Reposted from Dr. Roy Spencer’s blog
January 29th, 2021 by Roy W. Spencer, Ph. D.
Several lines of evidence suggest observed warming trends are not nearly as large as what you have been told.
It’s been almost eight years since I posted results on my analysis of the global Integrated Surface Database (ISD) temperature data. Despite finding evidence that urbanization effects on temperature measurements have not been removed from official land temperature datasets, I still refer people to the official products (e.g. from NOAA GHCN, HadCRUT, etc.). This is because I never published any results from my analysis.
But I’ve started thinking again about the question, Just how much warming has there been in recent decades (say, the last 50 years)? The climate models suggest that this should have been the period of most rapid warming, due to ever-increasing atmospheric CO2 combined with a reduction in aerosol pollution. Since those models are the basis for proposed changes in energy policy, it is important that the observations to which they are compared be trustworthy.
A Review of the Diagnosed Urban Heat Island Effect
The official datasets of land surface temperature are (we are told) already adjusted for Urban Heat Island (UHI) effects. But as far as I know, it has never been demonstrated that the spurious warming from urban effects have been removed. Making temperature trends be the same independent of urbanization does NOT mean urban warming effects have been removed. It could be that spurious warming has simply been spread around to the non-affected stations.
Back in 2010 I quantified the Urban Heat Island (UHI) effect, based upon the difference in absolute temperatures between closely-spaced neighboring stations having different population densities (PD). The ISD temperature data are not max/min (as in GHCN), but data taken hourly, with the longest-record stations reporting at just the 6-hourly synoptic times (00, 06, 12, 18 UTC). Because there were many more stations added to the global dataset in 1973, all of my analyses started then.
By using many station pairs from low to high population densities, I constructed the cumulative UHI effect as a function of population density. Here are the results from global data in the year 2000:
As can be seen, the largest warming effect with a change in population density occurs at the lowest population densities (not a new finding), with the most total warming at the highest population densities.
The Effect of Population Density on U.S. Station Temperature Trends
In 2012 I experimented with methods to removed the observed UHI effect in the raw ISD 6-hourly data using population density as a proxy. As you can see in the second of the two graphs below, the highest population density stations had ~0.25 C/decade warming trend, with a reduced warming trend as population density was reduced:
Significantly, extrapolating to zero population density would give essentially no warming in the United States during 1973-2011. As we shall see (below) official temperature datasets say this period had a substantial warming trend, consistent with the warming in the highest population density locations.
How can one explain this result other than, at least for the period 1973-2011, (1) spurious warming occurred at the higher population density stations, and (2) the evidence supports essentially no warming if there were no people (zero population density) to modify the microclimate around thermometer sites?
I am not claiming there has been no global warming (whatever the cause). I am claiming that there is evidence of spurious warming in thermometer data which must be removed.
Next, we will examine how well that effect has been removed.
How Does this Compare to the ‘Official’ Temperature Trends?
Since I performed these analyses almost 10 years ago, the ‘official’ temperature datasets have been adjusted several times. For the same period I analyzed 8-10 years years ago, look at how some of these datasets have increased the temperature trends (I used only CRUTem3 back then):
The CRUTem3 data produce a trend reasonably close to the raw, unadjusted 6-hourly ISD-based data (the correlation of the two datasets’ monthly anomaly time series was 0.994). Note that the latest USHCN data in the above graph has the most warming, at +0.26 C/decade.
Note that this is about the same as the trend I get with the stations having the highest (rather than lowest) population density. Anthony Watts reported qualitatively similar results using different data back in 2015.
How in the world can the warming result from NOAA be reconciled with the (possible zero warming) results in Fig. 2? NOAA uses a complex homogenization procedure to make its adjustments, but it seems to me the the results in Fig. 2 suggest that their procedures might be causing spurious warming trends in the data. I am not the first to point this out; others have made the same claims over the years. I am simply showing additional quantitative evidence.
I don’t see how it can be a change in instrumentation, since both rural and urban stations changed over the decades from liquid-in-glass thermometers in Stevenson screens, to digital thermistors in small hygrothermometer enclosures, to the new automated ASOS measurement systems.
