From Dr. Roy Spencer’s Global Warming Blog
by Roy W. Spencer, Ph. D.
As I spend more time working on a research project, the more time I have to reflect on things that others have simply assumed to be true. And in the process I sometimes have an epiphany than clarifies my thinking on a subject.
As I continue to investigate how to quantify urban heat island (UHI) effects for the purpose of determining the extent to which land surface temperature trends have been spuriously inflated by urbanization effects, there is one recurring theme I find has not been handled well in previously published papers on the subject. I’ve mentioned it before, but it’s so important, it deserves its own (brief) blog post.
It has to do with the common assumption that “urban” thermometer sites experience spurious warming over time, while “rural” sites do not.
Obviously, at any given point in time urban environments are warmer than rural environments, especially at night. And urbanization has increased around temperature monitoring sites over the last 50 to 100 years (and longer). Yet, a number of studies over the years have curiously found that urban and rural sites have very similar temperature trends. This has led investigators to conclude that temperature datasets such as the Global Historical Climate Network (GHCN), especially after “homogenization”, is largely free of spurious warming effects from urbanization.
But the conclusion is wrong…all it shows is that temperature trends between rural and urban sites are similar… not that those trends are unaffected by urbanization effects.
Instead, studies have demonstrated that the greatest rate of warming as population increases is for nearly-rural sites, not urban. The one-fourth power relationship found by Oke (1973) and others (and which I am also finding in GHCN data in the summer) means that a population density increase from 1 to 10 persons per sq. km (both “rural”) produces more warming than an urban site going from 1,000 to 1,700 persons per sq. km.
Thus, “rural” sites cannot be assumed to be immune to spurious warming from urbanization. This means that studies that have compared “rural” to “urban” temperature trends haven’t really proved anything.
The mistake people have made is to assume that just because urban locations are warmer than rural locations at any given time that they then have a much larger spurious warming impact on trends over time. That is simply not true.
Probably like 1 percent or less of all the weather stations in the world are actually good. But there’s people who believe the data NOAA that is producing with their silly homogenization technique is legitimate. If you ask me, I think they intentionally put thermometers near cities and heat sinks to produce a fast increasing temperature trend so they could say in 2023 that the models were accurate all along. How else could they have been so confident back in the late 80s to conclude that the warming being observed is mostly caused by humans? The warming was just beginning; not to mention there were several volcano eruptions further masking the trend. They couldn’t possibly know given our limited understanding of the climate. That’s what truly makes me skeptical about this whole thing.
One needs to look at each individual site be it rural or urban to see if it is being effected by say some kinda structure near by for example down wind of the prevailing winds. There is just to much variation to make generalization about either.
up wind
Hot air rises, so UHI has an effect on surrounding areas on windless days and nights.
Then on days with higher winds the sides down wind will be affected by heating as the air passes through the urban areas
On the other hand, if a station is near the edge of a city, then when it is upwind, the wind will have the affect of cooling the station.
A study was done a number of years ago that attempted to rate the impact of UHI by comparing “urban” stations against “rural” stations on windy vs. non-windy days.
Two huge problems, first they set windy as a binary. Below a certain wind speed, the day wasn’t windy. Above that wind speed, the day was windy.
The second was they didn’t check to see if the station was up wind or down wind of the city.
There were other problems, but either of the first two problems were sufficient to invalidate the study all by themselves.
Even Rural stations picked for their “remoteness” will have many cases of surrounding trees that researchers will not think about other than “if there are trees its rural”. Trees grow pretty tall over about 80 years, and reduce the view factor from the station to the cold night sky. It is more significant than a few ppm of CO2 in the air, if 15% of cold sky is replaced by ambient temperature tree leaves….from the radiative heat transfer viewpoint of a 2 meter above ground level thermometer station.
And what do we find?….multi decadal trends point to more nighttime warming than daytime warming….hmmm….indicating a temperature signal that, on average, the cold night sky is being obscured by a few percent per decade on about 1/2 of the stations…
Tall buildings or other structures can have the same impact. Especially in winter when the buildings are being heated.
It is one reason UAH is a better source than surface stations.
“It has to do with the common assumption that “urban” thermometer sites experience spurious warming over time, while “rural” sites do not.”
