Adjusting Pristine Data

by John Goetz

On September 15, 2008, Anthony DePalma of the New York Times wrote an article about the Mohonk Lakes USHCN weather station titled Weather History Offers Insight Into Global Warming. This article claimed, in part, that the average annual temperature has risen 2.7 degrees in 112 years at this station. What struck me about the article was the rather quaint description of the manner in which temperatures are recorded, which I have excerpted here (emphasis mine):

Mr. Huth opened the weather station, a louvered box about the size of a suitcase, and leaned in. He checked the high and low temperatures of the day on a pair of official Weather Service thermometers and then manually reset them…

If the procedure seems old-fashioned, that is just as it is intended. The temperatures that Mr. Huth recorded that day were the 41,152nd daily readings at this station, each taken exactly the same way. “Sometimes it feels like I’ve done most of them myself,” said Mr. Huth, who is one of only five people to have served as official weather observer at this station since the first reading was taken on Jan. 1, 1896.

That extremely limited number of observers greatly enhances the reliability, and therefore the value, of the data. Other weather stations have operated longer, but few match Mohonk’s consistency and reliability. “The quality of their observations is second to none on a number of counts,” said Raymond G. O’Keefe, a meteorologist at the National Weather Service office in Albany. “They’re very precise, they keep great records and they’ve done it for a very long time.”

Mohonk’s data stands apart from that of most other cooperative weather observers in other respects as well. The station has never been moved, and the resort, along with the area immediately surrounding the box, has hardly changed over time.

Clearly the data collected at this site is of the highest quality. Five observers committed to their work. No station moves. No equipment changes according to Mr. Huth (in contrast to the NOAA MMS records). Attention to detail unparalleled elsewhere. A truly Norman Rockwell image of dedication.

After reading the article, I wondered what happened to Mr. Huth’s data, and the data collected by the four observers who preceded him. What I learned is that NOAA doesn’t quite trust the data meticulously collected by Mr. Huth and his predecessors. Neither does GISS trust the data NOAA hands it. Following is a description of what is done with the data.

Let’s begin with the process of getting the data to NOAA:

From Co-op to NOAA

Mr. Huth and other observers like him record their data in a “B91 Form”, which is submitted to NOAA every month. These forms can be downloaded for free from the NOAA website. Current B91 forms show the day’s minimum and maximum temperature as well as the time of observation. Older records often include multiple readings of temperature throughout the day. The month’s record of daily temperatures is added to each station’s historical record of daily temperatures, which can be downloaded from NOAA’s FTP site here.

The B91 form for Mohonk Lake is hand-written, and temperatures are recorded in Farenheit. Transcribing the data to the electronic daily record introduces an opportunity for error, but I spot-checked a number of B91 forms – converting degrees F to tenths of degree C – and found no errors. Kudos to the NOAA transcriptionists.

Next comes the first phase of NOAA adjustments.

NOAA to USHCN (part I) and GHCN

The pristine data from Mohonk Lake are subject to a number of quality control and homogeneity testing and adjustment procedures. First, data is checked against a number of quality control tests, primarily to eliminate gross transcription errors. Next, monthly averages are calculated from the TMIN and TMAX values. This is straightforward when both values exist for all days in a month, but in the case of Mohonk Lake there are a number of months early in the record with several missing TMIN and/or TMAX values. Nevertheless, NOAA seems capable of creating an average temperature for many of those months. The result is referred to as the “Areal data”.

The Areal data are stored in a file called hcn_doe_mean_data, which can be found here. Even though the daily data files are updated frequently, hcn_doe_mean_data has not been updated in nearly a year. The Areal data also seem to be stored in the GHCN v2.mean file, which can be found here on NOAA’s FTP site. This is the case for Mohonk Lake.

Of course, more NOAA adjustments are needed.

USCHN (part II and III)

The Areal data is adjusted for time of observation and stored as a seperate entry in hcn_doe_mean_data. TOB adjustment is briefly described here. Following the TOB adjustment, the series is tested for homogeneity. This procedure evaluates non-climatic discontinuities (artificial changepoints) in a station’s temperature caused by random changes to a station such as equipment relocations and changes. The version 2 algorithm looks at up to 40 highly-correlated series from nearby stations. The result of this homogenization is then passed on to FILNET which creates estimates for missing data. The output of FILNET is stored as a seperate entry in hcn_doe_mean_data.

