MEDIA ADVISORY: 96% OF U.S. CLIMATE DATA IS CORRUPTED
Official NOAA temperature stations produce corrupted data due to purposeful placement in man-made hot spots
Nationwide study follows up widespread corruption and heat biases found at NOAA stations in 2009, and the heat-bias distortion problem is even worse now
ARLINGTON HEIGHTS, IL (July 27, 2022) – A new study, Corrupted Climate Stations: The Official U.S. Surface Temperature Record Remains Fatally Flawed, finds approximately 96 percent of U.S. temperature stations used to measure climate change fail to meet what the National Oceanic and Atmospheric Administration (NOAA) considers to be “acceptable” and uncorrupted placement by its own published standards.
The report, published by The Heartland Institute, was compiled via satellite and in-person survey visits to NOAA weather stations that contribute to the “official” land temperature data in the United States. The research shows that 96% of these stations are corrupted by localized effects of urbanization – producing heat-bias because of their close proximity to asphalt, machinery, and other heat-producing, heat-trapping, or heat-accentuating objects. Placing temperature stations in such locations violates NOAA’s own published standards (see section 3.1 at this link), and strongly undermines the legitimacy and the magnitude of the official consensus on long-term climate warming trends in the United States.
“With a 96 percent warm-bias in U.S. temperature measurements, it is impossible to use any statistical methods to derive an accurate climate trend for the U.S.” said Heartland Institute Senior Fellow Anthony Watts, the director of the study. “Data from the stations that have not been corrupted by faulty placement show a rate of warming in the United States reduced by almost half compared to all stations.”
NOAA’s “Requirements and Standards for [National Weather Service] Climate Observations” instructs that temperature data instruments must be “over level terrain (earth or sod) typical of the area around the station and at least 100 feet from any extensive concrete or paved surface.” And that “all attempts will be made to avoid areas where rough terrain or air drainage are proven to result in non-representative temperature data.” This new report shows that instruction is regularly violated.
For more information, or to speak with the authors of this study please contact Vice President and Director of Communications Jim Lakely at jlakely@heartland.org or call/text 312-731-9364.
This new report is a follow up to a March 2009 study, titled “Is the U.S. Surface Temperature Record Reliable?” which highlighted a subset of over 1,000 surveyed stations and found 89 percent of stations had heat-bias issues. In April and May 2022, The Heartland Institute’s team of researchers visited many of the same temperature stations as in 2009, plus many not visited before. The new survey sampled 128 NOAA stations, and found the problem of heat-bias has only gotten worse.
“The original 2009 surface stations project demonstrated conclusively that the federal government’s surface temperature monitoring system was broken, with the vast majority of stations not meeting NOAA’s own standards for trustworthiness and quality. Investigations by government watchdogs OIG and GAO confirmed the 2009 report findings,” said H. Sterling Burnett, director of the Arthur B. Robinson Center on Climate and Environment Policy at The Heartland Institute who surveyed NOAA surface stations himself this spring. “This new study is evidence of two things. First, the government is either inept or stubbornly refuses to learn from its mistakes for political reasons. Second, the government’s official temperature record can’t be trusted. It reflects a clear urban heat bias effect, not national temperature trends.”
An example of the bias problem
The chart below, found on page 17 of the report, shows 30 years of data from NOAA temperature stations in the Continental United States (CONUS). The blue lines show recorded temperatures and the trend from stations that comply with NOAA’s published standards. The yellow lines are temperatures taken from stations that are not compliant with those standards (i.e. near artificial hot spots). The red lines are the “official” adjusted temperature released by NOAA.
“If you look at the unperturbed stations that adhere to NOAA’s published standard – ones that are correctly located and free of localized urban heat biases – they display about half the rate of warming compared to perturbed stations that have such biases,” Watts said. “Yet, NOAA continues to use the data from their warm-biased century-old surface temperature networks to produce monthly and yearly reports to the U.S. public on the state of the climate.”
“The issue of localized heat-bias with these stations has been proven in a real-world experiment conducted by NOAA’s laboratory in Oak Ridge, Tennessee and published in a peer reviewed science journal.” Watts added.