That leads to the basic fallacy of this thinking, that all stations are experiencing “spurious” warming. They aren’t. It actually happened, and should be measured. Later you can try to work out if it fits within some framework you are thinking about. But measure first.
The problem with UHI is that urban areas may be oversampled, giving an unrepresentative sample to average. Focusing on rural stations fixes that. It is the only thing you should try to fix.
That number is likely extremely marginal. So while you are correct that not ALL stations are experiencing spurious warming, Dr. Spencer’s point is largely correct. How many so called rural stations are completely rural (as in miles and miles away from the nearest house in some grassland or forest that a person would have to make the trouble to traverse through just to write down some numbers). It’s way more convenient just to monitor it at your house which unfortunately sacrifices the quality of the measurements unless you are making absolutely no renovations while you are living there.
USCRN is pretty rural. But then, USCRN and ClimDiv give virtually the same average anomaly.
Ah, but give USCRN sufficient time to give meaningful climate-related data; we’re not even up to 20 years yet. Anyway it just goes to show everybody that the U.S. temperatures are not increasing as UN IPCC CliSciFi models predict.
How rural is “pretty rural”?
USCRN stations are very rural, very accurate, and unaffected by any local artificial heat sources. But the fact that ClimDiv now agrees with the new (since 2005) USCRN data suggests nothing about the accuracy of the earlier ClimDiv data. We would need to see the ClimDiv data before homogenization and other “corrections” and we would need to see if those adjustment processes changed when USCRN data became available. In other words, is ClimDiv data now adjusted so it agrees with the USCRN data? The fact that the raw ClimDiv data and all the corrections to it are not published along with the adjusted ClimDiv data makes me very suspicious.
Even the USCRN data has processing mistakes. Last year they updated the anomalies based on the new climatology (30 year typical climate data) and a protracted period of no warming became a protracted period of warming. I pointed out to them that that can’t be right. I never got a response explaining how that happened but they did fix it. If I or someone else had not noticed the change, how long would it have gone on showing a false warming trend?
I’m going to go check it again.
USCRN now shows a 12 year, 2 month warming trend of 0.013 °C per century = essentially no warming for 12 years and 2 months.
“The fact that the raw ClimDiv data and all the corrections to it are not published along with the adjusted ClimDiv data makes me very suspicious.”
The ClimDiv stations are mostly GHCN V4 stations. Unadjusted data is available for them. I did my own comparison of an average using unadjusted GHCN V4 stations with USCRN here. Again the answer is, very close.
Do most or any of the stations require someone to go there to write down the numbers?
You have reminded of two other sources of error.
The first is that the initial automated stations suffered from a problem that required the cable connecting the station to the remote recording equipment to be relatively short (was it 12 meters?). This meant that the newer, automated stations were closer to buildings than the older, non-automated stations were.
The other was that the automated equipment was powered. This meant that there was electronics, inside the weather station, giving off heat.
I would assume so. There is cloud data but wouldn’t that require Wi-Fi? How can there be Wi-Fi way out there?
aren’t there phones that get a satellite signal? I dunno.
You also have to take into account the land itself. How many people have land dedicated to just monitoring the temperature? I suspect very little.
there should be if they want us to spend gazillions of dollars to fix a problem based on those monitors
The only thing that you should, nay that you must fix, is the ludicrous anti-science that pretends that burning “fossil fuels” is even a modest concern, let alone an existential threat to anybody or anything.
Anthony Watt’s pictures of compromised weather sites are not strictly urban v. rural. Putting the weather station in a blacktop parking lot next to a set of air conditioning compressors will produce spurious readings anywhere.
We in the UK can be proud.
Our “Settled Science” geniuses from the MET Office and their BBC chums go for measurements from the side of a concrete runway at Coningsby whilst Tornado jets are taking off and landing!
A new 40°C record spike lasting just 60 seconds!
What’s not to like? And no, they don’t need no steeenking investigation!!
Good enough evidence for a few more Billions to spend on Ruinable Energy!
Yes, how much variation would you get?
Even at our place in Australia, a single dwelling on 150 acres, two thermometers on two different porches, both always in the shade, this morning registered a difference of 4c.
One 18c, the other 22c.