Now GISS wants to use the data,  but the NOAA adjustments are not quite what they are looking for. So what do they do? They estimate the NOAA adjustments and back them out!

USHCN and GHCN to GISS

GISS now takes both v2.mean and hcn_doe_mean_data, and lops off any record before 1880. GISS will also look at only the FILNET data from hcn_doe_mean_data. Temperatures in F are converted and scaled to 0.1C.

This is where things get bizarre.

For each of the twelve months in a calendar year, GISS looks at the ten most recent years in common between the two data sets. For each month in those ten most recent years it takes the difference between the FILNET temperature and the v2.mean temperature, and averages them. Then, GISS goes through the entire FILNET record and subtracts the monthly offset from each monthly temperature.

It appears to me that what GISS is attempting to do is remove the corrections done by NOAA from the USHCN data. Standing back to look at the forest through the trees, GISS appears to be trying to recreate the Areal data, failing to recognize that v2.mean is the Areal data, and that hcn_doe_mean_data also contains the Areal data.

Here is a plot of the difference between the monthly raw data from Mohonk Lake and the data GISS creates in GISTEMP STEP0 (yes, I am well aware that in this case it appears the GISS process slightly cools the record). Units on the left are 0.1C.

Even supposedly pristine data cannot escape the adjustment process.

Advertisements

  Subscribe  
newest oldest most voted
Notify of

Mr. Goetz has exceeded himself, let’s give him a round of applause! – Anthony

Lord Kelvin weeps.

evanjones

Yes, very good.
(And why on earth doesn’t GISS simply adjust NOAA raw data?)

George M

That is a wonderful piece of detective work, showing that these climatic bit-heads can’t leave even the purest data alone, but it leaves several loose ends. I believe that in the original post, a copy of a form was shown which indicates the presence of an MMTS. Which data is so thoroughly massaged, MMTS or the Min/Max? If, indeed, Mr. Huth’s data is still used, what happens to the MMTS data? I would not expect it to be there if it is not used for something.
Also, I hope Mr. Goetz has sent a copy of this by registered mail to Anthony DePalma of the New York Times, pointing out how futile Mr. Huth and his predecessors’ work has been.
[Reply by John Goetz: There are several interesting mismatches between the B91 and NOAA MMS data, in addition to the MMTS issue. For one thing, Mr. Huth told the NYT that he recorded the temperatures at around 4:00 PM every day. However, every B91 I looked at signed by Mr. Huth indicated the time of observation was 5:00 PM. Maybe a nit, but hardly pristine.]

Dave Dodd

Will someone explain to me (and perhaps other lurking newbies) whether “a pair of official Weather Service thermometers” can be read to a granularity of 0.1F? My experience with mercury thermometers leads me to believe that reading “accurately” to 0.1 degree involves a high degree of subjectivity. My old science teacher in HS would have required us to read an order of magnitude greater than that or 0.01 degree and then round back. WIll someone please enlighen me?
[Reply by John Goetz: They are read to an accuracy of 1 degree F, but the conversion process to C is where 0.1 C “accuracy” comes into play. I have yet to see a B91 with a temperature recorded in anything other than full degrees. Not proof, of course, that they don’t exist.]
REPLY by Anthony: John, you are close but not quite correct. The thermometer reading is read in 0.1F resolution, but then rounded to the nearest degree F at the time the observer makes the reading and writes it down on the B91 form.

Jeff B.

(And why on earth doesn’t GISS simply adjust NOAA raw data?)
Well that would give Hansen less opportunity to hide, and then find warming.

dearieme

They were nincompoops before they were crooks.

Jeff Alberts

REPLY by Anthony: John, you are close but not quite correct. The thermomter reading is read in 0.1F resolution, but then rounded to the nearest degree F at the time the observer makes the reading and writes it down on the B91 form.

Which means the margin of error is at least as much as the purported warming of the late 20th century. Yup, high quality.
Pardon me while I go puke.

Dave Dodd

Read to 0.1F and record as an integer — GREAT! Simple math rules! My old science teacher will rest peacefully! However, if your B91 data are integers and you average 10,000 integers you still get an integer as the end result, even if you change to a different metric system! The oft-cited AGW temperature rise of 0.7C/century is mathmatically incorrect, is it not? One often sees temps cited to 0.01 degree for data sets presumably from the nineteenth century. Somebody’s finagling!