“By contrast, NOAA operates a state-of-the-art surface temperature network called the U.S. Climate Reference Network,” Watts said. “It is free of localized heat biases by design, but the data it produces is never mentioned in monthly or yearly climate reports published by NOAA for public consumption.
The Heartland Institute, a free-market think tank founded in 1984, is one of the world’s leading organizations promoting the work of scientists who are skeptical that human activity is causing a climate crisis.
Heartland has hosted 14 International Conferences on Climate Change attended by thousands since 2008, published the six-volume Climate Change Reconsidered series by the Nongovernmental International Panel on Climate Change, and for 21 years has published Environment and Climate News. The Heartland Institute has also published several popular books on the climate, including Why Scientists Disagree About Global Warming (2015), Seven Theories of Climate Change (2010), and Is the U.S. Surface Temperature Record Reliable? (2009).
###
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Excellent work, Anthony Watts and others!
These links below show another way to look at the USHCN and nClimGrid data, using the daily values of Tmax. I realize this does not substantiate any claim about the long-term trend one way or the other, but in my view shows clearly that the attribution of any trend at all to slow increases in CO2 and other non-condensing GHGs is unsound, or at least very weak at best.
First, these are plots by date of the 5-year mean of the average of all USHCN daily Tmax for all stations reporting a value, for the years 1895-2021. So the data points plotted are from 1897 to 2019. The point is that the cycles and trends differ greatly by date, indicating “timed” influences that combine or cancel. I have previously shared these plots in comments on “open thread” and perhaps other posts on WUWT. The data is available here: https://www1.ncdc.noaa.gov/pub/data/ghcn/daily/hcn/
My plots are here:
https://drive.google.com/drive/folders/102pEDMex4dPGqADMkIqiNmPJ-YV1SdMR?usp=sharing
Second, these are similar plots of the 5-year mean by date from the nClimGrid database, for the reported values for CONUS daily Tmax. There is a “filenames.txt” document in this folder that gives the origin of the data. Each file can be located by prepending “https://” to the filename. Same result about the cycles and trends. It is lamentable that the nClimGrid database does not extend farther back in time. The data is from 1951 to 2021, so the plotted values are from 1953 to 2019.
My plots are here:
https://drive.google.com/drive/folders/1-M2olMyCGexz5vgNAAXJxu0DhGDeWaKM?usp=sharing
In each of these folders linked above, the best way to get the point is to click through the images one by one and note the changing shape of the plot.
So, the bottom line, in my view, is that EVEN WITH the siting issues, adjustments, and perhaps other factors, the daily data reveals that whatever is affecting the contiguous U.S. climate is not explained by slowly increasing concentrations of CO2. There must be other strong influences, and they look to me like timed cycles combining and canceling differently by date.
“There must be other strong influences…” At surface level, water vapor molecules outnumber CO2 molecules 24 to 1 and have been increasing 7 times faster than CO2 molecules. The increase in WV can account for all of the climate change attributable to humanity. http://globalclimatedrivers2.blogspot.co
David D,
These are fascinating and important graphs (day by day Tmax for a century).
I have done related work on Australian Tmax data, but more as station after station. Not national averages. Also, a lot of heatwave work based on daily Tmax
Here are some graphs. Numerical data are freely available.
A big update and expansion is nearly finished.
Note that some severe effects of homogenisation are everywhere visible among the graphs and troublesome, see data labelled ACORN that is version 2.2 from December 2021. Geoff S
http://www.geoffstuff.com/asixheatwave2022.xlsx
Geoff S, thank you for that data. I see what you mean. I look forward to your update.
I appreciate that you took the time to look at the plots I posted.
And yes, the 5-year mean of the average day-by-day Tmax data for a well-instrumented country like the U.S. cleans up the plot from the “weather” and leaves the unmistakable residual cycles and trends. By the way, the Tmin data gives a similar result, so the fact that I use Tmax is because heat waves get a lot of attention.