Conclusion
It seems to me that there remains considerable uncertainty in just how much the U.S. has warmed in recent decades, even among the established, official, ‘homogenized’ datasets. This has a direct impact on the “validation” of climate models relied upon by the new Biden Administration for establishing energy policy.
I would not be surprised if such problems exist in global land temperature datasets in addition to the U.S.
I’m not claiming I know how much it has (or hasn’t) warmed. Instead, I’m saying I am still very suspicious of existing official land temperature datasets.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
This is ther sort of article that should be posted to Everything Climate.
The US Climate Reference Network sites have (mostly) been in operation for 15 years now. In my state, the three locations are very remote; no UHI problem. Beyond my skill set to analyze their results, so please look at these questions. Has there been any significant trend with what they are reporting? Has there ever been a comparison of USCRN sites vs. nearby urban GHCN sites?
GHCN is now “controlled” by USCRN sites..
They cannot let their USCHN or ClimDiv fabrications warm at a different rate than the USCRN sites..
ClimDiv is forced to match USCRN, which it does very closely
The USA was somewhat affected by the 2015/2016 El Nino, which shows up as a bulge toward the right side of the graph below..
This causes a slight positive trend in the linear calculation.
(AGW apostles HAVE to use El Ninos to show warming, its all they have)
But as you can see, that bulge has all but gone, and USA temperatures are basically back down to the zero trend of before the 2015 El Nino .
USCRN has basically caused warming in the USA to CEASE.
That is what generally happens when things are measured properly.
Thanks, fred.
Yes, very informative. It looks like the 0.4 deg C / decade rise is fitted to UAH, USCRN looking more like 0.6
If you look carefully, you will see the USCRN surface temperature is responding more, in either direction, than the UAH data
As we progress further with the now cooling trend, both linears trend will come closer together, meeting at zero
It is ONLY the response to the El Nino that is causing that slight trend.
Yes, this is true. Surface data shows a lot more monthly variation than satellite data for the US. I’d guess this is because surface data is measuring just the land, whereas satellites are measuring a less exact area in the troposphere. Hence, temperature changes are diluted by temperatures from the oceans.
Spuyten Duyvil
” Has there ever been a comparison of USCRN sites vs. nearby urban GHCN sites? ”
Oh yes.
Two years ago, there was a heavy discussion here about a weather station at Anchorage International Airport
USW00026451 61.1689 -150.0278 36.6 AK ANCHORAGE INTL AP 70273
suspected of measuring 2 °C more than some stations nearby.
That motivated me to search, around that station, for a so-called pristine station, e.g. one belonging to the USCRN station set. Here it is:
USW00026563 60.7236 -150.4483 86.0 AK KENAI 29 ENE CRN 70342
Indeed: Kenai’s absolute data is at least 2 °C below that of Anchorage Intl AP.
No wonder when we look where it is:
https://tinyurl.com/yxtb8b54
But… so does their comparison look like:
https://drive.google.com/file/d/1OhCuDiAFUT80Ws4S8XopciaWQTp4rorn/view
That is the difference between looking at absolute data and looking at the departures from the mean of a common reference period.
*
A surprisingly similar result is obtained when averaging
https://drive.google.com/file/d/1yeAaITxUWgH2LpIwbvMwGWstVLMdzRhA/view
Linear estimates in °C / decade, for Jan 2004 – Jun 2019:
Buona notte, it’s a bit late here at UTC+1…
J.-P. D.
Again we see the strong 2015/2016 El Nino being the stand-out feature.
Note also that temperatures are back down to what they were before that El Nino event
Again we see that the AGW stall-warts HAVE TO USE THAT EL NINO to pretend there is a trend.
fred250
” Again we see that the AGW stall-warts HAVE TO USE THAT EL NINO to pretend there is a trend. ”
Are you able to do anything else than to insult people?
I did nothing else than answering the question
” Has there ever been a comparison of USCRN sites vs. nearby urban GHCN sites? ”
I’m not interested in showing any trend in whichever direction.
Jesus are you an aggressive, stubborn person!
J.-P. D.
Poor victim…. or is it LIAR… ?.
.
You calculated linear trends at the bottom, across El Nino events
If you want just show the comparsion, why did you add those in.
As I said…….
USCRN has brought USA temps under control.