Mike C

John,
I’ve seen this before. The V2 file you are using is mislabeled. It is actually late V1 stuff… you get TOB, Homogeneity and Filnet adjustments, except the Homogeneity adjustment is actually the version 1 Homogeneity adjustment (SHAP, for documented discontinuities and Karl et al 1988 for urbanization effects). Since this station had no documented discontinuities, there should be no changes for SHAP. Then they employ Karl et al 1988 to adjust for urbanization, which is basically averaging the local USHCN stations.
Since Hansen uses his own urbanization scheme (night lights), he subtracts the Karl et al urbanization adjustment then applies his nightlight adjustment, which for this station will be no adjustment because this is a lights = zero station.
So let’s try to untangle here a little:
Raw Data
plus
TOB
plus
SHAP aka high frequency variation (actually a value of zero)
Plus
Karl et al 1988 (urbanization) aka low frequency variation
Plus
Filnet
Then Hansen takes over;
Minus
Karl et al
Plus
Hansen Night Lights adjustment (actually a value of zero)
Hansen can only use the V 1 homogeneity adjustment scheme because it separates the SHAP and Urbanization adjustments.
You see, the USHCN V2.0 is Hansen’s little problem because the V2.0 Homogeneity adjustment is sold by the NCDC as adjusting for both high and low frequency variation (adjustments for station discontinuities AND urbanization). Hansen cannot subtract any urbanization adjustment in V2.0 then add his night-lights scheme.
Now, here is where it gets really, really funky: Go to the KBSF home page here:
http://home.earthlink.net/~ponderthemaunder/index.html
flip down to the story about how NCDC wants Urban Heat Islands in the USHCN record. Then click on the link to Claud Williams’ powerpoint presentation at the AMS, Jan 2006 and tell me if you figured out why neither NCDC nor Hansen have properly adjusted schemes.

evanjones

Well, with oversampling I guess you can fine it down, but I do think the 0.1 claim is a bit tight.

Bobby Lane

Are you kidding me? Back and forth and back and forth conversions? Estimates? Algorithms? Averages? I’m terrible at math myself, but this just screams for the existence of room for small inaccuracies that will add up overtime to big changes. I mean, if each day was adjusted up one one-thousanth of a degree over the period of a year you get nearly 4 tenths of a degree rise (.001 x 365 = .365) in that time span. No doubt real warming is taking place at times because of the LIA recovery, but such an addition continuing over time becomes rather extreme. This process reminds me of a game of Telephone. What was said in the beginning and what comes out in the end is bound, by accident or purpose (or both), to be different. Talk about parallel dimensions!

So, from the news before the “Hansen Returns” visit to Congress where “James Hansen, one of the world’s leading climate scientists, will today call for the chief executives of large fossil fuel companies to be put on trial for high crimes against humanity and nature”
It seems to me this poor, defenseless Mohonk data that never harmed no one has had barely speakable but high crimes committed against algorithms and nature. It may not be a martyr, but perhaps it can be a poster child.
So, do we have the raw Mohonk data? How about a graph?

JFA in Montreal

And kudos for your choice of picture on that post.
Very à propos ! 🙂

crosspatch

So if I take 100 readings with recorded temperatures in whole degrees, add them together and divide by 100, that give me a number to a precision of two decimal places! Every year the data gets more “accurate”!
/sarcasm

Eric Anderson

What Jeff Alberts said.

Mike McMillan

As I recall from my 1960’s non-digital-era geodesy course, you can get accuracy greater than your precision with a sufficient number of measurements, but this is just the opposite, rounding to 0.5 degrees when your precision is 0.05 degree. Not quite ‘measure with a micrometer, cut with a chainsaw,’ but I suppose it met the requirements of the time.
Obviously back in 1896 they didn’t know that mere tenths of a degree could spell life or death for a penguin, polar bear, or buckeye tree.