If the house hadn’t been there, they would both have registered the same temperature.
They may have read the same but that is not an absolute. Even with the house gone they could still read differently. Temperature depends on a lot of variables having to do with the microclimates for each.
Dr Spencer, I’ve always wondered what you get when if you filtered just on windy days, and compared temp anomalies year to year. To me it’s obvious on a still day you are measuring the atmosphere in your immediate vicinity, perhaps 10’s of square feet. On a windy day you are measuring the temperature of 1000’s of acres upwind.
I recall reading on Climate Audit around 2007 or so about someone filtering out windy days as they wanted to get rid of temperatures biased by a local ‘cool pool’, a lake. That to me is exactly what you would not want to do. A windy day would probably not eliminate UHI, but I bet it would have a very significant impact. Of course filtering would reduce the sample size, but there are a lot of samples over time and they ought to give you an honest result over a long period of time. I think it would be an interesting exercise anyway.
I expect rain days would have a similar effect.
I think we are talking about two separate but related issues. One is urbanization and the other is poor siting. It makes sense that the more populated an area becomes the more it would be exposed to the urban heat island effect. However just because an area becomes more populated doesn’t necessarily mean that the weather monitoring site is compromised. On the other hand a site that originally met all site requirements but construction has taken place and compromised the site is a different issue. The compromised site could even lose population but the buildings, parking lots, exhaust fans or whatever will always effect the readings in a bad way for us.
If a site is surrounded by a city of several million, then it really doesn’t matter whether the 100 meters immediately surrounding the site is pristine or not. The site will be contaminated.
The real way to look at this is to find place that have had long time weather site and look at not only population growth or lack there of, but also energy consumption growth of said population. The question has many variables to consider. Just installing curb and gutter and storm sewers will affect temperature readings(accelerated run off and less evaporative cooling.) My home town small very little growth still should have changes since it was once surrounded and had sloughs in town. Most are now drained.
nstead, studies have demonstrated that the greatest rate of warming as population increases is for nearly-rural sites, not urban. The one-fourth power relationship found by Oke (1973) and others (and which I am also finding in GHCN data in the summer) means that a population density increase from 1 to 10 persons per sq. km (both “rural”) produces more warming than an urban site going from 1,000 to 1,700 persons per sq. km.
“nearly rural is not an objective definition, i guarrentee you went looking for
differences FIRST and then defined “nearly rural”
a hint on population. go look at every population dataset, even the ESA ones, even the facebook 5 meter data. then look at population around airports.
population data does work for UHI because people dont cause UHI.
as Oke finally realized, the units are dimensionally wrong
nstead, studies have demonstrated that the greatest rate of warming as population increases is for nearly-rural sites, not urban. The one-fourth power relationship found by Oke (1973) and others (and which I am also finding in GHCN data in the summer) means that a population density increase from 1 to 10 persons per sq. km (both “rural”) produces more warming than an urban site going from 1,000 to 1,700 persons per sq. km.
GHCN station location data is not accurate enough to get good population data.
double check it against census track data
Awhile ago I reported on the BOM weather station at Giles, Western Australia.
This is a unique situation , the station being established during the 1950s to monitor weather as part of the nuclear testing program in central Australia.
It is still a manned station despite its extremely remote situation in the middle of the Australian continent and being isolated from any development, unlike almost every other weather station thus unaffected by UHI.
It thus is, arguably, effectively a controlled experiment where the only change over the nearly 70 years of existence is the universal rise in CO2 levels.
This unique and valuable record shows no significant rise in T min and no reduction in T max-Tmin. Therefore no response to the enhanced?greenhouse effect.
This is only one site but its unique credentials surely make it worthy of more detailed study.
Cherry picking is a constant issue in climate science due to the nature of the “beast” but a comparison of this data withy other stations of similar quality would be interesting (if there are any?)
Dear RL Moore,
Although not reported on http://www.bomwatch.com.au I studied Giles data and the history of the weather station in 2016 and again in 2019, and I should probability re-visit the dataset again.
The site was originally a ‘Federal weather station” set up by the Department of Supply and referred to as the 600-mile station (600-miles down-range from Woomera), towards Port Hedland which marked the ‘end’ of the rocket range.