David,
Could it be that some of the patterns you show arise from the cooling effects of rainfall and subsequent soil moisture changes that can linger on for years, also affecting local vegetation? That is an impression I get without a detailed study of the rain, so it is not to be taken seriously, Geoff S
Geoff, no doubt to the extent that the underlying influences may induce more or less rainfall, then the plots would include that cooling effect or the lack of it. Suppose the timed influences combine to promote more low pressure systems over the U.S. on some dates and fewer on other dates in the same year or series of years. The effect on temperatures would have contributed to the plotted results.
But I go back to the effect of the national averaging and the smoothing by using the centered 5-year mean. It tends to blur the effect of local variable precipitation and to reveal the nature of the underlying influences. For example, look at the USHCN plots for May 3, 10, 17, 24, 31. All the same month, a week apart, but the trends look very different. So in whatever way the timed influences may affect the weather – sunshine, humidity, cloudiness, precipitation, wind, high and low pressure systems – it shows up in the plot but it may not be possible to isolate one factor vs. another in reverse.
Just my thoughts. There must be a lot more to this phenomenon which we don’t understand.
David!
Wow! Nice plots. There does appear to be cyclic effects of some kind. Just a quick visual look seems to indicate a 20 year cycle of some kind. It also shows the 30’s as being much warmer than today even with lower CO2.
I’ll probably download some of these for later analysis and for use as assertions against the climate alarmists.
If you did graph the Tmin data you might pass along the site for them.
Thanks,
Here is the link for Tmin for USHCN. I didn’t do Tmin for nClimGrid. My sense is that there must be multiple timed influences of different periods, so the combinations produce some longer and some shorter apparent cycles.
About the 1930’s, yes, warmer especially in July and August for many of those dates. Not for all dates in the year though. And that is the key thing about what these plots imply.
https://drive.google.com/drive/folders/1_PhFDKTQ1Wtz-xGKyWNyA0djiBeqAv9H?usp=sharing
Some notes about the USHCN data when sharing:
The nClimGrid plots for Tmax for the shorter time period do not have any of these caveats.
I appreciate your interest.
OK, why can’t someone gather the data from the hundreds/thousands of quality personal weather stations that are live on the internet? Crowd source the data and partition it into geographic classes: Urban,Suburban, Agricultural, Recreational (golf courses?), etc. My 10 year old LaCrosse weather station is higher quality than the lab grade stuff we had when I was in school 60 years ago.
Yeh, strange how no-one does that…
I think that it would be more revealing to plot Tmax and Tmin from the corrupted and un-corrupted stations rather than Tmean. It would also be instructive to include the error bars. I’m not a fan of collapsing all the data into a single number. We know that Tmax and Tmin are increasing at different rates and for different reasons.
Are you sure about Tmax?
I thought it was just Tmin
TMax would go up from asphalt.
Yes, I’m sure. See Fig. 2 at
http://wattsupwiththat.com/2015/08/11/an-analysis-of-best-data-for-the-question-is-earth-warming-or-cooling/
The link to the Heartland Institute doesn’t work. Is the website still active? Thanks.
no issues that I can find?
I clicked the Heartland Institute plus another at random and they worked.
It pleases me to see the Oxnard station singled out for special mention.
I am shocked, SHOCKED, to learn that the climate activists in government are deliberately using bad data that shows what they want.
Who am I kidding, none of us are the least but surprised by this.
In any case, nowhere is immune from artificial bias such as the Urban Heat Islands
Thanks to modern industrial farming, those islands are now situate in vast Rural Heat Oceans
Those places are artificially heated exactly the same way as the cities are – by lowering of the Albedo and via the dessication of entire landscapes.
Water that would have been sitting just below ground in high water tables has been collected in reservoirs and piped into cities
Also, the water table has been lowered by farmers and others pumping out water for irrigation.
On top of that, literally, the soil organic matter has been systematically stripped out by Roundup and Tillage, so there’s no water/dampness in the surface layers, to both cool directly and for evapotranspiration by plants.
It’s amazing that temps haven’t risen more than they have.