Demesure

Very nice presentation John Goetz, thank you.
“The quality of their observations is second to none on a number of counts,”
The observers should look at what the GISS has done to their “second to none” Mohonk lake data: no monthly data in the online database since 2007 (filled with 9999) !
2007 999.9 -6.8 1.4 6.8 16.0 999.9 999.9 999.9 999.9 999.9 999.9 999.9

Bobby Lane

I need a little help with something. I am reading elsewhere (i.e., not on this site) but I cannot make sense of a certain statement. Regarding the Eocene period, it is stated that:
“At the same time, however, equatorial temperatures were found to be about 4K colder than at present.”
I thought that might be four thousand at first, but then I remembered I was dealing with temperature. So I assumed the K is degrees Kelvin. Well, I googled a converter so I could find out what that meant in terms I could understand (Fahrenheit). It converted 4K to -452.5. But somehow reading the statment that “equatorial temperatures were found to be about 453F cooler than at present” is a bit difficult to stomach. Am I making a methodological error?
Here is the paragraph in that paper from which it comes, which is interesting reading in itself.
“In the first example, the original data analysis for the Eocene (Shackleton and Boersma, 1981) showed the polar regions to have been so much warmer than the present that a type of alligator existed on Spitzbergen as did florae and fauna in Minnesota that could not have survived frosts.
At the same time, however, equatorial temperatures were found to be about 4K colder than at present. The first attempts to simulate the Eocene (Barron, 1987) assumed that the warming would be due to high levels of CO2, and using a climate GCM (General Circulation Model), he
obtained relatively uniform warming at all latitudes, with the meridional gradients remaining much as they are today. This behavior continues to be the case with current GCMs (Huber, 2008). As a result, paleoclimatologists have devoted much effort to ‘correcting’ their data, but,
until very recently, they were unable to bring temperatures at the equator higher than today’s (Schrag, 1999, Pearson et al, 2000). However, the latest paper (Huber, 2008) suggests that the equatorial data no longer constrains equatorial temperatures at all, and any values may have existed. All of this is quite remarkable since there is now evidence that current meridional
distributions of temperature depend critically on the presence of ice, and that the model behavior results from improper tuning wherein present distributions remain even when ice is absent.” (from page 10)

Jon

A question came to mind: how does TOBS handle DST?

Bobby Lane

Nevermind, I got it. I put in, say 98 degrees F, converted it to K, subtracted 4 degrees from the K result, and converted it back to F. That makes more sense. The point was just that the equatorial regions were cooler than at present (which was stated) but not necessarily cold (which is what my first, and incorrect, calculation made me think).

Jan RH

John G. states: “yes, I am well aware that in this case it appears the GISS process slightly cools the record.”
But, correct me if I’m wrong, doesn’t the fact that Mohonk-GISS temp’s are going from positive to negative just mean that GISS temp’s are getting comparatively higher as time goes by?
Which means that the GISS process produces warming in the record.

Leon Brozyna

The words that come to mind — Rube Goldberg.

MattN

“This article claimed, in part, that the average annual temperature has risen 2.7 degrees in 112 years at this station.”
And 2.7 degrees appears to be exactly the amount of the overall adjustment since 1896….
Reply – However, surrounding stations don’t show that amount of increase and some even show a decrease. See Calling All Climate Sleuths for an example. – Dee Norris

It all started with the 2.7 Fahrenheit (?) per hunderd years increase in the Mohonk Lakes data. Are the pristine data also giving that increase? How did NOAA change the increase, and, finally, what did GISS procedures do to it?

Chris H

“For one thing, Mr. Huth told the NYT that he recorded the temperatures at around 4:00 PM every day. However, every B91 I looked at signed by Mr. Huth indicated the time of observation was 5:00 PM. Maybe a nit, but hardly pristine.”
Maybe he really *does* take temp measurements at 4pm, but the time recorded in the database is adjusted for Summer Time? (Would make sense since Summer Time is a purely human convention, which might otherwise complicate analysis?)

Phil M

John RH
“John G. states: “yes, I am well aware that in this case it appears the GISS process slightly cools the record.”
But, correct me if I’m wrong, doesn’t the fact that Mohonk-GISS temp’s are going from positive to negative just mean that GISS temp’s are getting comparatively higher as time goes by?
Which means that the GISS process produces warming in the record.”
Yes – I’d agree with your reading of that
– as shown (Mohonk – GISStemp) shows that the GISS process produces warming….surprise!

A very good and entertaining post; I knew about the adjustments previously, but hadn’t realised how much this resembles a kind of sausage factory for numbers. The long-winded process reminds me of the “think of a number” trick that we used to astound our friends with at school (at age 6 or thereabouts): “Add 100, then take away 5, then take away the number you first thought of…” Magic! No wonder it all somehow “adds up” to Global Warming. :o)

Bobby Lane
A degree Kelvin is the same size as a degree Centigrade. The C scale has zero at the freezing point of water, the K scale has zero at absolute zero. Your conversion method was correct.