I’m not exactly sure, but it seems it was transferred to the Bureau in 1972. Plans at the National Archives of Australia showed the Stevenson screen was close to a building or workshop (which were fibro portables). I had a photo which vaguely shows the outline of a Stevenson screen, but it is out of focus and due to lack of detail I could not relate it to a particular position within the camp.
There is certainly reason to doubt the Bureau’s claim that “There are no known substantial moves since the site opened in 1956.”
Tmax stepped-up 0.61 degC in in 1979, which is consistent with a change in exposure (not a change in the climate), and each side of the step, there was no detectable trend. The next step-change in 2013 followed sealing of the Warakurna Road which runs past about 25m from the current Stevenson screen. The road was sealed after September 2011 and before October 2012.
Allowing for step-changes and rainfall, there is no residual trend attributable to the climate.
All the best,
Dr Bill Johnston
http://www.bomwatch.com.au
so..resurfacing a parking lot produces less rise in temperature than turning a cornfield into one? got it.
😀
hmmm…. got me thinking- if a new surface is added to the old one- not replacing it- so the pavement is now much thicker- I wonder if it’ll retain more heat and thus effect a nearby station? Such a consideration is probably irrelevant to the discussion but the thought came to me.
That reminds me of a street near campus when I was in college.
It had been repaved so many times that the surface of the road was about 3 inches higher then the tops of the man hole covers.
Eventually the complaints got bad enough that the city decided to do something about it.
They removed the man holes covers and the device they were seated on. They added a row or two of bricks to top of the chimney, then replaced the man hole.
It was now about a quarter inch above the top of the road surface, but a little bit of asphalt ramping up to the cover took care of that problem.
Hohe Warte in Vienna is one of the longest urban temperature records and dates back to the 18th century. Then Vienna was already a big city and surrounding buildings did not change. So the change in buildings makes a uhi trend.
I think that is only partially correct, Hans. There are also larger scale effects. As others have commented, there are advection effects and inversion effects as the extent of the urban area expands. Even if the instrument perfectly complies with siting standards and the immediate area remains apparently unchanged for several centuries, the areas upwind and even immediately adjacent certainly are using more energy today than in the 18th century. There is almost certainly more pavement, faster runoff to storm drains, less vegetation, etc. in the surrounding areas that affect the measurement site.
This is a trend I’ve noticed amongst so called “climate scientists”.
That being, their inability to handle complex matters. Instead of handling the concept of urban heating as the complex subject that it is, they simplify it to a binary, present or not present.
Instead of handling clouds as the complex subject that it is, they declare that simplified parameterization is sufficient to model highly complex cloud structures.
This habit of reducing complex material to overly simplified models and then convincing yourself that you actually understand the subject is same thing that the left does with complex things like economic matters, as well as all manner of social interactions.
Is poverty a problem? Then obviously the solution is to give poor people money. Problem solved, move on to the next one.
Violence is a problem? Outlaw guns. Problem solved, move on to the next one.
Have your solutions actually reduced the problems? Who cares, the models say they should, therefore they did. If you doubt that you are just a hater who wants poor people to suffer.
They can’t even understand the difference between the variation of experimental results and measurement uncertainty.
Experimental result variation: Even if you had infinitely precise and absolutely accurate measurement devices giving absolute true values the experimental results over multiple experiments can vary. E.g. chemical solutions not absolutely homogenous, quantum probabilities, non-identical genetics, ever so slightly different environmental conditions, etc. Experimental result variation is handled by using confidence intervals to describe the absolutely accurate and precise measurements.
This is not measurement uncertainty. Measurement uncertainties stem from measurement devices that are *NOT* absolutely accurate or infinitely precise. Therefore the true value is unknown. Where it might lie is the measurement uncertainty interval.
To a statistician these are exactly the same thing since they are described by the same statistical descriptors. Thus you can assume measurement uncertainty doesn’t exist and get the resulting uncertainty by analyzing stated values since they are considered to be absolutely accurate and infinitely precise.
Thus the inconsistency of statistical world and the real world.
Who would down vote that comment?
Two of the resident trolls have already stuck their heads out from under the bridge. Picking either would give one about a 50% chance of being right.
The UHI effect will be the build up that occurs very near a site. Doubling the population a km away might barely register.