Repeat: Deserts might have high temperatures but they are Cold Places – there’s so little energy there in them.
Peta,
The Sahara desert has come and gone several times without the Hand of Man.
We do not know the mechanisms in any useful detail.
Why bother with tiny matters like you mention when bigger issues abound? Geoff S
I blame government politicians, administrators and bureaucrats. They are inept and evil.
.Peta stop your moaning .
The worlds population has gone from 2 billion to near 8 billion in 80 years .
Grassland farming as practiced in New Zealand and many other countries builds soil depth and sequests carbon .
I know this from many years of farming and observation .
Red clay tracks bulldozed around the sides of hills for access soon grass over and black soil soon covers the clay .
Shallow soils I have chisel plowed to aerate the soil and then cropped and regrassed are now deep soils with the addition of lime and fertilizer and they grow longer in the dry spells.
I have read about farmers in Australia who are engaging soil scientists to measure the carbon that is building up in their soils to claim carbon credits .
Here in NZ we are up against the climate commission who will not recognize trees and other land in strips around rivers etc are sequestering carbon and they will not even consider soil carbon .
They are anti farming .It makes you wonder what the eat .
Did you just reply to the wrong post?
Thanks for posting your findings. Disappointing, but not surprising.
It is shocking to see important government agencies like NOAA and NWS with such a cavalier attitude toward standards – especially their own. Most standards are set to provide clear unambiguous requirements that, if not met, are disqualifying. In the private sector non-compliance with product standards is grounds for mandatory product recalls or cancellation of contracts. Being close to complying is no defense.
It seems the NOAA/NWS position on weather stations can be summed up in the phrase “good enough for government work”. What Anthony er.al. have demonstrated is that the historical data available is mostly poor quality and hopelessly biased. All the adjustments and “corrections” applied by the government bureaucrats amounts to trying to make a silk purse from a sow’s ear.
It is much the same here in the UK Temperature by night and by day are always
much warmer than the surrounding areas, Heathrow airport which is very close to London, is always two to three degrees Centigrade higher. and the met office and the media recognise this, yet they still quote the temperature here.
In 1974 The Met office allowed for this by adding 0.2 C and it has not been changed since then, although urban development has increased dramatically over the past 50 years. The media are very keen to report any extreme weather event
and add climate change as the reason. without checking the past records.
On average we get a very warm summer every 10 years based on the CET temperatures, which go back to 1650
I found it interesting that based on a 10 year running mean that summer temperatures were almost as warm 300 years ago 1727/ 36 was the warmest
I could find at 15.90 C and the warmest recent ten year period was 15.95
But winter temperatures were much colder.
Somewhere some climate scientist is thinking – “Who cares what the weather stations say, we’ve got computer models!”
Yes , and somewhere some weather station recorded –
“Harry Read Me”
That rationale has been used to dismiss historical pH measurements in the oceans.
Comparing stations at 2 points in time (2009-2022) we find the quality is dropping. But over an 11 year period there may be short term trends of interest, as station locations and environments change more than once. An example would be Panguitch, UT. The 2009 location is truly terrible.
The current location has an upgrade to CRN4, with cited issues of shading and irrigation. For a period of time between these two dates, there was a 3rd location that I photographed in 2010. I’m not competent to make the call, but perhaps it could have been as high as CRN2. At any rate it doesn’t seem to have a shading or irrigation problem.
My impression is that there is a real problem getting volunteers to take on this responsibility, and they generally will take what they can get.
Does the grass under the device ever green up? If it does then that will affect the calibration of the station!
Hi Tim,
Lacking evidence of irrigation, I presume the vegetation of the entire lot reflects the natural growth of the entire region. I should perhaps have included the other pictures of the site. The MMTS gauge is in the back ground of that picture. It is not as close to the house as it appears, but it is probably too close to qualify for CRM2. Here’s another shot from a different position:
I wonder if there is a way of suing them for knowingly publishing false and misleading data. The false info is being used to convince the country to spend billions of dollars to fight climate change.
That’s what the tort system is for.