Phil M

Rounding to integers (F)
– yes, it’s a pity that they do this
– which just goes to show that the temperature monitoring was never intended for the purpose for which it is now being used….
– but, by using large enough samples it *is* possible to recover the information that was lost in the rounding process.
– that’s the benefit of using a large sample set – the error introduced by the rounding can be effectively eliminated by using taking readings from many places
– although I do agree that there is an overall problem with the accuracy of the whole system, which Anthony has pointed out many times
– hence all the ‘correction’ factors that get applied….

Peanut Gallery(formerly know as the artist Tom in Florida)

This is a classic example of taking a simple thing, adding a large dose of government and presto:
awholebunchofstuffthatisallmixedupanddoesn’tdowhatwasoriginallyintended

Mike Bryant

Thanks John Goetz,
This is just mind-boggling. Of course, I’ve read of these adjustments, but to see them laid out like this…
I guess putting this data through all these acrobatics makes them oh so perfect.

Two words come to mind….
“Paralysis By Analysis”
Why in the world the information gets manipulated is beyond me…
http://www.cookevilleweatherguy.com

*LOL*…or was that 3 words?? Haven’t had enough coffee yet! 🙂

Mike Bryant

Like I said Phil, oh so perfect. Shame on you. I thought you were a scientist. Or was it sarcasm?

Phil M

Slightly off topic
– In September Satellite temps (lower troposphere, AMSU)
– it looks like this month is going to come out with an anomaly of around +0.2C
– the highest for this year…
It will be interesting to see if this coming winter is as cold as the last one
– or if we return to regular anomalies of around +0.2C….

John Goetz

Mike C says:

The V2 file you are using is mislabeled.

I don’t think so. GISS is clearly using that dataset as I confirmed reading their STEP0 code. For Mohonk Lake I also confirmed that it is the Areal version found in the referenced USHCN file, simply by subtracting Areal from v2.mean and getting 0.

John Goetz

Jan RH and Phil M:

But, correct me if I’m wrong, doesn’t the fact that Mohonk-GISS temp’s are going from positive to negative just mean that GISS temp’s are getting comparatively higher as time goes by?

Be careful. The graph is “GISS – Raw”. GISS gives a higher temperature in the early years and lower in the later years. This means that GISS slightly cools the record.
I have plotted the monthly trend for both Raw and GISS and run a linear trend through them. Both show an almost imperceptible warming trend – the only way to see the trend, IMO, is to look at the slope value for the trend equation.

MarkW

Is Hansen starting to lose it?
http://www2.ljworld.com/news/2008/sep/23/nasa_climate_expert_warns_dire_consequences_global/
“If we don’t get this thing under control we are going to destroy the creation,” said James Hansen,

Paul (04:19:39) :

Bobby Lane
A degree Kelvin is the same size as a degree Centigrade. The C scale has zero at the freezing point of water, the K scale has zero at absolute zero. Your conversion method was correct.

Yes, but your terminology is confused. Technically “degree Kelvin” shouldn’t be used. It may have happened when Centigrade was renamed Celcius (and cycle/sec Hertz, etc), but Kelvins were redefined to make them used more like other measurements.
The phrase “about 4K colder than at present” does have one confusion, i.e. K is for Kelvin and K is a prefix for a 1000 multiplier. The K in 4K isn’t a prefix, so it must be Kelvins. Consider “that pipe is 1m shorter than the old one.” You’d know that if the old pipe was 1.618 meters long, you’d know the the new pipe was 0.618 meters long. A melting ice cube’s temperature is 273 Kelvins or 0 degrees Celcius, not 273 degrees Kelvin. “About 4K colder than at present” is the same as “4 degrees C colder.”
In dog nights, it’s about one dog, i.e. a three dog night would be a four dog night.

Bill Marsh

Mark W,
“Starting to lose it”?