Only El Nino event warming
ABSOLUTELY ZERO Human CO2 warming
And you are wimpy little climate apologist !
Should I send you a pack of tissues ???
After he fed the bellworm troll – or was it the other way around – the worm rang the bell over and over again it’s plausible to become somewhat upset. On the other hand it would be fine if such arguments would stop after the third exchange. Think also of the lurkers.
I may have let my personal feeling show through in some of my posts recently, and for that I apologize. But I must object to the suggestion that I’m resonsible for fred250’s bad manners.
I try to be civil here, and try to stick to a specific point, yet in the case of fred250 I get immediately hit with multiple comments none of which are relevant to the point I was trying to make. They also often contain tiresome name calling, but I don’;t care about that as it’s only addressing my silly pseudonym. But I do feel the need to respond to his points, if only to avoid the impression that I think he has a point.
You deliberately post DIS-INFORMATION as an AGW troll.
Then can’t take being corrected
Stop trying to pretend otherwise.
its petty and childish to pull out the whinging “SJW victim response”
You remain POINTLESS and EMPTY
I have to make several posts to correct your multiple cases of continued nonsense.
Then spell out my DIS-INFORMATION. All I did was to quote the exact trends for UAH and BEST, and draw what seems like an obvious conclusion, that is contradicts the claim that most of the warming in BEST was caused by UHI.
I still don’t know if you agree with Dr Spencer or not.
BEST does what the scammers at BEST want it to do.
YOU made a comment about UAH trend, and I explained where it came from, the 2015/16 El Nino bulge, and that other than that, there has been no warming in the USA.
You STILL REFUSE to accept that fact, because you AGW mind-washing will not allow you to.
In fact, over the longer term, USA is COOLING, despite urban effects.
Isn’t USHCN great.
Shows up El Ninos events as the ONLY warming
And has stopped all the spurious “adjustments” that are in prior US data
Absolutely ZERO HUMAN CO2 warming signal over the USA
Thanks for that,
It shows that UHA is VALIDATED against unadjusted, un-urban-affected, reliable surface temperatures.
Hence anything that deviates wildly from UAH is a load of CROCK
ie You have just PROVEN that GISS et al….
…. Global Surface Temperature fabrications are a load of garbage. 🙂
a load of CROCK ?
Usually you would “a crock of …. ” not a something of crock.
Crock is derived from crockery : the material of the chamber pot once used for night-time emergencies. So a crock of … is a chamber pot full of ….
is a chamber pot full of … BS !
How did they manage to teach a bull how to shiite in a pot?
If GISS was printed on toilet paper, it would have a purpose. !
Once you are aware that Earth’s energy balance is set by two thermostatic responses; one for cooling at -2C and one for heating at 30C, then you know that any reported temperature trend that is not zero over decades to centuries is a measurement error.
Spencer and Christy still have to remove the bias in the UAH lower troposphere record if it is meant to have any meaning with regard to surface temperature.
So the answer to the question is a big NO; it is not “largely” spurious it it is simply WRONG if the trend for four or five decades is not zero.
There are trends of annual, decadal, centennial and millennial periods. There is no reason that a trend of four of five decades has to be zero. And what is this “bias in the UAH lower troposphere record?”
Compare the UAH over the NIno34 region with the data from the moored buoys. The buoys show now trend over the satellite era.
The Earth’s surface temperature is thermostatically controlled meaning there is every reason for there to bo zero trend over a number of decades. If a surface measurement shows a trend, there is a measurement system error.
Anyone claiming they can detect a surface trend over the last century is in la-la land. The records a century ago were so sparse they are meaningless from a global perspective.
UAH lower troposphere is a contrived number bearing little relationship with surface temperature. Have a look at the actual data rather than the anomaly. It is measuring something high in the atmosphere that has little relationship with the surface temperature. When it replicates the data from the tropical moored buoys I will give it some credibility.
I’m sorry, Rick, but you have taken a zero trend over a particular, small ocean area and use it to justify your assertion that any positive global temperature trend estimates must be due to faulty measurements. Additionally, you have not shown the UAH6 temperature estimates over the NINO 3.4 region.
Don’t be sorry Dave. Sad to be like that. Be proud of the fact that you are prepared to question my knowledge.