The government does an excellent job of hiding any of their data which would conflict with the establishment narrative of catastrophic climate change. Often it is there, but you have to look for it and if it’s contrary data, it won’t automatically be charted for you. Rather than trend plots, they prefer heat maps which can be made to look scary regardless of what the data actually shows.
The Climategate emails revealed one where the Australian BoM keeper of the records (Jones?) boasted to the other members of “The Team” how he snowed any requests for temperature records by sending unsorted file dumps.
Shows what the mentality is – stonewall by any means possible.
CONUS occupies less than 2% of the planet. The best assessment of average global temperature trend (no UHI effect) is obtained from satellite based instrumentation reported monthly by UAH at https://www.nsstc.uah.edu/data/msu/v6.0/tlt/uahncdc_lt_6.0.txt UAH 6.0 was tracked fairly closely by RSS 3.3 until they abandoned it to join the surface temperature measurement mob.
Be careful even of UAH. UHI can affect even the low troposphere as seen from the satellites. How would they compensate for this?
That’s nonsense. 72% of the planet is ocean. Any air temperatures that are raised in urban areas are probably quickly mixed away above a few hundred feet.
“That’s nonsense. 72% of the planet is ocean. Any air temperatures that are raised in urban areas are probably quickly mixed away above a few hundred feet.”
If that were true then how can gliders ride thermals to 25,000′ to 90,000′ over UHI sites, e.g. large cities or other heat sites like deserts? How can thunderstorms tower to 30,000′ – 70,000′?
Note that the UAH data is global – take satellite data from the whole lower troposphere. nClimdiv data is from (on land) point sources: surface stations where older sites have become more urbanized and newer sites are typically at airports. The nClimdiv data is likely to have much greater impact of UHI than the UAH data. UAH data is showing that the atmosphere is warming at 0.13C per decade, I don’t think that has been broken down to UHI, CO2, power plants, few point increases from cooling towers, agriculture, etc.
The US is 6.7% of the Earth’s land area, which is what’s being compared here. The global anomaly is predicted by using land stations, so it’s appropriate to compare land area with land area.
If the anomaly is predicted using all GHCN stations except for those in the US, and then the anomaly is predicted using the GHCN stations within the US what difference would you expect to see? Would the US be warmer? Cooler? The same?
Manchin capitulates; forfeits title as “The Dude”.
Saw that. Manchin didn’t even wait for the next inflation report.
Earlier he said he wouldn’t vote for more spending until he saw the next inflation report, but I guess some pressure from somewhere made him change his mind.
I wonder what it was that makes him vote against his constituents best interests by raising their taxes and raisng the inflation they have to pay. Joe looks like he’s gone off the deep end. His vote will be detrimental to the people he represents. Joe Manchin did not help his poltical future with this decision.
Just use UAH satellite data and ignore NOAA if it appears biased.
Excellent idea. That’s what I do. 🙂
Even UAH will pick up urban heat conducted/convected into the lower troposphere. There’s no way to compensate for this!
Not meaningful amounts.
This makes me so angry I could just spit, preferably at whomever the corrupt jerks are at NOAA and wherever else who are creating this idiotic CAGW panic and trying to bankrupt the whole world with China and Russia excluded, of course, because they aren’t dumb enough to fall for this nonsense and get with the (Great Reset?) “program.”
We now know we can no longer trust the FDA, the CDC, the NIH, the DOJ, the FBI, the White House, the CIA, the MSM, and, now, NOAA and it’s related agencies and NASA. I’m sure I left out many other deserving agencies in Government and in the private sectors, especially, corporations and the super wealthy who are buying our government lock, stock and barrel.
A corrupt administration corrupts all its agencies by putting corrupt people in charge of them.
how can you trust a government that verified the 2020 election?……..that biden got over 15 million votes more than obama while getting FEWER votes from blacks, hispanics, and women…..
by my calculations, if this trend continues by 2091 most temperature stations will be on fire
Intentionally or by alarming temperatures?
Re: ADJUSTMENTS
This is important.