Sydney Obs had a big increase in maximum temperatures of a degree between 1890 and 1920. The passing of the approach to the Sydney Harbour bridge being built very close to it did not result in a warming trend (slight cooling). The Cahill expressway curling around it post 1950 seems have done little but a hotel built just east corresponds to a third of a degree hotter after 1970.
There is a large warming trend after 2000 but I suspect that is dubious behaviour rather than UHI.
Well yes, it’s change in the site’s location which is important, which is why sites that have been urbanized for a long period show a uhi trend during the urbanisation period and then tail off.
Rural sites can be affected by one building built close by, or by the building of access roads.
This is a complete train wreck and its hasn’t a hope in he11 of going anywhere
EVERYBODY is confusing Heat (Energy) with Temperature
They are NOT the same thing
Urban Heat Islands are Urban Temperature Islands
They are actually deficient in Heat.
They are also:
So when comparing Urban with Rural – all those things have to be taken into account.
Even a simple as keeping well away from trees when ‘going rural’
If you were constructing a shelter-belt to protect from the wind, the general rule is that you obtain downwind shelter at a range of ten times the height of your shelter (trees, wall, fence whatever) Also to a range of twice its height upwind.
Any shelter will cause your thermometer to read high
And if you want to account for turbulence in the wind having an effect, you need to be 100 times the height away
The Very Real Problem is = Your bowl of breakfast cereal, lunchtime sandwiches and evening pizza, fries & ketchup. With beer.
Those things came about because of a guy operating a tractor with a plough attached
Ask him why he does that and you will be told frankly and in all honesty:
“To dry out the soil and warm it up so that his wheat, corn, barley & potato seeds grow better and faster”
IOW: he is creating an urban cityscape = low albedo and dry
Thus, the places you can put a Pristine Climate Station become rather limited – unless you it in a desert, a forest or a lake.
And tractors with ploughs are NOT = Natural Variation
They only significantly came into being following WW2.
In parallel with both the CO2 and temperature graphs as they went skyward.
If you have ever walked in the ‘canyons’ of New York City, you have experienced the dirt and litter being blown around and into your face.
When I last looked the GHCN uses 10,000 population as an upper limit for their Rural group.
See the obvious UHI response in villages way under that population.
Torok, S.J., Morris, C.J.G., Skinner, C. and Plummer, N. 2001. Urban heat island features of southeast Australian towns. Australian Meteorological Magazine 50: 1-13.
http://www.warwickhughes.com/climate/seozuhi.htm
I have scanned the pages at above link.
A temperature transect through the small centres of Camperdown pop 3315 and Cobden pop 1477 is shown on in Fig 3(b) on page 7.
Here is an experiment I carried out in 2010.
Two degrees C Urban Heat Island in small village of Barmedman, NSW, Australia 25May2020
http://www.warwickhughes.com/blog/?p=575
It really doesn’t matter, because the temperatures that control the climate are those above the oceans, that store 99.9% of the Earth’s surface heat and cover 70% of the surface to an average depth of 3,700 metres, etc., >10^24 Joules per deg of total warming of all this controlling ocean, most in the SH anyway, all of which was badly and sparsely instrumented in the legacy record pre 1979, but is now accurately, densely and consistently measured by the common instruments of Polar satellite constellation. Temperatures in some benighted extreme spot or UHO, in the US or elsewhere, are not 30 year global averages dominated by the atmospheric temps at sea. Are they? a tiny bit of land like the USA is a parochial and trivial consideration on a global oceanic scale, modified to create extremes by the low heat capacity, so very sensitive to conditions and with a land surface that is a flat channel between the Tropical GoM and the Canadian shield in the middle, where activists don’t live, etc. They get the oceanic Hurricanes, the middle gets the Tornadoes, sorta, and the Arctic blows into Texas, etc. LOcal weather, not global climate cahnge.
A sense of reality and proportion is required. And the ability to read and trust data that has not been “corrected”, homogenised, trendified, pastyoureyes by NOAA/NASA and the original data then deleted by their Ministry of Climate Truth. Trust your satellites Luke, ignore the forgers. Stronger things happen at sea. etc. The truth is not hard to find, if you stop allowing the narrative ro be controlled by last century’s legacy science arguments. Now of little merit scientifically. It didn’t happen as predicted because the models were designed to prove CO2 was a problem, not understand the climate. So were never deterministic, AKA real, science.
