explosive issue for you? are you just trying to get out of providing that which you require of others (and suggest to me) by flinging vitriol?
you have no actual data apart from a presumed slope from a difficult to read line. you yourself say that eyeballing such a line is insufficient. why don’t you just admit that you can’t back up this data?
Why the fuck do you keep asking me for Tamino’s data? If you want something from Tamino, fuck off and ask him for it. Why would you expect me to have figures used by a random blogger? I’ve told you how to recreate the graph so fucking go and do that, or shut the fuck up.
you are the one who cited it as evidence that GISS does not overstate warming.
but you never bothered to check its numbers and can’t come up with them. if i had pointed to that graph, you would certainly have asked me for the data.
i guess your standard of evidence is not terribly high…
given your repeated insistence throughout myriad postings on this site that others produce the numbers behind anything they cite, this seems a significant lapse for you to expect us to tolerate.
you are both a hypocrite and a charlatan, and worse, you are an uncouth boor that acts like a petulant child.
i had hoped that by engaging in a data centric conversation with you, i could elicit some manner of civilized discourse and delve into the science of the subject matter.
however, i see now that you are utterly incapable of such.
i wash my hands of you.
Can’t come up with the numbers? I’ve told you several times how to get them but you’re so retarded you can’t understand that. You’re a fucking moron. An absolute fucking moron. Follow the methodology and make the graph. What the fuck else do you expect people to tell you? You want someone to wipe your arse when you’ve had a shit? Do you dribble when you eat? Fucking pathetic.
Scientist,
are you really a 13 year old??
“You’re a fucking moron. An absolute fucking moron. Follow the methodology and make the graph. What the fuck else do you expect people to tell you? You want someone to wipe your arse when you’ve had a shit? Do you dribble when you eat? Fucking pathetic.”
Look, in March when the GISS avg tem for the month went up about .5c and everyone else went up .2c or more LESS, it can’t be covered by hand waving statistics and cursing.
GISS is not only warmer now, but, it is getting cooler in the past as they “adjust” their problem stations giving a stronger WARMING TREND!!!
Oh yeah, that also adjusts their reference period.
You lose. Now go suck your thumb.
PS: Why would anyone want to waste their time proving something pointless???
Why is it necessary to compare UHI biased urban stations with their rural neighbours which have NO such biase. I cannot understand why there is a need to mathematically tease out this UHI bias from the urban stations and use the resulting data when there are enough rural stations with uncontaminated data that could have been used in the first place. There are I understand an adequate number of historical rural stations available certainly in the northern hemisphere and especially in the US to totally discount all urban stations.
Please explain.
To be able to see signs of global warming linked to CO2 you need two things (1) accurate CO2 data and (2) accurate temperature data. There is accurate CO2 data but only I believe since 1958 at Mauna Loa, as for temperature, well there are thousands of high quality rural weather stations throughout the world and especially in the US and the northern hemisphere that have long histories and NO UHI bias. These stations are totally uncontaminated and provide temperature data from a totally natural environment like Mauna Loa does for CO2.
What do warmers such as Hanson/Giss do, well, they compare data from a highly UHI contaminated urban city weather station with CLEAN data from a neighbouring RURAL station. They then Use some secret algorithm put all the data through a computer which then gives them the result in degrees C. The odd thing is that the result always shows a steep rising temperature trend when the neighbouring RURAL CLEAN data shows only a flat or a slightly rising trend.
Would it not have been more logical to have discarded the contaminated data from the Urban stations and used only the clean data from the Rural sites. I believe the answer to this is that this method would not allow the use of this secret algorithm which makes it difficult to detect how much the data is being manipulated.
Bad data + good data = very good data, because it shows rising temperatures this is called Hansonisation.
Good data from rural weather stations = NO link between CO2 and temperature rise and we don`t want that do we. .