Read, learn and inwardly digest the significance of this study by colleague Chris Gillham.
It is mostly in pictures. But, oh what a sorry picture it paints.
The go back to the whole of his web site and learn more aspects of the really sorry state of the Australian climate record.
See how Australian data adjustment fails by a country mile to support the Zeke Hausfather mantra often stated for the USA (even in these WUWT comments) that adjustment has cooled the arming trend in raw data.
Geoff S
http://www.waclimate.net/acorn2/index.html
Nick Stokes, Griff, AlanJ, Simon, et al, you must be high temperature deniers.
Not logical thinkers.
Reality deniers; unable to distinguish fantasy from reality.
Trendologists, all.
In this NOAA experiment, reported on WUWT, siting, as distinct from UHI, was tested in consideration of the claim that conditions at many surface stations produce artificially high temperature readings. I remember a comment from Nick Stokes that this proved that siting problems were much more important than UHI. I can’t find the article but what I remember is that a number of measurement stations were placed at various distances from some human heat source (e.g. the output of a building air conditioner).
It was reported that the temperature readings were affected by as much as 0.62̊C. I believe the test sites were in a starlight line out to at least 1000 meters. It was not clear to me where that 0.62̊C increase was measured. This is considerably more (in the ‘it is getting so terribly hot’ world) than the numbers reported in the Oak Ridge experiment cited in this article.
UHI.
Those seeking references and a discussion might try this from 2018/12/20:
The Climate Sciences Use Of The Urban Heat Island Effect Is Pathetic And Misleading – Watts Up With That?
Good to see Nick defending the indefensible because it means the comments are hitting home..
What’s also quite weird: the adjusted data (red line) should be somewhere between the yellow (too hot, non compliant) and the blue line (compliant) stations. But the red line is above the yellow line. So there’s also “bias” in the “adjustment”.
One reason for the adjustments, after the start of building lots of airports, is to compensate for station moves from downtowns to airports. A more recent reason is the predecessor to ASOS (HO-83) having a warm bias in comparison to ASOS.
I noticed that the graph here is from 1979 to 2008. I found a graph of raw and adjusted with a much earlier start date and an end date of 2013, along with a comparison of USHCN and USCRN that showed USCRN warming at least as much as USHCN. https://climatefeedback.org/claimreview/claim-of-no-us-warming-since-2005-is-directly-contradicted-by-the-data-it-is-based-on/ If they have access to USCRN, shouldn’t the author of this WUWT article be able to show a graph of USCRN in comparison to USHCN?
Indeed so. NOAA shows that graph here (with ClimDiv, the update to USHCN). ClimDiv is compiled from these supposedly corrupted stations, but agrees with USCRN almost exactly.
Again, why bother going through the Bureau of Adjustors if it makes no difference?
Nick Stokes
“. ClimDiv [USHCN]is compiled from these supposedly corrupted stations, but agrees with USCRN almost exactly.”
–
Comparing one rotten apple with another rotten apple should show almost exact agreement.
The stations are adjusted to agree with each other and both lots use the same corrupted data.
In reality temperature variations between sites close in distance are quite different because of terrain and elevation.
Once a one size fits all approach is allowed the data must agree.
Nick shows this at his site where data convergence in the short term all matches but going back any reasonable distance in the past shows data set difference.
“The stations are adjusted to agree with each other and both lots use the same corrupted data.”
Things like this so often said here with absolutely no evidence or attempt at justification. But in fact it would be impossible to make such an adjustment. These are all modern AWS with data reported and posted almost immediately. There is no way that you could adjust those readings so that at some future time nationwide averages of independent systems would agree with each other.
So what about the calculation process? As mentioned above, it is perfectly possible for the ordinary citizen to compute the averages from those station figures. I do that, eg here. And the agreement is just as good as the NOAA displays:
Nick,
Can you please show the same data in actual degrees C, as measured in the first instance? Not in anomaly form. Geoff S
Geoff,
“as measured in the first instance”
It wasn’t measured in the first instance. It is a calculated spatial average. And as I have explained many times, before averaging you must subtract the mean.