Here is the UAH anomaly for Nino34:
http://climexp.knmi.nl/data/itlt_60_-120–170E_-5-5N_n.png
KNMI do not publish the UAH values but they RSS. Here is RSS for the Nino34:
http://climexp.knmi.nl/data/irss_tlt_120-170E_5–5N_n.png
As you can see it is a meaningless number as far as the surface temperature is concerned. The UAH data presented each month are not meaning with regard to the surface temperature.
Uh, Rick, the UAH data you present is for temperatures of the lower troposphere. Sea surface temperatures are determined by various instruments at different depths in the water. Two different metrics.
Anyway, the globe has been warming overall for the past 300 years or so. Prove me wrong.
Dr. Spencer,
could something be learned by comparing temperatures in populated area to a few surrounding stations on calm and windy days. I would suspect with a 30mph wind the temperature difference between the populated area and unpopulated nearby station would almost disappear. And that comparing upstream unpopulated to downstream unpopulated stations(upstream and downstream of the populated station) as wind speed changes might give a good indication of the union.
UHI not “union”
Same with Germany
Basically ZERO WARMING once UHI is properly accounted for.
Even better…
Important Climate Change Mystery Solved by Scientists – Rutgers EOAS
They managed to get rid of the Holocene Climatic Optimum…
Seasonal origin of the thermal maxima at the Holocene and the last interglacial | Nature
the abstract is so funny as it explains why they do it:
“This substantial model–data discrepancy undermines confidence in both proxy reconstructions and climate models, and inhibits a mechanistic understanding of recent climate change.” which translates to the data were a pain so we got rid of it so our theory now works all the time even when it did not.
Ironically when it comes to the Cult of Climate Change there is more uncertainty about the historic data than there is about the future data. We know that the past data needs to be adjusted otherwise people will realise that it was hotter almost 100 years ago. In Australia we have lost our Guinness Book of Records world record heat records for Marble Bar, due to random adjustments made with no rational explanation. However going forward we know with a certainty that despite what you experience and feel that the current year will be the hottest Summer on record ( even though the data would suggest it’s been the coolest for as long as I can remember).
The BOM already had the headlines ready for the two above average days that Sydney experienced last week. Saying heat waves are becoming a major factor in Sydney and it will become unliveable. I think calling two days of above average temperature ( ie above 30) a heat wave. I remember when I was at school in the 1960s and 70s when you’d get a week of temperatures in the high 30s and 40s. Now that was a heat wave ! Usually followed by thunderstorms.
it’s accepted pretty much by anyone who has followed the global warming mantra that the messaging surrounding this issue had to become climate change because even if there was some warming it was nowhere near what they predicted.
At one stage Tony Abbot was going to audit the BOM but chickened out and since that near miss it has been pretty much out of control in making adjustments to fit in to their narrative.
The only thing I really agree with is that most of the warming is man made. Unfortunately by man made I mean that some man made up these adjustments which has been responsible for most of the warming that has occurred.
There was a review of BoM and it showed all the past temperatures had been cooled.
The BoM are now undertaking a quality control system of their data – one of the recommendations from the review. There are now individual histories for each site. These are quite a rich source of information if you want to find the cause of an apparent step change in reading. Usually associated with installing electronic temperature monitoring replacing the LiG thermometers.
I had reason to investigate changes at Cape Moreton recently and found this data:
http://www.bom.gov.au/clim_data/cdio/metadata/pdf/siteinfo/IDCJMD0040.040043.SiteInfo.pdf
On page 14 you can see they moved to electronic temperature in August 1995. Soon after that, the data shows a step change:
http://www.bom.gov.au/jsp/ncc/cdio/weatherData/av?p_display_type=dataGraph&p_stn_num=040043&p_nccObsCode=36&p_month=13
Electronic transducers have been a great help in BoM measuring Global Warming.
RickWill
” On page 14 you can see they moved to electronic temperature in August 1995. Soon after that, the data shows a step change… ”
*
I don’t want to spend too much time on looking for data for BoM stations as near as possible near Moreton. Most don’t have the life span needed for a comparison.
Therefore, I looked where I feel at home: in the raw GHCN daily stuff.
Though having the same ids, the GHCN stations there have life spans differing from those in BoM’s database, but anybody might spend some time in finding similar BoM data, and redo my little exercise.