MarkW (05:37:56) :

Is Hansen starting to lose it?
http://www2.ljworld.com/news/2008/sep/23/nasa_climate_expert_warns_dire_consequences_global/
“If we don’t get this thing under control we are going to destroy the creation,” said James Hansen,

I think so. He seems to be both becoming more and more messianic in his speeches and seems to be reaching out to large forums. There was a rock & environmental festival in the spring (that got mostly rained out) that he spoke at. From the coverage it seemed he was trying to expand his flock. As he seems to be losing his support from science as reality, he seems to be drawing more and more on a faithful following. Evan suggested that Hansen not be forced from NASA lest he become a martyr, my sense is the sooner the better. I do think that any Hansen watcher look beyond the science to try to figure out where he’s going.

“science as reality”? I meant “science and reality,” though it seems to work either way. 🙂

Harold Ambler

I sent an e-mail to Benjamin Cook, the NOAA meteorologist whose data was used by the New York Times for its Mohonk House article, asking him about a year at the turn of last century shown to have a sub-freezing average temperature. Having lived in the Northeast for most of my adult life, I knew this was pretty unlikely! This is what he said:
“It turns out that the graphics people at the times converted between Celsius and Fahrenheit incorrectly, so the temperatures in the graph were way too cold (although the shape of the curve and the trends were the same). I’ve already notified them, and they said they would be fixing the online graphic.”
It was good of Benjamin to get back to me. Hopefully, the Times folks will do what they have promised.

Harold Ambler

P.S. The graph in the Times article, using the incorrectly converted figures, shows a 20-degree swing from the coldest annual temperature to the warmest. This also seems pretty surprising, and I have sent an e-mail to Benjamin asking him about it.

Successively rendering significant figures insignificant, then “creating” significant figures and adjusting them back into insignificance is mind boggling. At some point in this process, “data” disappears and is replaced by “number sets” which purportedly represent what the data sets shoulda/coulda/woulda looked like had they been collected timely from properly installed and calibrated instruments in the first place.
The suggestion that the denizens of this globe should invest more than $100 trillion to correct a “problem” projected to occur based on these “supple” number sets is laughable, at best.

Dan McCune

I just checked with http://www.surfacestations.org/ and this site has not been surveyed. Anyone near by who could verify the equipment is reliable as the measurments?
Surveyed?
Active?
lat
long
CRN Rating*
USHCN ID 305426
Station name MOHONK LAKE
Distance from Post office 0.1
dir_PO SE
State NY
lat 41.77
long -74.15
GHCN ID 72504006
Elev (ft) 1245
location MOUNTAIN HOTEL ON LAKE 4 MILES WNW OF PO AT NEW PALTZ, NY
MMS id 20026
Reply – Ummm, Yes. See Calling All Climate Sleuths – Dee Norris

Bill Illis

Anthony’s trip to the NCDC this spring allowed us to see how much they are adjusting the raw temperature records with these adjustments.
USHCN V2 has two adjustments for TOBS (increases the average trend by 0.2C) and the Homogenity Adjustment (also increases the trend by 0.2C). So the adjustments increase the overall temperature trend (in the US) by 0.4C compared to the raw data.
These adjustments are shown in Slide 7 and Slide 16 of the powerpoint given to Anthony by the NCDC (not shown anywhere else on the net that I have seen).
(this link locks up sometimes)
http://wattsupwiththat.files.wordpress.com/2008/05/watts-visit.ppt#256,1,U.S. HCN Temperature Trends: A brief overview
Original can be found in this post by Anthony.
http://wattsupwiththat.com/2008/05/13/ushcn-version-2-prelims-expectations-and-tests/
Of course all the adjustments in USHCN V1 are shown in this chart (0.55F or 0.3C) So Version 2 increases the trend by a further 0.1C compared to Version 1.
http://www.ncdc.noaa.gov/img/climate/research/ushcn/ts.ushcn_anom25_diffs_urb-raw_pg.gif

Phil M

John Goetz
– sorry, I mis-read the graph title (several times!)
Mike Bryant
– me? I wasn’t being sarcastic
– you can get data back (statistically) that’s been ‘lost’ in rounding…
– you just need enough spacial or temporal samples…
– which is what the GISS software it trying to do by using thousands of sites…
– but it is interesting that the trend we’re looking for is of the same magnitude (per century) as the rounding error.
– persumably the results were rounded to 1F because it was felt that this was the useful limit of accuracy of the data, which is also interesting (i.e. the raw data isn’t very accurate)

This is why I don’t trust the NOAA and GISTemp datasets. If they do this to temperature readings within the USofA, Lord knows what they do to temperature readings from the ROW. I still haven’t figured out what happens to Canadian data.