How do you spatially average stations on different sides of a mountain? How do you spatially average stations on different sides of a river valley?
The temperature variances in both cases are likely to be different so what does subtracting a mean buy you?
It takes out the major source of inhomogeneity. Different variance does not bias the spatial mean.
“It takes out the major source of inhomogeneity. Different variance does not bias the spatial mean.”
It certainty impacts the variance associated with that mean! Thus it impacts the statistical uncertainty of that mean Variances add when you combine independent, random variables. And each temperature measurement used to calculate that spatial mean is an independent, random variable whose variance is its measurement uncertainty. Thus the uncertainty of the spatial mean grows as you combine more neighbors!
Why do climate people always want to assume that temperature measurements have no measurement uncertainty? Why do they assume the stated value of each temp is 100% accurate?
“It certainty impacts the variance associated with that mean!“
Certainly not. Variance is just the expected value of sum of squares about the mean. You already subtract the mean in calculating variance.
Nick,
But you must not then vanish the uncertainty of the mean, you have to carry it forwards. Geoff S
You said different variances apply. Now you say they don’t?
You seem to be saying what you need to say in the moment, consistency be damned.
Nick,
Come on. How do you change a mean and not change the distribution in some form or fashion? If the distribution changes, the variance changes.
The only way to change the mean and maintain the variance is to move the whole distribution up or down. That means changing all the values in that distribution. Guess what, that’s why people say you are simply playing a numbers game to get the output you want.
“But in fact it would be impossible to make such an adjustment.”
If that data is recorded in a computers memory it is easily changed. It’s no different than voting records, you need a paper record as a backup. The paper is not so easily fudged!
It is posted on web sites. In detail, then daily data, monthly data etc. I described the process in my area at WUWT here.
But of course people who make these accusations never actually look at this record.
I don’t need to look at a record. Many posts have been made here and elsewhere showing multiple changes at locations in the early temperature records. This immediately says that nothing scientific drives the changes. They either spread previous changes outward, or someone has decided a different algorithm will give a result more in line with what the programmer expects. Either way you can not give a concrete scientific answer as to multiple changes.
It boils down to someone somewhere saying here is what I THINK the temperatures should like.
Nick,
Not true. I do.
You claim that “before averaging you have to subtract the mean.”
The mean, in common practice, is the mean of observations over a time period like 30 years. Each observation has its own uncertainty. The mean has an aggregated uncertainty that has to be carried through and represented in the numbers after subtraction, so if this is done correctly, there should be no great change to uncertainty because of homogenisation. So, why do we sometimes see much smaller uncertainty bounds on homogenised data than its parent data?
Geoff S
I have yet to see where anyone deals with the issue of showing a calculated mean that far exceeds the resolution of the original measurements. This is important as is dealing with the uncertainty because of resolution of the original measuring device.
Only mathematicians would consider showing calculations out to the one thousandth place when the original measurements were only to the integer place. They have no concept of resolution, only how many digits their calculators will display.
10/100 = 0.1, or do you think we should list a result of 0?
Yes! How do you KNOW it is not zero? The values should be shown as 10 +/- u if it is a measurement. Thus you should get 0 +/- u as the result. If 10 is an absolute count of objects then how do you get 0.1 objects? In that case you must use a different statistical description.
You are the perfect example of a mathematician/statistician who never considers the real world when working with numbers. Measurements are always 100% accurate and objects can be infinitely divided.
I don’t know what your numbers are chosen to represent. Please look up Significant Digit rules. There must be 10,001 sites that explain the rules and why. If you can’t find one, let me know. Therules were created to prevent people from displaying results that exceed the information contained in the physical measurements.
If I have a sum of ten numbers that equals 33 made up of all integer measurements. I simply can not do (33 / 10 = 3.3) and say that my result is 3.3. It ISN’T done in science. It IS done in mathematics. More specifically it IS done in statistical mathematics.
The final result can not exceed the resolution of the lowest resolution measurement.
This is the best quote I have from a well known university chemistry department. No one has ever refuted it.