Here is a comparison of Moreton with Brisbane Aero (which moreover seems to be the synthesis of 2 BoM stations):
ASN00040043 -27.0314 153.4661 99.9 CAPE MORETON LIGHTHOUSE 94594
ASN00040223 -27.4178 153.1142 4.0 BRISBANE AERO 94578
Yearly absolute data for 1957-2019
https://drive.google.com/file/d/1oD9jw27kzJIImbsTBDSyDFXmpw0QtiKG/view
Monthly absolute data for 1957-2019
https://drive.google.com/file/d/1eiYMT-urlY-t_IzUj-Aj_Mx820OkIW_D/view
Once more, an amazing similarity, as I detected many times when comparing nearby stations despite their difference in average temperature.
Thus, a step change in Moreton’s data? Hmmmh.
What I see since 1957 (Moreton’s start in GHCN daily) is
Is that really a reason to automatically suspect other people’s work?
J.-P. D.
I have no idea what is going on in figure 2 but it looks totally spurious. Dr Spencer leaves the graph up but does not comment.
Firstly, the OLS linear regression he has done on the scatter plot is meaningless. There is clearly massive uncertainty in the x variable and OLS is invalid. ( Just because Excel can fit a “trend” does indicate it has any statistical validity, it doesn’t ).
Ignore the fitted line and try to eyeball a best fit , it would be WAY steeper an obviously so. This is called regression dilution ( the actual slope was diluted by the x errors and is spuriously low ).
Now, the fact that the two regressions are damned near identical is blatantly a huge coincidence. The intercept is identical to 4s.f. and the slope only varies by 1 in 4s.f.
I find it deceptive to publish these two graphs. Any scientist knows that is a coincidence but publishing the two side by side without comment gives the layman in impression there is a very strong corroboration between the two.
Now Dr Spencer has done some great work and I respect his integrity in general. I do not understand what is going on here.
Roy Spencer – If the land temperature datasets are corrupted by UHI, then surely a discrepancy with satellite data would show up? Could you take all of the land temperature record as supposedly adjusted for UHI covering low population density areas and compare them with satellite data for the same areas.
What is the point of investigating official temperature data and finding it to be suspect? No one is going to pay attention to this work.
Following the position of the esteemed Dr. Spencer, I am not claiming there has been bias in selecting the science to drive political action (whatever the cause). I am claiming that there is suspicion of impartiality which must be removed.
IF the science is biased, THEN no amount of corrective papers will suffice to alter it. Such papers will be ignored, smeared if they are offered as evidence, and eradicated, together with the career of the authors, if they look like becoming influential in any way. That is the problem we need to deal with, and it will not be solved by a dispassionate consideration of the mathematics.
‘impartiality’ = ‘partiality’, of course. No edit is a bummer…
Let me first say that, “crap I lost a big post due to may own fault”. So I am going to make this short and sweet.
I “assume’”, and correct me if I am wrong. I assume that the satellite algorithm basically gives an integral of the entire temperature profile for the globe, i.e. a global temp.
Comparing this to a surface temp record may be erroneous. The satellite obviously can not measure the surface temperature at a fixed height, like surface stations do. There is probably a slice taken, say 50 meters and who knows how that compares. I suspect there should be a conversion factor at sometime in the future.
Surface stations – Whoa!
Problem 1 – The accuracy, precision, and uncertainty of measurements are totally ignored. How this affects the total range of temperatures is quite large and unknowable.
Problem 2 – The warmists like to assume that each station record is a “sample” of the global temperature population. By doing this they can invoke the Central Limit Theorem that insures the mean of the sample means gives an accurate calculation of the Global Average Temperature. However, ignore is the operative word again. The CLT requires Independent and Identical Distributed samples. If anyone thinks individual station records all have Identical Distributions, please show me.
Problem 3 – The homogenized, interpolated, extrapolated, infilled data is not real data and there is no guarantee that it would match any other determination of a global temperature.
A little off topic, but with the release of the the latest UAH for January, I predict that Lord Monckton’s pause will now be starting in August 2015, making it now 5 years and 6 months old. This also means August 2015 has the distinction of being both part of the old and the new pause.