Terrestrial measurement of the oceanic atmosphere – forgettaboutit, as they say in parts of Noo Joisy. We can do it so much better now.
Science moved on, climate change science still repeat its consensual beliefs and argues about legacy guesswork”science”, which does not happen as advertised since we started satellite measurements in 1979, along with many other measurable claims. Also see POlar bears, ice extents, GBR Coral. All fine. Nothing to see there either. S’obvious. How hard can it be to understand?
Wake up and smell the Bullshit. CO2s makes plants grow, and animals, at 400ppm more is good, up to over 1,000ppm when primates evolved 140Ma BP, for sure. If there is a small warming effect caused by humans, not natural solar wind driven cycles, that the natural feedback does not fully reverse, slowing the very well observed neo glacial cooling, great. We in the North will need time to get out from under the coming ice sheet. FFS.
But it all happens SO SLOWLY humans can easily move. It’s the nation state borders that will need redesigning (by the UN?!*!) that I see as the real adaption problem, in a few thousand years when it’s getting colder…. now THAT’s global climate change. No tropics even.
Haven’t I seen an Australian website about Met stations within a thousand kilometres of Alice which showed no warming for a century? Couldn’t have imagined it
As others have pointed out, the ground based sensor network was never designed to measure climatic changes, and isn’t fit for that purpose.
1) UHI contamination, and it’s growth over time. That is, even if we could accurately calculate what the impact of human activities is on a station today, figuring out what that impact was decades, or centuries ago is impossible.
2) Undocumented site changes, even something as seemingly trivial as the growth of trees 100 feet away from the sensor makes a difference.
3) Undocumented equipment changes. There’s the obvious swapping of one sensor for another, whether it be replacing one mercury thermometer with another or something more dramatic such as replacing a mercury thermometer with an electronic one. How do you disentangle the differences between two sensors, from changes to the local micro climate, from changes in the global climate.
4) Documented and undocumented location changes.
5) Changes in personnel. Both in level of training and level of dedication. Back when readings were taken down by hand, how do you know that the person actually went out every day and took readings and how often, when it was cold and wet, did they decide just to repeat yesterday’s reading. When using a mercury or alcohol thermometer, you have to get your eye exactly level with the liquid in the tube before you can get an accurate reading. What happens when there is a new person, who’s height is not the same as the previous person’s?
The glass thermometers only recorded the daily high and low, and then the readings were rounded to the nearest degree. Also because they had a lot of mass, they respond slowly to any changes in temperature. The newer, electronic equipment, the sensor is much lighter in weight, so it responds more quickly to a change in temperature. A quick change in temperature that would have been missed before, is now recorded. Also there is the change from getting two daily recordings, high and low, to getting multiple readings during the day.
Beyond that, there’s a global issue regarding the gross lack of coverage. 200 years ago, almost all of the sensors that did exist were in the eastern part of the US and Canada, and is western Europe. Even northern and southern Europe were only sparsely covered. The total number of sensors outside of N. America and Europe was in the single digits.
Even today the number of sensors available is still grossly inadequate for the job of measuring “global climate change”. Way too few and the few that are available aren’t spread evenly about the globe.
Trying to take this data and using it to determine if there has been any change to the Earth’s climate over the last 200 years is a fools errand. Trying to claim that you can determine this change to within 0.01C is scientific fraud.
Here is a link to an error riddled shambles by the great and the good.
Easterling, D.R. et al., 1997, “Maximum and minimum temperature trends for the globe”, Science, 277, 364-367.
https://www.warwickhughes.com/climate/easterling.htm
Putting aside for a moment the nonsense of Easterling et al claiming that stations from populations up to 50,000 are “non-urban”, the GHCN station / population is seriously flawed in that there is systemic understating of populations right through the GHCN station inventory file.
Easterling et al is condemned by deficient methodology, deficient datasets, straw clutching, excursions off into irrelevancies, pathetic attempts at salvage, in the end with none of its creators paying attention it features the most astonishingly error ridden colour plate (Figure 2) to appear in a modern climate Journal.